EPA-R2-73-185
MARCH 1973 Environmental Protection Technology Series
Control of
Hazardous Chemical Spills
by Physical Barriers
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of"
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-R2-73-185
March 1973
CONTROL OF HAZARDOUS CHEMICAL SPILLS
BY PHYSICAL BARRIERS
by
J. V. Friel
R. H. Hiltz
M. D. Marshall
Project 15090 HGP
Contract 68-01-0100
Project Officer:
Ira Wilder
Environmental Protection Agency
Edison Water Quality Research Laboratory, NERC
Edison, New Jersey 08817
Prepared for:
OFFICE OF RESEARCH AND MONITORING
U. S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D. C. 20460
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402
Price 90 cents domestic postpaid or 66 cents QPO Bookstore
-------
EPA Review Notice
This report has been reviewed by the
Environmental Protection Agency and
approved for publication. Approval
does not signify that the contents
necessarily reflect the views and
policies of the Environmental Pro-
tection Agency, nor does mention of
trade names or commercial products
constitute endorsement or recommenda-
tion for use.
ii
-------
ABSTRACT
The magnitude of potentially hazardous chemicals now being
transported through the country poses a serious threat to
the water ecosystem. Unless spills can be controlled at
their source, movement into the water system may be in-
evitable. Such control dictates the availability of systems
capable of forming dikes or flow diverting barriers either
as a portable system carried on the vehicle or a mobile
unit rapidly deployable to the site. In this regard, a pro-
gram was instituted to investigate the applicability of
foamed materials for forming such dikes and barriers. It
was successfully demonstrated that polyurethane could be
packaged in a portable unit and dispensed as a low density
rigid foam capable of diking liquids on a variety of sub-
strates (for example, dry concrete, asphalt, bare ground,
vegetation). Attempts to develop a rigid high expansion
system were not fully successful. A foamed concrete system
was also successfully evolved, which used mobile equipment
to build free form dikes. Modified surfactant foam was
also shown to be an effective cover over spilled chemicals
to control vapor release and fire hazards. In each case,
a field tested unit was demonstrated or shown to be feasible.
This report was submitted by MSA Research Corporation in ful-
fillment of Project Number 15090 HGP, Contract 68-01-0100
under the sponsorship of the Office of Research and Monitoring,
Environmental Protection Agency.
i i i
-------
TABLE OF CONTENTS
Section Page
I CONCLUSIONS 1
II RECOMMENDATIONS . 3
III INTRODUCTION 5
IV POLYURETHANE FOAM 9
System Output 9
Foamed Barrier Studies 12
Chemical Compatibility 19
Adhesion to Cold, Wet or
Contaminated Surfaces 20
Storage Tests 23
V LOW-EXPANSION INORGANIC FOAM 27
Preliminary Evaluations 28
Small Scale Barrier Pours 34
Field Tests 45
Summary 50
VI HIGH EXPANSION SYSTEMS 53
Water Soluble Systems 53
Conventional Rigid Foams 58
Evaluation Tests 60
Additional Development and Evaluation . . 63
VII HIGH EXPANSION FOAM COVERS 67
Foam Resistance to Wind 67
Foam Candidates 68
New Foam Development 73
Expanded Rigid Systems 78
VIII SYSTEMS SUMMARY 81
Polyurethane 81
Foamed Concrete 82
Foam Covers 89
IX REFERENCES 91
-------
FIGURES'.
No.. Page
1 PORTABLE POLYURETHANE SYSTEM WITH A CLOSE
UP OF NOZZLE ARRANGEMENT 10
2 CONCRETE PAD WITH CURB DRAIN 13
3 POLYURETHANE BLOCK OF THE CURB DRAIN 14
4 SINGLE POLYURETHANE DAM SET ON BARE GROUND 17
5 DOUBLE POLYURETHANE DAM SET ON BARE GROUND 18
6 TYPICAL TIME VS DELIVERY FOR 0.6 PCF POLY-
URETHANE FOAM SYSTEM 25
7 MASON FLOW MIXER 35
8 FLOW MIXER SHOWING MOYNO PUMP AND
DELIVERY SYSTEM 35
9 FOAM CONCRETE TEST POUR SHOWING HEIGHT AND
ANGLE OF REPOSE POSSIBLE WITH SILICATE SYSTEM 44
10 FIELD TEST POUR SHOWING BUILD UP OF THE DIKE
AND VIEWS OF THE BARRIER WITH IMPOUNDED WATER 47
11 POUR OF FOAM CONCRETE USING MSA BATCH
TYPE MIXER 49
12 PROPOSED EMERGENCY FIELD UNIT 52
13 LABORATORY SET UP FOR SCREENING OF HIGH
EXPANSION FOAM CANDIDATE COMPOSITIONS 69
14 LABORATORY FOAM GENERATOR SYSTEM 71
15 VAPOR SUPPRESSION OF TYPICAL HAZARDOUS
MATERIALS BY PECTIN MODIFIED FOAM COVER 76
16 SCHEMATIC OF FOAMED CONCRETE SYSTEM 83
vi
-------
TABLES
No.- Page
1 ADDITIVES TO THE BASE FORMULATION 21
2 SELECTED CHEMICALS APPLIED AS A PRE-COAT
FOR SORPTION OR DRYING 22
3 SUMMARY OF THE EVALUATION OF THE SET OF
GYPSUM FORMULATIONS WITH ACCELERATORS 29
4 SUMMARY OF THE EVALUATION OF PVA-GELLED
GYPSUM SYSTEMS 30
5 SUMMARY OF THE EVALUATION OF CEMENT-
GYPSUM SYSTEMS 31
6 SUMMARY OF THE EVALUATION OF CEMENT-
SILICATE SET SYSTEMS 33
7 MASON FLOW MIXER TEST DATA 37
8 SUMMARY OF DEVELOPMENT STUDIES 42
9 EFFECTS OF ADDED POLYMERS ON PVA-CONGO
RED SOLUTIONS 57
10 TESTING OF SUPPORT MATERIAL TO
IMPROVE ADHESION 65
11 COMPARISON OF THICK VERSUS THIN LAYERS
OF FOAM COVER ON LIQUID TOLUENE 72
12 TYPICAL PROPERTIES OF PHENOLIC SPHERES 80
13 POLYURETHANE SYSTEMS 81
14 MOBILE FOAMED CONCRETE EMERGENCY UNIT 85
vii
-------
SECTION I
CONCLUSIONS
The conclusions to be drawn from the work conducted during
this program primarily relate to the feasibility of the
systems evolved. In three of the four phases, the basic ob-
jectives of the program were realized.
In the first phase, that concerned with polyurethane, it was
demonstrated that a rigidized foam could effectively block
or divert spills of a number of chemicals if applied under
the correct conditions. These are water based liquids ex-
cept strong acids, nonpolar organics, chlorine, and ammonia.
Acids and polar organics are a question. The present formu-
lation is not effective against methyl alcohol but is against
acetone. The material is most effective on dry hard surfaces,
but provides limited control on dirt, gravel, or vegetated
ground. Adhesion to wet surfaces is poor and the diking of
flowing streams is not possible. Some improvement with ad-
hesion to wet surfaces was realized late in the program.
The urethane foam system can be packaged in a man-portable
unit to yield an average of 5 cfm. Foam can be generated
at densities as low as 0.3 lbs/ft3 but 0.6 to 0.7 lbs/ft3
is the optimum compromise between strength, adhesion, and
foam yield. In this range, 60 to 70 cubic feet of foam are
realized from the portable unit.
In the second phase, inorganic foam systems using concrete
and a silicate gelling mechanism were evolved which allow
the construction of barriers up to 2 ft in height without
forms or supports. The height is limited by the ability of
the gel set to support its own weight. Greater heights can
be achieved with the development of some concrete set which
has much greater strength but requires some 10 minutes cure
time for its development. The material can be generated
with commercially available continuous mixing equipment
which is compact enough to be truck mobile. Final concrete
density can be as low as 30 lbs/ft3. The set material
appears inert to most chemicals and attacked only slowly
by the rest. With the correct procedure it can block flow-
ing streams.
In Phase 4, soft detergent foam systems were shown to be
capable of providing an insulating cover to hazardous spills
reducing the ignition potential and toxic vapor release.
Imine modified medium expansion foam 100 to 200:1 served
adequately for most of the chemicals investigated. These
-------
imine foams were not successful with polar compounds. To
control polar compounds, a low expansion system 50:1 in-
corporating a gellant was necessary. Of the available
gelling materials, pectin provided the best results.
In Phase 3, attempts to evolve a high expansion (150:1 or
greater) foamed material for diking were unsuccessful.
Gelling of polymer foams such as polyvinyl alcohol (PVA)
did not provide a sufficiently rigid material to be useful.
Isocyanate-water foams rigidized but were permeable and
lacked integrity. At higher density they approximated ure-
thanes but were inferior in strength, adhesion, and im-
permeability.
-------
SECTION II
RECOMMENDATIONS
The initial program has defined the potential of three
systems: polyurethane, foamed concrete, and foam covers.
For each of these materials sufficient data has been gen-
erated to allow delineation of a commercial package. Before
such systems can be put into general use, however, further
developmental work,to optimize the units and define their
capabilities and limitations, appears warranted.
With the polyurethanes, significant changes in formulation
have been made to adapt this material to the diking applica-
tion. In this regard some problems have been experienced
with shelf life of the packaged systems. Changes in material
combinations would appear to be able to correct this de-
ficiency. The allowable shelf life must also be determined.
The polyurethane foam once rigidized has shown good resis-
tance to a variety of chemicals. The number of chemicals
tested was limited to chlorine, methanol, acrylonitrile,
toxaphene (90% in xylene) and phenyl mercuric acetate
(ammoniacal water solution). Only one material to date,
methyl alcohol, has shown any adverse reaction. An eval-
uation of a broader spectrum of material appears warranted,
however, to define the applicability and limitations of the
foam. These tests should also include evaluation of the
foam density. This property is determined in part by the
degree of crosslinking and hydroxyl substitution, features
which may be significant with respect to chemical inertness.
Towards the end of the initial program, two techniques were
developed which markedly improved adhesion to surfaces wet
with water. It is necessary to determine if these pro-
cedures are beneficial in situations where the wet surface
is due to the spilled chemical. Sufficient testing should
be conducted to define a procedure covering the broadest
range of wet surface conditions. The selected procedure
would then be incorporated into the portable unit, either
directly into the chemical package or as an auxiliary unit.
Like the polyurethane, the testing of the chemical inertness
of foamed concrete has been limited and a more comprehensive
spectrum of materials should be examined. Foamed concrete
does have the capability to dam flowing streams. The de-
tails for such emplacement varies with conditions and sub-
trate. Some further work appears warranted to determine
-------
the capability to place foamed concrete dams in existing
bodies of water such as creeks, drainage ditches and the like
The principal recommendation with respect to foamed concrete
is the adaptation of generation to batch-type mixers. The
majority of work to date has been performed with continuous
mixers. These are more versatile than batch operations, but
require level positioning and are not very amenable to in-
accessible areas. Batch mixers have advantages in this re-
gard. Some work has been performed to demonstrate feasi-
bility. Difficulties were experienced in achieving the
necessary control of materials' proportions. The development
of the procedures using the batch type unit would allow
flexibility in the selection between a continuous or a
batch mixer to suit the topography and other conditions at
the spill site.
The potential of soft detergent foam covers has been demon-
strated on small spills in the laboratory and in a few iso-
lated instances in the field. True proof of the advantages
of such a cover requires much greater testing on large scale
spills. Such parameters as rate of cover, adequacy of pro-
tection, rate of foam deterioration/rate of recover and at-
mospheric effects (wind, precipitation, etc.) must be
measured.
The medium expansion imine modified system can be generated
with commercially available generators. The low expansion
pectin-fortified system has not lent itself readily to the
common low expansion generators. It may be necessary to
devise a generating system for application of that material.
Foam pumps are a new item on the market that have the po-
tential of handling the pectin system and should be eval-
uated before undertaking development of a new generating
mechanism.
-------
SECTION III
INTRODUCTION
It Is immediately obvious in considering hazardous chemicals
that the principal deterrent to the damage of the ecology
from such materials is the strict control of ^their manu-
facture, storage and transportation. The first step, of
course, is strict rules designed to prevent the spills of
such materials. Even the most stringent rules, however,
will not completely eliminate the possibility of accidental
spills. Procedures must be available to control such dis-
charges such that they cannot damage the surroundings or
reach local waterways. Such control will be a major problem
if such spills occur in the transportation mode. With land
vehicles at least there are potential solutions which would
tend to minimize, if not eliminate, the possibility of such
spilled materials entering adjacent water systems.
In past efforts, the MSA Research Corporation (MSAR) has been
associated with two programs to study the problems of hazard-
ous chemical spills in unconfined areas. The first was con-
cerned with the spill of liquid rocket fuels in field com-
plexes', the second with spills or leaks of chemical warfare
agents during transportation or in the field?, in both of
these studies it was clear that immediate action had to be
taken if there was any hope of containing or controlling the
spill at all. Three basic steps were considered: (1) sealing
of the leak if possible, (2) containment of the spilled
material and (3) physical removal of the spill.
It was concluded from these studies that control of the spill
of the hazardous chemicals from transporting vehicles must
be considered as one major area in the prevention of hazardous
chemical materials from damaging the water ecosystem. Based
upon our prior work and our current knowledge in these areas,
it was our belief that the available technology would allow
the successful evolution of both portable and mobile systems
which could provide surface containment of spilled materials
at the site of the accident either by the formation of dikes
or flow-diverting barriers. Technology was also available to
provide in situ generation of surface covers to prevent •
vapor losses from such spills. Completely contained, the
spilled materials can then be cleaned up by a number of
techniques. Subsurface losses, seepage and the like, would
require additional controls.
In considering the erection of dikes or diversionary barriers
the Initial thought 1s usually to natural materials. Al-
-------
though significant use should be made of natural barriers
and depressions by diverting spills toward them, their use
to construct dikes or other barriers pose a number of prob-
lems. Not all types of surface materials are amenable to
the formation of dikes of any significant height. When the
ground is frozen or water soaked, movement by hand is diffi-
cult if not impossible, and under the worst conditions even
heavy equipment may not be sufficient.
To provide a general purpose system operable under all of
the environmental conditions to be experienced, it becomes
necessary to consider structures built up out of artificial
materials. Whether these materials form an enclosure or
divert the flow of spilled materials, the basic requirements
are the same. Materials should be resistant to chemical
attack, nontoxic, disposable and fire retardant if not in-
flammable. Most important, maximum benefit should be rea-
lized from as small an amount of material as possible using
portable or mobile equipment with small power requirements.
In assessing the requirements, the limiting value would
appear to be the necessity of obtaining large quantities of
material rapidly from small portable or mobile systems.
These conditions appear best satisfied by materials which
can be foamed or expanded. At present two material classes
possess the basic requirements. These are polymer foams
and so-called foamed inorganic materials.
Although a large number of polymers can be produced in the
foamed or expanded condition, only polyurethane appears to
possess the necessary characteristics for field operable
units. Formed by the exothermic reaction of two materials,
the heat generated volatilizes and expands a reaction product
or an added blowing agent which creates the foam. Poly-
merization occurs at the same time, setting the foam to a
rigid mass. The only power necessary is pressure to expel
and blend the two components. Other polymer foams require
mechanical agitation and may be of open cell structure and
thus form a porous medium.
There is a broad spectrum of inorganic materials which can
be produced in lightweight form by blending them with a pre-
formed detergent or protein foam. The major item in this
class at the present is foamed concrete. Foamed gypsum is
also an available material,as well as sodium silicate foam.
The latter, however, requires mechanical agitation, is
partially open-cell and subject to hydrolysis.
Once contained, rapid clean up of the spill is almost man-
datory. Vapor release to the air and seepage into the ground
-------
constitute a significant material loss which poses immediate
problems to the surrounding environment. Ground-water move-
ment can transport spilled material into the water system.
Large exposed surfaces of spilled chemicals also pose an
immediate threat, in terms of toxicity and flammability, to
the personnel attempting to control and clean up the chemical
as well as other individuals in downwind locations.
Isolation of the chemical surface along with control of boil-
off would appear to be desirable. The surface covering should
be such that it does not interfere with the clean up opera-
tion. In the past year foam has been shown to be effective
as a surface cover to isolate and control vaporization rate
of selected chemicals such as gasoline, ammonia and vinyl
chloride monomer.
Based upon the existing technology, a program supported by
the Environmental Protection Agency was undertaken by MSAR
in four phases to evaluate and potentially evolve foam
systems applicable to the containment of hazardous chemical
spills. The first phase was concerned with polyurethane
foam, the second with foamed inorganic materials, the third
with rigidized high expansion materials and the fourth with
soft foam covers. The following sections present a detailed
reporting of the work conducted in each phase, the program
accomplishments, and, where applicable, recommendations for
additional work to further evolve the concepts.
-------
SECTION IV
POLYURETHANE FOAM
One of the strategies for the control of hazardous chemical
land spills depends on the availability of a small, portable
system which could be carried on transporting vehicles and
utilized by the carrier personnel. Polyurethanes, already
commercially available as a portable foam dispensing system,
potentially offered a quick response solution. Polyurethanes
are resins derived from the catalyzed reaction between polyols
and isocyanates. Once formed, they are reasonably inert to
most other chemicals. The foamed form, which is currently
used in a wide variety of applications, is obtained by blend-
ing a low boiling liquid, usually a Freon, into the reaction
mass.
A typical portable, man-carried unit is shown in Figure 1.
This is a packaged two tank unit. This size delivers some
10 cu ft of foam with an expansion of 25:1 and an output of
approximately 1 cu ft/min. This unit would appear to be
quite suitable to fill the requirement for a small, portable
system to provide emergency diking. The only deficiency in
the unit is the low output rate. A discharge of 4 to 5 cfm
would be a more suitable situation for spill control. With
the existing formulations and portable packages as a base,
a program was undertaken to adapt this device for use in the
control of hazardous chemical land spills. Three major
items were considered:
1. Increasing the discharge rate
to an average of 5 cfm.
2. Limited evaluation of the
chemical compatibility of the
urethane with^hazardous chemicals.
3. Determination of the effectiveness
of the material as a barrier against
a variety of substrates.
