600283100
MOBILE SYSTEM FOR EXTRACTING
SPILLED HAZARDOUS MATERIALS
FROM EXCAVATED SOILS
by
Robert Scholz
Joseph Milanowski
Rexnord Inc.
Milwaukee, Wisconsin 53214
Contract Number 68-03-2696
Project Officer
John E. Brugger
Oil and Hazardous Materials Spills Branch
Solid and Hazardous Waste Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08837
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the
United States Environmental Protection Agency under Contract No. 68-03-2696
to Rexnord Inc. It has been subject 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.
-------
FOREWORD
The U.S. Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimonies to the deterioration of our natural environment.
The complexity of that environment and the interplay of its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem solution;
it involves defining the problem, measuring its impact, and searching for
solutions. The Municipal Environmental Research Laboratory develops new
and improved technology and systems to prevent, treat, and manage wastewater
and solid and hazardous waste pollutant discharges from municipal and com-
munity sources, to preserve and treat public drinking water supplies, and
to minimize the adverse economic, social, health, and aesthetic effects of
pollution. This publication is one of the products of that research and
provides a most vital communications link between the researcher and the
user community.
This report describes the laboratory feasibility investigations and the
design and construction of a full-scale, mobile system for treating
excavated, contaminated soils at a field site by sieving, disintegrating
the soil matrix, and cleansing the soil of hazardous chemicals. The system
is expected to take its place among other emerging methods such as in situ
treatment or incineration, for treatment of contaminated soils at the
actual site of a release under a wide variety of circumstances.
Francis T- Mayo
Director
Municipal Environmental Research Laboratory
iii
-------
ABSTRACT
Laboratory tests were carried out with three different pollutants (phenol,
arsenic trioxide, and polychlorinated biphenyls [PCB's]) and two soils of
different character (gravel/sand/siIt/clay soil and organic loam topsoil)
to evaluate a technique for the scrubbing or cleansing of soils conta-
minated with hazardous materials. The results of the tests were used in
the design and construction of a full-scale system for field evaluation in
cleaning excavated contaminated soils. The system employs two major soil
scrubbing components: a newly designed water-knife, screen stripper, and
soaking unit for breaking up soil lumps and stripping chemicals from larger
particles (<2.5 cm but>2mm in diameter); and a countercurrent chemical
extractor for washing the smaller particles (<2 mm)that pass through the
screens. The complete soils scrubbing system also requires the use of a
wastewater treatment system to permit recycling of the washing fluids and
an air cleaner when gases are evolved. The projected processing rate of
the complete system is 2.3 to 3.8 m^/hour [3-5 yd^/hr). (Treatment
ratesrange from 11.5-13.5 m3/hr (15-18 yd3/hr) when only the water-
knife unit is needed.) Treatment residues consist of skimmings from froth
flotation, a mixture of washed gravel, sand, and organic detritus
discharged from the knife-screen unit, fine particles from the extractor,
the spent washing fluids, exhausted carbon, and the other treatment
products (floes, sediments).
A potentially limiting constraint on the treatability of soils is their
clay content, since clays generally are good adsorbers and the particles
may not be so effectively scrubbed as are sands. In particular, clay lumps
that are only surface contaminated should be removed before full
disintegration is accomplished (since the uncontaminated particles could
then function as adsorbents and lower the efficiency of the cleansing).
Contaminant-saturated consolidated clay lumps, upon disintegration, will
release pollutants, however. Most inorganic compounds, almost all water
soluble or readily oxidizable organic chemicals, and some partly
miscible-in-water organics can be treated with water or water plus an
additive.
During limited laboratory extraction tests, phenol was very efficiently
removed from both organic and inorganic soils, whereas PCB and arsenic
oxide clung more tenaciously to the soils and were released far less
readily into the washing fluids. Until acceptable limits of residual
concentrations in the washed soil are agreed upon, the utility of the soil
washer in a particular situation cannot be assessed. Laboratory tests did
show that soil scrubbing vastly speeds up the release of chemicals from
soils, compared with the much slower rate that occurs under natural
leaching conditions.
1v
-------
A full-scale soil scrubber has been constructed and tested and is ready for
further demonstration. The cleansed soil can be returned to the site from
which the contaminated soil was excavated.
This report was submitted in fulfillment of Contract No. 68-03-2696 by
Rexnord, Inc. under the sponsorship of the U.S. Environmental Protection
Agency. This report covers the period December 1976 to April 1982,
and work was completed as of November 1982.
iv
-------
CONTENTS
Foreword ill
Abstract i v
Fi gures vi
Tables vii
1. Introduction 1
2. Conclusions 10
3. Recommendations 11
4 Selection of Scrubbing Methods, Soils, and Chemicals for
Testing 12
5. Preliminary Laboratory Testing Program and Results 26
6. Implications of the Preliminary Laboratory Testing Program
on System Design and Soil Treatment 62
7. Design and Construction 68
References 82
-------
FIGURES
Number Page
1. Schematic of soil scrubbing operations
2. Schematic of the water knife used for chemical stripping and
1 ump breakdown 14
3. Schematic of counter-current chemical extraction system with
two extractions 15
4. Schematic of hydrocyclone solids/liquid separator 17
5. Inorganic soil for laboratory experiments 23
6. Organic soil for laboratory experiments : 24
7. Schematic of water knife testing apparatus 28
8. Rotary screen and water kn ife 29
9. Typical soil breakdown data for preliminary water knife testing... 30
10. Laboratory froth flotation testing apparatus 32
11. Laboratory hydrocyclone testing apparatus 35
12. Loss on ign/ition (LOI) and adsorption of phenol versus soil
gradation ,for an organic soil 48
13. Soil and solvent movement during shaker tests with solvent recycle 55
14. Soil phenol reduction as a function of number of extractions
using two different extraction methods 61
15. Process flow scheme for soil scrubber 69
16. Fully constructed rotary drum screen scrubber 72
17. Soil loading and metering system 73
18. Soak zone description 74
19. EPA Froth Flotation Unit (beach cleaner) modified as a
counter-current chemical extractor for soil scrubbing 76
20. Process flow chart of counter-current chemical extraction system.. 77
21. Cost per volume treated for scrubbing phenol from soil 81
vi
-------
TABLES
Number Page
1. Properties of Soils Selected for Preliminary Laboratory Tests 25
2. Summary of Hydrocyclone Testing Results 37
3. Hydrocyclone Test Results for Solids Removal from Overflow 39
4. Comparison of Preliminary Settling Test Results with Hydrocyclone
Overflow Total Solids Concentration 40
5. Beaker Settling Test Results 41
6. Column Dosing Test Results for Phenol 44
7. Distribution of Phenol on Organic and Inorganic Soils by
Particle Size 45
8. Distribution of Arsenic on Organic and Inorganic Soils by
Particle Size 46
9. Water Knife Test and Submerged Washing Results for Soil
Particles in the Size Range of 2mm to 12.7mm 51
10. Phenol Distribution in Soil Samples Used for Water Knife and
Submerged Washing Tests 52
11. Single Extraction Test Results 56
12. Summary of Phenol Extraction Efficiency for Counter-Current
Extractions 59
vii
-------
SECTION 1
INTRODUCTION
NEED FOR NEW TECHNOLOGY
The leaching of hazardous materials from contaminated soils into groundwater
is recognized as a potential threat to the nation's drinking water
supplies. Such situations occur as the result of accidental spills of
hazardous substances and as releases from the many uncontrolled hazardous
waste disposal sites now known to exist throughout the country. Current
corrective technology is largely limited to the excavation and transfer of
such soils to suitably sealed or lined landfills where uncontrolled leaching
cannot occur.
Onsite treatment can often be a more cost-effective solution to the problem,
and avoids the need for transporting hazardus wastes. Research projects
have demonstrated, for example, that contaminated soils can be isolated by
grouting, slurry-wall trenching, installation of steel piling, etc., and
then subjected to in-situ leaching for treatment. The effectiveness of such
processes is limited by, among other factors, the permeability of the soil
in its undisturbed state.
An alternative process is desirable for those situations in which soil
permeability or other factors prevent effective cleanup by in-situ
leaching. This proposed technology, the subject of the current effort,
consists of excavation onsite, above-ground treatment of the contaminated
soil, and return of the treated soil to its original site. Excavation of
the soil from its natural state opens a number of options for improved
separation of contaminants through better mixing and the ability to use
different solvents. Such cleanups can also be carried out more quickly than
they could with the in situ leaching of a more or less compact natural
soil. An integrated engineered approach also makes it possible, or more
convenient, to incorporate the controls needed for air emissions and/or
treatment of the contaminated wastewaters generated by the process.
In passing, one may note that, once a soil is excavated, cleanup is not
restricted to the washing approach described herein. Washing with organic
solvents, processing with fluids at or above their critical points, wet air
oxidation, incineration, and other flushing and/or disintegrative processes
may be applicable. The prospective costs associated with each approach must
be based on a specific situation; it is not realistic to assume that there
is any universal "best" approach in terms of either cost or effectiveness.
-------
The purpose of this project was to carry out appropriate laboratory studies
and then design and construct a system capable of treating a wide range of
excavated contaminated soils. Ideally, the resulting system will be useful
for the correction of long-standing contamination problems (waste disposal
sites), as well as for emergency cleanup of spills.
SELECTION OF WASHING FLUIDS
A variety of washing fluids can be used to cleanse contaminated soil,
depending on the nature and properties of the chemicals to be removed
and of the soils themselves. An effective treatment fluid or solvent must
have these characteristics:
1. Favorable separation coefficient for extraction.
2. Low volatility under ambient conditions.
3. Low toxicity.
4. Safety and ease of handling in a field situation.
5. Recoverability.
Desirable Characteristics of Fluids
Favorable Separation Coefficient—The contaminant must be soluble in the
solvent and the solvent itself easily separated from the soil. The higher
the solubility of the contaminant in the fluid, the lower the volumes
required for extraction. The solvent should be effective for a large number
of contaminants and a wide variety of mixed hazardous substances.
'
Low Volatility—The solvent should not be extremely volatile, especially
under ambient conditions, nor should it be flammable. Excessive volatility
could result in significant solvent losses into the air and in causing
hazardous situations for operating personnel.
Low Toxicity—The contaminant to be scrubbed from the soil poses a potential
exposure hazard to treatment system operators, especially during those steps
of the process that are intended to reconcentrate the contaminants, e.g., in
the processing of spent extraction fluids. A treatment fluid should not
subject treatment operators to unnecessary, additional hazards. The
quantity of treatment fluid in the treatment system at any one time will
probably be many times greater than the amount of contaminant. The exposure
hazards from contaminants and treatment fluids will chiefly consist of skin
and eye contact diseases and inhalation effects.
Safety and Ease of Handling—When a mobile system incorporates potential
ignition sources such as an internal combustion engine generator and an
electrical power system, special equipment design and operational procedures
are required to minimize the potential for fire and explosion in the
presence of flammable and/or explosive mixtures, contaminants, or treatment
fluids. The transport and handling of such fluids is of equivalent concern.
Recoverability—The spent treatment fluids must be recoverable in a
field-oriented treatment system so that cleaned fluids can be recycled and
ultimately disposed of in an uncontaminated form.
-------
Types of Cleaning Fluid
Assuming that most contaminants can be classified as either organic or
inorganic, three basic solvent systems can be identified:
1. Water only.
2. Water plus an additive.
3. Organic compounds (neat or as mixtures).
When the system uses water or water plus an additive, most inorganic
compounds, almost all water soluble or oxidizable organics, and some par-
tially miscible-in-water organics can be treated. The important steps
involve obtaining good contact (high mixing energy) between the soil and
water and using the optimized amount of additive. These additives may
include:
1. Acids to improve removals of metal salts and oxides.
2. Bases to enhance safety and removel for cyanide compounds, solubilize
phenolics, etc.
3. Oxidants to aid in organic and reducing agent treatment.
4. Surface active agents to increase solubility of organics in the
water phase.
Organic solvents allow more effective removal of organic contaminants,
which action is a crucial consideration for compounds that have low water
solubility. However, handling these materials in a field situation can be
extremely difficult. Most are volatile and many are flammable. An organic
solvent such as hexane or Freon may not be effective for all contaminants,
or may extract humectic components and harmless organics from the soil and
lead to a fouling problem with washing fluids and components. Organic
solvent mixtures present special problems in recovery and reuse. Finally,
adapting field equipment for use with an organic solvent is, at best,
extremely difficult and costly. The advantage of using organic solvents to
remove certain organic contaminants did not appear to outweigh the
significant disadvantages of organic solvents for a system of the size and
cost contemplated. For these reasons, this project concentrated on the
development of soil washing technology that employs water plus an additive
as a washing fluid.
When the unit is constrained to use water+additive washing fluids, the
removal effectiveness for organic contaminants may be poor. Soil washing
with water is most efficient for compounds with a high water solubility and
a greater affinity for the water than the soil. This situation makes the
soil scrubber most efficient with short-chain organic compounds, other
organics with a relatively high water solubility, and inorganics that are
rather soluble in water. The inorganic removal efficiency may decrease in
those instances where the compound has a strong affinity for the soil compo-
nents, especially, clays. This phenomenon may be critical for pollutants
that are chemically sorbed onto certain soil types.
Acceptable discharge levels will determine whether on site soil washing is
an appropriate treatment method for any particular contaminant. The "clean"
-------
soil contaminant levels will certainly vary with the chemical involved.
For example, treatment of'an alcohol to a parts-per-million level may be
acceptable while pesticide by-products may require treatment to the low
parts-per-billion range before environmentally safe discharge of the soil.
When evaluating the potential usefulness of a washing process, these levels
of cleanliness must be defined beforehand. It may also be practical to
scrub the soil to a contaminant concentration level that will reduce the
grade of landfill required for disposal, i.e., from a hazardous landfill to
a sanitary landfill.
Care must also be taken to avoid compounding problems when scrubbing soil.
Transfer of contaminant to the water phase and subsequent water treatment
will often result in a solid waste for eventual disposal. The residual may
be spent carbon or a treatment sludge and the volumes involved may be quite
large. Problems and costs associated with disposal of these residues
should not exceed those that would have been encountered when handling the
contaminated soil alone.
CONCEPTUAL APPROACH TO SOIL SCRUBBING
A review of available literature on soil systems, soil/chemical inter-
actions, and treatment procedures produced little directly applicable
information. It was necessary to rely on the past experience in soil
handling and treatability of the project team and on common sense in order
to establish reasonable limitations and to identify the potential
applicability of a particular technique.
A simplified schematic of the basic operations involved in soil scrubbing
is shown in Figure 1. The first step in soil washing involves excavation
of contaminated soil, which can be accomplished using various types of
construction machinery and earth-moving equipment. The removed soil must
be reduced in size and the sofl structure or "fabric" disintegrated to
allow effective contact between the solvent and soil particles.
This type of pre-conditioning requires mechanical action on a wide variety
of soil types. After size reduction, the appropriately conditioned soil is
then ready for actual treatment in the scrubber.
The treatment action involves two mechanisms:
1. Physical stripping of that portion of the contaminant that is not
tightly bound to the soil particles. The contaminant may have
filled void spaces and be physically trapped in the soil system.
Removal can be effectively accomplished by stripping the contaminant
from the soil, possibly by using high pressure fluid rinsing.
2. Extraction of the tightly bound contaminant from the soil into a
solvent. This step allows transfer of the contaminant into a phase
more easily handled using conventional treatment processes. The
extraction efficiency may be increased by applying a countercurrent
approach and solvent use can be optimized if solvent treatment and
recycle are utilized.
-------
II OS NV313
< 3
o o
— a
o e:
LU CQ
a:
I-
x
o_
oi
r
>
i
o
LU
a.
oo
-i <
O LU
in
c
O
to
i-
1}
a.
o
a>
«*,.-.
> LU
< cc:
o <
X
o
I/)
o
o
o
oo
O)
\
1
±l± I
r\ T c~» '
;
o
oo
2
o
LJ
-------
The soil scrubbing approach is quite simple; however, there are significant
difficulties involved in applying these concepts to complex situations of
soil contamination. Soil types vary considerably both in their physical
characteristics and attraction for different contaminants.
SOIL TREATABILITY POTENTIAL
The varied geologic history throughout the United States has resulted in
complex combinations of near-surface soils, as exhibited by their wide
range of chemical constituents, gradations, and organic content. These
variables affect the soil treatability potential.
The treatability potential of any soil can be considered a combination of
its physical handling and sorption/desorption characteristics. For
example, it may be technically feasible to desorb a particular contaminant
from a clay soil, but the mechanical effort and energy required to disinte-
grate the clay to a particle size that allows the contaminant to be
desorbed efficiently may result in low cost-effectiveness and low
treatability potential.
The wide variety of qualities encountered in the soil systems can adversely
affect the processing equipment. Major difficulties or limitations can
result from:
1. The gradation of the soil particles, primarily the predominance of
large objects such as rocks or boulders. Large rocks, boulders,
tree roots, and branches cannot be easily broken down by conven-
tional size reduction equipment; they should, therefore, be
presorted and disposed of separately.
2. The clay content of the soil, primarily as related to large zones of
clay or cohesive materials. The cohesive and plastic nature of most
clays, combined with their very small particle size, makes it
difficult to rapidly break them down into treatable sizes. Most
conventional size reduction systems are designed to handle only a
certain maximum percentage of clay within a soil structure. Clays
can be broken down by freezing or drying combined with impact or
crushing techniques. However, these methods are energy intensive
and, in general, are not very practical for high volume materials
handling systems. Note that the plasticity of the clay results in a
low permeability so that the potential impact of a contaminant spill
on clay is much less than on a sand or gravel soil, unless the clay
is finely divided and dispersed (large surface-to-volume ratio).
The percentage of clay will directly affect the rate of soil treat-
ment both at the pre-treatment stage and during the chemical
treatment phase itself. Although this limitation of clay within a
soil system is significant, throughout most of the United States,
use of the scrubber unit is feasible and adaptable.
3. The root system typically developed in a thick topsoil can foul many
mechanical systems designed to break down soil into more easily
-------
handled sizes. Roots tend to be flexible and resilient, and wind
around most rotation devices. A common example which is understood
^,«, by gardeners is the way in which roots tend to wind around the shaft
of a rototiller. Large tree roots or branches are also difficult to
process with conventional soil handling systems because of their
bulk and resistance to breakdown.
