Evaporation Rate of Volatile Liquids. Final Report, Second Edition
Pace Labs., Inc., Minneapolis, MN
Prepared for:
Environmental Protection Agency, Washington, DC
11 Dec 89
V
C
-------
so;n • i oi
REPORT DOCUMENTATION
I, REPORT NO.
i.
PAGE
EPA/744-R-92-001
4. Title and Subtitle
EvapocatLon Rate o£ Volatile Liquids
Final Report
... Second Edition .
7.
Karl_0. Brauo/.KnowIton J._Caplan
1. Performing Organl* ation Name and Addreti
Pace Laboratories, Inc.
1710 Douglas Drive North
Minneapolis, Mti 55422
12. Spontorlng Grgftnlxation Nimi and Addrett
U.S. Environmental Protection Agency
Office of Pollution Prevention and Toxics
401 M St. SW (TS-779)
... .Washington, _DC 20460
IS. luppltmmtiry Noln
Developed under subcontract to PEI Associates/ Inc.
5 PB9 2-232305
5. Report Data
12/11/89 (Date of Issue)
ft.
B. Performing Organiiation Rapt. No.
10. Project/Talk/Work Unit No.
11. Contract(C) or Gr»nt(G) No
II. TjrP* of h.aori & Period Covered
Final Report
14.
1ft. Abstract {l imit: 200 word*)
This research was designed to cetarrnine whether the evaporation rate of a volatile
liquid could be adequately predic'-e from its cortmon "handbook" properties over a narrow
range of environmental condition^ ,iot to refine or amplify the classic theories of heat
transfer and mass transfer. - An ixperimental apparatus to measure the evaporation rate
under controlled conditions and nearly ideal presentation of the evaporation surface to
the controlled airflow was developed. The evaporation rate for base set of twelve
chemicals from four classes of organic liquids was measured. The relevant physical
properties of the liquids (e.g. vapor pressure/ surface tension/ viscosity, latent heat
of evaporation, etc.) were correlated with the observed evaporation rate.controlled
variables of temperature and airflow to determine the relationship between these variables.
Additional testing of water, n-butyl acetate, and three examples of low 'vapor pressure
alcohols was then conducted. Predictive equations for the low vapor pressure alcohols
were developed and presented for comparison with the base set of equations. The results
were compared with other referenced methods. The error of prediction of the empirical
equations for the range of conditions was generally less than + 12 percent, or generally
less than + 20 percent, depending on which predictive equation was used.
17. Document An»{y*.'% •, Devrrtpton
b. Identifier*/Ope r»-Cr»d«d Tirmi
c, COSATt field/Group
it. Availability luttnwm
». WeurMy Clatt (Th't Report)
21. Na, ol
Release Unlimited
Jlncla®if.ieC...„... _
M. kcurlty Clati CThlt Pa»al
Unclassified
n. Prtca
Sh fotfucftons en ft even#
OPTIONAL rOftM <72 <4-771
Ifsrmcrly WTIWi)
#1 Commtrc^
-------
This Page Intentionally Blank
-------
pace
PWOfUltOKAl ANALYTICAL CHtMifTMt « |M
-------
TABLE Of CONTENTS
EXECUTIVE SUMMARY
BACKGROUND AND RAi IOHALE
OBJECTIVE
1
2
4
APPARATUS S
Figure 1 - Schematic <.' System 7
Figure 2 - Tut Chimber 8
Figure 3 - Test Pari 9
Figure 4, 5. 6 - Photographs or Apparatus 10-12
EXPERIMENTAL PROTOCOL 13
Table 1 - Liquids Tested 13
Add1 Monal Test Work. 13
TEST PROCEDURE >«
EXPERIMENTAL MCTHOOS AND QUALITY ASSURANCE 15
TEST RESULTS 16
Tables II and III - Summary of Predictive Equations 17-18
"Still Air" Evaporation Rate Qualitative Discussion 20
Low Vapor Pressure Discussion 2
COMPARISON HI TH OTHER METHOOS 24
Table IV - ASTH Relative Rate Osta vs. IHE Data 26
Table V - Gray vs. IHE Data 27
Table VI - Hanna and Drlvas vs. IHE Data 29
Table VII - Hanna and Drlvas (Incorrect VP) vs. IHE Data 31
Table VIII - O'Brien & Stutzman vs. IHE Data 32
Figure 7 - Hade Equations vs. IHE Data 34
Table IX - EPA.'MacKay Equation vs IHE Data 37
Table X - Hu/Schroy Model vs. IHE Data 38
Table XI
Summary of Method Comparisons 40
PROBLEMS ENCOUNTERED 41
FUTURE WORK 42
REFERENCES 43-44
APPENDIX A - Data Presentation Ubles A1-A16
APPENDIX B - Air Velocity, Air and Liquid Temperature Graphs B1-B16
APPENDIX C - Data Correlation Graphs C1-C14
APPENDIX D - RELATIVE RATE EVAPORATION OATA
Appendix Dt - Relative Rate Evaporation Rate Paper by L.D. Wilson Dl-1 - 3
Appendix D2 - ASTM Method D3539-76 (Reapproved 1981 ) 02-1 - 11
APPENDIX E - Comnwntj on Paper by D.C. Gray E1-E18
APPENDIX F - Excerpt from Vapor Cloud Dispersion Models by Hanna... F1-F7
and DrWaj
APPEN0IX G - PEER REVIEW C0MHENT AND AUTHOR RESPONSE
Appendix GIA - Popendorf Peer Review Comments G1A-1 -
-------
EVAPORATION RATE OF VOLAIILE LIQUIDS
£XEOtfm_SUMHAR*
The evaporation rate of volatile liquids, as measured In a test duct with the
liquid surface flush with the bottom of the duct, can be predicted by an
equation of the form:
ER - K(MW)m(VP)n(V)P (Eq. 1)
Hhere:
ER - Evaporation Rate Ib/hr
K. m,n,p - Constant and Exponents
MW - Molecular Weight of liquid
VP - Liquid Vapor Pressure, at Liquid Temp., Inches Hg.
V ~ Alr Veloc1ty, ft/mln
The partial vapor pressure of the evaporating liquid In the Incoming air Is zero
excopt for water. Therefore, the equation Is shown to Include only VP, the
vapor pressure of the liquid at the liquid temperature. In the case of water,
the term (VP1-VP4) is used in place of VP. VPL is the vapor pressure of the
liquid (water) at the temperature of the liquid (water) and VPA 1s the partial
pressure of the liquid (water) In the incoming air.
If the specific equation for each class of chemicals and range of velocities Is
used, the error of prediction Is generally less than ± 12 percent. If the
equation representing all of the chemicals tested Is used, error Is generally
less than ± 20 percent.
This form of equation represents the range of air velocities from 100 feet per
minute (fpm) to 1,500 fpm (0.5 to 8 meters per sec., m/s), and air temperatures
ranging from 40*F to 140*F (4.4 to 60* C). The classes of compounds tested were
aliphatic hydrocarbons, aromatic hydrocarbons and aliphatic alcohols and
ketones. The equations presented for each chemical class were developed from
experimental work conducted on a base set of three examples from each class.
Over this range of conditions, a strong correlation between evaporation rate and
the other parameters studied, I.e. Reynold's Number, surface tension, diffusion
coefficient, latent heat of vaporization, and viscosity was not found. The use
of the Schmidt Number (which Includes air viscosity and diffusion coefficient as
parameters). In place of the molecular weight, yielded a slightly lower accuracy
of prediction.
The values of K, m, n and p In Equation 1. vary somewhat with the class of
compounds. A discontinuity was found between 100 fpm (0.5 m/s) velocity and 500
fpm (2.5 m/s) and higher velocities. At 100 fpm (0.5 m/s), the flow regime in
the duct was In the transition range between laminar and turbulent, whereas at
500 fpm (2.5 m/s) or more the flow regime was turbulent. The discontinuity Is
likely a result of these flow differences.
If a similar relationship can be developed for other classes of volatile
liquids, an estimate of evaporation rate (source strength) can quickly and
easily be made from knowledge of the identity of the liquid, Its temperature
1
J
-------
(and hence vapor pressure), and the local air velocity. The accuracy of such
estimates would be adequate for purposes of PMN procedures, atr pollution
Inventory and In-piant Industrial hygiene activities.
Evaporation rates were compared with those obtained by other Investigators.
Five other methods Involving 8 cases were compared. Three were on average
within lOt of HIE results but with a large range, 2 were significantly higher
and 3 were significantly lower. Comparability between methods are discussed In
detail In th-. .ectlon "Comparison With Other Methods".
Draft copies of this report v/ere furnished to six peer reviewers for comment.
Only two reviewers had substantive comment and these along with the I HE
responses are furnished as an Appendix.
This study was Initiated by Industrial Health Engineering Associates, Inc.
(IHE). PACE Laboratories, Inc. purchased IHE in 1987 and the project was
completed by the IHE Division of PACE. The study was performed for the Chemical
Engineering Branch, OTS-EPA, under subcontract to Pi I Associates, Inc.
fiACKGRQUiiD_AMLEAIl QUALE
Knowledge of the evaporation rate of volatile liquids exposed to air 1s
obviously necessary for quantitative estimation of air contaminant emissions In
a number of typical situations. Such situations range from emission sources In
Industrial manufacturing operations such as open surface tanks, container
filling, and indoor spills; to more general sources such as spills occurring In
transportation and from bulk storage facilities.
The objective of this research was to determine whether the evaporation rate of
a volatile liquid could be adequately predicted from Its common "handbook"
properties over a range of conditions representative of the typical factory
environment. The objective was not an attempt to refine the classical chemical
engineering approach to simultaneous heat-and-mass transfer phenomena.
There are several practical applications of data concerning evaporation rate?.
The Chemical Engineering Branch, OTS-EPA, npeds such data in their legal.y
mandated review of Pre-Manufacture Notification (PMN) for new chemicals. The
information furnished in PMN applications is frequently minimal; but one
category of information that Is reasonably complete Is the physical properties
of the .hemlcal itself. If sufficiently accurate evaporation rates can be
estimated from those properties, then (for volatile liquids) the adequacy of the
PMN permit conditions will be Improved, and possibly at less cost in terms of
professional technical effort. Beth EPA and manufacturer will run less risk of
overly conservative (and unnecessarily costly) or inadequately safe extremes
that could result from emission and exposure estimates made without this
Information (It should be noted that OTS Is legally required to make such
estimates regardless of data quality)
There are several other Important practical uses for better evaporation rate
data.
(1) Manufacturers use techniques similar to those of Chemical Engineering
Branch, OTS in contemplating manufacture of a new chemical and Its use
In customer's facilities.
(2) Fpldemlology studies attempt to reconstruct "historical exposures" of
exposed persons for situations In which there Is little or no direct
exposure data. However, the Identity of the chemicals, nature of
operations, etc. can be determined and "PMN type" estimates made using
reasonably accurate evaporation rate data.
2
-------
(3) Ventilation/Mr Pollution engineers would find good evaporation rate
information useful In designing exhaust ventilation systems and
selecting air pollution control equipment.
(4) Emission Inventory estimates (when used In lieu of stack, sampling)
would be Improved where simple material balance estimates are
1napproprlate.
(5) Emergency Respons«/Sp111 Team efforts would bt made more efficient and
less risky 1f evaporation rate of spilled liquids can be quUkly
estimated, Wearing self-contained breathing apparatus (SCBA)
Introduces an element of risk, to the wearer, espaclally in an emergency
situation, due to the reduced visibility and perception of the local
environment. The wearer's work, proficiency and efficiency Is
significantly reduced. Thus, If vapor concentrations are not
sufficient to require SCBA protection, Its unnecessary use Is an added
unnecessary risk, and cost.
(6) There are many Industrial hygiene applications for exposure estimates
from open surface tanks and other Industrial processes or operations In
the work place.
7) Estimates of vapor clouds from spilled pool liquids would be Improved
by better estimates of evaporation rates. This application has many
ramifications, Involving the general environment and the local
population, Implementation of emergency evacuation, etc.
Presently available evaporation rate data are of two types. One type Is a
tabulation of relative evaporation rates, I.e. the evaporation rate of one
liquid as compared to a "standard" liquid under specified conditions. No
quantitative evaporation rate can be computed from this Information without
Indication of at least one quantified value. Additionally, the test apparatus
used does not represent, nor attempt to represent, the physical environment
encountered In actual use of volatile solvents. A comparison of quantitative
and relative rate Information Is presented In the section "Comparison With Other
Methods."
The other approach is the solution of the simultaneous heat and mass transfer
phenomena represented by evaporation of the liquid. This is a tedious Iterative
computation requiring many assumptions, and frequently leads to highly
questionable results. It Is Impractical for emergency or routine dally use.
For process design work, the solution of simultaneous mass and heat transfer
problems is necessary. Conditions may cover a very wide raroe of temperatures,
pressures, etc. and the final process design includes suitable safety factors.
The time and effort required for such computations Is not a problem.
Many models have been proposed for evaporation rates from spills, as summarized
by Hanna and Drlvas (Ref. 10). In that reference the wide range of disparity,
as much as a factor of 3, between model prediction and evaluation Is described
In Chapter 6. Gray (Ref. 9) has shown a correlation of evaporation rate with
Schmidt number for some types of organic liquids. The correlation requires a
very complex mm'.puUUon of parameters and appears to have several other
deficiencies, as discussed in detail In Appendix E. The results obtained In
this study are compared to t^ese and other referenced methods In the section
"Comparison With Other Methods".
3
5
-------
QBJICim
The objective of this research was to determine whether the evaporation rate of
a volatile liquid could be adequately predicted from Its r.ommon "handbook."
properties over & range of conditions representative of the typical factory
environment. The objective was not an attempt to refine the classical chemical
engineering approach to simultaneous heat-and-mass transfer phenvonena.
There w&re two parts to this ohjectlve. One was to design and build an
experimental apparatus that would permit the measurement of evaporation rate
under controlled conditions and nearly ideal presentation of the evaporation
surface to the controlled airflow. The second objective was to conduct testing
on a base set of three examples from each of f ur classes of organic liquids
(aliphatic, aromatic, alcohols, ketones) to determine whether or not evaporation
rates can be adequately predicted from commonly avalUble physical properties of
the liquid.
It -was realized that under practical circumstances the relationship between the
evaporation surface and the airflow may frequently be far from Ideal. However,
a rigidly defined and reproducible technique was necessary for a determination
of the fundamental properties affecting evaporation, and for potential value as
a standard empirical test. The Initial research was restricted to a few pure
chemicals; It Is recognized that in practical use mixtures must also be
evaluated as well as non-Ideal flow conditions. However, this approach was
adopted as basic, and In the Interest of confining this Initial effort to
relatively modest proportions.
The physical properties of the liquids which presumably affect evaporation rate
such as, vapor pressure, surface tension, viscosity, latent heat of evaporation,
etc. were correlated with the observed evaporation rate and controlled variables
of temperature and airflow to determine the relationship between these variables.
This work has resulted in a reasonably accurate correlation of the type sought
Involving only knowledge of the class of p. compound, the molecular weight,
liquid vapor pressure and air velocity. The correlation covers the range of air
velocities from 100 to 1500 fpm (0.5 to B m/s) and air temperatures from 40°F to
140* F (4.4 to 60°C). These ranyes approximate the conditions where this
Information would normally be applied.
After completion of the base set of experimental work, and the development of the
base set of predictive equations, some additional work was conducted at EPA
request. The additional work Included evaporation testing on water, n-butyl
acetate and three examples of low vapor pressure alcohols (a vapor pressure
range of 1.0 to 0.1 mm Hg at room temperature was designated). The data from
this work Is Included In this report; however, it was not used for development
of the base set of predictive evaporation equations.
The additional work was for comparative and Investigative purposes. For
example, ASTM relative rate evaporation data Is based on n-butyl acetate.
Measurement of the evaporation rate of n-butyl acetate allowed calculation of
relative evaporation rates for the base set of test chemicals. The relative
evaporation rates calculated were compared to ASTM published relative
evaporation rate data for the same chemicals. The results are presented in the
section "Comparison With Other Methods". Low vapor pressure evaporation data
was assembled at EPA request to expand the range of experimental work and for
comparison to the base data. Predictive equations for the low vapor pressure
alcohols were developed and are presented for comparison with the base set of
equations. A low Vapor Pressure Discussion section Is Included.
Obviously, much further work is required for total general application of this
approach, as will be described later.
4
-------
AEEAMIUS
The test apparatus consisted of a test chamber or duct 4 In. high and 8 In.
wide (10.2 x 20.3 cm). Test liquid was held In a specially designed
removable 5.5 In. square I In. deep (14 x 14 x 2.5 cm) test pan. The pan was
designed to be easily removable from the test chamber for cleaning; or In
case a different material of construction was needed for some particular
liquid. Flow stralghteners were provided to secure a uniform flow profile.
Obstructions upstream and downstream of the test pan were kept to a minimum
to avoid turbulent eddy currents. A 1/4 In. (6.4 mm) diameter hot wire
anemometer and an 1/8 1n. (3.2 mm) diameter RTD were placed upstream while an
1/8 In.(3.2 mm) diameter Teflon feed tube, thin RTD temperature leads and the
level control fulcrum covered by a clear aerodynamic canopy were downstream
of the test pan. A temperature controlled base or pan chamber was provided
under the test pan for liquid temperature or base temperature control (in
these tests, the base temperature was controlled to equal the liquid
temperature). Airflow was provided by a remotely located blower. Heating
wa? provided by a hot water coll system and cooling was provided by a chilled
water col' system. An IBM XT computer equipped wtth a custom process control
program utilizing Macmlllan ASYST software was used to control and record
airflow, temperature and evaporation rate, etc.
The objective of the apparatus design was to provide for a uniform and
controlled profile of airflow at controlled temperatures passing over an
exposed liquid surface In such a way that turbulent eddy currents over the
liquid would be minimized; and to permit measurement of the evaporation rate
at equilibrium test conditions.
A schematic of the apparatus is shown In Figures 1, 2, 3 and photographs are
shown In Figures 4, 5 and 6. The pan was designed so that the liquid surface
Is In the same plane as the base of the duct. This eliminates turbulent eddy
currents at the liquid surface. In use, the liquid surface was about 0.2 mm
below the 'op of the pan. rhe streamlined "ripple dam" on the sides and down
stream e.;' jf the exposed liquid was to prevent any ripples, created on the
surface of the liquid by the air velocity, from spreading over the flat
surface of the duct.
Test liquid was fed to the test pan by a gravimetric metering system designed
to maintain the liquid level (± 0.1 mm) and to gravlmetrtcally measure the
feed rate to maintain that level. A float arid lever system as shown In
Figure 2 controlled the liquid level. A flag on the lever arm rode between
light emitting diodes (LEO) and depending on Its position, signaled the
liquid feed valve to operate. Evaporation rate was determined by maintaining
the liquid level In the test pan while monitoring the weight change of a
liquid supply container which was suspended from a sensitive (± 0.1 gram)
load cell. The weight change of the supply container over time (l.e feed
rate of test liquid) divided by the liquid surface area defined the
evaporation rate.
5
7
-------
The base or pan chamber temperature could be controlled. In this work, the
base temperature was controlled to be the same as the liquid temperature In
order to minimize heat transfer between the evaporating liquid and the pan
itself. The objective was to avoid biasing the data concerning simultaneous
heat and mass transfer at the liquid surface. By elimination of heat
transfer between the pan and the base, the alr-llquld interface becomes the
major heat transfer location thus providing a simplified model from which the
parameters affecting evaporation hopefully would become more apparent. Other
base temperature experiments can be done, to simulate spills on hot pavement
or controlled liquid temperatures of open surface tanks, etc.
