EPA-600/2-77-164
October 1977
Environmental Protection Technology Series
IN SITU TREATMENT OF HAZARDOUS MATERIAL
SPILLS IN FLOWING STREAMS
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document Is available to the public through the National Technical Informal?
tioitServfce, Springfield, Virginia 22161.
-------
EPA-600/2-77-164
October 1977
IN SITU TREATMENT OF HAZARDOUS MATERIAL
SPILLS IN FLOWING STREAMS
by
Gaynor W. Dawson
Basil W. Mercer
Richard G. Parkhurst
Battelle-Northwest
Richland, Washington 99352
Contract Nos. 68-03-0330
68-03-2006
Project Officers
Ira Wilder
Joseph P. Lafornara
Oil and Hazardous Materials Spills Branch
Industrial Environmental Research Laboratory - Cincinnati
Edison, New Jersey 08817
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory - Cincinnati, U. S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endores-
ment or recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
This report is a product of the above efforts. It documents the studies
conducted to evaluate the effectiveness of two methods of applying activated
carbon treatment to flowing watercourses which have been impacted by spills
of hazardous materials. As such it serves as a reference to those in state,
local and Federal Agencies, the transportation and chemical industries, and
others who are interested in the control of spills of hazardous materials.
The project is part of a continuing program of the Oil and Hazardous Mater-
ials Spills Branch, lERL-Ci, to assess and mitigate the environmental impact
of pollution from hazardous material spills.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
ill
-------
ABSTRACT
Two methods of applying activated carbon adsorption treatment
to flowing streams were evaluated under comparable conditions. The
first involved sub-surface introduction of bouyant carbon into the
water column followed by the floating of the carbon to the surface
and'subsequent removal using conventional surface skimming techniques.
The second involved the addition to the water of non-bouyant granular
activated packaged in porous fiber bags ("tea bags") which were attached
to floats. The bags were allowed to travel with the spill plume for
a given distance and were subsequently removed manually. Controlled
field experiments using n-hexone as the test chemical were conducted
at various flow rates in a specially modified abandoned irrigation
channel at the Energy Research and Development Administration's Hanford
site and showed that for "low-flow" non-turbulent conditions the bouyant
carbon technique was more effective in removing the chemical from
the water with only tolerable amounts of the carbon remaining in the
stream. As the flow and turbulence increased the pollutant removal
effectiveness of the "tea bag" approach improved.
This report was submitted in fulfillment of Contracts Nos. 68-03-0330
and 68-03-2006 by Battelle-Northwest under the sponsorship of the U.S.
Environmental Protection Agency.
iv
-------
CONTENTS
Foreword
Abstract
Figures
Tables
Acknowledgements viii
1. Introduction 1
Buoyant Carbon 1
Sinking Carbon 2
Purpose of These Studies 2
2. Conclusions 3
Buoyant Activated Carbon 3
Porous Fiber Bags 3
Comparative Evaluations 4
3. Recommendations 5
4. The Flowing Stream Test Facility 6
5. Description of Treatment Concepts 15
Buoyant Carbon 15
Sinking Carbon 17
6. Field Application Studies with Buoyant Carbon 19
General Information 19
Series I 20
Series I - Results and Discussion 20
Series II 25
Series II - Results and Discussion 25
Series III 28
Series III - Results and Discussion 28
Effectiveness of Carbon Containment Boom 30
7. Development of the Porous Bag System 33
Porous Bag Design and Fabrication 33
Porous Bag Regeneration Facility 35
8. Comparative Studies of Alternate Carbon Systems 40
9. Practical Aspects of in situ Treatment in Flowing Streams .... 45
References / g
v
-------
FIGURES
Number Page
1 Maximum stream flow rate for a given
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Maximum run time available to maintain
a given stream depth
Test section of canal
Reservoir section of Flowing Stream Test
Facility (FSTF) looking downstream
Test section of FSTF looking downstream
Test section of FSTF looking upstream
Quiescent section of FSTF looking downstream. . . .
Mechanisms for removal of contaminants by
surface applied floating activated carbon ....
Sampling site locations for test series #1
Sampling site locations for test series
#2 and #3
Booming floating carbon in quiescent
section of FSTF
Fabrication procedure for producing porous
bags
Completed porous bag
Standard float arrangement
Schematic of steam strip regeneration facility. . .
Sorption characteristics for carbon application
9
10
11
12
13
14
16
21
26
31
34
36
37
38
42
vi
-------
Number
1
2
3
4
5
TABLES
Apparent fraction remaining compared to
Apparent fraction remaining compared to
initial sample S . . . . .
Results of runs with spills in turbulent
Results of instantaneous spill conducted
at reservoir weir
Page
. . 22
. . 23
. . 23
. . 27
. . 29
Recovery of various sized particles from
application of 75 grams of gloating carbon ... 32
Comparison of in situ treatment efficiencies
at various flow rates 41
Comparison of in situ treatment efficiencies
at larger spill volume (flow rate ^10 CFS) ... 44
vii
-------
ACKNOWLEDGEMENTS
The assistance and advice provided by Mr. Ira Wilder and
Dr. Joseph P. Lafornara, EPA Project Officers, are gratefully
acknowledged. The authors also wish to express their apprecia-
tion to Battelle-Northwest personnel: Mr. James Coates,
Mr. Marvin Mason, Mr. Robert Upchurch, Mr. Gary Schiefelbein,
Mr. Gary Roberts, Ms. Nancy Painter, and Ms. Betty Thomas who
assisted in conducting the reported study and preparing this
document.
Vlll
-------
SECTION 1
INTRODUCTION
The use of activated carbon for in situ treatment of hazard-
ous material spills has been reported previously.1"1*'6 The two
major alternatives presently under development are the use of
buoyant carbons and the use of commercial grade sinking carbons
packaged in porous fiber bags attached to floatation devices.
Both systems have shown promise with respect to specific appli-
cations in spill situations.
BUOYANT CARBON
Initial work with buoyant activated carbon involved media
development, delivery package development, and field demonstra-
tion of the technique. A commercially available granular carbon,
Nuchar c-190, was found to display the desired buoyancy pro-
perties as produced. Subsequent laboratory tests confirmed that
delivery could be achieved if ballasted packages could be devised
that would release the media upon reaching the bottom of the
receiving water. Three feasible alternatives were identified:
containment in weighted plastic bottles, containment in unfired
clay containers, and incorporation with gravel ballast in an ice
matrix. In the first case, release of the media occurs through
the narrow mouth of the bottle. Ballast contained in the bottom
of the bottle holds the bottle in an upright position to allow
the media to be released. Release of the media from the other
packages occurs upon disintegration of the clay container in
water or melting of the ice cake. All three delivery packages
are considered potential alternatives at the present time.
Field demonstration was conducted using activated carbon con-
tained in weighted plastic bottles. A total of 835 Ibs (380 Kg)
of carbon was applied to a simulated spill of 78 Ibs (35 Kg)
of an emulsifiable oil solution of an organophosphate pesticide
in a ten million gallon water storage basin. The bottles of
carbon and ballast were dropped into the spill area from a heli-
copter. Carbon was subsequently collected at the surface through
use of an oil containment boom and pumped as a slurry to a
storage tank. Analysis of pretreatment and posttreatment water
samples taken in the spill zone showed that approximately 80 per-
cent of the pesticide was removed from the water. Carbon recovery
with a standard oil boom exceeded 90 percent.
-------
SINKING CARBON
Preliminary work with sinking carbons was directed to the
development of application techniques, and the optimization of
adsorption kinetics. The former was achieved through construction
of bags from open weave nylon fabric which was sufficiently tight
to hold granular carbon (8 x 30 mesh) and yet allow a maximum
amount of water to flow through the bag. Individual bags were
suspended from floats and placed in contact with contaminated
water. Tests in static tanks revealed that adsorption is exceed-
ingly slow in the absence of turbulence. Hence, performance in
lakes and backwaters would not be good unless wave action or
artificial mixing were prevalent. On the other hand, the presence
of currents could increase removal efficiency greatly. Indeed,
small scale testing in a race track configuration revealed high
levels of removal during a three hour contact time.