System Output
The majority of current applications for urethane foam re-
quire a high quality material. To achieve this quality, it
is necessary to obtain intimate mixing of the two components
in the system just prior to discharge from the nozzle. The
current design of-the spray nozzle utilizes helical vanes
within the body of the nozzle coupled with a small (1/32 in.)
orifice. Although this is extremely effective in obtaining
good mixing, it does severely restrict the rate of discharge
-------
FIGURE 1 -
PORTABLE POLYURETHANE SYSTEM WITH A CLOSE UP
OF NOZZLE ARRANGEMENT
10
-------
of the urethane material. If the mixing helix is removed
and the nozzle orifice enlarged to the maximum possible,
which is some 7/64 inch, discharge rates averaging better
than 7 cfm can be achieved, but the quality of the foam is
reduced.
Current urethane systems use a combination of the blowing
gas, usually Freon, and nitrogen to develop sufficient
pressure within the system to discharge the two components
through the nozzle. In the initial portions of this phase
of work serious consideration was given to increasing both
the quantity and the pressure of the nitrogen. Discharge
rates which could be achieved in the absence of the mixing
vanes raised the possibility that outputs in the range of
4-5 cfm could be achieved without the additional pressure.
Increasing the nitrogen pressure would require increases in
tank walls with an attendant weight increase. This is an
undesirable situation where a man-carried package is con-
cerned.
It was necessary to determine if the mixing system could be
changed without reducing the quality of the foam to an un-
acceptable level. It was certain that the diking system
did not require the foam quality of the existing applications,
but poorly reacted material might lack the strength or chemi-
cal inertness to survive in the diking application. Two
changes in the internals of the nozzle resulted in improved
outputs with a foam quality which appeared to be more than
adequate for the application. In one arrangement, the
helix was shortened from its normal 50 mm to 35 mm to pro-
vide less resistance to flow. The resulting average output
from the system was some 4 cfm. It is to be noted here that
discharge rate of the package decreases with time due to the
expenditure of the pressurizing gas; thus an average output
of 4 cfm represents an initial discharge rate of the order
of 7 cfm, decreasing to something less than 3 cfm as the
final amount of foam is discharged. Other discharge rates
show a similar variation.
The second arrangement which gave a slightly poorer quality
foam but an average output in excess of 5 cfm involved re-
moving the helix completely and replacing it with a double
cone. This provided a second orifice within the body of the
nozzle by narrowing the inside diameter of the nozzle from
16 mm to 4 mm over a 55 mm length.
When the 5 cfm average output was achieved, work on im-
proving the system output was terminated. Determination of
which of the two arrangements was acceptable could only come
from field studies which would provide a practical evaluation
11
-------
of the quality of the foam being produced. These field
studies have shown that both arrangements are acceptable.
The double cone system was selected because of its slightly
higher output rate.
Foamed Barrier Studies
Concurrently with the previous studies, work was initiated
on evaluating the ability of the urethane foam to bond to
a variety of substrates and to contain a significant liquid
head. This work was initiated before full evolution of the
portable unit since it was felt that the bonding ability and
ultimate strength of the foam were reasonably independent of
the rate of generation.
A variety of natural substrates ranging from bare ground to
dense growth were available for evaluation along with gravel
and paved surfaces. To provide a broader spectrum of con-
ditions, however, tv/o additional installations were fab-
ricated. The first was a concrete pad incorporating a
typical curb type drain; the second was an asphalt surface
incorporating a highway storm grate. This facility is shown
in Figure 2.
The initial work was with hard surfaces -- asphalt, concrete,
and hard packed dirt -- and significant success was realized
in applying the urethane and successfully containing sig-
nificant volumes of water 2 to 3 ft in depth. A circular
form of polyurethane 36 inches high and 30 inches inside
diameter was set up on concrete and maintained full of water
without leakage for a period of days. Similarly, polyurethane
was used to block the curb drain set in concrete as shown in
Figure 3. Sealing to asphalt was as good if not better than
sealing to the concrete.
Similarly, effective seals could be accomplished against
bare ground. The ability of the polyurethane to dike liquids
in that particular situation, however, depended upon the in-
tegrity of the bare ground. A breach of the dike eventually
occurred. The breach was always in the substrate just below
the level of the urethane dam and usually restricted to a
small localized area. The time required to achieve the
breach was a function of the head of water and the density
of packing of the substrate. Even though the dike was
eventually breached, several inches of water could be held
for an hour or more. In the case of bare ground, where poor
containment was experienced, additional downstream dikes
could be erected to extend the control time of the liquid
material.
12
-------
FIGURE 2 - CONCRETE PAD WITH CURB DRAIN
13
-------
FIGURE 3 - POLYURETHANE BLOCK OF THE CURB DRAIN
14
-------
As the test sequence moved from dry bare ground to gravel,
vegetated areas and wet surfaces, the effectiveness of the
urethane as a liquid dike diminished rapidly. With wet sur-
faces the problem was lack of adhesion and with the gravel
and the vegetated ground, the inability of the urethane to
penetrate more than a few fractions of an inch below the
top surface.
The difficulties of bonding urethane to wet surfaces proved
to be a major difficulty and is discussed in a separate
section of this report. The inability of the foam to pene-
trate very far into porous or vegetated surfaces proved more
amenable to solution. In polyurethane technology, the term
"cream time" is used to denote that interval between the
mixing of the chemicals and the onset of foaming. In normal
formulations, cream time varies between 2 and 4 seconds.
This means that some frothing has already begun as the
material leaves the nozzle. As frothing proceeds, material
viscosity increases, thus as the material strikes the surface
it does not have the fluidity necessary to penetrate and tend:
to bridge across the high points.
Cream time can be varied by changing the reaction rate and/or
the volatility of the blowing agent. Such variations were
studied in the program. It was found that significant change;
were possible which provided not only for a higher expansion
of the urethane without significant loss of its other de-
sirable properties but that variations in cream and rise time
alleviated the difficulties initially experienced with gravel
and vegetated surfaces.
By making minor changes in the catalyst from triethylamine
(TEA) to dimethylethylamine (DMEA) and a major change in the
blowing agent from a Freon type material to methylene chlo-
ride, a rigid polyurethane foam was achieved with a density
between 0.5-0.6 pcf rather than the normal 1-2 pcf. The
rise time of this particular form of polyurethane foam was
much more rapid than that of normal materials but the cream
time, that interval before the onset of foaming, was in-
creased to some 12 to 14 sees from the normal 2 to 4 sec
increment. It is this last feature which allowed a better
penetration into vegetated surfaces and gravel.
This particular system had one apparent drawback. Because
of the exceptionally fast rise and the concentrated heat
release, there was a tendency of the foam to shrink slightly
upon cooling from the reaction temperature. Field tests,
however, showed that the bonding strength to dry surfaces
of concrete, asphalt and bare ground was sufficient that
15
-------
the shrinkage did not adversely affect the material's ability
to contain liquids. Further, the cream time was sufficiently
long that penetration did occur into vegetation to the point
that effective seals could be achieved against the surface
as long as the vegetation was not too thick and matted. Im-
proved penetration was also achieved on gravel surfaces but
never to more than some 2 diameters of the stone. It is
doubtful if this material could be effectively utilized to
control liquid spills on a gravel surface.
The benefits due to the increased expansion and the extended
cream time of the new formulation without sacrificing any of
the desirable characteristics exhibited by the normal 1-2 pcf
foam led to the adoption of this particular material as a
standard material in this phase. The use of methylene chloride
as the basic blowing agent produced some changes in materials
viscosity. To maintain the 5 cfm delivery rate, and to get
good mixing in the nozzle of the correct proportions of both
resin and activator, minor changes were necessary. To achieve
the desired 1:1 resin-activator proportions, the cross-
sectional area ratio of resin and activator orifices was
changed to 2.3:1. Some mixing problems were encountered but
were overcome by adding 10% by weight of Freon R12 to both
resin and activator. Volatilization of this material in the
nozzle provided for good mixing. With the Freon-nitrogen
pressure system, 200 psi had to be used as the initial tank
pressure. With the methylene chloride, the pressure was in-
creased to 220 psi.
A significant number of field studies were run using this
material on a variety of substrates to block and hold water
with good results except when wet or extremely cold surfaces
were involved. Figures 4 and 5 show dams erected with this
type of polyurethane material. By expanding this material
into a rubber or plastic envelope it was possible to block
water flowing in culverts and other similar circular drains.
The final effort in this portion of the program was an eval-
uation of the ability of the polyurethane materials to block
or divert already flowing streams. These efforts were com-
pletely unsuccessful. The primary difficulty was the in-
ability to achieve good adhesion between urethane and a
surface wet by water or one of the hazardous chemicals.
Several mechanical aids were tried, such as the use of hold
down stakes; heavy filler, such as rock; or a support system
such as screen or netting. None of these were successful in
improving the ability of the polyurethane to combat the
flowing stream situation. In addition, they complicated the
construction of the dam appreciably. This is a major dis-
advantage when it is considered that the portable unit must
16
-------
FIGURE 4 - SINGLE POLYURETHANE DAM SET ON BARE GROUND
17
-------
ysrv 7
FIGURE 5 - DOUBLE POLYURETHANE DAM SET ON BARE GROUND
18
•
•
-------
be designed to be used by essentially untrained people in
emergency situations. Work in this area was thus terminated.
Chemical Compatibility
All existing data would indicate that polyurethanes, once
rigidized, would be essentially inert to a majority of the
long list of chemicals now appearing in the'hazardous chemi-
cal classification3. The substantiating data is relatively
meager, however, and it appeared necessary to conduct some
tests v/ith representatives of several classes of materials
which had a high order of ranking in the current listing of
transported hazardous materials. In addition, there is a
question as to the effects the hazardous materials might have
on the reaction and rigidizing of urethane foam. Thus, tests
were necessary in this regard, also.
Five chemicals representing types of hazardous chemicals
shipped in large quantities in the United States were selectee
for reactivity tests. These were chlorine, methanol, acrylo-
nitrile, toxaphene (90% in xylene) and phenyl mercuric acetate
(ammoniacal water solution). With the exception of methanol,
none of these materials exhibited any significant reaction
with the urethane foam once it had rigidized. Methanol did
not react but tended to be absorbed by the foam. Tests on a
large scale which were conducted as part of the study of
blocking flowing streams, verified the laboratory findings.
The foam behavior with methanol is not understood. It is not
a destructive reaction but rather one of absorption. The ab-
sorbed methanol does destroy foam rigidity, however. Tests
run with acetone, a polar compound like methanol and acrylo-
nitrile which is also polar in nature did not produce this
effect. Further work is necessary to explain the phenomenon
and to define what other materials might behave as methanol.
At the present, polyurethane is not recommended for contain-
ing methanol.
To determine the effect of the hazardous materials in their
liquid form on foam generation and rigidizing, a second
series of tests were run where foam was generated directly
into a pool of each of the chemicals. In each case there
was a reaction between the urethane and the hazardous chemi-
cal. It was never violent but it destroyed the urethane
before it could foam and rigidize.
To assess the effects of the hazardous chemicals' gas phase,
foam was applied to the edges of a tray 10 in. x 10 in. x 4
in. holding the chemical. It was possible for the urethane
to set up and bridge each of the chemicals, with the ex-
19
-------
ception of chlorine. It can thus be concluded that vapors
emanating from the chemical were not sufficient to impede
urethane formation. In the case of chlorine and ammonia the
heat of urethane reaction produced rapid boiling. Even with
boiling, the ammonia pool was covered and even an hour later
there was still liquid ammonia in the bottom of the tray.
With chlorine, however, the issuing vapors kept a blow hole
open in the urethane cover until all chlorine had been
vaporized.
A much broader range of chemicals needs to be evaluated, but
on the basis of this limited data it would appear that ure-
thane foam can control spills of a variety of hazardous chemi-
cals. It is not feasible to generate foam directly onto the
spilled materials but vapors in the surrounding area will
probably not have any significant effect on the generation
or set of the foam.
Adhesion to Cold, Wet or Contaminated Surfaces
The degree of success achievable by any hastily erected
barrier is directly linked to the ability to anchor the
barrier to the existing surfaces. As has already been ex-
plicitly stated, in its present state of development a ure-
thane type formulation is not practical on wet or contami-
nated surfaces. Cold surfaces also pose a problem but of
a lesser magnitude. The presence of water, alcohol, etc.
act as chemical interferants and disrupt the stoichiometry
of the applied urethane mix. Oil-like materials, merely
because of their physical presence, block any interaction
with the surface. Cold surfaces interfere with the exotherm
necessary for foam formation. A specialized effort was
directed to solve the problem of anchoring urethane to water
wet surfaces. Some degree of success was realized but time
did not permit an extensive development program. The tempera-
ture problem received only a cursory investigation and also
needs more work.
Wet Surfaces
Four approaches were taken in an attempt to alleviate the
poor adhesion to wet surfaces. The word poor is used because
there was always some attachment of the foam to the surface.
Adhesion was quite discontinuous, however, with those areas
of attachment exhibiting only a weak bond. The problem had
been determined to be interaction between the isocyanate and
the water to form COo. This generation of gas at the imme-
diate surface retarded adhesion. Reaction was not restricted
to the surface. Water tended to flow upward, probably as
a vapor, into the foaming mass. The further reaction in the
foam created gross voids which were the areas of no attach-
ment.
20
-------
Tests using wet surfaces were performed in the laboratory in
a rather expeditious manner. Patio block concrete was used
as a substrate. Water was added to the test area insuring
complete coverage. The test surface was then tilted to allow
excess water to drain free. This method insured a saturated
area without pools of liquid. A total of about 20 cc of
urethane mix was prepared including any additives under study.
For comparison, a test was also made in the immediate area
using the base formulation alone. All solutions were hand
mixed and poured onto the test area.
The first approach was a flash coat of isocyanate to consume
the water. A reaction occurred but the material formed had
no Integrity and blocked any possibility of adhesion of the
following polyurethane foam to the substrate test pad.
The second approach Involved the addition of a number of
materials Into the base formula to combat the water. These
chemicals are listed in Table 1. They represent three classes
TABLE 1 - ADDITIVES TO THE BASE FORMULATION
Npnionic Water
Wetting Agents Soluble Polymers Gelling Agents
Triton X-100 Acrysol GS 1607 Primafloc
Polyethyleneamine Rhoplex LC-40 PVA (72-60)
Amlsol Carbopol 941
Dow Latex 460 Acrysol ASE-60
PVA (72-60) Hycar 1571
Gantrez AN-139
of materials: (1) nonlonic wetting agents to improve spread
and penetration of the foam and increase strength of ad-
hesive bonding; (2) water soluble polymers to occlude the
water; and (3) gelling agents to perform essentially the
same task as (2). This total approach was taken on the
basis of empirical data emanating from the high expansion
materials phase, where such materials added to the water
side of the isocyanate-water system showed some improvement
in that material's adhesive quality. With normal urethane
none of the additives exhibited any improvement.
The third approach was the use of drying agents and sorbent
materials as a precoat on the wet surface before the appli-
cation of foam. These materials listed in Table 2 were
spread as a thin dry layer on the wet block. They were
21
-------
effective in tying up liquid water on the surface, but could
not effectively remove surface adsorbed moisture.
TABLE 2 - SELECTED CHEMICALS APPLIED AS A PRE-COAT
FOR SORPTION OR DRYING
Ethyl silicate
Organic ammonium silicate
Calcium chloride
Calcium sulfate
Drysorb
The fourth approach consisted of a pre-coat of the catalysts,
either TEA or DMEA. With both of these materials, excellent
adhesion was achieved. The reaction involved is probably
between the water and isocyanate and parallels that action
hoped for in the first approach. The excess catalyst present
due to the precoat restricts reaction to the immediate sur-
face with no gross voids formed nor prolonged C02 generation
to adversely affect the initial adhesion developed.
The success using a precoat of TEA was further verified by
field tests run on the concrete pad previously described.
The tests revealed an impediment in the bonding action. A
barrier formed in this manner was ineffective in holding
water for any significant period of time. It was concluded
that the excess catalyst at the surface resulted in a foam
containing an inordinate amount of water soluble products
such as amines. These products were slowly leached from the
foam by the solubilizing effects of the water. This erosion
of the foam structure by water eventually caused barrier
failure. Limited tests were made involving the addition of
materials which would contain or otherwise protect the sol-
uble fraction, but no progress was realized before the end
of the program. A change in the catalyst or the use of a
material which would render the products insoluble could
possibly alter the situation.
Cold Surfaces
The evaluation of the effect of temperature was primarily
limited to substrate temperatures. Urethane systems are
normally recommended for a 50°F or better storage and use
22
-------
temperature. Results with lower test temperature will be
discussed in the section on storage.
Polyurethane foam, both normal and low density, is capable
of containing low boiling liquids such as ammonia and chlo-
rine if the dike or seal is erected at surface temperatures
50°F or above. The erection of polyurethane barriers at
lower temperatures is known to suffer since the substrate
temperature robs the reaction exotherm.
A series of barrierswere erected on the concrete pad with
the curb drain during the winter months. As the substrate
temperature decreased, the expansion of the initial layer of
urethane foam decreased, the expansion of the initial layer
of urethane foam decreased, as did the strength of the in-
terfacial bond. The bond was always strong enough to block
liquid flow, but it tended to leak. Temperatures as low as
15°F were encountered during the sequence. There was no
correlation between the temperature and the rate of leakage.
In an effort to improve bonding to cold surfaces, catalyst
changes were made to accelerate the reaction and concentrate
the heat output. Of the materials tested, an addition of
0.10% lead napthanate had the best effect. It decreased
the cream time to some 3 seconds at temperatures as low as
15°F.
The accelerated reaction improved expansion and bonding at
low substrate temperatures. It was detrimental, however,
with respect to the ability to penetrate into vegetated
surfaces. Thus, at the present state of development,a
choice must be made between the ability to achieve good
bonding to a cold substrate or a vegetated one.
Storage Tests
At the termination of that portion concerned with poly-
urethane, three short term storage cycles were initiated.
Two sto'rage tests were at room temperature (75°F) and one at
40°F. As expected, the lower temperature storage adversely
effected the expansion of the foam as well as increasing
the cream time. It was decided that lower temperatures would
only exaggerate the situation. Since time was not available
to do more than had been accomplished with the lead naptha-
nate, work was terminated.