The operating range generally used was:
AIRFLOH : 100 to 1500 fpm <0.5 to 8 m/s)
AIR TEMPERATURE : 40'F to MOT (4.4 to 60°C)
HUMIDITY : 40*F <4.4'C> dew point or ambient
BASE CHAMBER TEMPERATURE : 40*F to I40*F (4.4 to 60°C), controlled
to liquid temperature
LIQUID TEMPERATURE : Dependent variable
EVAPORATION RATE : Dependent variable
6
-------
FIGURE \
-S5>
-------
FIGURE: Z
- FLovJ
SrtAtC,H7lN£K
R T.O. LE4CS
G/& T£/y)P
-FjEEO
TUS£
Rj .a l & 4 oS
L/Qu/o tiewA
resr /*»/~
¦LEYZL CPNTAOL.
plan v/e^
vJ/'/£>
U^tUNQ-
SCUM3
3
RXD. LCAOJ
6*S£ rewp
¦ r
XEAnoV4&Lt C0K£/<
/r=s^
T
103"
-2_
&43B GMrfe&Z
Hrcv AND CdojLJNC-j
eL£VA TIoa/
A.
c:i
L&.S/&L. COH7X&/L.
MT£ • {tfSUL4TKrV mr
Show Fa* CtlZ/rr
TEST CH$maC:K
I'
-------
FIGURE: S
StCT/ON A-A
T£ 5 T /V? A/
TgTn^j^agaaae——¦——i^—
9
u
-------
4-A Plan view of evaporation pan with
cover removed, (a) pan, (b) liquid
feed tube, (c) level control, (d)
LED sensor, (e) liquid temperature
sensor. Air flow right to left.
4-B Same as 4-A, with cowl over level
control.
Figure 4
10
\l
-------
5-A General view, feed side, (a) load
cell, (b) liquid container, (c)
feed valve. Air flow right to
left.
5-B Same side, elevation view of pan
and base chamber
(a) base chamber
(b) liquid feed valve
Figure 5
1?
-------
, • ¦ ?.v!
mam
6. Control station.
Figure 6
12
-------
EXEiRlMENTAL PROTOCOL
Anticipating that there would be significant differences In evaporation behavior
due to molecular composition, four simple classes Of organic liquid with three
examples of each were Initially studied as shown In Table I. Reagent grade
laboratory analytical chemicals were used.
J.ABLE I
Allohatlc
Hexanes
Heptane
Octane
Aromatic
Benzene
Toluene
Xylenes
Alcohols
Methanol
N-propanol
N-pentanol
Acetone
MEK
2-octanone
The basic experimental protocol Involved measuring liquid evaporation rate at
three conditions of air velocity, and three conditions of air temperature, ihe
base chamber was controlled to the same temperature that the liquid reached by
evaporative cooling, In order to prevent or at least minimize heat transfer
to/from the liquid except at the alr-llquld Interface. In general, the majority
of test data was collected within the following test ranges:
AIR. V[LOCIM.SMLS L.
100 (0.5)
500 (2.5)
1000 (5.0)
AIR TEMPERATURE. *F CO
45 (7.2)
70 (21.1)
100 (37.8)
Exceptions to this Involved test liquids which had either a very slow or very
fast evaporation rate. For example, the evaporation rate of n-pentanol and
2-octanone was slow and therefore both air temperature and velocity
modifications were made. Fast evaporators such as acetone and hexane
required testing at increased air temperatures because evaporative cooling
caused the liquid temperature to drop below the operational range of the
equipment (I.e. the base chamber could not be cooled enough to maintain a
base temperature equal to the liquid temperature).
Add 111 ona 1 Test Work,
Additional testing was later conducted for water, n-butyl acetate and three
low vapor pressure alcohols (1-hexanol, 1 -heptanol and 2-octanol). The
specific alcohols were selected because they are generally within the vapor
pressure range designated by the EPA (1.0 to 0.1 mm Hg at room temperature)
and because they are relatively Inexpensive, readily available, relatively
nontoxic and safe to handle, and are members of a class of chemicals which
had already been studied. The same bask test conditions were used for this
test work as for the base set of data; however, the low end temperature
(45°F, 7.2"C) was dropped and an additional high end temperature (120"F,
48.9°C) was added. This was done because pf the long test times required to
measure the evaporation rate for the low vapor pressure alcohols.
13
IS
-------
TESL_ERQCED.URE
For clarity In the discussion, the following terms are defined:
A "run" - the determination of evaporation rate of a liquid
under one defined environmental condition.
A "set" - the total runs at all planned conditions for one
comical substance.
A "batch" - the sets of data representing all three examples of one
mo,ftcular class of compounds.
Initially, It had been planned to conduct all runs of the set for a given
chemical In a single series. It was felt this would reduce set up time required
for changing test liquids, etc. However, 1n operation It was found to take less
time to change test liquids than to change and stabilize test conditions (air
temperature and velocity). The computerized temperature control system was not
successful and much manual intervention was required for successful operation.
These problems prevented a more systematic approach and Inhibited concurrent
data analysis.
The typical test procedure was as follows:
1) Drain and clean test pan from previous run.
2) Drain test liquid supply container of previously tested liquid.
3) Flush test liquid supply container and feed lines two times with a small
quantity of the next liquid to be tested.
4) Fill the test liquid container with test liquid.
5) Set the fulcrum/lever/float level control In place.
6) Start blower.
7) Start hot and chilled water systems as required for test condition.
8) Set and modify temperature controls as required until stable
temperatures are achieved (± 1* F, 0.55" C).
9) Initialize the fulcrum/lever/float level control, adjust as necessary.
10) Start data recording.
11) Run until steady state evaporation Is obtained. Steady state
evaporation was defined when the last reading varied less than 1 OX from
the average of the previous ten consecutive evaporation rate readings.
12) Stop data compilation.
13) Next run - go to 1.
The time required to complete a single test run varied from approximately one
half hour to over three hours depending on the test liquid, test conditions
and operational problems with temperature stabilization and level control
Initialization.
14
it
-------
£X££MMtMIAL_M£IJlQD5_AMD-i3UALirL_ASS!iRANCE
The following Instrumentation and calibration procedures were used for airflow,
temperature and evaporation rate measurements.
AlfL^ELQCiil• The duct centerllne air velocity was continuously monitored with
a Kurz Model 415-4 anemometer. A National Burea* of Standards (NBS) traceable
Dwyer Model 1430 Mlcrotector (point gauge) was used with a standard p1 tot tube
to measure the velocity head and verify the anemometer. The centerllne duct
velocity as measured with the anemometer was correlated with the average duct
velocity measured by pi tot tube traverse < E'A Methods 1 and 2). Calibration or
flow checks of this sort were made over the flow range of the apparatus prior to
start of testing and again near the end of testing. The flow profile was In all
cases quite uniform with less than 10% disagreement between the anemometer and
the average velocity.
Test air and test liquid temperatures were continuously monitored
using resistance temperature detection (RTD) assemblies. A National Bureau of
Standards thermometer was used over the range of test temperatures once prior to
the start of testing and again near the end of testing. Arfdltlonally, single
point dally checks were made with the NBS thermometer to verify proper RTD
function. Correlation between the NBS thermometer and the RTD's was 1n general
within ± 1 *F (±0.55*C).
LOAD-CELL: The test liquid evaporation rate was determined gravlmetrl cal ly. A
computerized level control was used to maintain a constant level of liquid 1n
the test pan. As liquid evaporated from the test pan, It was replenished by a
liquid supply container suspended on a load cell through a feed valve operated
by the level control. The evaporation rate was determined by change In weight
of the supply container. The verification of the load cell output was done with
Cfass S calIbratlon quality wights conforming to specifications established by
the National Bureau of Standards at the beginning, near the end and at several
times throughout the testing phase. The load cell correlation was 1n general
within 11 (I.e. within 0.1 grams for a 10 gram weight change).
EVAPORATION RATE: The function of the automated level control used to determine
evaporation rate was verified manually prior to the start of testing. A thin
stiff wire suspended from above the surface was used as a level Indicator to
manually monitor the Initial or "full11 liquid level and check the repeatability
of the automated level control. Evaporation rate determined using the manual
method correlated closely with the automated system (generally there was less
than 5t difference). For each run, a single liquid at constant temperature was
used. Therefore, the surface tension during the run was quite constant, and
error due to change 1n surface tension adherence to the wire point was minimal.
This was confirmed by multiple trials of the wire point and comparison with the
automatic level control.
Additionally, replicate testing was performed with the automated system to
assure data reliability. Because of the nature of the testing, I.e., steady
state evaporation rate under controlled conditions, little variability was
expected and therefore only two duplicate runs on two different chemicals had
initially been planned. However, In practice at least one r?pl1cate run was
made for each of the test liquids during the testing phase. There were
approximately fifty replicate runs made during the testing process with a
variance of approximately lOt between tests.
15
-------
ItSOESilLIS
The results, along with the parameters Judged likely to relate to evaporation
rate, have been computed and tabulated In Tables A1 thru A16, Appendix A. These
are:
Reynold's Numbpr, NRE
Schmidt Number, NSC
Vapor pressure, VP (Inches mercury)
Surface tension, ST (lb/ft)
Diffusion coefficient, DIF (ft2/sec)
Latent heat, LATH (Btu/lb)
Viscosity (of liquid), VISL (lb-sec/ft2)
fLre.SiS0teiUcJi.J2f—Bita: The first treatment of the data was to prepare graphs
depleting the relationship of air velocity, air temperature and liquid
temperature (Figures B1-B16, Appendix B). These curves served to depict the
nature of the relationship and disclose any "outlier" data points so they could
be rerun. As presented here, the graphs Include the rerun data.
Using computer graphics and computational programs, various combinations of the
parameters were attempted. The molecular weight, vapor pressure, and air
velocity as shown in Equation 1, provided the best fit. Substituting the
Schmidt Number for molecular weight provided the second best fit. Including the
other variables (NRE, surface tension, diffusion coefficient, latent heat, and
liquid viscosity) did not further Improve -the fit, although obviously several of
them Influence, directly or Indirectly, the evaporation rate.
Using computer graphics and analysis of variance programs, various combinations
of the variables and their coefficients and exponents were compared to the
data. The coefficient and exponents were refined by an Iterative process.
After each iteration the predicted values and the experimental values were
compared by a least-squares method (ANOVA) and the next Iteration was an attempt
to minimize the variance by making Incremental changes in the coefficients and
exponents. The decision-making parameter was the multiple correlation
coefficient.
Different values for coefficients and exponents were found to provide the
highest multiple correlation coefficient 1f the 100 fpm data was treated
separately from the higher velocities. This "discontinuity" Is attributed to
the fact that the lower velocity was In the transition range between Reynold's
number turbulent and laminar flow regimes. Accordingly two sets of equations
were derived as shown In Tables II and III. The 100 fpm (0.5 m/s) evaporation
rate equations shown on Table III do not have a velocity term. All velocities
In this case were 100 fpm (0.5 m/s) and therefore the velocity term was
Incorporated In the constant.
16
-------
TABLE NO. II
HIGH VELOCITY RANGE (500 I 1.000 fpm)
MOLECULAR HEIGHT AND SCHMIDT NO.
EVAPORATION RATE EQUATIONS
TEST
UflJUlS
ALL CHEMICALS*1}
E¥AtQBAUQtL£Ql>& HONS
0.000237(MW)(VPXV)° 625
0.00725(HSC)'-4S(VP)(V)0M
MULTIPLE
CORRELATION
CQLEEKIEML-8
0.978
0.972
PERCENT ERROR MEASURED
VS CALCULATED VALUES
WITH A STUOENT-T 9CI
12.36 t 2.13
14.88 t 2.33
REFERENCE
IlOiM
CI -A
CI -B
ALCOMOLS
(2)
0.0008617(MN)°-900-5
0.02672(NSC)'•25(VP)(V)0-46
0.987
0.988
10.38 t 5.20
11.52 i 4.96
C3-A
C3-6
KETONES*35
0.002998°-2S(VPXV)0-70
0.0069«7(NSC)°-450-65
0.0.221(*SC)°-60(VP)(V)0-65
0.984
0.984
7.99 ± 2.33
7.93 t 2-13
C7-A
C7-6
AROHATICS
HATER
LOW VP ALCOHOLS
(4)
0.00008622(VP)(V>0-55
C.0006447(MK)(VPl-VPa)°-89(V)0-56
0.0176 °-890-57
O.OOOOtS44(MH),-380-65
0.001486(NSC)
2.45
(VP)
0.81(yjO.65
0.983
0.981
0.982
0.98)
0.967
0.966
9.08 ±2.19
9.43 ± 2.35
8.52 i 5.33
8.61 t 5.33
17.71 t 8.87
16.24 ± 6.49
C9-A
C9-R
Cll-A
Cll-B
CI 3-A
C13-B
Footnotes:
(1) Mater and the low vapor pressure alcohol data was not Included In the "All chealalt" analysts. Pentanol
1000 and 1400 FPM <5.0 to 7.1 ¦/$) data was used In error analysis.
(2) Pentanol 1200 and 1400 FPM (6.0 to 7.1 «/s) data was used In alcohol error analysis. Tht low vapor
pressure alcohols were not used 1n development of the general alcohol equation.
(3) 2-octanon# data Is not Included 1n ketone error analysis.
(4) Low vapor pressure alcohols used were: l-hexanol. l-heptanol and 2-octanol.
-------
TABLE HO. Ill
LOW VELOCITY RANGE (100 fpa)
MOLECULAR HEIGHT AHO SCHMIDT HO.
EVAPORATION RATE EQUATIONS
TEST
Llamas
ALL CHEMICALS<2)
£¥AK>R&LIOtUQUAHQ«('}
0.0069(KH>' 08(VP>° 98
0.312CNSC)1•55(VP)0.99
PERCENT ERROR MEASURED
HULTIPLE VS CALCULATED VALUES
CORRELATION WITH A STUDENT-T 901 REFERENCE
COEfflCIEHI. R2 CONFIDENCE INTERVAL £IUiR£
0.934 15.13 ±3.18 C2-A
0.925 16.85 ± 3.17 C2-B
ALCOHOLS*3)
0.0487(MH)0-59(VP)089
0.390CNSC)°-79(VP)0-88
0.960
0.959
7.64 ± 5.39
7.71 ± 5.33
C4-A
C4-B
KETONES
2.212(KH)-°-34(VP)0'95
0.517(NSC)_0-,38(VP)106
0.981
0.978
3.19 t 1.50
3.50 i 1.86
C6-A
C6-B
ALIPHATICS
0.0592
-------
In planning the project, It had been assumed that Reynold's Number or Schmidt
Number (or both) would be Important parameters In any resulting evaporation rate
equation, but this proved not to be the case. The experimental data Includes
only one' test air velocity (for a given liquid) that Is not In the Reynold's
Number turbulent range. This velocity Is 100 fpm (0.5 m/s). The result Is that
a flow regime variance (laminar vs. turbulent) was recognized and separate
equations for 100 fpm data (0.5 m/s) were developed; however, the scope of the
project did not allow full exploration of the laminar flow or transition
regions. Because only one velocity condition (i.e. 100 fpm, 0.5 m/s) was
tested In the laminar or transition region, equations developed for the 100 fpm
(0.5 m/s) category have no velocity component as velocity Is a constant for the
data set.
The evaporation rate equations with the values of the coefficients and exponents
and the accuracy of the prediction are shown In Table No. II and III. The
correlations between the measured evaporation rates and those predicted by
equation are Illustrated In Figures C1-CI4 of Appendix C. The starred values on
the graphs represent the calculated values on the X-axis plotted against the
measured values on the Y-axis. The forty-five degree line represents the
situation of perfect agreement or a correlation coefficient of 1.0 for the
equation versus measured values. The location of the starred points relative to
the line Is an Indication of the difference between the measured and calculated
values (I.e. a point above the line Indicates the measured evaporation rate was
greater than the corresponding calculated value).
Over the range of conditions studied, Equation 1 promises to provide a fast,
easy and adequately accurate method of predicting evaporation rate. If the
Identity of the chemical Is Known, It 1s necessary only to determine the air
velocity and the liquid temperature (hence the liquid vapor pressure) In order
to estimate the evaporation rate.
In practice, the liquid temperature will be determined by a number of factors.
In this work, the base temperature of the apparatus was kept at the liquid
temperature caused by evaporation of liquid from the surface. This technique
was used in order to simplify the physical situation so that the majority of
heat transfer took place only at the alr-llquld Interface.
In reviewing the tabulated data (Tables A1-A16) It can be noted that In ge:v?rol
a change In air flow rate while maintaining the same air temperature has little
affect on the resulting liquid temperature. Although this occurence Is not
exact throughout the data, It Is observable. Minor variance In control
conditions, particularly the matching of the base temperature to the liquid
temperature, undoubtedly affects the outcome. This same phenomenon was noted 1n
G.T. Sergeyev's work (Ref. 22).
A predictive equation was offered by V.G. Matsak (Ref. 16) for liquid
temperature; however, application of the equation to this data did not result In
accurate prediction of liquid temperature.
19
j. i
-------
A qualitative explanation of this phenomenon Is that the film heat transfer
coefficient varies with air velocity to the exponent 0.25 to 0.66 (depending
upon the flow regime) whereas the mass transfer coefficient varies directly with
the velocity. Therefore, the heat transfer Is limiting to the overall process.
Apparently, In the comparlslon of two evaporation rates at different velocities
and the same air temperature, the additional heat transfer due to Increased
velocity 1s (almost) exactly offset by the additional heat loss due to Increased
evaporation, resulting In the same temperature differential between air and
11quId 1n both cases.
In practice, In most situations the liquid temperature will be known or easily
measured. For example, for open surface tanks the liquid temperature 1s known
or easily measurable. For Indoor spills, the liquid temperature can be assumed
to be the same as the floor temperature, or It can be measured.
For outdoor spills, the liquid temperature can be measured, assuming that the
spill can be approached from the upwind side; or If the spill is of negligible
depth, the temperature can be assumed to be the same as the surface
temperature. Outdoors, solar heat gain can be of major Importance. The heat
flux from the sun can easily be as high as 250 BTU per hour per square foot (790
J/S m2); which 1s, for example, 2-1/2 times the heat transferred from air to
toluene to cause Its evaporation In a test at 500 fpm (2.5 m/s) air velocity and
71.7°F (22'C). These conditions yielded an evaporation rate of 0.56 pounds per
hour-square foot (7.62 X 10"4 kg/m2 s), which is equivalent to 100 BTU per hour
per square foot (315 J/S m2).
Since solar heat gain can have a major effect on the phenomena of simultaneous
heat-and-mass transfer, It would seem prudent to measure the liquid temperature
anyway to eliminate tho influence of this factor.
Effective Size o' Pool• well known (Ref. 9, 10, 18, 21, 24) that the size
of the pool makes a sign difference In the average evaporation rate per
unit area. At low alt velocities, the airflow boundary layer at the liquid
surface on the downwind sldf f the large pool will have a high concentration of
vapor which causes a cc .to Mng reduction of evaporation rate as compared to
the upwind side. i lrar r,ow conditions will, of course, Increase this
difference by reduci > mixing and dilution In the boundary layer.