PURPOSE OF THESE STUDIES
Early studies of both the buoyant and sinking carbon systems
were focused largely on concept development and limited testing
in ponded waters. Historical data, however, indicate that a
preponderance of spills occur in flowing waters. Dawson and
Stradley have estimated that 82 percent of all freshwater spills
occur in rivers and streams.5 Of the remaining 18 percent, some
spills will actually involve reservoirs on navigable waterways.
It was the purpose of these studies to evaluate the two most
promising in situ treatment techniques operable in flowing waters.
Of particular interest with regard to the use of buoyant acti-
vated carbon were several technical issues: the need (or lack
thereof) for ballast and packaging; the efficiency of contact;
probability of unsightly carbon buildup along stream banks; and
the efficiency of spent carbon collection. If the need for bal-
last were eliminated as a result of media suspension sponsored by
natural turbulence, delivery systems could be greatly simplified.
The major question to be resolved with respect to sinking carbon
was the ability of natural currents to supply sufficient tur-
bulence to enhance the kinetics of adsorption.
The work reported herein was performed under two separate
contracts with the U. S. Environmental Protection Agency and was
directed to the evaluation of both buoyant and sinking activated
carbon systems in flowing streams. All work was conducted in a
simulated flowing stream maintained on the Hanford Atomic Energy
Reservation. Early studies to verify the potential for appli-
cation of floating carbon without ballast were followed by
parallel evaluations of the two systems under varying conditions.
-------
SECTION 2
CONCLUSIONS
The removal of a soluble organic hazardous material from a
flowing stream environment has been demonstrated by two different
approaches one utilizing buoyant activated carbon, and one
employing sinking carbon in porous fiber bags suspended from
floats. The following conclusions are based on results of these
studies.
BUOYANT ACTIVATED CARBON
Buoyant carbon can be effectively employed on flowing streams
without the use of ballasted packages. Carbon can be applied
directly to the surface or slurried and injected beneath the
surface.
Natural turnover is sufficient to provide intimate contact
of carbon with contaminated water in shallow streams.
Floating carbon was capable of achieving 50 percent removal
at a carbon to contaminant ratio of 10:1 under the range of
flow rates and spill conditions studied.
Some floating carbon was captured in eddies and debris along
the side of the water body, but carbon recovery is generally
in excess of 90 percent. Containment booms must be placed
in a quiescent stretch where velocity components will not
exceed the capabilities of the device. Simple oil booms are
sufficient if weather and current conditions allow operation.
POROUS FIBER BAGS
Porous fiber bags can be effectively applied to spills in
flowing streams, but removal is directly related to the tur-
bulence and current structure in the receiving water. Re-
moval rose from insignificant levels at low flow rates to
20 percent at a flow rate of 15 cubic feet (425 liters) per
second.
Fiber bag efficiency appears to be limited by kinetic con-
siderations. Removal may improve with longer contact times
than those which can be achieved in the test facility
employed here.
-------
Up to 25 percent of the fiber bags were lost during the tests
as a result of shore capture and snagging in shallow areas.
This effect diminished with higher steam velocities.
Fiber bags can be collected after use with simple booms or
a wire strung perpendicular to the flow just below the sur-
face of the stream. Positioning in a quiescent stretch is
not required.
Fabric for the fiber bags must be carefully selected to avoid
decomposition by the material being removed. At the same
time, use of a heat and pressure resistant material allows
regeneration in the bag, thus avoiding the necessity for
empty and refill sequences.
Steam stripping was found to be an adequate method of regen-
eration for the methyl isobutyl ketone employed in this
study.
COMPARATIVE EVALUATIONS
Buoyant carbon was superior to porous fiber bags in terms of
removal efficiency under the spill and flow conditions
tested.
Little difference between methods with respect to media
loss along shorelines was noted, but fiber bags can be
retrieved more simply than buoyant carbon and under more
extreme flow conditions.
Fiber bags can be loaded, unloaded, and handled with greater
ease than buoyant carbon.
Both approaches rely heavily on the ability of the response
team to locate and trace the movement of the contaminant
plume.
The ability to inject slurried buoyant carbon at depth
renders this approach more attractive than fiber bags for
use in deeper channels.
Buoyant carbon will be more greatly affected by adverse
weather conditions than will fiber bags.
Both approaches may be difficult to apply to very large
spills simply as a result of the logistics of ferrying
large quantities of carbon to the site.
-------
SECTION 3
RECOMMENDATIONS
A survey should be conducted to assess the availability of
buoyant carbons for use in spill response.
An air deliverable slurry injection system should be
developed and tested for routine application of buoyant
carbon to spills.
A study of the treatment effectiveness of porous bag packaged
carbon should be performed at a site where longer contact
times than were possible for these tests can be achieved.
Further work should be conducted to develop remote sensing
and/or other techniques for the identification, location, and
monitoring of spills. Some consideration should also be
given to the development of methods for marking spill plumes.
A decision framework is needed to determine when spill
response is warranted and what the most effective means of
response is for specific spills on a real-time basis.
-------
SECTION 4
THE FLOWING STREAM TEST FACILITY
All tests were conducted in the Flowing Stream Test Facility
(FSTF) which is an abandoned irrigation canal located on the Atomic
Energy Commission Reservation at Hanford, Washington. The par-
tially cement-lined canal was .taken out of operation some thirty-
four years ago when the federal government appropriated the land.
Since 1973, however, it has been the subject of renovation efforts
aimed at equipping it for use as a model stream for hazardous
materials spill research.
For the purposes of the work reported here, renovation largely
consisted of efforts to provide and control the flow of water.
Two nearby wells were deepened, reactivated, and fitted with gaso-
line driven pumps capable of producing 200 and 600 gallons (757
and 2271 liters) of water per minute, respectively. Aluminum and
"transite" irrigation pipes were installed to carry water to a
reservoir formed in the upper 1000 feet (305 meters) of the canal
by the construction of a permanent weir with a screw-controlled
drop gate. This reservoir is followed by 2200 feet (671 meters)
of test section and an additional 200 feet (61 meters) of quies-
cent water. The test section of the canal has a trapezoidal
cross-section with a 5 foot (1.5 meters) base, a 15 foot (4.6
meters) top, and a 5 foot (1.5 meters) altitude and a bottom
slope of 0.00024. The quiescent section was widened to approxi-
mate a 20 foot (6.1 meters) wide by 5 foot (1.5 meters) deep
rectangular cross-section. A second weir with an optional over-
flow or overflow or underflow gate was installed to control flow
in this section. After passage through the second weir, the
water is released to a sandy basin in the adjoining desert.
Since the integrity of the original cement lining was
breached by various plant forms, temporary linings were installed
in portions of the canal for the present program to prevent
seepage. The reservoir was lined with sheets of heavy duty
polyethylene sheeting sealed together and covered with soil and
gravel to prevent wind damage. The test section was treated with
a slurry of bentonite clay to seal off major infiltration routes.
The quiescent section of the facility was overlaid with a single
fused sheet of 30 mil(0.08 centimeters) polyvinyl chloride.
During actual field studies, the flow was controlled by com-
bining the reservoir water with the pump discharge. Total flows
of 0-16 cfs (453 liters) were achieved. Flow characteristics
-------
for the canal are illustrated in Figure 1. Run time relations
are presented in Figure 2. Actual flow rates during test runs
were determined both by depth sticks in the run stretch and a V
notch weir. A schematic diagram of the test facility appears in
Figure 3. Figures 4 through 7 illustrate the various features of
the canal.