The initial tests at ambient were conducted over a six week
period. No change was noted in the materials or the foam
generated from them until the sixth week. At that time
separation had occurred in the resin mix with two liquids
23
-------
formed. They could be reblended by agitation but separated
again within 24 hours. Since the only major change that had
been made in formulation was the methylene chloride, that was
assumed to be the contributing factor.
A second test package was made in which the methylene chlo-
ride was added to the activator instead of the resin. Eight
weeks of storage with this arrangement showed no problems.
With methylene chloride in the activator changes in the de-
livery ratio of materials was necessary from 1:1 to 1.3:1
activator to resin. At this ratio good quality foam of an
average density of 0.6 pcf could be obtained at an average
rate of 5 cfm.
As the final step in the "polyurethane foam" phase of the
project, two units using the methylene chloride formulation
and the foregoing mixing arrangement were shipped to the
Technical Project Officer. One was the refillable labora-
tory unit and the other a non-rechargeable portable unit
typical of that to be expected as a commercial unit. The
refillable laboratory unit has a capacity of 2.1 gallons
and delivers 25 cu ft of foam. The typical delivery rate
versus time curve is shown in Figure 6. The prototype of
the commercial unit has a total material volume of 6.4
gallons and delivers approximately 65 cu ft of foam at 5
cfm. The initial nitrogen pressurization is 220 psi with
some 10-15% ullage. The total weight of the system including
the carton is 65 Ibs.
The prototype urethane package can be activated by opening
the two flow-control valves. Triggering of the main valve
at the nozzle disperses the foam which initially will appear
as a free flowing liquid.
Polyurethane foam will serve a number of functions such as
construction of barriers, diverters, plugging of sewer drains
covering storm drains and damming of drainage ditches, etc.
The single factor which will exert the most influence in de-
termining the usefulness of foam will be the type of surface.
On dry, firm surfaces such as concrete and asphalt and at
temperatures above 30-35°F no difficulties should be ex-
perienced. On this type of substrate tests have shown that
4 foot high barriers can easily be constructed in which at
least a 3 foot head of water can be contained. Surfaces
which are cold and/or wet (water, solvents, etc) will impair
the quality of the initial .deposit of foam. This layer of
low quality foam can serve as a buffer zone upon which addi-
tional foam can be successfully applied. Unfortunately,
cold, wet surfaces impair the bonding or adhesion and the
24
-------
5.0
4.5
= 4.0
ro
en
a
s.
3.5
3.0-
2.5
(5 ft'/mln limit assuming 0.6 pcf foam)\P
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180
1 m1n 2 m1n 3 m1n
Time (seconds)
FIGURE 6 - TYPICAL TIME VS DELIVERY FOR 0.6 PCF POLYURETHANE FOAM SYSTEM
-------
low quality foam which results is prone to leakage. Grassy
and weedy surfaces also prevent the penetration of the foam
and thus the formation of a good seal. Leak-plugging and
strengthening can be achieved to a certain degree by selected
addition of more foam, but in general the performance of
barriers built on this type of surface is unpredictable and
most frequently merely limit the flow of liquid.
The volume of foam available in a single unit may not be
capable of impounding material volumes carried by current
transportation vehicles. Sixty-five cu ft of foam can form
a barrier 1 ft high by 1 ft wide by 20 ft in diameter which
would impound some 2000 gallons. Reducing the height to
6 inches would increase this to 4700 gallons. The 12 inch
base is necessary for good strength on the substrate. The
primary function appears to lie beyond full impoundment.
There is sufficient material in 65 cu ft to be successfully
used as a flow diverting barrier to direct the spilling
chemical away from a water system and/or into a naturally
occurring or otherwise available collection site. It can be
used to supplement these areas into more effective impound-
ment. Where leaks are slow it can provide impoundment until
more effective equipment can be brought to the site.
In many cases spilled material ends up following existing
courses to water systems. The foam is able to effectively
seal sewer openings and storm drains to prevent the chemical
from entering the water system in that way. On major high-
ways and railroad right-of-ways drainage paths including
ditches, culverts and conduit are used to divert and drain
water. The specified volume of foam is capable of sealing
these paths to effectively contain the spill. There are
examples of chemical spills which have entered urban sewer
systems or have reached rivers by flowing through railroad
drainage systems.
The chemicals which can be controlled by urethane foams need
to be clarified further. Data on specific chemicals is
limited. Water based liquids with the exception of strong
acids are contained as are nonpolar organics. Data is also
positive for chlorine and ammonia. Polar solvents are a
question. Acetone is controlled but methyl alcohol is not.
If the foam is applied and rigidized prior to contact with
the spilled chemical, it is doubtful that any catastrophic
reaction will occur. Reacting chemicals can break bonds
destroying the adhesion to the surface or they may even seep
through the foam mass as is the case with methyl alcohol.
Thus in most cases urethane will exercise some control if
applied correctly. Even in those 'cases where control is not
effected it is not expected that the spill situation would
be exaggerated. Guidelines in this general area are almost
completely lacking, however.
26
-------
SECTION V
LOW-EXPANSION INORGANIC FOAM
The objective of this phase of the program was "to evaluate
and develop current expanded inorganic materials such as
foamed concrete as barrier-forming materials* for containing
hazardous chemical spills".
Foamed concretes and foamed gypsums are common materials of
construction, with well developed formulation and appli-
cation technology. To use these foams for spill containment,
it was necessary to provide them with Instant setting cap-
abilities which allow barriers to be built up without forms.
To accomplish this, it was considered necessary to impart
a gelling action to the foamed material as it was delivered
to the barrier location. Thus, the basic mechanism to be
evolved was a gellant system to provide a uniform, semi-
rigid foamed mass to standard foamed concrete and foamed
gypsum mixes.
Assuming success in this regard, optimum gellant systems
would be defined along with the associated concrete or
gypsum formula. These compositions would be examined for
their ability to block and hold spills of hazardous chemicals,
along with the ability to form structures of appropriate
shapes and heights. The ultimate system would then be demon-
strated in the field.
Foamed concrete is one type of lightweight concrete. It is
an established construction material with a density range
usually given as 25 to 100 pcf. (More recently the American
Concrete Institute (ACI) has defined the density range of
50 pcf or less to be low density concrete.)
Lightweight concrete can be either precast, or cast-in-place.
Its largest use is in industrial and commercial buildings
where field placed, low density concrete is commonly used
for fills, thermal insulation and roof decks.
Because of the basic behavior characteristics of these con-
cretes under impact and shock loading, they have military
applications as shock dissipators in protective construction
schemes.4 They are also being considered in a number of new
non-military applications.
Foamed concrete is produced by introducing controlled quanti-
ties of air, water and foaming agent under pressure into a
foam nozzle and blending the resultant foam with cement
slurry or cement-aggregate slurry in a variety of mixing de-
vices, either batch or continuous. The foam must have suf-
ficient stability to maintain a structure until the cement
27
-------
sets to form a matrix of low density concrete. Foam gene-
ration and proportioning into the cement slurry are regulated
to achieve control of final strength and density. Up to 80%
of the final mix volume may be air, added as pre-formed foam.
The resultant foamed product can be pumped distances up to
1000 feet, to a height of 300 feet. The pumping parameters
are simply a function of the slurry viscosity, pump speci-
fications and slurry set time.
Existing foamed concretes have similar fluidity and set times
as normal concretes and thus require forms. There are, how-
ever, a number of fast set gypsum and concrete formulations
which have never been formulated in the cellular, or foamed
mode. REG-SET, a commercial fast set concrete, for example,
can reportedly be made to take its set within 5 minutes
after mixing with water. "Hardstem" and "Hardstop", both
gypsum formulations, were developed as rapid set formulations
by the British for pour-in-place mine bulkheads.5 More re-v
cently, Halliburtono developed both fast set gypsum and con-
crete-based formulations in their investigation of materials
suitable for sealing abandoned mines to prevent acidic mine
drainage. Various additives and sodium silicate were em-
ployed to achieve extremely rapid sets. This latter work,
although employing normal density materials, provided a
suitable basis for our study of foamed materials. Our pro-
gram consisted of:
1. Preliminary formulation evaluation
2. Small-scale barrier pours
3. Field testing
Each of these areas are discussed in detail in the following
sections. .,
The objective of this laboratory study was to develop leads
for possible formulations or techniques that could be de-
veloped into an instant-set pour system with slurry mixing
equipment. Therefore, commercially available slurry mixing,
preformed foam and pumping equipment were employed without
modification in order to expend more effort upon the formu-
lation development and the feasibility of the technique as
a spill control measure.
Preliminary Evaluations
The evaluation studies were conducted on a small scale, on
the order of 1/4 to 1/2 Ib of material per experiment.
Studies with gypsum formulations considered accelerators
28
-------
and gel-set formulations employing the water soluble polymer
polyvinyl alcohol (PVA) and a gelling agent.
Similar studies were also conducted on two fast-set cement
formulations developed by Halliburton Company. One of the
Halliburton formulations consisted of a mixture of cement,
bentonite, fly ash and gypsum. It employed bentonite to
obtain high viscosity and gypsum to obtain a fast set. A
second, more rapid-set formulation, employed cement and
bentonite for structural strength and sodium silicate to
impart the rapid set to the mixture. -A specialty gypsum
(Hydroperm) and a regulated-set cement (Huron Cement
Company)7 were also investigated.
Table 3 summarizes our studies to accelerate the set of
gypsum using an accepted accelerating agent (Na2S04) at
several concentration levels. Set times were reasonably
rapid when compared to normal plaster formulations. How-
ever, when using larger amounts of sodium sulfate, the time
to set was not nearly fast enough for this intended appli-
cation.
TABLE 3 - SUMMARY OF THE EVALUATION OF THE SET OF
GYPSUM FORMULATIONS* WITH ACCELERATORS
Quantity of Na?SOy| Remarks
none
all blended at once - 7 min.; re-
sistance to flow, 10 min.; jello-
like; 21 min.; set.
0.6 g gypsum slurried with 62 g H20 and
blended with foam from remaining
ingredients - 6 min.; heat but not
set; 17 min.; no structural in-
tegrity, plaster set in little balls.
0.8 g all blended at once - stirred 1 min.;
2 min.; no flow; 4 min.; lost sheen;
7 min.; hard set.
1.6 g same as above - no flow in 2 min.;
set in 4 min.
*Standard mix - gypsum, 100 g; water, 77 ml; protein
concentrate, 0.6 g
29
-------
TABLE 4 - SUMMARY OF THE EVALUATION OF PVA-GELLED GYPSUM SYSTEMS
CO
o
Gypsum Ho°
(9) (ccT
100
100
100
100
200
100
72
72
PVA
Foam or Foam
Concentrate
Gellant
0.6 g protein sat'd-
cone. soln.(4 cc)
100
100
57
72
3
2
2 cc MSA
cone.
2 cc pro
cone. soln. (2 cc)
2 cc protein
cone.
Congo Red
0.3 g
74 (combined into preformed Congo Red
foam) 0.8 g
(combined into preformed
foam)
Congo Red
1.0 g
(88 g of preformed foam with sat'd-l
PVA) soln. (4^cc)
(80 g of preformed foam with
PVA) soln. (2"cc)
Remarks
stirred 1 min. - no foaming action; added
Na2B407 soln., stirred 1 min. giving rubbery
mixture; 8 min. lost sheen; 16 min., hard set.
no foaming action on vigorous stirring.
stirred 1 min. - poor foaming, added ^38407;
7 min., no flow; 15 min., friable but set;
1 hr., tough, hard structure.
stirred 1 min. with 37 cc H^O - added congo
red in 35 cc f^O; approx. 50% increase in vol.
(20% of expected use); 2 min., still flows;
11 min., set.
slurried ingredients would not blend well with
preformed foam.
formed foam gypsum of reasonably light density-
14 min., no flow
stirred 1 min. after adding ^28407 to give
viscous mixture; 3 min., set; 12 min., hard.
same as above, but PVA ael appeared to
separate - no structural integrity.
-------
The gel time for polyvinyl alcohol-containing gypsum formu-
lations was also slower than desired (Table 4). Polyvinyl
alcohol solutions can be gelled to a rigid structure with
basic materials such as borax, and various dyes such as
Congo red. Although we obtained some highly viscous mix-
tures by adding small amounts of saturated solutions of
Na2B407 to slurries containing polyvinyl alcohol, at no
time did we feel that the time to set was short enough to
allow us to build barriers of practical heights.
Our evaluation of the cement-gypsum formulation developed
by Halliburton Company was initially encouraging, but failed
to develop into a practical fast-set system. The formu-
lation, as developed, is a two-slurry system; one containing
cement, fly ash and water; and the other, gypsum and benton-
ite. In Halliburton's application, the bentonite-containing
slurry is premixed to allow the bentonite to hydrate before
mixing with the cement-fly ash slurry. Our initial experi-
ment conducted in this manner was encouraging (Table 5); the
formulation resisted pouring after only one minute and took
a hard set after five. For the intended emergency spill
control application, however, the two-slurry system was
complicated and time consuming. Thus, our efforts were
TABLE 5 - SUMMARY OF THE EVALUATION OF
CEMENT-GYPSUM* SYSTEMS
Initial Set-Time
Type of System (non-pouring) Set-Time
(1) two slurry 1 min 5 min
(2) single slurry 4 min 9 min
(3) single slurry with
REG-SET cement 5 min 15 min
(4) single slurry with
added ^804 (2.5 g) 2 min 5 min
(5) single slurry with
added K2S04 (5>0 9) 4 min 5 min
plus protein foam
*Standard mix - cement, 20 g; gypsum, 40 g; bentonite, 4 g;
fly ash, 40 g; water, 62 ml.
31
-------
directed toward using this formulation compounded as a single
slurry system, without prehydration of the bentonite. As
noted in the second experiment, the time to a non-pouring
formulation increased from one minute to four and the hard
set time from five minutes to nine.
No particular improvement was noted in Run (3) when the fast
setting REG-SET/ cement was substituted for the Type I Port-
land cement used in the first two experiments. This was not
too surprising since REG-SET cement, although much faster
setting than the Type I cements, will not normally set before
five minutes. Adding potassium sulfate as an accelerator
to the REG-SET cement did not appear to materially aid the
initial set; however, once initiated, a hard set rapidly
developed.
The last experiment (5) was somewhat encouraging in relation
to our ultimate goal. Protein-based foam was added to the
slurry to achieve a final wet set density of 35 pcf without
significantly altering the set time of the mix. Unfortunately,
none of the above formulations set fast enough for our appli-
cation, but this experiment gave hope that if a fast set mix
could be developed, the addition of preformed foam to achieve
the desired low-density barrier material would not affect the
set time.
Our most encouraging result came from a modification of an-
other of Halliburton's specialty formulations developed for
sealing mine entries (Table 6). This system was also a two-
slurry formulation containing water and cement in one slurry,
and a water, bentonite, sodium silicate mixture in the other.
As mentioned previously, we felt that the two-slurry system
was not practical for our application, but that the addition
of a solution of sodium silicate to the foamed cement formu-
lation as it exited the nozzle was entirely feasible. We
anticipated forming a lightweight slurry of cement, bentonite,
water and preformed foam, and adding the silicate solution
by means of a water ring at the nozzle to achieve a fast
set.
Our initial laboratory scale studies (runs 1-4) were dis-
couraging until we discovered that gel formation occurred
within seconds after adding the sodium silicate (run 5), and
that any additional agitation only served to break up the
exiting gel. Once formed, the gelled material resisted
flow and appeared to have reasonable strength. In addition,
preformed foam made from MSA salt water foam concentrate was
compatible with the cement-bentonite mixture and encouraging
samples of foamed concrete of 37 and 45 pcf were obtained
(runs 6-7).
32
-------
TABLE 6 - SUMMARY OF THE EVALUATION OF CEMENT-SILICATE SET SYSTEMS*
CO
Run No.
1
2*«*
3
Preformed Foam
Added (cc)
8**
600
400
400
400
500
Gel Time
(seconds)
5-10
5-10
Cement
Density (pcf)
2-3
2-3
2-3
37
45
Remarks
Components slurried, silicate
added while stirring.
Same as above
MSA salt water foam concen-
trate (2 cc) slurMed with
components; expanded 2 times
original volume.
Preformed foam mixed with
slurry, silicate added while
stirring, weak gel structure.
Same as above; gelling
occurred before complete
addition of silicate, addi-
tional stirring broke up gel.
SlurMed to about 800
silicate added all at
while stirring.
SlurMed to about 650
silicate added all at
while stirring.
SlurMed to about 750
silicate added all at
while stirring.
cc,
once
cc,
once
cc,
once
* Standard mix - Portlant Type 1 cement, 192 g**; benton1te.il g;
water, 189 ml.
** 192 g of solids In Run 2 made up of cement (20 parts), fly ash (40 parts),
gypsum (40 parts).
*** Mix contained 231 g of Portland cement.
-------
Experiment 8 summarized in Table 6 was an attempt to deter-
mine if the cement-fly ash-gypsum formulation evaluated in
Table 5 would also gel with sodium silicate. This was
found to be the case.
Small Scale Barrier Pours
Small scale barrier pours using field-scale equipment were
conducted using a Mason Flow Mixer, a continuous slurry
generator developed by Hoge-Warren-Zimmerman of Cincinnati,
Ohio and a Model 2000 Slurry Rock Dust Distributor, a batch
slurry mixer manufactured by MSA for use in coal mines. The
formulation development work and pouring techniques were de-
veloped with the Mason Flow Mixer.
The Mason Flow Mixer is shown in Figures 7 and 8. The unit
consists essentially of a hopper for powdered solids, a
screw feed for solids metering, a slurry chamber and a slurry
pump. For producing foamed products, preformed foam is
added either in the slurry pump or immediately downstream of
the slurry in the delivery hose.
The unit is capable of producing slurry in the range of 5-10
gal/min. It requires 220 V power and a water supply, but
can also be powered by a gasoline engine.. This particular
unit is adapted for laboratory experimentation. Controls
are available to vary the water-to-dry powder ratio, the rata
of slurry production, and the speed of the slurry pump. On-
off controls for the water and powder feed as well as auto-
matic wash out capabilities are available on a remote control,
hand-carried push-button device.
In operation, water is started through the mixing chamber
and slurry pump, followed by the activation of the dry powder
feed. Approximately 20 sees is required to blend the dry
material into a slurry and deliver it through the slurry pump
into the delivery hose.
