600281214
May :981
MODIFICATION OF SPILL FACTORS AFFECTING AIR POLLITION,
Vol. I - An Evaluation of Cooling as a Vapor Mitigation
Procedure for Spilled Volatile Chemicals
by
J . S. Greer
S . S . Gross
R.H. Hi Hz
M.J. McGoff
MSA Research Corporation
Division of Mine Safety Appliances Company
Evans City, Pennsylvania 16033
Contract No. 68-03-2648
Project Officer
John E. Brugger
Oil and Hazardous Materials Spills Branch
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABuRATOR1--
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI , OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental
Research Laboratory, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation
for use.
i i
-------
FOREWORD
The U.S. Environmental Protection Agency was created be-
cause of increasing public and government concern about the
dangers of pollution to the health and welfare of the American
people. Noxious air, foul water, and spoiled land are tragic
testimonies to the deterioration of our natural environment.
The complexity of that environment and the interplay of its
components require a concentrated and integrated attack on the
problem.
Research and development is that necessary first step i',
problem solution; it involves defining the problem, measuring
its impact, and searching for solutions. The Municipal En-
vironmental Research Laboratory develops new and improved tech-
nology and systems to prevent, treat, and manage wastewsrer and
solid and hazardous waste pollutant discharges from municipal
and community sources, to preserve and treat public drinKing
water supplies, and to minimize the adverse economic, social,
health, and aesthetic effects of pollution. This publication
is one of the products of that research and provides a most
vital communications link between the researcher and the u s e ^
commu n i ty.
This report addresses the evaluation of cooling as a
mechanism of reducing the release of toxic or flammable vapors
from a spilled chemical. The results show the capabilities of
a variety of coolants and demonstrate that dry ice is the best
choice for most situations.
Francis T. Mayo, Director
Municipal Environmental Research
La bora tory
-------
ABSTRACT
Spilled chemicals that pose a hazard to the land and water
ecosystem can also provide a significant vapor hazard. Although
the vapors released by such chemicals may ultimately be dispersed
in the environment with little long-term effects, they do pose a
hazard to life and property downwind of the spill.
Among the vapor amelioration techniques that have been con-
sidered is the use of a coolant to lower the temperature of a
spill and reduce its equilibrium vapor pressure. This program
has been conducted as a feasibility study of that mechanism.
Four potential coolants were examined: water ice, dry ice,
liquid carbon dioxide, and liquid nitrogen. Further evaluation
based on laboratory studies and limited scaled-up tests estab-
lished dry ice as the most versatile coolant choice.
Water ice does not cool sufficiently. Liquid nitrogen and
carbon dioxide require large quantities of material, and produce
a dense obscuring cloud that has some adverse implications.
Dry ice avoids the problems presented by the other coolants
and is readily available at a reasonable cost, but a method is
required for crushing and distributing the dry ice to the spill.
A prototype unit was developed consisting of a crusher and a
pneumatic conveyor to perform these functions.
A pool of diethyl ether with 2.23 m2 (250 sq ft) of surface
was cooled to -60°C (-76°F) using 408 kg (900 Ib) of dry ice eed
at a rate of 13.6 kg (30 Ib) per minute. A measurable reduction
in downwind vapor concentration was realized. Pool temperature
was still below -10°C (14°F) two hours after dry ice discharge
was terminated.
This program has established feasibility of the mechanism,
but additional work is necessary to establish practicality, de-
fine materials to which cooling is applicable, and optimize the
dispensing equipment.
This report was submitted in partial fulfillment of Contract
No. 68-03-2648 by MSA Research Corporation under the sponsorship
of the U.S. Environmental Protection Agency. This report covers
the period 29 October 1976 to 4 December 1980, and work was com-
pleted as of 15 May 1981.
i v
-------
CONTENTS
Foreword 1 ]"'
Abstract 1V
Figures Y ]
Tables vl.l
Abbreviations and Symbols v i 11
Acknowledgment xi
1. Introduction 1
2. Summary and Conclusions 3
3. Recommendations 6
4. Program Plan 9
Phase I - Coolant Evaluation and Selection 9
Phase II - Equipment Development 10
Phase III - Field Demonstration 11
5. Coolant Identification, Evaluation and Selection . . . 12
Identification and Selection of Coolants 12
Evaluation of Coolant System Characteristics
f o r S p i 1 1 Application : 9
Laboratory Studies for Data Clarification 24
Selection of Candidate Coolants 43
6. Scaled Up Testing of Selected Coolants 46
Scaled Up Test Design 46
Coolant Tests 47
CoolantSelection 51
7. Equipment Development 54
Review and Evaluation of Commercial Equipment. . . . 54
Integrate and Test Prototype 56
EquipmentRedesign 60
8. Field Demonstration 67
Demonstration Test Design 67
Field Test Operations 57
Spill SceneCleanUp 70
Evaluation of Cooling as a Mechanism of
Vapor Hazard 71
References 73
-------
FIGURES
Number Page
1 Factors influencing the temperature of an
accidentally spilled material 20
2 Laboratory test apparatus 25
3 Candidate coolants added to 105 g glacial
aceticacid 27
4 Time-temperature changes observed when carbon
dioxide was placed on representative
hazardous materials 23
5 Time-temperature changes observed when liquefied
nitrogen was placed on representative hazardous
materials 29
6 Candidate coolants added to acetone 3 C
7 Subscale field test results with ethyl ether -
temperaturevs time 48
8 Crusher/snow blower combination 58
9 Diagram of spill impoundment area 59
10 Field demonstration unit 65
11 Dry ice distributor. . . 66
12 Cooling rate vs time for dry ice treated
ethyl ether pool 69
-------
TABLES
Number Page
1 Candidate Cooling Materials and Their Properties
to be Used for Evaluation 14
2 Hazardous Materials with Acceptable Thermal
Properties 16
3 Representative Hazardous Materials 23
4 Results of Screen ing Tests 32
5 Vapor Concentrations Measured Over a 250 cm2
Pool of Acetone 34
6 Comparison of Measured and Calculated Evaporation
Rates Using Dry Ice (T=-78°C [-110°F]) 36
7 Comparison of Evaporation Rates for Repre-
sentative Hazardous Materials 37
8 Coolant Storage Losses 39
9 Selected Properties of Candidate Coolants 40
10 Spill Coolant System Comparisons 44
11 Cryogen Comparisons 52
12 Shredder/Grinder (W.W. Grinder Co.,Model 4-G). ... 55
13 Diethyl Ether Vapor Concentration Before and
After Application of Dry Ice 61
VT i
-------
ABBREVIATIONS AND SYMBOLS
A -- experimentally fitted constant (not dimensionless)
B.P. -- boiling point
BTU -- british thermal unit
°C -- degrees centigrade
c -- concentration
ca 1 -- calories
cm --centimeter
D -- diffusion coefficient
E -- evaporation rate
F -- film thickness
°F -- degrees Fahrenheit
ft -- feet
g - - gram
hr -- hour
Hg -- mercury
HP -- horsepower
in. - - inch
Real - - kilocalories
Km - - kilometer
«m -- mass transfer coefficient (equal to D/RT)
KW -- kilowatts
kg - - kilogram
v i i i
-------
k p h • -- kilometers per hour
L, 9. -- liquid phase
LOX -- liquid oxygen
1 b -- pounds
M -- molecular weight of diffusing species
M.P. -- melting point
m -- meter
mm -- millimeter (millimeters of mercury or torr when used
to designate pressure)
mi - - mile
min -- minute
ml -- milliliter
mph -- miles per hour
oz -- ounce
2
N -- newtons (1 newton - 1 kg m/sec )
P -- atmospheric pressure
P -- vapor pressure
v a p
p', p1' -- vapor pressures of diffusing species at liquid s^r-
face and far removed
p p m -- parts per million
q -- amount evaporated
reefer -- refrigeration
R --gas constant
rpm -- revolutions per minute
S, s -- solid phase, S also used as area of spill
T -- temperature
TLV -- threshold limit value
-------
t
torr
x
t i me
unit of pressure (1 torr - 1 mm Hg
d i stance
Greek Letters
radius of surface of simulated spill
-------
ACKNOWLEDGMENT
This work was performed under subcontract to the Environ-
mental Monitoring & Services Center of Rockwell International,
Newbury Park, California. The authors wish to thank George R.
Schneider, Project Officer, Environmental Monitoring & Services
Center of Rockwell International, and John E. Brugger, Project
Officer, U.S. Environmental Protection Agency, for their direc-
tion and support.
-------
SECTION 1
INTRODUCTION
Spilled chemicals that pose a hazard to the land and water
ecosystem can also provide a significant vapor hazard. Although
the vapors released by such chemicals may ultimately be dis-
persed in the environment and, because of degradation or dilution
have few long-term effects, they do pose an immediate hazard to
life and property downwind of the spill. In addition, a hazard
is posed to those responding to the spill who must remain in the
area for the duration of the incident.
The vapor hazard from spilled chemicals takes two forms:
(1) the release of toxic fumes that pose a life hazard even at
low concentrations (parts per million), and (2) the release of
flammable vapors where minimum dangerous concentrations are
usually above one percent. Some chemicals may exhibit both
hazards, but toxicity, with its lower allowable concentration,
will be the controlling feature.
The great difference in minimum hazard levels creates two
distinct problems. In the case of flammable vapors, small in-
crements of reduction can be meaningful. For toxic materials,
the ability to provide meaningful mitigation of the hazard may
only lie with reduction of the equilibrium vapor pressure.
T I
"here are a number of possible vapor amelioration techniques
which have received consideration. A review of techniques con-
ducted under U.S. Coast Guard sponsorship (Greer, 1976) shows
that a number of mechanisms are impractical or ineffective for
field situations. The techniques of mechanical covers, induced
air movement, vapor scrubbing and vapor phase reaction are in
these categories. The techniques of foam cover and liquid phase
modification are the only methods which in their present state
of development have had any practical demonstration and can pass
the criteria of cost, availability, deployment, and application.
The degree of vapor control that can be experienced has been
shown to be sufficient to be beneficial where flammable vapors
are the hazard (Gross, 1980). Where vapor concentration from a
spill can be expected to be much greater than the TLV, signifi-
cant reduction of the toxic hazard does not appear possible
within existing technology.
-------
A promising mechanism of va,jor control is the use of a
coolant to reduce the temperature of the spill. This method
would reduce the equilibrium vapor pressure and the rate of
vapor release per unit time.
This technique has been addressed in the prior referenced
Coast Guard program (Greer, 1976) and also in a report to the
USEPA (Brown, 1980). In neither study was any systematic in-
vestigation conducted. On the basis of these studies, EPA
inaugurated a detailed program to evaluate the potential of
cooling and if warranted to conduct a simulated spill scenario
to define feasibility. This report presents a comprehensive
documentation of the work that was accomplished, the conclusions
reached, and recommendations for further validation of the pro-
cedure including additional equipment evaluation.
-------
SECTION 2
SUMMARY AND CONCLUSIONS
This program has established the basic feasibility of the
cooling concept. A literature review and data evaluation pro-
vided an extensive list of potential coolants. Practical con-
siderations of cost, availability, safety and field handling
reduced the candidate list to four materials: wet ice (solid
water), dry ice (solid C02), liquid C02 and liquid nitrogen.
The application to spills of N2 and C02 in their liquid
form can produce significant reductions in the temperature of
the spill but their use is accompanied by certain disadvantages.
Both liquid C02 and liquid nitrogen produce a dense obscuring
cloud above the spill surface. While this has the benefit of
providing a nonflammable atmosphere above the spill, it has the
disadvantage of being nonbreathable. Effective vapor phase con-
trol requires continued application of the coolant. The boiling
of the liquid coolant can exaggerate the vapor release from tne
spilled material. There is a tendency of liquid C02 to sink
into the spilled material, restricting the amount of cooling at
the surface.
Liquid C02 can be converted to solid C02 by expansion of
the liquid to atmospheric pressure. Conversion efficiency is
only 15%, however, and the generation of the obscuring non-
breathable cloud still occurs.
Wet ice has certain advantages but its cooling ability is
restricted. The minimum achievable temperature is only 0°C
(32°F) unless the ice is super cooled prior to use. Ice may
react with some materials and will increase the volume of the
spill.
Laboratory tests showed dry ice to have the greatest po-
tential for effective, persistent cooling with small material
losses and minimal cloud formation. The best results were ob-
tained where the dry ice was crushed and applied as an aggregate
of particulates.
The standard commercial form of dry ice is a 25 cm (10 in.)
square block. Large scale field application of dry ice requires
a mechanism to crush the blocks and apply the particuiates to
the spill surface. After survey of commercial equipment and a
-------
limited testing program a commercial shredder/crusher was se-
lected to convert the dry ice blocks to a suitable form. The
speed of rotation and the configuration of the tines were
altered to achieve a reasonable yield of particles within an
acceptable size range. The size distribution ranged from fine
participates to coarse material of 0.635-1.27 cm (1/4 to 1/2 in.)
size. Because of sublimation losses which averaged 15 percent,
a true size range was never measured.
Several methods were evaluated for dispensing the crushed
dry ice onto the spill surface. A snow blower was originally
selected. Its operation reduced the effective dry ice discharge
to 65 percent of that fed to the crusher. The combination of
snow blower/crusher was not an satisfactory application device.
The dry ice discharge distance was insufficient to permit ope-
ration from a single location and the machine was difficult to
maneuver around a spill.
The crusher was then combined with a pneumatic conveyor
consisting of an auger, air blower, and a hose. The auger fed
the dry ice particles from the crusher into an air stream for
delivery to the spill via the hose. This apparatus was mounted
on a wheeled frame to provide mobility but the hose discharge
of the dry ice allowed an extended discharge pattern from a
single location. About 50 percent of the dry ice fed to the
crusher was effectively applied to a spill by the hose.
Field tests were conducted using diethyl ether as a haz-
ardous spill simulant. Seven hundred liters (200 gallons) were
spilled into a 7.6 by 6.1 meter (25 ft by 20 ft) impoundment.
Dry ice was' charged at a rate of 13.6 kg/min (30 Ib/min) for 30
minutes with an effective application rate to the spill of 6.8
kg/min (15 Ib/min). The spill temperature was reduced to about
-60°C (-76°F) in about 20 minutes. The data showed that the
diethyl ether vapor concentration in the vicinity of the spill
was reduced to 25 percent of the pre-treatment value. Due to
wind effects an absolute measure of the vapor concentration re-
duction could not be obtained.
This program has shown that dry ice can significantly re-
duce the temperature of a spilled liquid with a concomitant re-
duction in the spill vaporization rate. Crushing commercially
available dry ice blocks to small particles and distributing
these particles over a spill surface can be achieved with com-
mercially available equipment. The equipment assembled and
modified for this program requires some further optimization in
terms of operation, configuration and materials of construction.
The tests conducted were not sufficiently extensive to clearly
show practical, efficient operation in a real-time spill scenario,
but they do support continued investigation and further eva"!uatior
of the cooling concept.
-------
It is reasonably clear that the cooling concept is primarily
for use on spills that pose a toxic vapor hazard rather than a
flammable/explosive vapor hazard. It has been well substantiated
that aqueous foams provide effective mitigation for those ma-
ter ial's presenting a flammable hazard. The use of fire-fignting
foams is a well developed technology in common use by emergency
organizations. Foam cannot provide the degree of vapor control
necessary where toxic levels are in tne parts per million range,
however. This provides a basic guideline for further evaluation.
-------
SECTION 3
RECOMMENDATIONS
The continuing evolution of the cooling concept for vapor
suppression of spilled volatile chemicals rests primarily on its
practicality in real-time spill situations. Although additional
testing with a broader range of chemical classes may be desirable
along with a more accurate appraisal of the degree of benefit
provided, the first question to be asked after acknowledging
feasibility concerns logistic practicality and the first recom-
nvendation is to address that question. The first approach is to
review actual spill scenarios to determine if the cooling con-
cept could have been employed. This should be done with know-
ledge of equipment and material needs, the realities of deploy-
ment, and the limitations of maneuverability.