A collection boom was constructed in the quiescent section
for retrieval of the floating media and porous bags. The boom
was formed by sections of 2 x 4's strung on a nylon rope. A
plastic skirt was attached to each segment such that it extended
three inches (7.6 centimeters) into the water and three inches
(7.6 centimeters) above. When buoyant carbon was employed, media
was pumped from the front of the boom with a gasoline operated
diaphragm pump. The collection port was funnel-shaped and sat
just below the water surface in front of the boom. Holding tanks
were maintained for drying and weighing the retrieved buoyant
media. All porous bags were collected by hand and returned to
the laboratory for regeneration.
-------
00
FIGURE 1. MAXIMUM STREAM FLOW RATE FOR A GIVEN STREAM DEPTH
-------
2.5
- O
VD
2.0
a
0)
a
(9
0)
1.0
w
.5
50
100
_L
JL
150 200
Run Time, Min.
250 300
350 400
FIGURE 2. MAXIMUM RUN TIME AVAILABLE TO MAINTAIN A GIVEN STREAM- DEPTH
-------
ROAD
FILL
RESERVOIR
RESERVOIR DAM AND SLUICE GATE
COLLECTION
AREA
DISCHARGE
OUTLET DAM AND SLUICE GATE'
FIGURE 3. TEST SECTION OF CANAL
-------
FIGURE 4. RESERVOIR SECTION OF FLOWING STREAM TEST FACILITY (FSTF) LOOKING DOWNSTREAM
-------
FIGURE 5. TEST SECTION OF FSTF LOOKING DOWNSTREAM
12
-------
FIGURE 6. TEST SECTION OF FSTF LOOKING UPSTREAM TOWARDS RESERVOIR
13
-------
FIGURE 7. QUIESCENT SECTION OF FSTF LOOKING DOWNSTREAM
-------
SECTION 5
DESCRIPTION OF TREATMENT CONCEPTS
BUOYANT CARBON
Major problems are associated with the application of pack-
aged buoyant media for flowing streams. As a result of movement
of the contaminated plume with the current, timing of media
release becomes critical. If media were delivered in packages,
the release of the media from the packages would have to be timed
exactly to coincide with the passage of the plume in order to
achieve effective treatment. In order to avoid this problem, two
alternative methods of application were explored: 1) surface
application of the media with contact dependent upon the natural
turnover of the stream water, and 2) subsurface injection of
slurried media.
Surface application relies on two mechanisms to provide inti-
mate contact between the contaminated water and the buoyant sorp-
tion media. Both of these are related to the natural turnover
of the water as it flows downstream. The first mechanism involves
the vertical velocity components of the flow itself which dis-
perse fine media particles downward where they contact contaminated
waters and sorb the contaminant. In the case of the second mech-
anism, the larger particles float on the surface and sorb contam-
inant from the deeper waters as the latter comes to the surface
and rolls back to the bottom. The two mechanisms are concep-
tualized in Figure 8.
With subsurface application, the media is slurried and pumped
into the deeper portions of the contaminant plume. Intimate con-
tact is achieved initially through the dispersion of the slurry
itself in the receiving waters and its subsequent ascent to the
surface.
Optimal particle size will depend in part on the mode of ap-
plication anticipated. Since surface collection of spent media
in quiescent reaches is the mode of retrieval, particles must be
sufficiently buoyant (a function of particle diameter when density
is held constant) to rise to the surface during residence in the
selected quiescent zone. On the other hand, if particles are too
large they will not be carried down into the water column or will
rise too quickly after subsurface injection to achieve the
required contact. These system requirements, therefore, determine
15
-------
LARGER PART I CLfS
CONTAMINANT
PARTICLES
FINE PARTICLES DEPLOYED BY VERTICAL VELOCITY COMPONENTS
CARBON
f
CONTAMINATED
WATERS \
\ v
VERTICAL VELOCITY COMPONENTS CARRY CONTAMINATED WATERS TO THE SURFACE
FIGURE 8. MECHANISMS FOR REMOVAL OF CONTAMINANTS BY
SURFACE APPLIED FLOATING ACTIVATED CARBON
16
-------
physical limitations on the acceptable particle-size range for
buoyant sorption media. They, in turn, are influenced by the
velocity components of the receiving waters.
Field studies were conducted to determine the approximate
mesh ranges of Nuchar C-190 which would rise to the surface when
applied to flowing streams. All samples greater than 250 mesh
were found to be sufficiently buoyant to be recovered in quiet
waters. The breakpoint for large particle sizes was more diffi-
cult to define. Whereas carbon in the size range greater than 50
mesh was found to stay on the surface with no mixing under calm
conditions, a slight wind was sufficient to ripple the surface
and initiate movement of the particles into the water column. To
facilitate testing, selection was oriented to assure all particles
could be recovered on the surface and many would mix under varying
environmental conditions. For practical purposes, the optimal
working range was defined as 50 x 250 mesh. While this includes
many large particles which may never mix to a significant degree,
it will be far less costly to obtain commercially than a narrow
size range.
SINKING CARBON
The use of sinking activated carbon for in situ treatment of
waters requires removal of the spent carbon from the bottom of
the watercourse, abandonment of the spent media at the bottom, or
incorporation of the carbon in a package which allows retrieval
at some later time. The latter approach (packaging) is the most
practical and was selected for development here. Each package
must be designed to allow contact with contaminated water without
releasing media. These requirements are similar to those which
have led to the development of tea bags. Indeed, the tea bag con-
cept is very appropriate for application of activated carbon.
Activated carbon is placed in a porous fiber bag with a pre-
selected thread count just sufficient to hold the smallest granule
sizes to be used. Contaminated water can thereby flow into the
bag, contact the carbon, and flow out. As noted earlier, the bags
do not perform well in static water because of a lack of flow
through the bag and media. In flowing streams, however, the
natural turbulence should be sufficient to constantly exchange
waters in immediate contact with the bag. Contact is further
stimulated by filling the bags only partially full and thereby
leaving ample room for the media to fluidize. In this expanded
state, adsorption kinetics are enhanced.
Adsorption will be greatly affected by carbon particle size
as well as the aforementioned factors related to water exchange.
The finer the particles employed, the greater the ease of
fluidization and surface contact, and the more rapid the
adsorption. Fine particles, on the other hand, require fine
17
-------
materials to contain them which, in turn, discourages water flow.
In balancing.these interests, researchers at Calspan Corporation
determined that commercially available granular carbons were ade-
quate.6 These are marketed in two size ranges: 8 x 30 mesh and
12 x 40 mesh. Early investigations were conducted with the larger
carbon. The smaller 12 x 40 mesh material was selected for use in
flowing streams to enhance contact with the water.
Retrieval of the "bags" can be provided for by attaching them
to floats or other bouyant devices. Visual observation of the
floats allows constant knowledge of where and how the bags are
moving. At the end of the contact period, bags can be retrieved
by placing a boom or other collection device across the channel
such that the floats are snared .and held against the current.
Since the floats allow the bags to move with the current, the
carbon is kept in constant contact with the contaminant plume.
Thus, contact times can be made as long or as short as desired
depending upon where collection booms are deployed. A final
degree of treatment is also provided at the boom where the snared
bags are analogous to a fixed carbon bed through which the con-
taminated water must flow.
18
-------
SECTION 6
FIELD APPLICATION STUDIES WITH BUOYANT CARBON
GENERAL INFORMATION
Field studies were conducted in the Flowing Stream Test
Facility at flow rates ranging from 3.98 to 4.45 cfs (113 to
126 liters). Spills were simulated with a solution of 1816 grams
(4 Ibs) methylisobutyl ketone (hexone), 550 grams (1 Ib) of
methanol, and 45 grams (0.1 Ibs) of rhodamine dye. The methanol
was employed as a bridge solvent to enhance the solubility of the
dye and the hexone. The dye was included to allow visual moni-
toring of the spill plume and hence facilitate the timing of media
application. Data and isotherms for the adsorption of this
solution on buoyant carbon and porous bags are given in Section 8.