Effect of Freeboar Open surface tanks almost always have considerable
freeboard, I.e. vertical distance from the edge of the tank to the liquid
surface. The effect of this "obstacle" would be to Increase the turbulence of
airflow at the liquid surface. It Is possible that such turbulence would
overcome or discount the observations we made In this duct apparatus showing
thai: a 100 fpm (0.5 m/s) velocity (In the transition range of the Reynold's
Number) resulted In separate evaporation rate equations.
"Still Air" Evaporation Rate
Reviewers of the first edition requested a qualitative discussion of liquids
evaporating In still air, and application of the equations derived from this
study to the "still air" case.
20
u.
-------
Absolutely still air would require an airtight, perfectly Insulated container
(whether a room or bottle) with an opening to the surrounding of small diameter
as compared to Its length, I.e. a capillary tube, to prevent pressure build-up.
A liquid exposed for evaporation would evaporate at a rate controlled at first
by diffusion In the container volume and later by a diffusion through the
capl1lary.
A more m#an1ngful comparlslon would result from adopting the ventilation
engineer's definition of "still air", l.t. air movement random in direction at
speeds 1n the range of 10 to 20 fpm (0.05 to 0.1 m/s). This condition is
observed In tight, Insulated rooms with no mechanical ventilation. The a 1 r
movement Is caused by minor temperature differences and wind pressures acting on
very small leaks, etc.
Wade (Ref. 24) made determinations of evaporation 1n "still air" The
evaporation pan, apparently not Inside an air duct, was surrounded on all sides
by vertical barrier walls of an unspecified height. It seems reasonable to
assume that, If air movement In the room was minimal, velocities In the range of
10 to 20 fpm (0.05 to 0.1 m/s) or less would have existed at the evaporating
pan. Hade further suggested as air velocity Increased, the Importance of the
small amount of evaporation due to "still air" conditions diminished to
Insignificance when compared to the larger evaporation rates under forced draft
conditions. Test conditions 1n this study Involved substantial air velocities
aid evaporation rate formula presented are velocity dependent. A discontinuity
was found between the 100 fpm (0.5 m/s) data and the higher velocity data;
however, elaboration as to the exact point of discontinuity or Investigation of
evaporation rate at low velocities has not been made.
The ASTM apparatus (Ref. 1) operates at a calculated air velocity past the
evaporation filter paper of about 0.8 feet per minute (0.004 m/s). Tho
uniformity of this airflow Is unknown, no special effort Is made to secure
uniformity. It Is deduced that this very low controlled airflow Is for the
purpose .of removing vapor from the region of the evaporation filter, thus
avoiding significant vapor pressure at the evaporation surface. however,
relative evaporation rates using the ASTM apparatus were found to be
consistently higher than this work (see Comparison with Other Methods section)
which Is no doubt due to the greatly extended surface for evaporation
represented by the small filter fibers and the wlcklng action bringing the
1lquld to that surface .
The experimental apparatus available could be modified to provide '-"or a
calculated air velocity In the range of 10 to 20 feet per minute over the
evaporation pan by providing an appropriate small diameter flow measuring device
such as a venturl or orifice In the air supply duct. Placing a uniform
resistance such as a sheet of filter media Immediately upstream of the
evaporation pan should give a reasonably flat flow profile so that the actual
air velocity over the liquid surface approximated the calculated velocity. As
the liquid evaporates, the cooling of that portion of the apparatus would
perhaps disturb the calculated profile at such low velocities. The more
volatile the liquid, the worse this disturbance. In the present stage of the
project, time and funding did not permit modification of the test apparatus and
low velocity evaporation tests (less than 100 fpm, 0.5 m/s) have not been made.
71
-------
The experiment described above would not correctly simulate the "real world"
I.e. the case of an open tank, or a spill In a room where "still air" conditions
exist. The evaporating vapor Is significantly heavier than air, and also cold
relative to the air. This causes the vapor to remain at the floor, or even to
"faU" to the floor from the Up of an open tank. The presence of the cold
vapor Itself would be one of the factors Influencing the otherwise Linton "still
air" movements In the room. The "still air" situation would be significantly
different from a directional air speed of 10 to 20 fpm (0.05 to 0.1 m/s> as
would exist In the modified test apparatus described above. However, there 1s a
different but Important reason why the apparatus should be modified and tests
run at two or three air velocities below 100 fpm (0.5 m/s). As explained
previously In the section "Test Results", the discontinuity In best fit
equations between 500 fpm (2.5 m/s) and 100 fpm (0.5 m/s) air velocities was not
anticipated. Many situations In Industrial plants Involve air velocities below
100 fpm (0.5 m/s), and the results of this research to date leaves a gap between
some velocity at or above 100 fpm (0.5 m/s) and common Industrial situations
ranging between 10 and 100 fpm (0.05 and 0.5 m/s). Tor example, on an
open-surface tank ventilated by an exhaust hood on one side, the "capture"
velocity on the far side 1s commonly only about 35 fpm (0.18 m/s). Therefore,
although the above described modification and experiment would not exactly
simulate the "real" world, meaningful Information may be gained from It.
A second modification could be Investigated which simulates more nearly an "air
tight" room with random air currents. To accomplish this, the existing
apparatus could be modified by blocking off the upstream and downstream ducts
and providing the downstream duct with a "pressure release" hole. A
non-dlrectlonal anemometer could be Installed over the pan to measure air
movement. Very small air jets could be provided at selected locations to create
the desired air movement If necessary. A small sampling air volume, equal In
volume to the air Jets, could be used to monitor vapor concentatIons so that
they would not exceed the lower explosive limit (LEL).
This option Is also rejected for the following reasons:
i) The amount of liquid evaporated to reach the LEL would be too small an
amount for the level control-gravimetric feed systems to quantify.
b) The ratio of liquid surface to "room" volume would be unreal Istlcally high
as compared to a typical "real world" scenario. Using a smaller pan would
make problem (a) even more difficult.
Thus It Is concluded that although meaningful information might be gained, the
present experimental apparatus cannot be modified to realIstlcally represent
"still air" evaporation.
The authors recommend that a "still air" experiment would be more meaningful and
easier to design 1f conducted In a mode) room designated or modified for that
purpose.
22
-------
Low .vaecr— Extsiu.te._D.IsxussJon
After completion of the Initial study. EPA requested that additional tests be
conducted on low vapor pressure liquids. A vapor pressure range of
aonroxlmatew 1 o to 0.1 mm mercury at room temperature was specified, and there
was no other direction regarding liquids to he tested. Lower vapor pressure and
corresDondlng Increase in molecular weight without change In type of molecular
structure could be accomplished only with aliphatic hydrocarbons or alcohols.
Alcohols were chosen on the basis of availability and cost.
The equations predicting evaporation rate of low vapor pressure alcohols are
^ 1 an1 f 1 can11 v different In coefficients and exponents than for the high vapor
nressure qroup and the correlation coefficient Is lower. Further developmental
work, was attempted on the low vapor pressure alcohol 500/1000 fpm molecular
wpinht eauatlon. Trials of other parameters as terms In the equation showed
latent heat to be an Important factor In evaporation rate. The best fit
equation of this type Is:
ER - 832(MW> (VP>°-80(V)°-65
-------
Comparisons of the evaporation rates determined In this project with rates
computed by other available methods are discussed below.
Relative EvaDorAtljoD-RalftJ^ta .(Ref. 1 -aad-2iI
The original project scope did not Include considerations of "relative
evaporation rate" techniques. In fact, frustration with the non-quantltatlve
aspects of widely available relative rate data, and with the fact that
experimental techniques for relative rate testing were not representative of
most actual applications, was part of the motivation for the present work.
EPA requested comparison of the results of this work with that of other
Investigators, Including "relative rate" data. The comparison of with "relative
rate" data proved to be quite interesting, and 1n conjunction with this work may
prove to be quite useful.
Relative evaporation rate Information available generally compares the
evaporation time of a "standard" chemical (ethyl ether, n-butyl acetate and
toluene are typically used) to the evaporation time of other chemicals. Test
apparatl utilize.! vary somewhat; however, all employ an absorbing medium (filter
paper) onto which a quantity of solvent Is dispersed. Evaporation rate Is
determined by time and qualitative (observation of the disappearance of the
solvent spot) or quantitative (gravimetric method using spring deflection)
methods.
Lynn D. Hllson (Ref. 25) presented relative rate evaporation data. The
reference 1s Included In Its entirety In Appendix Dl. Wilson's work relied on
visual observation of the disappearance of solvent applied to filter paper. The
basic Information supplied Is a list of solvent evaporation rates as ratios of
drying time compared to ethyl ether. It Is this work that 1s the basis of the
"Relative Evaporation Rate" table In the ACGIH Ventilation Manual and Is also
used for Material Safety Data Sheets. It is also perhaps useful In premllmlnary
tests for various diluents of coatings.
ASTM Designation D3539-76 (Reapproved 1981) Standard Test Method for Evaporation
of Volatile Liquids by Shell Thin-Film Evaporometer (Ref. 1) Is also a source of
relative evaporation data (see Appendix 02). This apparatus utilizes delivery
of a known quantity of solvent to a filter paper disk suspended from a
calibrated steel spring. Air at 77 ± 0.5*F (25 ± 0.25*C) and 0 to 51 relative
humidity Is admitted to the evaporating chamber at a rate of 21 L/mln (0.75
ft. /mlrO. Per Information garnered from the manufacturer (Falex Corp.) of the
Shell Thin-Film Evaporometer, the cross-sectional area of the apparatus Is 8
inches by 16 Inches (20.32 X 40.64 cm). The resultant air velocity through the
apparatus Is therefore less than 1 fpm (4.0 X 10"3 m/s). The filter paper disk
plus added solvent cause the spring to elongate. Readings of the position of
the filter paper disk are made at convenient Intervals of time as the sample
evaporates and the disk returns to Its original position. The ASTM Standard
notes two methods of data reporting.
24
-------
The first expresses evaporation 1n terms of percent weight evaporated versus
time. A relative evaporation rate Is calculated from the 901 weight evaporation
time for the test solvent and for n-butyl acetate. The second method expresses
evaporation rate quantitatively In terms of grams evaporated per second per
square centimeter of evaporating surface area. This alternate method 1s noted
as being not exact because the evaporation rate Is not linear and the
evaporating area not constant throughout the evaporation period.
Because tho ASTM method furnishes more precise control of the experimental
parameters than does that of Wilson, a determination was made of the actual
evaporation rate of n-butyl acetate so that comparison of relative ratss could
be made. Quantitative evaporation rate measurements for n-butyl acetate were
made with the IHE test apparatus. The evaporation rate was measured at an air
temperature of 77 ± 1*F (25 ± 0.56"C) and three air velocities: 100, 500 and
1000 fpm (0.5, 2.5 and 5.0 m/s). Corresponding data for the chemicals tested
under this project were taken from the air temperature versus evaporation rate
graphs shown in Appendix B. A relative evaporation rate was then calculated
using the quantitative data from the n-butyl acetate measurements and the
evaporation rate graphs. These relative evaporation rates were then compared to
those listed In the ASTM Standard (see Table IV).
The ASTM relative evaporation rate Is consistently higher than the IHE relative
rate data. The difference Is attributed to airflow velocity and evaporation
mechanisms. The greatly extended surface for evaporation represented by the
small filter fibers and the wlcklng action bringing the liquid to the surface
are likely causes Increasing the absolute evaporation rate. However, the
relation between the two data sets for relative rate Is fairly consistent, as
seen In Table IV, the ratio ASTM/IHE averages 1.41 and ranges from 1.16 to 1.67.
25
-------
TA#U IV
AST* HL4HVC MTt
VMSUS
IHC HtASUMO DATA
JJlLJtioorALlfifl. DiU_
CtitiAtii
H«th»noI
1 -froplDOl
Mr Vtl.
ls»
100
SOO
1000
100
SOO
1000
Oitalctl 8-01 y 1 Acttlt*
t*»p. !»*• £v»p. *»tt
1W_UZL-Z1^ iJt/hf....ft^-77'f
l-P»nUnol 100Q
o.s<
0.4s
0,#)
o.n
0.41
0.S2
0.04
O.M
o.u
O.M
a.»
O.M
O.M
O.M
m *»l»tW»
Hit.
I.so
1.2S .1.17
1.34
0.71
0.71 Av«.0.77
0.71
0.12
ftflit W*
MTM DtUkWt *»t# *»Uo
ut. asim/ihi.
1.1
o.H
0.2
1.S3
I.It
1.67
1 -Htiinol 100 0.015
SOO 0-023
1000 0.042
1-Hipttnol 100 O.OOI
SOO 0.010
iooo o.oij
2-Oct4f>ol 100 0.010
500 0.011
1000 0.016
0.34 0.042
0.S2 0.044 Avq.O.OSO H/AC1)
O.M 0.0(4
O.M 0.022
0.S2 0.014 A»J.0.020 M/AC1)
0.64 0.020
0.36 0.021
o.a 0.021 Avj.0.024 amu>
0.66 0.024
Acttont
ME*
100
SOO
IOOO
100
SOO
1000
2-Oet»noo« 1000
I .55
1.92
2.76
1.06
1.20
1.65
0.11
O.M
0.S2
0.64
0.16
O.H
O.M
O.M
4.31
3.64 Avg.4.0<
4.11
4.31
2.31 A«9.2.U
2.SO
0.17
S. 7
3.1
A/A(1)
1.40
1.51
N«I4»«
Ktptiri*
Octini
70
MX
IOOO
100
SOO
1000
100
SOO
1000
I.H
2.0«
3.20
0.»
1.10
1.64
0.11
O.M
O.H
O.M
0.S4
0.64
O.M
O.S«
O.M
O.M
0.52
O.M
$.31
4.00 A*|.4,
4.tS
75
2.SO
2.12 A»j.2.37
i. 44
1.03
1.04 Avj. 1.00
o.n
7 .1
*M(i)
1.4
l.W
1.40
100 1.12 O.M 3.11
500 '•*> 0.S2 2.M Mj.J.OO j.j , t7
1000 1.99 O.M 3.02 ,l7
Tolutft* 100 O.U O.M 1.21
500 ° M 1.23 Avj.1,32 j.0 ,
IOOO O.M O.M 1.45
Xyhnt 500 0.27 O.U 0.S2
1000 ° O.M 0.7J A»f.0.43 0.77 1.22
H*t«r too O.U O.M O.U
55? ®-" O.JSA»|.0.33 0.34 I Q*
1000 0.22S O.M 0.34
Avtrift 1.41
**"«• 1.14-1.47
1) */A: ASTK r»l»tl*« *v»pr»tUon ritt 4iti i»t mllibH
C0HVEIK10HS:
ft/aln X 0.00SOt • s/i
<*F - 321 X 1/9 « *C
Ib/hr f»z X O.OOI36 . Xj/i J-
26
21
-------
, 9).
Gray made an attempt to provide a comprehensive approach to the subject. A
detailed analysis of his p»p«r Is Included In Appendix E. As will be seen front
that material, we were unable to understand (without further study) how his
method would be used In practice. Gray analyzed his own method (p. 707 ) with
the following results
Ratio of Expected/Observed (E/0)
Xylene Average of 23 runs E/0 - 1.37
Range 0.73 to 1.8A*
N204 Average of 13 runs E/0 - 1.04
Range 0.38 to 2.21
The ranges depicted are unacceptably large.
Gray experimented with xylene, which we also used, and tabulated his data, A
comparison with IHE data and equation Is shown In Table V.
TABLE V
GRAY VS. IHE DATA
GRAY
LJ-auLd XiLlm
Air Vel. ft/mln 492
Air Temp. *F /5
Liquid Temp *F 70
Evap Rate lb/hr ft2
Measured 0.21
Calculated by Gray Equation NA
Calculated by IHE Equation 0.30
Variance X
Calc. to Measured 42X
CONVERSIONS:
ft/mln. X 0.00508 - m/s
(#F - 32) X 5/9 - #C
lb/hr ft^ X 0.00136 - kg/s
27
El
IHE
Xylene
500
70
64
0.22
NA
0.26
18%
-------
Hannft ^nrl Drlvas (Ref_. WJ—(A-LS-CL.S£j6l_Apj2Cnd 1X F..)
oresent equations for calculating the evaporation rate from a
sr m . 3,>. ih. b«ic e^ion ts
attributed to Fleischer (Ref. 6) In his SPILLS computer model:
. n /DT (H t D 4-10)
Qa " Ap s m/RTa
Where:
Q. - gas emission rate, kg/s
k - mass transfer coefficient, m/s
*2 - pool area, m2 _
Pw - vapor pressure of liquid at t(a)
HS - molecular weight
r - gas constant (8.314 joules/mol - K)
T - ambient temperature 'K
Q
Banna and Drlvas suggest quantifying Kg by the equation
kg . 0.00482 NSC-0'67 U°'7fl 0-°-" 4"'2>
where NSC - Schmidt number
u - windspeed, m/s
d - depth of pool, m
a second equation for quantifying kg Is also offered as Eq. 4-21.
Hanna and Drivas, In discussing this model on p. 44, state:
"As shown In Table 4-2, the difference among the emission rates calculated
bv the six models for the "Edgewood" Hz04 tests Is approximately a factor
of three for both tests. However, almost all the models greatly
overpredlcted the observed emission rates for N2O4, with a median
overpred1c11 on of 1 factor of about three.
We applied these equations to our data, making the appropriate unit conversions
to make the data comparable. The results are shown In Table VI. In general the
ratio of the rate calculated by the Fleischer "SPILLS" equation to I HE
experimental data ranged from 1.4 to 10.1 for turbulent air flow conditions, and
usually was less than 1.0 for the Indetermlnant or laminar flow conditions
represented by the 100 ft/m1n. tests. It Is prooable that for the large scale
outdoor conditions of the H&D report, only turbulent airflow was considered. It
is Interesting to note that the ratios we obtained average 2.76 as compared to
the phrase "approximately a factor of three" 1n the above quotation. Whether
this has any significance Is not Known.
28
Js
-------
TABLE VI
HANNA b DRIVAS (REF 10) VS. IHE DATA
Evaporation
Rate
lb/hr ft2
HiD
Air
Equation
IHE
Veloclty
Temp.
with IHE
Measured
Ratio
Liquid
ft/mln.
•F
data aDDlled
Evap. Rates
H&D/IHE
Hexane
1000
100
18.6
3.89
4.8
Hexane
100
120
4.52
2.28
2.0
N-Heptane
1000
100
6.1
2.2
2.8
N-Heptane
100
52
0.20
0.57
0.35
Octane
1000
100
2.24
1 .06
2.1
Octane
100
50
0.063
0.17
0.37
Benzene
1000
100
11.8
2.37
5.0
Benzene
100
70
1.0
1.12
0.89
Toluene
1000
100
4.0
1 .51
2.7
Toluene
100
54
0.20
0.35
0.57
Xylene
1000
120
1 .90
1 .20
1.6
Xylene
100
100
0.19
0.40
0.48
Methanol
1000
100
9.7
0.96
10.1
Methanol
100
51.5
0.43
0.48
0.89
N-Propanol
1000
100
2.4
0.60
4.C
N-Propanol
500
70
0.46
0.38
1.2
N-Propanol
100
45
0.058
0.013
4.5
1 -Pentanol
MOO
100
0.80
0.33
2.4
1 -Pentanol
1000
44
0.06
0.03
2.0
Acetone
1000
109
27.4
2.93
9.4
Acetone
Methyl ethyl
Methyl ethyl
2-0ctanone
2-0ctanone
ketone
ketone
100
1000
100
1000
500
71
120
70
1000
89
1.92
17.0
0.98
0.23
0.095
CONVERSIONS:
ft/mln X 0.00508 - m/s
<*F - 32) X 5/9 - *C
Ib/hr ft2 X 0.00136 - kg/s m2
1 .55
2.59
1 .01
0.17
0.13
Average
Range
1.2
6.6
0.97
1 .4
0.73
2.76
0.37-10.1
29
31
-------
In the Initial calculations for comparing IHE data to the Fleischer "SPILLS"
equation In Hanna & Drlvas, one of the authors made a mistake. He used the
vapor pressure at the liquid temperature rather than at the "ambient"
temperature as required by Fleischer (Ref. 6). Hhen this mistake was
discovered, no more such Incorrect calculations were made. It Is Interesting to
note, however, that the agreement between IHE data and Fleischer calculated rate
using the "wrong" temperature was much better than the agreement when the
Fleischer equation was correctly used. See Table VII as compared to Table VI.