Desorption Characteristics
The soil/chemical interactions that affect sorption/desorption are
complex. The desorption of a contaminant from a soil is related to, though
not necessarily a direct function of, the sorptive ability. Adsorption
onto a soil surface is a phenomenon based upon physico-chemical equilibrium
conditions between the contaminant, soil water, and soil itself. The
solubility of the contaminant in water and the reactions between the
contaminant and the soil components determine the equilibrium that is
established.
Researchers have evaluated attenuation or adsorption of contaminants in
soils rather than contaminant desorption (much work has been done on
• modeling dispersion/diffusion of pesticides in soil). Although the
sorption/desorption relationship is not well-established, the generalities
and methods of adsorption may be appropriate for a basic understanding of
desorption. The existing research does suggest the following unquantified
relationship:
A = f(C, pH, Or, Co, t). The definition of the parameters follow.
*w»
"A" is the attenuation of a given contaminant by the soil in question,
expressed as percent removed from solution. Attenuation is proportional to
five parameters, the first being the clay factor "C". The clay factor, in
turn, is a function of four, interrelated factors:
1. clay content of the total soil volume
2. type of clay mineral in the soil
3. exchange capacity of the clay
4 specific surface area of the clay.
The directly proportional functional relationship between attenuation and
increasing clay content implies an inverse relationship between attenuation
and an increasing content of granular material in the soil.
The term "pH" refers to the pH of the soil/contaminant environment. This
includes both the contaminant waste stream and the natural soil pH condi-
tions. The attenuation is affected by the pH of the leaching solution and
is a function of the specific contaminant and the retention mechanism
involved in its removal from solution.
"Or" is the organic content of the soil. The organic content is most
important when evaluating organic contaminants that tend to readily adsorb
onto organic matter in the soil. Whether the organic matter is in the form
•— of living roots or dead decaying material may be important because of the
W differences in structure and composition.
-------
"Co" is the contaminant concentration factor. At low levels, the concen-
tration of the waste stream may influence the attenuating mechanism
operative in the soil system. However, the situation is somewhat different
at high contaminant concentrations. The soil, in many cases, may be
incapable of attenuating the applied dosage even when all of the factors
needed are present, simply because it does not have the required surface
area or sufficient exchange capacity.
The term "t" is a time factor. Attenuation is directly related to the
contact time between the contaminant and the soil, all other conditions
being constant. The attenuation achieved during a given contact time
should be proportional to the permeability of the soil. The permeability
of the soil is directly related to its grain-size distribution, which ties
in with its mineralogical composition.
Soil contamination situations can be grouped into two types as differ-
entiated by a time factor: immediate problems and long-term situations. A
spill occurrence is a problem which, when responded to immediately,
involves only a short contact time between soil and contaminant. Long-term
situations of soil contamination, on the other hand, result from initially
untreated spills or from improper chemical waste disposal practices.
The contaminant contact time defines the rate at which equilibrium will be
established. For spill situations in which the contaminant can be
relatively quickly contained or removed, the soil/chemical interactions may
be minimal. Conversely, in a long-term situation where leachate slowly
leaks from a landfill, the soil/chemical equilibrium reactions may be well-
defined and established.
The time effects are most prevalent in cohesive or clay-like soils. Since
the permeability and flow rate are low in a clay or silt structure, the
contaminant will require a longer time to contact the soil and to be
adsorbed onto the soil surface. With time, the amount adsorbed will
increase and the subsequent time for desorption may also increase. In a
sand or gravel soil, on the other hand, a spill will flow almost
immediately through the soil structure, contacting the soil surface and
moving quickly onward. In this latter situation, the effects of time are
less significant.
PROJECT OBJECTIVES
The purpose of this project was to develop and construct a mobile system
for demonstration of the extraction of hazardous materials from spill-
contaminated soils. The goals of the soil-scrubbing system are threefold:
1. To clean contaminated soil sufficiently to make it suitable for
redepositing into the place from which it was removed
2. To separate hazardous materials from the extracting fluids to the
extent that the fluid is non-contaminated
8
-------
3. To recover the hazardous material in such a form that it can be
handled, reprocessed, and/or disposed of by environmentally
acceptable methods.
This project did not explore technology for breakdown of large soil lumps
in any detail. Rather, the assumed inlet situation was free-flowing soil
which an entry screen would restrict to 2.5-cm (1.0-in.) particles or
smaller. The project did not address the spent washing fluid treatment
system, which will probably consist of a physical/chemical treatment system
such as EPA possesses and has already demonstrated for field use (14).
Cleaning of the air contaminants generated by the process was not
addressed, except to provide containment of these emissions with
connections for ducting to an appropriate air cleaner. The primary
emphasis of this project was on the scrubber itself, specifically on those
components that permit soil particles to be effectively washed through
physical and chemical reactions.
Prior to design and construction, alternative approaches were evaluated and
selected approaches were tested on a laboratory scale. This report dis-
cusses those evaluations and their influence on the ultimate construction
of the components for demonstration.
-------
SECTION 2
CONCLUSIONS
The following conclusions can be drawn from the work carried out during
this program and the knowledge gained during that effort:
1. Spill-contaminated soils can be excavated and treated onsite, using
extraction with water or aqueous solutions for many pollutants frequently
encountered in such situations.
2. A system capable of decontaminating 2.3 to 3.8 m3/hr (3-5 yd3/hr) of
soil has been designed and constructed and is now available for field
testing by the U.S. Environmental Protection Agency (EPA). If finer (<2mm)
are absent or do not require treatment, the processing rate (water-knife
drum screen unit) is 11.5-13.5 m3/hr (15-18 yd3/hr).
3. Water knives function as a compact and economical means of achieving
effective contact between contaminated soil particles and extractant.
4. Countercurrent extraction is effective in removing certain adsorbed
contaminants from soils and, for the size of equipment needed, hydro-
cyclones are the preferred devices for separating the extracted solids from
the extractant.
5. Laboratory experiments demonstrate that soil characteristics
(particle size, distribution, organic content, pH, ion-exchange properties,
etc.) are very important factors in the removal of contaminants.
6. In addition to the percent-contaminant removed, the allowable level
of pollutant remaining in the soil is an important consideration in
determining when adequate decontamination has been achieved, since the
final concentration affects the options available for disposal of the
residual solids.
10
-------
SECTION 3
RECOMMENDATIONS
Based on the observations made during this investigation, several
suggestions are offered for future work.
1. Laboratory screening tests should be performed on a wider range of
typical compounds and mixtures encountered in hazardous substance spill and
release situations to ensure that high levels of decontamination can be
achieved with this process. Conversely, lab tests should be performed
before the complete system is field-mobilized.
2. The results of this study apply primarily to spill situations.
Contaminated soils found at waste disposal sites may exhibit different
extraction characteristics because of the extended soil/contaminant contact
time and of weathering and in situ reactions. Studies are needed to
establish whether and to what extent such changes affect the
decontamination process.
3. Other extractant solutions and solvents should be explored to learn
whether the efficiency of the process can be improved without damaging the
equipment or increasing the hazards to which the workers may be exposed.
4. A wider range of soils should be examined to determine what changes
in the system are practical to better cleanse soils with characteristics
(e.g., greater cohesiveness and adsorptive properties of clay- or silt-rich
soils) that differ significantly from those of the soils already tested,
and to evaluate the tendency of disintegrated clay (clay particles) to
function as adsorbents.
11
-------
SECTION 4
SELECTION OF SCRUBBING METHODS,
SOILS, AND CHEMICALS FOR TESTING
As the first stage in the development of a soil scrubber, it was necessary
to make preliminary scrubbing equipment choices and to experiment with
.these methods on a laboratory scale using selected chemicals and soils.
Physical Stripping
The first step in the processing train involves physical stripping of
contaminants that are not physically or chemically adsorbed on the soil
particles. Since a significant amount of contaminant could be entrained in
soil void spaces, this effort may result in high treatment efficiencies.
However, since pretreatment devices would probably not be capable of fully
breaking down soil lumps, a secondary objective for this equipment involves
further size reduction below 2.5 cm (1.0 in.) in diameter.
Both contaminant stripping and lump breaking require a relatively high
level of energy input for the greatest efficiency. Several systems were
evaluated and all but one were eliminated from further consideration.
Commercially available rotary scrubbers were eliminated because the action
of the slowly rotating cylinder was not considered sufficiently energetic
to perform stripping and lump breakdown. The log washer, generally used in
the aggregate industry for washing clay and silt from sand or gravel, was
also evaluated. The device is basically an inclined rotating screw
conveyor with internal beater paddles that increase the cleaning action.
Unfortunately, the rotational speed of the arms is quite low, which reduces
the unit's effectiveness in breaking down clay lumps. In addition, the
equipment requires large volumes of water for proper operation, a condition
that could significantly impact its ultimate usefulness in the field.
The attrition scrubber uses high energy scrubbing forces applied by a dual
propeller mixer. The pitches of the two blades are opposed, which causes a
substantial amount of mixing turbulence. An attrition scrubber is intended
for use on sand-size particles and may not be suitable for breaking down
clay lumps. Efficient use of this device requires feed concentrations of
70-percent solids, which is a difficult constraint to meet in a natural
soil system contaminated with chemicals. Wet soil can foul the unit and
cause serious operational difficulties. A large amount of gravel in the
soil will cause undue wear and vibration on the equipment. Also considered
but rejected were: horizontal vibrating screens, fixed blade mixers,
concrete mixers, aggregate scrubbers, ball mills, jaw crushers, hammermills
and non-clog hammermills (1).
12
-------
Another high energy system uses water knives to create a powerful fluid
stream for contaminant stripping and lump breakdown. The water knives are
small, flat-blade nozzles that function by creating a high velocity stream
of water in the shape of a long narrow line (Figure 2). Washing is
performed by directing the nozzle stream at a moving layer of material.
The relative motion between nozzle and material can be achieved by:
1. Moving the material past a stationary nozzle on a conveying
mechanism.
2. Moving the nozzle over the top of a stationary layer of material.
3. Moving the material on a conveyor and moving the nozzle over the
material.
The knife-like hydraulic action is achieved by a combination of the force
and thinness of the water stream. The washing is instantaneous. The
contact of the knife with the material can be so gentle that it can clean
peaches, or it can be so "sharp" that it will strip scale layers from
steel. In the soils scrubbing application, the water knife seemed capable
of both lump breakdown and scrubbing. The volumes of water and pressures
needed for proper nozzle operation can be estimated from design parameters;
however, the actual effectiveness of the scrubbing technique for treatment
must be determined from laboratory testing.
Extraction and Separation
Small soil particles will most likely be more affected by chemically bound
contaminants than large particles. As a result, the physical stripping
process alone may not be effective for treatment of small particles, and an
additional step may be necessary. Chemical extraction in a counter-
current mode was chosen as a treatment method. This process is carried out
in a series of extraction steps in which the contaminated soils are given
sufficient time to contact solvents of increasing purity. A schematic of a
simple system incorporating/two extraction steps is shown in Figure 3. The
purpose of countercurrent extraction is to maximize the concentration
gradient between the soil and the solvent while minimizing the volume of
solvent required through solvent reuse. Therefore, the flow of soils from
step to step is opposite to the flow of solvents, thus the name counter-
current.
Besides the extraction tanks, the countercurrent extraction system also
requires solid/liquid separators after each extraction step. Various units
may be applicable, including centrifuges, belt presses, vacuum filters,
filter presses, and hydrocyclones. A gravity separation procedure could
also be used between extraction stages but is likely to be inefficient,
time consuming and, therefore, impractical. Centrifuges and filters are
often limited by solids loading and are extremely expensive, very large,
and often difficult to operate. They do perform efficient separations;
however, they are not suitable for multiple separations in a mobile unit.
13
-------
;?*
5
w
w
Q
OJ ^
s|
^s
°£
-------
Q
U
z
-~ e o
Q H-I H ol
w => <
_i O- a: •
O M < O
C/J J 0, Z
5
a
> i-
•J o
o
B,
a
>-H
O- O i
M H
35
O
CJ '
C
O
X
(U
(O
u
O)
o
KJ
o
O)
M- *
0 E
O*
•
u
oo
O)
s_
o
UJ
z
8
15
-------
The hydrocyclone, on the other hand, is less efficient but has a high
capacity and relatively low cost. It utilizes fluid pressure energy to
create centrifugal fluid motion, permitting separation or classification of
suspended solids. It is also capable of separating fluids of different
specific gravities. A sketch of a commercially available hydrocyclone is
shown in Figure 4. Feed slurries enter a tangential inlet at the top of
the device, then spiral their way around the cylindrical walls proceeding
downward. Centrifugal separation occurs during the period of spiral ing and
those materials having a higher specific gravity migrate toward the outer
walls. When the spiral flow reaches the lower part of the cyclone it
undergoes a direction change and begins to spiral upward toward the
outlet. The result is a spiral within a spiral. The materials along the
outer wall of the cyclone do not participate in this inner spiral action
but, rather, continue downward where they are discharged either
continuously or intermittently from the bottom of the cyclone. The solids
concentration in the outlet streams vary with, among other things, influent
solids type and concentration, particle size, specific gravity, and fluid
viscosity. Hydrocyclones do "lose" some fine solids in the overflow. The
amount of material discharged in this manner and- its impact on treatment
efficiency requires definition by lab testing. Mixing and chemical contact
of particles with the solvent are enhanced by the action of the
hydrocyclone.
CHEMICAL SELECTION
Once the preliminary scrubbing equipment choices were made, it was necessary
to choose contaminants and solvents to be used in the preliminary laboratory
testing.
Contaminant Choices
Four major criteria were addressed to select contaminants for preliminary
laboratory testing. These criteria are that the selected chemicals should
be:
1. representative of a wide range of contaminants that could impact
soils systems.
2. classified as hazardous and presenting a high risk of spillage.
3. involved preferentially in existing soil-contamination problems,
which could be partially or completely alleviated by successful soil
scrubbing.
4. chosen to span the range of difficulty of removal.
On the basis of the above contaminant-selection criteria, the following
three contaminants were selected for preliminary laboratory testing:
16
-------
INFLOW
INFLOW
TOP VIEW
OVERFLOW
UNDERFLOW
SIDE VIEW
APEX DISCHARGE
CLAMP VALVE
Figure 4. Schematic of hydrocyclone solids/liquid separator.
17
-------
Water soluble organic - phenol.
Water insoluble organic - polychlorinated biphenyls (PCBs).
Relatively insoluble inorganic - arsenic oxide.
Phenol
The selection of phenol as a soil contaminant for the laboratory testing is
well-justified. This chemical is one of the few water-soluble organic
compounds that meets the majority of the selection criteria. Phenol is a
very hazardous substance that is highly toxic to humans when ingested or
absorbed through skin. Fatal doses as low as 1 g have been reported;
however, the average fatal dose is 15 g (2). The drinking water standard
for phenol (to avoid taste and odor problems) is 0.001 mg/1. This
substance is also extremely dangerous to the aquatic environment. The
provisional water quality standard for the protection of fish species is
0.01 mg/1. The critical concentration for fish toxicity is 0.1 mg/1,
although there are species that can tolerate much higher concentrations (3).
Phenol is ranked extremely high in a ranking system which is used to define
potential risks from hazardous materials spills (4). In terms of the total
quantity of hazardous chemicals shipped by various means, phenol is ranked
46 (of 257). When ranked according to the overall risk (degree of hazard
and possibility of spillage) of spills of soluble hazardous chemicals,
phenol ranks first. During a 2.5-year period from January 1971 to June
1973, approximately 293,000 L (77,410 gal) of phenol were spilled in 19
reported incidents. Six additional incidents of spills occurred; however,
the quantity spilled was not reported. Three spills resulting from train
derailments accounted for 85.3 percent of the total volume spilled during
this period (5). These characteristics justify the choice of phenol for
testing.
Polychlorinated Biphenyls (PCB)
The major reasons for the selection of PCB as a contaminant for the
laboratory testing are the existing soil contamination problems with this
material and the hazardous nature of the substance. Unlike many other
hazardous chemicals, the probability of spills of PCB during transport
are currently quite small since the sale, production, and shipment of PCB
was drastically reduced since mid-1971. The majority of all the PCBs
manufactured during 45 years of production are still in use, however,
primarily in electrical transformers and capacitors, heat transfer systems,
and hydraulic systems. When this material is replaced, it is transported
to storage and disposal sites. Fortunately, due to stringent government
regulation, accidental spills are becoming less frequent; however, there is
a significant portion of the total PCB production that remains in the
environment (air, water, soil, and sediments). Also, a large fraction of
this material is present in dumps and landfills (6). The existence of
specific problem areas, such as Chatham County, North Carolina, and other
smaller areas throughout the country, provide the justification for
laboratory testing of PCBs.
18
-------
There is little dispute as to the hazardous nature of PCBs. Although it
does not exhibit acute toxicity (toxic effects from high level, short-term
w- exposure) it does accumulate in many biological species exhibiting chronic
toxicity even at very low concentrations. The chronic or long-term effects
of PCB exposure are similar to the chronic effects of DDT exposure. The
extremely stable nature of PCBs and their persistence in the environment
make this problem much more serious than that encountered with DDT. Unlike
DDT, the quantity of PCB present in the environment has not decreased
greatly since the initiation of regulations on the production and disposal
of these materials (6).
PCBs also meet the other selection criteria for an experimental soil
contaminant. This chemical represents a unique group in that it is
extremely insoluble and heavier than water. Solubilities of PCB range from
0.24 mg/1 for Aroclor 1242 (primarily two, three, or four substituted
chlorine atoms) to 0.0027 mg/1 for Aroclor 1260 (primarily five, six, and
seven substituted chlorine atoms) (6). Therefore, the laboratory approach
for testing the feasibility of removing PCB with a soil scrubbing system
with water as the solvent had to examine additives such as surfactants to
increase PCB solubility. As a result of the extremely low solubilities and
the lack of information on how to remove PCB that is bound to organic
materials within the soil, this material indisputably meets the "difficulty"
selection criterion. PCB also represents an extremely persistent group of
chemicals. Its stable nature is one of several properties that made it
ideal for use as dielectric and heat transfer fluids. The ability to
degrade PCB has been demonstrated in some microbial species, however, the
rate at which degradation proceeds is extremely slow. PCB is among the
most persistent of all potential organic soil contaminants.