Preformed foam was produced from commercial foam units using
both protein and detergent-based concentrates. The essential
elements of a foam unit are a tank for foam solution, a pro-
portioner, and a refiner column. A 4 to 6% concentration of
the foam concentrate in water is fed under pressure to the
proportioner where it is mixed with air and fed to the foam
refiner column. The proportioner may consist of either
metering valves for both air and foam solution, or an ori-
fice plate for the solution with a valve to regulate the air.
The refiner consists usually of a 1 to 2 inch diameter copper
column 1-1/2 to 2 feet in length, full of a suitable packing
34
-------
FIGURE 7 - MASON FLOW MIXER
FIGURE 8 - FLOW MIXER SHOWING MOYNO PUMP AND DELIVERY SYSTEM
35
-------
material (e.g., stainless steel wool, Raschig rings, Burrell
saddles, etc.).
The proportion of air to dilute concentrate solution deter-
mines the foam density. The total pressure on the system
(usually 45-90 psig) determines the rate of foam production.
Protein and detergent-based foam were each used initially in
the studies. The protein foams were produced by Mearl
Corporation -- Mearlcel for gypsum-based and Mearlcrete for
cement-based mixtures. MSA's detergent-based foam concen-
trate was later found to be at least equivalent to the protein-
based concentrates and compatible with both gypsum and concrete
based mixtures. Therefore, it was used almost exclusively in
our test pourings and field trial studies.
A summary of our preliminary development runs is catalogued
in Table 7. Some of this early work was conducted simul-
taneously with the laboratory studies and similar disappoint-
ing results were noted. They served, however, to confirm
our laboratory results and to gain experience with the con-
tinuous mixer.
The initial unsupported barriers were produced with gypsum
formulations (Runs 1-13). Set times of the order of minutes
were achieved by adding accelerators, but the set was not
fast enough to produce a non-flowing slurry.
Development work with commercial quick-set cement formulations
was carried out with Huron Cement Company's Regulated Set
Portland Cement (RSPC) (Runs 14-17). This is a special Port-
land cement supplied with an accelerator that produces an
initial set in 5 to 15 minutes, depending primarily on water
content and temperature. With added foam, the set time in-
creased considerably and the cement was of poor strength and
quality. The retarded set was likely due to the protein-
based foam since some organics (e.g., proteins, sugars,
starches, etc.) are known to retard the set of Portland
cement.
The preliminary development studies on Halliburton-based
formulations are summarized in Runs 18-32. The gypsum based
formulations (Runs 18-21) set too slowly for our application.
Following the lead of our laboratory studies, however, form-
ulations containing sodium silicate as a fast gelling agent
showed promise, and were slowly developed into barrier forming
materials.
36
-------
TABLE 7 - MASON FLOW MIXER TEST DATA
GO
-J
Run No.
,1
2
3
4
Components
100 parts gypsum (molding plaster),
62 parts water
100 parts gypsum, 62 parts water,
13.5 parts Mearlcel foam 2 pcf
71.5 parts gypsum, 62 parts water,
13.5 parts Mearlcel foam 2 pcf
Slurry No. 1, - 100 parts gypsum, 62
parts water. Slurry No. 2, - 13.5
parts water foam containing 2.5% PVA,
Borax added via Shot-Crete nozzle
100 parts gypsum, 62 parts water,
13.5 pacts foam. Na2S04 added with
foam
100 parts gypsum, 62 parts water,
13.5 parts foam, 0.1% Na2S04 added
with foam
Same as 6 except start gypsum before
foam
100 parts gypsum, 62 parts water,
0.1% NaoSO. with 16 parts foam
added Snot-Crete nozzle
Same as 8, Increased foam and
Na2S04 content 50%
Set Time
5 m1n
7 m1n
16 m1n
Remarks
5 mi n
4 mln
Low angle of repose, density - 112 pcf.
Low angle of repose, poor strength, density
32 pcf.
Density - 32 pcf.
PVA plugged the foam generator filter. Borax
does not mix well through Shot-Crete nozzle.
Very poor foamed plaster - wet,soupy.
No foamed structure produced.
No foaming action, plaster plugged In refiner
and control valve.
Low angle of repose, no flow after 2 minutes,
density - 48 pcf.
Too much foam, poor mixing, poor angle of repose,
-------
TABLE 7 (continued)
CO
00
Run No. Components Set Time
10 100 parts gypsum, 62 parts water, 3 mln
4.5 parts PVA, .08 parts Borax and
16 parts foam added to nozzle
11 100 parts gypsum, 62 parts water, 1.5 mln
4.9 parts PVA, .08 parts Borax
and 16 parts foam added at nozzle
12 110 parts gypsum, 62 parts water, 5 m1n
4.9 parts PVA, .08 parts Borax
with 16 parts foam added at nozzle
13 110 parts gypsum, 62 parts water, 10 mln
4.9 parts PVA, .08 parts Congo Red
with 16 parts foam added at nozzle
14 Huron Reg-Set Cement, 2 turns on no set
powder feed, foamed with protein-
based preformed foam
15 100 parts Reg-Set Cement, 62 parts 20 min
water, 13.5 parts foam with 4X
silicate added at .nozzle
16 100 parts Reg-Set Cement, 62 parts 25 mln
water, 5.5 parts dry bentonlte,
13.5 parts foam 1n 20% Na2S103
protein added through Shot-Crete
nozzle
17 Same as 16 except foam and silicate 45 min
added just after Moyno pump
18 20 parts Portland cement, 80 parts no set
flyash, 42.5 parts water, MSA salt-
water foam (1.9 pcf)
Remarks
Poor strength to foam structure, poor angle of
repose, density - 35 pcf.
Low angle of repose.
Low angle of repose.
Low angle of repose.
- 33.6 pcf, low angle
No set 1n 90 mln,.density
of repose.
Very friable structure.
Low angle of repose, poor strength, foamed mass,
42 pcf.
Homogeneous foamed concrete of poor strength.
Poor foam-concrete, fly ash settled to bottom
in 15 mln.
-------
TABLE 7 (continued)
Run No.
Components
Set Time
Remarks
eo
vo
19 20 parts Portland cement, 40 parts
fly ash, 40 parts gypsum, 42.5 parts
water. MSA salt-water foam (2.0 pcf)
5 gal/m1n
20 20, 40, 40 parts respectively of
Portland cement, fly ash, gypsum.
2.43 pcf foam blended In using In-
line spiral mixer
21 Slurry No. 1, 20, 40, 40 parts re-
spectively of Portland cement, fly
ash and gypsum with 30 parts water.
Slurry No. 2, 5.4, 30 and 64.5 parts
bentonlte, silicate and water
22 Slurry No. 1, 65.7 parts cement, 34.3
parts water. Slurry No. 2, 5.4, 30
and 64.5 parts bentonlte, silicate
and water. Ratio of Slurries 1:2
foam was 1:0.83:1
23 Same as 22 Increased foam delivery
ratio
24 Same as 22 and 23
25 Bentonlte left out of formulation.
Cement slurry of 66 parts cement, 34
parts water and foam (2.7 pcf).
Silicate solution (502 grade 40* 1n
water) added with water ring at
shot-crete nozzle. Feed rates
87 Ibs/mln slurry, 10 Ibs/m1n sili-
cate solution
4 m1n
no set
3-4 sec
3-4 sec
gel
<15 sec
gel
Very poor foam-slurry mixing. No strength
after set.
Much more homogeneous foam cement, rather slow
set and lacks strength, friable, density -
42.4 pcf.
Gelled, but too liquid. Low angle of repose.
No set observed 1n 24 hrs, density - 45 pcf.
Low angle of repose. Gelled rapidly but lacked
foam structure, density - 90 pcf.
Set 1n 9 m1n, low to medium angle of repose.
Density of foamed cement varies from 93 Ibs to
34.5 pcf.
Poor foam blending. Set to a friable structure,
will not support weight of a man. Density -
35 pcf.
Very soupy formulation, low angle of repose
Density - 46.6 pcf.
-------
TABLE 7 (continued)
Run No.
Components
Set Time
Remarks
26
27
28
29
30
31
32
Slurry of 67 parts cement, 33 parts
water at 87 Ibs/mln, foam (3.3 pcf)
and SOX silicate solution at 10 Ibs/m1n
Same as 26
Same as 26 and 27 except silicate rate
doubled. Slurry of 67 parts cement,
33 parts water at 87 Ibs/min. Foam
(1.5 pcf) at 15 gal/min and BOS sili-
cate added at nozzle at 23 Ibs/min
Same as 28 except more low density
foam used. MSA Super Foam (3.2 pcf)
at 16.5 gal/m1n
Same as 29 except that MSA Super Foam
Introduced into Moyno pump along with
cement slurry
Same as 28 except MSA super Foam (20
gal/min) introduced 1n slurry line
Immediately downstream of Moyno pump
Same as 31 in all aspects
3-4 sec
gel
3-4 sec
gel
3-4 sec
1-2 sec
1-2 sec
gel
1-2 sec
1-2 sec
Excellent angle of repose although Inadequate
foam content. Density - 68 pcf.
Produced dike 8 in. high, 39 in. wide and 7 ft
long. Density - 42.5, good foamed cement structure,
medium angle of repose.
Poor foamed cement consistency, rather soupy.
Gel supported a 12 in. high dike, 2' in. wide.
Friable cement, density - 71.6 pcf.
Gel supported a 12 in. high dike. Excess displace-
ment of first poured material occurs. Very strong
cement.
5 gal. MSA Super Foam in 17 and 30 seconds causes
Moyno funnel to overflow. Output of Moyno pump
too low for this method of foam Introduction.
Highest degree of foam character in cement formu-
lated to date. Gel supported a 12 in. high dike.
Density of cement at beginning and end of pour
33 pcf. Very strong dike.
Gel supported 15.5 In. high dike, 5 ft wide and
5 ft long. Base of dike slid as in 31. Foam
cement density - 32 pcf.
*Note: Silicate used before dilution contains 37.IS Na20-3.34 S10,
-------
Because of the extremely fast gelling action of the sodium
silicate, it was necessary to modify the generating and de-
livery system. The cement-bentonite slurries were formed in
the Mason Flow Mixer and combined with preformed low expansion
detergent foam downstream of the slurry pump. The silicate
was subsequently blended into the foam slurry by the use of
a shotcrete nozzle as it exited the delivery hose. A water
ring, consisting of two circular rows of holes in the shot-
Crete nozzle, was used to inject the silicate solution.
In the initial formulations the development work stressed
variations in the silicate concentration, bentonite addition,
type and quantity of detergent foam formulation and tech-
nique. Quick gelling mixtures were obtained having densi-
ties of 32 to 93 pcf. In these runs two problems persisted:
one was the tendency of the unsupported barrier base to be
displaced by subsequent layers; and the second was the dif-
ficulty in obtaining a uniform foam delivery rate. Both of
these problems are related; the first is a function of foam
density and silicate concentration while the second is caused
by variations in downstream slurry pressures. Air pressure
was used to produce the foam, and to introduce it into the
cement slurry mix. The rate of foam production and its
addition is a function of the pressure drop across the
metering system. Thus,as the viscosity of the cement slurry
varied, the back pressure in the feeder line varied which
affected the foam addition rate. This condition produced
concrete of varying densities and support strength. In the
first three runs (22-24), bentonite added to the slurry mix
to increase its viscosity produced a high back pressure,
resulting in foamed concrete in excess of 90 pcf. In Runs
25-32, therefore, the bentonite was not used. Although some
improvement was noted, considerable density variations still
persisted indicative of a foam feed-rate problem.
Although the presence of problems was recognized, it was
apparent that the sodium silicate-gelled foamed cement offered
promise for unsupported barrier formations. Accordingly, the
remaining research and development effort centered on this
formulation, and detailed data was gathered in an effort to
delineate the conditions for best barrier production. This
data and subsequent data for field test runs are summarized
in Table 8.
The basic cement formulation for these studies contained 5%
by weight bentonite. The slurry feed without bentonite was
too thin to effect a reasonable barrier build-up. A concen-
tration of 5% bentonite, however, materially aided in in-
creasing the viscosity of the slurry without causing undue
pumping problems. Starting in Run 59, additional lime was
also added. Free lime is present in Portland cement. With
41
-------
TABLE 8 - SUMMARY OF DEVELOPMENT STUDIES
Dun No.
r Mixer Feed/Mn.
^ement(c) water(w) w/e
Obs) (Ibs)
Preformed roam/mln.
Kate Density water
(gallons) (pef) (Ibs)
St 1 1cate/m1n.
Neat Sol'n Total Mater
(Ibs) - (Ibs)
Total
Hater
Total
w/c
Cement
Silicate
C/S
Run
Time
(m1n)
Dike
Hts.
(in)
Concrete
Density
(pef)
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
4G
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
M
65
tt
68
69
70
71
72
73
74
7!
76
77 *
78
79
58.2
54.4
55.2
55.2
58.2
58.2
58.2
58.2
52.5-43.2
44.4
44.4
58.0
52.5
52.5
52.5
52.5
52.5
52.5
56.5
56.5
56.5
46.0
47.0
47.0
47.0
47.0
47. U
47.0
47.0
52.0
?2.0
52. n
52.0
M
52.0
52.0
52.0
52.0
52.0
52.0
174.0
131.0
121.0
91..0
68.0
77.0
29.1
29.1
29.1
29.1
29.1
29.1
29.1
29.1
29.0
28.1
28.1
28.5
28.5
28.2
28.2
28.2
28.2
28.2
28.
28.
28.
28.
28.
28.
28.
28.
20.4
28.4
28.4
2C.2
28.2
2R.2
28.2
\l\\
28.2
28.2
28.2
28.2
28.2
28.2
123.5
71.5
85.5
56.9
42.3
38.0
0.50
0.53
0.53
0.53
0.50
0.50
0.50
0.50
O.CO
0.63
0.49
0.49
0.54
0.54
0.54
0.54
0.54
0.50
0.50
0.50
0.62
0.60
0.60
O.CO
O.£0
0.60
0.60
0.60
0.60
0.54
0.54
0.54
8:13
0.54
0.54
0.54
0.54
0.54
0.54
0.71
0.71
0..71
0^625
0.625
0.50
IS. 2
-15.0
15.6
13.6
15.6
IS. 6
13.9
13.9
15.6
11.9
14.3
16.7
16.7
18.2
19.2
16.1
16.1
11.9
15.0
15.6
IS. 6
21.2
20.8
14.3
28.6
20.0
20.)
10.)
14.3
1.3.6
12:8
12.0
18.8
13.0
19.6
13.0
14.6
9.8
19.4
43.8
26.0
38.1
22.0
3.2
2.8
2.4
2.2
2.5
2.1
2.1
2.0
4.1
2.1
3.5
2.0
2.4
2.1
2.2
2.7
2.8
2.7
1.8
2.7
3.0
3.7
3.1
3.6
4.2
2.5
2.0
8.1
4.8
2.2
2.8
2.4
2.1
Lost
5.2
4.8
4.7
4.1
4.1
'.5
4.5
4.5
2.5
4.0
6.1
6.1
3.9
4.9
3.7
5.8
3.5
5.6
10.0
5.0
6.2
6.2
9.1
7.8
6.8
7.0
6.0
4.6
17.4
11.8
11.5
13.7
4.65
7.34
8.02
12.8
3.01
5.2
4.2
3.4
.
9.0
12.6
8.2
8.0
5.4
11.:
25.3
15.1
12.2
11.9
11.6
11.7
15.3
15.6
15.1
16.1
15.2
12.5
11.7
11.7
15.1
12.1
4.7
4.6
12.3
8.8
11.3
8.3
16.6
16.2
17.1
15.6
16.9
10.9
10. 9
16.7
12.5
23.4
16.4
15.6
Lost
16.5
11.7
14.4
14.4
23.4
18.7
40.8
33.8
29.5
19.0
16.2
16.2
21.3
21.3
21.5
20.8
22.2
21.2
20.3
16.3
16.3
21.0
8.0
3.1
6.3
17.2
12.2
15.7
11.6
23.2
22.6
23.9
21.8
23.6
15.2
15.2
23.2
17.4
32.6
22.4
19.4
24.4
21.7
-
23.0
16.3
20.0
20.0
32.5
26.1
57.0
47.0
41.1
26.5
51.4
51.4
54.2
55.5
53.6
57.1
53.8
54.9
54.4
49.4
55.4
42.4
40.4
42.3
51.2
47.0
50.1
44.6
69.0
62.8
63.8
63.9
56.65
50.9
- 51.62
64.25
48.7
66.0
54.8
53.1
58.1
53.3
-
60.2
50.5
56.4
56.2
160.0
109.0
167.8
119.0
95.6
64.0
0.88
0.88
0.92
0.95
0.92
0.98
0.92
varied
1.22
1.11
1.05
0.81
0.77
0.81
0.98
0.83
0.88
0.79
1.47
1.33
1.35
1.35
1.20
I'.OB
.10
.23
0.94
.27
.05
.02
.12
1.02
•
1.16
.97
1.082
1.08
.92
1.33
1.39
1.31
1.41
0.83
5.02
4.65
3.61
4.21
3.73
3.85
3.61
3.83
3.79
3.79
3.48
4.34
11.22
11.54
4.27
6.42
5.00
6.81
2.83
2.90
2.75
3.01
'2.78
4.31
4.31
3.11
4.16
2.22
3.17
3.74
2.97
3.33
.
3.15
4.44
3.61
3.61
7.4
C.4
2.97
2.7 -
2.3
4.0
6.7
5.5
2.8
6.1
6.0
6.2
5.8
7.5
9.5
6.5
8.0
4.5
9.1
6.0
5.0
6.0
9.0
9.0
7.5
7.3
3.0
5.0
5.75
9.2
7.8
5.0
3.75
1.0
4.2
7.6
5.0
5.5
5.7
5.0
7.5
4.0
4.6
8.5
9.75
2.5
4.9
4.0
4.5
9.5
11.6
24
16
16
17
20.5
20
19.5
26
6
12
17
16
21.5
15
6
6
19
15
16
17
4
16
18
25. 5
19
16
8
.