The "incorrect" treatment yields a much smaller range of ratios and the average
ratio was 1.09.
The term "ambient temperature" could be ambiguous when temperature differences
between air and evaporating liquid range from 20* to 60*F. It seems most
reasonable to Interpret "ambient temperature" to mean air temperature in the
circumstances of Intended use of the Fleischer equation.
Two major differences between a small evaporation pan and a large outdoor pool
seem apparent. For a large pool, the boundary layer proceeding downv* nd will
have an Increasing vapor concentration, thus creating a "back pressure" against
the vapor pressure and slowing the evaporation rate. On the otherhanrl, ripples
or waves can develop to a greater extent, so that the actual liquid surface Is
greater than the plan area of the pool, thus Increasing the evaporation rate.
Although the phenomena counteract each other, It seems unrealistic to assume
they cancel. However, It Is Interesting that the average ratio shown In Table
VI so closely coincides with the ratio quoted from Hanna and Drlvas quoted above.
-------
TABLE VII
HMNA_OfiIVAS_V^_IilE.J)AIA
INCORRECTLY USING LIQUID TEMP FOR VAPOR PRESSURE DATA
Evaporation Rate
lb/hr ft2
H&O
Liquid
Air
Veloclty
ft/mln.
Air
Temp.
Of
Equation
with IHE
IHE
Measured
Ratio
H&D/IHE
Octane
1000
100
1.37
1 .06
1 . 3
Octane
500
70
0.48
0.48
1 .0
Octane
1000
50
0.07
0.17
0.42
Benzene
1000
100
2.99
2.37
1 .26
Benzene
500
79
1.61
1 .47
1 .1
Benzene
100
70
0.42
1.12
0.37
Xy1ene
1000
120
1.09
1 .20
0.91
Xy1ene
100
100
0.017
0.40
0.04
Methanol
1000
120
2.07
1.15
1 .8
Methanol
100
52
0.25
0.48
0.53
1-Pentanol
1400
100
0.43
0.33
1.3
1-Pentanol
1000
44
0.08
0.03
2.7
Acetone
1000
109.5
4.46
2.93
U
Average 1.09
Range 0.04-2.7
CONVERSIONS:
ft/min X 0.00508 - m/s
CF-32) x 5/9 - *C
Ib/hr ft2 X 0.00136 - kg/s m2
31
1)
-------
QlB_v J.en_.&_S t ut zman _(Re.L_18)
O'Brien and Stutzman presented some data on benzene, which we also tested. The
comparison Is shown In Table VIII. The comparison 1s not meaningful because
these authors heated the liquid.
TABLE VIII
O'B & S
IHE
Liquid
ienzeoi_
Air Vel. ft/ ml n
293.0
100 500
Air Temp. *F
100.0
100 101
Liquid Temp *F
81 .7
45 52
Evap Rate Ib/hr ft2
Measured 0.86 1.27 1.82
Calculated by IHE Equation 3.04* 1.27 1.75
Variance X
Calc. to Measured 343X -0- 4X
* due to high liquid temperature.
CONVERSIONS:
ft/mln X 0.00508 - m/s
CF-32) X 5/9 - °C
lb/hr ft2 X 0.00136 - kg/s m2
32
3f
-------
H&de Afle_L__241
Wade's work was similar to this project In many ways. He derived a general
equation for evaporation which applies to forced and natural (still air)
convection of the form:
E - 1CT7 H0'71 ([9.8 log_1(-0.011 V>](Pe-Pd>1l25 + (Hade Eq. 16)
1.57 V0-85 >
Where:
E - Rate of evaporation in gm/sec from a surface 8.9 cm long by 8.9 cm wide
M - Molecular weight of liquid
V - Tangential velocity of the air stream In cm/sec.
Pe - The vapor pressure at liquid surface temperature, mm Hg
Pd - The partial pressure of vapor In the Incident air stream, mm Hg
Wade further states the portion of the equation relating to natural convection
can be neglected where air velocities exceed 260 fpm (1.3 m/s) and indicated P^
will be zero In all cases other than water. Thus the equation reduces to:
E - 1.57 X 10"y M0-71 Pe V0-85 (Wade Eq. 9)
The I HE physical data (liquid molecular weight, liquid vapor pressure, air
velocity) was applied to Wade's equations and the resulting evaporation rates
compared to the IHE measured evaporation rates. The results of this comparison
are shown In Figure 7. Wade equation 16, (for both forced and natural
convection) was applied to the IHE 100 fpm (0.5 m/s> data (Figure 7A) Wade
equation 9 was used on the 500 and 1000 fpm (2.5 and 5.0 m/s) data (Figure 7B).
The lines on the graphs represent the locus of points for perfect agreement with
the Wade equation. The starred points represent the calculated values of
evaporation rate from the Wade equation on the X-axls plotted against the IHE
measured values on the Y-ax1s. Wade equation 16 universally under-estimated the
measured evaporation rate In the 100 fpm (0.5 m/s) case. A maximum percent
difference of 79X was found with an average difference of 63X (Wade equation
results versus IHE measured values). Hade equation 9, applied to the 500 and
1000 fpm (2.5 and 5.0 m/s) data, more closely approximated the IHE measured
values with an average deviation of 19X (Wade equation versus IHE measured
values). This equation had a multiple correlation coefficient of 0.92 versus a
0.98 coefficient using the IHE equation on the same data.
Raw measured evaporation rate data could not be directly taken from Wade's
paper; therefore, comparisons between experimentally measured values (IHE vs.
Wade) could not be made.
33
If
-------
FIGURE 7
WADE EQUATIONS VS I HE DATA
1.40
.280
r
i
* *
*
*
*
M ...
* *
*
f
rf
X
/'
r
...
1
FIGURE 7A
0.5 m/s (1OOFPM) DATA
,200 .600 1.88 1.40 1.80
Wade Eq. calculated Evap. Rate, g/min
Evap. Rate = 6x 10~6M0-71 {(9.8 1og"1(-0.011V))(Pe-Pd)1 •25 +
1.57 V0-85 (Pe-Pd)>
C
e
CD
ra
cc
cx
T3
>
0>
U
3
CO
fO
a;
4,50
3.50 -¦
2,50 --
1.50 J-
,500 x
,¦*
/
/
/
/
/..it.:
* >
/ *
s
y* * *
*
j j *
Jrlr. *
*
w *
*
F 1
1 . > i .
. : .L.i
.500
1.50 2.50 3.50
FIGURE 7B
2.5 to 5.0 m/s DATA
(500 to 1 COOFPM)
4.50
Wade Eq. calculated Evap. Rate, g/min.
Evap. Rate = 9.42 x 10"6M0-71 PeV0-85
NOTE: The constants in
WADE's Eqs. were
dimensionally modified
from g/s to g/min.
34
-------
Ua £^eA-Ie^ni£4lJiuj.dam^9iLKflLZMjlLjtojdyili^5JL-lj}.
The Technical Guidance for Hazards Analysts: Emergency Planning for Extremely
Hazardous substances published by the U.S. Envlronmental Protection Agency (Ref.
23) presenti an e-tjuatSon Attributed to dement (Ref. 5> and In turn referenced
to MacKay and Matsuga (Ref. 15) for evaporation of liquids. (These references
and their rsference lists were searched for an aquation of this form, without
success. This Situation v&s probably caused by i transcrfption error, but is
not of Importance because several equations of that general form are found In
the literature. Further, alt equations of this form use the tnass transfer
coefficient, whereas the thrust of this research effort was to find an empirical
method of predicting evaporation rates mhlch avoids ilia theoretically correct
tut complex and assuAiptfon-sensItlve simultaneous heat-and-mass transfer
phenomena. The research premise was that the simplified method might be more
accurate, and easier to apply, for the smail range of temperatures and pressures
Involved In liquid evaporation at ambient conditions.)
The EPA/Mackay equation fs shown In full and 1n its simplified form belo*.
OR . 150^e.cMnlJ:tmjLll_(&lJ^L^2Z5Uai^m2) (EPA EQ. 1)
(R) (Tt + 273) (760 mHq/itm) (454 g/lb)
where:
vR - P.-a~e af releii-e to a1 r ' Its* ml ii).
MH - Molecular Weight .
K m Gas phase mass transfer coefficient (cm/sec).
A - Surface area of spilled material (ft^).
VP - Vapor pressure of material at temperature T! .
K 13 determined from the value of K for a reference material. In this case water.
K - (^refX^ref/^)1 '3 (EPA EQ. 3)
^ref >• Kwter - 0.25 (U) 0-78 (EPA EQ. 1)
where: u - vfndspeed m/s
Combining equations 1, 3 and 4 and constants results ^n:
Oft ,
(8) (T1 + 273)
35
31
-------
The EPA/MacKay equation, except for the Inclusion of the mass transfer
coefficient, Is generally of the same form as IHE equation which uses molecular
weight, liquid vapor pressure and air velocity as primary components. However,
the EPA/MacKay equation uses T1 which 1s defined as the temperature at which
the chemical Is stored. The use of the storage temperature is somewhat
questionable because In the case of a spill, evaporation rate Is only affected
by the liquid storage temperature until the Uquld 1s spilled where upon It
comes to thermodynamic equilibrium with the surroundings. In the case of most
spills It can be assumed this happens fairly rapidly.
For the purpose of comparison to the IHE equation, a portion of the IHE data was
converted to the proper units and applied to the EPA/Mackay equation. Tabic IX
Illustrates. In this application, the equilibrium liquid temperature and vapor
pressure were used from the IHE data. The evaporation rate units resulting from
EPA Equation 7 are Ibs/mln. These were converted to lb/hr ft^ to match In units
with the IHE measured data.
36
-------
TABLE IX
EP_AZHacKay-IflUA.t \ oruyx^IUL^U-
A1 r
Air
Llq
Veloc1ty
Temp
Temp
L1 a u 1 d
£lZmln_
•F
*F
Methanol
100
70.5
39.2
Methanol
100
120.8
47.6
Methanol
500
100. /
43.2
Methanol
500
119.7
47.5
Methanol
1000
100. 1
40.8
Methanol
1000
120.4
48.2
Evaporation
IMiiLliL
EPA Equation IHE
with IHE Measured Ratio
0.06 0.52 0.16
0.08 0.75 0.11
0.25 0.73 0.34
0.28 0.85 0.33
0.41 0.96 0.43
0.50 1.15 0.44
MEK
100
69.6
0.14
1.01
0.14
MEK
100
100.3
57.2
0.17
1.14
0.15
MEK
500
70.3
41.6
0.38
1.17
0.33
MEK
500
120.1
57.8
0.59
1.68
0.35
MEK
1000
69.6
41 .7
0.68
1.67
0.41
MEK
1000
100.7
52.6
0.90
2.04
0.44
Heptane
100
51.8
47.7
0.07
0.57
0.12
Heptane
100
99.6
62.8
0.11
1.26
0.09
Heptane
500
70.5
49.1
0.25
1 .00
0.25
Heptane
500
99.8
64.7
0.39
1.58
0.25
Heptane
1000
41.1
38.3
0.30
0.92
0.33
Heptane
1000
69.0
49.8
0.45
1.48
0.30
Toluene
100
54.4
53.2
0.04
0.35
0.11
Toluene
100
99.5
69.9
0.07
0.88
0.08
Toluene
500
71.7
54.7
0.17
0.56
0.30
Toluene
500
100.0
71 .1
0.27
0.98
0.28
Toluene
1000
69.9
54.4
0.29
0.85
0.34
Toluene
1000
100.0
70.2
0.45
1.51
Q.3Q
Average A11 Data 0.27
Range 0.08-0.44
Average 100FPM Data 0.12
Range 0.08-0.16
Average 500-1000FPM Data 0.34
Range 0.25-0.44
ConversIons:
ft/mln. X 0.00508 - m/s
CF-32) X 5/9 - "C
lb/hr ft2 X 0.00136 - Kg/s m2
3?
-------
It can be seen from tabulated data In Table IX that the EPA/MacKay equation
universally under estimated the IHE measured data. Additionally, a marked
difference can be seen when looking at the 100 fpm data as compared to the 500
and 1000 fpm data. In the 100 fom case, the EPA/MacKay equation results were on
the average 0.12 that of the IHE measured data with a range of 0.08 to 0.16.
The 500 and 1000 fpm case although showing a much closer prediction to the
measured values, still had an average factor difference of 0.34 and a range of
0,25 to 0.44. The difference between the 100 fpm and 500/1000 fpm case Is
likely due to flow regime differences as previously discussed. The source of
difference In general between the EPA/MacKay and IHE equations Is unclear;
however, It It Interesting to note that Hu and Schroy (Ref. 26) also developed a
model based In part on MacKay's work. This Is a two part computerized
theoretical model based on heat and mass transfer. Application of a portion of
the IHE 100 fpm data to this model was made by Philip C. Dell'Oreo In a review
of the IHE report (see Appendix G). Table X Illustrates.
IMLO
Evaporation
lb/hr ft2
Liquid
Air
Veloclty
ft/mln
A1r
Temp
1E_
Llq
Temp
* F
Wu/Schroy
Model with
IK£_DAiiL__
IHE
Measured
Ratio
idimy/lUE
Methanol
100
70.5
39.2
0.178
0.52
0.34
n-propanol 100
70.5
52.9
0.046
0.24
0.19
Acetone
100
70.7
36.2
0.448
1.55
0.29
MEK
100
69.6
48.8
0.261
1.01
0.26
Hexane
100
70.7
37.3
0.437
1.91
0.23
Heptane
100
70.5
50.2
0.154
0.98
0.16
Octane
100
69.6
55.8
0.056
0.32
0.18
Benzene
100
70. 1
40.6
0.243
1.21
0.20
Toluene
100
69.5
51.3
0.109
0.40
0.27
Xylene
100
100.0
83.4
0.063
0.40
Average
Range
0.16
0.23
0.16-0.34
Conversions:
ft/mln. X
0.00508
Ml
m/s
CF-32) X
5/9
-
*C
lb/hr ft2
X 0.00136
¦
Kg/s
m2
38
-------
As Table X Illustrates, the evaporation rates predicted by the Wu/Schroy
equation arp on the average 0.23 that of the I HE measured values with a range of
0.16 to 0.34. This Is similar In magnitude to the overall average factor
difference between the IHE (100, 500 and 1000 fpm) data and the EPA/Maci'ay
equation; however, It Is not as comparable to the 100 fpm EPA/MacKay data (le
the EPA/MacKay equation differed from the IHE 100 fpm measured values by an
average factor of 0.12).
A definite difference between the heat and mass transfer theoretical models and
the IHE experimental values exists. The difference appears to be somewhat
constant and maybe attributable to a variety of factors Including model input
parameters (heat transfer coefficients, etc). Because It Is a computerized
model which was not available further analysis was not made.
39
-------
S.uipm ft rx -
1.04
0.38-2.21
Hanna & (afr temp)
(10)
2.76
0.37-10.1
Dr 1vas (1iquld temp)
1 .09
0.04-2.70
(3)
O'Brien & Stutzman
(18)
N/A
N/A
(4)
Wade (100 fpm)
(24)
0.37
0.21-0.54
(500 & lOOOfpm)
1 .08
0.58-1.60
(5)
EPA/MacKay
(5, 15, 23)
0.27
0.08-0.44
Wu/Schroy
(Note 6)
0.23
0.16-0.34
(6)
1) ASTM method measures relative rates but not actual rate. Actual rate of at
least one substance also tested for relative rate must be known for most
applIcatlons.
(2) Presents Gray's own analysis of measured vs. calculated. IHE-measured rate
for xylene at one set of conditions were almost identical to Gray's
result. S&e Table V.
(3) Results using liquid temperature for base of liquid vapor pressure Instead
of air temperature as stated by authors.
(4) N/A - not applicable. Authors heated the liquid.
(5) Wade equations vs. IHE data.
(6) Wu/Schroy method, personal communication from Dell'Oreo. See Appendix G.
40
rt-
-------
E&QPLEHS ENCOUNTERED IN THE HQRK
1. In planning the experiments It was assumed that Reynolds Number and/or the
Schmidt Number would be a parameter of Importance In evaporation rate
equations and therefore, only one air velocity below the Reynold's Number
"turbulent" range was planned. The best fit equations, however, have a
velocity term but not a Reynold's Number term. With only one air velocity
below the turbulent range there can be no velocity term In the low velocity
equations because velocity was a constant at 100 fpm. Velocity In these
low ranges probably 1s an Important parameter, but 1s not shown in this
vork.
?, The computer automated control system Is not successful. Much manual
Intervention Is required for successful operation. The preferred system at
this stage of Investigation would be more conventional process controls,
with the computer serving as a data logger and computational aid.
3. The level control, although very accurate, Is exceedingly awkward and
sensitive. Setup changes for liquids of differing densities are
time-consuming. The presence of the equipment Inside the test duct
represents an obstacle to air flow and also introduces secondary problems.
A different system, depending only on light beams and physically external
to the test duct, would be preferable.
4. Obtaining evaporation rate data by measuring loss of weight Is, for a slow
evaporating liquid, very time consuming. Improvement 1n the control system
as described above would reduce the cost but not the calendar time. In
order to determine the underlying scientific parameters, some of the data
was taken at hljher temperatures than would normally be encountered In the
environment to speed the process.
One solution would be to provide a larger evaporating surface, but It would
seem preferable to attempt to reduce the size of the equipment Instead of
increasing 1t. Another solution would be to measure the vapor
concentration In the exhaust air, but this has the disadvantage of
requiring caref|i calibration of the sampling and quantification
Instrumentation for each individual compound tested, and would Introduce
other possible sources of error encountered In sampling a moving air
stream.
It appears that the best solution would be to provide for better automatic
control and to allow the test to continue for the necessary time. With
better control, the technician In attendance could accomplish other
productive work. However, the option to quantify the amount of evaporated
vapor should be kept open for further study.
5. The long test times required to measure low vapor pressure liquid
evaporation rates worked In combination with a slight load cell drift to
necessitate use of a manual level control method when testing slow
evaporating liquids. Tills condition could be corrected with Improvement of
the level control and stabilization of the load cell.
6. It Is difficult to achieve adequately low temperatures I.e. below 45°F with
the current setup. This problem Is circumstantial, not fundamental, and
can be overcome 1n several ways.
41
*3
-------
£UIUR£JOfiK
1. It would be highly advisable to modify the apparatus to explore the air
velocity region between 20 and 100 fpm. This can be done by Installing a
low flow rate orifice or venturl In the upstream duct, and a
flow-distributing resistance such as a perforated plate or a shee'. of
high-density filter media Just upstream of the evaporation pan.