Arsenic
The inclusion of arsenic compounds as soil contaminants for the laboratory
testing is to provide information on the feasibility of removing hazardous
inorganic materials from soils with a scrubbing system. Arsenic belongs to
the chemical group V-A. This group includes typical non-metals (nitrogen
and phosphorus), a typical metal (bismuth), and elements that are often
referred to as metalloids (arsenic and antimony). Arsenic generally
exhibits chemical properties intermediate to the extremes exhibited by
other members of this group. Arsenic compounds occur in many forms
including halides (e.g., AsCl3, AsFs), arsenites (e.g., AS03 =),
arsenides (e.g., Zn3As2, MgaAsz), sulfides (e.g., As2S3),
oxides (e.g., As20a, As205), and oxygen acids (e.g., HaAsOs) (7). Two
common oxidation states of arsenic in solution are As(III) and As(V). The
ionic forms and solubility of arsenic present in solutions are highly
dependent on oxidation potential (Eh) and pH. Other ions in solution and
minerals associated with soil solids influence the solubility and ionic
form of As.
Arsenic trioxide (As20s) is the most important arsenic compound in
terms of manufacturing technology and is utilized to produce nearly all
other arsenic compounds. These compounds include insecticides, weed
w 19
-------
killers, rodent poisons, sheep dip, wood preservatives, Pharmaceuticals,
and hide preservatives. The oxide is also used in the production processes
for commercial and ornamental glass, Paris Green, and enamels (7,8).
Although there are a large variety of arsenic compounds that could be
utilized as potential soil contaminants for the laboratory testing, arsenic
trioxide was the most desirable because of its widespread use.
Nearly all compounds containing arsenic are extremely toxic to many animal
and plant species. This is verified by the fact that a large number of
arsenic compounds are used in an assortment of pesticides and herbicides
(7,8). Also, the potential hazard posed by arsenic compounds in general
has been evaluated through rankings in terms of the total quantity of
materials shipped and the overall risk from a spill. Of the 257 hazardous
chemicals considered, arsenic compounds ranked only 221 in the annual
quantity shipped. However, when ranked according to the potential risk
from spills in aquatic systems, arsenic compounds have an intermediate
ranking of 98 (1). There have also been several occurrences of soil
p contamination problems with arsenic compounds, although these occurrences
5 " are not well-publicized in comparison with other contamination problems
such as those with PCB.
tj Solvents
Several classes of additives were considered:
Acids.
Bases.
Polar organics (short chain alcohols).
Oxidizing agents.
Reducing agents.
Surfactants.
Two major considerations for the selection of additives were cost and
associated hazards. Extraction of contaminants with an extremely high or
low pH solvent would require large quantities of additives. Also, the
"clean" soil may not be suitable for return to the spill site without
additional treatment. Likewise, additives formulated with extremely toxic
components could not be selected.
Solvent for Phenol
Since phenol is a fairly soluble organic (8.2 percent @ 15°C) and is also
quite polar, proposed solvent conditions included water alone (neutral pH),
water in a pH range from 1.0 to 12.0, methyl alcohol and water, and two
aqueous surfactants.
Methyl alcohol was selected as an additive to water since phenol is very
soluble in alcohols. Two surfactants (Tween 80 and MYRJ 52 from ICI Inc.)
were selected for testing since these materials are manufactured for the
purpose of increasing the solubility of organics in water by forming an
emulsion.
20
-------
Solvents for Arsenic
As discussed previously, the solubility of arsenic ions in a soil/water
system is influenced by oxidation potential, pH, other ions in solution,
and materials associated with soil solids. Therefore, a range in solvent
pH, an oxidizing agent (sodium hypochlorite), and a reducing agent (sodium
bisulfite) were selected for laboratory testing. At low Eh (oxygen
depleted soils). As(III) is the primary form of arsenic in natural soils
and the solubility is primarily a function of pH. Below a pH of 8,
cationic forms of As(III) predominate, e.g., AsO+l. Various anionic
species (^AsOJ1"2) are present at pH values above 9.0. Arsenic (III)
oxide (As203) was selected as the soil contaminant; therefore, a range
of solvent pH was considered a viable means of extracting the arsenic from
soil. Arsenic can be present in a wide range of oxidation states (-3 to
+5), so both a strong oxidizing agent and a strong reducing agent added to
water were selected as solvents for laboratory extraction testing.
Solvents for PCBs
In a review of available literature, only one solvent additive was
identified as increasing the solubility of PCB in water, namely, a
commercially available surfactant called Tween 80.
The solubility of PCB in water varies depending on the number of chlorine
atoms in the mixture. For example, the solubility of monochlorobiphenyls
range from 1.2 to 5.9 mg/1 at 2QQC, while the solubility of the
decachlorobiphenyl is 0.015 mg/1. The solubility of the decachlorobiphenyl
in a 0.1 percent Tween 80 solution is 5.9 mg/1 and exceeds 10 mg/1 in a 1.0
percent Tween 80 solution.
On the basis of this information, only water and water plus Tween 80 were
selected for laboratory testing.
SOIL SELECTION
Two representative soils with differing adsorption capabilities, as well as
somewhat variable physical handling potentials, were selected for the
laboratory experiments. Since a variety of chemicals were to be evaluated,
the broadest range of adsorption potential in terms of organic content was
desired. Thus, an organic soil and a non-organic soil were selected.
The nature of the soil structure was also important in the selection of
soil types to be evaluated. A soil mass may contain a wide variety of
gravels, sands, and clays. The range and proportion of soil types exhibit
varying degrees of cohesion between soil particles. The amount of energy
required to reduce the soil mass to treatable sizes is a function of the
cohesiveness of the soil. A granular sand and gravel soil, being
non-cohesive, can be readily separated by sieving or light crushing
forces. Silt and clay, because of their cohesive nature, require much more
energy for size reduction. Both the organic topsoil sample and inorganic
gravel, clay, and silt mixture chosen for testing exhibit cohesive, as well
as non-cohesive, tendencies.
21
-------
Soil Descriptions
Sand with Gravel, Silt and Clay—A naturally occurring sandy material from
the Pioneer Pit in Kenosha, Wisconsin was selected as a representative
non-organic soil. The clay and silt content was sufficient to provide some
cohesive clay lumps, although the material would generally be defined as
cohesionless (Figure 5).
Organic Topsoil~The organic soil was manufactured topsoil, produced by the
Liesener Topsoil Company of Jackson, Wisconsin. It consisted of moist peat
and humus, with little sand. The organic content was 18.4 percent. This
material represents a rich, friable, agriculturally excellent topsoil
(Figure 6).
Physical and Chemical Characteristics
The physical and chemical characteristics of both soils were evaluated.
Physical testing included specific gravity, organic content, unit dry
weight, permeability, and gradation. These tests were useful in defining
the soils and in establishing correlations with other soil types. Table 1
summarizes the properties of the selected soils.
22
-------
T
Figure 5. Inorganic soil for laboratory experiments,
23
-------
Figure 6. Organic Soil for laboratory experiments.
24
-------
TABLE 1. PROPERTIES OF SOILS SELECTED FOR
PRELIMINARY LABORATORY TESTS
Parameter
Topsoil
Sand with gravel,
silt, and clay
Physical characteristics
Specific gravity, g/cm3
Organic content, %
Natural conditions:
Moist density, g/cm3
Water content, %
Laboratory proctor density:
Standard method, pcf*
Modified method, pcf
Optimum water content, %
Permeability:
Falling head at natural
density condition, cm/sec
Atterberg limits:
Liquid limit
Plastic limit
Plasticity index
Gradation:
Percent passing, %
3/4" sieve, 19.1 mm
No. 4 sieve, 4.76 mm
No. 40 sieve, 0.420 mm
No. 200 sieve, 0.074 mm
Chemical characteristics
Cation exchange capacity,
MEQ/100 g
PH
Conductivity, umho/cm
Alkalinity, % CaCOs by weight
2.21
18.40
0.74
93.40
1.00
48.70
1.8 x 10-6
74
None
Non-plastic
100.0
100.0
82.5
56.0
26.0
6.9
860.0
0.61
2.65
1%
2.07
4.20
2.23
7.20
7.4 x 10-7
Non-plastic
Non-plastic
Non-plastic
100.0
60.3
22.5
14.9
2.3
7.9
400.0
5.28
*pcf = pounds per cubic foot.
25
-------
SECTION 5
PRELIMINARY LABORATORY TESTING PROGRAM AND RESULTS
INTRODUCTION
Once the preliminary scrubbing equipment selection and the choices of
chemicals and soils for laboratory testing had been made, the actual
testing began. The preliminary laboratory testing program had three
objectives, namely to:
1. define critical operating variables for potential scrubbing
components.
2. develop laboratory procedures that could be used to predict
scrubbing efficiency.
3. test three hazardous chemicals representing a wide range of
characteristics and potential for being removed by soil scrubbing
methods.
The laboratory testing was divided into two main parts. First, equipment
testing was performed to establish certain scrubbing system
configurations and operating variables. The actual chemical extraction
tests were then performed and incorporated much of the information
obtained in the equipment testing.
EQUIPMENT TESTING
Prior to beginning the chemical extraction testing, some preliminary
equipment testing was needed to define important operating variables for
potential scrubbing components. The equipment evaluations centered on
three components:
1. Water-knife size reduction and stripping apparatus.
2. Froth-flotation mixing and contact unit.
3. Hydrocyclone liquid-solids separator.
Water Knife Experiments
Testing of the water knives involved determining an effective contact
configuration and operational mode including nozzle type and size and
water flow-rate and pressure. Initial tests were done with a local
subsoil containing large amounts of clay lumps. No chemical contaminants
were introduced in this phase of the testing.
26
-------
Contact Configuration—Several different configurations were tested to
determine which one allowed the most effective stripping and lump
w breaking. Initial tests used the water knives to scour soils laying on a
moving, impervious belt. Little lump breaking was accomplished since the
high pressure water from the knife propelled the lumps away from the water
stream before any breaking up action could take place. Supporting the
soils on a screen and even using a secondary screen to sandwich the soil to
hold it immobile did not improve the results. It was concluded that the,
water knife could not function as a lump breaker unless the movement of the
lumps was constrained so that they could not move away from the knives.
After some discussion of the problem, a tilted, rotary screen was selected
as a device which could provide the necessary constraint. A schematic of
the water knife and rotary screen assembly is presented in Figure 7 and a
photograph of the rotary screen is shown in Figure 8. A U.S. Sieve No. 10
screen (2-mm mesh openings) was selected to retain as much coarse material
as possible in the basket for stripping while allowing the fine soil pass-
through to be transported to the extraction process. This screen choice
represented the smallest mesh that would probably not be subject to
plugging problems.
Operational Mode—Various operational modes were tested using about 2.7 Kg
(6.0 Ib) batches of the local subsoil. The soil was placed into the rotary
screen and physically stripped for different time periods at a rotational
speed of•about 20 rpm. At various time intervals, residual soil weights
were recorded and observations were made.
The first tests were performed at a water flow rate of 4.5 1/min (1.2 gpm)
***•*"•• and a water pressure of 4.9 Kg/cm2 (70 psi). The tilted orientation of
the rotary screen caused the soil to remain in a small pile near the bottom
of the screen. Slow rotation of the screen (10 - 20 rpm) continuously
turned the soil mass over to permit the knives to contact more of the
soil. In addition, the water knife was moved up and down across the soil
pile so that it scoured the entire area of the pile approximately 20 times
per minute. As the flow from the water knife contacted the soil, a violent
scouring action occurred and much of the loose soil was forced through the
screen rapidly. Figure 9 shows that the rapid initial breakdown of the
soil reduced the retained volume of the soil by two-thirds during the first
two minutes of operation. After this time, the soil reduction rate began
to diminish and all that remained in the rotary screen was a small amount
of gravel and grayish clay lumps. Effective breakdown of these clay lumps
by the water knives continued, but more slowly. A soaking step was found
to be helpful in softening the lumps for easier and faster disintegration.
It was concluded from the preliminary lab tests that three cycles are
needed for best performance of the water knives on the 2-mm screen:
27
-------
to
I—I
2
a;
u
(O
s-
(O
Q,
Q.
rO
-------
Figure 8. Rotary screen and water knife.
29
-------
c
o
CJ
HI
CO
2
I—i
oi
X3
C
01
c
Ol
4->
o>
i-
O)
+->
03
'1HDI3W 1IOS Q3NIV13H
O)
3
Ol
30
-------
1. A primary cycle using high pressure fluid recycled from the
collection sump. During this period, small soil particles and soil
lumps that can be broken down to less than 2 mm pass through the
screen.
2. A secondary cycle composed of soaking and of stripping with water
knives and spray nozzles.
3. A rinse cycle using clean water at low pressure as a last step
before discharge of the retained sand and gravel.
Froth Flotation Unit Testing
One alternative piece of equipment for use as a mixer/contactor during
chemical extraction was the available EPA froth flotation unit (13). To
determine the feasibility of utilizing the existing mobile froth flotation
unit as part of the scrubbing system a laboratory size froth flotation unit
was obtained from EPA for use in the testing (Figure 10). Prior to actual
chemical extraction testing, preliminary mechanical tests were performed to:
1. Familiarize test personnel with the operation of the unit.
2. Determine the range of unit speed and air flow settings for adequate
mixing action.
3. Determine the range of solids concentrations that can be kept in
suspension with the unit.
4. Qualitatively evaluate the froth flotation unit for use as a mixing
cell in countercurrent extraction.
The mixing action of the unit was first observed using water alone. The
pattern of mixing was observed through the Pyrex mixing cell. Different
mixing speeds and air flow settings were tested, and it was quite apparent
that the parameters were interrelated. Speeds of 1200 to 3100 rpm were
tested. When the air flow was too low at a particular mixing speed, the
mixing was very turbulent. Increasing the air flow to a certain point
resulted in a very smooth mixing action. Once the point of smooth mixing
was achieved, increases in air flow had little effect on mixing action. As
the froth mixing speed was increased, the air flow necessary to obtain
smooth mixing action also increased.
Testing continued using the organic and sand/clay soils with solids con-
centrations in the range of 10 to 30 percent. The results indicated that,
as solids concentrations were increased, higher mixing speeds were required
to achieve smooth mixing conditions. Tests conducted at similar concentra-
tions indicated that the same speed and air flow settings were required for
31
-------
Figure 10. Laboratory froth flotation testing apparatus,
32
-------
'***'
both soils. An air flow of 470 ml/min was sufficient to maintain a smooth
mixing action up to an operating speed of 2400 rpm during all of the tests.
The quantity of froth produced was generally minimal.
For the organic soil, the best mixing actions were obtained with the opera-
tional parameters at the following levels:
Optimum Optimum
speed air flow
Percent solids _ (rpm) _ (ml/min) _ Comments _
11.5 1800 - 2100 470 Little change in mixing
when air flow increased
to 640 ml/min.
19.1 1900-2100 470 When speed increased
to 2400 rpm, air flow
needed to be increased
to 580 ml/min to main-
tain smooth mixing
action.
These results indicate that there were few changes necessary to maintain
adequate mixing in this range of solids concentrations. The relationship
between air flow and mixing speed observed during the testing with water
was also observed when the speed was increased above the optimum to 2400
Test results for the sand/clay mixture are summarized as follows:
Speed Air Flow
Percent solids (rpm) (ml/min) Comments
18.4 1900-2100 470 Air flow of 600 - 640
ml/min required when
speed increased to
2400 rpm.
31.0 2700 - 3100 580-610 Some difficulty in
keeping all particles
in suspension—required
balanced adjustment of
speed and air flow
settings.
Solids concentrations higher than 30 percent were tested with this soil and
the froth flotation unit provided an adequate mixing action. However, flow
settings seemed more critical and required some adjustments to ensure
smooth mixing action at high solids concentrations.
Basically, these results indicate that the froth flotation unit provides
sufficient mixing action during chemical extraction of soil particles less
than 2 mm in size. Adequate mixing was achieved at solids concentrations
up to 30 percent when proper rotational speed and air flow were maintained.
33
-------
Hydrocyclone Testing
Preliminary equipment evaluations had identified the hydrocyclone as a good
candidate for separating the soil particles from the solvent after the
extraction steps. In the laboratory testing program, the efficiency of
this separation method had to be established for the soil types of
interest. After contact with manufacturers, it was determined that
accurate and comparable results could not be obtained from a small
laboratory-sized unit. The principal problem was the small size of the
apex discharge, scarcely larger than the maximum size of the particles in
the hydrocyclone feed water (less than 2 mm particles).
Therefore, a full size system, which operated at a flow rate up to 227
1/min (60 gpm) and a pressure drop of 1.4 Kg/cm2 (20 psi), was obtained
for test from Krebs Engineers. According to the manufacturer's
specifications, the unit could make a particle separation around 20 - 25 u
on a silt/sand mixture with a specific gravity of 2.6. The actual
separation efficiency—including the amount of solids entrained in the
overflow and the percent solids concentration in the underflow—had to be
established prior to the chemical extraction testing. This set of tests
was limited to soil and water with no chemical contaminant introduced into
the system.
Equipment Setup—A large tank and mixer was utilized to mix the soil
(particles less than 2 mm) and water initially and to receive both the
underflow and overflow from the hydrocyclone (Figure 11). This setup was
required to minimize tank capacity requirements since the hydrocyclone was
operated at inflow rates ranging from 95 to 189 1/min (25 to 50 gpm). A
large tank mixer was utilized to keep the soil in suspension in the holding
tank and a 7.6-cm (3-in.) submersible pump was utilized to feed the
hydrocyclone.
The underflow contained the concentrated soil particles whereas the
overflow contained only the fine particles, which could not be separated
with this hydrocyclone. The apex was a rubber tube that delivered the
underflow. The flow ratio of underflow to overflow and the underflow
solids concentration could be varied with a hose clamp on the apex.
Test Conditions—Three variables were identified for the laboratory testing
of the hydrocyclone, namely:
1. Inflow rate or pressure drop (interrelated variables)
2. Inflow solids concentration
3. Apex setting.
Inflow rates at pump pressures of 0.70, 1.40, 1.80 Kg/cm2 (10, 20 and 25
psi) were evaluated in the tests. The flow rates associated with these
pressure drops varied on the basis of soil type and inflow concentration.