12
16
j
18
18
(1)
13
8
3
13
27
13
13
14
14
14
14
35.1
34.8
35.0
37.0
34.0
34.0
32.0
34.0
34.0
33.0
. 39.0
35.0
37.0
28.3
29.0
36.0
33.0
38.0
36.0
38.0
41-52
36.0
36.0
40. C
38. C
41.0
22.6
32.0
31.4
47.0
51.0
39.2
38.8
35.0
34.6
27.5
34.6-
36.0
47-
48.5
43.3
31.5
55.5
31.1
35.0
42.0
32.0
29.0
37.0
-------
our high sodium silicate addition, and resultant calcium
silicate formation, however, we felt that the final cement
blend would be deficient in lime. This additive proved to
be beneficial and thus the cement mixture from Run 59 on
consisted of 90 parts cement, 5 parts bentonite and 5 parts
1 i me.
Our detailed study evaluated the water to cement (W/C) ratios
of the slurry and final mix (adding in the water contributed
by the foam and silicate solution) and the cement to silicate
(C/S) ratio, as a function of dike height. Consistent dikes
of over 16 inches were poured in Runs 33-40 with concrete
densities of 32 to 37 pcf. A typical dike pour is shown in
Figure 9. Gel structure breakdown, with a resultant slurry
slide, limited the pour height to approximately 2 feet on a
12-16 inch wide base. Once the initial gel structure was
broken, slides continued to occur along the same break line.
Short asbestos fibers (5% by weight) were introduced into the
powder feed in Runs 41-43 in order to mitigate the slide prob-
lem. Back pressure surges, however, as determined from a
pressure gauge in the delivery line, resulted in foam feed
problems and a necessity for adding increased water. The
total effect was to lower the dike height.
At the recommendation of Hoge-Warren-Zimmerman, the manu-
facturer of the Mason Flow Mixer, the preformed foam was
added to the slurry in the Moyno pump chamber to overcome
the varying feed rate of the foam generating unit. In order
to carry the increased capacity of the slurry and the foam,
it was necessary to increase the pumping speed of the Moyno,
by a factor of four, to its maximum. Driving the Moyno
pump at such high speeds caused troubles not previously en-
countered. Pumping the water-cement slurry and the pre-
formed foam resulted in electrical overload and breaker
cutout. In addition, it was very difficult to balance the
flow of water-cement slurry and preformed foam delivery so
as not to overflow the Moyno pump. Runs 44-59 summarize
efforts to introduce the preformed foam and cement-water
slurry simultaneously through the Moyno pump. These runs
were very short lived; no sooner were flows established
than the Moyno pump would overload and spill foamed concrete
onto the floor.
Although homogeneous foamed concrete barriers of excellent
height and strength were occasionally produced, the elec-
trical overload problems made this procedure impractical.
The slurry mixer automatically shut down several times during
each run and was restarted only with difficulty to continue
the run.
43
-------
FIGURE 9 - FOAM CONCRETE TEST POUR SHOWING HEIGHT AND
ANGLE OF REPOSE POSSIBLE WITH SILICATE SYSTEM
44
-------
To avoid the Moyno feed and electrical overload problems, a
pump was installed on the preformed foam unit to meter and
pump the foam into the concrete downstream of the Moyno pump.
This enabled us to reduce the speed of the Moyno and to add
larger quantities of preformed foam consistently in order
to produce lower density foamed concrete. At the same time
a larger diameter delivery line was installed to alleviate
high feed line pressures. The change enabled production of
foam concrete as low as 23 pcf (Run 60). The resulting con-
crete structure lacked good integrity, however, which may
have been due in part to a poorly blended foam slurry mix.
Similar results were observed in Run 61. This method of
foam introduction greatly improved the operation in that it
eliminated the electrical overload problem.
The one inch (I.D.) spiral mixer-delivery hose system was
installed in the above system to provide additional blending
action of foam and slurry. This action produced reliable
barriers of significant height (Runs 62, 63, 65). Run 64
had a malfunction in the silicate feed system.
As mentioned previously, Run 59 presents the data for a formu-
lation consisting of 90 parts cement, 5 parts bentonite and
5 parts lime slurried and gelled with silicate to produce a
16 inch high, 9 inch wide and 6 foot long barrier with the
qualities of block wall. This was a most successfully pre-
pared barrier, emboding all the requirements of the program --
quick gel in 2-3 seconds, high angle of repose, no displace-
ment during pour and exceptional strength 1-1/2 hrs after
placement.
Field Tests
Field tests were conducted with the Mason Flow Mixer on the
ability of the quick-set foamed concrete formulation to block
and impound flowing liquids. The powder feed for the tests
consisted of:
Parts by Weight
Cement 90
Bentonite 5
Lime 5
The unit was placed on a flat bed truck for the tests and
connected to water, air and power supplies at the test site.
A mobile compressor provided the necessary air.
A problem with the unit, as far as its utility in an emergency
system, was found during these tests. In order for proper
45
-------
water mixing to occur, the unit must be nearly level. A
slight slant rearward allows the water to flow backwards out
of the mixing chamber without contacting the powder feed.
The powder, water and silicate feed rates were held constant
throughout the runs, and as closely as possible to conditions
for Runs 62-65, which gave reasonable dike heights. Some
problems were encountered during sub-freezing weather, causing
changes primarily in foam feeds. The run data are summarized
as Runs 66-73 in Table 8.
In the initial test (Run 66), a concrete barrier 16 to 18
inches high and 8 feet long was erected across a depression
at the same time as water flowed into the impounded area at
a rate of about 12 gal./min. The foamed concrete was ex-
cellent (39 pcf) and gave no evidence of disintegration
even though the gel set alone was providing the barrier
strength. Very small leaks appeared under the barrier
approximately one hour after water flow started but days
later had not enlarged. The pictorial progress of the test
is shown in Figure 10.
In Runs 67-68 water flow was started through a depression at
a rate of about 17 gal./min before the dike formation was
attempted. Run 67 was unsuccessful. The water flow immedi-
ately washed away the freshly poured foamed concrete. To
overcome this situation, the flowing water was allowed to leak
through a gap on a section left open in the foam concrete
barrier. The barrier was erected from both sides of the de-
pression to the gap and allowed 2 minutes set time. A piece
of sheet metal was then inserted as a sluice gate in the gap
to stop the flow. Immediately upon application of the sluice
gate, foamed concrete was poured behind it to strengthen and
tie together the two barrier sections. In approximately 4
minutes the reservoir filled to a depth of about 10 inches
and at that same time the barrier around the sluice gate
failed.
In Run 68 a barrier was successfully built across a water
flow of 15 gal./min in 7.5 minutes. The barrier averaged 10
to 13 inches in height, 9 inches thick and 9 feet wide. The
impoundment was full (10 inches deep) in 22.5 minutes.
The difference between Run 67 and 68 in the ability of the
barrier to stop flowing water is apparently a function of the
hydraulic head build up rate on the barrier. Run 68 filled
to a depth of 10 inches in 22.5 minutes, whereas Run 67,
which failed, had a smaller impoundment area, and filled to
a similar depth in only 4 minutes. The slower fill rate
obviously allowed additional strength to develop in the dike
material.
46
-------
A * . *
FIGURE 10 - FIELD TEST POUR SHOWING BUILD UP OF THE DIKE AND
VIEWS OF THE BARRIER WITH IMPOUNDED WATER
-
f
-------
At ambient temperatures below 32°F, further field test work
was adversely affected. At these temperatures, the foam
solution lines and various water lines on the slurry mixer
tended to freeze (Runs 69, 70). The incorporation of either
aliphatic or polyhydric alcohols in the foam solution allevi-
ated the freezing problem but unfortunately these anti-
freeze materials retarded the cement set and the foam concrete
tended to be easily dissociated by flowing water.
The use of inorganic solutes, CaCl2 and NaCl, in high con-
centration to lower the freezing point of the detergent foam
solution resulted in poor concrete quality (CaClg) or poor
foam quality (NaCl). A barrier to a 15 gal./min flow of,
water using MSA salt water foam and 5% CaCl2 solution pro-
duced a 13 inch high barrier of 47 to 48.5 pcf concrete
(Run 71). The concrete was of poor integrity, slow to gel
and tended to dissociate.
Run 72 was conducted on an above-freezing day without added
anti-freeze material in the foam, to test our feed rates
on the equipment. The barrier height of 27 inches was one
of the best dikes produced.
Run 73 was run in sub-freezing weather using sodium hydroxide
to lower the freezing point and Derifat 160-C, a foam agent
from General Mills, Inc. The foam cement tended to gel
rapidly but was lacking in strength.
The field tests using the Mason Flow Mixer established the
utility of foamed concrete as a barrier for hazardous material
spills and the feasibility of erecting barriers under field
conditions. The two problems encountered, however, orien-
tation sensitivity and cold weather dike formation, need
further investigation.
Following the field tests discussed, above, an MSA Model 2000
slurry rock dust distributor was evaluated as an example of
a batch operation foam concrete generator. The unit, shown
being used for making a typical barrier installation in
Figure 11, consists of a slurry tank of approximately 200
gallons capacity, and a Moyno pump for slurry delivery. Both
the Moyno and the slurry tank agitators are powered with
variable speed hydraulic motors, which are in turn powered
by an electrically driven hydraulic pump. A pump was employed
to introduce the detergent foam downstream of the Moyno pump.
The hose delivery system and silicate blending nozzle used
on the Mason Flow Mixer were also employed on this system.
48
-------
FIGURE 11 - POUR OF FOAM CONCRETE USING MSA BATCH TYPE MIXER
-------
The advantages of using a batch system, rather than the con.-
tinuous unit like the Mason Flow Mixer, are simplicity and
ease of field operation. A batch unit eliminates the need
for metering the powdered solids and water feed. This unit
also would operate in off-level positions.
The data for the runs with the slurry rock duster are summa-
rized as Runs 74-79 in Table 8. Improved weather conditions
during this test period permitted the use of the detergent
foam without added anti-freeze.
The ability of the foamed concrete to form barriers through-
out the six experiments was generally inferior to that finally
developed from the continuous mixer. Problems were experi-
enced initially with extremely rapid slurry feeds and bal-
ancing the preformed foam and silicate solution feeds to ,the
slurry. In order to obtain good powder mixing within the
slurry tank with the small agitators normally used to slurry
the rock dust, an abnormal amount of water had to be employed.
Whether due to this or to unbalanced preformed foam or sili-
cate feeds, barrier heights exceeding 14 inches were never
realized. Poor gel formation, resulting in slides, Inevita-
bly held batch processed dikes to the above height limit.
Only in the last run (Run 79), when the water to cement ratio
was held at 0.50 in the slurry tank with additional stirring
to allow for mixing, did we obtain foam concrete of quality
similar to that obtained from the Mason Flow Mixer. However,
we could not form more than a 14 inch barrier before running
short of foam concrete. Additional material was available in
the hopper, but at the lower water to cement ratio it was
difficult to achieve flow from the hopper into the Moyno
pump at a rate sufficient to stabilize the feed of preformed
foam and silicate.
Summary
The feasibility of foamed concrete to produce a dike which will
hold back or impound a typical spill of liquid has been demon-
strated. Tests on such varied substrates as clay, shale,
chipped limestone, grass end weed-covered ground have been
successful. Screening tests have shown that such hazardous
liquids ss methanol, 1,1,1.trichloroethane, phenol, acetone
cyanhydrin and acrylonitrile do not affect freshly poured
foam concrete.
The rate of build up of hydraulic head on the barrier system
is probably the single most important factor to successfully
impounding a spill. An extremely high liquid velocity would
probably not allow barrier formation in the flowing stream.
In such a spill, however, most of the contents would likely
50
-------
be lost and the damage done before emergency equipment could
be brought to bear. Where low flow velocities are present,
or a large Impoundment area available to allow filling over
a period of 15 minutes or more, a foam concrete dike could
be successfully employed.
Although our experience with the batch type mixers was limited,
we believe that the batch type operation offers the most simple
approach to an emergency field unit. A possible design for
such a unit is shown in Figure 12. The unit is trailer mounted,
suitable for a pick-up truck operation. It would produce
approximately 50 ft3 of foamed concrete at 45 pcf per batch
within 30 minutes of delivery to the site. Repeat cycles
would take approximately 25 minutes each.
Three such batches (150 ft^) could be produced in approxi-
mately 80 minutes to build barriers of such sample dimensions
as:
2 ft x 2 ft x 38 ft
1.5 ft x 2 ft x 50 ft
1.5 ft x 3 ft x 33 ft
Raw materials for one batch are:
10 bags of cement (940 Ibs)
30 gallons of sodium-silicate solution
1.5 quarts of foam concentrate
Material for one run would be transported on the unit. Ma-
terial for three batches would weigh approximately 4900 Ibs
and require about 1800 Ibs of water (215 gal.) at the site.
Additional material could be brought in if needed.
The sodium silicate and foam concentrate would be the only
material required to be stored in quantity for multiple
batches. Enough Type I cement could be stored for 1 to 3
batches, with additional cement picked up as needed.
51
-------
MVOCAUI-IC PUMP
Scale: 3/4" = TO"
FIGURE 12 - PROPOSED EMERGENCY FIELD UNIT
52
-------
SECTION VI
HIGH EXPANSION SYSTEMS
In the majority of spill situations the demand for control
materials will exceed the supply. An alternative to pro-
viding larger portable or mobile units is to increase the
expansion ratio of the original starting material. The com-
mercially available foam formulations of urethane and foamed
concrete available at the beginning of the EPA program offered
an expansion of only up to 35:1. It was of interest to exam-
ine the use of material which would offer at least 3 to 4
times these expansion ratios. This phase of the report deals
with the screening,selection and evaluation of highly expanded
materials capable of serving as a suitable hazardous material
land spill barrier. Some of the work performed in this phase
served as a basis for the higher expansion urethane reported
in the previous phase.
The search for candidate starting materials focused on ma-
terials which were polymeric in nature or capable of being
polymerized or gelled to a highly expandable, rigid chemi-
cally inert foam structure. There were two candidate ma-
terials -- water soluble polymers and conventional rigid foam
systems. The water based systems were capable of high ex-
pansion but needed development in rigidizing. The rigid
foams had not been developed for high expansion.
The primary requisite for each candidate system was that it
be amenable to foaming. Methods of generating a foam varied
considerably due to the oeculiar physical and chemical prop-
erties of the test solutions. For the most part, conven-
tional foam generating hardware was used and modified as re-
quired.
Water Soluble Systems
Polyvinyl Alcohol - Borax Gel
Primary emphasis in the water soluble polymers focused on
solutions of polyvinyl alcohol since such systems can be
solidified after foaming. In general, different grades of
PVA exist and the properties are dependent upon the molecular
weight and the degree of hydrolysis. Maximum film strength
and solvent resistance are generally obtained from the high
molecular weight, highly hydrolyzed grades. All grades of
PVA are water soluble but the types of interest were generally
only soluble in boiling water. Prepared PVA solutions of
5% and greater are extremely viscous and not amenable to con-
ventional foaming.
53
-------
A 1% solution of PVA can be readily foamed, especially with
the aid of a surfactant. The foams produced possess no out-
standing properties although the mass of foam may have a'
somewhat longer life than normal surfactant foam. The in-
teraction of borax with PVA is rather striking. For example,
as little as 0.1% of borox in a solution of 5% PVA causes
immediate gellation. Physically, the entire solution forms
a firm, solid mass.
All attempts to produce a high expansion foam from solutions
of PVA and borax were unsuccessful. The major difficulty
with this system was the rapid rate at which gellation
occurred. Instant gelling occurred in the solution if'there
was any delay time in foaming the solutions after they were
mixed. Similarly, any solution holdup, runoff or drainage
almost immediately formed a firm immovable mass which caused
plugging or fouling in the generating hardware.
It was obvious that this system posed serious mechanical
problems. Admittedly, it may have been possible to generate
foam using a more highly developed mixing chamber. However,
an examination of the PVA-borax gel prepared without foaming
indicated a texture that was firm but somewhat brittle.
Ideally, materials best suited for foam are somewhat free
flowing and elastic. These properties are generally impor-
tant in preserving structural integrity of the foam while the
excess water drains or evaporates. In view of these mechani-
cal and physical weaknesses it was considered that the PVA-
borax system was a poor choice for further development.
PVA-Congo Red Gels
Solutions of PVA can also be gelled by the addition of Congo
red. This system forms a thermally reversible gel in that
below 40-45°C the solution is solid and above which it is
liquid. The warm liquid solution foams exceptionally well
with or without the addition of a foaming agent or sur-
factant. The properties of the PVA-Congo red gel were found
to be more consistent with the needs of this program. In
general, the gels were extremely adhesive, tough and elastic.
Foam masses dried to an extremely tough durable cover which
adhered well to smooth surfaces such as glass and metal.
Solutions for foaming were prepared as 5 wt % solutions of
PVA in water (duPont Elvanol 72-60). In preparation to
foaming, the solution was raised to 70-80°C and Congo red
added at 0.3-0.4 wt %. In some cases a liquid surfactant
was added. The surfactants were generally maintained at
about 1-3 wt % in the final solution. Generating low ex-
pansion foam from a 40-50°C solution could be accomplished
by a number of methods. However, the foams blown could not
be considered for diking systems. In general, foams blown
from solution at 45°C were of low expansion requiring about
54
-------
5-10 minutes to develop a reasonably non-flowing structure
but, some several hours to develop a strong matrix. The
dried masses were surprisingly strong and resistant to water
and^other chemicals, such as acetone.
The results of preliminary tests were encouraging but the
time-required to develop strength along with the need for a
warm solution were recognized as major weaknesses. In
addition, preliminary results were obtained at a relatively
low expansion of 30:1. Studies were conducted to circumvent
the need for elevated temperatures and to examine the foam
quality produced at higher expansion. Hopefully, eliminating
the temperature requirement would also benefit the set time.
Several attempts were made to blend the two separate solutions
together immediately before foaming. In these tests the in-
gredients were at room temperature. Aqueous solutions of
PVA at 6% and Congo red at 0.4% were charged in separate
pressure tanks. Each tank was pressurized to 100 Ib with
nitrogen. Needle valves at the exit end metered the flow,
from each tank. The output of both tanks were blended to-
gether immediately before entering the high expansion prop-
erties system. Results of these runs showed the PVA Congo
red foams without added surfactant could not be foamed. With
an added foam agent, a high expansion foam (150:1 expansion)
was formed which possessed poor structural strength and no
apparent life or stability.