2. It would be advisable to Improve the apparatus to reduce the labor cost of
using It. Items In this Improvement list would be to Improve the method of
measuring differences In weight of the feed container, Improve the level
control system, Improve process control system/computer Interface, and
Improve the chiller system. The control function and the data logging
function of the computer should be separated.
3. Experiments should be conducted with a simulated "freeboard" In the
evaporation pan, I.e. the liquid level should be maintained at some
predetermined distance below the edge of the tank. This would simulate the
situation In open surface tanks. Relationships may be found between the
tuvbulence parameters and the evaporation rate.
4. Other classes of organic molecular structure should be Investigated.
5. Evaporation rates of mixtures of liquids should be determined and
predictive techniques developed, If possible.
6. A few large scale tests should be made to develop a relationship between
these duct tests and larger scale environmental phenomena.
7. Study should be conducted with base temperatures simulating cold or hot
pavement or controlled temperatures tank temperatures.
8. Study expanding the experimental scope within the already tested chemical
classes and liquids should be conducted.
9. The relationship between liquid temperature, vapor pressure, and
evaporation rate should be further Investigated.
10. A difference In equation form was found for the low versus high vapor
liquids studied. Detailed exploration of the Involved parameters and
development of applicable equations requires further study.
42
v/
-------
REEIBINCLS
1) ASTM, Standard Test Method for Evaporation Rates of Volatile Liquids by
Shell Thln-FIlm Evaporometer, ASTM Designation 03539-76 (Reapproved 1981)
2) Bishop, E.C., W.Popendorf, 0. Hanson and J.Prausnltz Predicting Relative
Vapor Ratios for Organic Solvent Mixtures, Am. Ind. Hyg. Assoc. 0.
43(9):656-661 (1982).
3) Brutsaert, W., A Model For Evaluation as a Molecular Diffusion Process Into
a Turbulent Atmosphere, J. Geophys. Res. 70: 5017-5023, 1965.
4) Carrier, W. H., The Theory of Atmospherlc Evaporation - With Special
Reference to Compartment Dryers, J. Ind. Eng. Chem., 13, 432-438, 1921.
5) Clement Associates, Inc. Mathematical Models for Estimating Workplace
Concentration Levels: A Literature Review, EPA Contract Ho. 68-01-6065
Oct. 1981.
6) Fleischer, M.T.: SPILLS, An Evaporation/Air Dispersion Model for Chemical
Spills on Land. Shell Devel. Center P.O. Pox 1380, Houston Texas 77001,
1980.
7) G1111 land, E. R., Diffusion Coefficients 1n Gaseous Systems, Ind. Eng.
Chem., 26. 681-685, 1934.
8) G11111 and, E. R., and T. K. Sherwood, Diffusion of Vapors Into Air Streams,
Ind. Eng. Chem., 26, 516-523, 1934.
9) Gray, D. C. , Solvent Evaporation Rates, Am. Ind. Hygiene Assoc. J., 35,
695-710, 1974.
10) Hanna, S. R., Drlvas, P. J., Guidelines For Use Of Vapor Cloud IU_sfi.ejisJ_£D
HOiJiJ.S, Am. Inst. Chem. Eng., N.Y., N.Y., 1987.
11) Hlnchley, J. W. , and G. W. Hlmus, Evaporation of Currents In Air, Trans.
Inst. Chem. Engrs., 2, 57-64, 1924.
12) Jasper, J. J., Physical Chemistry Reference Data, Vol. 1, 1972.
13) Lewis, H.K., The Evaporation of a Liquid Into a Gas, Mech. Eng., 44,
445-446, 1922.
14) Lurle, M., and N. Mlcha 11 off, Evaporation From Free Water Surface, Ind.
Eng. Chem., 28, 345-349, 1936.
15) Mackay, Douglas and Matsugo, Ronald S. Evaporation Rates of Liquid
Hydrocarbon Spills on Land and Water, The Canadian Journal of Chemical
Engineering, Vol. 51, August, 1973.
16) Matsak, V.G., and G. Sanltarlya, Vapor Pressure and Evaporation of
Substances In Moveable Air, NTIS Doc. 0TT 6413166, i-15, Gigyenad
Sanltarlya, pp 35-41, FTD-TT 1044/1+2, (Russian) 1963.
17) Naln, V. P. S., and J. R. Ferron, Prediction of Binary Diffusion
Coefficients for Polar Gas Mixtures, Ind. Eng. Chem. Fundam., 11, 420-421,
1 972.
43
-------
18) O'Brien, L. 0., end L, F. Stutzman, Mass Transfer of Pure Liquids From a
Plane, Free Surface, Ind. Eng. Chem., 42, 1181-1187, 1950.
19) Perry, R. H., and D. W. Green. Perry's Chemical Engineers Handbook. 6th
Ed; McGraw-Hill Book Co., N.Y., N.Y., 1984.
20) Popendorf, W., Vapor Pressure and Solvent Vapor Hazards, Am. Ind. Hyg.
Assoc. J. 45 (10): 719-726, 1984
21) Powell, R. H., and E. Griffiths, The Evaporation of Water From Plane and
Cylindrical Surfaces, Trans. Inst. Chem. Engrs., 13, 175-198, 1935.
22) Sergeyev, G. T., Heat and Mass Transfer During Evaporation of a Liquid in a
Forced Gas Flow, NTIS #TT 6225831, 1-6, translated by Air Force Sys. Comm.,
FTD-TT 62-327-1+2+4, 1962.
23) U.S. EPA, Federal Emergency Management Agency, U.S. DOT, Technical Guidance
for Hazards Analysis: Emergency Planning for Extremely Hazardous
Substances, supplement to NRT-1, Hazardous Materials Emergency Planning
Guide, 1987.
24) Hade, S. H., Evaporation of Liquids In Currents of Air, Trans. Inst. Chem.
Engrs., 20, 1-13, 1942.
25) Wilson, L. D., Evaporation Rates of Solvents and an Improved Method For
Their Determination, Paint and Oil Chemical Review, 1955.
26) Wu, J.M., and J.M. Schroy,: Emissions from Spills, APCA Spec. Conf. on
Control of Specific (toxic) Pollutants Pittsburg, 1979
44
-------
APPENPIX-A
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
METHANOL
AIR
VELOCITY
FPM
AIR
TEMP
F
TEMP LIQ.
TEMP BASE
F
NRE
NSC
VP
in. Hg.
ST
(TO-3)
ib./U.
DIF
ft.2 /sec.
LATH
3TU/lb.
VISL
(io-5)
!b. sec./".
MEASURED
EVAP. RaTl
Ib./hr. ft.
£VAP. RATE FROM
EQ. USING
ib./h'. ft.2
NSC
v.w
100
51.5
35.7
42.5
4640
0.9621
1.30
1.63
1.659
529.9
1.64
0.48
0.48
0.48
100
70.5
39.2
43.5
4330
0.9628
1.48
1.62
1.776
5?R.n
1 .60
0.52
0.53
0.53
100
99.7
45.8
46.6
3940
0.9637
1.85
1.60
1.950
524.7
1.52
0.58
0.65
0.65
100
120.8
47.6
47.8
3730
0.9613
1.98
1.60
2.067
523.8
1.50
0.75
0.69
0.69
500
45.3
34.0
42.5
24180
0.9571
1.25
1.64
1.600
530.7
1. 66
0.54
0.55
0.55
500
70.2
38.9
42.9
22090
0.9649
1.45
1.62
1.738
528.3
1.60
0.61
0.65
0.63
500
100.7
43.2
44.2
20120
0.9599
1.70
1.61
1.917
526.0
1.55
0.73
0.75
0.74
500
119.7
47.5
45.6
18790
0.9294
1.95
1.60
2.122
523.7
1.50
0.85
0.83
0.85
1000
44.2
32.2
41.8
48050
0.9615
1.20
1.64
1 .fi03
53(1.?
1 FQ
n ri
0.73
0.74
1000
70.7
38.3
43.6
43980
0.9883
1.40
1.63
1.704
528.6
1.61
0.92
0.88
0.86
1000
100.1
40.8
37.9
39-: 30
0.9544
1.60
1.62
1.932
527.1
1.57
0.96
0.97
0.99
1000
120.4
48.2
46.9
37,570
0.9620
2.00
1.60
2.050
523.5
1.49
1.15
1.22
1.23
METHANOL MOLECULAR WT. = 32 E?A CONTRACT NO. 68-02-4248
iHE JOB NO. 747-004
DATE: 4/5/88
BY: K.0. Brcun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
N-PROPANOL
AIR
velocity
FPM
AIR
TEMP
F
TEMP. L:Q.
TEMP BASE
F
NRE
NSC
VP
in. Hg.
ST
(TO"3)
lb./ft.
DiF
(io-)
ft.2 /sec.
LATH
8Tu/;b.
V1SL
(io-5)
lb. sec./ft.
Ml aSU R
evap. rate
lb./hr. ft.
EVA P. RATE PR 0U
EQ. USiNC
'.b./hr.
NSC
uw
100
45.3
41.9
40.9
4810
1.482
0.20
1.70
1.039
358.5
1.11
0.13
O 11
0.13
100
70.5
w.y
51.6
4400
1.511
0.32
1.67
1.113
354.8
1.07
0.24
0.20
0.20
100
100.4
69.1
69/1
3940
1.512
0.59
1.62
1.245
349.0
1.01
0.31
0.34
0.34
>
no
500
45.0
4b.4
45.7
24440
1.455
0.23
1.69
1.004
357.4
1.10
(1)
0.19
0.17
0.18
500
70.3
59.1
59.9
22140
1.510
0.42
1.65
1.107
352.2
1.04
0.38
0.33
0.32
500
100.2
/1.7
71.5
20120
1.510
0.64
1.61
1.219
348.0
l.'OO
0.51
0.50
0.49
1000
45.0
43.6
43.3
48620
1.507
0.22
1.70
1.011
356.0
1.10
(1)
0.19
0.24
0.24
1000
70.5
61.4
61.3
43980
1.509
0.44
1.64
1.116
353.5
1.04
0.48
0.47
0.48
1000
100.3
70.0
70.3
40460
1.510
0.60
1.62
1.212
347.5
1.01
0.60
0.64
0.65
^ ^Note same evaporation rates at 500 and 1000 fpm 45°F condition. EPA CONTRACT no. 68-02-4248
Higher liquid temperature at 500 fpm condition probable cause. 1HE JOB NO. 747-004
s ^ 1 DATE: 4/5/88
PR0PAN0L MOLECULAR WT. = 60.1 BY: K.O. Broun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
1—PENTANOL
AIR
VELOCITY
FPM
AIR
TEMP.
• F
TEMP. LIO.
TEMP.BASE
¦ F
NRE
NSC
VP
in. H g.
ST
(io-3)
lb./ft.
DIF
(io-*)
ft.2 /sec.
LATH
B TU /¦; b.
VISL
(10"5)
d. sec./ft.
MEASURED
evap. rate
ib./r)r- ft.
EVA3. RATE FRCU
EO. 'JSi'JC
-J
KSC
VI iV
1000
44.1
(1)52.2
52.1
48120
1.923
0.04
1.82
0.780
272.5
11.83
(2)
0.03
0.06
0.06
1000
70.3
bl.J
67.7
43980
1.927
0.06
1.77
0.874
268.4
8-80
0.06
0.09
0.09
1000
101.7
93.8
93.6
39950
1.925
0.19
1.68
0.963
263.0
6.20
0.24
0. 30
0.29
1200
44.7
(1) 52.2
51.5
57510
1.927
0.04
1.82
0.802
272.5
11.80
(2)
0.03
0.06
0.07
1200
70.0
by. i
69.0
52680
1.923
0.06
1.77
0.877
268.7
9.00
0.11
0.09
0.10
1200
99.3
88.5
87.4
47920
1.932
0.16
1,70
0.960
264.2
6.67
0.27
0.25
0.27
1400
44.7
(1) 43.0
48.8
68500
1.931
0.03
1.83
0.787
273.2
12.40
0.04
0.06
0.06
1400
70.1
69.4
69.2
61403
1.932
0.07
1.76
0.874
268.5
9.00
0.16
0.14
0.14
1400
100.3
92.2
91.8
56970
1.924
0.18
1.69
0.946
263.4
6.33
0.33
0.35
0.35
(1) Liquid temp, higher than air temp., approx 2 1/2-3hr test run.
(2) Sarr.e evap. rates. Slow evaporation rate, same liq. temps, and
low velocity differentia"! contribute.
PENTANOL MOLECULAR WT. = 88.1
l?A CONTRACT NO. 68-02-4248
¦HE JOS NO. 74 7-00 A
uATt: 4/5/88
Bv: K.O. 3roun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
ACETONE
AIR
VELOCITY
FPM
AIR
TEMP
F
TEMP. LIQ.
TEMP.BASE
F
NRE
NSC
VP
in. Hg.
ST
(10 -1)
lb./ft.
DIP
(io-4)
ft.2 /sac.
LATH
BTU/lb.
ViSL
(10-5)
lb. sec./ft.2
MEASURED
evap. rate
Ib./'hr. ft.
£V/kP. RATE fROU
EQ. USiNC
!b./hr. !'2
use
uw
100
52.9
32.1
44.1
4620
1.467
2.75
1.80
1.094
248.6
0.86
1.41
1.43
1.45
100
70.7
36.2
45.2
4340
1.471
3.00
1.78
1.159
247.7
0.84
1.55
1.57
1.56
100
99.6
37.1
42. C
3980
1.470
3.20
1.78
1.265
247 4
0.84
1.73
1.63
1.68
500
71.1
32.8
42.7
22380
1.470
2.80
1.80
1. 127
239.4
0.70
1.92
1.79
1--. 80
500
100.4
36.8
39.7
19910
1.481
3.10
1.78
1.256
247.5
0.84
1.94
1.99
1.99'
500
120.1
38.7
40.0
18690
1.485
3.30
1.77
1.334
247.0
0.83
2.06
2.12
2.12
500
141.1
40.9
40.4
17800
1.466
3.50
1.76
1.419
246.5
0.82
2.12
2.24
2.25
1000
59.3
30.0
42.1
45550
1.495
2.60
1.81
1.114
239.8
0.71
2.80
2.7 2
2.71
1000
109. 5
35.5
40.1
38,910
1.469
3.10
1.78
1.295
247.6
0.84
2.93
3.22
3.23
1000
126.5
38.0
40.1
37,060
1.470
3.20
1.77
1.360
247.2
0.83
3. 58
3.33
3.33
1000
134.4
40.7
40.4
36,130
1.469
3.50
1.76
1.395
246.5
0.82
3.60
3.64
3.65
EPA CONTRACT NO. 68—02—4248
ACETONE MOLECULAR WT. = 58.1 IH£ JOB NO. 747-004
DATE: 4/5/88
BY: K.O. Broun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
METHYL ETHYL KETONE
AIR
VELOCITY
FPU
AIR
TEMP
• F
TEMP HQ
TEMP BASE
• F
NRE
NSC
VP
in. Hg.
ST
OO"1)
lb./ft.
Dir
(10-")
f t.,J/ sec.
LATH
3TU /lb.
v;sl
(io-5) ,
lb. sec./ft.
u£aSURE3
Eva p. Rate
¦ b./Sr. ft.
E V)*?. RATE >"30"
EG. U5.SC
".2
f.SC
100
52.6
44.2
47.2
4660
1.671
1.65
1.78
0.950
215.4
1.04
0.84
0.82
0.33
100
69.6
' 48.8
49.2
4360
1.675
1.90
1.76
1.014
214. f.
1.01
1.01
0.95
0.95
100
100.3
57.2
58.8
3990
1.672
2.40
1.73
1.111
213.0
0.90
1.14
1.22
1.19
500
/it; n
36.9
41.9
24160
1.668
1.35
1.81
0.919
216.8
1.09
0.92
0.91
0.91
500
70.3
41.6
41.8
22390
1.675
1.50
1.79
0.988
215.9
1.06
1.17
1.02
1.02
500
100.9
48.7
48.2
20370
1.667
1.90
1.76
1.090
214.5
1.01
1.39
1.29
1.29
500
120.1
57.8
55.9
18990
1.668
2.40
1.72
1.170
212.9
0.96
1.68
1.63
1.62
1000
44.1
38.2
45.8
47860
1.664
1.40
1.81
0.930
216.5
1.08
1.35
1.54
1.54
1000
70.0
41.7
44.5
45430
1.674
1.55
1.79
1.002
215.9
1.05
1.67
1.71
1.70
1000
100.7
52.6
53.2
40050
1.625
2.10
1.75
1.106
213.8
0.99
2.04
2.28
2.31
1000
120.4
55.4
53.6
37860
1 .672
2.30
1.73
1.171
213.3
0.97
2.59
2.53
2.53
EPA CONTRACT NO. 68-02-4248
!HE JOB MO. 747-004
DATE: 4/5/83
BY: K.0. 3rcun
MEK MOLECULAR WT. = 72.1
N
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
2-0CTAN0NE
AIR
VELOCITY
FPM
AIR
TEMP
F
TEMP LIQ.
TEMP.BASE
F
NRE
NSC
V?
,r,. Hg.
ST
(10
lb./ft.
Dip
(io;«)
ft.2 / sec.
LATH
3TU /.b.
visl
C'O"5) .
;b. sec. /ft.
V.'. E A. SiJ ^ L O
EVAP. PA —
ib.A'. f:3
£vi,?. rate thou
£3. USING
\SC
500
89.4
87.6
87.0
20610
2.462
0.07
1.72
0.720
166.3
NA
0.13
0.06
0.08
500
119.5
107.9
106.2
18720
2.469
0.13
1.66
0.801
164.3
NA
0.24
0.11
0.10
en
mnn
44.4
(1M 48"7
liR) 48.9
48030
2.456
0.01
1.85
0.628
170.5
NA
0.04
0.01
0.01
1000
73.9
f1p\ 74.8
!1B' 75.1
43630
2.404
0.04
1.77
0.690
167.7
NA
0.10
0.05
0.05
1000
100.1
97.5
97.3
40050
2.459
0.09
1.69
0.752
165.3
NA
0.17
0.12
0.11
/ \ * • -r i j 4. ^ i -k • +• , u CONTRACT \'0. 68-02-4248
(1A) Note Air Temp, lower then liquid temp,-eoul ibrium not met, 3 hr. run .Hr Jog NQ 747_004
(18) Note Air TEmp. lower then liquid temp.-eaulibrium not met, 2 hr. run date': 4/5/68
OCTANONE MOLECULAR WT- - 128.2 BY <.0. 3rCo-
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
HEXANE
AIR
VELOCITY
FPU
AIR
TEMP
F
TEMP LIQ.
TEMP BASE
F
NRE
NSC
VP
in. Hg.
ST
(TO"3)
!b./ft.
D!F
(¦o-4)
ft.2/sec.
LA 7H
3TU/lb.
ViSL
(iu i ,
is. sec./ft.
*.'EaSU=.E3
EvaP. SaTZ
, 2
. t./.-.r.