Two soil concentrations were utilized for each soil type. The organic soil
34
-------
Figure 11. Laboratory hydrocyclone testing apparatus.
35
-------
was tested at concentrations of approximately 5 and 10 percent dry weight
while the inorganic soil was tested at 2.5 and 5 percent solids. A maximum
inflow solids concentration of 10 percent was suggested by Krebs En-
gineers. Krebs Engineers recommended that the appearance of the underflow
after it discharged from the apex should be cone-shaped rather than
rope-like. Underflow "roping" results from excessive tightening of the
hose clamp on the apex. Two extremes of the apex setting were evaluated
during the testing:
1. An open setting where no pressure was applied to the apex with the
hose clamp.
2. A closed setting where the underflow was still cone-shaped but any
additional tightening of the hose clamp results in solids roping.
For each set of test conditions (3 pressure drops x 2 soil types x 2 soil
concentrations x 2 apex settings = 24 tests), samples of the inflow, under-
flow, and overflow were collected and analyzed for total solids. Overflow
and underflow rates were also measured for each test condition.
The results indicate that the least sensitive parameters were pressure drop
and inflow rate. The following trends were observed with increasing
pressure drop or inflow:
1. Underflow, as a percentage of the total flow, decreased.
2. The percentage of the inflow solids in the underflow decreased.
3. The underflow solids concentration decreased.
Although these trends were consistent, the differences observed as a
function of pressure drop did not have a significant impact upon test
results. Therefore, the data have been summarized by averaging results at
the three flow rates.
Table 2 contains the summarized data from the hydrocyclone tests. For the
organic soil, the percentage of the total flow as underflow ranged from 4.9
to 18.5 percent and the percentage of the inflow solids concentrated in the
underflow ranged from 49.6 to 74.3 percent. Tests conducted with the open
apex setting had the highest proportion of inflow solids (63.4 and 74.3
percent) as underflow; however, higher solids concentrations were obtained
in the underflow with the closed apex setting.
The inorganic soil results showed similar trends as the organic soil;
however, the hydrocyclone was generally more effective in concentrating
solids in the underflow in these tests. The percentage of total inflow
solids in the underflow ranged from 67.6 to 82.3 percent. Underflow solids
concentrations were more than two times higher with the apex closed rather
than open.
Particle-size gradation testing was performed for both the inflow and under-
flow for the inorganic and organic soils. Combined sieve and hydrometer
tests were conducted. Testing indicated that the fine fractions were
removed from the organic soils in the particle size range of 5 to 300 u.
36
-------
I
^
i
s
**
s
58
O ^3
O CM CM O O
P^r^^r^r^
O) •*-»>*
o» e
w X
o o
li
M *
28
M
m at >«» <«> oo o » >e
Si "S "S "S "S
i
«
sl
s|
8
u u u u
..
* s- i t ? ?
OIO1O1OOO
S
u
•J: £'
M >—
S "i
8 -5
X
*
M
i
wi
«
•8
I
1 1
2 g
1 2
e o
§ §
37
-------
The majority of change was in the 10 to 200 jj range. The hydrocyclone had
no effect on the removal of clay-size particles, namely, those smaller than
5 ju. For the inorganic soil, the particle size gradation removed in the
underflow contained a higher percentage of fine particles than the inflow.
The fractions of particle gradation removed were in the range of 5 to
1000 >i, especially noticeable in the 10 to 100 ju range. This observation
implies that the hydrocyclone removed the coarser fraction from the
inorganic silty, fine-to-medium sand soil as expected.
Additional hydrocyclone testing was conducted to determine whether the
overflow solids could be effectively removed by an additional hydrocyclone.
Tests were conducted with organic soil overflow with a concentration of
about 1.7 percent solids and inorganic soil overflow with a concentration
of about 1.5 percent solids. Table 3 lists the results from these tests.
The majority of the tests were conducted with the apex setting open since
the purpose of the tests was to optimize total removal of solids. The
results indicated that approximately 25 percent of the solids in the
organic overflow and 20 percent of the solids in the inorganic overflow
could be removed as underflow. However, the concentration of these
underflows was only between 2.0 and 3.8 percent, which indicates that the
same hydrocyclone is not an effective means of concentrating the fine soil
particles in the underflow. Underflow concentrations as high as 9.0
percent solids were obtained with the closed apex setting, but due to the
low rate of underflow with this setting, only 6.5 percent of the inflow
solids were removed as underflow. These results show that the same model
hydrocyclone cannot be used to clarify the spent solvent used for
extraction of contaminants from soil.
Hydrocyclone Simulation Testing
To accomplish a laboratory scale evaluation of the proposed counter-
current extraction scheme for the removal of contaminants from fine soil
particles (less than about 2 mm in diameter), it was necessary to develop a
bench technique to approximate the solids removal achieved with the hydro-
cyclone. A hydrocyclone could not be used for the chemical testing since
the smallest unit readily available for this testing operated through a
flow range of 95 - 189 1/min (25 to 50 gpm), much higher than can be
efficiently utilized in the lab scale study. Scale down using a smaller
hydrocyclone would not be representative of separations obtained in the
larger hydrocyclone. Since the 95 - 189 1/min (25 - 50 gpm) flow range was
anticipated to be utilized in the full-scale scrubber unit, simulation of
this device should give the most valuable results.
Settling tests were tried as a method for simulating solids removal with
the hydrocyclone. Initially, samples of the inflow used for the hydro-
cyclone tests were mixed, poured into a graduated cylinder, and allowed to
settle for specified periods of time. Results from the hydrocyclone
testing were used to define the relative proportions of underflow and
overflow for a given soil type and concentration, and the solids
concentrations observed in the overflow and underflow were utilized for
comparison purposes. Preliminary tests were conducted assuming both an
38
-------
i-X
I/I C
o o
\St (J
.0
OT3
5§i
r*.
v c*j f^ o 9 oo
S|
00
SJ-
55 ee e; ooi 01 S da a
§
T 9
IS |"
I
u
O Sj (NJ ->»<
I
1 I
•I-
»«
•I-
M
I I
i ^
i i
2 S
1 i
o o
S §
o
«
"3
jj
IO
u
39
-------
open and closed apex setting with the organic soil. With a 5-percent
organic soil inflow and closed apex condition, the underflow volume
averaged approximately 5 percent of the total volume for the three flow
rates tested with the hydrocyclone. Therefore, after the specified
settling time for the simulation tests, 95 percent of the total sample
volume in the graduated cylinder was removed using a vacuum trap. This
liquid was considered overflow and analyzed for total solids. When
simulating the open apex setting, approximately 85 percent of the total
sample volume was assumed to be overflow, as determined from the
hydrocyclone tests.
Results of the preliminary tests are presented in Table 4. The overflow
concentrations were utilized for comparison because of difficulties
associated with obtaining a representative sample for total solids analysis
of the sludge (underflow) in the bottom of the graduated cylinder. All of
the overflow concentrations from the settled samples were close to those
obtained from the hydrocyclone. The overflow concentrations for the open
apex setting simulation were lower than those for the closed setting since
only 85 percent (rather than 95 percent for the closed setting) of the
total sample volume was considered overflow. The results from the
hydrocyclone testing indicate that the overflow concentrations are not
affected significantly by the apex setting.
TABLE 4. COMPARISON OF PRELIMINARY SETTLING TEST RESULTS
WITH HYDROCYCLONE OVERFLOW TOTAL SOLIDS CONCENTRATION
Soil type
Organic
Apex
setting
Open
Open
Closed
Closed
Settling time,
minutes
5
20
5
20
Overflow concentration, %
Settled
2.76
2.08
2.85
2.42
Hydrocyclone
2.36
2.36
2.32
2.32
These results indicate that settling was a useful technique for simulating
the solids separation achieved with a hydrocyclone. However, the settling
technique and times had to be modified so that these could be used in con-
junction with countercurrent extraction testing. Additional settling tests
were conducted with both the organic and inorganic soils but a large beaker
was used for settling the samples which allowed easy removal of the settled
solids (underflow). A closed apex setting was assumed for all of the
beaker settling tests; and inflow samples, from the hydrocyclone testing,
were retained and utilized for these tests. Table 5 lists the overflow
concentrations obtained with the settling tests and from the hydrocyclone.
The results indicate that short settling times are sufficient to achieve
the desired overflow concentration. With the organic soil, a 10-minute
settling time provided excellent results for the 5-percent solution,
40
-------
however, the results indicate that a 5-minute time may be required for the
10-percent solution. With the inorganic soil, the 5-minute settling time
provides a good comparison to the hydrocyclone results for both soil
concentrations.
The results of the settling tests indicate that, for the soils being tested,
settling is a feasible technique for approximating solids removal with the
hydrocyclone. The purpose of these tests is not to exactly duplicate the
hydrocyclone, but to provide a technique for determining the effects of
incompleted soil/solvent separation on the countercurrent extraction system.
TABLE 5. BEAKER SETTLING TEST RESULTS
Mixed
concentration Settling time Overflow concentration (%)
Soil type
Organic
Inorganic
(%)
5.0
5.0
10.0
10.0
2.5
2.5
5.0
5.0
( mi n . )
10
20 •
10
20
5
10
5
10
Settled
2.31
1.99
2.66
2.43
0.80
0.64
1.77
1.52
Hydrocyclone
2.32
2.32
3.69
3.69
0.77
0.77
1.67
1.67
CHEMICAL TESTING
Once/many of the mechanical operating parameters were established, chemical
testing began. The testing involved several different steps designed to
establish test conditions and treatment efficiencies:
1. Column dosing studies, to define the amounts of contaminant to be
added to the soil for subsequent laboratory tests.
2. Contaminant distribution tests, to determine the distribution of
contaminants according to particle size.
3. Water-knife stripping/soaking tests for particles greater than 2 mm,
to determine the ability of this process to strip contaminants from
large particles.
4. Extraction tests for particles less than 2 mm, to determine the
ability of the countercurrent chemical extraction process to extract
chemicals from the soils.
41
-------
Analytical Techniques
Accepted analytical procedures were utilized for all parameters involved.
The specific methods were chosen to provide accurate results with a minimum
amount of analytical time. A colorimetric procedure was utilized to
quantitate for phenol in order to save analytical time. Although the
procedure was straightforward for water samples, soil analysis involved a
multiple extraction of phenol into water. The water used in the soil
extractions was analyzed and the results summed to obtain the total
quantity of phenol in the soil sample. A 2-g sample of soil contaminated
with phenol was mixed with 500 ml of clean water and shaken for 15
minutes. The liquid/solid separation was accomplished by centrifuging and
the liquid was analyzed for phenol. The soil was resuspended in 500 ml of
clean water and the process repeated until the concentration measured in
the extract was insignificant compared to the initial concentration. The
final result was reported as a summation of the phenol analyzed in each
fraction (mg/kg dry soil).
Arsenic analyses were conducted utilizing atomic absorption flame tech-
niques as presented in several sources (9, 10). Soil samples for arsenic
analysis were digested with a mixture of hydrochloric and nitric acids.
PCB samples were analyzed by a gas chromatography-electron capture detector
(11). Hexane was utilized to extract PCB from soil samples for analysis.
Standard cleanup procedures were utilized in the analysis.
Column Dosing Studies
The purpose of these tests was to define the amounts of contaminant to be
added to the soil for the subsequent laboratory tests. The approach
involved simulating a spill condition as much as possible. The liquid
contaminants were applied to the surface of the soil columns and the fluid
was allowed to drain for 24 hours. In this way, the time factor associated
with a spill event was simulated since the contaminant (i.e., pollutant)
was primarily retained in the void volume of the soil with less being
sorbed on the particles themselves.
Test Methods—A Plexiglas (8.9-cm (3.5-in.) diameter x 15.2-cm (6.0-in.)
high) column was packed with soil using a range of soil moisture conditions
and dry bulk densities, which simulated natural conditions. A stock
solution of contaminant at a volume equal to the soil void volume for the
given soil condition was poured onto the column. After 24 hours, the
volume of the liquid passing through the column was measured and analyzed
for contaminant concentration. The quantity of contaminant entrained by
the soil was calculated on a dry soil weight basis. These tests were
conducted for both the organic and inorganic soil types.
Stock solutions of contaminant were chosen to simulate conditions expected
during a spill. Since technical grade phenol is often shipped as a liquid
having a purity of 87 percent, this concentration was used. Arsenic
trioxide formed a saturated, aqueous solution (9600 mg/1) within 24 hours,
thus simulating the dissolution of solid arsenic oxide after a rainfall.
42
-------
The PCB concentration occurring during a spill could vary significantly
depending upon the situation. A value of about 75,000 mg/1 was used for
PCB testing.
Column Dosing Results—Phenol was the first of the contaminants tested.
Each soil type was dosed under three conditions, namely, concentration of
cantaminant, soil water content, and dry bulk soil density. The conditions
represent a wide range of natural situations, including high and low soil
densities and water contents. Table 6 summarizes the results. For both
soil types, the intermediate dry bulk density and lowest water content
(tests 2 and 5) resulted in the largest entrainment of phenol. These
dosing conditions were then utilized for all further laboratory tests with
phenol since they represented worst case conditions. These worst case soil
conditions were then used for the dosing studies performed on both arsenic
oxide and PCB. The dosing rates determined from these tests are as follows:
Arsenic oxide PCB
Soil type mg As/gm dry soil' mq PCB/gm dry soil
Inorganic 0.75 3.0
Organic 5.00 25.6
Contaminant Distribution Tests
These tests were conducted to determine the distribution of the sorbed
contaminants according to soil particle size. Only soils contaminated with
phenol and arsenic oxide were used because of hazards associated with
excessive handling of soils highly contaminated with PCB.
Test Methods—The methodology for the distribution tests differed for
phenol and arsenic because the more volatile phenol could be lost during
drying. For phenol, the tests were performed by separating dry soil
fractions using a sonic sieving device and then contacting the fractions
with phenol stock solution. In order to assure contact of the phenol and
the soil fraction, a sample of the fraction was placed on filter paper and
then phenol solution was poured over the fraction. It remained in contact
with the soil for 18 hours. The soil fractions were then rinsed with water
to remove phenol from the void volume and analyzed. For arsenic,
contaminated soil from the column tests was dried at 60QC, size fractions
were separated with the sonic sieving device, and the fractions were then
analyzed.
Contaminant Distribution Results—Tables 7 and 8 list the results of the
distribution tests for phenol and arsenic, respectively. The results are
presented in two ways:
1. Actual contaminant (phenol or arsenic) concentration per unit weight
(kg) of dry soil.
2. Percent of total soil contaminant in each size fraction.
Differences of less than 20 percent were found when the fractionated phenol
values were summed and compared to the total amount measured on the non-
43
-------
o
1 . 1
LU
a.
a:
O
*•*"
oo
*~
•^
oo
UJ
cc
,
oo
~*
'
CD
t— i
oo
O
o
^s
3
o
0
10
UJ
O3
^^
2^
^~*
•P™
o
•Q (/>
co
c >
•i- S_
*O J-
£ +•> E
co E o
a. i—
o
c
a
a
0
•M
c
£ a
«-— •
•a
O)
•o
•o
tO
"o
c
a. E
3
"o^
^•>
;—
o
>fc<^™«
V. E
a o
-M
IB C
i- O)
CD -t->
-(->£• —
^-^^
3 -M CO
.a T- E
(/) "X
>> C 0
S-
4-^
r-—
•p-
O
4_> •
(/I O
CO C
1—
Px CO
00 00
1— ^-
o o
r— 00
^- CVJ
rx. un
CO CVJ
co «a-
"*" ^
O 00
O 00
O rx,
CVJ r—
CM CVJ
• *
^^« ^^
LD O
<— *t in
*f co
oo o cr> cr>
rx co oO CT*
^^
C^ ^f CO l"s^
• • • •
i— CO Lft 00
O r- fx O
en co cy> o
CO CO 00 r—
P^
u
£ U U U
C7I £ £ £
^K (C 1C (O
O 7> C7) O^
E i- i- S-
I-H O O O
CO «* LO tD
cu
•
•r~
*— *
r-
•r-
O
>j
S-
"O
4-
O
•M
•r~
cu
5
aj«— v
o
>> to
JD
-a co
ai -3.
•a
>§,
•o o
o
i. r—
CO
•u s-
fO CO
3 OL
^L _ — .
o •--
o
+-> 10
r-
O> >>
•i- S-
CO "O
CO 0)
-l-> fx.
•
(/) 1 —
tO LO
"O IO
O) £
Iw 'O
•r- CO
•+- e
• CO
CO T3 +J
1 % fK
** ^
4- (/Id)
-Q S-
r— CO O)
^ Q.
II -r-
o «a-
CM tO •
(£> CO
o s- en
0
O 4— H—
o
5^ ^j
£ +•>
CO CO £
E -U CO
*x £ 4->
0) O £
S^ O O
0
CO S_
CO +J CO
-|_> ro +J
o 3 us
z * 2
44
-------
LU
GO
UJ
_J
-
CO
t/)
_J
t— 1
£2
^o
0
1—4
z
cn
Q£
o
z
1—4
o
z
o
1— »
1
o
z
o
_J
o
z
UJ
T:
o.
z
o
^—
=>
CO
*"•*
Or*
GO
Q
•
fs.
UJ
oo
£
i
p—
O
c
4_> •> f~™
CO *r-
0. O
I/)
P—
IO
• J
0
jQ
C
•o >> o
0) J3 -P-
4-* 4*^
IO i— U
i— O
0 N4
•P»
m
^3
0) •— O) C
«^ ft
•*• ^rf
0| CX 0
o
c
4-» t-
fll — «
*i/ *^^
£o
0) M
a.
01?
IM «— •
•p»
(M C
O
r~ «f»
•i" 4>^
o u
GO IO
<4_
c
o
•r-
+J
0
03
i.