PVA-Metal Ion Gels
Other reagents reportedly able to gel PVA were also examined.8
These included chromium (III) and titanium (IV). Chromium
(III) was generated in the final mix by incorporating sodium
chromate in a slightly acid solution in one tank and a 6%
PVA solution containing a small amount of sodium sulfite in
the second tank. The intent was to combine the acid and
sulfite together and reduce the chromium (VI) to chromium
(III). The chromium (III) was then to gel the PVA. In a
similar manner attempts were made to gel the PVA by the use
of titanium (IV) added in the form of potassium titanium
oxalate. The results were similar to those obtained with
Congo red. Foams could not be produced using the high ex-
pansion system without the presence of a foaming agent. Even
with a foaming agent the freshly generated foam failed to
develop any structural strength. At expansions greater than
150:1, which could only be achieved with the use of added
foaming agents, the life of the foam was exceptionally short.
There was no true gelling and generally total collapse
occurred within one hour.
55
-------
It appeared that the PVA-gellant systems were not amenable
to conventional high expansion generation methods. Another
mode of operation was explored. The two streams consisting
of 6% PVA and 0.4% Congo red were metered together at a cross.
The blended streams were simultaneously mixed with a flow of
nitrogen and the combined output was fed to a one foot by one
inch diameter chamber of metal helices.
The foam produced by this system was of low expansion (*\»30:1)
but with improved life. However, the foam was a free-flowing
mass with no structural strength.
Miscellaneous PVA Studies
In additional efforts with PVA systems, several attempts
were made to minimize the use of some of the water by dis-
persing the Congo red in a mixture of methylene chloride and
Freon 12. The intent was to use the Freon as a blowing
agent. The PVA and the anhydrous Congo red solution were
mixed and foamed into a warm tray to volatilize the Freon.
Foams produced in this manner were of low expansion (30:1)
and in all but one case were relatively quick to shrink and/
or collapse. In one singular case a stable structure was
formed which when thoroughly dried was exceptionally tough
and durable. The matrix of this material consisted of
bubble sizes averaging 1/4 inch diameter. This was not re-
produced in subsequent tests. It was strongly suspected
that this gel was similar to the thermally reversible gel
discussed earlier.
In addition to the above approach, numerous support materials
were added to the Congo red-PVA system in an attempt to build
immediate strength into the foamed mass. Table 9 lists the
additives tried but all such attempts failed.
V
Emphasis shifted to other water soluble polymer systems,
particularly those which could be employed at higher initial
concentrations. Of special interest were the Acrysols of
Rohm and Haas. This class of materials is acidic and when
neutralized forms soft gels at concentrations of 1-2% of the
neutralized polymer. The Acrysols are supplied at ^28%
concentration in water. ASE60 was used as representative
of the Acrysols.
Neutralization and subsequent gellation can be induced by
the addition of a number of basic materials including sodium
and ammonium hydroxide, amines, etc. Neutralization by
sodium carbonate was of special interest since carbon dioxide
would be liberated to enhance the foaming action.
56
-------
TABLE 9 - EFFECTS OF ADDED POLYMERS ON PVA-CONGO RED SOLUTIONS
Base Material
96 ml of Additive
Remarks
in
240 cc of 5X
66 cc of 3% Congo
Hater
0.51 Poly H-295
PCS Glue,2%
Stymer S, 51
Methyl Cellulose,21
Acrysol GS, 31
Low molecular weight
PVA, 5*
Methocel MC. 21
Gels fast, develops a strong elastic
foam.
Good foam, relatively stable (^16 hr),
weak structure.
Foams well but unstable.
Cells Immediately, no foam.
Forms a gel with Congo red. Foamed
mass weaker than PVA-Congo red.
Good foam but unstable.
Poor foam, weak structure.
Gels with Congo red alone. Inferior
to PVA system.
240 cc of 5X
50 cc of 4S Congo Red
Ami sol. 31
Gelgard
PVP-K-30, 10%
(Polyvlnylpyrrolldone)
Amisol, 121
Carbopol 941, O.SJ
Slow to set, excellent foam quality
after 16 hr.
Similar to PVA Congo red alone
Unstable foam
Good foam, extremely tacky,>4 hr
stability, weak structure
Unstable foam
(i) Elvanol 72-60 (duPont)
(i) Congo red dispersed in water
-------
A number of runs were made in which ammonia vapor and vapors
of various amines such as triethylamine and diethylenetri-
amine were metered into the air stream. The results were not
fruitful. When the amine or ammonia was used in lesser
amounts, the generated foam was no better structurally than
the ASE60 used alone. At higher stoichiometric amounts of
the basic vapors, gellation occurred in the generator causing
immediate clogging.
Use of inorganic type base such as sodium hydroxide and
sodium carbonate produced a somewhat less elastic type foam.
With the addition of sodium carbonate, the liberated CO?
caused foaming of the gelled solution but in all cases the
generated foam possessed a wet strength no greater than
shaving cream.
When it was apparent that none of the systems investigated
showed promise for the further development of a foamed mass
with sufficient wet strength to serve as a spill barrier or
diverter, this phase of the program was terminated. This
was prompted additionally by the fact that a parallel pro-
gram conducted intermittently with the studies described
above but focusing on a conventional rigid foam formulation
was showing considerable promise.
Conventional Rigid Foams
The most simplified polymeric system from an application
point of view is undoubtedly the rigid urethanes. The
nominal working density of conventional polyurethane is
about 2 pcf although densities of the blown foam can approach
that of soft pine or about 22 pcf. Although there have been
reports of urethane as low as 1 pcf, which offers an ex-
pansion of about 75:1, there was little if any information
of lightweight material in the 0.25-0.50 pcf range.
An alternative to the urethane is the urea-formaldehyde type
formulations. Several contacts with industrial applicators
of this type system indicated that this finished foam would
not retain sufficient structural strength. A sample of a '
urea-formaldehyde foam at approximately 0.6-0.8 pcf was re-
ceived and was found to be exceptionally weak. In fact,
the material hardly survived shipment to the laboratory. A
concerted effort was therefore directed toward the develop-
ment of a urethane type formulation with the basic urethane
properties but a density of less than 1 pcf. This work was
initiated in parallel with that of the first phase on this
program. Formulation investigations to yield urethanes less
than 1 pcf were performed in this phase along with preliminary
tests. When it became clear that densities less than 0.5
58
-------
to 0.6 were not practical, further development and testing
of the higher densities were transferred to the first phase.
Low Density Urethane
The chemistry of urethane foam is complex. In brief, the
essential ingredients of a urethane formulation consist of
a polypi, polyisocyanate, a surfactant, catalyst and blowing
agent. In operation, the polyol and the isocyanate in the
presence of a suitable catalyst react to form the plastic
polyurethane. The reaction is accompanied by the evolution
of considerable heat which tends to boil the added blowing
agent,in turn causing the plastic mass to foam. The sur-
factant or cell control agent regulates the foaming phase
so that uniform cell structures are obtained. Although
these are the necessary ingredients of the foam, numerous
combinations of ingredients are possible due to the avail-
ability of a selection of polyols, catalysts and blowing
agents. As stated earlier, blowing agents most commonly
used are the Freons; however, any low boiling component
could be considered. Water is not only a reactant (polyol)
but can also serve as a blowing agent since the reaction of
the isocyanate with water yields carbon dioxide.
Laboratory development studies were performed to assess the
effects of varying the essential ingredients. In practice,
all ingredients except the isocyanate were first added and
mixed with a stirrer. The isocyanate was injected into the
stirred solution using the calibrated barrel and plunger of
a syringe. A representative type formulation for a 2 pcf
density urethane foam is shown below:
Activator Parts by Weight
Isocyanate - Mondure MR (Mobay) 138
Resin Mix
Polyol - Poly G435 (Olin) 129
Blowing Agent - Fluorocarbon II 40
Catalyst DABCO 80-20 (Hoodry) 1
Surfactant DC 193 (Dow Corning) 1
Familiarizing runs were made using a base formulation and
merely incorporating more blowing agent into the mix. The
excess blowing agent did produce a somewhat higher expanded
foam but the foam was of poor quality as evidenced by the
relatively large and uneven cell sizes. Eventually, with
increasing amounts of blowing agent, additional amounts of
catalyst were required to initiate the reactions. Foam
59
-------
densities ranging from 0.8 to 1.0 pcf could be obtained by
this approach; however, the solution now was unwieldy with
the excess blowing agent and catalyst.
Water was next added to the mix with triethylamine (TEA)
found to be the most effective catalyst. Numerous attempts
then followed to obtain the proper combination of polyol-
water and the required amount of TEA catalyst. The work was
performed by maintaining a master mix of the polyol, sur-
factant and fluorocarbon in one solution and varying the
amount of water and/or catalyst. Excellent foam textures
could be produced in the range of 0.6 to 0.7 pcf. However,
all foam was found to shrink upon standing. Shrinkage was
attributed to the pressure differential formed in the in-
dividual cells when the gas phase cooled. The degree of
shrinkage was sufficiently serious to indicate the likeli-
hood that the applied foam would tear itself away from the
surfaces upon which it was deposited. Ultimately, it was
found that non-shrinking foams could be formed by using
methylene chloride in place of the fluorocarbon. With some
modification in the polyols and the amount of TEA catalyst,
it was possible to form a high quality foam in the range of
0.6 pcf (expansion ratio 115:1).
A secondary development of the study occurred when it was
found that water with TEA could react with the isocyanate to
yield a foam-like material. By optimizing the amount of TEA
catalyst it was found that high quality foam down to 0.29 pcf
could be routinely made in the laboratory. Visually, there
were no apparent differences in the water/TEA/isocyanate foam
generated at 0.3 and 0.6 pcf when compared to material with
the conventional 2 pcf density. The lightweight material
was, of course, weaker but considerably less brittle than
the high weight material.
Evaluation Tests
The screening studies had been unsuccessful in defining a
useful water soluble polymer system, but the urethane type
materials were sufficiently encouraging to move candidate
systems on to further evaluation. The failure of the water
soluble polymer approach to yield a promising candidate was
attributed to the use of excess water associated with these
systems. Realistically, the foam masses which were formed
consisted of 95-97% water. In general, the excess water
provided the necessary flexibility in the generated foam to
allow for shrinkage during the subsequent drainage or evapo-
ration stages. At the same time the presence of available
water inhibited the formation of the required rigidized
structure. Maximum strength would not be achieved until the
foam masses approached dryness and this required several
60
-------
hours at best. The only systems which miqht circumvent the
difficulties involved either increased temperature and/or
the handling of highly dangerous peroxide type chemicals.
The results obtained with the low density urethane-type foams
were quite encouraging. The expansion with the lightweight
material (230:1) was undoubtedly much greater than could
possibly have been achieved with a practical'water soluble
polymer system.
The evolved formulations were of particular interest for
another reason. Both formulations (0.6 pcf, 0.29 pcf) util-
ized water. It was now possible, if required, to incorporate
into the urethane mix the water soluble polymers which by
themselves were found ineffective. In this respect the water-
isocyanate reactions were the quick set mechanisms lacking in
the water soluble polymer systems. At this point the possi-
bilities for a high expansion polymer system appeared very
promi sing.
Preliminary spray tests using conventional urethane spraying
systems were conducted on both formulations. The results of
these tests are described separately.
Water-Isocyanate Mix (0.3 pcf)
The extremely low density formulation consisted of the
following:
Activator Parts by Weight
Mondure MR (Mobay) 10
DC 195 -v.0.25
Resin
Water 2
Triethylamine (TEA) 2
The two reagent solutions were stirred thoroughly immediately
before charging to two separate pressure pots. Each pot con-
taining the mix was pressurized to 100 Ib with nitrogen.
Throttling valves at the end of each tank metered the flow
from each side. Both streams were fed to a conventional
Binks 18 FM Mine Gun which is standard urethane hardware.
Spray tests were conducted indoors at about 70°F. Optimum
foam generation using this system consisted of 3.5 parts
activator to 1 part water-catalyst mix.
The foam mix was sprayed onto both smooth cardboard and con-
crete. The freshly sprayed material appeared as a free-
61
-------
flowing liquid and tended to run off vertical surfaces. Cream
time or the time for initiation of rise was approximately 3
seconds. After the initial period, the rise of the applied
foam was remarkably fast. Time to achieve full rise was
about 5-7 seconds.
It was found that the applied foam did not adhere well to
either the cardboard or concrete surfaces. In addition, con-
siderable amounts of vapors of TEA were evolved during the
actual spraying and subsequent reaction. Sections of the
applied foam showed a density of 0.35 pcf which approached
the laboratory data.
It had been suspected that in the formation of a lightweight
foam much of the cell structure would remain open. It was
of interest to determine the resistance of the specialized
foam to penetration by liquids.
Sections of the sprayed foam were immersed under 8 inches of
water containing a dissolved dye (Congo red) and a wetting
agent (Aerosol OT). Time of immersion was about 16 hours.
By visually examining sections of the test foam it was found
that depth of dyed liquid penetration was generally about
1/8 inch into the foam interior. Some sections of the test
specimens showed deeper penetration but these occurred at the
interface between two distinct layers of foam. Thus, to
build up a layer of foam 4 inches or more, it was necessary
to apply the spray in at least two passes allowing the first
coat to expand fully. Maximum dye penetration was observed
along the surface between the two layers.
These tests of foam generated with conventional hardware
showed the major weakness to be the lack of adhesion to smooth
surfaces. In all other respects the generated foam was
structurally sound.
Methylene Chloride Mix (0.6 pcf)
The formulation for a 0.6 pcf foam consisted of the following:
Activator Parts by Weight
Mondure MR 1.3-1.4 per 1.0
of resin mix
62
-------
Resin Parts by Weight
Polyols G460 24
UI800 3.5
Surfactant DC 195 1
Methylene Chloride 8.2
Water 11.4
DMEA 11.4
Spray tests were conducted similar to those described with
the lighter material. Optimum foam production was achieved
at a ratio of 1.3-1.4 activator to 1 of resin mix. Fumes of
DMEA were minimal.
. T
The sprayed solution was free flowing with a cream time of
about 4 seconds and reaching a full rise in about 8 seconds.
Sections of the foam showed a density of about 0.6 pcf and
were obviously much stronger than the 0.35 pcf material
sprayed earlier.
The sprayed and cured foam was found to be only weakly sealed
to smooth cardboard. Whole chunks of foam could easily be
freed from the cardboard surfaces. A much improved adhesion
was evident on the smooth concrete. In some regions it was
apparent that adhesion to concrete was stronger than the co-
hesive force within the foam itself, Subsequent liquid pene-
tration tests showed no significant penetration by dyed water.
Overall, the 0.6 pcf density material showad a definite im-
provement over the 0.3 pcf formulation, especially in the
adhesion on concrete. However, when compared to a conven-
tional 2 pcf material, the 0.6 pcf formulation was under-
standably inferior, both with respect to adhesion and strength.
The results of these initial spray tests indicated that the
0.6 pcf formulation possessed the capability of substituting
for the 2 pcf but in terms of strength alone was obviously
less sufficient.
Additional Development and Evaluation
Unquestionably, the major weakness of the water-isocyanate
foam was its apparent lack of adhesiveness. This was es-
pecially true because of its extreme lightweight and in-
herently weaker structure. Additional laboratory studies
to improve adhesion consisted of adding water soluble polymers,
especially film forming varieties, to the original water-
isocyanate mix. The polymers were added to the water phase
of the mix with the intent that the reaction with isocyanate
would remove all water,leaving a dried film of polymer to im-
prove adhesion and possibly cohesion. The results in general
63
-------
were unsatisfactory, however. Materials tested and the re-
sults are given in Table 10.
A series of spray tests were conducted on concrete in which
the foam was applied 10 inches high in a 3 foot wide circle
to form a container. Water was added to the inner area.
With a four inch height of water, seepage at the concrete-
foam interface was apparent. In addition, water was found
leaking through channels which had inadvertently formed in
the barrier wall. With about 5-6 inches of water, the barrier
uplifted from the concrete. There was no evidence of fractur-
ing or tearing of the foamed mass indicating that structurally
the foam would serve well as a barrier if the problems of
fixing it firmly to the substrate were solved. These results
clearly indicated that because of the buoyancy the light-
weight material probably required exceptionally good ad-
hesion qualities, even to dry surfaces. Although the higher
the expansion the greater the amount of material delivered
from a portable unit, the unsolved problems would appear to
limit the practical urethane systems to densities of 0.6 pcf
or greater.
64
-------
TABLE 10- TESTING OF SUPPORT MATERIAL
TO IMPROVE ADHESION
Added Polymer
Rhoplex B60A
Polyethyleneamine
Geon 351 (Latex)
Hycak 1570
PVA 570 (PVA 72-60)
Sodium silicate
PVA 205K
Ami sol
Acrysol ES
PHmatloc
Concentration*
50%
50%
50%
50%
5%
6%
10%
10%
10%
10%
Results
Good foam, poor adhesion
Good foam, 0.5 pcf
Poor foam
Did not foam
Good foam, poor adhesion
Good foam, poor adhesion
Good foam, poor adhesion
Good foam, improved ad-
hesion
Poor foam
Poor foam
*denotes the amount of solid polymer or additive
in the water portion of the mix.
65
-------
SECTION VII
HIGH EXPANSION FOAM COVERS
Spilled material contained either by artificial or natural
barriers can still pose a major threat. Release of dangerous
vapor can cause a toxicity hazard not only in-the proximate
location but in downwind areas as well. A potential fire and
explosive threat may exist perhaps accentuated by the move-
ment of men, equipment and existing circumstances. The
danger to the clean-up crew may slow down that operation
allowing time for ground water movement or other means of
moving the contaminant into the water system. This phase of
the report deals with the evolution of soft surfactant type
foams which when applied to a spilled contaminant would form
a protective covering.
Foam Resistance to Wind
Prevailing wind speeds are a major consideration in the
application of detergent foam in open areas. Since foam
covers will perform partially as a function of their thick-
ness, it is necessary to have information on the effect of
wind speeds. Preliminary tests were, therefore, conducted
to determine the resistance of foam heights to various wind
speeds. Tests were conducted in a wind tunnel in which foam
was applied onto a liquid substrate. Trichloroethylene was
contained in a tray measuring 1.5 feet wide by 2.0 feet long
by 1.5 inches high. Various heights of high expansion foam
were applied by conventional means onto the surface of this
simulated contaminant.