Three series of tests, each composed of multiple runs, were
carried out during the course of the program. Each series dif-
fered in the manner in which the spill was simulated and in which
the contaminant plume was allowed to develop prior to treatment.
Nuchar C-190 carbon was used in all tests with carbon mesh size
varied for certain of the test series. For each series of tests,
a spill was conducted without application of floating media to
establish background levels for the contaminant plume and the
effects of natural dilution. This was necessary because, in
addition to dilution, sorption onto plants and material in the
test canal occurred during the course of the study. Removal was
then defined as the difference between concentrations for treated
and untreated samples taken at the same location.
Three sets of samples were taken during the trials. Samples
were taken across the entire width of the stream at a depth of
six inches (15 cm) with some provision for a larger sample input
at the deeper center portion of the flow. All samples were
stored in glass bottles and refrigerated until analysis. The
first, at sample site S, was taken just upstream of the carbon
application site. The second, at sample site C, was taken at the
head end of the quiescent reach (approximately 15-25 minutes
of carbon contact time). Sample site D was located at the lower
end of the quiescent reach just behind the carbon collection boom.
Samples at this site may not be completely representative since
the 20-foot channel width and unpredictable currents make it
difficult to obtain composite samples. A schematic diagram
19
-------
showing the locations of the sampling points is presented in
Figure 9. Three analytical techniques were initially employed to
characterize the spill. Dye content was measured using a GK
Turner Model 111 fluorometer. Total organic carbon content was
monitored through use of a Beckman Model 915 total organic carbon
analyzer. Hexone measurements were made with a Perkin-Elmer 900
gas chromatograph.
SERIES I
For the first series of tests, the hexone solution was spilled
over a ten-second period at a point 200 feet (61 meters) downstream
of the reservoir. This location was selected to assure that all
artificial turbulence from the sluice gate was damped. Forty
pounds (18.1 kilograms) of floating carbon was applied to the
stream at a point 230 yards (210 meters) from the reservoir
(approximately 10 minutes' flow time). Large grain size (12 x 40
mesh) Nuchar C-190 was employed for all tests. Runs involving
both surface (Run 2) and subsurface (Run 3) application were con-
ducted in this series. Surface application consisted of sprinkling
the carbon on the water as the contaminant plume reached the appli-
cation point. For the subsurface application case, a carbon
slurry of approximately 10 g/1 was prepared prior to the spill and
was then pumped to the bottom of the stream as the contaminant
plume passed.
Series I - Results and Discussion
Results of hexone, rhodamine dye, and total organic carbon
analyses are presented in Table 1. The total material figures
were derived from concentration measurements taken at specified
time intervals as the plume passed the sampling point and summed
for the plume at the flow fate noted.
The apparent fractions remaining in Runs 2 and 3 as compared
to the blank Run 1 are given in Table 2. Table 3 shows the
apparent fractions remaining when compared to the initial samples
at site S. Several observations are in order.
It is evident from the data that a large fraction of the
hexone cannot be accounted for. Only 26 to 34 percent of the
original 1816 grams (4 Ibs) of hexone was detected in the initial,
untreated S samples. Similarly, only 39 to 45 percent of the
original TOG was detected at site S, while virtually all of the
dye was accounted for. The data suggest that the apparent loss
in TOC can be attributed almost entirely to the apparent hexone
loss. At site S, the loss in hexone, 1200 to 1535 grams (2.6 to
3.3 Ibs), translates into a theoretical TOC loss of 864 to 1052
grams (1.9 to 2.3 Ibs) as compared to a measured TOC loss of
826 to 926 grams (1.8 to 2 Ibs).
20
-------
ROAD
FILL
RESERVOIR
RESERVOIR DAM AND SLUICE GATE
STREAM FLOW
COLLECTION
AREA
CARBON APPLICATION
SAMPLE SITE C
OUTLET DAM AND SLUICE GATE-
BOOM
SAMPLE SITED
DISCHARGE
WOOD FLUME FOR
CANAL CROSSING
FIGURE 9. SAMPLING SITE LOCATIONS FOR TEST SERIES #1
-------
TABLE 1
ANALYSIS OF SERIES OF #1 SPILL PLUMES
Run #2
Subsurface Run #3
Run #1 Blank Slurry Application Surface Application
6761 1/min (3.98 cfs) 9173 1/min (4.45 cfs) 9173 1/min (4.45 cfs)
Sample Site S
Total Hexone (g) 616 481 539
Total Organic Carbon (g) 687 591 587
Total Dye (g) 44.0 38.! 34>9
Sample Site C
Total Hexone (g) 670 643 508
Total Organic Carbon (g) 872 559 751
Total Dye (g) 37 10.0 1.5
Sample Site D
Total Hexone (g) 752 223 341
Total Organic Carbon (g) 737 271 574
Total Dye (g) 26.7 1.4 0.76
-------
TABLE 2
TEST SERIES # 1
APPARENT FRACTION REMAINING COMPARED TO BLANK RUN
Run t 1 Run # 2 Run # 3
Sample Site S
Hexone
Organic Carbon
Dye
Sample Site C
Hexone
Organic Carbon
Dye
Sample Site D
Hexone
Organic Carbon
Dye
APPARENT FRACTION
Sample S
Hexone
Organic Carbon
Dye
Sample C
Hexone
Organic Carbon
Dye
Sample D
Hexone
Organic Carbon
Dye
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
TABLE 3
.78
.86
.86
.96
.64
.27
.30
.37
.05
TEST SERIES # 1
REMAINING COMPARED TO
Run # 1
1.00
1.00
1.00
1.08
1.27
.84
1.22
1.07
.61
Run I 2
1.00
1.00
1.00
1.34
.94
.26
.46
.46
.04
.95
.86
.79
.76
.86
.04
.45
.78
.02
INITIAL SAMPLE S
Run # 3
1.00
1.00
1.00
.86
1.28
.04
.58
.76
.02
23
-------
Since methanol is more volatile than hexone and apparently
little methanol is unaccounted for, evaporation does not appear
to be a plausible explanation of the apparent loss. A more
reasonable explanation is that the dye and the methanol are both
water soluble to a greater extent than hexone and hence can be
expected to be selectively extracted from the hexone into the
water. The hexone would then form a lighter floating layer of
hexone and hexone-saturated water solution which would mix more
slowly with the remaining water column. There would then be a
gradient from high hexone concentrations at the surface to lower
concentrations with depth below the surface. Hence, samples taken
at a six-inch (15 cm) depth may underestimate hexone content con-
siderably. This incomplete mixing hypothesis would also explain
the apparent production of hexone in the blank run as the plume
moved downstream. That is, with movement downstream, turnover
and vertical dispersive forces would slowly bring the hexone to
an isoconcentration state thus making the amount of hexone at
the six-inch depth level more representative. This would be
especially true of the site D samples since it is downstream of
the containment boom which would stimulate vertical currents.
Stratification of this nature could well affect removal
efficiency. Whereas the apparent removals were 64 and 45 percent
hexone, respectively, more than half of the original spill was
on the surface of the flow where the bulk of the carbon remained.
Hence, all of the missing hexone may well have been sorbed on the
carbon by the end of the run. Removal of dye was consistently
better than 90 percent. This does not allow for corrections
required as a result of differences in absorption onto plants
and soil in the treated runs as compared to the blank run, but
there is no reason to believe such a correction would be very
large.
There appears to be no advantage to use of slurry application
over surface application with the larger grain size carbon. This
may in part be an artifact of the mode in which the slurry was
pumped into the ditch, but more likely it reflects the quick
rise time for the large particles and hence a minimal amount of
increased contact with submerged waters. The advantages of sub-
surface injection are likely to become apparent in deeper streams
where surface application leads to contact with only a portion of
the contaminated plume.