*4-
0)
M
•r-
(/I
C
**~
x—^
cn
E
"•V*
^— ^
on
cn
•*-*•
c
o
^->
y
t-
w
N
10
C
z
X
^f
CO ^^ CO
p— r- CO
i— VO O-
O O O
rC co
^~
CMtx.,-
i— CM CM
r>. in ci>
•— co vn
O O O
tO O^ ^D
co in o
O 0-—
in
vo
CM
vo
«el- i—
«3- CO
O 0
i— CM
O O
§2
r- O
CO CM
C»0
co O
r"»
o
8
,_
§§
CM CM
o o in
f^> CO ^**
in i—
O O r-
tj- in vo
CO CM
0 O 0
in co oo
r^ co in
oo"
in i— o
cn in m
o in
•— co
0 0
in ^- o
o m o
p— CO O
O O r—
u
•r-
C
cn
t.
o
c
in
•^
r_
CM
0 0
CM in
CO CM
,_T
0 O O
in in o
O CM CO
CM
«• o
^8
8 8
0 CP
i— CM
§"o o
o o
CM CM CM
r~*
X
C
0
•r~
4J
U
fB
«w
i.
<*-
ai
M
•P"
«/>
C
•P"
(•^
•p*
o
VI
«4-
O
4^
C
0)
u
vv
Q^
•^ ^
(/)
«s
CV
.^J
*o
^v
u
•
a>
rw
(/>
r"~
IO
4J
O
cn
o
^
• 10
c
o c
•p* *r—
O i—
ia o
J_ C
vp* VI)
jr
C CX
M-
•o o
W
t- 4J
3 C
in 3
IO 0
W E
«= J
•
^J
c
a>
L>
&
o
• 1.
^•^
c
o
4J
u
J_
c
•
fj
c
o
u
c
o
0) 3
JE o *r
JC JC
v»- 4-> cn
O 3
r- O
i— J.
IO •/> JC
4J $ 4->
0 0
ai s to
-c u x
4-> ••- 0
c t.
>> > IO •*—
J3 cno
i- (/i
0°0
AI ^"\ ^™
V/ O C
JC Vf_ 10
ex cn
cn t.
•o E o
0) C
IO l^^k
r- fv> J_
U VO*H_
r- VO
10 i— cn
0 • E
^3 W ^^
• ^«*
a> T- •
3 r-
IO i/l "O
> C C
o ia
JC T-
O 4J>-^--^
•O U C C
AI «< c e
w *w c E
t. 3 3
Cn«4- r— r-
C O O
"- W U O
•O M
•p» -f— s s
> in i— i—
-r- O O
•0 r- C C
>s«J= 2
>> 10 jc jc
-Q CX CX
•o o -o -a
W •*_
-------
co
cc
•^
U
tr~
^Z 'O
cu
0 r—
S- T-
cu o
a. to
lO
o
4->
jz
>> c
.a o
O 4->
•i- O
53 'O
CU i-
10 «4-
S-
rO CU
PSl
to
O IO
•^> .
C O
cu c
l/l O
S- U
o
c
4^ *^m
c
u £-
s_ o
cu
a.
N
to c
o
o o
oo
c
s*-~^
01
E
cn
"•s^
Ol
£
•^—*r
c
o
'£>
o
s_
**"
cu
IM
-
C
2
X
IO
cu
a.
>n
+J
1—
•^
O
oo
•—
,
oo
CX3
00
00
l>*
r_
LO
O
0
U
c
C
CU
u
S-
cu
Q.
10
lO
•o
0)
(O
p—
3
O
(O
0
4^
• c
cu cu
i— U
Q. i.
E CU
IO Q.
to
•r— p—
0
to
r-
fQ X^^
+•> c
0 0
cn u
^ CU
C 5-
3 3
O to
E (O
< CU
.0 E
o
li
*^
o «a-
vo J=
cu cn
<"• f" 3
4-> cn o
3 S_
M- o .c
o s- 4->
.c
p™ 4-* P~
IO
0 "~ •£
4J l/> O
S S-
0) 0
-C S- =
4-> _ 0
_Q -i- lO
c cn
A A3 ^»
cn cn o
E s- c
O -i-
<* S S
c
o «*-«»-
•i- O O
-M
0 E E
IO 3 3
S- to to
CU •— t—
N 'I- •!-
••- O O
to to to
>> U 0
^^ «^* «r~
c c
O
4-> E
(0 V-
r— 00 O
3 • >4-
u oo
i— oo cn
10 PN. E
u
•> LO
O CU 00
* ^^
0} •<- r—
3
> c «
o •
^ »^ ^— ^^^
0 4J C C
•0 o E E
CU IO 3 3
S- i— r—
cn<4- o o
C 0 O
•t- a>
•a N = s
•r- -r- O O
> t/1 •!—•!-
"- EC
"O i— CU CU
r— VI V}
>> 10 S_ t.
.a 4->
fO ^™ *O fO
i— 10 ^ r—
3 O 3 3
0 •<- O 0
•—£•—•—
CJ c" c_5 C_5
0 O = =
•
•a
cu
c
o
•r~
4^
0
-------
fractionated sample (Table 7) (viz., ((208100-166774)7(208100))xlOO = +20%
((1250-1471)/(1250))xlOO = -18%). This recovery was quite high considering
the natural soil variations and the difficulty of phenol analyses on
soils. The comparisons were even better in the case of arsenic (Table 8)
(viz., ((806-783.8)/(806))xlOO = +3% and ((200-178.5)/(200))xlOO = +11%).
Phenol Distribution—The results for the phenol distribution tests onto
organic soils indicate that approximately 90 percent of the phenol was
sorbed by the larger particle sizes (0.595 to 2 mm). It was hypothesized
that this result was caused by a high organic content in these larger size
fractions. To further evaluate this, the organic content or loss on
ignition (LOI) test was performed on the organic soil. Samples were heated
to 5500C for four hours and then reweighed (ASTM Test Designation
D-2974). The differing organic content in the coarse and fine fractions
was then apparent. When the LOI values and'phenol adsorption for different
size ranges are presented on the same graph (Figure 12), it can be seen
that a direct relationship exists for size ranges between 0.9 mm (U.S.
sieve No. 20) and 0.25 mm (U.S. sieve No. 60). This result suggests that
woody fiber in the sample sorbed large amounts of phenol. For particles
smaller than 0.25 mm, a direct relationship between the organic content and
phenol sorption was not observed.
The absence of a direct correlation between organic content and phenol
uptake in the fine fraction (<0.25 mm) indicates chemical, as well as
physical, difference between the coarse and the finely divided organic
fractions. The organic fines are most probably composed of the
decomposition products resulting from microbiological attack on plant
tissue. These products are commonly termed "humic" or "fulvic" acids and
are characterized by large surface-to-mass ratios and an absence of
internal structure. The particles are generally the size of clay
minerals. By contrast, the woody fibers still retain much of the open
structure and porosity of the original plant material. The high phenol
sorption in the coarse organic fraction appears to result from the
penetration by the phenol into the woody fibers and retention in the
internal pore spaces. Conversely, the finely divided organics seem to
adsorb in a manner similar to clay slurries. The net result implies that
phenol uptake in organic soils may be directly related to the degree of
decomposition of the organic materials. Removal of phenol from the
internal pore spaces of woody fibers will probably require soaking. Since
the phenol appears to have been absorbed by the woody fibers, a wash of the
particle surfaces would probably be only partially effective. The
conclusion that phenol sorption in organic soils is controlled by the
penetration into internal pore spaces of woody fibers also suggests that
phenol adsorption would be very sensitive to water content of the fibers.
If the internal pore spaces were occupied by water at the time of phenol
application, very little sorption would be anticipated immediately. Since
either dry or wet conditions may exist in natural soils, the amount of
phenol sorption onto larger soil particles at a specific spill site cannot
be easily predicted.
Inorganic soil fractions up to 12.7 mm were evaluated to determine phenol
adsorption (Table 7, Inorganic). Similar calculations were made with these
47
-------
C
o
•o
(O
i-
00
C
E
r-
O
t/1
i-
O
c
O)
C r—
O T-
•i- O
-(-> CO
a.
s_ o
o •!-
CO C
T3 (O
(O a>
i.
-a o
c
(O C
(O
o o
C
o
C
a>
CO
CO
o
'anaaosav lONana) ^ 'NOIIINOI NO SSOT
O)
i-
en
48
-------
data as had been made with the data from the organic soil tests so the
results could be compared. The results indicate that in the particle size
range less than 2 mm, approximately 60 percent of the phenol was sorbed by
particles less than 0.105 mm (viz., (840/471)xlOO = 57%).
The coarse sand and fine gravel material (2 to 12.70 mm) sorbed about three
times more phenol on a weight basis than the soil fraction less than
0.105 mm. This result was unexpected because the small particles have a
much higher external surface area per unit weight than the gravel.
However, the dry, limestone-type gravel found in this soil has a greater
total surface area than that calculated solely on the basis of average
particle-size diameter. The limestone has a high internal surface area
because of pores within the particle itself. These internal pores greatly
increase the surface area exposed to phenol in the dosing tests and this
condition can account for the higher sorption found during the testing. It
should be noted that this sorption was increased because only dry soil was
used in the testing. The natural water content of the soil would have
filled some of these internal pore spaces and, therefore, decreased the
amount of available surface area for chemical adsorption.
Arsenic Distribution—Arsenic distribution on the organic soil (Table 8)
indicated that nearly 78 percent of the arsenic was associated with soil
particles ranging from 0.595 to 2 mm. The results are similar to those
obtained in the phenol testing and can possibly be explained by the
retention of arsenic within the pore spaces of the woody fiber present in
these size fractions. The smaller soil fractions sorbed approximately the
same amount of arsenic per unit weight but contributed differently to the
total based on the weight of that soil fraction present in the sample.
This is the same trend observed with phenol indicating again that the
surface area is not the only phenomenon controlling adsorption.
With the inorganic soil, the smallest size fraction (less than 0.105 mm)
had the highest arsenic concentration (420 mg As/kg). When weighted
according to soil particle size distribution, this fraction contained
approximately 22 percent of the arsenic. The fraction between 0.354 and
1.00 mm contained about 40 percent of the arsenic. However, this result
can be explained by recognizing that this size fraction represented 45
percent of the total soil being tested. The higher sorption onto smaller
particle sizes is more consistent with expected results, although it is not
consistent with the sorption found with phenol. The difference could be
caused by the difference in testing methods. For phenol testing, the
individual soil fractions were exposed to the phenol and then analyzed.
For arsenic, the entire soil mixture was exposed, then fractionated and
analyzed. The tech- nique used for phenol would encourage sorption onto
the large particles since there was no other sorption media available.
Conversely, placing the arsenic onto mixed soils would allow preferential
sorption to high external surface area particles of small size and less
sorption to internal pores of the gravel.
49
-------
Water-Knife Stripping/Soaking Tests
The basic equipment operating methodology was established in previous
testing. These results indicated that the water-knife apparatus was
effective in breaking down clay lumps when a sequence of soaking and water-
knife stripping was used.
Chemical tests were then required to establish the efficiency of the water
knives in stripping contaminants from large soil particles (>2 mm). Pre-
liminary chemical testing indicated that the submerged (soaking) cycle was
also important for effective removal of contaminants sorbed to larger
particles.
Test Methods—A 1000-g sample of contaminated soil was placed into the
rotary screen and stripped by the water knife with clean water for one
minute. This allowed removal of contaminated soil particles that were less
than 2 mm in size and also stripped any contaminant present in the void
volumes. The percent-removals for treatment-efficiency calculations were
based on the amount of contaminant sorbed to the large soil particles
measured at this point in the testing. The bucket containing the larger
than 2 mm soil particles was then submerged in rinse water for time periods
that varied from 15 to 120 minutes. The screen was rotated while it was
submerged to encourage solvent/soil contact. Samples were collected at
various intervals, rinsed, and analyzed. After 120 minutes, the entire
contents of the bucket were rinsed with a low pressure nozzle for one
minute, and samples were collected.
Water Knife/Submerged Washing Test Results--Water-knife test results for
soils contaminated with phenol, arsenic oxide, and PCB are presented in
Table 9. The percent-removals were calculated after the water knife had
stripped the contaminant from between the void spaces and removed particles
less than 2 mm. Therefore, the results represent actual treatment of the
sorbed materials on the particles greater than 2 mm.
Treatment using water-knives and submerged washing was most efficient for
phenol (Table 10). The removals were highest for the inorganic soil since
almost 98 percent of the phenol was removed from the soil after a 15-min
submergence time. Neither arsenic nor PCB was stripped with the same ef-
fectiveness even after a 120-min submergence. Removals of phenol from the
organic soil were lower, probably because of the stronger affinity of the
organic soil for the organic phenol. This trend is not apparent in the
arsenic results, which is as expected since arsenic oxide is an inorganic
contaminant.
Residual amounts after varying submergence times do indicate continuing
removal with time; however, the amount of contaminant removed per unit time
becomes less and less. This indicates a decreasing benefit with time and
has an impact on the final design considerations. The usefulness of the
final low velocity rinse is seen when the residual concentrations at 120
minutes are compared. The final rinse water appears to remove or replace
contaminants retained in the washing water that remains entrained in the
void volume. A much cleaner product results from the final rinsing.
50
-------
UJ
u
1—
Q.
S
g
u.
l/> _
is
|S
* CM
°lfc
S*
|l
is
^ "*
2*
UJ *
u. ••
UJ
i
en
S
CO
u
«
I
u
1
WI
i
Is
u S
cu 1
B. i-
*
o o
J*
£1
§5
£ E
It
•
o e
«4<
AJ ^»
^
i
e .
•
u
U J<
i*
|
u
1
1
£
2
1
•» O
CM in
i
O O O
V ^ *~
.— — OJ
en ^
CO CM
IV. CM O
I* • *
CO ^* ^p
Icn ro w
w M *"
^ CM
Ok <7k
°. •. *.
CO •» «•
in P«. 10
I I
1 9i 5»
'Is
**-^ 5 * s *
^"oT 3 e a e
i! T1 «r
- u -sii •=•§
« oi _e « _e «
S_c T "5 f "5
— "5 £ » ° *
u
"s
1
«
CM
O
J,^
CM
CM
f^
i
CO
3
00
CM
1
||
e «
f*
IO
CO
CM
O
o
^_
CM
in
O
3
—
si
•~
si
•o
S.
1.
3 WI
wi e
•w
"• *
§1
en
2
^_
i
o
A
00 f»
00 rv
O
^J » .CM
— 5 <•?
00 l>-
in o
en 55
§g
v .1
- | S
«
?S5 , S,
15 - fc
3 *" fc— J S
WI *• *» 0) 3 C
01 O "S C """u
*» i — •• r* OI
3 ** k 4-> •Q
•J * _« OI _C 10
Ss S| si
u
10
e»
CO O O
in W en
in in in
i 1 .1
CM O IO
2* S d
fv QO en
2 2 S
P» in en
eo in n
^ 1 1
CA en ^
K 01 G 01 2 01
2 wi 2 wi s wi
3 e 3 e wi e
WI •»- WI **• ^
1. L. OI U
01 41 *»
*»-0 *• -O 3-O
3 e 3 c c e
G X 10 X — *
m
CM
in
CM
~
•-
00
eft
oo_
IO
?$
(A -M
l|
g |
51
-------
TABLE 10. PHENOL DISTRIBUTION IN SOIL SAMPLES USED FOR
WATER KNIFE AND SUBMERGED WASHING TESTS.
Soil type
Inorganic
Organic
Phenol
concentration,
units
mg/kga
mg/sample
mg/kga
mg/sample
Soil fractions
Complete
soil
48,300
42,700
452,200
254,410
<2 mm
710
626
196,240
110,400
>2 mm
10,770
9,520
18,210
10,250
Soil void
spaces
36,820
32,554
237,750
133,760
a Values listed are mg of phenol in a given soil fraction per kg of
complete or non-fractionated dry soil.
Final concentrations of contaminants in the soils varied significantly.
Treatment of the phenol-contaminated inorganic soil resulted in the best
removals and almost the lowest, post-treatment contaminant concentrations.
Concentrations in the organic soils were much higher, however; the mass of
soil involved was considerably less. Only five percent of the organic soil
was within the size range of 2 to 12.7 mm treated by the water-knife.
Approximately 50 percent of the arsenic was removed during this processing
step for both soil types. Residual arsenic concentrations for the in-
organic soils were an order of magnitude less than those on the organic
soil, because of the higher initial dosages onto organics and the strong
sorption of arsenic oxide to the wood chips and vermiculite present in the
organic soil. Residual arsenic concentrations were not significantly
changed after 30 min of submergence.
PCB treatment removals were the poorest among all the chemicals tested.
The initial concentration was low (only 14 mg/kg) and resultant residual
concentrations of approximately 10 mg/kg resulted in only 28 percent
removal. The residual concentrations were the lowest for all the chemicals
tested. Because of the previously noted trend of even less effective
treatment for organic soils, especially with phenol, additional testing of
PCB removal on this substrate was judged to be unnecessary.
When evaluating these results, it should be emphasized that the water knife/
submerged washing process is performing two functions:
1. Removal of entrained contaminant and separation of contaminated soil
particles less than 2 mm from larger soil particles.
2. Removal of contaminant sorbed on the soil particles greater than
2 mm.
52
-------
Removal efficiencies presented in Table 9 address only the latter functions.
If the removals had been calculated based on amounts of contaminant present
on the total mass of soil, the percentages would have been much higher
because of the relatively high amount of entrained contaminant. Using this
overall calculation approach would not have allowed as effective a
comparison of actual treatment efficiency among contaminant types. How-
ever, the benefits of the water-knife apparatus as a separation device
should not be overlooked. Entrained contaminant is transferred from the
soil to the water phase using the water-knife. The amounts involved can
range from 50 to 75 percent of the contaminant present, which immediately
results in a high removal and meets certain goals of the scrubber. Fine
soil particles are separated in this step, which also "removes" the
contaminant from the surfaces of the larger particles. These processing
benefits are crucial to the operation of the unit, although they are not
readily apparent when one evaluates only percent removals.
Removal of Entrained Chemicals—The importance of the removal of the
contaminant present in the void spaces and sorbed to particles less than
2 mm is further illustrated when the distribution of phenol onto the two
soil types is further evaluated (Table 10). This relationship in Table 10
was calculated using the results of the fractionation tests and the amount
of phenol dosed onto the soil (see Table 7).