The availability of a piece of equipment, even though
limited, allows actual response to a real spill situations. By
assigning the current equipment to a spill response operation, a
regional emergency response unit of EPA or a contractor response
unit such as EERU, it would be available for deployment in a
spill situation. As an alternate to that, the assigned unit
could conduct mock up spill response based on real-time situations
to determine practicality. Such evaluating procedures would be
conducted concurrently with the theoretical analysis.
Should the analysis of the practical aspects of the cooling
concept be positive, the next recommendation is for equipment
modification. The unit developed in this program has a practical
limit of 13.6 kg (30 Ib) per minute dry ice feed. This limit is
imposed basically by the size of the auger and the configuration
of the hopper. More rapid feed through the crusher causes ma-
terial buildup in the hopper which results in bridging over the
auger. This blocks further material feed into the auger.
Some mechanical improvement appears possible by changes in
the size of the hopper, the slope of the hopper walls and the
capacity of the auger. There would appear to be limitations on
the size increase the equipment can experience without affecting
ease of deployment and maneuverability. Thus consideration needs
to be given also to devices which prevent bridging such as vi-
brators, moveable vanes or similar devices mounted within the
hopper.
-------
A second area for mechanical improvement is in the configu-
ration of the crusher. The current unit uses a hammermill design.
It produces a large number of fines which tend to fully sublime
in the discharge hose. A reduction in fines would increase the
effective discharge percentage, while an increase in the number
of coarse particles would assist in the uniformity of the dis-
charge pattern and the projection distance. The potential bene-
fits would appear to justify an examination of different arrange-
ments or spacing of tines or the substitution of spikes in place
of the tines.
Certain material changes have been indicated. It is certain
that a review should be made of hose materials which remain
fairly flexible at the temperatures experienced in the discharge
line. Low friction materials may be a benefit on the sides of
the hopper.
The final recommendation concerns the applicability of the
concept to chemical classes. There is the temptation to recom-
mend more extensive testing to establish in quantitative terms
the degree of vapor control that could be achieved. There is a
question as to how successful such efforts would be.
The rate of vapor release from a spill is dependent on its
equilibrium vapor pressure. The amount of vapor released is in-
fluenced by time, the size of the spill surface, the wind ve-
locity and to a lesser extent by the depth of the spill and the
topography of the area. Since each spill scenario is different,
the most meaningful data is the degree of temperature reduction
that can be achieved. This will be influenced by the ambient
temperature and probably by the depth of the spill and the wind.
Thus it would seem that additional evaluation should first
look to the possibility that the application of the dry ice re-
sults in a reasonably constant temperature reduction. If this
is the case it also allows the determination of which materials
can be converted to solids.
Regardless of whether constant temperature reduction is
likely, it seems questionable whether effort should be expended
to attempt to interrelate spill size, ambient temperature and
wind with the degree of reduction which can be effected in the
downwind vapor concentration by applying dry ice to the spill.
At this point in time such a recommendation does not appear
justified.
Consideration should be given to classes of materials which
respond differently to temperature than the simulant material.
Three classes warrant consideration: 1) those which would solidify,
2) those whose boiling point is below ambient, and 3) those which
fume in moist air.
-------
Based upon thermodynamic theory it would be expected that
the temperature of materials solidified by the application of a
coolant would not be significantly reduced below their freezing
point. The combination of the heat of fusion and the poor heat
transfer expected between the subliming dry ice and the solidi-
fied spill would appear to preclude measurable temperature re-
ductions below the freezing point. This assumption needs to be
validated, however.
With liquefied gases the question is how significant will
the heat of vaporization be in affecting the degree of cooling
that can be achieved. Materials such as chlorine and vinyl
chloride monomer would be likely candidates for study. An in-
vestigation of such materials is recommended.
Materials which react with water and are not amenable to
control by dilution with either water or foam may also be candi-
dates for vapor control by cooling. Typical materials are sulfur
trioxide and anhydrous hydrogen fluoride. This type of material
reacts strongly with water and direct application sorely aggra-
vates the vapor release. Foam is a better mechanism for dilution
but does not provide any short term benefits in vapor control.
Cooling with dry ice might have a benefit beyond simple
cooling. Because of its density it tends to displace the air
above the spill. The moisture is also displaced. Reduced
vaporization could be helpful if only in controlling visibility.
The opposite situation may occur, however. In the field
tests with diethyl ether, ice crystals appeared in the liquid.
It is probable that some of this resulted from fractional
crystallization of the water impurity in the ether. It is also
possible that some of it developed due to condensation of the
moisture in the air. If such condensation is significant, the
precipitation of water ice into reactive materials would work
against cooling due to the heat of reaction or solution. It is
recommended that some analysis be made of these possible com-
peting effects.
These recommendations were derived from the work to date.
With more knowledge, other recommendations may become obvious.
-------
SECTION 4
PROGRAM PLAN
The program had as its objectives the development of
mechanisms to modify the vapor hazard of spilled volatile chemi-
cals through reduction of the surface temperature of the spill.
The plan to implement that program had three phases -- the
evaluation and selection of coolants, the development of equip-
ment to dispense the coolant, and a field demonstration of the
fully developed system.
PHASE I - COOLANT EVALUATION AND SELECTION
The first phase covered six months of effort and was sub-
divided into three main tasks. To some extent the tasks over-
lapped each other rather than running consecutively.
The first task defined the available coolants. All po-
tential candidates were identified, their physical and chemical
properties compiled and refrigeration capabilities identified.
These values allow definition of theoretical capabilities of
each candidate. Other factors which influence the practical
capabilities including on-site generation, long-term storage
considerations, system costs, materials losses during appli-
cation, potential environmental effects and heat transfer
efficiencies under field conditions were identified and assessed.
This task encompassed an eight week period.
The second task evaluated each candidate using the data
acquired in the first task. The evaluation was conducted with
consideration of two additional factors, spill size and the type
of vapor hazard: flammable or toxic. Based upon existing know-
ledge best coolant candidates differed markedly as a function
of these two factors. This task covered six weeks, beginning
in the sixth week of the phase. It was expected that at the
end of task two there would be certain factors which could bene-
fit from further clarification. Environmental effects and heat
transfer efficiencies were anticipated to be in that category.
The third task entailed the conduct of limited laboratory
tests using representative volatile chemicals to provide the
desired clarification. These tests were conducted over a ten-
week period beginning in the tenth week of the phase using a
test procedure previously successfully employed in assessing
-------
other forms of vapor hazard mitigation.
Upon completion of the third task, specific selections were
made of coolants deemed most pro in ising. A report defining this
coolant and reasons for its selection was submitted to the con-
tracting officer. This report set forth the requirements for
effective application of the coolant. This included equipment
specifications and a specific and detailed program to develop
this equipment. This served as a detailed plan for Phase II and
supplemented this program plan.
PHASE II - EQUIPMENT DEVELOPMENT
The selection of the coolants and delineation of the re-
quirements of the equipment for application allowed the second
phase to be initiated. In this phase, conducted in five tasks
over a 9-month period, equipment to dispense the coolant and
apply it to the spill surface was developed.
The first task provided a review of commercially available
equipment to determine if the requirements could be met with
existing units. This review surveyed a broad range of equipment
designed for similar applications. This task encompassed a
period of 8 weeks starting in the 30th week of the program.
In the second task a specific equipment design was evolved
and the necessary equipment components acquired. Where possible,
commercially available components were employed. Modifications
or fabrication of nonstandard items were accomplished as required
to build the equipment to meet the preset requirements including
safe operation under real-time spill conditions. This task was
conducted over a period of 10 weeks starting in the 34th week of
the program.
As purchased or fabricated parts were received, equipment
fabrication was initiated. Assembly began in the 42nd week of
the program and continued for 18 weeks.
Upon completion of task three, the fabricated equipment was
operated under simulated conditions. Modifications and changes
were made as indicated during the shakedown operation. This work
comprised task four and was conducted over an 8-week period
starting in the 60th week of the program.
With the determination that the developed equipment met de-
sign specification, the final task was initiated. In this task
manuals were prepared for the operation, maintenance and trouble
shooting of the equipment. Suppliers and parts lists were pro-
vided for each individual component in the assembly. This task
covered 4 weeks beginning with the 68th week of the program.
10
-------
PHASE III - FIELD DEMONSTRATION
Upon completion of the equipment development phase a full
scale field demonstration was conducted using a simulated vola-
tile hazardous material. This phase conducted over 8 weeks com-
prised three tasks.
The first step was a detailed plan and procedures for con-
ducting the field demonstration. With approval of the plan, the
field test was set up and conducted. This test verified the oper-
ation of the equipment, the efficiency of the coolant material
and recommended mechanisms of spill removal. Full documentation
was made of the data gathered along with photographic documen-
tation of the test operations. These tasks took 6 weeks starting
in the 70th week of the program.
The final task conducted over the final four weeks of the
program was the preparation and submission of a comprehensive
technical report covering the total effort on the control. A
film documentation of the field demonstration in the form of
slides and a narrated 16 mm film was included with the final re-
port.
This submission concluded the program with the exception of
those changes which might be required in the format or contents
of the final report.
1 1
-------
SECTION 5
COOLANT IDENTIFICATION, EVALUATION AND SELECTION
The objective of this first phase of the program was the
selection of coolants that possessed the correct combination of
properties to cool and reduce the vaporization rate of hazardous
material spills.
IDENTIFICATION AND SELECTION OF COOLANTS
If cooling the spilled material is considered as a general
refrigeration process, then any substance capable of going
through the Joule-Thompson cooling cycle could be considered as
a candidate to be used in the proposed system. A listing of
materials meeting this criterion resulted in 60 candidate
coolants. The fact that the coolant system is being designed
for use as a response to spills of hazardous materials sets
safety as an initial consideration. Applying the criterion that
the coolant should not contribute a health or safety hazard at
any time during the spill response, the removal or ultimate dis-
posal of spilled substances reduced this to 50 materials. Six-
teen others were rejected for logistic reasons, which included
available in quantity, transportaole, and ability to handle with
a minimum of training.
Physical and Chemical Properties
For the 34 substances identified in the prior task, a list
of properties were collected as given below:
Boiling point Critical properties
Freezing point Threshold limit values
Heat of fusion Liquid density
Heat of vaporization Vapor density
Flammability limits Cost
The sources used in collecting these data included the
Handbook of Chemistry and Physi(^s_ (1978), Bureau of Mines B u 1 1 e -
tin No. 503, National Fire Protection Association Fi^re Cp_d_es ,
Vol. 13, No. 325M, the 1978 Edition of Threshold Limit Values
fo_r Chemical Substances and Physical Agents in the Workroom En -
v i r o n m e n t, technical information circulars supplied by manu-
facturers and technical papers in which properties for individual
materials were reported. Costs were calculated on the basis of
12
-------
prices quoted for ton quantities of the cooling materia"'. These
data were compiled into the format shown in Table 1 to facili-
tate correlation and evaluation.
Preliminary Evaluation
A screening process was initiated to select the most pro-
mising coolants from among the 34 candidates. Using the cri-
teria of cooling capability, availability, safety in handling
during both application and cleanup, and the environmental
acceptability, the candidates were given a merit rating.
Three other factors were acknowledged as important. These
were application efficiency, the ratio of coolant used versus
the amount which reaches the spill surface; logistics, the de-
ployment of equipment and materials to the spill site and sub-
strate heat loads, heat inputs from the ground or containment
structure. Because they were difficult to measure, they could
only be assessed in general during this preliminary evaluation.
Materials costs were calculated and have been shown in Table
1. They were not a consideration in the initial screening, how-
ever.
Cooling Capabilities --
Cooling capacity can be calculated as the combination of
the available sensible heat of the coolant and the latent heat
resulting from any change of state that may occur. These values
are provided as part of the coolant properties given in Table 1.
The available cooling capacity basically determines the
volume of coolant required in any spill situation. The true
effectiveness of a coolant is the ultimate temperature which can
be obtained by its use. These two factors, cooling capacity and
minimum achievable temperature, determine the basic capability
of any material to be an effective coolant.
To the above factors must be added application efficiencies.
During handling and delivery of coolants to the spill surface
certain losses will occur. The greater these losses are, the
poorer will be the total cooling capacity. The efficiencies
cannot be accurately assessed without some experimental data.
Availability and Storage Characteristics --
Availability is defined in terms of ready supply in large
quantities through the U.S. on fairly short notice. It was
clear that many of the materials in the list of candidates could
not meet this requirement.
As an alternate to availabiity, the potential of storing
13
-------
OO
UJ
a_
O
o
i—
u~> CC C Cj
Oxj CC UD
) O cr o
. — '-) ,C
*"-* O ijCi Osj "XT LT) UD ^-^ CT'
— — 0 0
CT- r^. r~~. r\! CD O~i LO •
OJ ~ rsj f\}
CO Oo •— LJ
•O-L," o: ;r 3: c
o
~" <*c* ro oo O
2: u_ o O f%.
cc exj cc
QJ OJ .
QJ c -c T:
X: i£> O O
QJ a* r ~
• — C QJ
-
-c .£: c -c
— 'J OJ
14
-------
materials at strategic sites or including such storage within
the resources of spill response units was considered. In this
case long term storage in low loss, low maintenance facilities
was a necessary requirement.
Safety and Environmental Effects --
One of the major requirements set forth early in the pro-
gram was that the coolant should not contribute to the inherent
hazard of the spill. This requirement referred to the coolant
as any possible interaction of the coolant with the spilled ma-
terial or the environment, and to the ultimate disposal of the
spill contaminated debris or residues.
A number of chemicals, as listed in Table 2, were rejected
as coolants because of their inherently hazardous nature. Other
potential coolants, with safety or environmental deficiencies,
were not initially deleted since they were considered to be less
hazardous than the potential spill materials and might provide
benefits that would offset their disadvantages. A final de-
cision on these coolants required an in-depth review and com-
parison with other candidates.
Selection
Although the collection and compilation of data concerning
cooling materials was done on an individual basis, in the early
portion of the selection process it was clear that some screen-
ing could be made on the basis of chemical classification. These
broad classifications included: rare gases, oxygen-containing
cryogens, general organic materials, freons, commercial re-
frigerants and hydrocarbons.
Rare Gases --
The limited availability and cost of the rare gases are the
major problems associated with their use. Although logistical
problems may be encountered to varying degrees with all cooling
materials, the rare gases have the greatest problem and are not
considered convenient materials to be used for spill response.
Oxygen-containing Cryogens --
Liquid air and other cryogenic materials that contain
oxygen were rejected as coolants because of their potential re-
activity with many substances considered to be in the low-boiling,
hazardous materials category.
General Organic Materials --
Materials, such as alcohols and ethers, which can be con-
sidered within a general classification of organic materials,
15
-------
TABLE 2. HAZARDOUS MATERIALS WITH
ACCEPTABLE THERMAL PROPERTIES
Vinyl Chloride B.P. -13.9°C
Cyanogen Gas B.P. -20.5°F
Boron Trimethyl B.P. -20°C
D i s i 1 a n e B.P. -14 . 5 ° C
Methyl Phosphine B.P. -14°C
Methyl Amine B.P. -6.5°C
Bromo Amine M.P. -7°C
1,2,3,4 Tetramethyl Benzene M.P. -6.25CC
Hydrogen Cyanide M.P. -13.4°C
Methyl Nitrite M.P. -17°C
1 6
-------
were rejected as contributing to health and safety hazards to-
wards response personnel and the general public. Pure hydro-
carbons, such as the alkanes, were tentatively rejected as
coolants because of their fla mm ability. The tentative status
was used since the hydrocarbons could be applicable if the use
of a hydrocarbon coolant did not add to the hazard already pre-
sent because of the spilled material. Because less hazardous
candidate materials were found that possess a cooling capability
equal to or better than that provided by hydrocarbons, the hydro-
carbons were ultimately rejected.
Freons --
Freons (chlorofluorocarbons) were initially set aside in
case no viable. candidate coolants could be defined. Ultimately,
the freons were rejected because of present concerns and liti-
gations surrounding their impact upon the environment. The pro-
duction of these chemicals may be limited and open end uses,
such as would be envisioned for spill response, have been cur-
tai 1 ed .