Comparison of TOC, hexone, and dye levels reveals no constant
relation between any two parameters. This might be expected since
the three components vary in their solubility, tendency to strat-
ify, and adsorptive properties. The use of gas chromatography
for hexone detection proved very satisfactory with good reproduci-
bility on field samples and standards. Therefore, since hexone
was the major component of the spill, hexone analysis was selected
as the major measurement basis in subsequent trials. TOC and dye
measurements were used sparingly to provide confirmation on
hexone data.
24
-------
During the first series of runs, several environmental
factors were found to complicate spill response activities. It
was found_that wind conditions greatly influenced the effective
contact time achieved during any single run. The larger, more
buoyant carbon particles stayed at, or very near, the surface
throughout the test period. Wind moving in the direction of the
flow accelerated the floating media to a velocity much greater
than that of the spill plume itself. Thus, after a short contact
period, the carbon passed the contaminant plume and moved down-
stream ^in contact with relatively unaffected waters. Under calm'
conditions, a similar development was observed to occur over a
longer period of time as a result of the greater relative velocity
of the surface waters to that of the deeper layers. Conversely,
when winds prevailed in a direction counter to flow, the carbon
stayed with the contaminant plume throughout the test period.
Difficulties were also experienced in collecting the spent
carbon. The carbon was easily detained behind a wooden boom
fitted with plastic skirts. When an attempt was made to pump
the contained carbon to a holding facility, however, it was found
that abundant plant debris quickly plugged the lines. The debris
was excessive as a result of the intermittent flow pattern in the
test facility. During dry periods, various wind-blown plant
forms accumulated in the run stretch. Initiation of flow then
scoured these weeds throughout the 2 to 3 hour test period.
This should not prove to be a major problem in natural streams
if flooding is not occurring. Additionally, coarse screens
preceding the intake line were found to remove most of the
plant debris and thus enable collection of the carbon.
SERIES II
For the second series, the dye-hexone-methanol solution was
added to the turbulent waters at the foot of the reservoir weir
with a variable speed pump over a five-minute time period. The
pumping rate was adjusted to simulate the concentration gradient
observed in well-developed plumes. Carbon addition was carried
out at a distance 50 yards (45.7 meters) downstream from the
simulated spill. Hence, carbon contact time prior to booming in
this series of runs was 50 percent longer than in previously
reported runs. The locations of the application point and
sampling sites are indicated in Figure 10.
Series II - Results and Discussion
Results of runs conducted in this manner are presented in
Table 4. The coarse carbon employed was the standard 12 x 40
mesh Nuchar C-190 applied in previous tests. The fine carbon
was 60 x 230 mesh Nuchar C-190. The latter was produced by
first grinding the commercial carbon and then sieving it to
specifications. Each test involved the use of forty pounds of
carbon and four pounds of hexone as in previous trials.
25
-------
ROAD
FILL
RESERVOIR
RESERVOIR DAM AND SLUICE GATE
STREAM FLOW
M
CT\
SPILL
SITE
APPLICATION
COLLECTION
AREA
SAMPLE SITEC
OUTLET DAM AND SLUICE GATE
BOOM
-SAMPLE SITED
DISCHARGE
WOOD FLUME FOR
CANAL CROSS ING
FIGURE 10. SAMPLING SITE LOCATIONS FOR TEST SERIES #2 AND #3
-------
TABLE 4
TEST SERIES #2
RESULTS OF RUNS WITH SPILLS IN TURBULENT SECTION OF THE FLOWING REACH
to
Hexone Percent Rexona Percent Hexone Percent Percent Hexone
at Sample Remaining at Sample Remaining at Sample Remaining Hexone on Accounted for
Run Site S Run/ Site C Run/ Site D Run/ Carbon at Site D
Mo. Description of Run (total g) Background jtotal q) Background (total g) Background (total g) I Total/1816 g
1 Background 1986 100 1650 100 1408 100
2 Slurry Addition of
78.0
Coarse Carbon
1618
81
1222
74
1592
113
219
99.7
Surface Addition of
Coarse Carbon
925
47
1186
72
1242
88
162
77.3
Surface Addition
of Fine Carbon
987
SO
1201
73
1170
83
185
74.6
-------
From the data of Table 4, it can be seen that subsurface
slurry addition of the coarse carbon was not as efficient as
surface application of the coarse or fine carbons. The fine
carbon effected a better overall removal than the coarse carbon.
It is interesting to note that removal declined with travel down
the canal. This is thought to be a result of desorption as loaded
carbon traveled past the contaminant plume and became exposed to
"uncontaminated" waters. Subsequent work in the laboratory
revealed that desorption does indeed occur. This will not be
the case with all hazardous substances since some have very
shallow sorption isotherm slopes, and others undergo irreversible
adsorption.
Desorption apparently did not occur in prior runs because of
the shorter distance traveled and the stratification of the
hexone. Had the last test series been terminated at sample
point S, removal would have been comparable with that reported
for the prior runs. It is also interesting to note that much
more hexone is accounted for at all sample sites than in test
series 1 or 3. This further substantiates that better mixing
was achieved when the hexone was pumped into the water column,
and therefore samples taken at depth were representative of the
plume.
Samples of spent carbon were collected at sample site D.
These were subsequently eluted with four methanol rinses to
desorb hexone. (Previous laboratory work indicated >95 percent
recovery can be achieved with a series of four methanol washes.)
The methanol was then analyzed with the gas chromatograph to
determine the total hexone contained in the forty pounds of
carbon applied to the spill. This input then was added to the
quantity measured in the water to complete a material balance
on the hexone. Recovery was typically 74-78 percent except for
Run 2 where an extraordinary 99.7 percent was accounted for.
This figure is believed to reflect a non-representative sample
of carbon.
SERIES III
The third series of runs was performed to evaluate instan-
taneous spill application. For this series, the dye-hexone-
methanol solution was spilled instantaneously (duration ^10
seconds) at the reservoir weir. Carbon application and sample
sites remained the same and can be seen in Figure 9. Forty-eight
pounds (21.7 Kg) of fine carbon (60 x 230 mesh Nuchar C-190) was
then spread on the surface.
Series III - Results and Discussion
Results of this application are presented in Table 5. It
would appear that hexone stratification again occurred. As in
test series 1, the recovery of hexone at sample site S during the
28
-------
TABLE 5
TEST SERIES #3
RESULTS OF INSTANTANEOUS SPILL CONDUCTED AT RESERVOIR WEIR
to
Bexona Percent Hexone Percent Hexone Percent Percent Hexone
at Sample Remaining at Sample Remaining at Sample Remaining Hexone on Accounted for
Run Site S Run/ Site C Run/ Site D Run/ Carbon at Site D
No. Description of Run (total g) Background (total g) Background (total g) Background (total g) E Total/1816 g
1 Background 794 100 1364 100 1418 100 ~ 78
2 Forty-eight Pounds
Fine Carbon Applied
at Surface 714 90 1045 77 722 51 118 46
-------
backgournd run is quite low. This suggests that it was not the
turbulence at the weir that eliminated stratification in the
second test series so much as the means of introducing the hexone
to the water. In Series II, when the hexone was pumped into the
water the discharge end of the hose was placed down into the
water column. This apparently created much better mixing and
minimized the effects of stratification. The removal obtained in
the final test is comparable with that noted in the shorter runs
of test series 1. Removal clearly is enhanced with stratification.
This no doubt reflects the greater contact between the carbon and
the concentrated portion of the hexone plume at the surface. The
poor recovery of hexone in the material balance is similar to
results obtained in test series 2.