It can be seen in Table 10 that, for both soil types, most of the phenol
was entrained in the void spaces. Although a greater percentage of phenol
was entrained in the void spaces of the inorganic soil than in the organic
soil, 76.2 percent ((36820/48300)x!00) versus 52.6 percent
((237750/452200)xlOO), the entrained quantity of phenol in the organic soil
was six times greater than (237750/36820) in the inorganic soil. Since
these calculations are based on maximum sorption conditions from the
fractionation tests, actual amounts entrained may be even higher. Large
particles (greater than 2 mm) in the organic soil contained a large
fraction of the phenol (22.3 percent) while this same size fraction
contained only four percent of the phenol in the inorganic soil. This is
due to the fact that there only was a small fraction of large particles
(5.7 percent of the total) present in the organic soil.
CHEMICAL EXTRACTION TESTS
These tests involved contaminant extraction from small soil particles
(<2 mm) using a water plus additive solvent system to:
1. Determine the most effective solvent for each contaminant.
2. Establish potential removals of each contaminant using a
countercurrent extraction system.
Two types of tests were run: single extraction tests for solvent screening
to establish single step removal efficiencies, and multiple extraction
tests to measure the improvements that sequential extractions offer on
removal efficiency and solvent usage.
53
-------
Test Methods
For single extraction tests, a Barrel! shaker with a 500-ml flask was
utilized to mix the contaminated soil and solvent (total volume of 200 ml
and dry soil solids concentration of 2.5 to 5.0 percent) for three hours.
The mixture was separated by centrifugation and the solvent analyzed for
contaminant.
The procedure for multiple extractions was identical to that for the
single-step extraction tests except that the contaminated soil was
extracted four times instead of once and each aliquot of solvent was
recycled in a four-step countercurrent direction (Figure 13). The final or
fourth extraction step for the soil was accomplished with clean solvent
while the first extraction step used the most contaminated soil and solvent
indicative of a countercurrent system operating at steady-state conditions.
The multiple extraction tests were further expanded to simulate the con-
taminant removal attainable with a full-scale soil scrubber chemical ex-
traction system. Procedures for these tests were similar to those used for
the four-step extraction tests; however, the froth flotation unit was used
as the mixing cell and settling was used to simulate the soil/solvent
separation of a hydrocyclone. The volume of the soil/solvent mixture used
in these tests was three liters.
Extraction Test Results
Single Extractions—The results of the single step solvent extractions
using different solvent systems are summarized in Table 11. Solvents
included water only and water plus additives that appeared suitable for the
particular contaminant. In these tests the emphasis was placed on removal
into the solvent and no additional rinsing of the soil was performed. It
was the purpose of these tests to provide a maximum removal into the
solvent by maintaining the complete mix for three hours. The
solvent-to-soil ratio was high so that solubility limitations were not
encountered and good contact between solids and fluid could be maintained.
Phenol—Removals of phenol from inorganic soils into the solvent phase were
most effective. Removals with all of the solvent types were very high and
the best extraction occurred with the inorganic soil using water as the
solvent. Because of possible experimental errors, the differences among
solvents for this soil system were probably not significant. The inherent
advantages of using water alone in a soil system (e.g., cost, operational
simplicity, lower corrosion potential) make this the preferred solvent.
Phenol removals from the organic soil were not so high but large quantities
of phenol were also extracted into all of the solvents tested. In this
soil, the most effective solvent system was water with sodium hydroxide
added to maintain a pH of 11. However, the handling difficulties
associated with this rather alkaline solvent system could be a major
disadvantage. Water alone, with a removal efficiency not significantly
less than the alkaline solvent, would probably be a suitable choice for the
organic soils also.
54
-------
'«**»»
O)
o
o
cu
O)
>
-u
s
IS)
OJ
i-
0)
01
c
S-
3
T3
c
HI
o
c
O)
-a
c
IT)
o
CO
OJ
S-
-------
TABLE 11. SINGLE EXTRACTION TEST RESULTS
Contaminant
Phenol*
Phenol b
Arsenicc
Arsenicd
PCS*
PCS*
•Initial soil
^Initial soil
cinitial soil
^Initial soil
•Initial soil
'Initial soil
Soil type Solvent type
Inorganic Water + HgSOa
Water + H2S04
Hater + HjSO*
Water + NaOH
5* MeOH
IX Tween 30
1* MYRJ 52
Water
Organic Water + HjSOa
Water + H2S04
Water + H2S04
Water + NaOH
5X MeOH
IX Tween 80
IX NYRJ 52
Water
Inorganic Water + HjSOa
Water + NaOH
Water + NaOH
7.5X NaHSO*
5. OX NaCl
Water
Organic Water + HgSQ*
Water + NaOH
Water + NaOH
7.5X NaHSQi
5. OX NaOCl
Water
Inorganic 0.1X Tween 30
1 .OX Tween 80
water
Organic 0.1X Tween 30
l.OX Tween 80
Mater
dose was 48.3 mg phenol /g dry inorganic
dose was 452.3 mg phenol /g dry organic s
dose was 0.75 mg As/g dry inorganic soil
dose was 5.0 mg As/g dry organic soil.
dose was 3.0 mg PC8/g dry inorganic soil
dose was 25.6 mg PC8/g dry organic soil.
Initial
PH
1.0
3.0
5.0
11.0
6.7
6.4
6.3
6.9
1.0
3.0
5.0
11.0
7.1
7.1
7.1
7.1
1.0
10.0
12.0
1.0
10.0
12.0
soil.
oil.
Supernatant
contaminant
concentration,
nKj/1
1,120
1,100
,100
,060
,100
,120
,180
,190
16,300
18.000
18,600
20,000
18,600
16,200
15,300
17,600
32
16
18
23
22
16
425
310
405
405
285
375
61
110
72
850
366
418
Percent
removal
92.8
91.1
91.1
87.8
91.1
92.8
97.8
98.6
72.1
79.6
82.3
38.4
82.3
71.6
67.7
77.8
85.3
42.7
48.0
61.3
58.7
42.7
85.0
62.0
81.0
81.0
57.0
75.0
20.8
37.5
24.6
20.8
33.8
48.3
Soil residual
contaminant
concentration,
mq/g
3.48
4.30
4.30
5.39
4.30
3.48
1.06
0.68
126.2
92.3
80.1
52.5
92.3
128.5
146.1
100.4
0.11
0.43
0.39
0.29
0.31
0.43
0.75
1.90
0.95
0.95
2.15
1.25
2.38
1.38
2.66
20.3
19.5
13.2
56
-------
Arsenic—The maximum removal of arsenic into the solvent phase was less
efficient than for phenol when inorganic soils were extracted (85.3 percent
as compared to 98.6 percent) but approached the extraction efficiency of
phenol on the organic soils (85 percent as compared to 88.4 percent).
Arsenic removal from organic soils was significantly affected by the type
of solvent used; water with sulfuric acid appeared to be the best solvent
(85-percent removal at pH 1.0) whereas water alone or water plus NaOH
removed less than half of the arsenic. When acid is used as a solvent in
the full scale system, the equipment used for chemical extraction must be
constructed with suitable corrosion-resistant materials. Care is also
required in the handling and use of the acid. When acid is not used, the
next best solvent for the inorganic soils is the sodium bisulfate (7.5
percent NaHS04) with a 61.3-percent removal.
With the organic soils, the differences among solvents were not so
dramatic; and water alone, as well as water plus NaOH or NaHS04, produced
results almost as good as the water plus sulfuric acid.
Another concern with the arsenic results was the low concentrations of this
material measured in the solvent. The solubility of arsenic in the solvents
tested is many times higher than the levels obtained during any of the ex-
traction tests. This result implies, at best, the need for large volumes
of clean solvent—and possibly longer contact times to better approach
equilibrium—to reduce residual soil concentrations to lower levels. The
result also raises questions about the extent of further reduction in soil
arsenic levels with multiple extractions. With lower residual soil
concentrations, solvent volume requirements may be even higher. It was
also significant that the concentrations measured in the different solvents
did not vary greatly. These relatively small differences are important,
however, when comparing extraction efficiency by solvent type. Because of
the relatively small amount of arsenic in the soil system, the measured
differences in solvent concentrations do result in large differences in
removal efficiencies.
Problems occurred with the analysis of arsenic contaminated soils, particu-
larly for the inorganic soils. A strongly acidic solution was required to
obtain total digestion and reproducible results. Digestion of soils was
attempted with nitric acid alone, mixed nitric and hydrochloric acids, and
nitric acid with hydrogen peroxide.
Use of an extracting mixture of nitric and hydrochloric acid and
quantitation by atomic absorption spectroscopy yielded reproducible arsenic
analyses. However, the difficulties in analyses and the requirement for
such a rigorous digestion illustrate the potential difficulties associated
with complete removal of arsenic into a more moderate solvent system, which
could be used in the soil scrubber. Even a mixture of water and acid at
pH 1, which gave the best comparative removal, does not provide the optimal
type of extractant that is needed for complete arsenic removal.
57
-------
PCB--PCB removal into the three solvent-types tested was quite low. A
maximum removal of 48 percent was noted for the inorganic soil. It should
be observed that the PCB concentration in the supernatant greatly exceeded
the solubility of PCB in any of the solvent mixtures. This indicates that
the shaker tests were actually performing a physical separation rather than
a purely solvent extraction. PCB that had been entrained in the soil was
removed and actually formed a third phase, which was noted by the
laboratory analyst. Because the procedure used in the laboratory testing
does not provide the longer times necessary for soil/chemical bonding, the
test results are difficult to apply to remedial-action type situations,
which are probably those situations for which scrubbing of PCB from soils
is most needed. In those situations, much of the entrained PCBs would be
gone and the remaining contaminant would be more tightly bound to the soil
itself. Depending upon the type of soil and many other factors, the
remaining PCB concentration may be either high or low. Lab-scale treatment
of PCB in a remedial-action-type situation would have to be modified to
accurately predict the impact of the critical time factor-on removal
efficiencies.
Multiple Extractions—Phenol was selected for establishment and
demonstration of the efficiency of multiple extractions and its effect on
solvent usage. Arsenic was not used for the demonstration because it
appeared that further extractions would probably not.be effective in
reducing the soil arsenic levels substantially because of the small amount
of contaminant that the solvent was able to remove in a single extraction
and the awareness of the tight bonding of the arsenic to the soil, as shown
by the difficulty in removing arsenic from the soil by acids for analysis.
PCB was not used for the demonstration because of its low efficiencies of
initial removal. Its hazards also detracted from using it for this initial
demonstration.
The first test run with phenol was a quick check to demonstrate the po-
tential for solvent reuse. Two aliquots of soil were contaminated with
phenol and extracted; the same volume of water was used for each extraction.
In one experiment, an aliquot of soil (# S-l) was washed twice with clean
water (# W-0) and the washings (# W-l and W-2) were saved. In the second
experiment, a fresh aquilot of soil (# S-2) was first washed with used wash
water (# W-2) from the second washing of the first aliquot of soil (# S-l)
then rewashed with clean water (# W-0). The results indicate that used
solvent could remove nearly as much phenol from contaminated soil as clean
solvent, so solubility limitations did not impede removals in this case.
The more detailed four-step, countercurrent extraction tests were then
conducted.
The results show that, for phenol, multiple extractions were far more
effective than a single extraction in removing the chemical from the soil
(Table 12). Multiple extractions continued to remove phenol at high
efficiencies so that after the fourth extraction the percent of initial
phenol retained by the soil had been reduced to a small fraction of a
percent. A log-normal plot of retained phenol versus number of extractions
indicates that the efficiency of extraction had not yet diminished sig-
nificantly after four extractions and that further extractions could be
58
-------
i
P
H-
X
^J
^C
i
UJ
1
s
u.
UJ
o
IT
u.
Ul
s
X
g
X
Q.
^*
>•
S'
1
t/>
CM
UJ
Si
JE
*o
S
44
e
01
tn
g
1
O £
g •Jj
— .O
10 i—
° —
||
II
•— i 1/1
^3
44 -O
C V
O) >
u o
S. E
01 Z
CL U
01
1
in
i
1
si
W f"
u 10
t. 44
01 01
a. «-
•4-»
5
** *-» "^
— ^ ^ to j£ CM i— ro co 91 1— ro ^ in 9*
<— S Oi <-^ «— O^ ^"OICM
S
u
o 25
^^i • • •
i^^ ^* i^^ un ^— co *• co co ^? tp 13 ^y ^s
** IOCOCM O O in CM
CM 9i^ un
CO CM
91 CM
*•* • •
in o in in in^ omin9iO9QO
CM O 91 C
1 11
r*> 01
«-*^ u f • I. 44
fo mi— e i.
in «> i-
§jir ^
U 44 C 1- VI
• 01 O
44 f e > m >»
WI44 O— 1.
ft Ol 01 « O>«
44 44 S. Ol O >
« o e Si—
f « Z- O
44 I. f f «x m
•^ 44 u 44 01
X 44— E Ol
•a « x in e
•o e J3 api-
a* « -o «Sr-c
44 44 01 C
IOCM in x • Q
* c .c
1*. «1- 44 o 01
wo •— vi "et—
10 oii- a «
X44 44 X 01
« e t. •
— C 3CM O-—
*- Ol c i-
^h *^* C XI "^ O •
-^ in o c m vi
01 iO .. 01
•—•oo at x —
•>" f^ in u CM > -r-
fO a CM * C *^ ^** 44
C X4h4k 01 Ol O U
••- 3 in in «
S"— 44 VI OT»> 1.
«— O44
44 ai e V » 01
u'o'oi^ "5jc — — 01
VI VI O 4444 O O 44
44 in «j o in in 10
in^^ •> 43 i. u
o -am .c 14. e S —
E oi«i44 «n^ o 5 —
i— i~ U t. O.
oi44i— x aif-- 01 •*- 3
£44— 4CCOIC T3
44OIO-O 44«in — T3
in in 01 9II-. e ai E
* o'j: J. (.CM i- > o
ink in o o--. s o «.
e Of * <»- _j g 6 <*.
o— o s ••LD S at
•«- 44 oi"- \o i- u ai
UOIJ3I— —"OCM »- •— io
10 01 i— in in o o
(.34410 o .. e e in
4J t- X — T3 >, O> 01 Ol —
44 01 "3 i— J=
•w c 01 3 e 91-. -a -o 44
oij=er oiin_i 01 01 c
<»-cj44ai ft— in in in ai
o in CLi— 1>». o o t.
V~~-l.v "10 in ^ <4- i*. °~
«•»<*. o> —CM •• o o c
•*» 43 44 • 91 —
f •*^*O — 1^ O 44 44
uoi-ocoecin
10444410 — • ,= 6 V 01
OlCiOf "^^U U 3
Oil. Olfi— k i- •—
e><044 .c 3 s « 01 a
7ssl f e| * * *
59
-------
expected to further reduce the phenol content from the soil (Figure 14).
W
Reduction in soil phenol resulted in a residual soil product concentration
of about 30 mg/Kg after four extractions using the shaker test and
centrifugation.
The phenol present in the solvent increased progressively with the number
of extractions, however, the majority of the phenol was removed in the
first stage by contacting the contaminated soil with previously used
solvent. The incremental increases in phenol concentration were much lower
in the subsequent soil extractions. Fresh solvent was used in the last
(4th) extraction.
When these results were further evaluated, it was determined that the re-
moval of phenol after three extractions with recycled solvent was
equivalent to the removal provided by two clean water extractions. Four
extractions provided even better removals than two clean water extractions
with half the water requirement.
The use of froth flotation and settling as an extraction apparatus to estab-
lish the impact of incomplete soil/solvent separation as would occur with
the use of hydrocyclones resulted in a similar trend of reduction of
chemical level in the soil, but with somewhat less efficiency. Removals
were more efficient from inorganic soils than from organic soils, as the
previous single extraction test data had predicted (Table 11).
ft**
W However, the final concentrations of phenol in the soil products are much
different for the two soil types (79 mg/Kg for the inorganic soil and 2560
mg/Kg for the organic soil) due to three factors:
1. The organic soil was dosed with approximately ten times more phenol
per gram of dry soil than the inorganic soil.
2. The concentration of dry soil in the froth unit for the organic soil
was 10 percent versus 5 percent for the inorganic; therefore, there was
less solvent per gram of dry organic soil.
3. The organic soil retained the phenol more tightly because of the
nature of the soil.
An additional factor that should be considered in examining the results is
the loss of a portion of the soil in the overflow (simulated). For the
inorganic soil, approximately 25 to 30 percent of the soil dry weight was
lost in the simulated overflow during the four-step extraction process.
Losses of up to 40 percent were observed in the overflow during the organic
soil extractions. Although settling was utilized to simulate soil/solvent
separation, these soil-loss values should be indicative of the expected
soil loss when hydrocyclones are actually used in separation/processing.
60
-------
100 (T
-4
M
O
<
H
r^t
5
a:
eu
10
1.0
0.1
0.01
FROTH FLOTATION AtlD SETTLING
O = ORGANIC SOIL
E = INORGANIC SOIL
SHAKER AND CENTRIFUGATION
A= INORGANIC SOIL
EXTRACTION NUMBER
Figure 14. Soil phenol reduction as a function of number of extractions
using two different extraction methods.
61
-------
SECTION 6
IMPLICATIONS OF THE PRELIMINARY LABORATORY TESTING
PROGRAM ON SYSTEM DESIGN AND SOIL TREATMENT
INTRODUCTION
The information obtained from the preliminary evaluation and the laboratory
testing allowed an initial evaluation of the soil scrubbing concept and
provided a basis for the design. Since initial laboratory testing was
limited to three different chemicals, broad generalizations regarding the
scrubber's range of application were difficult to establish. Instead, the
results were considered indicators of the effectiveness and usefulness of
soil scrubbing on the particular chemicals tested and the classes they
represent. The same concept holds true when extrapolating the impact of
different soil types on the scrubber unit. Implications on system design
based on the equipment review and the laboratory testing are presented
below, followed by implications from the chemical testing.
IMPLICATIONS FROM EQUIPMENT TESTING
Water-Knife Stripping
Physical stripping of both free contaminant and that bound to soil
particles using a water-knife system was found to be effective. Breaking
down small, cohesive, clay lumps can be accomplished using water knives in
conjunction with a soaking cycle.
The most effective configuration was found to be an internally water-knife
washed rotary screen unit which passed particles less than 2 mm in size and
provided the action to flush (strip) contaminating chemicals from the
particles greater than 2 mm in diameter that were retained in the screen
chamber. This system allowed the retained soil to be confined in a pocket
so that the force of the water knives could be fully utilized. A
three-step operation, which included an initial stripping with the water
knives to remove fines and entrained contaminant, a submergence step for
lump softening and treatment, and a final rinse, was found to be
effective. Recycle of water for the first high pressure stripping appeared
desirable to reduce water requirements in the system. The final rinse
utilized a lower pressure, clean water stream to remove contaminated water
from the soil surfaces.