The foam formulation consisted of polyethyleneimine, poly-
vinyl alcohol and MSA long lasting foam agent. This particu-
lar formulation was selected since it reasonably approached
the type of ingredients which were expected ultimately to be
used as a vapor barrier. The expansion of the foam used in
these tests was about 200:1.
The test results were not able to establish clear-cut boundary
conditions. Sufficient experience was gained to indicate that
at 9 mph windspeeds all high expansion foam masses >4 inches
could easily be uplifted and completely displaced from a liquid
surface. The maximum allowable height in a 5 mph wind was
about 4-6 inches. Even in this latter case it was strongly
suspected that the ability of the foam to anchor onto the
sides of the tray and the outside surfaces contributed some
stability to the total foam mass. A foam height of 6-8
inches under the same 5 mph windspeed was obviously dis-
placed from the liquid surface.
67
-------
These tests indicated vulnerability of high expansion foams
to wind in the 5-10 mph range. Thus, maximum vapor suppres-
sion and protection should be accomnl i shed with foam heights
no greater than 3-4 inches.
Foam Candidates
In general, foamed masses are delicate systems whose prop-
erties are continuously changing with time. The stability
of a foam is dependent upon a number of factors but the
greatest threat to the life of the foam is posed by the chemi-
cal and physical properties of the contaminant. The numerous
types of potential contaminants, each with its own unique
properties, render it impossible to predict specific foam
qualities suitable for all possible contaminants.
A screening program was designed in which various preparations
and methods could be evaluated at least on a comparison basis.
These screening studies were conducted in the system shown in
Figure 13. A stainless steel holding tray (3-7/8 in. wide
by 6-5/8 in. long by 3/4 in. high) was housed in a plexiglass
box measuring 6-1/2 in. wide by 18 in. long by 9 in. high.
A sweep of 50% RH air entered at one end of the box and
purged the escaping vapors to the mixing chamber at the exit
end. Volume flow of sweep gas was 0.6 cfm. All tests were
conducted at ambient conditions
A typical test consisted of charging 120 cc of liquid con-
taminant to the holding tray. The exposed liquid surface area
was 25.6 square inches. With the charged tray in place, the
air sweep was activated. The free evaporation of the con-
taminant liquid normally resulted in depressing the tempera-
ture of the bulk liquid. In most cases, subsequent thermal
equilibrium was established at a temperature several degrees
less than room temperature. Samples taken at this time were
indicative of vapor evolution from an uncovered source. At
this point the test foam was applied by filling the entire
box to the required height. The amount of contaminant vapor
present in the gas phase was taken as a measure of the
effectiveness of the foam cover.
The total volume flow of sweep gas was maintained constant
for all runs. The significance of this is that as the
applied foam collapsed the void space increased thereby de-
creasing the velocity of the sweep across the foam surface.
It was felt impractical to maintain a constant velocity of
the sweep gas especially where foams were collapsing rapidly.
The variation in sweep velocity was of significance only in
instances where deep beds of foams were examined.
68
-------
FIGURE 13 - LABORATORY SET UP FOR SCREENING OF HIGH EXPANSION FOAM CANDIDATE COMPOSITIONS
-------
Initially, vapor analyses were performed using MSA detector
tubes. These devices are selective, rapid and reasonably
quantitative. As work progressed and the inadequacy of the
detector tubes became apparent, analyses were performed using
the gas chromatograph. This instrument is extremely selective
and offered a capability of coping with a several thousand
fold change in concentration.
Conventional Foams
A conventional fire-fighting foam was evaluated to determine
the covering capability of material which would be readily
available. MSA standard fire fighting foam was selected as
a typical material and toluene as the contaminant. The
selection of toluene was based on its relative chemical in-
ertness or absence of reactive functionalities (-OH.^C = 0,
-C^JJ , etc). Toluene is lighter than water and is also a
very poor solvent for water;hence it would not be affected
by the foam drainage.
A 6.5 inch height of foam was applied to the toluene from a
3% solution of MSA foam agent using the generator shown in
Figure 14.
The collapse of the foam was evident almost immediately. After
about 15 minutes of contact with the liquid toluene, the
initial 6.5 inches of foam cover had completely collapsed
exposing the liquid surface. The normal collapse rate for
this foam in the absence of environmental effects is 8-10
inches per hour. A single sample taken 12 minutes into the
run showed a toluene content of 125 ppm in the sweep gas
phase versus a value of >800 ppm for the uncovered toluene.
Actually, the 800 ppm level is the maximum quantity of toluene
which could be reliably detected by the detector tubes. Sub-
sequent analyses by the gas chromatograph showed toluene
vapor evolved from an uncovered tray in this type of test
to be several times this value.
An additional test was performed using a 6% solution of foam
agent. The results were similar -- complete collapse of the
foam within 20 minutes. The rate at which this foam collapsed
was surprising especially in view of the presumed inertness
of the selected contaminant. These data must not be con-
strued as demonstrating the complete ineffectiveness of
existing foam agents. Actually, the limited data collected
showed a very significant reduction in vapor evolution up
until the foam collapsed. In cases of emergency where other
facilities and materials may not be available, the use of
70
-------
FIGURE 14 - LABORATORY FOAM GENERATOR SYSTEM
71
-------
existing foam agents could provide a useful service. These
laboratory data Indicate that the beneficial effects would be
short lived but repeated applications could be made.
Long Lasting Foams
Since vapor suppression was obviously associated with the life
of the foam, a proprietary preparation of MSA was of special
Interest. This formulation consists of polyethyleneimine,
polyvinyl alcohol and MSA foam agent and in the past had been
the source of foams with lives of over 3 months or more.
Table 11 shows the performance of an 8 inch layer of this
foam generated at an expansion of about 200:1. The results
can be compared to a 0.5 inch layer using the same formulation
TABLE 11 - COMPARISON OF THICK VERSUS THIN LAYERS
OF FOAM COVER ON LIQUID TOLUENE
Elapsed Toluene Vapor Content (ppm)*
Time 0.5 1n. Cover 8 in. Cover
(mln) Low Expansion (4:1) High Expansion (200:1)
Uncovered >800 >800
30 34 20
45 12
60 45
110 12
140 55
175 50
230 50 Foam collapsing
240 100
260 300, liquid surface
exposed
16 hrs 250
*Toluene vapor content in air sweep
but dispensed at an expansion of about 4:1. Toluene served
as the contaminant in both cases. In the latter case, the
fine foam was generated using blow pipes equipped with fine
fritted ends. The results obtained in both cases showed a
marked improvement. Of equal Importance was the significant
vapor suppression achieved by exceptionally thin beds of the
72
-------
foam, The data shows excellent vapor suppression up to 16
hours which was the effective life of this covering. Ordi-
narily, the collapse of this foam during a similar time
period but without the presence of a contaminant would have
been less than 0.1 inch.
New Foam Development
The results of these preliminary tests showed:
1. the limitations of existing fire
fighting foam agents
2. the effectiveness of thin but
densely packed beds of foam
3. the adverse effects on foam life
posed by the contaminant.
Emphasis at this time shifted to a foam cover which could
cope with the more chemically reactive species such as alco-
hols, ketones, etc. These materials are not only produced
in large quantities but were known to be effective poisons
for surfactant foams.
The most difficult class of compounds to smother effectively
are the low molecular weight, volatile, highly polar solvents
with methanol as an outstanding example. Most foam prepa-
rations collapse instantly on contact with the alcohol. With
continued addition of foam solution the subsequent foam
drainage alters the alcohol surface tending to decrease the
adverse effects giving a false sense of foam stability. On
a practical basis, however, where large spills were involved,
such dilution would not occur.
For the purpose of further materials screening, methyl alcohol
was selected as the test contaminant. Three approaches were
taken which showed considerable promise. These included:
solution of silicates which formed a relatively hard layer
of solid silicates on the surface of dry methanol, solutions
of pectin, which formed a gel on the surface and solutions
of Acrysols which also formed a thick rubber-like coating on
top of the methanol. Both these latter types of cover event-
ually dried to a hard impervious crust. All three materials
were best applied from a low expansion system.
Silicate System
The silicate solution employed was a commercial grade water
solution containing approximately 30% of dissolved sodium
73
-------
silicate. This solution was diluted with approximately an
equal portion of water and a foaming agent added to maintain
its concentration at about 4%. Foams blown from this formu-
lation collapsed on contact with dry methanol; however, a
danse, crystalline structure resulted. If applied at a high
expansion ratio, insufficient material was available to form
a continuous blanket. In that case the crystallized segments
sank. At low expansion, the generated foam "froze" into a
solid mass with sufficient entrapped air to remain buoyant.
Subsequent tests showed that the silicate formulation was
effective for methanol but was not suitable for other non-
alcohol type materials, polar or otherwise.
Acrysols
The Acrysols are a class of acid,type polymers which when
neutralized by a base form a solid gel. A foam of this
variety is difficult to prepare since gellation occurs
immediately upon contact with the base. Low density foams
could be produced by using solutions of sodium bicarbonate.
The COg liberated from the mix served as a blowing agent to
expand the foam.
In contact with alcohol the gelled foam formed a thick coat-
ing which tended to float on the surface of the alcohol.
Some dissolution of the gel occurred but the rate was slow.
This type of polymer was also an effective cover for ammonium
and amine type compounds where gellation occurred uponfcon-
tact. Like the silicate, it was limited to a small group of
materials.
Pectin Type Foams
The most widely applicable foam was prepared from citrus
pectin. Pure citrus pectin is a powder which dissolves
readily in water provided sufficient mixing is available.
It is commercially available, in preparations offering a range
of set times and thickening power. Chemically, citrus pectin
is a weak acid and maximum thickening is achieved at a pH
of about 3.1. Once the added powder is adequately dispersed,
wetted-out and dissolved, the process of thickening begins.
Normally, a 1.5 to 2.0% solution of pectin in water becomes-
viscous within 30-60 seconds. On a practical basis some
thickening of the solution begins immediately following the
introduction of the powder. Foams generated immediately
after dissolving the pectin showed exceptional capabilities
of minimizing the evaporation of a number of substances.
Similar to the above systems, maximum results were obtained
by applying the pectin in a low expansion (30,:1) form. It
was successful in part because the contaminant was swamped ?
74
-------
with the pectin solution. Immediately upon contact, the foam
collapsed slightly but entrapped finely divided air bubbles,
forming a tough impermeable buoyant skin on the surface of
polar contaminants. With the non-polar type, contaminant
skin did not form but the foam remained intact behaving some-
what similar to conventional type foam agents. In both cases,
however, the foam eventually thickened and dried to a perma-
nent structure.
The data obtained with pectin solutions is presented in
Figure 15. The data is shown as the percent reduction of
vapor found in the gas phase with the uncovered samples used
at 100%. Effective reduction in vapor evolution was achieved
for a variety of compounds with different chemical entities.
The results achieved with these formulations and especially
with solutions of pectin were exceptionally promising. They
were able to be floated on a variety of materials,even
methanol and acetone with quite low liquid densities, with
good control of vapor release.
These results were obtained by mixing 4 gms of citrus pectin
(Eastman Chemicals, Cat. No. P2569) in 200 cc of water and
adding 4 gms of a foaming agent such as MSA salt water foam
agent. The pectin powder was added to the liquid solution
of water and foam agent in a blender and mixed at high speed
for 10 seconds. Foam was blown using a blow tube with a fine
fritted end to produce small bubbles. Foam produced in this
manner was generally expanded 2-5 fold. It was difficult to
control the expansion ratio since the properties of the so-
lution were changing rapidly. Normally, the height of the
cover rarely exceeded 0.5 inches above the liquid surface.
Analyses in these tests were performed using a gas chromato-
graph. Samples of the contaminated sweep gas were taken in
a 1 cc hypodermic needle and injected directly into the gas
chromatograph.
There are some inherent disadvantages to the use of pectin.
Water is the only solvent and because of the viscosity prob-
lem, the maximum concentration of pectin in water is limited
to about 1%. Further, once the pectin is added to the water
the properties of the resulting solution begin to change.
Thus, the characteristics of the solution exiting at a foam
nozzle depend upon the pectin dwell time in the line. For-
tunately, dwell time can be controlled by inserting the pectin
into the water line near the nozzle, at the pumper or at in-
termediate locations.
Introduction of the pectin into standard systems posed some
problems, however, A powder eductor was examined as the
first means of introducing pectin into a hose line. This
75
-------
TOO
01
•M
ra
cc.
•a
a*
i.
a*
>
o
u
c
c
J_
Ol
>
o
I I I
c
Ol
u
J_
0*
a.
1.1 I
Acrylonltrlle
i i i
123456
Time - Hours
123456
Time - Hours
100
100
a>
•M
•o
a*
t-
a>
>
o
u
c
Acetone
u
t-
01
at
o>
j-
o
u
c
Xylene
I I i
i 1
Percent
V
• >w Toluene
12345
Time - Hours
12345
Time - Hours
FIGURE 15 - VAPOR SUPPRESSION OF TYPICAL HAZARDOUS
MATERIALS BY PECTIN MODIFIED FOAM COVER
76
-------
device consists of a large funnel attached to the intake of
an eductor. Sufficient pectin could readily be injected into
a 10 gpm water line to maintain a pectin concentration of well
over 2%. The exiting water solution was collected and showed
the pectin to be finely dispersed with no obvious lumps or
aggregates of undissolved powder. To add both foam concen-
trate and pectin into the line at the same time, two approaches
were examined: (1) a completely powdered system containing a
solid surfactant blended into the pectin and (2) a stabilized
suspension of pectin in a liquid surfactant. The latter system
would be more consistent with existing foam dispensing equip-
ment.
Subsequent laboratory studies showed powdered amphoteric sur-
factants to be excellent foaming agents for pectin. A powdered
mix was prepared by blending 1 part surfactant to one part
pectin. The problem of preparing a free-flowing, stabilized
suspension of pectin was only partially solved. The best formu-
lation consisted of equal parts by weight pectin, a liquid
alkylaryl surfactant and propyl alcohol as a viscosity modi-
fier. The resulting suspension was only temporary and complete
settling of the pectin occurred within 24 hours.
Several attempts were made to spray the two pectin concen-
trates using conventional fire-fighting type generators. The
solid blend of foam concentrate and pectin was evaluated in a
Feecon type low expansion nozzle. Materials were pre-blended
in a 1:1 weight ratio. Two tests were completed with the
funnel eductor installed at two different locations along the
hose line to provide a dwell time of approximately 5 and 50
seconds. Output foam of the low expansion nozzle was directed
onto 2 feet wide by 4 feet long trays containing about a 2 inch
depth of dry methanol. The output foam in both cases was ex-
ceptionally poor when compared to a conventional foam agent
used alone. The pectin foam was dispensed as a heavy, wet
foam which caused excessive splashing and dilution of the
alcohol. Some foam cover did develop with the shorter dwell
time. However, the solution runoff was exceptionally gritty
indicating the pectin had not completely dissolved. In both
cases, dilution of the alcohol had occurred to render the
test results inconclusive. It was obvious that a major prob-
lem existed in dunlicatinn the quality of the foam used in
the laboratory tests.
An attempt was made to produce a high expansion foam starting
with the liquid "suspension" or slurry of pectin and sur-
factant described earlier. The generator was of conventional
design and capable of expansion of 200-300:1. The liquidized
pectin was uplifted into the line via an eductor at a location
yielding a dwell time of about 20 seconds. A reasonably high
77
-------
expansion foam was produced but was found to collapse readily
on contact with the alcohol. Samples of the runoff and drain-
age again indicated that the added pectin was not completely
dissolved.
Small scale tests were next performed in the laboratory using
a miniature spray system. The device consisted of a 0.5 gpm
spray nozzle and a flat screen of about 1.5 inch diameter.
The entire spray head or generator was housed in a cylinder
measuring 1.5 inch diameter by 3 inches long. The system
operated batchwise; that is, a foam solution was prepared and
charged to a pot. Pressurized nitrogen or air (100 Ib psi)
fed the solution to the generator. The study portion of the
effort consisted of charging the pot reservoir with water,
then when all was in readiness, quickly adding the pectin
mixes followed by foaming. Various time delays were examined
between adding the pectin and the actual foaming. Results •'
showed that optimum foam cover was developed with a 30 second
dwell time. Foams produced at this time were thick creamy
mixtures which formed excellent covers on methanol and acetone.
With holding times greater than 30 seconds, the expansion ratio
decreased until at one minute the foam output was only slightly
higher than 1 to 1. Even at the lowest expansion the foam
(or aerated solution) formed an excellent cover. If dropped
from any height, this material sank deeply into the alcohol but
immediately surfaced and remained buoyant.
The results of these tests indicate that although quality
foam can be produced, a requirement would exist to maintain
control over the operating conditions. Undoubtedly, the
source of the difficulty can be traced to the rate at which
the pectin is dispersed, wetted out and finally dissolved.
If the foaming process is carried out prematurely, insuffi-
cient pectin is dissolved to form an effective cover. On
the other hand, unnecessary delays in foaming allow the
solution to thicken in which case only high energy.type
generators would be effective.
Expanded Rigid Systems
Urethane
The success indicated by the skin or film formers led tofan
examination of other type materials which could be floated
onto methanol, acetone, etc. The use of more rigid light-
weight foams which could be generated in place and floated
onto the spill was next examined. Since urethane was avail-
able it was of interest to adapt this form to the present
need. Undoubtedly, urethane foam could be generated at the
spill site and merely floated onto a contained spill. Thus,
78
-------
thin sections of urethane could be generated and "welded"
by additional urethane mix to generally cover an easily
accessible area. However, this approach similar to stretch-
Ing a plastic sheet across the spill would be at a serious
disadvantage in rough terrain or where tall weeds or grass
were-present and might hinder cleanup operations. To cir-
cumvent this difficulty, a limited effort was applied to
the generation of urethane "snow". The intent was to spray
tiny 'droplets of a quick rise urethane mix into the air and
have the material expand fully by the time it fell onto the
solvent. If such spheres were sufficiently small they could
effectively flow Into and around most obstructions.
Discrete but irregularly shaped balls of urethane were
successfully generated by using an excess of gaseous Freon
to propel the liquid mix from the nozzle gun. An additional
line carrying the Freon gas was fed to the mixing nozzle
to break up the liquid mix into fine droplets. The generated
urethane was allowed to free fall onto trays of alcohol. Al-
though some urethane was sufficiently blown and cured to float
on the. alcohol, most of the deposited urethane was still in
the blowing stage with the result that it sank or dissolved
into the alcohol. The output of foam was small and consump-
tion of gas was extremely high and it was concluded that this
approach was a relatively poor risk for further development.