Samples of carbon were taken at each sampling site to inves-
tigate desorption. The total hexone accounted for on the carbon
was 0.02 ounces (0.67 g, 4.76 ounces (135 g), and 4.16 ounces
(118 g) for sites S, C, and D, respectively. Some desorption
appears to have occurred, but the effect was generally overwhelmed
by the effects of stratification. Desorption in this case may
be the result of either passage back into the water column or
volatilization to the atmosphere. Some degree of the latter
would tend to explain the poor material balance results. The
potential for volatilization from the carbon is greater than
that from the water itself since the black carbon absorbs a great
deal of solar radiation and thus heats the hexone directly. In
this respect, the carbon may act as a wick withdrawing hexone
from the water and releasing it to the atmosphere.
During the various test runs, it was noted that carbon loss
along the sides of the stream was not significant. Carbon
recovery with the booming system, on the other hand, was very
effective with in excess of 90 percent of both the fine and the
coarse carbon accounted for. The degree to which simple booms
can hold and direct floating carbon movement is illustrated in
Figure 11.
EFFECTIVENESS OF CARBON CONTAINMENT BOOM
An independent evaluation was made to determine the extent
to which various size fractions of floating carbon are carried
beneath the surface of the water. Seventy-five gram (2.65 ounces)
samples of carbon were slurried and added to the center of the
stream flowing at 5 cfs (141.6 liters) in the run stretch of the
FTFS. A skimming device was then employed to collect all carbon
found in the top 0.5 inches (1.3 cm) of water some 25 feet
(7.6 meters) downstream. Results are presented in Table 6. All
size fractions tested appear to be capable of mixing below the sur-
face in this section of the facility. The ratio of surface resid-
uals shows significant differences in this parameter.
Based on this preliminary work, test conditions were
selected for the parallel evaluations between floating carbon
30
-------
I
FIGURE 11. BOOMING FLOATING CARBON IN QUIESCENT SECTION OF FSTF
31
-------
TABLE 6
RECOVERY OF VARIOUS SIZED PARTICLES FROM APPLICATION
OF 75 GRAMS OF FLOATING CARBON
Mesh Range 40x60
Total Wt. 1.6954
Recovered
Percent of 2.26
Sample Applied
Ratio* 1.00
60x70 70x100 100x200 200x325
1.4215 0.6308 0.2356 0.1615
1.90
0.84
0.31
0.838 0.372 0.139
0.22
0.095
*Ratio of weight recovered to weight of 40x60 mesh sample
recovered
and sinking carbon in porous bags. The spill application employed
in test series #3, instantaneous spillage at the tail race, was
deemed the most advantageous for further testing. It simulates
actual spill conditions most closely, and minimizes variations
between tests which might result from attempts to pump the
hexone into the canal in a simulated plume. At the same time,
the flow pattern at the tail race spreads the plume out analogous
to a dispersed plume that has had time to develop. Spills made
further down the canal remained concentrated during the brief
run times available.
32
-------
SECTION 7
DEVELOPMENT OF THE POROUS BAG SYSTEM
Many of the important conceptual aspects of applying sink-
ing carbon in porous bags were addressed in early work.1*/6
Therefore, preliminary efforts in this study focused on specific
design considerations with respect to large scale production and
application in the FSTF. These activities were divided between
two objectives: bag design, and regeneration.
POROUS BAG DESIGN AND FABRICATION
While previous studies have shown that removability is
largely a function of carbon mesh size, bags for use in the FSTF
were designed to enhance removability by maximizing responsiveness
to water currents. Original specifications called for bags to be
shaped like long thin fibers with outside dimensions of
1 (2.5 cm) x 12 (30.5 cm) inches. Practical considerations,
however, dictated expansion to a width of three inches which
allowed greater ease in loading and larger amounts of carbon
per bag.
Pursuant to the goal of minimizing absorption kinetic
problems, a finer 12 x 40 mesh Filtrasorb carbon produced by
Calgon Corporation was selected for field testing. This
necessitated the use of a fabric with an ASTM mesh count of 51.
In order to allow steam regeneration of spent bags, fabric types
were sought that would withstand heat, moisture, and the hexone
solvent. After review of various commercially available materials,
a polyester monofilament screen cloth produced by Kressilk
Products, Inc. was selected for construction of the bags.
General fabrication procedures for producing the porous bags
are given in Figure 12. Sheets of screen cloth were cut to
28 (.71.1 cm) by 12 (30.5 cm) inches and sealed on three sides
with double stitching (Step A). All stitching was accomplished
with a commercial upholstery machine using heavy duty dacron
thread. Vertical double stitching was then added at 3 inch
(7.6 cm) intervals forming eight consecutive bags side by side
(Step B). Individual bags were then separated by cutting between
the double stitching (Step C). Each bag thus formed was then
charged with 62 ounces (17.5 grams) of Filtrasorb (Step D). Bags
were sealed with stitching and carbon spread evenly inside each
(Step E). Horizontal spaces were then cleared at three inch
33
-------
1
1
1
1
1
1
1
1
1
!
,."~i r _ _ i i
1
'-
.
>...__ _ j
r~_i;
,.,
i
j
J
tr:::
.<
t----l
/! [ I
It i .
/' i i
li '
II' ' '
{/I i
r '
U' i
r * i
1...J
1--.J
Step A
Step B
Step C
mjmm
Step D
Step E
Step F
Step G
FIGURE 12. FABRICATION PROCEDURE FOR PRODUCING POROUS BAGS
34
-------
intervals and stitching added to form four 3 inch (7.6 cm) squares
in each bag containing equal quantities of carbon and sufficient
free space to allow fluidization (Step F). Finally, grommets were
affixed to the one inch margins at the top and bottom of the bags
(Step G). In excess of 700 bags were produced in this manner.
A typical bag appears in Figure 13.
Grommets on each bag facilitated attachment to floats and
linear attachment of bags for deep waters. Wooden floats were
constructed from 15 inch (38.1 cm) long 2 by 2's for the tests
conducted here. Eye hooks were placed in the end of each of
these and a loop of twine threaded through forming a complete
circle around each float. Six bags, three each side, were
then threaded onto each float.1 A completed float set is
pictured in Figure 14.
POROUS BAG REGENERATION FACILITY
Because of the cost of producing the porous bags it was
determined that carbon should be regenerated after use rather
than discarded. This can be accomplished in one of two ways:
1) separation of carbon from the bags followed by regeneration,
and 2) regeneration of the carbon while still in the bag. The
latter involves the least amount of time and labor, but is limited
in that thermal regeneration is precluded. For the purposes of
the work reported here, it was determined that hexone and methanol
are sufficiently volatile to be removed by steam stripping. Since
the nylon fabric in the bags was guaranteed for temperatures up
to 300° F (148.9° C), no separation from the bags was necessary.
A steam generation unit and contact tank were constructed
for all subsequent regeneration work. A schematic of the
facility is pictured in Figure 15. Spent bags of carbon were
separated from the wooden floats and stacked in the stripping
tank. Packing was added as needed to minimize short circuiting.
Bags were then steamed over a 12 hour period.
The stripping unit was tested early in the program to assess
effectiveness. The bottom square of three individual bags were
suspended in beakers containing 1.7 ml hexone in one liter of
water. After stirring for one hour, bags were retrieved and two
units placed in the steam stripper for regeneration over a four
hour period. The third bag was held out for control purposes.
After regeneration, carbon samples were removed from the bags and
residual hexone extracted utilizing four rinses of methanol. The
extract was then analyzed by gas chromatography while the carbon
was dried at 103° C (217.4° F) and weighed. Analysis indicated
that 70 percent of the absorbed hexone was removed during the
four hour period. Based on these results, a 12 hour exposure
was selected as adequate for regenerating bags used in field
trials.