Froth-Flotation Mixing and Contact
The froth-flotation system provided smooth, efficient mixing using a com-
bined mechanical- and air-mix system. The two mechanisms were interrelated
with optimum settings established by soil type and solids concentrations.
62
-------
Solutions containing 30-(or greater) percent solids were effectively mixed
with this system. The amount of froth generated in clean soils was
minimal. The process limitations of applying froth flotation to the
scrubbing system, however, include the added difficulty that would be
encountered in controlling emissions from volatile contaminants and the
possible foaming problems when surfactants are used as additives for
chemical extraction.
Hydrocyclone Liquid/Solids Separation
Hydrocyclone testing also had a direct impact on the scrubber evaluation.
The equipment operation was quite simple, and only a manual apex adjustment
was needed to control solids concentrations in the underflow.
With a closed or tight apex setting, the solids concentration in the
underflow ranged from approximately 30 to 50 percent. This concentration
should be acceptable for discharge of cleaned soils, possibly with some
auxiliary drainage on sand beds.
Mounting hydrocyclones above the mixing tanks of the modified oily beach
sand cleaner was no problem and the units are capable of operation over a
relatively wide flow range. Another advantage is that changes in the flow
rate do not have adverse impacts on the solids concentration in the
underflow. The major disadvantage of the hydrocyclone is the amount of
solids that is entrained in the overflow of the unit. Solid recoveries of
70 percent were measured; however, 30 percent of the solids passed out of
the system with the overflow. The solids concentrations in the stream
ranged from 0.7 to 3.7 percent. Treatment of this stream with a similarly
sized hydrocyclone was ineffective. It is likely that an additional, more
efficient hydrocyclone or that chemically assisted gravity separation will
be needed prior to any additional treatment for solvent recycle.
IMPLICATIONS FROM CHEMICAL TESTING
The results of the different tests performed during the preliminary
laboratory studies had an impact on the design and operation of the soil
scrubber. An evaluation of their effects are presented in the following
section.
The chemical test results were somewhat inconclusive since a defined system
endpoint has not been established for the chemicals tested. The most
effective treatment, in terms of percent removed, was found for phenol.
Neither arsenic oxide nor PCB was scrubbed with as high a removal
efficiency. In all cases, the residual soil concentrations were reduced;
but, without established, acceptable concentration limits, the results are
difficult to evaluate.
Contaminant Distribution
Varying sorption with particle sizes was expected; however, the results
obtained during the testing were not so straightforward as anticipated.
The larger soil particles in both the organic and inorganic soil sorbed a
greater fraction of contaminant than would be predicted based on nominal
63
-------
surface area. This result can be partly explained by the open structure of
the coarse, woody fibers in the organic soil and by the dryness and pore
structure of the limestone gravel in the inorganic soil.
On the basis of these experiments it does not appear fully justified to
assume that small particles will adsorb more contaminant per unit weight
than large particles. A number of variables besides particle size appear
to impact how much contaminant sorbs to a particular soil including:
1. initial moisture content of the soil
2. internal structure of the soil particle
3. nature of the soil material
4. manner in which the contaminant contacts the soil
5. duration of contact time between contaminant and soil.
It is difficult to predict soil/chemical interactions, which emphasized the
need to make the soil scrubber as flexible as possible.
Effectively treating soils by concentrating on a predetermined soil
fraction is also difficult. With the extreme variations of soil gradation
and of chemical sorption to different soil materials, an inflexible
approach would significantly limit the scrubber's effectiveness and
usefulness.
Contaminant distribution information, nonetheless, is still very useful
because the data may be applied to estimate the overall scrubber
effectiveness when the contaminant sorption by particle size and the
particle size distribution of the soil are known. Both large rocks and
boulders and extremely fine clays will escape treatment because of system
limitations. Knowing the fractional distribution of particles between
these two limits permits calculation of a theoretical system efficiency.
Water-Knife Stripping
The submergence step, which has been found to be necessary for both
softening lumps and cleansing of the larger particles, was studied. A
submergance time of 15 min provided good removal without greatly slowing
down the processing rate. However, the differences in soil and contaminant
types emphasize the need for flexibility in system operation. The three-
step approach, i.e., initial stripping, soaking, and rinsing, was
effective; however, the final rinse prior to discharge had to be performed
with low-velocity spray jets for the most efficient removal of the
contaminant.
Chemical Extraction
The single solvent extraction tests provided a way to compare solvents and
to establish preliminary removal efficiency ranges for soil treatment. The
contaminant tested that had the best results in these tests (phenol) was
further tested in a multiple extraction scheme, which simulated the
four-step extraction system proposed for the prototype soil scrubber.
Multiple extractions resulted in continued, efficient removal of phenol
from soil to low residual values. In these tests, solvent reuse in a
countercurrent scheme was shown to be an effective approach.
64
-------
Other Considerations for Chemical Treatment
W
Residual Soil Contaminant Levels—A second and possibly more critical
factor that must be further addressed involves the residual contaminant
levels remaining in soil after treatment. This factor is of particular
importance when one evaluates the preliminary laboratory results.
Acceptable soil concentration limits are not always available, so it is
difficult to determine whether the treatment has been "successful."
Calculation of percent-removal is one way to begin the evaluation, but
these numbers can be misleading since different initial amounts of material
were used. Therefore, it was necessary to reassess the results from a
slightly different point of view.
It became apparent that the soil scrubber system itself might be able to
establish an acceptable upper limit of residual contaminant concentrations
in the soil. Since the scrubber is designed to remove materials into a
water solvent under forced conditions, the residual remaining on the soil
should have limited mobility in the natural environment. It is expected
that the material left sorbed to the soil after the sequential extractions
will be very tightly held and have minimal environmental impact. The
release of the chemical may continue but at a rate so extremely low that
there is no significant environmental danger. If concern still remains
regarding the impact of a specific chemical, disposal to a sanitary
landfill or other site with more controlled access may be an acceptable
compromise.
""*«* The post-processing disposal option concept can be further evaluated by
considering the three chemicals tested in the laboratory study. Phenol
removals after step extraction were high (>99 percent for both organic and
inorganic soils) with lower residual concentrations on the inorganic soil.
These results were anticipated because of the organic-chemical-to-organic-
soil attractions. Since the phenol was not so effectively desorbed from
the organic soil, or from the inorganic soil it is apparently not so
mobile, i.e., it is more tightly held. This observation appears to confirm
the hypothesis that the chemical that is removable is scrubbed and that
which remains may be quite tightly bound to the soil. Using this concept,
the natural attenuation or treatment capacity of the soil then works in
favor of the treatment approach rather than against it.
The same concept is more graphically illustrated when considering the other
two chemicals tested, arsenic oxide and PCB. Arsenic has a strong affinity
for the fine inorganic soil particles, probably due to an ion exchange type
mechanism. It is also held quite tightly to the organic soil. The
relatively high residual concentrations of arsenic after treatment and the
relatively low arsenic concentrations in the solvent confirm the strength
of the soil/arsenic attraction. Another indication of this interaction can
be found when considering the rigorous analytical procedure needed for total
arsenic analyses. The soil sample was completely digested in a mixture of
nitric and hydrochloric acid before a consistent arsenic measurement could
be obtained. If arsenic is so difficult to remove during analysis, it is
**** not likely to be mobile in the natural environment. Studies of the natural
Nw attenuation properties of soils tend to confirm this result.
65
-------
Multiple extractions of soil contaminated with arsenic have not been run;
but it is expected that, unlike phenol, the extraction efficiency will
decrease substantially with each successive extraction. The probable
effect will be that, after only a few extractions, very little additional
arsenic will be drawn into the solvent and further treatment by soil
scrubbing will be marginally useful in effectively reducing soil arsenic
levels. If additional arsenic cannot be forced out of the soil by rigorous
stripping and extraction, it should be quite immobile in the natural
environment.
Treatment of PCB using the scrubber approach provides the most difficult
challenge for effective treatment. These materials have an extremely
limited water solubility and a strong affinity for soils. Laboratory tests
indicated that some PCB was removed into the water phase but much of what
was being processed remained in the soil. However, the initial sorbed
amounts of PCB were quite low since the lab approach, simulating a spill
situation, allowed only a 24-hour contact time between contaminant and
soil. Use of higher initial levels of sorbed PCB may show better percent-
removals.
Considering the hypothesis that soil treated with multiple extractions to
the point such that no further removal would have only a minimum impact on
the environment, application of soil scrubbing to PCB-contamination
problems may be promising. Available disposal methods for PCB-materials
are limited to landfilling in specially approved sites, incineration,
and—possibly--some novel processes. Treatment with the soil scrubber
prior to disposal may allow relaxation of some of the originally applicable
restrictions. It may be possible to scrub the soil and then dispose of the
residuals in a sanitary1 landfill or in a less-than-totally-secure toxic and
hazardous site. If the concept of repeated scrubbing to reach a stable
soil residual is accepted, the usefulness of the soil scrubber could be
almost unlimited.
FURTHER LABORATORY TESTING
Two types of soil contamination exist that can be handled by the soil
scrubbing system. The first is a spill situation that involves an
immediate response to a freshly spilled chemical-onto-soil system. In this
situation, a large amount of contaminant may be physically entrained and
only a small fraction actually sorbed to the soil. The goal is then to
promptly remove the contaminant into the water phase and replace the soil
at the site. The second type of soil/chemical problem is already existing
contamination arising from improper chemical disposal. Usually this
situation is a longer term problem so that a much larger portion of the
contaminant is tightly sorbed to the soil. The goal in this instance is
also removal of the contaminant. However, it may be acceptable to mainly
remove the mobile fraction and then dispose of the residual soil in less
rigorously designed landfills.
The laboratory work concentrated mainly on the more immediate situations
and has been used to preliminarily evaluate the treatment of different
chemicals and soil types. Using this approach, treatment was found to be
most effective for phenol, a water soluble organic compound.
66
-------
Neither arsenic nor PCB were as effectively treated when considering percent
removals. From a removal efficiency standpoint, the soil scrubbing system
will probably be most useful for chemicals that have a higher water
solubility and a lower affinity for soil than PCB's or arsenicals.
To confirm this result, some additional testing using chemicals with a
higher potential for effective treatment should be undertaken. Possible
candidates include certain aromatics and heterocyclics, medium chain length
organics, and metals (copper, nickel, cadmium, zinc) in cationic states.
The program requires an abbreviated version of the preliminary testing
performed to date. The same two soil types can be dosed in a manner to
simulate spill conditions. Then, a water-knife test program can be
undertaken using a set submergence time of 15 or 30 minutes. Finally, the
extraction tests can be performed to select solvents and, when appropriate,
to evaluate countercurrent extraction. The data obtained from these tests
should give a broader indication of the potential range of application of
the scrubber for spill and release (remedial) situations.
Another approach to consider involves further investigation of the concept
of stabilizing the contaminant level in the soil by scrubbing. To
accomplish this, the laboratory approach should be modified somewhat.
Existing contaminated soil samples (rather than newly dosed samples) should
be utilized in the testing so that the impact of time on soil sorption can
be evaluated. Since both PCB and arsenic oxide represent existing soil
contamination problems, either or both of these contaminants could be
tested. Once samples are obtained, an initial analysis of the contaminant
concentration should be made. Then, samples with a range of contaminant
amounts can be tested. Some preliminary soil characterization will also be
needed. Typical soil testing for particle size, organic and water content,
and Atterburg limits would be needed to broadly establish soil types.
The actual tests would be similar to those performed during the preliminary
laboratory study. A water-knife test, using the three-stage treatment
approach, is the first step. Then, material less than 2 mm can be tested
for extraction effectiveness. PCB solvents screened can include water and
water plus Tween 80 or other detergents/surfactants. If arsenic is
evaluated, then redox solvents may be most promising. The extractions can
be performed using successive amounts of solvent until a constant
concentration in the solvent is achieved. Then the treated soil can be
exposed to the extraction procedure (EP) specified under the Resource
Conservation and Recovery Act (RCRA).
When the extract from the EP does not contain contaminants in concentra-
tions more than 100 times the primary drinking water standard, the soil may
possibly be classified as non-hazardous. In a full-scale system, disposal
to a sanitary landfill or even replacement at the site may then be
acceptable, depending upon the situation.
When EP is an acceptable indicator of the soil's contaminant level, this
measurement will provide a criterion from which to judge the effectiveness
of soil scrubbing. Therefore, further evaluation including the EP
procedure would be extremely valuable in assessing the feasibility of the
soils-scrubbing concept for any type of soil/chemical contamination.
67
-------
SECTION 7
DESIGN AND CONSTRUCTION
INTRODUCTION
Upon completion of the laboratory experiments, the treatment sequence for
the mobile soil scrubber was finalized. Following is a description of the
flow scheme and a discussion of some of the major design considerations.
Photographs of the fully constructed prototype soil scrubber components are
included within the text.
PROCESS FLOW SCHEME
The sequence of operations for the soil scrubber is as follows (Figure 15).
Contaminated soil is deposited into a feeder through a screen that
separates oversize material (>2.5 cm). The feeder introduces the soil at a
uniform rate into the drum screen scrubber where water knives break down
soil lumps and mechanically strip chemicals from the soil. Following
initial stripping with recycled wash spray, soil particles greater than
2 mm, which are retained within the drum by the screen, enter the soaking
and final rinse zones of the drum before being discharged from the far end
of the drum and returned to the site of the excavation or otherwise
disposed of. Soil particles of this size dewater relatively easily, and
further dewatering beyond that which occurs at the outlet screen of the
drum screen scrubber is probably not required.
Particles less than 2 mm, which were not retained within the drum by the
screen, form a slurry with the washing fluids, which thickens in the bottom
of the wash-spray-collector troughs. This slurry is continuously withdrawn
from the troughs and pumped to the countercurrent chemical extractor for
further treatment. In the extractor, the soil is processed in a four-step
continuous sequence wherein the soil is contacted by washing fluid of
increasing purity. The concentration gradient between the chemically con-
taminated soil and the washing fluid forms the primary driving force for
transfer of contaminants into the washing fluid. Within each of the batch
chambers of the extractor, the soil/washing-fluid slurry is well-mixed by a
froth-flotation technique, which employs both mechanical mixing and fine
bubble aeration to achieve good contact between soil and washing fluid.
Hydrocyclones are employed between the four stages to partially/
sequentially separate the soil from the washing fluid. There is also a
hydroclyclone in the solids discharge line at the output of the treatment
system. Unlike the drum-screen scrubber, the product of the countercurrent
chemical extractor retains a portion of the final rinse water as it exits
the underflow of the last hydrocyclone. A drying bed may be necessary
before redepositing this portion of the soil at the site, depending on the
68
-------
_,
o
CO
Q
pa
CO
Z>
CC
0
CO
2.
CNJ
+
j
to _l
~ 0^
E Q_ ^C^^
s; to
CO
o
h—
Q
1- O
XO
1 1 1 Of
LI-
CC
CC LU LU
•CC OL < Q
< LU O
•z. i ^ — -J i-o
< u ~ <_) too:
Ul CO O
_i — cc < z:
<_> C3 — X O
l-
— u
El
(.
u
c
»~
u
<_
«:
3
0
»a
>
C
L
h
<
_^
^
£
J
J
J
J
J
:
«
3
J
3
1
«
L
f.
U
t
o
03
ty.
LU
Q_
CO
U
O
CL.
\f>
69
-------
characteristics of the site and the method for reintroducing the soil.
Likewise, spent washing fluids are also discharged through a hydrocyclone
(the overflow of the hydrocyclone in the effluent of the first washing
step). Since the hydrocyclone is not fully efficient in removing
particles, some of the very fine particles (especially clay) are carried
into the spent washing fluid stream. These particles must subsequently be
removed by chemical treatment, clarification, and/or filtration prior to
introducing the spent washing fluids to carbon adsorption columns or other
polishing units where the extracted chemicals are removed before the
washing fluid is recycled.
The froth-flotation method of providing intimate contact between soil
particles and washing fluids results in the production of a froth, which is
skimmed from the top of each of the four extraction chambers. This froth
is a treatment residual that must be ultimately disposed of, along with the
fines separated from the spent washing fluid.
Froth-flotation can also result in the sparging of volatile chemicals from
the wash water. The air discharge from the froth-flotation units, as well
as that from the drum screen scrubber (the drum is not a forced-air system
but can produce fugitive air emissions), is collected and processed, when
necessary, before discharge through a stack.
WATER-KNIFE DRUM SCREEN DESIGN AND OPERATION
Preliminary, small-scale testing during the first phase of this project has
shown that thin, intense jets of water from water knives are capable of
effectively breaking up even compacted clay lumps, especially when the
spraying cycle is combined with a soaking step. Stripping can be
effectively accomplished at a water jet pressure of 4.2 kg/cm2 (60 psi),
a low enough pressure to permit direct recycle of stripping fluids without
undue problems with nozzle clogging. Tests showed that effective use of
water knives also required that the material being stripped be constrained
to remain beneath the water knives. Laboratory work demonstrated that the
necessary constraint was achieved when the soil was rotated within a
drum-like screen.
Mobility Requirements
Initially, efforts were made to procure a commercially available,
modifiable rotary screen, that could be converted to accomodate stripping,
soaking, and rinsing. Standard trommel screens are expensive and did not
meet the design requirements. Extensive modifications would have been
required to meet the process objectives. The decision was then made to
design a rotary screen specifically tailored to accommodate the stripping,
soaking, and rinsing processes.
Since it is advantageous to process as much soil per hour as possible, the
largest, practical size drum-screen is desirable. However, since most
states limit semi-trailer size to 13.7-m (45-ft) long, 2.4-m (8-ft) wide,
70
-------
and 4-m (13.5-ft) high, the proposed trailer-borne drum-screen scrubber had
to fit within that size envelope for transport without special permits.
When allowance in the width is made for sprockets, hoods, drive unit, and
piping, the maximum allowable drum shell diameter is about 1.4-m (4.5-ft).
The drum-screen unit was thus designed for over-the-road transport without
special hauling permits.