Commercial Refrigerants --
Commercial ref r i g.eran ts , such as ammonia, methyl chloride
and sulfur dioxide have been rejected as health and safety
hazards. Their use as spill coolants could be considered for
enclosed areas, but probably would require measures to reclaim
or otherwise recover the coolant for acceptable disposal.
Miscellaneous Coolants --
Slurry mixtures of carbon dioxide particles in liquid
nitrogen or argon and some procedures for producing them have
been patented by Smith and Townsend (1968). The claims made for
these slurries imply that they could be considered alternatives
for this investigation. Action on such mixtures was deferred in
this program since it would require working with patented ma-
terials and procedures.
Sulfur hexafluoride possesses favorable ther mo physical
properties. Although it has an identifiable toxic i ty , the TLV
is high at 1000 ppm. After detailed consideration it was de-
cided that the use of large quantities in a spill situation
could result in unsafe downwind concentrations of the hexa-
fluoride. This, coupled with uncertainties of availability and
cost, resulted in rejection of sulfur hexafluoride from con-
sideration.
Refrigeration Systems --
In addition to specific coolant materials, consideration
was given to two approaches which fall in the category of re-
17
-------
frigeration systems. Water or ammonia absorption systems and
steam-jet refrigeration units with large cooling capacities are
available which provide safe, dependable cooling. They do not
present environmental problems, but they are large and can re-
quire special services such as high pressure steam. Because of
logistic problems, size, weight, services, etc., such systems
for field operations were deemed impractical.
Acceptable Coolants
The consideration of toxicity, f1ammabi1ity , cooling cap-
acity and availability already discussed in this section re-
sulted in the elimination of most of the materials appearing in
Table 1. Four candidate cooling materials remained for further
evaluation: liquid nitrogen, liquid carbon dioxide, solid carbon
dioxide and water.
Wet Ice --
Water is the most innocuous of all the materials considered
It should present neither health nor safety hazards as a result
of its use in quantities that would cool a typical spill. Water
could be used in either' the liquid or solid state, but a sub-
cooled solid form would be the most effective means for cooling
a spill.
There are three possible problem area which can arise with
the use of water as the cooling material: 1) the total liquid
volume of the spill would be substantially increased, possibly
causing excessive spread or overflow of containment devices,
2) the perceptible heat load could be increased by the heat
capacity of the water, and 3) exothermic reaction between the
water and the spilled material.
These problems could restrict the use of water ice as a
coolant on some spills. Many hazardous materials are miscible
with water, however. Dilution with water could offset the
stated difficulties. Assessment of these effects would require
experimental measurements.
Carbon Dioxide --
Carbon dioxide has acceptable thermophysical properties
for the proposed application, and has the additional benefit of
being a refrigerant. It could be used in either the solid or
liquid state. Storage as a liquid permits the use of its re-
frigerant properties to minimize storage losses anc allow long
term storage at remote locations. Solid C02 is commercially
available, however, in a multiplicity of locations within the
U.S.
The application of large volumes of carbon dioxide to a
18
-------
spill will result in the generat on of large volumes of CO2 9as
over the spill. This can have both desirable and undesirable
effects. The gas would remain close to the surface to condense
the vapors of the hazardous material and would inhibit the ig-
nition and burning of the spill vapors, if they were flammable.
On the other hand the carbon dioxide vapor presents a problem
for response personnel since its density could result in the for-
mation of an obscuring and nonrespirable gas mixture in the spill
vicinity. This would require the monitoring of ambient oxygen
concentrations.
Liquid Nitrogen --
Liquid nitrogen has the potential to cool most low-boiling,
hazardous materials below their freezing points which could
facilitate their separation and removal from the ground.
Liquefied nitrogen has the unique characteristic among the
candidate coolants selected for responding to accidental spills
of being less dense than many of the low-boiling, hazardous ma-
terials. It would appear possible, therefore, to float this
coolant on the surface of many spills and take advantage of the
insulating properties of the hazardous materials to reduce heat
transfer to the ground beneath the spill. The total heat load
might thus be significantly reduced.
Storage and dispensing of liquefied nitrogen would present
some difficulties or require some specialized equipment, but
both equipment and experience are available from present com-
mercial manufacturers.
Liquefied nitrogen has the problem of creating a non-
respirable gas mixture and obscuring cloud in the spill vicinity.
Nitrogen is a simple asphyxiant and would require large volumes
and poor mixing to create health hazards, but monitoring ambient
oxygen concentrations would be necessary for safe operation.
EVALUATION OF COOLANT SYSTEM CHARACTERISTICS FOR SPILL
APPLICATION
An outline of the factors influencing the temperature of
accidentally spilled materials is presented in Figure 1. The
temperature of the spilled material, which is the primary con-
cern of this program, is shown to be directly related to its
environment, its own thermophysical properties, and the amount
spilled. The presence of the coolant and the ratio of the two
sets of thermophysical properties and amounts determine the rate
at which any temperature change occurs and the ultimate tempera-
ture which can be attained.
Factors affecting the temperature of a spilled material
have been divided into two categories: 1) those which are
1 9
-------
----- i-g
o
OJ
ra
4->
.a
13
GO c —
ra
-o T-
0) 1-
i — Ol
i- ro
a. s:
-13
>, 0)
rO -f
4-1 a.
ai !
•o i
<
ro
M— \
O
01
.3
l!
53 !
h- !
4J
C
OJ
1 £
>
LU
jj
C
o
,5
4-1
01
a.
o
a.
4-1
c
3
M.
Ol
4-1
t/1
01
4-1
e
o
3
l-O
cn
o
u
5
c
u
LU
r- O)
•i— ITJ
t/) P
-C
06
rO
<
0)
•^
•4-J
4-) i-
c Q_
s. a»
flj ^J
ra O
3: 0
•^- c
CLT-
i- ro
cn u
o o
Ol _!
^
o
in
ra
01
on
Ol
1 C
•i— C
ra O)
z: 4-j 01
CD
J
o
4-J
>, 00
^ -^
> .a
4-1 4-1
CO C
13 <_> O
4-)
C
cn
C 1—1
o h
4-1 cn
CO) —1
cn c: LU
>
t—
OJ
c
^
U1
i-
Ol
a.
2
j=
-a
u
.rt
D_
c:
s_
i_
t—
5
3
o. c
O O)
Q- a
D
ra
Z
4-1
,— 5
4J CL
ro CT
O LU
4-1
S_
a.
ij
QJ
S-
=3
-M
ro
S_ i —
ai ns
Q.-I-
E S-
OJ O)
OJ E
T3
cu
O CL
c: oo
cu
i/i ai
i- T3
o -r-
4-i O
U O
03 n3
20
-------
considered to be uncontrollable, and 2 ) those which are subject
to some degree of control. The factors which are considered un-
controllable in accidental spills of lew-boil ing, hazardous ma-
terials include the environmental factors such as season, wind,
rain, location, substrate characteristics and structure and the
spill itself, including: amount, rate of spillage, vapor pres-
sure, heat of vaporization and thermal conductivity.
Representative values were selected for these factors to
establish a range of conditions under which a proposed cryogenic
cooling system could be applicable.
Nine factors associated with the effectiveness of a coolant
are included among those which are considered controllable. They
play an integral part in selecting a cooling material. These
factors include the amount of coolant, its rate and effective-
ness of dispersal, phase transition temperature, latent heat of
phase change, effective rate of heat transfer, liquid or vapor
density and solubility or miscibility with the spilled material.
Heat Transfer Assessment
Two separate approaches may be used to define heat transfer
within an accidental spill treated with a cryogen. The rate can
be estimated by theoretical approximations, or it can be measured
under simulated conditions. Theoretical approximations may be
made with standard heat transfer equations, or more refined
estimates can be made from models, such as described by Bell
(1978), which require computer calculations.
Basic heat transfer calculations yield heat transfer rates
of 3300 Kcal/m2hr (1230 BTU/ft2hr) during the initial cooling,
and 950 Kcal/m2hr (350 BTU/ft2hr) to maintain the cooled con-
dition.
The alternative approach for defining a typical heat load
is a pragmatic evaluation of experimental data published con-
cerning spills of cryogens. These data have been reported by
Drake and Reid (1975), Kosky and Lyon (1968) and Board, et al
(1971). The data reported for boiling rates have been measured
in various units of heat transfer, but generally fall within the
range of 2,000 to 11,000 Kcal/m2hr (740 to 4060 BTU/ft2hr).
The work published by Drake and Reid (1975) is a synopsis
of experimental data obtained from spills of a liquefied natural
gas on various substrates. These data are fitted to the equation:
q = A t1/2
in which q is the heat transferred, t is the time from initial
spillage in hours and A is an experimentally fitted constant.
Values of A, obtained for typical substrates vary considerably,
21
-------
but generally fall in the range cited above.
This extensive range of heat transfer rates indicates the
variability which would occur under ambient conditions. Since
the objective of this program is to establish the working
parameters for candidate systems to treat accidental spills,
the task of selecting a representative heat transfer rate in-
cluded the selection of realistic conditions while retaining a
reasonable safety factor.
The approach used an analysis of experimental data with
conservative estimates where data were unavailable. A heat
transfer rate of 6700 Kcal/m2hr (2460 BTU/ft2hr) was selected
to apply during the first half hour and a rate of approximately
2000 Kcal/rn2hr (740 BTU/ft2hr) to maintain a steady state con-
dition, for the remaining three and one-half hours of the pro-
posed spill response operation. These are clearly and purposely
higher than the rates obtained from the calculations noted above
These were selected to insure a conservative situation.
The assessment of candidate cryogens, as cooling materials,
was simplified by using two assumptions: 1) the average specific
heat of the spilled hazardous material is 0.5 cal/g°C and 2)
the cooling capacity of candidate cryogens will be approximated
by the latent heat of the phase change with subsequent corre-
lations of these cooling capacities by converting thermal data
into cost.
Estimation of Coolant Potential
Estimations of the potential of coolants on spills of
hazardous materials were made by comparing equilibrium vapor
pressures at two temperatures, 0°C and 25°C (32°F and 77°F).
The purpose of this exercise was to determine the degree of
benefits which would accrue by cooling to 0°C (32°F) and con-
versely to assess the need for lower temperatures for meaning-
ful mitigation of the vapor hazard of spilled materials. These
comparisons provide a common basis from which to make general
comparisons of the various cooling materials but are insufficient
for estimating efficiencies or calculating vaporization rate re-
ductions.
Representative hazardous materials were selected for this
evaluation of coolants. These materials and the properties
needed for estimating coolant effectiveness are presented in
Table 3. Effectiveness must reference defineable end points.
The f1ammabi1ity limits and TLV of the hazardous materials were
selected as the reference points. Lower flammability limits
would average about 1 percent vapor concentration, which would
correspond to an equilibrium vapor pressure of approximately 7
torr. Threshold limit values are in the range of 10 to 100 ppm,
which corresponds to equilibrium vapor pressures in the range of
22
-------
0.01 to 0.1 torr. Assessment of tt,e relative effectiveness of
coolants is also influenced by the average vapor concentrations
which would cause asphyxiation. Asphyxiation would require a
coolant vapor concentration greater than 12 percent, which would
correspond to a vapor pressure of approximately 90 torr.
The vapor pressure comparisons shown in Table 3 show that
cooling to 0°C (32°F) can reduce the vapor levels of most ma-
terials below their lower flammable limit. There are other more
universally available techniques, foam blankets and direct water
dilution, which are equally effective in mitigating the flammable
vapor hazard.
For cooling to be a useful mechanism of vapor control, with
the anticipated difficulties of logistics and field operation
along with cost, it must provide benefits beyond those of pre-
sently available techniques. It would be desirable to reduce
vapor concentrations below the TLV but this may not be possible
for all materials under any set of conditions.
Based upon the assessment of the coolants system require-
ments, it was decided that 0°C (32°F) was the maximum tempera-
ture which could be considered. The preference was for a coolant
system which could produce subzero temperatures at the surface
of a spilled material. This did not result in the elimination
of water ice from consideration but it prejudiced the selection
towards liquid nitrogen and the two forms of carbon dioxide.
LABORATORY STUDIES FOR DATA CLARIFICATION
To complete the process of selecting coolant candidates,
experimental data were required. The experimental investigations
undertaken had two objectives:
1) to observe the characteristics and
results of placing candidate coolants
on representative, hazardous materials
2) to provide information on the relative
effectiveness of the candidate coolants.
These investigations included the measurement of cooling and
heating rates, minimum temperatures, vapor concentrations, and
evaporation rates. The materials selected as hazardous spill
simulants were readily available, and were considered representa-
tive of compounds which might be encountered in a typical spill
response situation.
Determination of Achievable Temperatures
The first series of tests was done in the apparatus shown
in Figure 2a to observe the behavior of the candidate coolants
24
-------
Optional n
Thermometer h
II
1
Stainless
Steel
Screens
Sifter
Filler
Hazardous Material
Stainless Steel Dish
a. Time-Temperature
Glass Ampoule
Container
Hazardous
Material
JL
b. Vapor Concentration
Exhaust Hood
Pyrex
Dish
Balance
Hazardous
Material
c. Evaporation Rates
Figure 2. Laboratory test apparatus
25
-------
and the representative hazardous spill materials when they were
placed in contact. One hundred milliliter (3.38 02) samples of
acetic acid, gasoline, methylene chloride, lacquer thinner,
naphtha, toluene and vinyl acetate were placed in the 10 cm
(3.94 in.) diameter, stainless steel dish. A 9.5 cm (3.74 in.)
diameter sifter filler was positioned about 7 cm (2.76 in.) above
the liquid surface to deliver a homogeneous layer of solid water
or carbon dioxide particles to the liquids in the dish. Liquid
nitrogen was poured over the surface from a weighed Dewar flask.
The amount of each coolant used in the tests was calculated
to remove approximately 1200 Kcal/m2 (135 BTU/ft2) over a 1/2
hour time span from the liquid samples in the stainless steel
dish. The total amount of coolant was added within the first
1/2 to 2 minutes of the test, however. Coolant loadings used
for the tests were about 1.5 g/cm2 (3.1 lb/ft2) for wet ice,
1 g/cm2 (2 lb/ft2) for solid carbon dioxide and 2.4 g/cm2 (4.9
lb/ft2) for liquefied nitrogen.
The temperature of the spill simulant was measured with a
low temperature thermometer placed at the rim of the stainless
s t e e 1 d i s h .
Minimum temperatures were achieved in approximately 5 to 8
minutes with wet ice (subcooled to -18°C [0°F]), in about 1/2
to 5 minutes with solid carbon dioxide, and in about 1 to 3
minutes with liquefied nitrogen. The observed minimum tempera-
tures of the spill simulant differed by an order of magnitude
depending on the coolant used, i.e., for wet ice -8 to -12°C
(18 to 10°F), for dry ice -50 to -75°C (-58 to -105°F) and for
liquefied nitrogen -150 to -180°C (-238 to -292°F). Attempts
to increase the cooling rate and/or reduce the minimum tempera-
ture obtained by using salt water ice rather than subcooled ice
were not beneficial. Typical time-temperature curves are shown
in Figures 3,4,5 and 6.
The cooling rates observed in these tests indicate that a
suitable performance level can be attained if the coolant can be
dispersed evenly over the spill area. Heat transfer rates, calc-
ulated from the amount of materials used and the observed tempera-
ture changes, were in the range of 500 to 4500 Kcal/m2hr (180
to 1660 BTU/ft2hr). These rates are in the lower part of the
theoretically calculated range, but were measured on small scale,
experimental test units.
The minimum temperatures obtained with subcooled, wet ice
were about -12°C (10°F). This would be adequate for reducing
the flammability of such materials as isopropyl ether, methyl
ethyl ketone and some lacquer thinners, that have flash points
exceeding -12°C (10°F). This minimum temperature would not be
maintained throughout the four-hour response operation and would
be inadequate to reduce the flammability of materials such as
26
-------
D
3-
C
E
CD"
P
03
o
4-J
CL>
Cn
LO
O
T3
O)
T3
o
C
u
QJ
-i-l
03
-a
-a
c
ru
S-
Z)
CD
27
-------
o
(D
CD
Q.