35
-------
Ul
o
FIGURE 13. COMPLETED POROUS BAG
-------
FIGURE 14. STANDARD FLOAT ARRANGEMENT
-------
V.ilVr
O
Stoan
Generator
Local
Garbage
Can
e
Stean
H.,0
Presure
^e 1 i o f
Valvi
Q
O
1/4" mesh
Screen
1 '?." Perforated
Pi DC
Condensate
Drain
FIGURE 15. SCHEMATIC OF STEAM STRIP REGENERATION FACILITY
38
-------
A preliminary run was performed on the FSTF to assure that
the bag design selected was compatible with flow conditions. No
problems were encountered with bags catching on obstacles on the
bottom of the channel. Up to 25 percent of the floats caught on
weeds or other intrusions along the shoreline during the run.
Aside from this, no operational problems were r.oted, and
subsequent studies were directed to the parallel evaluation of
floating carbon and porous bags.
39
-------
SECTION 8
COMPARATIVE STUDIES OF ALTERNATE CARBON SYSTEMS
As was the case for field application studies with floating
carbon, all comparative evaluations were conducted in the FSTF.
Sample collection and media application sites were the same as
those identified in Figure 10. Tests at each flow rate were
conducted in series and included one run each with the application
mode to be studied and a blank background run to which no carbon
was added. Hexone was spilled instantaneously at the tail race
of the reservoir gate.
Buoyant carbon was wetted and air dried prior to each run.
Application was made by manual distribution on the surface of the
stream. Porous bags were similarly air dried prior to testing,
but no wetting was deemed necessary since bags were cooled in the
steam stripping tank prior to removal. Floats were manually
placed in the stream for testing.
Field evaluations focused on determining the comparative
performance of systems with variation of two parameters flow
rate and spill concentration. Flow rate studies were conducted
for discharges of 5 (141.6 liter), 10 (283.2 liter), and
15 (424.7 liter) cfs. Results of test runs are presented in
Table 7. In general, buoyant carbon was more effective than
porous bags for removing hexone at all flow rates tested.
Differences in removal rates at specific flow rates were suffi-
ciently small to suggest an origin of sampling discontinuities.
Removal averaged 50 percent. The porous bags, on the other hand,
displayed an increasing removal capability with higher flow rates.
This implies that poor removal at low flow rates results from
poor adsorption kinetics. Increased turbulence at higher flow
rates apparently overcomes these difficulties. It is emphasized
that this effect is related to the turbulence in the stream and
not the flow rate percent. In this respect, velocities associated
with these flows are a better determinant of the conditions which
sponsor removal with porous bags than are the flows themselves.
For the flow rates tested, average velocities were in the range
0.84 (25.6 cm)-1.17 (35.6 cm) ft/sec with greater surface
velocities.
Sorption characteristics for the hexone solution on floating
carbon and a porous bag are given in Figure 16. Tests were
conducted in polyethylene bottles placed on a shaker for
40
-------
TABLE 7
COMPARISON OF IN SITU TREATMENT EFFICIENCIES AT VARIOUS FLOW RATES
Flow Rate
4.45
4. 45
6.49
6.49
12.6
12.6
10*
15
15
15
Hexone at
Mode of Sample Site S
Run (Total g)
Background
Floating
Background
Porous Bags
Background
Porous Bags
Floating
Background
Porous Bags
Floating
794
714
665
811
1,441
907
2,739
1,860
2,003
1,089
Percent
Remaining
(Run/Bkgrd)
.100
90
100
100+
100
63
100+
100
100+
59
Hexone at
Sample Site C
(Total g)
1,364
1,045
2,469
1,678
2,265
1,644
1,783
2,849
2,634
1,634
Percent
Remaining
(Run/Ekgrd)
100
77
100
68
100
73
100+
100
92
57
Hexone at
Sample Site D
(Total q)
1,418
722
1,344
1,916
2,135
1,939
1,201
2,818
2,379
1,299
Percent
Remaining
(Run/Bkgrd)
100
51
100
100+
100
91
62
100
84
46
Operational problems caused rapidly declining flow rate vhich led to overestimation of
hexone remaining at each sample site. Actual removal believed to be somewhat better
than reported.
-------
10
1000
900 -
800 ~
a
Ck
c
H
c
Q)
u
o
700 -
600
400
300
200
100
0
10:1 Carbon/Hexone
Ratio
ti»» I I I 1 1 1 1
20
40
60 80
mg/g
100
110
FIGURE 16. SORPTION CHARACTERISTICS FOR CARBON APPLICATION APPROACHES
-------
30 minutes. Results were measured based on total organic carbon
content. At an application of 10 to 1 carbon to hexone,
theoretical limits are 80 percent removal for buoyant carbon and
60 percent for porous bags. At the same time, the buoyant carbon
displays a greater potential for removal during a 30 minute
agitation period than the porous bags. It is apparent that,
kinetic difficulties associated with porous bag sorption increase
contact requirements to more than 30 minutes.
A set of test runs was also conducted utilizing a greater
spill volume and the same 10 to 1 carbon to hexone ratio. The
latter was performed under the same operating procedures as
previous investigations. Ten pounds of hexone solution were
spilled instantaneously at the reservoir tailrace. Subsequently,
100 pounds (45.4 kg) of floating carbon or activated carbon in
porous bags was added. The entire set, including the background
run, was made at a flow rate of 10 cfs (283.2 liters). Results
are compared to those standard runs at 10 cfs (283.2 liters) in
Table 8.
Sampling anomalies continued to produce apparently erroneous
estimates of removal at the upstream sampling sites, but final
removal levels were comparable for high and low spill volumes
in both approaches. It would appear that within the limits of
this evaluation, size of the spill has little or no effect on
removal efficiency for the flows tested here. A larger test
facility would be required to better delineate effects of spill
size. From the sorption characteristic data in Figure 16, one
would expect better removal under circumstances where the sorption
is more concentrated, i.e., closer to saturation.
43
-------
TABLE 8
COMPARISON OF IN SITU TREATMENT EFFICIENCIES
AT LARGER SPILL VOLUME (FLOW RATE vLO CFS)
Hexone at
Spill Quantity Mode of Sample Site S
(Ibs) Run (Total g)
4
4
4
10
10
10
Background
Porous Bags
Floating
Background
Porous Bags
Floating
1,441
907
2,739
5,350
4,174
6,740
Percent
Remaining
(Run/Bkgrd)
100
63
100+
100
78
100+
Hexone at
Sample Site C
(Total g)
2,265
1,644
1,783
5,337
7,650
3,445
Percent
Remaining
(Run/Bkgrd)
100
73
100+
100
100+
65
Hexone at
Sample Site D
(Total g)
2,135
1,939
1,201
5,316
4,711
3,118
Percent
Remaining
(Run/Bkgrd)
100
91
62
100
89
59
-------
SECTION 9
PRACTICAL ASPECTS OF IN SITU TREATMENT
IN FLOWING STREAMS
Field trials have shown that buoyant sorbents and, porous
fiber bags, if given sufficient contact time, can be effective
in removing organic materials spilled into flowing streams.
Removal, however, is highly dependent upon the prompt location of
the contaminant plume, even dispersion of media or bags over the
surface, and favorable environmental conditions.
Wind will prove to be one of the major obstacles to efficient
use of floating media or bags. Not only will air movement result
in significant effects on the contact period, it can severely
hamper collection efforts by herding the buoyant carbon or bags
away from the boom. If aerial application is attempted, wind
complications will be further amplified. It is also important to
note that collection booms have a limited operation range and
cannot deal with excessive currents (^5 knots). Therefore,
quiescent or slower moving reaches of the stream must be sought
for media retrieval.
Use of buoyant media also carries the potential for leaving
unsightly carbon residuals along shorelines and beaches. Similar-
ly, fiber bags may be left along the shore or hung up on shoals.
While these effects were minimal during field trials, they must
be considered prior to application in any public waterway.