After considering various means for transporting and mounting the unit, it
was concluded that the best choice was a skid-mounted unit designed for
transport on a drop-deck trailer with removal and placement on ground level
during operation. The overall dimensions are: 9.8-m (32-ft) long, 2.4-m
(7-ft, 10-in.) wide, and 2.4-m (8-ft) high. The empty weight is 6,340 kg
(14,000 Ib).
Drum Feeding
The loading system consists of a hydraulically lifted bucket, a feed
hopper, and a soil metering paddle wheel (Figure 16 and 17). Workers load
the bucket with a tractor-mounted front-end loader and rake the soil
through a coarse bar screen (ca. 7.5 cm (3 in.)) on the hopper that screens
out oversize material. Retained materials can be washed while still on the
screen using a hose sprayer, and then removed for final disposal. Soil and
liquid passing through the coarse screen fall onto a second bar screen (2.5
cm (1 in.)) where there is additional spraying and removal of chunks too
large to drop into the skip hopper. The skip hopper method permits the
feeding of soils with any fraction of liquid content, from dry materials to
slurries. When the bucket contains a sufficient quantity of material, the
operator elevates the bucket above the feed hopper and dumps it
hydraulically. This action allows precise control of the dumping rate.
The feeding of the hopper is thus under control to prevent overfilling and
spillage or clogging. The metering paddle wheel in the feed hopper throat
turns at 4 rpm when metering the maximum feed rate of 13.8 m3/hr (18
yd3/hr). (The variable speed, paddle-wheel-drive motor permits metering
lesser quantities of soil at lower feed rates.)
Drum Screen Construction
Once inside the drum, the soil moves in a controlled fashion from the feed
end to the discharge end through a combination of drum inclination and
rotational speed. The duration of the three cycles of stripping, soaking,
and rinsing is established by the length of drum devoted to each process.
The rotating drum screen soil scrubber unit itself is 6.4-m (21 ft) long
and 1.4-m (54-in.) in diameter (Figures 17 and 18). The initial spray zone
is a 1.2-m-(4-ft) length of reinforced screen HYCOR Contra-Shear™ that
can be backwashed. The next (soaking) section is 4.6-m (15-ft) long and
terminates in a baffle having a 0.56-m (22-in.) diameter opening through
which the soil and water pass to the rinse zone, which is 0.6-m (2-ft) long
and is similar in construction to the initial spray zone. Both screened
zones are equipped with water-knife nozzles. The initial spray zone is
fitted with 15 internal knife spray nozzles (13 gpm @ 60 psi) and 6
external screen backwash knife elements (10.6 gpm @ 40 psi). The soak zone
71
-------
•u o
JZ (/>
CD
•r- T3
S- C
rO
O
•M C
O)
•M O)
«4- $-
Ol U
r— CO
*"^ OJ
•r- C
S -*
O I
JZ S_
en CD
4->
« rO
S- S •
0) >>
3 S- E
S- O) O>
O tJ CO
(/) "O CO
01
OJ
(J -i- -
S- O) -O
T3 Q.
Q.r—
>> O -i-
s- i: O
IO CO
•M TJ
o oj -o
J_ OJ 0)
<*- J3
•> JD
^-> •— O
C 0) CO
3 >
O
T3 ^-- C
O) O)
4-> 4J U
U (U I
3 J* I
I- O r-
•M 3 O)
00 JD C
C (O
O TJ Q.
U QJ
o
•— -r- S-
i— i— 4->
3 C
U- >» O
i— O
-
vo o -i->
O) 3*-"
I- <0
3 S- 4->
72
•r- >, C
LL. j= 3
-------
>
O)
o
(1)
(/I
o
5-
o
OJ
O)
c
•r-
i-
OJ
-t->
O)
13
c
fO
O)
c
o
oo
O)
S-
73
-------
A. Drun cross section
B. Drun Isonetric
Figure 18. Soak zone description.
74
-------
has one wide angle sprayer (24 gpm @ 40 psi). There are eight 3CK>
W full-cone spray-pattern nozzles (10 gpm @ 40 psi) in the final screen rinse
section, (1 gpm = 3.785 L/m; 1 psi = 6.9 kN/m2), some of which can be
adjusted for screen backwashing.
Once inside the first screened section (initial spray zone) of the drum,
the soil is subjected to water-knife stripping with recycled water.
Preliminary tests showed that the minimum stripping time in this section
should be one minute; however, because of uncertainties about the
cohesiveness of the soil to be scrubbed, the initial stripping was
lengthened to four minutes.
The soil then is flushed for 15 min by a countercurrent flow of water in
the soaking zone. A baffle separates the soaking zone from the final
(rinse) zone. The effective baffle height is determined by drum pitch
angle and soil thickness. For a 3-degree pitch of the soil scrubber and a
design soil thickness of 22-cm (8.6-in.) at the end of the soaking zone, a
baffle height of 41-cm (16-in.) is required. A channel of water is formed
by the soil boundary and the drum wall. Soil tumbles into this
free-flowing reservoir to be cleansed by flushing and tumbling. Flights
are located at the end of the soak zone to lift the soil over the baffle
into the rinse zone.
Lower pressure, clean water sprays are utilized to rinse the soil for two
minutes in the final screened section of the drum. Clean soil tumbles from
"— the drum into the discharge chute for collection, possible drying, and
W replacement at the site.
Drum-Screen Capacity
Another design-limiting factor is the capacity of the EPA-ORD mobile
physical-chemical treatment system (activated carbon treatment trailer
(14)), which is utilized to treat washing fluids for reuse. The nominal
processing capacity of the basic granular activated carbon unit is 6.3
L/sec (100 gpm) but can be increased six-fold with lowered residence
times. Floe-settle and sand- anthracite filtration processing is part of
the treatment process. For many of the soil compositions and conditions
that will be encountered, such a rate of wash fluid recycle permits a
processing rate of 2.3 to 3.8 m3/hr (3 to 5 yd3/hr). However, when the
soil contains a large amount of free flowing granular solids (e.g., material
from railroad of highway beds), much higher treatment rates, perhaps as great
as 13.8 m3/hr (18 yd3/hr), are feasible.
COUNTERCURRENT CHEMICAL EXTRACTION UNIT
The principal elements of the semi-batch, countercurrent chemical extractor
for soils consist of well-mixed contact tanks and soil/washing-fluid
separators (Figures 19 and 20). A simplified schematic of a countercurrent
chemical extractor is shown in Figure 20. A soil slurry containing about
10 percent solids is fed and moves through the system in a forward
direction (from extraction tank 1 to tank 4) [left to right in Figure 20].
w
75
-------
QJ
O
lO •>
i — to
Q. a.
4-> E
C C 3
OJ -r- Q.
&-
i- I- CD
3 OJ C
O Q.-I-
i- CL X
OJ O M-
•i-» x: E
3 T3 T3
O C C
O «3 ,
••- O) O
M- W O
O 5
CT
OO ^: - I/)
O J- ^
O) (O
co cr>4->
c
•r- O -r- .
O J= i. l_
<— W) 0) O
(/> 4->
E » I 10
at en c i_
+-» c -r- a>
!/)•(- C
>>_a i_ a)
CO J3 3 O)
3 o
c s- •*- «—
O (J 0)
o o
• —
u.
i_
-C O
••-><*-
o
4. t_
U. O
«t o
Q. fl3
UJ $_
•M
X
• ai
cr>
ai u
J- -i-
3 E
UL o
a> a
-u
S-
(/> ro
00
4->
T3 O
(U 3
w -o
3 o
+j a.
o
C C
1T3
-4-> a>
3 i—
.a o
76
-------
-M cn
_c: i/> s.
01 s_
^: =#= i—
to o
E C +J
O) fO O)
4-> •!-> i—
to C
>> ....-
i^
-i-
X TJ -C
O) ••- Wl
3 (O
U JC
•i- ^: to
E to CD
0) <0 S-
•M C (U
C »
3 S- ^T
O 3
(J t— »
to j_
O •!-
o o
o «
o i- ••>
S- OJ I/)
Q- X O)
•r- C.
E O
»• x: o o>
o +J >, c
C»
3 =«fe >, (J
cn ^: o
•i- ^. S-
U- C -O T3
no c >>
•»-> 10 ^:
-------
Fresh water is introduced at the opposite end of the system (extraction
tank 4). The function of tank 4 is to rinse residual contaminant and
chemical additives from the soils before the last solid/liquid separation
(hydrocyclone 4) and final discharge. Fluid flow through the system is in
a reverse direction to the flow of soils (i.e., from extraction tank 4 to
1). Chemical additives may be injected during extraction steps 2, 3, and
4. Following solid/liquid separation (hydrocyclone 1) after the first
extraction, the spent washing fluids are removed for treatment and recycle.
Contact Tanks
To be effective, the extraction tanks must provide thorough contact between
the soil and the solvent. Mixing enhances this contact and is, therefore,
crucial to proper extraction. Two types of contact tanks were considered
for the soil scrubber: a propeller-mixed tank and the existing EPA mobile
froth flotation unit (oily beach sand cleaner) (Gumtz, 1972). The latter
alternative was chosen (Figure 19). Advantages of the EPA-ORD froth-"
flotation unit are the good mixing potential of the combined pneumatic/
mechanical mixing system and the use of a skimming mechanism for removing
floating materials and foam. The drawbacks include greater difficulty in
control of volatile substances and possible foaming problems when
surfactants are used as additives.
Required modifications to the EPA-ORD mobile unit included changing inlet
and outlet configurations, mounting hydrocyclones and pumps, coating the
tanks to provide chemical resistance, emplacing baffles to reduce fluid
short-circuiting, and installing an air contaminant enclosure. The tanks
are about 1.7-m (5-ft) high (including sloping bottom), 1.7-m (5-ft) long,
and 2-m (6-ft) wide. The modifications are complete; the converted
countercurrent chemical extraction unit is shown in Figure 19. (The unit
can easily be reconfigured to serve its original function as a beach sand
cleaner; indeed, the sand hopper (Figure 20, far left) is not used in the
soil scrubbing system but has not been removed.)
System Flow Equalization
A special feature has been incorporated into the system to eliminate
problems from non-uniformity in pump flows. Without this feature (unless
all of the pumps are perfectly matched in flow), the fluid levels in the
tanks would not remain constant and excessive effort to adjust valves
manually would be necessary unless complex servo instrumentation was
provided. Flow equalization is accomplished by allowing a small amount of
fluid to overflow from each tank back to the previous tank, i.e., from
right to left in Figure 20.
Solid/Liquid Separation
The preliminary evaluation indicated that hydrocyclones were quite
acceptable for the solid/liquid separation step. The hydrocyclones are
rubber lined and are all the same size, but do have various size vortex
finders and apex valves to accommodate the conditions that exist at each
extraction tank. When necessary, vortex finders can be changed in the
field.
78
-------
Washing Fluid Recycle
Washing fluid recycling is essential to reduce operating costs and to make
the soil scrubbing concept economically feasible. The spent washing fluid
from the countercurrent chemical extraction system is cleaned (treated) and
returned to the system. Note that cleaned washing fluid is also needed for
the nozzles of the water-knife drum-screen scrubber. The state-of-the-art
of treatment of contaminated water is quite advanced; mobile units are
already available for these situations. The use of the existing EPA mobile
physical/chemical treatment unit with the soil scrubber system is one
suitable way of achieving washing fluid recycle capability.
Treatment Residuals
In addition to the debris removed by the initial screening, the residuals
from the soil scrubbing system consist of spent carbon in the washing fluid
recycle system, soil fines carried out of the soil scrubber with the spent
washing fluids, floe-settle and filtration wastes, and foam and skimmings
from the froth-flotation tanks.
Emissions of air contaminants from the soil-scrubbing processes are
controllable using enclosures and ventilation, followed by air scrubbing or
other forms of air cleaning. Some, but limited, fugitive emissions can be
expected during soil excavation and feed to the system. Personal
protection to prevent inhalation hazards, as well as skin and eye contact,
is required for all system operators.
TREATMENT COSTS
For comparative purposes, the costs for scrubbing soil with different
levels of phenol contamination have been estimated and are presented in
Figure 21. Higher dollar requirements at elevated levels of phenol
correlate primarily with the additional expense associated with the
cleaning and recycling of washing fluids.
The estimated costs include those for: fuel for the diesel-powered
generators that furnish energy to operate the systems, two operators for
each process, front-end loader rental, chemical additives, and activated
carbon. The estimated costs do not include those for: startup and
shutdown, disposal of treatment residuals, maintenance and repair,
depreciation, clean water supply, freight, and shipping.
SUMMARY
A mobile system for treating excavated, contaminated soils at a field site
by sieving the soil, disintegrating the soil matrix, and cleansing the
soil is ready for shakedown, evaluation, and demonstration. The system is
expected to take its place among other emerging methods such as in situ
treatment or incineration, for treatment of contaminated soils at the
actual site of a hazardous substance spill or release under a wide variety
of circumstances.
79
-------
Laboratory work shows that non-aqueous (organic) solvent soil-washing
systems have special advantages in processing soils contaminated with
highly water-insoluble hazardous substances but that significant design and
engineering problems must be overcome to make such systems economically
attractive.
80
-------
100
7'j
ee.
UJ
Q.
a
25
LXPLCTED
OPERAiING
RANGE
r~ MAXIMUM
\ SYSTEM
\ CAPACITY
i i i i mi
l I I i in
I i
PHENOL CONTAMINATION, PPM
1000
Figure 21. Cost per volume treated for scrubbing phenol from soil,
81
-------
REFERENCES
1. Huibregtse, K. R. and Kastman, K. H., "Construction and Preliminary
Testing of a System to Protect Groundwaters from Hazardous Spills",
Proc. of 1980 National Conference on Control of Hazardous Material
Spills, Copyright Vanderbilt U., Nashville, Tennessee 37232,
pp. 77-81.
2. The Merck Index, Merck and Co., Inc., Rahway, New Jersey, 1976.
3. Ottinger, R. S., et al , Recommended Methods of Reduction, Neutraliza
tion, Recovery or Disposal of Hazardous Material, Volume II -
Toxicologic Summary, U.S. EPA 670/2-75-053-6, U.S. Environmental
Protection Agency, Cincinnati, Ohio, 1973.
4. Dawson, G. W., et al , Control of Spillage of Hazardous Polluting
Substances, Federal Water Quality Administration, U.S. Department of
the Interior, Washington, D.C., 1970, 89 pp.
5. Rockwell International, Environmental Services, Invitation to
Propose, EPA-8, In-Situ Treatment of Hazardous Spills in
Watercourses, Request for Proposal, March 9, 1979.
6. PCBs in the United States, Industrial Use and Environmental
Distribution, PB-252-012, National Technical Information Service,
U.S. Environmental Protection Agency, Cincinnati, Ohio, 1976.
7. Gould, E. S., Inorganic Reactions and Structure, Henry Holt and
Company, New York, 1955.
8. The Encyclopedia of Chemical Technology, Van Nost, Rand & Reinhold
Co., N.Y., 1966.
9. Standard Methods for the Examination of Water and Wastewater,
American Public Health Association, American Water Works Associa-
tion, Water Pollution Control Federation, Washington, D. C., 1976.
10. Methods for Chemical Analysis of Water and Wastes, U.S. EPA
600/4-79-020, U. S. Environmental Protection Agency, Cincinnati,
Ohio, 1979.
11. Bellar, T. A., and J. J. Lichtenberg, Methods for Benzidine,
Chlorinated Organic Compounds, Pentachlorphenol and Pesticides in
Water and Wastewater, Interim report, U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1978.
82
-------
12. Public Law 96-510. "Comprehensive Environmental Response,
Compensation, and Liability Act of 1980".
13. Gumtz, G.D., 1972. Restoration of Beaches Contaminated by Oil.
EPA-R2-72-045 (September, 1972).
14. Gupta, M.K., 1976. Development of a Mobile Treatment System for
Handling Spilled Hazardous Materials. EPA-600/2-76-109.
83
-------
TECHNICAL REPORT DATA
(I'lcase read Instructions on the reverse before cr»ni>lctin/>/
1. REPORT NO.
4. TITLE AND SUBTITLE
MOBILE SYSTEM FOR EXTRACTING SPILLED
HAZARDOUS MATERIALS FROM EXCAVATED SOILS
7. AUTHOR(S)
Robert Scholz and Joseph Milanowski
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Rexnord, Inc.
5103 W. Beloit Road
Milwaukee, Wisconsin 53214
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory—Cin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO
10. PROGRAM ELEMENT NO.
TEJY1A
11. CONTRACT/GRANT NO.
68-03-2696
13. TYPE OF REPORT AND PERIOD COVERED
Final 12/76-4/82
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officer: John Brugger (201) 321-6634
Laboratory tests were conducted with three separate pollutants (phenol, arsenic
trioxide, and polychlorinated biphenyls (PCB's)) and two soils of widely different
characteristics (sand/gravel/silt/clay and organic loam) to evaluate techniques for
cleansing soil contaminated with released or spilled hazardous materials. The tests
show that scrubbing of excavated soil on site is an efficient approach for freeing
soils of certain contaminants but that the effectiveness depends on the washing fluid
(water + additives) and on the soil composition and particle size distribution. Base
on the test results, a full-scale, field-use system was designed, engineered,
fabricated, assembled, and briefly tested; the unit is now ready for field
demonstrations. The system includes two major soil scrubbing components: a water-
knife stripping and soaking unit of novel design for disintegrating the soil fabric
(matrix) and solubilizing the contaminant from the larger particles (>2 mm) and an
existing, but re-engineered, four-stage countercurrent extractor for freeing the
contaminants from smaller particles «2 mm). The complete system requires auxiliary
equipment, such as the EPA-ORD physical/chemical treatment trailer, to process the
wastewater for recycling; under some circumstances, provision must be made to confine
and treat released gases and mists. The processing rate of the complete system is
2.3 to 3.8 m3/hr (4 to 5 yd3/hr), though the water-knife unit (used alone) can
process 11.5 to 13.5 m3/hr (15 to 18 yd3/hr).
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
I'.IDFNTIFIERS/OPFN ENDED TERMS
COSATI Held/Group
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19
?1 NO OF PAGES
2O SFCURITY CLASS (This r>af!<'I
UNCLASSIFIED
EPA Form 2220-1 (9-73)
84
------- |