EVAP. 9ATE FROM
ZO. US-NC
• b ^
-fc.sc
100
70.7
37.3
42.5
4470
1.873
2.13
1.38
0.885
163.2
6.25
1.91
1.82
1.83
100
100.2
39.6
41.5
3980
1.884
2.30
1.37
0.987
162.3
6.00
2.10
1.98
1.97
100
120.3
43.4
41.2
3780
1.881
2.55
1.36
1.042
162.2
5.60
2.28
2.19
2.19
100
140.1
49 .9
48.7
3540
1.884
3.05
1.33
1.109
161.2
5.03
2.32
2.62
2.62
500
45.1
32.2
41.7
24120
1.879
1.85
1.40
0.817
164.0
6.80
1.82
1.87
1.88
500
68.3
35.9
42.2
22260
1.882
2.10
1.39
0.884
163.4
6.40
1.98
2.13
2.14
500
100.3
40.8
42.7
20060
1.886
2.40
1.37
0.979
162.3
5.85
2.20
2.44
2.44
500
109.6
43.8
43.0
19200
1.887
2.55
1.35
1.022
162.2
5.55
2.56
2.59
2.60
1000
45.6
29.0
43.1
48440
1.881
1.80
1.41
0.813
164.5
7.20
2.45
2.86
2.88
1000
70.2
36.6
45.1
45010
1.890
2.10
1.38
0.901
163.4
6.30
3.05
3.35
3.35
1000
100.4
42.5
43.8
40980
1.850
2.50
1.36
0.977
162.4
5.70
3.89
3.93
3.99
1000
121.2
45.9
44.0
37930
1.884
2.70
1.35
1.037
161.8
5.40
4.49
4.30
4.31
EPA CONTRACT NO. 63-02-42-8
HEXANE MOLECULAR WT. =86.2 ;he JOB NO. 747-n04
DATE: 4/5/88
3Y: K.O. Broun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
HEXANE
AIR
VELOCITY
FPM
AIR
TEMP
F
TEUP. LIO.
TEMP.BASE
F
NRE
NSC
VP
in. Hg.
ST
(10-*)
lb./ft.
DlF
ft.5 /sec.
LATH
BTU/lb.
V1SL
CO"5)
lb. sec,/ft.
MEASURED
EVAP. RATE
lb./hr. f(.J
EVAP. RATE FROu
EO. US1NC
n?
NSC
500
120.7
46.0
44.2
18780
1.885
2.70
1.35
1.046
161.8
5.40
2.63
2.74
2.75
500
139.9
52.1
48.7
17710
1.888
3.20
1.32
1.107
160.9
4.80
3.33
3.25
3.26
innn
130.1
4B.b
49.4
36510
1.887
2.95
1.34
1.076
161.5
5.15
5.06
4.70
4.71
HEXANE MOLECULAR WT. = 86.2 EPA CONTRACT no. 68-02-4248
IHE JOB NO. 747-004
DATE: 4/5/88
BY: K.O. Broun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
N- HEPTANE
AIR
VELOCITY
FPU
AIR
TEMP
F
TEMP UQ.
TEMP BASE
F
NRE
NSC
VP
in. Hg.
ST
(10
lb./ft.
OIF
(10"«)
ft.2 /sec.
LATH
BTU/!b.
V1SL
(TO"5)
ib. sec./ft.
MEASURED
lVAP. RATE
Ib,/hr. ft.
= . RATE FSOu
EC. USiNC
ib ./nr. il?
N'SC
U*l
100
51.8
47.7
47.8
4750
2.207
0.75
1.46
0.707
160.4
0.99
0.57
0.71
0.70
100
70.5
50.2
50.0
4450
2.203
0.85
1.45
0.755
160.1
0.98
0.80
0.81
0.80
100
99.6
62.3
62.2 _
4030
2.207
1.23
1.40
0.833
158.4
0.91
1.26
1.17
1.16
500
45.2
38.7
40.2
24470
2.203
0.56
1.49
0.687
161.7
1.04
0.77
0.62
0.61
500
70.5
49.1
48.3
22470
2.200
0.80
1.45
0.749
160.2
0.98
1.00
0.89
0.87
500
99.8
6417
64.3
20060
2.206
1.30
1.39
0.837
158.1
0.90
1.58
1.39
1.42
1000
41.1
33". 3
43. 1
1
48860
2.202
0.55
1.49
0.689
161.7
1.05
0.92
0.96
0.94
1000
69.0
49.8
49.6
43690
2.208
0.85
1.45
0.767
160.2
0.98
1.48
1.49
1.45
1000
100.5
64.9
64.2
40230
2.307
1.30
1.39
0.835
158.1
0.90
2.21
2.34
2.22
heptane molecular wt. = 100.2
EPA CONTRACT NO. 68-02-4248
!HE JOB NO. 747-004
DATE: 4/5/88
BY: K.O. 3roun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
OCTANE
AIR
VELOCITY
FPM
AIR
TEMP
F
TEMP. UQ.
TEMP BASE
F
NRE
NSC
VP
i"n. Hg.
ST
(10-3)
lb./ft.
DIP
(io-4)
ft.2 /sec.
LATH
3TU/lb.
ViSl
(TO"5)
!b. sec./ft.
MEaSURED
Evap. rate
2
Ib./hr. ft.
EVAP. RATE fRCM
EO. USiNC
itj.Air. !t.^
KiSC
100
49.2
44.0
42.2
4800
2.442
0.19
1.57
0.633
160.4
3.25
0.17
0.19
0.19
100
69.6
55.8
53.5
4440
2.372
0.28
1.53
0.704
158.4
2.85
0.32
0.28
0.28
100
100.0
79.7
79.1
4000
2.371
0.60
1.44
0.781
155.5
2.35
0.64
0.60
0.61
500
45.0
43.2
42.2
24310
2.365
0.18
1.57
0.644
159.9
3.28
0.21
0.21
0.21
500
70.1
63. U
62.2
22240
2.366
0.36
1.50
0.704
157.5
2.58
0.48
0.42
0.42
500
100.4
81.4
80.5
19970
2.369
0.63
1.43
0.783
155.3
2.25
0.82
0.73
0.81
1000
44.4
44.6
45.6
48920
2.328
0.19
1.57
0.651
159.8
.'.22
0.30
0.35
0.34
1000
70.2
61.4
60.9
43560
2.464
0.34
1.50
0.690
157.8
2.65
0.53
0.61
0.62
1000
99.6
79.7
79.4
39760
2.370
0.61
1.44
0.786
155.5
2.15
1.06
1.11
1.11
OCTANE MOLECULAR WT. = 114.2
EPA CONTRACT NO. 68-02-4248
!HE JOB NO. 747-004
DATE: 4/5/88
BY: K.O. Broun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
BENZENE
AIR
VELOCITY
FPM
AIR
TEMP
F
TEMP. LIQ.
TEMP.BASE
F
NRE
NSC
VP
in. Hg.
ST
(io-3)
Ib./fi.
DiF
(10-*)
ft.2 /sec.
LATH
BTU/ib.
VISL
(io-5)
lb. sec./ft.
MEASURED
Ev'AP. PA'1
ib./hr.
EVA P. RATE FROM
tO. U&NC
i b./n r. f: ^
N'SC
U H
100
70.1
(1) 40.6
42.4
4450
1.737
1.35
2.11
0.959
191.4
1.69
1.12
1.13
1.11
100
100.3
44. /
43.1
4010
1.754
1.50
2.09
1.054
190.8
1.64
1.27
1.30
1.27
100
120.4
50.1
48.9
3800
1.739
1.80
2.07
1.122
190.0
1.57
1.59
1.57
1 .59
500
78.7
44.5
4fi .
21370
1.726
1.50
2.09
0.997
190.8
1.64
1.47
1.41
1.42
500
101.1
51.5
52.6
20030
1.741
1.85
2.06
1.062
189.9
1.55
1.82
1.77
1.75
500
120.3
55.6
54.4
18650
1.742
2.10
2.04
1.141
189.2
1.51
1.96
2.01
1.98
1000
44.9
(1) 40.1
56.8
48810
1.726
1.3u
2.12
0.879
191.5
1.69
1.83
1.79
1.80
1000
79.3
(1) 41.b
42.6
42800
1.732
1.35
2.11
0.999
191.3
1.68
2.08
1.87
1.87
1000
99.8
47.9
47.3
39940
1.740
1.70
2.08
1.066
190.4
1.60
2.37
2.38
2.35
1000
117.5
bb. 1
54.1
38360
1.736
2.10
2.04
1.112
189.3
1.51
2.72
2.93
2.90
EPA CONTRACT NO. 68-02-4248
(1) Liquid temperature below freezinq. (H£ JOb no. 747-004
BENZENE MOLECULAR WT. = 78.1 DATE: 4/5/88
BY: K.O. Broun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
TOLUENE
AIR
VELOCITY
FPU
AIR
TEMP
F
TEMP. LIQ.
TEMP BASE
F
NRE
NSC
VP
in. Hg.
ST
(10-J)
lb./ft.
DIP
(to-;)
ft.2 /sec.
LATH
BT'J/Ib.
VISL
Oo-5) ,
lb. sec./ft.
MEASURED
EVA.P. P.aTE
ib./hr. fi.
EVAP. RATE FROM
£Q. USiNC
lb ,/r^r.
K'SC
100
54.4
53.2
51.9
4720
1.933
0.54
2.02
0.812
180.0
1.35
0.35
0.50
0.48
100
69.5
51.3
49.4
4440
1.935
0.52
2.03
0.862
180.2
1.36
0.40
0.48
0.46
100
99.5
69.9
69.9
4040
1.938
0.91
1.95
0.946
177.8
1.22
0.88
0.91
i
0.94
500
44.8
41.9
43.0
24150
1.934
0.38
2.07
0.793
181.4
1.45
0.42
0.45
0.45
500
71.7
54.7
54.5
22080
1.930
0.58
2.01
0.869
179.8
1.33
0.56
0.68
0.68
500
100.0
71.1
70.8
19330
2.002
0.94
1.94
0.957
177.7
1.21
0.98
1.19
1.11
1000
44.3
4^. "J
45.9
47550
1.937
0.39
2.07
0.804
181.3
1.44
0.74
0.68
0.67
1000
69.9
54.4
54.4
44590
1.932
0.57
2.02
0.860
179.8
1.34
0.85
0.98
0.98
1000
100.0
70.2
69. 9
39700
1.935
0.92
1.94
0.964
177.8
1.22
1.51
1.59
1.59
TOLUENE MOLECULAR WT. = 92.1 EPA CONTRACT NO. 68-02-4248
IHE JOB MO. 747-004
OATE: 4/5/88
BY: K.O. 3rcjn
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
XYLENE
AIR
velocity
FPU
AIR
TEMP
F
TEWP. L10.
TEMP BASE
F
NRE
NSC
VP
in. Mg.
ST
(TO"1)
lb./ft.
ft.: /s«c.
CD
VISL
OCT5)
lb. sec./ft.
measured
Evap. RaTE
it>.A', 't.2
evap. rate fROu
£0. USNC
Ib./hr. ft2
NSC Uw
100
100.0
83.4
83.2
4040
2.112
0.33
2.01
0.S6S
175.0
1 . 55
0.40
0.35
0.35
100
120.4
92.3
91.2
3760
2.111
0.43
1.96
0.933
174.0
1.47
0.64
0.47
0.48
.... -
500
45.8
46.0 ¦"
44.4
24310
2.110
0.10
2.17
0.722
179.1
1.99
0.13
0.14
0.14
500
70.3
64.3
53.3
22110
2.112
0.18
2.09
0.793
177.1
1.75
0.22
0.2 5
0.2 6
500
100.2
80.2
78.1
20000
2.114
0.30
2.02
0.S76
177.4
1.58
0.46
0.42
0.43
500
120.3
96.3
95.8
18790
2.114
0.48
1.96
0.933
173.5
1.42
0.85
0.68
0.69
1000
44.2
[ 45.2
45.9
47610
2.114
0.09
2.17
0.736
179.3
2.01
1
0.18
0.19
0.19
1000
69.1
63.6
63.2
44640
2.116
0.18
2.09
0.?$i
177.2
1.76
0.34
0.37
0.38
1000
79.8
72.8
72.1
42780
2.115
0.23
2.06
O I
CO 1
t—t
iCJ
176.2
1.66
0.55
0.48
0.48
1000
99.6
tfb. 4
85.0
39700
2.117
0.35
2.00
O.SS!
174.8
1.52
0.85
0.73
0.73
XYLENE MOLECULAR WT. = 106.1
EPA CONTRACT NO. 68-02-4248
IHE JOB NO. 747-004
DATE: 4/5/88
BY:
K.O. Brci.
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
XYLENE
AIR
velocity
FPU
AIR
TEMP
F
TEMP. LIQ.
TEMP BASE
F
NRE
KSC
VP
in. Hg.
ST
(10 -1)
lb./ft.
DIF
Ocr<)
ft.2 /sec.
LATH
BTU/lb.
VISL
(io-5) 3
lb. sec./ft.
MEASURED
EVAP. RATI
Ib./Hr. 't.
EVAP. RATE fROu
EQ. USNC
lh./hr.
use
1000
120.3
100.2
QQ.fi
37610
2.115
0.54
1.94
0.931
173.1
1.39
1.20
1.12
1.13
XYLENE MOLECULAR WT.
106.1
EPA CONTRACT NO. 68-02-4248
!HE JOB NO. 747 — 004
DATE: 4/5/88
BY: K.O. Srcun
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
WATER
AIR
VELOCITY
FPU
AIR TEMP.
DEW PT
F
TEMP. UQ.
TEMP.BASE
F
NRE
NSC
VP HQ.
VP AIR
in. Hg.
ST
(10-1)
lb./ft.
d;f
00-4)
ft.2 /sec.
LATH
BTU/lb.
V1SL
OCT5)
lb.-soc./ft.
MEASURED
EYAP. RATE
Ib./hr.-ft.2
EVAP. RATE n?OU
EQ. USMC
lb.A"", ft?
NSC
MW
100
49.6
33.9
50.2
49.7
5184
0.616
0.37
0.19
5.08
2.547
1065.2
2.61
0.02
0.C2
0.02
100
69.3
34.2
61.2
60.9
4235
0.617
0.54
0.20
5.02
2.695
1059.0
2.30
0.075
0.07
0.07
100
100.3
35.8
72.0
72.5
3557
0.617
0.79
0.20
"4.96
2.996
1052.9
2.07
0.10
0.10
0.10
500
45.0
36.0
45.0
45.0
24694
0.617
0.30
0.21
5.11
2.490
1068.2
2.74
0.05
0.04
0.04
500
70.7
38.0
62.4
62.2
22264
0.618
O.b/
0.22
5.01
2.719
' 1058.3
2.28
0.14
0.15
0.15
500
99.9
50.0
73.6
73.6
19664
0.620
0.84
0.35
4.95
2.992
1052.0
1.96
0.20
0.20
0.20
1000
4b.4
33.0
4b.8
46.1
36832
0.618
0.31
0.19
5.11
2.494
1067.7
2.72
0.07
0.08
0.08
1000
69.2
34.9
58.9
59.2
44076
0.618
0.49
0.20
5.04
2.714
1060.3
2.36
0.21
0.19
0.19
1000
99.8
45.0
73.4
73.2
48514
0.619
U.83
0.31
4.95
2.992
1052.1
1.99
0.30
0.31
0.31
EPA CONTRACT NO. 68-02-4248
IHE JOB NO. 747-004
DaTT 9/7? /PR
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
1—HEXANOL
AIR
VELOCITY
FPM
AIR TEMP
F
TEMP. UQ.
TEMP BASE
F
NRE
NSC
VP
in. Kg.
ST
(10-3)
Ib./ft.
DIF
(10"<)
ft.2 /sec.
LATH
BTU/lb.
VI5L
CO'5)
ib.- sec./ft.
UEASUREX)
EVAP. RATI
lb./hr.- ft.2
EVA?. RATI FROW
EQ. USJNC
IbVnr
NSC
MW
100
70.0
69.4
66.9
4422
2.115
0.0295
1.79
0.792
260.4
NA
0.0114
0.0197
0.0195
100
99.8
89.7
87.8
4040
2.114
0.0709
1.73
0.867
256 8
NA
0.0331
0.0344
0.0342
100
119.6
107.1
106.5
3788
2.107
0.1378
1.68
0.928
253.8
NA
0.0534
0.0522
0.0524
200
71.2
70.0
69.1
8782
2.119
0.0307
1.79
0.796
260.3
NA
0.0152
0.0174
0.0177
200
98.6
91.9
91.2
8124
2.110
0.0768
1.72
0.864
256.4
NA
0.0393
0.0363
0.0368
500
71.6
70.5
69.8
21853
2.126
0.0315
1.79
0.797
260.2
NA
0.0179
0.0325
0.0327
500
102.9
93.6
91.7
20038
2.112
0.0827
1.72
0.875
256.0
NA
0.0664
0.0700
0.0708
500
119.7
108.7
108.6
18919
2.121
0.1467
1.67
0.923
253.4
NA
0.1126
0.1125
0.1120
1000
71.3
70.4
70.0
43390
2.147
0.0315
1.79
0.795
260.2
NA
0.0363
0.0523
0.0513
1000
99.6
92.7
91.5
40285
2.123
0.0787
1.72
0.866
256.3
NA
0.0896
0.1069
0.1167
1000
120.0
111.4
110.5
37839
2.121
0.1594
1.66
0.923
253.0
NA
0.1978
0.1888
0.1877
1-HEXANOL MOLECULAR WEIGHT =
NA = NOT AVAILABLE
102.18
EPA CONTRACT NO. 68-08-0112
PACE PROJECT NO. 890501.315
DATE: 9/69
BY: K.O.B.
~ - CALCULATED VALUES FOR THE 200 FPM CASE WERE DETERMINED USING THE 500/1000 FPM EQUATIONS
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
1-HEPTANOL
AIR
VELOCITY
FPM
AIR TEMP.
• F
TEMP. UQ.
TEMP.BASE
• F
NRE
NSC
VP
in. Hg.
ST
(10-3)
lb./ft.
DIF
(10—)
ft.2 /sec.
LATH
BTU/ib.
'
V1SL
(io-3)
lb.- sec. /ft.
MEASURED
EVAP. RATE
Ib./hr. ¦ ft*
EVAP. RATE rsou
EO. U9NG
Ib./pr. V?
NSC
WW
100
70.6
70.0
70.0
4420
2.2S^
0.0033
NA
0.734
236.7
NA
0.0068
0.0058
0.0058
100
99.8
93.1
92.5
4050
2.280
0.0137
NA
0.803
233.3
NA
0.0184
0.0145
0.0143
100
118.0
102.8
100.5
3820
2.280
0.0233
NA
0.854
231.9
NA
0.0219
0.0203
0.0202
500
70.5
70.8
70.7
22050
2.288
0.0035
NA
0.734
236.6
NA
0.0080
0.0066
0.0067
500
100.4
92.9
92.4
20060
2.274
0.0135
NA
0.812
233.4
NA
0.0227
0.0193
0.0198
500
120.2
109.7
108.9
18810
2.274
0.0333
NA
0.866
229.2
NA
0.0483
0.0401
0.0408
1000
46.7
50.1
50.7
43200
2.510
0.0008
NA
0.683
239.6
NA
0.0064
0.0039
0.0032
1000
70.4
70.6
70.5
44040
2.297
0.0034
NA
0.732
236.6
NA
0.0088
0.0102
0.0103
1000
99.4
93.8
92.0
40810
2.277
0.0142
NA
0.797
233.2
NA
0.0350
0.0317
0.0324
1000
120.2
111.2
109.8
37500
2.273
0.0359
NA
0.869
230.6
NA
0.0831
0.0669
0.0680
EPA CONTRACT NO. 68-08-0112
1-HEPTANOL MOLECULAR WEIGHT = 116.21 PACE PROJECT NO. 890501.315
NA = NOT AVAILABLE
-------
EVAPORATION RATE OF VOLATILE LIQUIDS
DATA SUMMARY SHEET
2-OCTANOL
AIR
VELOCITY
fPM
AIR TEMP
F
TEMP. UQ.