An attempt was made to apply shredded urethane foam onto
alcohol. Invariably, all attempts to shred the cured urethane
resulted in the production of great deals of powder with no
cell structure. This material in contact with alcohol was
wetted out and rapidly sank.
Microballoons
An examination was made of the applicability of tiny hollow
spheres both as a cover and as a support media to provide
additional buoyancy. The properties of these phenolic type
balls are given in Table 12. In the first test a layer of
spheres up to 1/4 inch deep was applied over the methanol.
Analyses showed no significant vapor suppression occurred
during a one hour test. In the second test the spheres were
added to the pectin foam after the powdered pectin was first
dissolved in the blender. The weight ratio of pectin to
spheres was 2:1. The spheres had a tendency to float on
top of the applied foam rather than be incorporated into the
foam matrix as intended, and the pectin-sphere combination
was no more effective than pectin alone.
79
-------
TABLE 12 - TYPICAL PROPERTIES OF PHENOLIC SPHERES
(BJO-0930, Union Carbide Plastics Co.)
Average particle size (dia.), inch 0.0017
Particle size range (dia.), inch 0.0002-0.005
Bulk density, pcf 3-5
True sphere density 12
80
-------
SECTION VIII
SYSTEMS SUMMARY
The purpose of the program was the evolution*of systems which
could be used in the control of hazardous chemical spills
from a transporting vehicle. Although the development effort
cannot be considered complete, significant progress was made
with the polyurethane and foamed concrete systems. Both
systems can be considered for control in a number of potential
situations. The following paragraphs will synopsize the statu
of both materials, the equipment, their characteristics and
the current limitations.
Polyurethane
The basic objective of the polyurethane development was the
adaptation of existing commercial equipment to the spill con-
trol application. Two portable units have been evolved. The
small unit is hand portable, the larger is available as a
back pack or cart mounted unit. Pertinent details are given
in Table 13.
TABLE 13 - POLYURETHANE SYSTEMS
Part No.
Vol. of foam delivered
Delivery rate (avg.)
Weight (gross)
Size (in.)
Storage time
(tentative)
Storage temp.
Useful temp, range
(substrate)
Model I
510028
22-25 cu ft
5 cfm
27 Ibs
20x10x5
6 mos.
above 50°F
15° to 120°F
Model II
210029
50-55 cu ft
5-7 cfm
65 Ibs
20x20x10
6 mos.
above 50°F
15° to 120°F
81
-------
Operating instructions are provided on the package. The poly-
urethane is dispensed from each unit in a spray cone as a
thick liquid. There is a delay time of some 8 to 12 seconds
after generation before foaming begins. The foam rigidizes
rapidly after expansion. The rigidized foam is inert to a
wide variety of chemicals and when bonded to the substrate
can effectively contain or divert chemical spills.
The polyurethane formulation bonds well to dry surfaces; such
as cement, asphalt and packed dirt. The material also bonds
to itself,thus new foam can be effectively deposited over old
foam. Control on dirt may be limited with failure occurring
due to liquid seepage through the substrate at the dirt-
urethane interface.
Some control is exercised on gravel, rocky or vegetated
ground. Some penetration will ultimately occur below the
substrate-urethane interface. In such situations double or
triple barriers can be set to extend the time of control.
Temperature also has an effect. The unit temperature at the
time of use should preferably be above 50°F, but at least
above 40°F. Substrate temperatures can be as low as 1$°F.
At the lower temperatures expansion may be reduced at^least
for that material applied directly-to the substrate.
At the present time good adhesion is not obtained on wet
surfaces nor can flowing streams be directly blocked. If
the foam can be mechanically locked onto the surface the
problem of wet surfaces can be overcome. This is the case
with sealing storm sewer grates, manhole covers, etc. By
generating foam into a plastic or rubber bag, pipe, culverts
and the like can be plugged even when liquid is flowing.
Although the chemical nature of polyurethane would indicate
resistance to most chemicals, actual evaluation is limited.
Control has been demonstrated for water base liquids except
strong acids, non-polar organics, and some selected materials
as chlorine and ammonia. Polar compounds are a question.
It has been established that the present formulation is not
effective against methyl alcohol.
Foamed Concrete
The objective of the foamed concrete task of the program was
to demonstrate the feasibility of erecting a fast-set foamed
concrete dike and to outline the general requirements for an
emergency field unit. Both objectives have been achieved.
82
-------
Foamed concrete of density about
an extremely fast set (2-3 secon
sodium silicate to form a gelled
strength to build dikes 2 feet h
ment requirements are simple. A
matic of Figure 16 would consist
a cement-water slurry, a slurry
ator, sodium silicate solution,
The preformed foam is metered in
it blends to produce foamed cone
is injected as the foamed concre
ducing a chemical reaction with
rapid set.
40 pcf can be made to take
ds) with the addition of
structure with sufficient
igh or better. The equip-
unit following the sche-
of a mixer for blending
pump, preformed foam gener-
storage tank and nozzle.
to the slurry stream where
rete. The silicate solution
te exits the nozzle, pro-
the cement to cause the
Silicate
Storage
Slurry
Mixer
Shotcrete
Nozzle
FIGURE 16 - SCHEMATIC OF FOAMED CONCRETE SYSTEM
83
-------
A foamed concrete barrier has strength enough to impoun'd
liquid immediately after being placed, and flowing water
of slow stream velocities has been successfully pooled.
The initial gel set has surprising strength, but to impound
liquids to any appreciable depth, such as two feet, behind
the barrier needs additional strength from the hydraulic
set of the cement. This comes on slowly after the initial
gel set and is a function of many factors including water
temperature, water/cement ratio and type of cement. Under
ideal conditions significant strength from the hydraulic
set could be attained in less than 5 minutes. Thus, the
rate of build up of hydraulic head on a newly poured barrier
is probably the single most important factor to successfully
impounding a spill. For this reason, broad, flat rather
than narrow, high-pitched impounding basins are favored.
The type of substrate is not an important factor. Tests
on clay, shale, chipped limestone, grass and weed-covered
ground have been successful. In addition, such chemicals
as methanol, 1,1,1 trichloroethane, phenol, acetone cyan-
hydrin and acrylonitrile do not appear to affect the gel
set action.
As might be expected in a water-based system, extremely
cold, sub-freezing working conditions make the equipment
operations difficult. The preformed foam and solution lines
tend to freeze first. The use of freezing point lowering
additives in the solution has only had moderate success.
A suggested design for an Emergency Field Unit was shown in
Figure 12. The unit consists of the basic components de-
scribed earlier, driven with hydraulic motors. A gasoline
engine furnishes power to the hydraulic pump and a small
air compressor. The components are all commercial off-the-
shelf items. The parts list is given in Table 14.
The unit is trailer mounted, suitable for a pick-up truck
operation. It will produce approximately 50 ft3 of foamed
concrete at 45 pcf per batch within 30 minutes of delivery
to the site. Repeat cycles would take approximately 25
minutes each.
Three such batches (150 ft3) could be produced in approxi-
mately 80 minutes to build barriers of such sample dimen-
sions as:
2 ft x 2 ft x 38 ft
1.5 ft x 2 ft x 50 ft
1.5 ft x 3 ft x 33 ft
84
-------
TABLE 14 - MOBILE FOAMED CONCRETE EMERGENCY UNIT
Item Quantity Description
1 1 Ingersol1-Rand Air Compressor, Type 30,
Model 71T2XGT with gas engine drive,
Wisconsin Model V465D engine and V-465D
power take-off clutch. Gasoline tank
included with motor.
2 1 Hopper similar to MSA P/N A91959, 400#
3 1 Agitator blade similar to MSA P/N 379689
Agitator B91962
4 2 Bearings, P/N 66802
5 1 Slurry pump, Moyno 3L4, P/N 68572
6 4 Motor, hydraulic, Char-Lynn Size AA
7 1 Pump, hydraulic, commercial shearing
P30 Series, 1 1/2 in. gear
8 1 Silicate proportioning pump. Viking Model
HL124S, carbon bushing and stuffing box,
all iron, built in relief
9 1 Foam pump, Viking Model HL156, all iron,
buiIt in relief
10 1 Foam tank, similar to pressure tank paint
container
11 1 Foam gun
12 1 Hydraulic tank, P/N 456313
13 4 Couplings, T.B. Woods #6, P/N 625462
14 1 Speed reducer, Dodge No. 11
15 1 Valve, hydraulic - 4 way 3 position
Double A P/N R6-175-FF-N
16 1 Valve, hydraulic - 3 way 2 position
Double A P/N R6-175-FZ-N
85
-------
Table 14 (Continued)
Item Quantity Description
17 1 Valve, hydraulic - flow control Double A
P/N QXA175 with P15-06 sub plate
18 1 Filter, hydraulic P/N 625468
19 1 Strainer, suction P/N 69806
20 1 Breather and filler cap P/N 69081
21 1 Water control valve P/N 69412
22 1 Air on-off valve, 1 1/2 in. size
P/N 69452
23 1 Gauge, hydraulic press, 0-3000 psi P/N 56414
24 1 Gauge, air pressure, 0-150 psi P/N 68052
25 1 Gauge, water pressure, 0-100 psi
26 1 Battery, Electric, 12V
27 2 Mounts for hydraulic motor with Viking Pump
28 1 Mount for hydraulic motor with Moyno pump
29 1 Mount for hydraulic motor with Dodge speed
reducer
30 Hydraulic hoses
31 Hydraulic fittings
32 Compressor piping
33 Moyno pump piping
34 1 Sheave, hydraulic pump P/N 625869
35 2 V-belt P/N 69753
36 1 Mount for hydraulic pump
37 1 Trailer, Model 12001-2 double axle
12,000 Ib capacity
86
-------
Table 14 (continued)
Item Quanti ty Description
38 2 Bearings, pillow block, 1 15/16 in. dia
P/N 66058
39 1 Stub shaft for mounting drive sheave
40 1 Coupling, T.B. Woods #7, P/N 63951
87
-------
Raw materials for one batch are:
10 bags of cement (940 Ibs)
30 gallons of sodium-silicate solution
1.5 quarts of foam concentrate
Material for one run would be transported on the unit. Ma-
terial for three batches would weigh approximately 4900 Ibs
and require about 1800 Ibs of water (215 gal.) at the site.
Additional material could be brought in if needed.
The sodium silicate and foam concentrate would be the only
material required to be stored in quantity for multiple
batches. Enough Type I cement could be stored for 1 to 3
batches, with additional cement picked up as needed.
A detailed description of the procedure for placing the
mobile unit in operation, with estimated time commitments
for 2 men, is as follows:
Operation Time (tnin)
Start gasoline engine, slurry and 2
slurry mixer
Attach water source, fill tank to 8
level mark (70 gal.)
Add cement (10 bags) 5
Add 1.5 quarts of MSA Reg Foam con- 4
centrate to 10 gal. foam tank --
fill with water to level
Attach silicate drums to pumper 2
Attach foam cement and silicate
quick connect hose to unit 2
Total 23 min
The first cycle including set up and a run time of 10 minutes
should require approximately 33 minutes. Additional cycles
should be possible in less than 25 minutes.
88
-------
Foam Covers
The evolution of a foam system capable of providing an
effective cover for all chemicals was not fully completed
during the course of the program. One prime class of ma-
terials, polar organic compounds,proved difficult. A pectin
fortified foam was evolved which was effective for this class
of materials but a generating system suitable for field use
was not delineated. As a consequence, no large scale foam
evaluations were conducted.
The tests which were conducted, coupled with prior work, do
provide some useful data on foam covers. Although it is
preferable to utilize a foam designed for use as a foam
cover, in emergency situations the type of foam generated
from fire fighting foam generators and agents may provide
some benefit. Such foam should be used with caution. It
should be applied to the greatest depth possible preferably
greater than 12 inches deep and topped off or recovered with
new foam when the depth collapses to 12 inches. It can pro-
vide some degree of control on neutral or alkaline aqueous
solutions and nonpolar organic compounds of low volatility.
The polar compounds as well as acids will destroy it. High
volatile liquids will be afforded only partial protection
since vapor release will cause holes or chimneys in the
cover. With liquids with boiling points below 0°C a frozen
layer will tend to form between the foam and the chemical
which will slowly minimize the chimney effect.
Not all foam agents will provide suitable foam covers. In
some cases a combination of foam and chemical characteristics
will actually exaggerate the spill hazard. Commercially
available fire fighting foam agents provide a broad range
of collapse rates from 8 to 90 inches per hour as well as
wide latitude in water drainage. Some drain as much as 70%
of their water in 2 minutes. The better agents give up less
than 50% of their water in 30 minutes.
The consequences of fast collapse are obvious. Those of
fast drainage are not as clear but more serious. The loss
of water causes two problems -- it reduces the weight and
integrity of the foam and drains water into the underlying
chemical. Dry foam is more subject to wind shear effects.
Thus wind can cause disruption of the foam cover. In some
cases foam masses can absorb quantities of the underlying
chemical. If these are tcxic or explosive, foam masses
torn loose and blown away by the wind can be hazardous of
themselves. Foam masses containing explosive vapors can
be detonated.
89
-------
It should be realized that foam blankets serve two pur-
poses -- they isolate the surface from external effects
which is important for flammable 'or explosive liquids, and
reduce the vaporization rate of the chemical where such
vapors are toxic or explosive. They do not control vapori-
zation by providing an overpressure but by insulating the
liquid surface from radiant energy. Water draining into
the contained spill has an adverse effect on this insulating
action. If the spilled chemical and water are miscible,
the heat of solution can actually adversely effect the spill
situation. This is truly a problem in the case of materials
such as anhydrous ammonia where the heat of solution is very
large.
By using a long lasting foam such as the polyethyleneimine
system of MSAR (Type L Foam), collapse and drainage are con-
trolled to low rates. Drainage is of the order of 40% in
30 minutes with collapse rates of 0.1 inches per hour>dr
less. The foam has greater integrity, is less susceptible
to wind effects and is useful with acidic liquids. These
foam agents can be used in standard high expansion foam
generators. Air aspirating units such as the MSA Mini-X
which deliver 250:1 to 350:1 foam are preferred since'they
provide a heavier foam density and can be operated from a
fire pumper.. These hand-held units can deliver up to 7000
cfnuof foam.
It should be cautioned that the use of a foam cover should
in no way change the procedures which would be employed to
handle the chemical without it. Its purpose is as anvextra
precaution. Further, whenever a foam cover is being applied,
either initially or as a recover, those people applying the
foam should wear full protective gear.' Body protection-such
as the Acid King clothing of Wheeler Protective Clothing
Company with bottled oxygen or the butyl suit of MSA with
a self-contained Chemox rebreather is recommended.
90
-------
SECTION IX
REFERENCES
1. MSA Research Corporation, "Sampling for Red Rock",
Rome A1r Force Base, Contract No. AF30(635)-12289.
2. MSA Research Corporation, "Safety 1n Transport and
Handling of Chemical Warfare Munitions, Eglln Air
Force Base, Contract No. AF08(635)-3674.
3. Modern Plastics Encyclopedia, Vol. 48, No. 10A,
1971-1972.
4. Hoff, G. C., "New Applications for Low Density Con-
cretes", Miscellaneous Paper C-70-8, U. S. Army Engineer
Waterways Experimental Station, Vicksburg, Mississippi,
May 1970.
5. United Kingdom Mines Rescue Service, Nottinghamshire,
Derbyshire and Leicestershire District, "Use of Gypsum
Powders in the Construction of Stoppings", National
Coal Board, Hobart House, London (1969;.
6. Halliburton Company, "New Mine Sealing Techniques for
Water Pollution Abatement, FWPCA Contract No. 14-12-453,
March 1970.
_ts •
7. Huron Cement Company, Alpena, Michigan.
8. Dreyrup, A. S. (to E. I. duPont deNemours & Co.) U. S.
Patent 3,492,250 (Jan. 27, 1970).
91 4U.S. GOVERNMENT WMNTING OFFICEU973 5U-153/Z39 1-3
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
7. R,.
No
w
T:r'
Control of Hazardous Chemical
Spills by Physical Barriers
5.
I*
8.
^ On;.
Friel, J.V.,
On - •• ;i
MSA Research
P u a n c P •{•!•»/ I
Hiltz, R.H., and Marshall, M.D.
Corporation
'annex/I \ran-i a 1 £ f"l "3 Q
w. ',....
FPA
EPA
15090 HRP
68-01-0100
Typ- Rep. ind
F ' ''Oil C- •??'.•:!
12. S; nsorir" Organ* >tion
Environmental Protection Agency report
number, EPA-R2-73-185, March 1973.
i6. Abs-ract The magnitude of potentially hazardous chemicals now being
transported through the country poses a serious threat to the water eco-
system. Unless spills can be controlled at their source, movement into
the water system may be inevitable. Such control dictates the avail-
ability of systems capable of forming dikes or flow diverting barriers
either as a portable system carried on the vehicle or a mobile unit
rapidly deployable to the site. In this regard, a program was insti-
tuted to investigate the applicability of foamed rvterials for forming
such dikes and barriers. It was successfully demonstrated that poly-
urethane could be packaged in a portable unit and dispensed as a low
density rigid foam capable of diking liquids on a variety of substrates,
Attempts to develop a riqid high expansion system were not fully
successful. A foamed concrete system was also successfully evolved,
which used mobile equipment to build free form dikes. Modified sur-
factant foam was also shown to be an effective cover over spilled
chemicals to control vapor release and fire hazards. In each case,
a field tested unit was demonstrated or shown to be feasible.
17a.
*Barriers, *Dikes, *Dams, *Water Pollution Treatment,
*Water Pollution Control, Hazards, Chemicals, Accidents
1 ~b Identifiers
*Spills, *Hazardous Materials, Hazardous Chemicals
;:-. COIV RRFit hi f Gro-i?
05G
IX. .i-ai 'a '>!',"•• 19. Security
fRepo
1. St. -if y
Class.
C 5.
21. No. of
'ages
" P"'*
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHJNGTON D. C. 2O24O
Research Corporation
. '. 'J '-' \ 9 7 I •
------- |