E
CO
60
Time, minutes
Figure 4. Time-temperature changes observed when carbon dioxide
was placed on representative hazardous materials.
28
-------
o
o
CD
13
03
i—
CD
d
E
CD
-125 -
-150 -
-175
-200
30
Time, minutes
Figure 5. Time-temperature changes observed when liquefied
nitrogen placed on representative hazardous materials.
29
-------
oo
0)
D
c
E
CD
E
o
4-J
OJ
o
rd
-a
OJ
-a
T3
O
O
U
O)
-o
T3
UD
GJ
S_
00 '
30
-------
wou
acetaldehyde , acetone, isopropylamine, ethyl formate or gaso-
1 i n e s .
The minimum temperatures obtained with dry ice were near
its sublimation temperature of -78°C (- 10 9 ° F). This minimum
temperature would be adequate for reducing the f1ammabi1ity of
most materials (gasoline -41°C [-41°F]), but would have to be
considered a diluent for liquefied fuel gases.
The minimum temperature obtainable with liquefied nitrogen
Id provide the most protection against ignition and would
approach the TLV for many hazardous materials. Some uncertainty
remains regarding the reduction of the spill vapor concentration
during the initial cooling phase. Soon after the addition of
coolant to the spill simulant, the smell of the vapor from the
spill was noted to intensify. This may be attributed to the en-
train ment of droplets of spill simulant in the vigorously bubbl-
ing coolant gases produced by the rapid phase-change caused by
the initially large spill-coolant temperature difference. The
general test results are compiled in Table 4.
Determination of Achievable Vapor Concentration
A series of tests was conducted with the apparatus shown in
Figure 2b, to measure the influence of candidate coolants on
vapor concentrations over a simulated spill. Acetone was se-
lected as the representative soil! material, because it has an
appreciable vapor pressure (175 torr at 20°C [ 6 8 ° F ] ) , a relatively
high threshold limit value (1000 ppm) and is a commonly trans-
ported solvent. The tests were beset by a number of problems;
although every effort was made to control the test variables.
The anticipated results were intended to provide indications
of the relative effectiveness of candidate coolants for suppres-
sing evaporation and to provide some measure of the reduction in
the size of the area that would be threatened by the dispersion of
a hazardous vapor cloud from the spill. Hhile vapor concentra-
tions were consistently reduced by the addition of the coolants,
the data were incapable of quantitative interpretation.
There are three possible sources of error in these experi-
ments: 1) substrate heat content, 2) variable air velocity over,
the spill, and 3) miscellaneous equipment problems. Steady
state temperatures of the spills indicated that the substrates
used in these tests did not permit adequate modeling of the heat
losses occurring from an open spill. Air velocities were
measured in. the exhaust hood. The air velocity varied between
0.5 and 1 Km/hr (0.3 to 0.63 mi/hr). This may have influenced
the observed vapor concentrations.
31
-------
oo
1
OO
LLJ
'
C3
S!
i i
._
•L:—
LU
LU
LY!
( )
L/ J
Lu
" — '
OO
1 —
1
ZID
00
LU
ct:
LU
1
CQ
^_
U
00
O
O
i
-tj
c
•
e>
c_>
OJ ' —
•2 •—
4-> i-
Tl D
4-J 4-J
C TJ
D 3:
on
U • —
<— < —
CL- —
>^
cn p —
c: c;
•r- U
NJ ^>
OJ U
4-> Ol
OO i- 00
•_ u_ 01
°*- ' — U
TJ 0)
Ti 4-J U—
i_
C Tl -
III Q- 3
> O
1) " r
TJ oo
oo oj
4-J E OJ
T3 1- !_
O O Tl
<4- TO
-a =i
aj ci o
0 Q —
• -- . — o
u
0) i-
-C I- O
4-> O CL
CL TJ
O; — T) I>
o >
TJ
OJ - C
— ^T) TJ
0 c
<_/! .,_ ^
-C O
- I/I -f—
en TJ •*->
.5^5
Jd LO -r—
00 S_
TJ U O
r— - Z^ i~j
CX C TJ
uo tyi >
„
oo
u
"^
x — .
o . —
— -TU d
t/l U
'>- O O
U -Q i-
na Tl ••-
5UC
~o
• .—
u
"0
LJ
•r—
4-J
D
OJ
f-
=
TJ
OJ
4-J
d
4-J OJ
c s-
oj TI
E CL
c a.
Tl
s_ ij
c o
o
-L-> 4-J ^^
4-J CI t '
0 U -f-
tl
O f~i "O
4-J a. a
TJ O
•u u o
O E oo
cn o '— U
'/I O 'M
OJ Q- U
o -cr TJ u
•r- 4-J > i_
•r— ' )
00 3 TJ
•a c ~a
en TI -1—
en ii e)
c: •-- c: er
•r- -C O ••-
r~ 4_) -r^ i
00 O 4_)
T! U T3 U-
' — o*- rsl O
CL •• —
OO "O '»- 4-J
C O on
U TJ Q_ O
> T3 E
•— en >
oo c;
OO 'r- TJ
u 5 •*-
'J '-0 CL
X O TJ
OJ 4- •„
Ol
U
TJ
X
O i —
- — -TJ d"
00 OJ
— c en
t- o o
11 -Q S_
TJ Tl -1-
3 O C
1)
I —
O
t/l
Ti
"TJ
•r- •
4-J
s_
t;
CX
•
"O
• >—
ZJ
cr
, — -
•a
OJ
LJ
T3
cn i —
C CX
•»- oo
JC T-
0
(i» >,
4-J
u c
• — u
4-J ',_
4-J T5
•i— CX
• — ex
Tl
rr)
en
d -O T3
•i- d U
-C T) ?—
1/1 )X. en
T) en o c
. — cr 4— -•—
CX •(— M
tn jd ~u a)
CO 13 OJ
•U T) O >-
•i— CX O
00 OO
OO i—
U >U O
'-J E CX
X O TJ
1) OO >
Ln
a-1
~o
X'
0 r—
— --C "c"
LO OJ
-— - d en
x. c o
O) _C X.
TJ T3 T-
2 u c
U
^
•I—
-Ct
4_>
•^
u
•—T-
f_)
TJ
0
"T!
C
C
CX
T3
>
T3
CX
,_
en
c^
•t—
(_)
13
-c;
u o
U L.
TJ ex ai
14— f^~i
s_ - u
z; *o u
l/> 1) i-
4--J U-
e -i-j
O U 4-J
3 o
OO C
4— > 'jO
TJ U OO
O ' — 1)
i — U O
'-t— -r— TJ
4_>
OJ '— "
U T! TJ
••— CL Z3
O
- . ^ , —
e^ a-* u
d d
. i — . ,_ ',__
-c jd o
t/!
CX CX
UO !jO '—
u
•u u en
r- • - '*_
C 0 TJ
00 OO p
OO
— -
u
"O
X , —
o -_-•
"t/r~5 c
^ - 'U
5- O O
OJ -O U.
ra 'TJ ••—
sue
TJ
5
—
'13
TJ
IJ
>^
_C1
4-J
jy
=
'fl
4-J
aJ
4~)
d
OJ
-o .-
O CX
— cx
U T)
L. -d
O J-J
CL- —
TJ J
>
d
-C O
4-J T-
•r- 4-J
li T)
^;
en C
d O
• P— <4_
_Jd
4-J ~O
O 13
x. o
U- , —
(J
u
E ~d
o d
LO TJ
cn
d ~o d
••— d e>
jd T) ••—
I/) 4-J
T: en TJ
1 — d M
CX-— •-—
'/I .d X.
'jO O
u -o ex
> ' — T.
LO u
d en
x. o o
D -O '*.
TJ ,
T3 C
> OJ
>.
"O -U
•^-«
CX '-O
T) 'U
L- NJ
OJ
- O)
d V.
O 'J-
-r_
T .'S
^ ^J
x, cr
o ••—
U— i —
.d •—
4-J F
O TJ
'_
U— TJ
d
-C TJ
4-J
••— TJ
2 ej
0
en i —
d u
-r—
C* X.
oo O
•en TJ CL
d • — T;
— CX >
-d 00
t/l ' —
T! Tj i —
• -— T!
CL o E
I/I OO OO
— -
— -
OJ
-a
X r
o - -
,-^.TJ d
oo 1)
X- C O
-u _a j_
"TJ T) -t^
2 U d
u
"U
—
o •+—*
•r- C
-4-J OJ
T) £
NJ C
•r— T—
U TJ
0 U
CL 4-J
TJ d
> 'U
"O O)
•T— P
Q.-C
T3 -i—
X- (/I
00
Jd O 00
4-J cx a)
•i— ^
2 01 U
E OJ
TJ O 5-
TJ OO 14—
i-J
4~J jd -o
D 4-J •>—
3 •— n
3 cr
OO -r—
u -o -—
• — u
0 E OL)
•r- JV. JT
4-J O 4->
cn i. u-
d TJ 4-
•<— CX "O O
jc n
00 • " O OO
T) en - — d
— d o o
CX-.— -r-
00 Jd rj>_ 4-J
t/l O i—
1) T) CX O
> — T) CX
OO i^l
uO •
OJ u , —
(._) ^- ^3
x c E
•U oO '../I
00
*—- '
u
TJ
^ ,
O -— •
-^••5 •"-
oO OJ
'- O O
0) -0 U
TJ TJ T—
2 u c
XI
TJ
TJ
"O
, —
^
-^_
-------
Gas samples were drawn into evacuated glass containers from
locations about 5 cm (2 in.) above the liquid surface. The in-
fluence of the escaping, cold vapors may have been responsible
for the lack of reproducibility obtained from these samples.
The glass containers were cleaned and evacuated prior to taking
a sample. Some leaks were found in the stopcock seals after the
tests with solid water were completed but it could not be deter-
mined if this had occurred before or after the analyses were per-
formed .
Data from these tests are compiled in Table 5. The data
clearly show that vapor concentrations can be reduced by the
addition of coolants to spills of hazardous materials. Vapor
concentrations were reduced to 25 percent of the untreated values
by adding wet ice and to 10 percent or less of the untreated
values by adding dry ice or liquefied nitrogen. It was not poss-
ible to determine the contributions of convective mixing or sub-
strate heat transfer with this experimental arrangement,however.
Determination of Evaporation Rates
The third series of tests was performed with the apparatus
shown in Figure 2c. The objective of these tests was to circum-
vent the experimental errors of the vapor concentration tests by
a direct measurement of evaporative losses. These measurements
were used to compare the relative effectiveness of each candidate
coolant for reducing the hazards caused by vapors generated from
accidental spills.
An accurate description of the evaporation and dispersion
of vapors from an open spill requires the use of a mathematical
model incorporating terms for all point sources. Several models
have been developed by Sutton, Pasquill, Turner and others,
referenced in a review paper by Slade (1968). Most of these
models, and more recent variations, require computer calculations
for solution, and were considered beyond the scope of this in-
vestigation.
All of these models rely upon the basic laws of diffusion
and the interchangability of heat and mass transfer equations.
Typical mass transfer equations such as,
dq/dt = DS dC/dX; and
E ^ KmPvapS/F
where q = amount evaporated; t = time; D = diffusion coefficient;
S = area of source; C = concentration of diffusion species; X =
distance from source; E = evaporation rate; Km = mass transfer
coefficient; Pvap = vapor pressure and T = absolute temperature,
are related by Km = D/RT, where R = gas constant and that vapor
concentrations are conveniently expressed in terms of partial
33
-------
<=c
C£
LU
>.
0
a
LU
C£
13
OO
< LU
UJ ^
2: o
1
r^^
OO LU
0 eC
t— (
h- u_
<=C 0
a;
i — i
^. o
LU O
CJ Cu
•^.
0^
CJ ^
u
o;
o o
D_ LO
eC CM
>
LO
LU
1
ca
1—
on
OJ
S-
n3
i_
cu
CL
E
OJ
h-
oo
33
O
•r—
S-
ra
>
ca
en
E
o
• t—
4-J
03
S_
4-1
E
OJ
O
r"
O
CJ
cu
E
O
CL)
O
e£
-M
O
0
O
. — ,
O Lu
o o
oo r-~
i i —
4_
O
1 00
1
C2j • — •
^
Q.
CL
Oi
cr,
CJ ^-^
0 Lu
CO o
1
CHj —
E
CL
CL
r^
O
"^
^-,
O Lu
0 0
LO OO
1 CM
+
<2J
£
CL
Q.
O
r-^.
LO
CM
CJ Lu
o o
CM «d-
i LO
-f-
ca —
,_
CL
CL
O
CM
LO
to
CM
O
O
0 Lu
CD C
<* O
i — CM
1 CM
f
ca - —
£
CL
CL
cn
ro
0 —
0 Lu
LO O
CM CM
1 i —
1
ca- —
£
CL
CL
ro
i —
, — .
O Lu
O O
co r^-
i i —
+
ca — -
"
CL
0.
CO
CO
O Lu
0 0
r-. n
r— CD
+
ca ^ — '
^
cz
Q.
Q.
CO
c^f-
c^.
CM
:z
34
-------
pressures. The working equation used to correlate measurements
made in these laboratory tests was:
E = AKmyln
This equation represented the particular geometry and charac-
teristics of our tests in terms of E - evaporation rate (g/sec);
A = proportionality constant; Km = mass transfer coefficient
(sec/cm); y = radius of simulated spill (cm); P = atmospheric
pressure (N/m2) and p' and p" - vapor pressures at liquid sur-
face and far removed (representing vapor concentrations).
Similar equations can be developed, starting from the Max-
well diffusion theory, to follow the form:
- RTF p_pl
in which the additional terms are defined as S - area of spill
(cm2); D = diffusion coefficient (cm2/sec); P = atmospheric
pressure (torr); R = gas constant (cm3 mm Hg/mol/°K); F = film
thickness (cm) and M = molecular weight of diffusing species
(g/mol ) .
The initial tests in this series provided experimental
evaporation rates for representative spill materials that could
be compared with theoretical values. Glacial acetic acid, ace-
tone, methylene chloride, and isopropanol were selected for the
tests. The results of these tests are presented in Table 6.
Acetic acid and acetone were selected as the representative
hazardous materials for several tests in which the evaporation
rate was determined after the addition of the candidate coolants
The tests were prepared by weighing 100 g (3.5 oz) of each ma-
terial into a pyrex dish with a soil interliner. The amount of
coolant used was calculated to absorb approximately 30 Kcal (119
BTU). This value was calculated from heat transfer approxi-
mations of 1200 Kcal/m2 (440 BTU/ft2) over an area of 250 cm2
(39 in2).
Glacial acetic acid was selected because it would freeze at
approximately 16°C (61°F), with a heat of fusion of 45.9 cal/g
(82.6 BTU/lb), and pose a formidable heat load for the candidate
coolants. Acetone was selected because its thermophysical prop-
erties are typical of many low-boiling, hazardous materials. The
results of these tests are compiled in Table 7 for comparison.
Coolant Availability, Sto r a g e a n d A p p 1 i c a t i o n
An evaluation of the cooling capabilities and other perti-
nent data does allow appraisal of the potential candidates, but
any coolant selected must be readily available or storable and
35
-------
TABLE 6. COMPARISON OF MEASURED AND CALCULATED
EVAPORATION RATES USING DRY ICE (T - -78°C [-110°F]
Calculated Experimentally
Stefan Equation D e t e r,m i n e d
Material (g/cm2/sec) (g/ciWsec)
Water 0.0003 0.000013
Acetic Acid 0.0002 0.000012
Acetone 0.0039 0.0002
i-Propyl Alcohol 0.0077 0.0004
Methylene chloride 0.0080 0.0001
36
-------
C£
o
u_
CO
LU
I— CO
eC I
C£ eC
i — i
s: a:
o LU
. . t
I — I |
H- <£.