TEMP BASE
F
NRE
NSC
VP
in. Hg.
ST
(io-3)
lb./ft.
DIF
(10-)
ft.2 /sec.
LATH
3TU/lb.
VISL
(10-*)
lb.-sec./ft.
tCASURtO
EVA P. RATE
lb. Air. • ft?
EVAP. RATE rsou
EQ. USING
b./V 'ft?
NSC
MW
100
70.8
70.3
70.5
4400
2.437
0.0041
1.80
0.691
260.3
NA
0.0080
0.0079
0.0079
100
99.7
95.1
95.3
4030
2.400
0.0180
1.72
0.766
255.8
NA
0.0234
0.0195
0.0203
100
119.6
111.7
111.4
3760
2.437
0.0423
1.66
0.808
253.0
NA
0.0308
0.0350
0.0351
500
70.8
70.4
70.0
21960
2.437
0.0041
1.80
0.6S2
260.2
NA
0.0088
0.0087
0.0085
500
100.7
90.9
89.6
20030
2.433
0.0142
1.73
0.760
256.6
NA
0.0245
0.0238
0.0242
500
120.1
110.0
108.6
18920
2.432
0.0388
1.67
0.805
253.3
NA
0.0509
0.0536
0.0540
1000
72.3
71.8
70.1
44050
2.437
0.0045
1.79
0.690
259.9
NA
0.0142
0.0148
0.0151
1000
100.3
93.5
91.2
40290
2.429
0.0165
1.72
0.757
256.2
NA
0.0368
0.0419
0.0427
1000
120.1
112.9
113.0
37560
2.432
0.0446
1.66
0.811
252.7
NA
0.0849
0.0941
0.0947
EPA CONTRACT NO. 86-D8-0112
2 - OCTANOL MOLECULAR WEIGHT = 130.23 PACE PROJECT NO. 890501.315
NA = NOT AVAILABLE DATE: 9/89
BY: K.O.B.
-------
AEEEMPJULB
AIRJ'aC)Cm,^B_ANiL_LIQUIIIJLEMPJRAIU.RLmPi[S
ut
-------
-------
-------
CJ
O
-------
iiiiilillililiiiilliili
11'') 1'' ¦ • i ^ ; : l: 'f ¦ ? j •
iill j ui; | li:. 11UU
¦j-1-n*pii I p r^| H7;1 '
-------
-------
-------
i'iili'ij'iji'li'j1;:;!!;!!!
-rtrr. , .
Mil!
! I
il5plip
I:-: ti" ¦;';: ;r'tS-• :t-¦¦ ii:"-•:
• ill; '
...C>
iijlliiii
f&i:
... ^WhI
4^njl-
liiunjip
l:: tr1 '¦
triK^r
liili
)1IS
iUi
lit
I'a
i
Kiil
1
Ijiilii'L
ii>a
LLilliLil
-Ui
ill
<:£- s^rerf-\
:: Xj- ¦
¦ ::; | O
.J.: i t i t •' i: l
:i:; i:
_^N- ; —'
!i!
iFIiinir
in;
iiiiUililliUiiLiuu
IhlliUi
ilii ijj:
'-i M :!:i
Ml
MSMm
Mill 1111 '
jig
11
:Tn int
I • H t f I
'liilil!
il'i
!i!»
uii
li-ililiiiiiJi'"'""' '
¦ , I "
tijil-jii
::I:I:::: !!i:^ lilti::' — i::iMi:ii'!•!il;¦ I| •
1
:ii!
IliMiiiiiii
ij jjljlji.
Hi Bife
ii::; illi'.j!::;':
Lu—-Iiliil i;:
' "i ;ii!ii;;:
! •. ! -I
liiliite
¦ ¦ i V>»!' • I ,k.. «¦„ .¦ ¦ I . . » , -
j;:; i
itrn^trr^TiTrr
i'H! Ilii
U;:i|!ll!i;!;|:iU!il:;
tj;iu;.i¦
iiiiilMillii
.iJliilMiiiiiilli
lMiiiiiiMll.
IMIMMM.
i;i;11:7;11;;;::;;.: ;i:ji
"j-! i-i-f-f) i-J -
l!|jljij
;;;;! i :t; ! s; m:
' 11'::':1 11:::
i•r •: ;1~
Juiiiiliii.
-------
-------
-------
-------
i!I;j£:vA'p6 f^A T 'iOM ¦ ~fcA-TE :¦: lb /SG-FT,
' 4 ' i •• 1-" -1 : ¦ . i. ¦ i ¦ . ¦ i ¦ : ¦ > i l A. •!- l • 1 t
. • i,.:
Ui_
1 ,.l, , k ¦ ¦ *
6
..ft
MfJI!
1 •:11 i i I::
11:: |:rj: (t'Ii :;i:.::i
ijj.J.i—frj-
~*r
~-s-
..Jhu
-1
i:::
1 •:•!_¦•;••• r,
_^_u.
-------
iiiiii;;::: viiiji J :1;ii^rvt: vj!:/i:: '
^XlOK;}::$;AY££ . A # /HAr. ¦ £J
_i—f t' ,i
( *3
o
"W
¦ I k
ilii lililiii
n|: TuT I I;
I ii
ill!
11 '¦;
lili
iiiiHiiLp
Si'iiv'p'-Ti^r , .i ••• , j...,
-i-
i.
hr
i.! •
i i«11 t
!#
i|!|
UiiJUUi
nTiT i:ii
!Ul
fm-
;;u
iilMillSp
h;
i I: I ¦ I; ij lit!
iiiililiiiLj iiilU MlliilMSlilMli
iiii ii;
III: pi} ;!p
•• • lit* i »
i 'j;i »Mi !*«
i;*
iiilliiii
I-":..:::: ^ ¦- I
sun;
I::'.!:::'
.;Xi:
r>>:
SMS
ftferivi': ST
i;i;, :I :i :ni :, ,,,
)!»' 11f i I ¦::: I::! t j11; i I:! i; j;:! i I:; i:
r »
ilillilii'ii-iiJiuiiijililfllipii!;
.i: pii;-::': j : i. i
1,'ii i:in
¦JiLiL
litepiislsit
•iii
ilii |ii;li',;! i1 1 .TT1 jMi TTTiTT^•:i;iii::ij.i
j a- •• :- ¦ i!::ii;I:I;f ' !¦' ¦ ;'i —':I
•'i!; ;!i! • I.
.|i. ::
1::::l I: i: • !¦¦¦
u.
III!
Ill
ilpiriiMp
iiiiliiifii!!
Mi
•!!ii i I
ii;-h*M: iiii !i.: :
si'•!'-|;Svir;;;-"
-1:" i.s: :;; -; t_::
!!=!
~ • ii • | • • . ' ¦ « • | T . . • f
' i • i: 11 ¦:;;
!-Sill1 ill'
liii'!'::':;:!;!.;;:
irr' li'i'l ilii
iMliiLHSfHTiHirhltrlTPT
::i:i:ii ::;t
¦ ! *•
UU-L
: 1::.
: i j-tltil i:
::::j:: ::i::! :
.: i.::; r::";
illSiim
''iil.'jijjlii:
:!i!
iiiillii
• I::;
lili
:;:. k t i! J;:::»':: - :::: s i i:: i! i:.
¦ ' : I ' ! ! ? 1 ! I ' I 1 1 - * T . 1 ' ' ' 1 ' : 1
: 1;::
-------
-------
-------
-------
-------
-------
-------
; C>
"I.:
iiM
Hiiif
B
ill
liiJlliii
;iiULiU
liiiiii;
m
nim;;
m
ui; |
"HI1"
rrttnrr
:!!•;!
iliilliii
l!l'fi!il
Ti H:i
ij; iiiiiliiii i'lliiii
till
5 !!
ii;!
iiii'.ii
III
Hi!
iiii
!l Hi HI!
1
liil.ijiiliill
iiii
feijiQiii
mwm
;i:: »«iSj|:lX>;i
* i'n i mi
I
iillli
»» USIiilit::;;-
SHI
¦!l: jjllij!!.
11:: J
;:u
¦l::],:: i::!:
1B1
iiii'li!
I
!ii!
I!
jiiliili
Jirllii
Si!
ii
H ii
u;l
iiHjili;
!l!l
iiii
ft!
m;
iiii
iiiili'il ijiil'-ilill'iiliiiil
¦': i:1:: I: :¦
ill
iiii
::iil '¦
~ ::L
lijii iiijiii!!
i'iilliiriiii-iiiii iiii
MiliiiiliiliSl
j j|ij:Ii:iI: =:j i:1;:l:j:;'Ii: j jjji:• • ji;!'j:;jS:• :j::'',11* 'V ¦: '|:;: !j: :.j:
:i i:::
i}iliiiii!;ii-L::ii:ii!l!i;
if ¦
Iiii
•f
ill'iillL
Hi
iiii
Miijir
-u'^y-I'-ii
¦tlTVMjiti iiiiii-niim
I
iiiiMiMiiiii
iiiihli
miL.
iiliiijiiliyiliiilltiiiiiii-lliii i'ii ii!ljjii;j::;!!;;::|i:::j!:iii: ;ii!ijli;ii:: jj: jl::'j!:::i: ! ¦i i": j: :i :j -:: :l
[tf-X? i.: iUaiiiiiiOl
!-"-1
: i i
liiiliilikiiilul- j|j!i!h!|i:S;;i|! !!i!{Mjiii&^
illuE Sp=Libi;li!i}13U ^
liTTTTT: :.;i ,u;^:::. | ' V:;-Uj. -.pv ¦; ¦¦ ¦
LiiiL'iiliii'iit:l • vli:Ljtii• •V
mdl: !¦;-! Lii-j U :Tvli iU
;^!;iii!!i;!ii.;?r^:' iimi'f;!!:-ixiiiiiiiii1:;
•1
; :¦i¦ . I..;.! .
j j;;* j j;.;j;jIi j j:;j j L-:;!^' i; ;••;j:;;; j:;: jj. j:: j ;.; j..; ;jI;::': j \;:} •;:' |:::: j. . j. -
riliiliriiiii-.iiiii:
ill HIT1'!'
tutihiHi
«!!ii;jii|ijji
!;!:i!i
:i:l : !'i:i
i;n il ; lit
ttTirrrnfttfT
M: t'ju ¦ i;j:s;•t;
iilUM juillill
Ivi'li'rlM'ilii'iili;
iiiiijli
iiiiiiiii
iiii
iilijiil!
:UU.i
i'iiiliiiiliiiii
u; iiii;Iiiiiiiii^
iiiiiiiiiiiijii::::;:;!.!. ;i:iiiii:i/^:ii'i;
11.U, • i• h 11!!::i I; 1;:'
!:!;: i:':, i:.: i
i:: i i :7t rrr
-------
: . J .: _l_
5
J._Ui_i
<>
-t.—i—-i,:. I
Cx>
"LS"
-fx
&
f-.
z
iiiijiiip!:! :::
lU.Uil
iliSiilllilii
lite
;!i:li;!i
:-:;ii"i
Jjsi&liiii.
ili:&: iiiiiiii:
::i
iiiUi
"ill':'
U." !•!¦
niili
;1T;: (
JJOII
illii::i!!i!!!li!ii
.llliiLULiliJ
ii i^ ' _
^|;|i:! ¦ p¦ -1_11j
-------
-------
-------
-------
-------
-------
-------
-------
t
X
m
x
>
2
o
r
-s.
-------
¦i i . ;
¦ ' !
sa rr;
:¦« I
. o
-1—
*
:*!
.XL
; ¦ i
i: .-¦!
Ht < ¦
P
o:
» . I ' .
.... r|;
r7:T
4LU
• I
*t "
: j.
i
iv
: :>»
i"'
:.L
f:: }:•
: ¦ : :l
:G?
m
I: H
I::::
ill!
£•
•0!
hf:
v::i:
a,b
-------
-------
-------
-------
APEfNDDLC
£WATL0^C^LCUL^JED.VIRS.ES_E^E£RIi!£NTALLY_RLA5UR^D
DMA.
ff
-------
ALL CHEMICALS1'
SOO £. 1 ,000 PPM DATA
MOLECULAR WT. EQUATION
JZ
4.53
3.50
figure CI-/5
.500 1.50 ' 2.50 '' 3.^0 1 4.1)0
Calcula^od Evap. Rates(1b/hr ft^
Cvap. Rate = O.U0O237 (MW)(VP)(V)°-525
ALL CHEMICALS ^
500 & 1,000 FPM DATA
SCHMIDT NO. EQUATION
FIGURE Cl-
2.58 ' 3.ii8 ' 4.U '
.500 1.50
Calculated Evap. Rate, Ib/hr ft^
Evap. Rate = 0.00725 (NSC)1-45 (VP)(V)°'64
"ALL CHEMICALS" does not include water or low VP alcohols.
CI
-------
ALL CHEMICALS '
100 FPM DATA
MOLECULAR WEIGHT EQUATION
FIGURE C2-h
1. 50 2.59 ' 3^F" ' 4.)0
Calculated Evap. Rate, Ib/hr ft
Evap. Rate = 0.0069
(MW)1-08^)0'98
All CHEMICALS ^
100 FPM DATA
Calculated Evap. Rate, Ib/hr ft
Evap. Rate « 0.312 (NSC)1'55^)0,99
"ALL CHEMICALS" does rot include v.'ater or the low VP alcohols.
-------
i.33
SC5 i 1 ,«C0 fw C*T".
MOUWlill? yr. (QUAliC-N
"¦ FIGURE
I. <3
i. 58
?SS
U1L
.,ir: i.
Jh ;
../X.
,23« I. £'3 i.ii 1,88
CalcuiSierf £vjo. Rite. Ikskr ft?
Emp. Sate • 0.0008617 (HW,ia'9-:vP)(V'3°'S
•"iatc: Fnt»j".o'! ' iOO i'90 r?H iita »ai jis-
ALttBK&S
500 & 1,006 f?S EftTA
S£lfw'-5T HC EQUAT;SH
H?
FlCUftt €3-
.596-
.m
¦s
A
mi
—; 1—¦—f-—¦—p>—f-——(—¦*-£—*
.ik m% i.aa us t.w
C-3''ci/l4te4 Eyap, Kits, lyi""
Evip. Rite = 0.O3S? (H5CV''a(VP5{v;G-46
*ttote: Pentasio? iZOO 3n4 1430 fpf? used
-------
9GQ
s:
.n
.
UJ
x:
QJ
i-
,700 -
L. 500
.300 T
QJ
2C
. 1B8 J-
ALCOHOLS
IGO FPM DATA
MOLECULAR WEIGHT EQUATION
: : : : t
¦ \ \ '/\ :
: : ; : yS-f ; ;
i 'y/J: : :
¦; ; : : :
'yi, ; ; : :
i i ! i i 1
'IfiURE C4-A
.300
.500 .700
Calculated Evap. Rate, lb/hr ft
Evap. Rate = 0.0487 (MW)°'59(VP)0'89
.909 1
-.700 -r
03
Cxi
-.508
'O
Q>
rry
QJ
2:
.300 -
.100 -
t
:
ALCOHOLS
100 FPM DATA
SCHMIDT NO. EQUATION
¦i
*/ :
,100
.300
.509
T?W"
.300
2
Calculated Evap. Rate, lb/hr ft
Evap. Rate » 0.39 (NSC)°-79(VP)0'88
FIGURE C4-E
-------
KETONES
500 & 1 ,000 TP/! DATA
MOLECULAR WT. EQUATION
.. : : .... ;
1
. • • /
: : : : V' :
; • v....:
. /¦
\ •: j
¦
\ .: ;
: *•'*" i : :
... : j ;
# : • : :
r ' | , ,
/ • • '
V'
; i i. . J i
—
¦
FIGURE C5-A
Calculated Evap. Rate, Ib/hr ft^
Evap. Rate = 0.002998 (MU)0•25(VP){V)0*7
KETONES
500 & 1,000 FPM. 0A7A
SCHMIDT NO. EQUATION
Evap. Rate * 0.006947 (NSC)0-45 (VP)(V)0•70
C3
-------
1,88 --
1.40 -
1.89 -
.m -¦
KETONES
100 FPM DATA
MOLECULAR UT. EQUATION
: *
; :
' ; ; :
: : : :
1.40
.280 .680 1
Calculated Evap. Rate, lb/hr ft
Evao. Ra te = 2.212
2
rIGUKE C6-A
(mw)"°-34(vp)0-95
.(09 -
.209 -
.209
KETONES
100 FPM DATA
SCHMIDT NO. EQUATION
-+¦
: : : :
: : : * :
: : : / :
; U*' I
;
i/i j
\
\
.. /r*
/
r
: : : :
;
: : : :
j
• ; | ;•
; : : :
1.1
FIGURE C6-G
.680 1.99 1.49
2
Calculated Evap. Rate, lb/hr ft
Evap. Rate » 0.517 (NSC)"n-138(VP)l-°f
-------
4,5a --
JZ
o
O.
*3
>
T3
CD
U
73
wo
OJ
>:
3.50 t
2.5G
1.59
.502
AL 1 PHATICS
500 1 1.000 FPM DATA
MOLECULAR WT. EQUATION
*•
i'
* "
*
: /\
¦v-
FIGURE C7-A
* .¦>'
¦;r
.508 ' 1.50 a.'50 3. jQ 4.^8
Calculated Evao. Rate, lb/hr ft'
Evap. Rate = 0.002412 (MW)°"45(VP)(V)°*65
4.53 f
3.50
I.52
1.50-
ALIPHATICS
500 & 1,000 FPM DATA
SCHMIDT NO. EQUATION
:/ :
V/ •••
: V:
: /***.
t •/
r>r
A\
:
.500 4-...
FIGURE C7-D
/ •
—i i 1 1 1 1 ! i—
.500 1.50 2. 50 3.50 4.ij0
Calculated Evap. Rate, lb/hr ft2
Evap. Rate = 0.01221 (NSC)0•&(VP)(V)°-6S
C7
-------
0J
*->
cc
•o
V-
3
4,59
3.50 --
2.50 T
1.50
500
AUPHAIICS
100 PPM DATA
MOLECULAR WT. EQUATION
: :
: \/
*
\ y.
* / '
*iJ
M
/.
V ¦
*
FIGURE Co-A
. 500 ' 1.50 ' 2.'50 ' 3.1)0 1 4.^0
Calculated Evap. Rate, lb/hr ftf
0 0
Evap. Rate = 0.0592 (MW) (VP)
sz
¦O
QJ
U
=}
to
-------
4,50
AP.OHAT ICS
SOO £ 1,000 TPM OATA
MOLECULAR WT. EQUATION
FIGURE C9-A
3.50
.500 1,50 2.58 3.^0 08
Calculated Evap. Rate, lb/hr ft^
Evap. Rate = 0.0000862 (HW)1'35(VP)(V)°"55
AROHATICS
500 S 1,000 FPM DATA
SCHMIDT MO. EQUATION
C9
-------
AROMATIC!)
100 TPM DATA
-1.88 "
-O
jl.40
03
a:
21.00
*o
CJ
u
§. 600
u
.208 -f
: : : ;
: : :
. y
: : /
*
. y . .
\Y\ : : -¦
¦ s
: : :
7- : ¦
j j ;
. . : v vi ......