The studies made to date suggest that removal efficiency will
be greatly affected by scaling. Small spills such as those
employed in the test program amplify the sensitivity to dosing
and environmental considerations. Larger spills are marked by
much larger spill plumes and higher concentrations. The dilute
edges of the plume represent a much smaller percentage of the
total spill. Therefore, removal in the center of the plume
where the carbon is most efficient and where movement of the
carbon does not separate it from the contaminated water is a
greater part of the total removal. This indicates that average
removal is likely to be much better on a larger scale than that
noted in static water testing. A small acid spill in a semi-
confined basin revealed very poor removal when compared to
laboratory work. A much larger pesticide spill, however, resulted
in removals comparable to those obtained in confined column work
in the laboratory.2 The edge and dilution effects become very
45
-------
important as the size of the spill is reduced, and subsequently
removal is less efficient in small scale applications.
While some of these factors suggest that removal may in
fact be better than is suggested by field trials, one important
feature of actual spills will complicate response greatly. That
is the location and tracing of the contaminant plume. In the
field trials, dye was used to facilitate accurate application.
In the field, response personnel may find it very difficult to
locate the spill and to define the boundaries of the plume. Thus,
all the media could be applied in the wrong area and hence be
totally ineffectual. This underscores the need for systems to
monitor spills. Recent work with remote sensing devices and
detection kits could satisfy this need.7'8'9'10
It has been noted that the hexone employed in the testing
program to date has a tendency to stratify when released in
water and, in so doing, complicates sampling. Volatility and
reversible adsorption characteristics add to the uncertainty of
analytical results. While these properties make hexone a
difficult material to study carefully, it must be realized that
they are shared by many hazardous substances and hence
reflect real problems encountered in the field. Thus the data
is complicated by incomplete recovery of all material and the
necessity to look at apparent removal as opposed to absolute
removal. The rhodamine dye, on the other hand, represents a
conservative substance. It mixes well in the water, it undergoes
no rapid degradation or volatilization, and it adsorbs onto the
carbon with little apparent desorption. Removal is consequently
much better for this substance (>95 percent versus 20-50 percent).
Many hazardous substances will behave as the rhodamine dye when
spilled and will therefore show much higher apparent removal
efficiencies.
It is obvious that under no conditions will removal ever be
complete. Therefore, one cannot assume that a response effort
will eliminate a spill. It will only reduce its impact. Certain
circumstances can be expected to maximize that reduction, and
those are the cases where response should be promoted. From
the observations made in this study, buoyant carbon is preferable
to porous bags in the flowing stream system when contact times
are very short (30 minutes) just as it was in static waters.
Because the major difference appears to be one of kinetics, the
longer contact times achievable in natural waters may change this
somewhat.
Application of either buoyant carbon or porous bags may pose
a logistics problem. Dr. Allen Jennings of the U.S. EPA Hazard-
ous Substances Branch estimates that spills average 3,500 pounds
(1.6 grams). At a ratio of 10:1 carbon to contaminant, this
would require 35,000 (15.9 grams) of media or 17.5 tons (15.9 kg).
If air transport is desired, this will require specialized
46
-------
equipment. Larger spills may exceed the present transport
capacity and thus suggest the potential use of containerized
media supplied by shuttle to the response component.
All of the considerations offered here have bearing on the
use of in situ response techniques on real spill events. They
are pointed out to facilitate informed decisions about spill
mitigation. Despite the restrictive nature of some of these
observations, both the floating carbon approach and the porous
bag approach are deemed promising for given spill scenarios.
Present information needs center on the evaluation of these
techniques under actual spill conditions.
47
-------
REFERENCES
1. Shuckrow, A. J., B. W. Mercer and G. W. Dawson. "The
Application of Sorption Processes for in situ Treatment of
Hazardous Material Spills," in Proceedings of the 1972
National Conference on Control of Hazardous Material Spills,
Houston, TX, March 21-23, 1972.
2. Mercer, B. W., A. J. Shuckrow and G. W. Dawson. "Treatment
of Hazardous Material Spills with Floating Mass Transfer
Media," U.S. Environmental Protection Agency, EPA-670/2-
73-078, September 1973.
3. Mercer, B. W., A. J. Shuckrow and G. W. Dawson. "Application
of Floating Mass Transfer Media to Treatment of Hazardous
Material Spills," presented at the 46th Annual Water
Pollution Control Federation Conference, Cleveland, OH,
October 4, 1973.
4. Ziegler, R. C. and J. P. Lafornara. "In Situ Treatment
Methods for Hazardous Material Spills," in Proceedings of
the 1972 National Conference on Control of Hazardous
Material Spills, Houston, TX, March 21-23, 1972.
5. Dawson, G. W. and M. W. Stradley. "A Methodology for
Quantifying the Environmental Risks from Spills of
Hazardous Material," presented at the AIChE Conference
Boston-Sheraton, September 8, 1975.
6. Pilie, R. J., R. E. Baier, R. C. Ziegler, R. P. Leonard,
J. G. Michalovic, S. L. Pek, and D. H. Bock, "Methods to
Treat, Control and Monitor Spilled Hazardous Materials,"
U.S. Environmental Protection Agency, EPA 670/2-75-042,
June 1975.
7. Kirsch, M. J. J. Vrolyk, R. W. Melvold, and J. P. Lafornara,
"A Hazardous Material Spills Warning System" in Control of
Hazardous Material Spills, Proceedings of the 1976 National
Spills, Proceedings of the 1976 National Conference on the
Control of Hazardous Material Spills, Information Transfer,
Inc., Rockville, Maryland, April 1976.
8. Silvestri, A., A. Goodman, L. M. McCormack, M. Razulis,
A. R. Jones, Jr., M. E. P. Davis, "Detection of Hazardous
Substances" in Control of Hazardous Material Spills,
48
-------
Proceedings of the 1976 National Conference on Control of
Hazardous Material Spills, Information Transfer, Inc.,
Rockville, Maryland, April 1976
9. Silvestri, A., A. Goodman, L. M. McCormack, M. Razulis,
A. R. Jones, Jr., M. E. P. David, "Development of a Kit for
Detection of Hazardous Material Spills Into Waterways."
Department of the Army, Edgewood Arsenal Special Publication
ED-SP-76023, August 1976.
49
-------
TECHNICAL REPORT DATA .
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-77-164
3. RECIPIENT'S ACCESSIOr
TITLE AND SUBTITLE
In Situ Treatment of Hazardous Material Spills in
Flowing Streams
5. REPORT DATE
October 1977
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
G.W. Dawson, B.W. Mercer, R.G. Parkhurst
8. PERFORMING ORGANIZ,
PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle-Northwest
Richland, WA 99352
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
68-03-0330 &
68-03-2006
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Lab.
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
- Gin., OH
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Two methods of applying activated carbon adsorption treatment to flowing
streams were evaluated under comparable conditions. The first involved
sub-surface introduction of bouyant carbon into the water column followed
by the floating of the carbon to the surface and subsequent removal using
conventional surface skimming techniques. The second involved the addition
to the water of non-bouyant granular activated packaged in porous fiber bags
("tea bags") which were attached to floats. The bags were allowed to travel
with the spill plume for a given distance and were subsequently removed
manually. Controlled field experiments using n-hexane as the test chemical
were conducted at various flow rates in a specially modified abandoned
irrigation channel at the Energy Research and Development Administration's
Hanford site and showed that for "low-flow" non-turbulent conditions the
bouyant carbon technique was more effective in removing the chemical from
the water with only tolerable amounts of the carbon remaining in the
stream. As the flow and turbulence increased the pollutant removal
effectiveness of the "tea bag" approach improved.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Water Treatment, Activated Carbon
Treatment, Hazardous Materials, Decon-
tamination, Water Pollution
Hazardous materials spil!
clean-up, Activated
carbon "tea bags", bouyai
activated carbon, "In
situ" hazardous chemical
spill treatment
13B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
58
2O. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
50
U.S. GOVERNMENT PRINTING OFFICE: 1977-757-140/6583
------- |