=C S
Qi
O CO
Q- ID
•=C O
> CD
LU 0£
cC
U_ M
0
o: (—
- — -CO Li_
oo r — o
S_ - — P O
0 CXJ r-
4- O II r—
O 1
-»^^ L--I-- ^ ^
»^~*, f— ^-— ^
CD
CO
-\
tM
E
u
"~\ CJ ^ — •
en - — -o u_
— • coo o
CXJ
i-D
1
O ro
i —
x sr
CXJ
V
LT> IT>
1 1
O O
1 — 1 —
X X
CXI CO
V
en en
i i
O CD
i — i —
X X
CD O n co in co
4-3 CXJ
rd HI h- 1 O
1
c:
••-
rd
S— C_?
O 0 -~ -
a. ro LU
rd CD CXJ 0
> C CO
LU o ii r-^
z: •• —
h-
4->
C
ro
r—
^3
^
•r-
co
*3~ *3~
1 1
O O
1 — 1 —
X X
CXI CXI
.
1 —
rd •
-(_> 13
"O C CD
•i— O) 4-3-
O E ra
rd -r- i —
s- rs
U CD U
•r- CX r—
4-3 X ra
CD CD U
CJ ^ — ^ * — ^
ct
•
l/l
ra
CD
E ro
.
4-> C
o
c
en tn
i i
o o
i — i —
X X
*3~ CTi
LO 3- ro
i i
O O
r— i —
X X
CXJ CTi
ro
rd — ^
4-> ~O
C CD
CD -1-3
E rd
•i — i —
CD S- 13
C CD O
O Q- r—
+-> X ro
CD CD CJ
U — ^ ^ — ^
-------
techniques must be available to ^ p. p 1 y the material to the spill
in a suitable form. Some consideration of these factors has been
given to potential coolants earlier in the program. An in-depth
review was made of the four coolants.
As a starting point, a survey letter was sent to a cross
section of companies engaged in the marketing of the coolants
and/or equipment for handling the coolants. The consensus of
suppliers was that C02 systems be used. The advantages cited
for their recommendation were the related usage, ready avail-
ability, and the existence of handling equipment which could be
modified for this application. Spent C02 gas is not considered
hazardous per se. It is heavier than air and stays near ground
level until gradually dispersed by wind action. Depths of cold
C 0 2 gas at the point of application are expected to range from
0.5 to 1 meter (1.5 to 3.0 ft) maximum. They are usually
visible, because of moisture condensation from humid air. For
safety, oxygen concentrations should be monitored at the spill
site since high C02 concentrations will displace the air and
could pose an asphyxiation problem to personnel.
Dry ice was suggested as a feasible alternative if a crush-
ing operation (with related packaging disposal) and particulate
dispersal equipment could be evolved. Coolant quantities and
costs could be lower with a properly developed system.
Subcooled wet ice was also feasible. For mobile operation,
system design would require some equipment modification for
optimum field use.
Liquid N2 was considered to be the least feasible of the
three candidate coolants because of high storage losses and prob-
lems of delivery to remote spill sites. Coolant storage and
typical losses for the coolants considered are presented in
Table 8. A summary of selected properties for the candidate
coolants is presented in Table 9.
Solid Water --
Ice is commercially made and stored in either the wet (0°C
[32°F])or cooled (-18 to -28°C [0 to -20°F]) condition in forms
ranging from 45 to 150 Kg (99 to 330 Ib) blocks to 1 to 6 mm
(0.03 to 0.24 in.) thick flakes. When stored at 0°C (32°F) in
insulated bins, ice will experience losses of 1 to 2 percent per
day due to melting. Cooled ice is kept relatively dry by a re-
frigerated annulus in the storage bin, which keeps losses almost
negligible and keeps the ice free-flowing for one to two months.
Handling equipment for high volume operation will require
some special design and/or modification. Shavers or crushers
with related conveyors may not be a problem. Transfer up to 150
meters (490 ft) in plants is commonly accomplished with medium
pressure blowers using cooled air. Such equipment may require
38
-------
TABLE 8. COOLANT STORAGE LOSSES
Coolant
L N2
L C02
S C02
dry ice
S H20
wet ice
Container
Trailer
Storage
Trai ler
Skid w/ reefer
Box w/blocks
Box w/pellets
Bin w/ reefer
Coolant Loss
%/24 hr Day
0.8
10.0
0.3
0
1.7
3.0
0
Remarks
LOX, trailer specifications
Liquid Carbonic quotation
Chemetron skid without re-
frigeration unit
Mechanical refrigeration usea
Chemetron @ 21°C (70°F)
Est. @ 2X block rate
Flake ice, held at -17 to
-12°C (0-10°F)
Bin, insulated
1.0
Est. @ 1/2 S C02 rate
39
-------
GO
\ —
2:
=:-
o
o
o
LU
^
a
i — i
o
eC
1 !
o
CO
LU
1 — i
1—
a:
LU
n
0
a:
D_
a
LU
1—
CJ
LU
I
U-i
GO
en
1 1 i
— j
CO
i —
00 i —
4-J i —
00 ••-
0 Q.
*J i. r— —
C 0 03 «rt
03 l*- U —
o a.
o >>
O 1—
c
c
\
•w*>
00
3 —
c -r-
O l-
-^ 0)
i ** Jz.
-o u
aj en c:
J-> C OJ
> O3 -i— •(— • — -
E ' — (J **
•r- O •(-
-M O <4-
00 (_J 4-
UJ UJ
1 cn-^^-
c -^ cn
j ^ Ja ^
o a. 03
; O 03 c_)
f 1 c ) •
OJ
3 ^~~
oo £
00 -t-l
OJ 03
: Q.
OJ
i.
=i
03
i- CJ
OJ 0
CL- — •
E
I—
d)
-4-3 OO
03 4-J
"O C
• i— 03
"O r~-
C 0
03 O
c_> o
o o
O O U")
O O 0
(— Ol r—
.000
4_> ±J 4-t
O O O
o o en
CO f**"- f***
o o o
CNJ CNJ CNJ
4-1 4-14-1
000
CO CNJ r—
i — CNJ
(
CNJ CNJ CNJ
CNJ CNJ CNJ
4-1 4-1 4-1
CO CNJ r—
co ro n
•— CNJ
O O 0
en cri co
r~~- «3~
en en ^o
r^^ co *~o
i —
I (— r—
CO LT)
. .
O r^ oo
, — p-^
i i
OJ O)
u u
•r— -1 —
4_) >^
cu s_
2 Q
O O
O O
LO U~>
CNJ CNJ
0 O
4-J 4->
O O
0 0
LO tn
o o
CNJ CNJ
4- 1 4-1
O O
en en
CNJ CNJ
CNJ CNJ
4-1 4-1
o o
l r~—
CNJ CNJ
un o
*^J" I —
r~-* CNJ
vo (-^
kH) ^o
^.
,
0 r-
CNJ
CO LO
r^ co'
i i
T3 0)
•r- CJi-O
3 ro O)
CT i- -C S
•r- o oo o
, — *-) 03 C
OO i — OO
4~
O
OJ *« 4->
c -a LO
O T— *3"
-Q X
i- O
03 •(—
0 Q
O
O
O
OO
o
•*-*
o
o
CNJ
o
CNJ
4- 1
O
0
CNJ
CNJ
CNJ
4- 1
0
CNJ
CNJ
O
n
CO
^
Tt
!
lvD
en
1 —
^
OJ
•a en
•r- O
Z> S-
cr -^-J
•1 T~
40
-------
as much as 5 kg air/kg ice (5 1 b air/lb ice) depending upon sys-
tem characteristics. Air lock modifications at the spill dis-
persal point may well suggest a low pressure fan system, as used
in snow blowers, highway mulchers, or silage blowers. These
would discharge directly over the spill area. Sized ice could
be further pulverized by a fan blade and cracking nozzle attach-
ments. High noise levels, requiring ear protection would be ex-
pected.
The simplest ice system would use a local supply of small-
sized ice, delivered to the spill site by the supplier's truck.
Deliveries could be expected within 2-4 hours, under normal work-
ing conditions, but delays of 16 hours or more could occur during
off hours or during weekends.
A crusher unit would permit the use of large blocks. Re-
finements, such as a blower unit, would be needed to increase the
effectiveness of the system.
More complex and costly ice systems could be devised for 1-4
hour response time supplying cooled flake ice. Flake Ice Company
has a mobile ice generating system rated at 20 tons/day (18,000
kg). Six of these systems would be required to reach 5-tons/hr
(4500 kg), the initial cooling rate. They would also require
electrical power and water supply accessories.
A more feasible approach may be to use a mobile cooled ice
storage bin system with field cooling and dispersing accessories.
The bin could be filled by local suppliers or by an ice maker at
the bin storage site.
Solid Carbon Dioxide --
Dry ice is commercially available in either block (23 kg
[50 1 b]) or pellet (1 cm [0.39 in.] diameter) forms, usually
packaged in insulated boxed containing 1100 to 1300 kg (2400 to
2900 Ib). Storage of the solid form of carbon dioxide usually
incurs a loss of 1.5 to 3 percent per day due to sublimation,
which would suggest procurement from local suppliers when needed.
Crushing or shaving operations are usually performed just prior
to use; 5 to 10 percent losses are experienced by suppliers during
in-plant operations.
These high sublimation losses preclude storage beyond 1 to
2 weeks and have been a major factor in the decline of distribu-
tion depots. Storage and handling now favor the "no-loss" liquid-
to-snow carbon dioxide systems but delivery times may be somewhat
greater than expected from solid water suppliers. Portable liquid
units are still not as universally available as block ice.
Handling equipment for spill applications can be similar to
that used for applying water ice. A dry ice handling system
would have the versatility of being applicable with water ice.
41
-------
Production of solid carbon uioxide at the spill site does
not appear to be a feasible approach. Production would best be
accomplished by the conversion of liquid to snow.
Liquefied Carbon Dioxide --
Liquefied carbon dioxide is also commercially available from
storage in insulated tanks, where it is held as a liquid at -18°C
(0°F) and 15,200torr (20 atmospheres) pressure. The liquid carbon
dioxide can be flashed to atmospheric pressure, through expansion
nozzles, to produce solid "snow"). Maximum conversion efficiency
is reported as 45 t 3 percent. Delivery times should be within
2 to 20 hours during the normal work schedule, but may become a
limiting factor during off hours or weekends.
Equipment is commercially available which has been designed
for analagous service and could be adapted into the proposed
system for treating accidental spills with a minimum of modifi-
cation. Liquefied carbon dioxide would be transported to the
spill site in insulated tank trailers. These are available in
16,000 to 23,000 kg (35,000 to 51,000 Ib) capacities. They meet
ASME/DOT specifications (including baffles) for road service,
but do not include self-refrigeration units. Filling with sub-
cooled liquid and insulating with 10 to 15 cm (3.9 to 5.9 in.)
of polyurethane are enough to keep losses negligible, however,
for the anticipated working times. The most economical system
would use the tank trailers from local suppliers to supply the
distribution system manned by the spill response team.
A fast-response system would include a mobile tanker fitted
with a se1f-refrigeration system and dispersing hoses and nozzles.
A typical unit which is available is a 5-ton (4500 kg), skid-
type tank fitted with a self-refrigeration unit, two 30.5 m (100
ft) reels of 3.8 cm (1.5 in.) hose and playpipe nozzles.
Liquefied Nitrogen --
Nitrogen is transported and used in the liquid state for
many commercial and industrial operations. It is normally trans-
ported in vacuum-insulated tank trailers at -196°C (-320°F) with
pressure relief valves set at approximately one atmosphere. These
trailers have capacities of 18,000 to 21,500 kg (40,000 to 47,400
Ibs) and normally carry such accessories as a pump and flexible
stainless steel hose which are capable of transferring this load
in approximately one hour.
Transporting and storage losses are between 0.8 and 10 per-
cent per day, which would preclude extended storage. The more
feasible approach would be to use the supplier's tank trailer as
the source of coolant, and couple this to a distribution hose
and dispersal accessories which are part of the spill response
unit.
42
-------
Liquefied nitrogen is available within 2 to 20 hours in
most sections of the country. Some supply problems could arise
from the quantities required for larger spills and when needed
on weekends and holidays.
The dispersal equipment for liquefied nitrogen would be 30
to 150 meters (98 to 490 ft) of 5 cm (2 in.) flexible hose, with
couplings. This hose is available in 4.5 to 7.6 meter (15 to 25
ft) lengths, which would weigh between 13 and 45 kg (30 to 100
Ib) each. Suppliers may be able to include extra hose with the
trailer, on occasion, but it would be preferable to include
extra lengths of hose and couplings with the spill response unit.
During the first half hour of cooling with the liquefied
nitrogen system, at least 2 to 4 operators would be needed to
assemble and deploy the hoses. Subsequently, the maintenance
cooling operations could be handled by 1 or 2 men.
Dispersal of liquefied nitrogen to produce a uniform cover
over the hazardous, volatile spill may be difficult. The com-
bination of very low temperature and low density should make it
possible to pour the nitrogen on the surface and have it spread
over the solidified spill. Pouring may not be a practical
approach. It is not possible at present to determine the losses
which would result if the LN2 would be sprayed onto the surface.
The alternative systems and components available for hand-
ling the three candidate coolants have been summarized in Table
10 for comparison. The costs, compiled in the last column, are
estimates of the total equipment expenditures required to prepare
a system for handling a spill of 1000 kg (2200 Ib) covering an
area of 50 m2 (540 ft2) .
SELECTION OF CANDIDATE COOLANTS
With the data derived from the laboratory studies combined
with the basic thermophysica1 properties it was expected that a
selection could be made of not more than two coolants to be
carried forward in the program. The possibility that two cool-
ants might be selected existed because of cost and availability
factors. For spills of flammable materials where vapor concen-
trations of 1 percent or less would be acceptable (since gen-
erally this lean a mixture will be below the lower f1ammabi1ity
limit), a coolant with a lower cooling potential could be se-
lected if cost and availability were favorable. For spills of
toxic materials, where vapor concentrations below 100 ppm are de-
sired, cost and availability factor would be secondary to the
cooling capaci ty.
The results of the laboratory work clearly showed that C02
met minimum requirements of cooling capacity. Thus, a firm com-
mitment was made to proceed with C02- A decision could not be
43
-------
o
00
D_
O
oo
oo
o
o
o
c u
05-
O X
<_)
-yl ^3
O -
^ 2. °
ai u
C ^/1
r3 =
O Q
CVJ CVj I
3D
c
o
C/l K 1/1
c/) •— O .' t/l —
•_ a cj '-a
aj s_ -^ | u
--- — a) s- •— — o — o
a. a ~ 2 , CL =:
u -a — .— 3 TJ
O i—
• — i — LO
1
r>» TT ' O O
<— CXJ , -DO
CXi r*».
i
I
1
I
i o o
CXI LO O
i— r-*. | cxj r-^
ii i i
CO CXt lOO
1 ~ ^
1
o '
J-J
T3 ;
i_ +J
— • 1) ' —
o ci 1 c: -a
cu '*. . .co a.
C CJ OJ U
O Qi I <-j' O
'^j — • <— «
"""•»•. ^Xl \
u o fl*J u
OJ <_? , r— — .
5 -~ ' "~ T;
r— "3 J3 *O
^r LO i ^r '.o
1 ' , ' '
i
' n
i O
1
^_ N
-
CO
—
1
•^— '
CXJ
rr
cxj n cxi
i i '
« — CXI CXI
0 O
•i— C_J
"*-* ^^
_«C 3 '— .
U O O
C3 •*— >,
'— U U
•y) C
en •— O
i- O '_>
U '— . "--*
-r— •— U i_
r^
'*. 31
CJ C
_^; -.—
™ •*-»
T3 *->
U.
i/r
1_ T3
OJ C
r—
o. cu
CL '-O
13 '~>
i
0
CO
CTi
m
i)
— o'o
13 i- O
3T" -*—1 CT>
_ -„ „
-J m 1
-------
reached on whether C02 was best ,n liquid or solid form; that de-
cision depended in part on methods of storage and application.
Firm decisions on liquefied nitrogen and wet ice (either at
0°C [32°F] or subcooled) could not be made without additional
testing on a larger scale than was possible in the laboratory.
The need for these scale-up tests reflect concerns regarding
both the effective application of liquid nitrogen and the ability
of wet ice to provide sufficient temperature reduction in real
spill situations. Because of these uncertainties a series of
scaled-up tests were conducted.
45
-------
SECTION 6
SCALED UP TESTING OF SELECTED COOLANTS
The inability to select a final coolant candidate based on
the review of the-rmophys i cal properties, the laboratory test
data, and other relevant information, indicated that further
testing and an evaluation of the logistic factors was required.
Accordingly, subscale field tests were conducted to provide
additional necessary information.
SCALED UP TEST DESIGN
Test Size and Configuration
The conclusion reached at the end of the prior phase of the
program was that a larger sized test was necessary to obtain
data on a practical scale, A test spill size of 9.3 m2 (100 ft2)
was selected. This was in the form of a pit 3.05 x 3.05 x 0.30
meters (10 x 10 x 1 ft) deep. The bottom was leveled with packed
sand and lined with a polyethylene sheet to prevent percolation
of the spilled liquid into the soil. This allowed renewal of
the pit after each test by replacing the liner.
Hazardous Spill Simulant
In selecting a spill test chemical, a material with a high
vapor pressure at ambient temperature appeared most desirable
since cooling effects would be most significant. A review of
several potential materials resulted in the selection of di-
ethyl ether as the simulant. It is a nontoxic, volatile liquid
with very limited water solubility which limits the interference
in the vaporization rate caused by dilution of the spill with
water from melting wet ice.
Test Measurements
The basic measurement taken in each of the subscale field
tests was the temperature reduction of the spill simulant, di-
et h y 1 ether. Thermocouples were located in the spilled liquid
and at a depth of 0.30 meters (1 ft) in the ground below the
plastic liner.
Ether vapor concentratrations were measured downwind of the
46
-------
spill with combustible gas detec ors. Weather instruments were
installed to measure wind direction and velocity along with
ambient temperature.
Oxygen indicators were installed to determine the effect of
CO? and nitrogen application on breathability downwind of the
spill.
Procedures for Coolant Application
Since these were still preliminary tests to select a candi-
date material, complicated equipment to apply the coolants was
not justified. Both wet ice and dry ice were crushed and applied
to the diethyl ether pool as particles using hand shovels. The
particle size distribution was not measured for either wet ice
o r d ry i c e .
Liquid carbon dioxide from portable tanks was discharged
through snow horns. This resulted in a combined discharge of
solid C02 snow and C02 vapor.
Liquefied nitrogen was applied to the spill as a liquid
spray. The liquid was delivered to the test site in a tank
truck. A 3.8 cm (1-1/2 in.) stainless steel hose connected to
the tanker was used for delivery of the nitrogen to the spill
surface.
COOLANT TESTS
All coolant tests were run with 208 liters (55 gallons) of
ether in the pit giving an average depth of 2.5 cm (1 in.). Dis-
cussions of results for each test are given in the following
sections. The results of the tests are plotted in Figure 7.
Wet Ice
For the wet ice tests, 380 kg (850 Ibs) of ice was applied
manually. The crushed ice was evenly distributed to give a
visible ice layer over the spill surface. The spill temperature
reduction was fairly rapid from 27°C to -5.5°C (80°F to 22°F).
The test duration was four hours. Over this time the
temperature of the spill remained fairly constant and there was
no need to add additional ice. In going from 27°C to -5.5°C
(80°F to 22°F) the equilibrium vapor pressure was reduced from
580 torr to 150 torr. The. measured vapor concentrations were in
line with the equilibrium vapor pressure. Initial vapor concen-
trations were in the range of 5 percent. Some variation was ob-
served which appeared to be caused by wind fluctuations. After
the ice was applied, the vapor concentration declined to between
1 and 2 percent. This is close to the lower flammable limit of
1.85 percent for diethyl ether. It would be expected that vapor
47
-------
30
20
10
0
-10
-20
O -30
o
£ -40
3
To
55 -50
Q.
| -60
-70
-80
-90
-100
-110
-120
-130
9.3m2 (100 ft2) pond
208 ,e. (55 gal) diethyl ether
-40 -30 -20-10 0 10 20 30 40 50 60 70 80 90 100 110 120 130
Time, min.
Figure 7. Subscale field test results with ethyl ether.
-------
concentrations would be well bel w the lower flammable limit down-
wind of the spill.
After the addition of ice the vapor level was still well
above the TLV of 400 ppm. Some reduction in the area downwind
of the spill wherein the vapor concentration exceeded the TLV of
400 ppm would be expected.
Dry Ice
In the dry ice tests, 252 kg (555 Ibs) of crushed dry ice
was added manually to the diethyl ether pool. Efficiency of
application was about 90 percent with most losses occurring
through sublimation. Temperature reduction was from 20°C to
-85°C (68°F to -120°F). This is equivalent to a reduction in
vapor pressure from 442 torr to 0.45 torr.
No meaningful diethyl ether vapor measurements were ob-
tained. The C02 interfered with the combustion type gas de-
tectors. Estimates of the diethyl ether concentration based on
equilibrium vapor pressure put the level in the range of 600 ppm
by volume but this value is not consistent with values measured
in subsequent tests. The values obtained during the liquid C02
coolant tests were eventually used as a reference.
The 252 kg (555 Ibs) of dry ice maintained the spill temp-
erature between -80°C to -85°C (-110°F to -120°F) for 1.3 hours.
At that point the spill began a slow increase in temperature,
reaching -20°C (-4°F) in fifty minutes. At this point the test
was terminated.
Liquid Nitrogen
Approximately 1020 kg (2250 Ibs) of liquid nitrogen was
dispersed onto the spill from the tank truck. The application
of the cryogen to the spill surface resulted in intense boiling
and the formation of a condensed moisture fog that obscured the
test pit and made visual observations impossible. For this
reason the discharge was terminated after twenty minutes to ob-
tain a view of the spill. After the coolant discharge was term-
inated, several minutes elapsed before the fog dissipated suf-
ficiently to permit observation of the pit. At that time it
was clear that the diethyl ether had been frozen. No additional
liquid nitrogen was applied.
The temperature at the beginning of the test was 13°C (55°F).
Unlike the wet ice and dry ice tests, the spill temperature con-
tinued to decrease after termination of coolant application. The
minimum temperature recorded, -120°C (-184°F), occurred twenty-
five minutes after coolant application had ceased.
There was no leveling off of the spill temperature as was
49
-------
observed in tests with wet ice a d dry ice. After reaching a
minimum, the temperature immediately began to increase. It pass-
ed through the -85°C (68°F) range achieved with the dry ice less
than one hour into the test, and returned to ambient temperature
less than two hours into the test.
The equilibrium vapor pressure of diethyl ether at -120°C
( -18 4 ° F) is about 0.01 torr. Due to the inability to use com-
bustible gas detection in the dry ice tests, gas samples were
taken during this test and analyzed on a chromatograph. In the
downwind direction the pretest diethyl ether concentration was
9.3 percent. As the spill passed through the temperature mini-
mum, a downwind level of 116 ppm was measured. This is well be-
low the TLV but still above the odor threshold. Maintenance of
vapor concentrations below the TLV would require continued appli-
cation of L N 2 . The 1020 kg (2250 Ibs) of L N 2 added to the spill
in this test held the diethyl ether concentration below the TLV
for no more than 40 minutes.
Liquid Carbon Dioxide
Liquid carbon dioxide was discharged to the spill through
a snow horn. Three hundred forty kilograms (750 Ibs) of CO2,
the volume of the storage unit, were discharged at a rate of
90.7 kg (200 Ibs) per minute. Expansion through snow horns con-
verts the pressurized, refrigerated liquid to a combination of
vapor and solid particles. Conversion efficiencies to particles
as high as 50 percent have been reported but in these tests the
conversion efficiency appeared to be no better than 15 percent.
It was subsequently confirmed that the observed 15 percent
conversion is consistent with the type of snow horns normally
available. Their purpose is fire extinguishment where the vaoor
is the effective material. Snow formation is used as a means to
project the C02 beyond the reach of plain vapor discharge. Higher
conversion efficiencies are achievable only with special show
horn designs.
The spill temperature achieved by the addition of dry ice
formed from liquid C02 by a snow horn approximated that attained
with the crushed dry ice. There was no maintenance of the low
temperature, however. Almost immediately upon conclusion of
solid C02 snow addition, the spill temperature began to increase.
This is attributable to the poor conversion to snow (15 percent)
which is the only effective coolant form.
Like the nitrogen application, the use of liquid C02 pro-
duced a heavy obscuring vapor cloud above the spill. Gas samples
taken downwind of the spill and analyzed on a chromatograph gave
a pretest diethyl ether value of 8300 ppm and a post application
value of 96 ppm. Both of these values are lower than those ob-
served in the liquid nitrogen coolant tests. The lower diethyl
50
-------
ether concentrations were probab y caused by differences in
ambient conditions -- wind speed, wind direction and air tempera-
ture. The significant feature in these test sequences is the
ratio of the vapor concentration before and after treatment of
the spill with coolant,
Since similar low temperatures were recorded in both the
dry ice and liquid C02 tests, the vapor concentration measure-
ments obtained are probably applicable to the dry ice test also.
COOLANT SELECTION
The basic importance of the data from the subscale tests is
th-e difference in temperature achieved in the ether by each
coolant and the maintenance of that temperature after application
was stopped. On the basis of these data, crushed dry ice would
be judged the superior spill coolant. Other factors need to be
considered, however. A complete review of the four coolant can-
didates was made and a summary of the advantages and disadvant-
ages of each is presented in Table 11.
Analysis of the total data clearly showed that liquid CO2
was the least desirable of the four materials. The poor con-
version efficiency and obscuring gas cloud are serious impedi-
ments when equivalent temperatures can be achieved with dry ice
without those difficulties. Efforts were made to identify and
obtain snow horns of better efficiency. Although the C02 in-
dustry quotes higher efficiencies, none were able or willing to
provide MSA with either a price of equipment or a design to
build to.
Choosing one from the other three coolants required assess-
ing advantages and disadvantages. These are summarized for each
coolant in Table 11. Wet ice was finally eliminated basically
because of its limited temperature reduction. There were very
few hazardous volatile materials with vapor concentrations that
could be reduced below the TLV by cooling to near 0°C (32°F).
Concentrations might be reduced below LELs but other more readily
available techniques such as foam application can also provide
such reductions.
The use of liquid nitrogen can result in the greatest re-
duction in spill temperature and for highly toxic materials with
low TLVs, it may be the only coolant which would be effective.
The problems of logistics, the need for large quantities, and the
poor persistence of temperature reduction led to the selection of
dry ice as the best all around coolant material.
Dry ice was chosen as the optimum material. It is appli-
cable to the broadest range of volatile chemicals readily avail-
able in a suitable form, and with the least complications. This
selection does not indicate that the other coolants are unsuit-
51
-------
(~n
~z.
o
on
»—i
Cu
O
Z
LLJ
CD
o
!>—
{—}
.
r— —
I
UJ
— J
cn
"-H
L
r —
o*
QJ
CT
TC
4~>
C
t)
^
>T3
O*l
•F—
Q
on
QJ
cr
T
C
"3
c£
C
GJ
cr
o
^__
CJ
4-J
Q ._ .,—
••- CL CL
on i/o i/i
• -—
> U- F —
o o
oo o
QJ >i U
U 4-J
n ••- o
~O C 4-J
a; • —
1— o T3
Qj •(— QJ
cn >-, > ~o
03 • — QJ
i- 4-J C QJ
O C ••— C
4-J 03
oo u on oo
•r- — QJ
C U- QJ -r-
C QJ -i- O
(/) CJ1 , — 4_J i )
QJ T- c
on un c 03 "O
on — -
— =3 >> 0
O x 1- on
1— ' — • O QJ
O u cn c
CL -a i- 03
;> Z^ O3 1 4-J
i — cxj ro
OO Qj
oo -Q
ro ••—
jQ t/l
oo
"O O
C CL
12
O on
CL QJ
L-
S- Z)
QJ 4-J
C QJ
0 CL
CXJ QJ
0 4-J
( >
2
5C O
-d
4-J >*)
F
QJ E
CL O
O3 i-
OJ 4->
_C X
( ) L i ]
i — CXJ
GJ C
•r- QJ
4- cn
QJ O
Z5 i-
CT4-J
CXJ
O oo
O
"O ~O
•i— 5^
r— O3
0 N
on 03
jC
O3 U—
o
aj
4-J QJ t-
U E 0
QJ C CL
U- -i— 03
U- 03 >
OJ S-
4-J CXJ
4-> c z:
00 QJ
O >>
t J QJ JO
i — •
on _Q "O
n^ ,p_ .,—
on ZJ
4-J on cr
O O ••-
^ CL. ••—
•cj- to
i
I ••- >> QJ
-r- 00 4-> Jt: O
, .,— -, — 4-J 4-J
CL > C
CL '•— C "O
O3 on u O QJ
QJ -r- -r-
C U > CXJr—
00 0 CL
CL "O C LJ CL
Z) QJ -r- T3
i- T3
Qj >i— — •—
L/! 4-J > l/l O3
0 C QJ =3
, — • (J o3 (.J
CXJ*'~ *i~ C -C O3
O CL 4- QJ 4-J
LJ oO ••— Cn 4-J
C >-, QJ C
~C) QJ cn x > ZJ
-.- _c T- o •— Q
cr ~a c TJ
•r- O T^ C GJ •—
O r— X O '•-
U C r- >,-r- QJ CL
Qj O U 4-J CL on on
JZ -r- ••— l/l QJ '•—
cn 4-J cxj . — i- on QJ
••- ro O ••— 4- O 03 -C
~T UO-Q O2I-d4->
. — CXJ CO
-
•1 —
CL
on
U—
O
c:
o
•F—
4-J
(_)
ZJ
-a
GJ
i_
QJ
ZJ
T3
on S—
QJ GJ
oo CL
1/1 E
O QJ
i — 4-J
QJ QJ
Cn-—
rO -Q
O C
4~> O
OO LO
03
O QJ
z: o:
i — CXJ
QJ
•r- GJ
4- c ~a
QJ O -r-
^ ja x
cr ^ o
__J 0 Q
c
o
•I —
4-J
O3
U
•r—
CL
CL
03
QJ
O
14—
QJ
^
^
03
i/i on
QJ Qj
on o
oo GJ
O C
r—
cn
GJ C
cn-F-
O3 TZi
V- -—
0 -r-
00 (J
^- cxj
L. rxj
o o
CXI CXI
O 2: X3
c_) i QJ
4-J , — on e_) o ••—
C GJ <4-
"3 Ct 4— j^; "O - — QJ
>CXJO O4-JOJJOZJ
*o z: •-— -r- 4-) 03 cr
03 4-J cn 3 o i — •(—
~O O3 C QJ -r- i
4-J QJ (_) ••— on T-J 03
00 -F- -r- CXJ. — E O > C
O M— i — O O QJ i- O3 O3 QJ
UQJCLCJO. — CL .r;, —
ZJ CL O JZ) >-, -4-J ~U
i-CTTJ'U OGJ- — C
QJ -i— •(— ~O i- -Q -r— 1^ O3
4->. — oOZ3-^CL "OGJ-iC
4-J on cr CL c o3 '4-
Qj 1~ QJ-f— T3OO3QJO3O
CQ o i • — cc:^:c_)ct:oo4-j
' — • CXI CO ^J" LT) '•jQ l^~--
c
o
_Q
A3
t-J> QJ
"O 'i—
••— X
O •(—
CO Cj
, —
1 r—
QJ -r-
l- CL
t/1
4~>
O *4-
C O
on Qj
••- E
Z3
i — • O
•r- >
CL
on "O
QJ >1 ZJ OO
J^ ^- Q- 4->
4~> 4-J •(— i
C r— QJ
^ QJ E
O -r- QJ
0 -C QJ
Qj -i— -4-t O
i_ l4_ ., —
Z) 4- on
4-J ZJ QJ QJ
ra on on -C
i- 03 4-J
QJ -O QJ
CL QJ i- c
E LJ (J QJ
QJ Z5 C JZT
i— -o •— • i
•— CXJ
!
QJ
~s
, —
O3 "O
> QJ
03 4-J
Q) U
lO >v QJ
•r- O
O T3 i-
4-> 03 CL
QJ
4-J S- GJ
QJ 4->
4- on c
T3 O O3
00 S U
' — cxi ro
0)
52
-------
able. In certain spill scenario: the cooling capability of wet
ice might be adequate or the extremely low temperature capa-
bility of liquid nitrogen might be required.
53
-------
SECTION 7
EQUIPMENT DEVELOPMENT
The selection of dry ice as the prime coolant brought with
it a need for equipment to crush the commercially available 23 kg
(50 lb) blocks to a suitable particle size, and to dispense the
particles over the spill surface. Since this was primarily a
feasibility study, extensive development of sophisticated op-
timized equipment was not warranted. This phase was constrained
by directions to develop a prototype unit to allow the demon-
stration of feasibility using existing commercially available
equipment to the greatest extent possible. The evolution of the
necessary equipment for the dry ice distribution system was con-
ducted within those guidelines.
REVIEW AND EVALUATION OF COMMERCIAL EQUIPMENT
Two basic components were defined for the desired system:
a crusher and a conveyor. Certain auxiliary equipment could also
be identified; these were primarily power train items whose se-
lection was secondary to the two major components.
Crushers, Choppers, Hammermi 1 1 s , et_c_.
To provide a list of equipment types to be examined, a
questionnaire was sent to a broad cross section of manufacturers
of crushing, grinding, shredding and chopping equipment including
those concerned with the generation of dry ice crystals for cloud
seeding and fog dispersal. The majority of commercial equipment
available (hammermi11s , soil and mineral grinders, and wood
chippers), readily converted dry ice blocks into fine particles.
The problem was that these devices produced a particle size dis-
tribution that contained far too many very small particles. This
resulted in high coolant losses from sublimation and also se-
verely limited the ability to project the crushed dry ice.
The basic design of equipment classified as shredders or
chippers allowed modifications both of the speed (rpm) of oper-
ation and the number and spacing of tines. Through variation
of these features, significant changes could be made in the size
distribution of the dry ice particles. A number of modifications
were tested with a wood chipper and a compost shredder. The
initial effort was concentrated on the wood chipper since it con-
tained an integral blower. Ultimately, however, the shredder was
54
-------
found to be capable of producing a more desirable size distri-
bution and it was selected as the crusher portion of the system.
The unit selected was a Model 4G Shredder/Grinder manu-
factured by the W.W. Grinder Company, Wichita, Kansas. This
model was chosen because it was the smallest unit of that design
which would accept 25.4 cm (10 in.) cubes of dry ice. The
crushed dry ice is discharged by gravity through a screen and
down a chute located beneath the rotating tines. The dimensions
and pertinent features of the unit are given in Table 12.
TABLE 12. SHREDDER/GRINDER (W.W. Grinder Co.,Model 4-G)
Power Tra-Tn 0.67 KW (9 HP) at 3200 rpm
Hopper Size - Mill 43.2 x 43.2 cm
(17 in. x 17 in.)
- Top 89 x 91.5 cm
(35 x 36 in.)
Screen Opening 5.1 cm (2 in.)
Mil 1 Speed Max. - 2500 rpm
Min. - 600 r p in
Conveyors, Blowers, etc.
A device was needed that could take the crushed material
from the shredder and distribute it over the spill. The equip-
ment compatible with most of the crushing units that were in-
vestigated involved blower mechanisms. In some cases mechanical
conveyors (belts, buckets and the like) were available. Pneu-
matic conveying,where particles are blended with a carrier air
stream rather than distributed by simply blowing, was an alter-
nate mechanism.
Mechanical conveyors were deemed unsuitable since tney
could not project the dry ice particles over the spill area.
Only minimal consideration was given to this approach.
While pneumatic conveying had certain advantages in mobility
and flexibility of the discharge port, it was anticipated that
excessive dry ice sublimation losses would result. The blowers
appeared to have more potential as a dry ice dispersion mechanism
than did pneumatic conveyors.
A number of blower types were evaluated and tne simple snow
blower design appeared to be the best for this application. For-
age blowers and similar units that were considered could project
crushed materials for greater distances. The snow blower was con-
sidered to be adequate for the spill size to be employed in the
55
-------
demonstration phase of this prog:am. In addition, the snow
blower incorporated an auger mechanism to feed the snow to the
blower whereas the forage blowers did not. In this application
it was believed that the auger feed would simplify mating the
blower unit to the dry ice crusher.
The initial selection of a snow blower was based upon the
mechanical needs of the ultimate equipment, but it was in-
fluenced by a desire to make the equipment as simple as possible
while utilizing existing commercial items. The snow blower
appeared to provide both a pick-up and distribution capability.
INTEGRATE AND TEST PROTOTYPE
It was reasonably clear that even though the blower could
project the dry ice particles over some distance, manipulation
and movement of the equipment at the test site would be necessary
The crusher came mounted on a four wheeled frame to provide mo-
bility. Integration of the snow blower into that frame would
facilitate simultaneous movement of the two components. With
that accomplished, testing of the unit could be conducted.
Component Integration
The crusher placement on the wheeled frame was such that a
large open area existed below the discharge screen. A receiving
tray was built into that area, which was closed on the bottom
and three sides. The open side faced out from the crusher. The
blower was mated to the crusher so that the auger pick up pro-
truded into the tray and the blower body provided closure of the
fourth side.
In operation, the crushed dry ice was fed by gravity into
the tray where it was picked up by the auger, fed to the blower
and discharged through a port at the top of the snow blower. A
flexible hood was attached to the discharge port to allow some
variation in the direction of discharge without moving the en-
tire crusher-snow blower unit.
The snow blower frame was attached to the crusher frame.
The method of attachment raised the blower so that its wheeled
chassis was clear of the ground. Movement of the crusher frame
thus moved the total unit.
Preliminary tests with this unit indicated certain other
refinements were needed. An inclined baffle was incorporated
into the crusher discharge to better direct the dry ice to the
pick-up auger of the snow blower. The discharge port of the
snow blower was moved from the top to the side of the unit to
eliminate the bends in the discharge path. This benefited the
discharge pattern and the throw distance.
56
-------
Tests with dry ice blocks revealed two- deficiencies. The
crusher could process dry ice blocks faster than the snow blower
could discharge the dry ice particles and the discharge pattern
was not uniform. It was decided that for the feasibility study
a slower feed rate into the crusher would be acceptable, with
the understanding that the equipment could be optimized if the
test results were negative.
The uneven dry ice discharge pattern from the snow blower
reflected the particle size range. The larger particles could
be projected several feet. The majority of the dry ice was dis-
charged in the form of fine particles and the distance was a
maximum of 3 meters (10 ft). For the feasibility tests this
also was held to be adequate although well short of that neces-
sary for a practical system. During these tests single blocks
were fed through and the output collected. Weighing the output
showed an average recovery of 75 percent.
The unit, comprised of the crusher/snow blower combination,
is shown in two views in Figure 8.
Preliminary Field Tests
To further evaluate the unit assembled in the prior phase,
a field test was conducted. The test parameters were based
upon an enlargement of the 9.3 m2 (100 ft2) subscale tests
previously conducted.
The test impoundment was a 46 m2 (500 ft2)steel pan, lined
with polyethylene. Due to environmental decisions by local
authorities, the test had to be conducted in a remote area even
though the only consideration outside of the immediate test area
was the odor. In retrospect this resulted in the test being per-
formed in terrain and under conditions that might be more typical
of an actual spill site.
A diagram of the test impoundment showing the location of
gas samplers and the wind direction is given in Figure 9.
To conduct the test 1630 kg (3000 Ibs) of ether were poured
into the test impoundment. Air samples were taken to charac-
terize the vapor concentration at five points around the spill
impoundment area (Figure 9 ) during unrestricted evaporation be-
fore the application of dry ice. The samplers draw 1 liter
(0.26 gallon) per minute through a 250 ml (0.07 gallon) sample
bottle. The samplers were operated for 5 minutes to insure a
representative sample. At that time they were sealed off and
the entrapped material analyzed by gas chromatography.
After the first set of gas samples were taken, dry ice
particles were added to the spill using the crusher-blower. The
crusher was initially fed the commercially available 23 kg (50 lb]
57
-------
• ' .-• "'•'&'•*•''- 'jiv* v
" i.. '&>' - •- •
.^ y~'-. •• ,»x...
Figure 8. Crusher/snow blower combination
58
-------
6.1m (20 ft)-
LO
C\l_
E
CD
Positions of sampling tubes
labelled 1 through 5
Wind predominantly from east
Area 7.6m x 6.1m (25 ft x 20 ft)
1C
>~ NORTH
04
Figure 9. Diagram of spill impoundment area
59
-------
dry .ice blocks. As noted earli. ,•, the crusher capacity is well
in excess of that of the snow biower, and the intermittent feed-
ing of the dry ice tended to stall the snow blower. To provide
a more uniform feed, the blocks were manually broken into smaller
pieces which were then fed to the crusher.
The short projection distance of the snow blower along with
wind changes forced excessive movement and manipulation of the
distribution system. The nature of the ground made this action
more difficult than had been anticipated. The low oxygen created
by the C02 vapor caused stalling of the gasoline driven engine
of the blower.
The mating of the two components had used a simple design.
It was inadequate for the conditions encountered and the linkages
failed. During the latter half of the test, the crushed dry ice
had to be applied to the spill manually.
The test was carried to completion even with these diffi-
culties since the test data, spill temperature and vapor concen-
tration measurements were felt to be meaningful. Seventy four
blocks of dry ice were added (1680 kg [3700 Ibs]) to the spill
over a 35 minute period. Assuming an efficiency of 75 percent
per prior measurements the effective application was 1270 kg
(2800 Ibs). Weather conditions at the site were clear with a
light variable wind and an ambient temperature of 27°C (80°F).
The temperature of the spill pool was rapidly decreased
from 20°C (68°F) to -63°C (-81°F). The lowest spill temperature
recorded was -67°C (-89°F) .
The vapor samples taken at various points around the spill
gave significant values for four positions both before and after
the dry ice application. Sample 2 was lost in the pretest
sequences and Sample 1 was lost in the post test sequence. The
results are given in Table 13. Samples 1 and 2 were taken down-
wind of the spill and Samples 3, 4 and 5 were taken upwind as
shown in Figure 9. The upwind-downwind status is somewhat
questionable since the wind was light and variable and all of
the sample ports were 30.5 cm (12 in.) above the spill surface.
The reduction in the vapor concentration brought about by
cooling the spill was significant, but the lowest diethylene
concentration found was still about an order of magnitude
greater than the TLV. Although the test could be considered
successful in terms of the ability of dry ice to reduce the
vapor hazard from the diethyl ether, it was a failure as far as
the equipment design and performance was concerned.
EQUIPMENT REDESIGN
The analysis of the preliminary field test results con-
cluded that the crusher portion of the equipment was satisfactory.
60
-------
TABLE 13. DIETHYL ETHER VAPOR CONCENTRATION BEFORE
AND AFTER APPLICATION OF DRY ICE
S a m p 1 e
Location No.
( see Fig. 9 ) Before After
1 13.9% (139,000 ppm) lost cap
2 lostcap 0.5% (5,000 ppm)
3 11.7% (117,000 ppm) 0.9% (9,000 ppm)
4 8.3% (83,000 ppm) 0.3% (3,000 ppm)
5 40.2% (402,000 ppm) 0.8% (8,000 ppm)
61
-------
Some optimization might be possile with tine size, shape,
spacing, rotation, etc., but the existing design appeared ade-
quate for the objectives of this program.
Th.e dispensing portion of the unit was the problem area.
Reconsideration was given to the two alternatives, mechanical
conveying and pneumatic conveying. Mechanical conveying was
ruled out. Its projection problems were considered to be worse
than those encountered with the blower and of sufficient magni-
tude that they were not offset by a potential of uniform dis-
tribution of dry ice independent of particle size range.
This left pneumatic delivery as the only viable alternative
The existing technology would indicate that such delivery is
feasible and that commercial equipment existed which could be
adapted to this application. The drawback came from anticipated
sublimation losses due to high air to particle volume ratios and
long flow paths in hoses. In the absence of any other viable
alternative, it was decided to test a pneumatic delivery system
in combination with the crusher.
Selection and Adaptation of a Pneumatic Conveyor
As part of its mining equipment line, MSA markets a pneu-
matic conveyor system designed for crushed limestone. This unit.
termed a "rockduster" consists of a blower, a hopper, an auger
feed screw and a hose discharge. It is driven by an electric
motor in an explosion-proof housing meeting Bureau of Mines
standards.
A small unit with a capacity of 540 kg (1200 Ibs) per
minute throughput was available. In the interest of time as
well as cost along with the availability of engineers familiar
with the unit and spare parts and optional equipment, a de-
cision was reached to use this unit for the evaluation of pneu-
matic conveying.
Some investigative tests were run by abutting the two units.
In the first run about 68 kg (150 Ibs) of dry ice was fed
through the crusher-duster. About 65 percent of the charged
material was delivered as particles. This was a greater sub-
limation loss than with prior combinations but was considered
to be acceptable.
Using 30.5 m (100 ft) of hose it was possible to reach the
far side of the 7.6 m (25 ft) wide test pit from the near side.
Particulate distribution was uneven as anticipated ranging from
0.32 cm to 2.54 cm (1/8 in. to 1 in.) in diameter. Only the
larger dry ice particles could be projected to the far side.
As with the snow blower, the capacity of the crusher is
significantly greater than the capacity of the pneumatic unit.
62
-------
Excessive feed through the crusher results in a build up of dry
ice particles in the hopper. The particles bridge and weld,
shutting off feed to the auger at the bottom of the hopper. This
bridging is easily broken up mechanically but it is not an accept-
able situation.
Several modifications were made in the rockduster in an
attempt to alleviate the bridging. The sides of the hopper were
made as smooth as possible but the reverse slope of the walls,
the main factor, could not be readily changed. Some air from
the blower was diverted back into the hopper to create agitation
in that area. It was beneficial but not a total answer.
The major modification was increasing the size of the auger
for a faster solids throughput. This provided some improvement
of throughput but resulted in some cases in plugging at the
nozzle-auger interface. This appeared to be due to the decreasing
distance between flights along the length of the auger. This
tended to compact the solids as they move through the auger
section. Because of the nature of the dry ice, the particles
self weld and agglomerate.
The size of the auger is 42.9 cm (16 7/8 in.) long with the
flights in the hopper being 8.9 cm (3 1/2 in.) diameter de-
creasing to 4.8 cm (1 7/8 in.) diameter in the feed tube ex-
tending off the hopper. Plugging resulted because the 4.8 cm
(1 7/8 in.) diameter section had flights decreasing from 5.1 cm
(2 in.) pitch spacing to 3.8 cm (1 1/2 in.) spacing. Another
auger of the same size was installed with 5.1 cm (2 in.) pitch
spacing on all the flights. This eliminated the compacting
effect of the decreasing pitch of the flights and the plugging
at the end of the auger.
To minimize back pressure the standard 3.8 cm (1 1/2 in.)
hose was replaced with a 5.1 cm (2 in.) inside diameter hose.
Broadcasting distance was not appreciably affected and large
particles were discharged a distance of 15-23 meters (50-75 ft).
To minimize fines production, crusher speed was reduced from
about 900 rpm to 600 rpm by pulley wheel change and an increase
in the size of the integral screen. These had some influence
but it was small. Any significant improvement in the operating
characteristics appeared to lie with major changes in crusher
and/or conveyor design.
After completing modifications, the unit was operated con-
tinuously for approximately 19-20 minutes with a throughput of
260 kg (575 Ib) dry ice or a rate of about 13.6 kg (30 lb)/min-
ute. Dry ice blocks used were 25.4 cm (10 in.) square by about
2.54 cm (1 in.) thick (2.40 kg [5.3 lb]/ea) fed at a rate of one
block about each 10 seconds. Measurement of dry ice input to dry
ice output from the discharge hose indicated about a 50 percent
conversion. The basic limitation was a dry ice feed of about
63
-------
13.6kg (30 Ibs) per minute. For the field demonstration to be
conducted, this was deemed adequate.
Two further modifications were made in this combination of
units which had no relation to system capacity or efficiency.
One modification mated the two units into a single entity and
the other unified the power source.
The rockduster comes mounted on a three-wheeled frame. To
make the dry ice distribution system, the crusher was removed
from its wheeled frame and mounted above the hopper of the rock-
duster so that both units were not supported by the three-wheeled
frame.
The rockduster is driven by an explosion-proof electric
motor but the crusher has a gasoline engine. The engine can be
compromised by the presence of C02 vapor and it could be a source
of ignition of flammable material spills such as the diethyl
ether test spill liquid. To simplify the system, the gasoline
engine was replaced with an explosion-proof electric motor. Both
motors could be driven by a remote diesel/electric generator.
The final unit as deployed for the field demonstration is
shown in three views in Figure 10, with an isometric schematic
drawing shown in Figure 11.
64
-------
hi g u r e 10. Field demonstration un 11
65
-------
HOPPER
GRINDER
HOSE. 15.2 m
(50 ft)
GRINDER STARTER
5HP. 460 VAC. 3 0
CRUSHED DRY
ICE DISCHARGE
BLOWER & AUGER
MOTOR STARTER
5HP, 460 VAC. 3 (J
Figure 11. Dry ice distributor.
66
-------
SECTION 8
FIELD DEMONSTRATION
The final phase of the program was the conduct of a demon-
stration of the feasibility of cooling as a technique to control
the vapor hazard from spilled volatile chemicals. This demon-
stration was designed to use the crusher-pneumatic unit evolved
in prior phases of the program.
DEMONSTRATION TEST DESIGN
The final demonstration test design paralleled that pre-
viously used to evaluate the crusher-snow blower combination.
The test involved a spill area of 46.5 m2 (500 ftz) in the form
of a polyethylene-lined steel pan 7.6 x 6.1 x 0.3 m (25 x 20 x
1 ft) deep. The impoundment was set up in the same manner and
same area of the prior field test.
Diethylether was chosen as the spill simulant for reasons
previously described; it is a highly volatile, essentially non-
toxic and basically nonwater soluble liquid. Seven hundred
fifty-seven liters .(200 gallons) of ether was set as the spill
volume. Eight hundred sixteen kilograms (1800 Ibs) of dry ice
coolant in the form of 25.4 x 25.4 x 2.54 cm (10 x 10 x 1 in.)
slabs was available for the test.
The spill site was monitored at five points around the
periphery of the test pit using gas samplers. The samplers draw
one liter (0.26 gallon) per minute through a 250 ml (0.066 gallon
sample bottle. Samples were designed to be taken after charging
the ether to the test pit and again after maximum temperature re-
duction was achieved.
In addition to gas samples, spill temperature measurements
were also planned. Weather instruments were incorporated into
the test plan to measure wind direction and speed along with
ambient temperatures.
FIELD TEST OPERATIONS
The field test was conducted on a clear day with an ambient
temperature of 21°C (69.8°F). Wind speed averaged 32.2 kph (20
mph). Due to unevenness of the test site, difficulty was ex-
perienced in leveling the test pan. This caused an uneven dis-
67
-------
tribution of the ether in the tray. The ether was restricted to
the upwind end of the tray covering an area 7.6 x 3.05 meters
(25 x 10 ft). This caused the sensors on the downwind side of
the tray to be about 3.05 meters (10 ft) away from the edge of
the spill rather than at the edge.
After initial vapor concentration samples were taken, dry
ice was applied to the liquid in the tray using the crusher-
pneumatic conveyor system developed during the program. The
crusher-pneumatic system was located about 25 meters (80 ft) up-
wind of the spill area. A diesel generator set located about
152 meters (500 ft) from the dry ice unit provided power to the
two electric motors.
Thirty and a half meters (100 ft) of hose was connected to
the discharge port of the pneumatic unit. Dry ice blocks were
fed to the crusher at an average rate of 13.6 kg (30 Ibs) per
minute. The crushed dry ice particles were discharged through
the hose onto the spill surfaces using 2.3 m 3 / m i n (80 ftVmin)
of air at 31.5 m/sec (100 ft/sec) using a 3.8 cm (1.5 in.)
diameter hose. Preliminary tests had given an application effic-
iency of 50 percent for a 6.8 kg (15 Ibs) per minute application
of dry ice.
The temperature of the spill pool was monitored during dry
ice application and for an extended period after application
was terminated. A maximum temperature reduction of -65°C (-85°F)
was achieved after 20 minutes of discharge. The method of appli-
cation used to distribute the C02 over the spill surface required
continuous movement of the discharge across the spill surface.
The maximum of -65°C (- 8 5 ° F) was achieved during the time of
direct dry ice application to that area where temperature was
being monitored. As the discharge is moved over the surface of
the spill the time interval between discharge at a specific
point and the return of the discharge to that point was of the
order of two to three minutes. In that interval temperature
variations of as much as 10°C (18°F) were observed. Thus, the
avarage temperature at any point in the spill pool was about
-60°C (-76°F). Dry ice was discharged for 30 minutes, consuming
408 kg (900 Ibs). Because of sublimation the amount actually
applied to the spill is estimated at 50 percent of that quantity
or 204 kg (450 1bs).
When dry ice discharge was stopped the temperature of the
pool increased within 30 minutes to -30°C (-2 2 ° F). The rate of
rise slowed with time as would be expected. It leveled off to a
slow rate of rise above -20°C (4°F) and was still below -10°C
(14°F) after two hours. A plot of the temperature-time profile
is given in Figure 12.
Ether vapor concentrations were measured before dry ice dis-
charge was initiated and again after 30 minutes of application.
68
-------
o
o
a.
s-
O)
_c
-I-J
O)
•4-J
OJ
"O
Qj
c
o
o>
69
-------
The values obtained vary as a result of location and wind effects.
Thus the observed vapor concentrations provide only a relative
measureofeffectiveness.
Sample stations 1 and 2 were located at the upwind edge of
the spill pool and locations 3, 4 and 5 were at the downwind
edge of the impoundment which was approximately 3 meters (10 ft)
downwind from the edge of the spill pool as previously noted.
The analysis of the samples gave the following results in
ppm:
Sample Sta t i_on Before Dry Ice After Dry I ce
1 6,400 80
2 10,800 350
3 3,800 80
4 2,100 430
5 2,300 540
The scatter is too great for any absolute quantitative
analysis but it is clear that the cooling has markedly reduced
the rate of vapor release fromthe ether. The sublimation of
the dry ice causes surface agitation which probably exaggerates
t he vapor release. The temperature reduction decreased the
equilibrium vapor pressure from 535 torr at20°C (68°F) to 4 torr at
-60°C (-75°F).
SPILL SCENE CLEAN UP
The final task of the program concerned an appraisal of the
clean up procedures which would be appropriate for the cooled
fluid and those interferences which might occur. For the sit-
uation where the cooled spill remains liquid, such as with the
diethyl ether, the benefit of the coolant is to slow materials
loss to the environment through vaporization. This does allow
evaporation of the spill at a rate which would be slow enough
that the TLV would not be exceeded for the life of the spill or
would be exceeded only at the immediate spill area.
The number of materials which might be treated in the above
manner is probably small. It is also probable that only small
spills could be allowed to evaporate although spills in remote
or difficulty accessible areas might be best dissipated through
slow evaporation.
The obvious course of action is to pick up the spill or
chemically treat it in place. The cooling process should not in-
terfere with pickup by vacuum trucks, the EPA Emergency Collection
System or similar procedures using collection tanks or bags. The
70
-------
dry ice particles that might be entrained in the spill are small
enough to pass through hoses and valves. C02 release within the
container may necessitate procedures for venting over pressure
and/or limit the volume which could be drawn into a given con-
tain e r s i z e .
Chemical treatment needs to be addressed on an individual
basis. Chemical reactions which are exothermic would be counter
productive to the cooling process. Such action would need to be
evaluated with reference to the benefits of chemical treatment.
Since C02 is not highly soluble in most of the materials to
which the cooling process might be applied, there should be
little retention to interfere with ultimate disposal procedures
such as containment or incineration.
If coolant application would convert the liquid to a solid,
clean up procedures would change. If the temperature of the
spilled material is decreased well below its freezing point, it
would be possible to pick up the material mechanically as a
solid. Except for small spills where shoveling by individuals
could be employed, mechanical pick up may be problematic. Heavy
equipment is not designed to skim materials nor to work rapidly.
Solid to solid heat transfer is considerably slower than
solid to liquid and in practice it may be difficult to subcool
the spill very far. If subcooling is difficult, cooling could
be used to hold the spill in a frozen condition until clean up
equipment adaptable to a liquid is in place. At that time,
cooling would be discontinued. As the spill warms and melts the
cold liquid can be picked up and transferred to temporary con-
tainment or treated chemically if that is advisable.
In terms of clean up and/or recovery of a spilled chemical,
the presence of the coolant does not appear to offer any serious
interference. It must be constantly recognized, however, that
the carbon dioxide can pose a life hazard to personnel in the
spill area due to oxygen displacement. All personnel working
directly on the spill would probably require some sort of
breathing protection due to the nature of the spill. Self-con-
tained breathing apparatus would be preferred to canister gas
masks. Canister units cannot protect against the aphyxiation
hazard of high C02-low oxygen atmospheres. The spill area at-
mosphere should be continuous monitored. The spill area must
always be considered dangerous. Impairment of work skills or
conscious processes due to reduced oxygen levels can have dra-
matic consequences .
EVALUATION OF COOLING AS A MECHANISM OF VAPOR HAZARD CONTROL
The work which has been conducted during this project has
well established the feasibility of cooling to reduce the vapor
71
-------
hazard from spilled volatile chemicals. There are some areas
where improvements could be made in the dispensing equipment but
these are not beyond technology. The question which remains un-
answered concerns the practicability of the coolant system. A
series of questions need to be answered in assessing practic-
ability:
• To how many spills would the cooling
technique be applicable?
• Of these, what number could be treated
by an alternate procedure?
• Do the alternate procedures give commensurate
results or do they possess advantages in terms
of logistics, deployment or application which
would justify some reduction in efficiency?
The answers to these questions allows determination of the
number of spills for which cooling would be chosen as the most
effective procedure. This gives a base to establish the de-
sirability of investment in the equipment necessary to effectively
utilize the coolant technology. The technology is completely de-
pendent on the ready availability of both coolant and equipment
and the ability to deploy and use it at the spill site. Until
these factors are established, any evaluation will be incomplete.
72
-------
REFERENCES
Bell, G.E., "A Refinement of the Heat Balance Integral Method
Applied to a Melting Problem", J. Heat Mass Transfer 21,
1357 (1978).
Board, S.J., Clare, A.J., Duffey, R.B., Hall, R.S., and Poole,
D.H., "An Experimental Study of Energy Transfer Processes
Relevant to Thermal Explosions", J. Heat Mass Transfer 14,
1631-1641 (1971).
Brown, D., et al, "Techniques for Handling Landborne Spills of
Volumetric Hazardous Substances", Battelle Columbus Lab-
oratories, Contract 68-02-1323, Task 52.
Drake, E.M., and Reid, R.C., "How LNG Boils on Soils", Hy^ro.-
carbon Proc. 54- (5), 191-194 (1975).
Greer, J.S., "Feasibility Study of Response Techniques for Dis-
charges of Hazardous Chemicals that Float on Water", Report
No. CG-D-56-77, Contract DOT-CG-51870-A (1976).
Gross, S.S. and Hiltz, R.H., "Evaluation/Development of Foams
for Mitigating Air Pollution From Hazardous Spills'1, EPA
Contract 68-03-2478 (1980).
Kosky, P.G., and Lyon, D.N., "Pool Boiling Heat Transfer to
Cryogenic Liquids: I. Nucleate Regime Data and a Test of
Some Nucleate Boiling Correlations", AIChE Journal 14 (3),
372-377 (1968).
Slade, D.H., Editor, Meteorology and Atomic Energy , U.S . A . E . C .
Div. of Tech. Info~Report No. TID 2419~0 (July 1968).
73
-------
TECH.V' vL REPORT DATA
i Please reed Instruct: > uii tne reverse before cornptcr.n^i
1. REPORT NC. 2.
4. TITLE AND SUBTITLE
MODIFICATION OF SPILL FACTORS AFFECTING AIR POLLUTION
Vol. ] - An Evaluation of Cooling as a Vapor Mitigation
Procedure for Spilled Volatile Chemicals
7. AUThORlS)
J.S. Greer, S.S. Gross, R.H. Hiltz, and M.J. McGoff
9 PERFORMING ORGANIZATION NAME AND ADDRESS
MSA RESEARCH CORPORATION
Division of Mine Safety Appliances Company
Evans City, Pennsylvania 16033
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati , Ohio 45268
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
May I 981
5. PERFORMING ORGANIZATION CODE
3. PERFORMING ORGANIZATION REPORT "C
80-197
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-03-2648
13. TYPE OF REPORT AND PSRIOO COVSREC
Final - Sep 1978 - Dec 1980
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
Monitoring Agency: Rockwell International, Newbury Park, California 91320
15. ABSTRACT
Soilled cnemicals that pose a nazard to the land and water ecosystems can also provide a
significant vapor nazard. Altnougn the vapors released by such chemicals -"ay ultimately be
dispersed in the environment with little long-tern effects, they do pose a hazard tc life a
property downwind of the spill.
and
3
1
Among tne vapor amelioration techniques that hav
re of a spill and reduce its _._
feasibility studyof tnat mechanism.
,, ,__ ,___ ...e been considered 1s the use of a coolart to
ower the temperature of a spill and reduce its equilibrium vapor pressure. This program has
een conducted as a
filter ice dees not cool sufficiently. Liquid nitrogen and carbon dioxide requ're large
quantities of material, and produce a dense obscuring cloud that has sore adverse indication
3ry ice avoids the problems presented by the other coolants and is readily available at a
-easonable cost, out a method is reouired for crushing and distributing tr.e dry ic = tc the
spill. A prototype unit was developed consisting of a crusher and a pneumatic conveyor to
perfjr-i these functions.
A pool of diethyl ether with 2.23 -n (250 sq ft) of surface was cooled to -5C°C i-7:'', using
408 Kg (900 Ib) of dry ice fed at a rate of 13.6 kg (30 Ib) per minute. A measurable
reduction in downwind vapor concentration was realized. POO! teiroerature was still below
-!0=C (14°F) 2 hours after dry ice discharge was terminated.
This program has established feasibility of tne mechanism, but aaditional work is necessary
to establish practicality, define materials to which cooling is applicable, and ootim'ze the
dispensing equipment.
5
a
j
j
e
j
Q
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Evaporation Control, Hazardous Materials,
Chemical Spills
IS. DISTRIBUTION STATEMENT
Release to pub! ic
b. IDENTIFIERS/OPEN ENDED TERMS
Spill Control, Hazard-
ous Material s Spil 1 s ,
Coolants, Vapor Pressure
Reduction, Diethylether ,
Dry Ice Distribution
System
19. SECURITY CLASS (Tins Report/
UNCLASSIFIED
20. SECURITY CLASS /This pa%ej
UNCLASSIFIED
c. COSATl FiiriJ/Group
11/07 Coolants
07/04
06/20
21 . NO. OF PAGES
79
22. PRICE
EPA Form 2220-1 (R«». .1-77)
PREVIOUS EOITION IS OBSOLETE
74
------- |