: * :
. » • >
j j j
—i—i—i—i—f-
- -j
i
.200 .600 1.1
Calculated Evap. Rate, lb/hr ft
Evap. '?jte = 0.000136(MW)l,98(VP)1,26
F1GURL CIO-A
1.88
JZ
n
*o
a:
1.40
O.
<¦0
>
1.00-
TD
OJ
U
=3
Of
£ •
.200 -
AROMATICS
100 PPM DATA
SCHMIDT NO. DATA
/
TifciF
A
"i:
1,1
l.<
-t-—h
1.1
FIGURE C10-B
Calculated Evap. Rate, lb/hr ft
Evap. Rate ¦ 0.2686(NSC)2(VP)1-12
CIO
/'»
-------
.500
409
309
^ 200
a
LkJ
cc
ZD
oo
S100
I h
WATER
500 & 1000 FPM DATA
MOLECULAR WT. COUATION
%
/
/
A
FIGUr\L CI I - A
.100 ,200 ,306 .400 .500
Calculated Evap. Rate, lb/hr ft^
Evap. Rate = 0.0006447(<-IW)(VPl-VPA)0.89(v)0*56
500
C\J
4-J
-.300
<
a:
<.200
Q
UJ
£.100
Ui
/*
WATER
500 & 1000 FPM DATA
SCHMIDT NQ. .EQUATION
ih
FIGURE Cll-[
.100 .200 .300 .400 .500
Calculated Evap. Rate, lb/hr. ft^
Evap. Rate = 0.01760 (NSC)(VPL-VPA)°-89{V)°-
Cll
a*
-------
2&3 -
U
£2
' 4>yrK —
. - i t' I? ,
120 I
<
ctr
O-
<
Q
LjJ
CX.
£.848
LU
:< -L
WATER
100 TPM DATA
MOLECULAR WT. EQUATION
Figure C12-A
.040
.120
.160
I W i U I UU U I i u W I i, V <
Calculated Evap. Rate, Ib/hr.ft?
Evap. Rate = 0.0115 MW(VPL~VPA)1 *20
.200
.200 T
(\J
+J
.160 --
s_
-C
-Q
LxJ
,120 --
J—
<
en
T
a.
<
>¦
UJ
.030 1
Q
I
Q£
rD
oo
<
.040 --
e:
/
.040
.120 .160
Calculated Evap. Rate, lb/hr.ft2on
Evap. Rate = 0-3344NSC(VPL-VPA)x'
Figure C12-B
.200
C-12
m
-------
(N
4J
U-l
)-l
JC
XI
nj
a
a
ra
>
w
TI
a;
u
3
U)
a3
CJ
£
180
140
LOW VAPOR PRESSURE ALCOHOLS
500 6. 1000 FPM DATA
MOLECULAR WT. EQUATION
(1
*
/
/
: /
'
/f
*
/.
* ¦ "• FIGURE C13-A
Calculated Evap. Rate, lb/hr ft
Evap. Rate = 0.00001544 (MW) ' (VP)u*ou(v)
¦p
u
.c
0)
-p
X
a
(O
>
w
T3
-------
CN
4J
4-1
V-i
x:
v
+j
n3
os
,870 r-
.050 -'
LOW VAPOR PRESSURE ALCOHOLS
100 FPM DATA
MOLECULAR WT. EQUATION
1)
. FIGURE C14-A
.010 ' .030 ,050 .070
Calculated Evap. Rates, lb/hr ft
Evap. Rate = 0.000261 (MW)^*^2(VP)0 *^4
* ,090
u
x:
a,'
(0
a,
n3
>
W
T3
;
: ^4
•
...JS.i i ;
—i—i—i—i—i—
i
FIGURE C14-B
.He
m .058
.070 .wb 2
Calculated Evap. Rate. J-fe/hr ft
Evap. Rate = 0.0299 (NSC)^•^5(VP)
1-hexanol, 1-heptanol & 2-octanol
C14
in
-------
APPENDIX. Q
fi£LAUVO&.I£_tMEQBAUO.OAIA
.01;... "CV.AP0RTATI0N BATES. QF._SQLYENT$_ANP. AN
IMPROVED_ METHOD FOR THE J OET.ERMIMIJ.QN.
B2.L._A5JM !) E SIG N A T LQN J) 3539-73-JBEAPPj?QV [D.J38 J).
:\s.ta.npard„ Tis.ijiEmQQs.. m sy.mmm
QFJffliALLU_LIQUi-QUX_£)! ELI TttUbfiJJJ
mmohLmi
»•*
-------
HI; " E.y^£QB. AIIQH
IMPROVED METHOD FQFLTHfcIE.DETERMINATION" „
6X.Xj.Dj_HILS0.tl
ii C
-------
ta^iagpoa'cjirsori fcteoaaes 01 stoaventj
sarad Am fjarapa'O^ed $&eflh®d! ifo*
¥SueIr ©efsa'a^janaSaon
frp'oducca i.om
_^P5l .IV.HIjHIe copy
b)f Lr nri D. YViliort
Wilton trvd«vlfial Hyflion# Oftd RtltOftK Lobo*otori«%
Chicogo, rilinott
Knowledge of evaporation r.iics or
tli ying times of solvents and thinners
u^cd in paint, lacquer and vatnish
manufacture is absolutely vital in
tontiolling "humidity blushing"
caused by condensation of moisture
resulting from the cooling effect ot
evaporation.
Direct and indirect methods have
boih been used in determining evap-
oration rates Indirect methods in-
clude determination of the variation
of vapor pressure with temperature,
and the reduction in temperature
with time under controlled cvajxjra-
tion. These methods are tedious and
somewhat cumbersome, and as a
. result more direct and simple meth-
od; have been devised, all falling
within two broad- classifications in-
volving either volumetric or gravi-
metric procedures. Loss in weight
or loss in volume is plotted against
tunc ami curses of absolute evapora-
tion rate dms obtained. A more
practical method is to report a rela-
tive result as i'nc ratio ol. the times
reS^'rg^ ^or evaporation of_ cither
un11 mass or unit volume of solvent
c^iTpaTetT to some standard such as
eTTicr. toluene or n-butyl acetate.
Perhaps one of the most simple
procedures for determining evapora-
tion rates is to observe the time
required lor evaporation ot' unit vol-
ume of solvent from an absorbing
surface such as filter paper. This
\metliod is currently in use by at
/least one solvent supplier, and has
oecn used previously in Germany.
The principaTThfTiciTlties associated
with tliis method have been (a) ob-
taining a readily reproducible vol-
ume, (b) maintenance of constant
conditions of temperature and hu-
midity, and (c) suitable means for
i ' " — -ii.— - - """ 11 ----
rtudf wot lupporlcd bf o gran* frem
Notion*! AncKi'otion of Muluol Co»-
X wo\ty Cpmpon»ci( Chicogo, Illinois.
viewing the disappearance of the
solvent spot on the absorbing
medium. This study is concerned
with the development of the ap-
paratus and method for obtaining
reproducible results, together with
actual data on a representative num-
ber of solvents using the filter paper
spot technique.
In correcting these difficulties the
first requirement was to select the
most, convenient, temperature and
humidity controls under which to
conduct the experiments. ThcJErd*
cyj Standard _ Stocfc Catalogue,
Section IV part 5, specifies a !tm-
pcrnture__of J3.5 rt .2?. F. and a
humidity,of. 50.% ±..A% for inspec-
tion, sampling and testing of paints,
varnishes, lacquers and related mate-
rials. Accordingly, the experiments
were conducted in an air-condi-
tioned room having these conditions
oL temperature and humidity. The
difficulty of obtaining reproducible
volumes was solved by the use of a
hypodermic syringe fB-D Yale.-Tu-'
berculimyringe, 5~j> ml.) suitably
mounted so mechanical stops (or the
plunger head would allow delivery
of 0.1 ml. The actual volume deliv-
ered would make no difference of
course, but best convenience of spot
sue and evaporation period were met
with a volume of 0.1 ml. Weight cali-
bration with water revealed the ac-
tual volume to be 0.105 nil.
oc
DC
d
c
Drawing of cquipm«nt used Jn tnHnj -
tvoporotion rctfi. All componmfi «ri
ovailobt* oi standard product! mode W
laboratory equipment monulodurtn.
Preliminary trial-, with different
makes and grades of filter paper it |
vealed lit at evaporation rates varirJ
between the various papers, prob
ably as a function of difference in-
porosity. Since it was intended ttut
a large number of solvents should
be investigated with several determi-
nations for each in the interests oi
accuracy, the cost factor became ot
significance. After suitable investi^'
tion, 11 cm. circles of E. H. Sargent
and Co. No. 500 filter paper were
TA8LE I
ABSOLUTS DRYING TIME OF REPRESENTATIVE SOLVENTS
Dry in
g Tim* in
SaconJt
Solvent
Run Not. 4 -
t inc.
EtKer
15.1
15.0
15.3
15.1
15.4
15-2
Acetone
28.6
28.4
28.8
28.4
28.0
28.-*
Beniot
42.6
41.7
43.0
41.5
42.6
42.3
Corbon Iclrochlofide ...
39.6
40.1
40.0
39.9
39.7
39.9
Methyl chloroform
40.1
40.9
40.5
40.4
40.6
40.5
Ccllovolvt .
429
428
425
430
425
427
6
PAINT OIL Cr CkEMlCAL REVIEW
Oee«™btf I, 1^55
-------
Jnil (0;t,
I hi: (lucttion ol nuini:uj
rc^oni!)ly r i ft ill CVI .UHI i'iij; turfite
wn revolve*! by cljmpin^ the circle
o( j.apci between ( die | ic i i" | > 11 ¦
cry of ill c tlic wile c i i c I f \
bem^ wcWIrd io i he anm of in
Uiihiwiy wire ten lube holder (torn
(he I u be c I :i in pi proper had
been rcniovetl. A horiioiu.il evap-
orjiin); position n at choicn n'ncc a
u ell <1 efined circle of lolveiu wh thus,
obtained and random ^rivitatioual
cffccu which would appear in ver-
tical hinging of (lie paper ituii
minimized.
I-inally, the mailer of a suitable
light soince wu i nveit igated. Kven-
tiially the Umveiwl Light Source
supplied by \V. M. Welch Co. under
their number S680 was accepted and
stj n
EXPLANATION OF ABBREVIATIONS
SUPPLIERS:
C6C—Caibirfc and Carbon
OA Olomon^ Allioli
EMS—E„ H. S«f9*nt and C«.
E K-—-C a »t r«« n Kodolc
JTB—J. T. BeWc r Ckemiccl Cc.
Mol Mo III* ckr oc *[H$,CPJ ^ *9
Ay_ C ! On 11 r 11 ( ( t K , P r O C I 5.2
Atlyt Alcohol (f.K I 8.5
A my I A^ rlOtf I J rn.Pi/» ) 11.6
t\0-Amyj Acovc 20O
G?r>tyl AJcoly>l (M,Rcog) Co 500
Rcnzyl ChlOfic5« (EK) 57.8
Dc^iyl fcxmof* (EK1 ....—....CO 150
n-Butyl A.cctOte (EMS) .IT__ 7.8
iio-Butyl Aol (Comml 19.6
ivo - Dutyl AlcoKol (Ml 16.3
ice• Dutyl Alcohol (Shell) 12.3
ler-Butyl Akotol (EMS,Tech) Solid
n-Gulyl O'jlyrole (EJC.Tech) 32.9
Bufyl Corbitol (EHS.Tcch) Hygroscopic
Bury! Cel'osolve (CGC) ca 85
n-Butyl Formate (EK) 5.2
n-Buryl Locfofe (EK1 co M5
n-Butyl P(Opi^ote (EK) 16.5
Buiylerve Glycol ( EHS ) MyQro*cop«c
Butylenc Glycol Methyl Acetote
(CCrCI ca 75
Corbitol (CMS) Mygrovcopic
Corbon Dl^lfide (M.Techl 1.6
Corbon Tconol (MCB) co 150
Cyclohexononc (EK) 22.2
Cyc'ohcxene (EK) ... )3.4
Cydof^exyl Acetofe (EK) co 70
Cydopentoror* (EK) 1 1.7
0-Cymene (EKI . 3).6
Decolin (EK.Tech) 4R.1
Dioceto-* Aicohol (CHS.Ttch) „co AO
01-iso-Butyl Keior%< (CGC) 30.0
Dibutyl Phiholote (EHS,Tech)
- Above 2^X3
0-Dichlorobeniene (Mol) iQ 7
iym*DIchlorO€thyl Ether (EK) co 200
1-2-DichtofO^thytene (Moth) 2.5
0-D«chto QO
Oimeihyl Phtholote I CHS J.L Above 20O
Dioitone (CMS)
£O'chlo/ohydrin (6KI
Ethyl Aceiote (CHS.85%) .
Ethyl Alcohol f£MS.D#noH.
Cthyl Amine ICK(33%)...
5.ft
no
2.7
. ci 7
Some at wolrr
inly
»n»y - r * - .
dflKcrwfcU cwo-pol«t.
inol
IWalKI — ca 36
Methyl n-Butyl Ko«
(Moth.Tech) ca 38
Methyl Eihyl Ketone (SONJ ) 2.7
Methyl Formate (EK
Methyl Loctote (EHS,Td Oxn>),— 4.8
iio-Propyl Alcohol e (EHS).... 19. \
Tetrochloroethylene (DA) 6.6
Tetrohy^Jrofuron (EK) 2.0
Tetrolin (EHS.Tech) co 100
Toluene (EHS.Comm) 4.5
Tributyl Phoxphote lEHS) Above 100
Trichloroethone lEK) 12.6
Tr»chIoroethyterve (EHS.Tech) 3.1
I.T Trichlorpropone (EK)«_„.-i 28.1
ricreiyl Phojphote (EHS) Abo^gt 200
i nethonol Amine (CHS,Tech) ..Hyorovcopic
Trlethyler.e Glycol lEHS,Tech)
- - - -.Hyorovcopic
Turpentine (Comm. Steom) co 375
V,M.6P Nophtho (EHS) 7.1
Xylene (EHS,Tech) 6.0
-------
Hi n'offucnl !iu,n
be*I AvnilAhii- copy
paper I'lic syringe was loaded I))'
laising (lie pli'nger ami carefully
<1 iscl>.11);n\i; r\(ru fluid lo the upper
nice 11.11¦ t> .tI slop. The evteiior of (lie
syringe was wiped with Kleenex
pi ioi lo final disibarging lo inMil e
ionstn( y of . Tabt.
lisi^ the evaporation rates of 151 ..
vents as the ratios of drying I'm
compared to ether, which was cho-.-n
since decimal or fractional value*
are thus largely eliminated. Wh' i -
possible, the manufacturer o in-
plier and the state of p'iri
rioted. Where no such not. ,s
found, the sample is to be i Jed
ai a typical laboratory chemical of
reasonable purity.
You get better paint
4 ways with
Calcium Stearaie #40
Developed especially for paints,
Witco's Calcium Stearate £40
assures you . . .
1. More constant viscosity over a longer period of time
2. No seeding difficulties
3. Excellent brushnbility
A. Better pigment suspension witliout increasing visco3ity
AjA our Technical Seruice Laboratory to h*lp you
tolvt >.5«r special paint problems.
Monvfocturvd in "Wilco'l ChicoQO, K/ooilyn and lot AngcUi plonh
rgS) WiTCO CHEMICAL COMPANY
122 East 42nd Street, New York 17, N.Y.
35 Ytan ' CW,<»«? ' • Akr.»-Onr«lo«W • Ariont. . Hourly
of Growth t»« *"«•••• • JonfrontlKO • Undo* «n
-------
P2; A.S TM_r ESIGNAII QtLD.353 &=29_J R.EA£PRQV£ QJiSJJ.
"STANDARD TEST METHODS FOR EVAPQRAI1QN
QL.YCMULE.. LIQUIDS. .EL SHELL.IHIR-tI.LM
EVAPOROMETER"
-------
Dt.tysaMow: 0 MM - Tt (Bnp»n»H 1W1>"
H"
AMtntcAM tocierv po« to-hmo amo matiaiai*
trtM
noxlMsd fr««H •«« Amml took mt AJTTM iaaAKk, Ca*yrl0tf ASTM
K m< *ad I* ** awrmw in * lm« i Ma>. mM appaar In Mm wit adMan.
Starxlartl T*«t Method* to.*
EVAPORATION PATES Of VOLATILE LIQUIDS BY SHELL
THtH-RLM EVAPOftOMETER1
** Hot* — E*un>l m+d tkm tfdt d«^ Dmakir tMI.
L Soft
1.1 Tbew eoeiho«lj cover the determiattioe
of lb* r»K of evaporation of all voUdk liquid*
of low rwx*ty Tb( m«0»O(il bavt b»«JI ap~
ptiac! io rolatiic tk)a>di *¦<& W or^kjuc Wcqnar
«Vv«au a»d ikiaocrv to pain I tad rarauis
kydrocarfcoa tkiucrt, to cicaaar'i aapbtk*,
muaenl cptriu, ud inwctictdc «pr*y b*M o«fc.
1J Tb« artJbods for th* daimiuaetioe of
rraportttoe taU uu| Uw ikia-(Ua rrapon>-
ctclcr lie
Sactfcwu
M«Okod a"-MuhI Rxordiaj ) to 10
MnWcxj I-Ammiw IlKunkai II to 17
I.) Tlx method* »f* (united oafy by tW
ntooaity of ihc volatile bqutd whack mad b«
wlTVacniiy krw to pemit lb* diapesmg of La
accurately mc&M/ad unptc (rem » lymjt.
I '• ff&cnkU DmataU
2.1 ASTM Slmmdmrdi.
Dm Tc«uforSpecific Oravityofladmacriaj
Aoautic Hydrocarbon! aod RcUlcd Ma-
Uruli
E I Spccjtcaiioa lor ASTM Tberaoaoan1
MrTHOO A-tVA.rO«ATX>N 1ATI USINC
thi TwiN-nuM tvAro«OMrrui
MANUAL UCOIUKNC
X S ¦ i iryaf M«d»d
) I An accurately mcaMiird aliquol of ike
folilik Uquid a drip<&*ed from « rjmaj* aJ a
uniform ratr crrti a period of 10 * 2 > oelo a
filter paper dial raapcodad from a calibrated
•teal apriaf. Tka material It Jiatiiti—¦< a*
rraaly m pnwftl* gaidad by a pndW Im
drew* arowtd lk« Sk* ptfm (m 4J). Air «
77 ± 0J*F (2J ± «nfoi»>ad M aacb m% la
JJ To okufa wprcd»rilgty »«fwihiwWct—dto ¦pod&trwr-
ky i1rlwfcn r — rffcfW AWM «a< H ta^ ^aa »a
*mm «f fm Hn r«k taafay hr Mai Ta*arfap.
•« Timifit » C«apa«rtaa Baa af lataac *V
>NnM X> PMai T I I i/T
-— liTin in 11 Ha.««. Yam p. IX*. j
WMyitrx nai<>^r*w»a
jt*+4Asnt »i iIm a >wii»«m«4. I
105 S
A- 1
-------
OUM
1 Ajftvhm
4 I £ThU-film 'rvftporvia**
titr* u tkowm U F^v I ft*d 2.
4.3 C*m#w- r/*y|
iXm hpf+rvtw* *m4 « MJT
co*o*«t/»t>o« llM comid W mxa^om, lfca( k, ft
tcrv-w but. Hcwtxf, M iuj b« 6*«4rtN*
io fcrx*M iLf uhrtru»««( ia I U'-oru**}
b*lky. mc
4 ) /*/rr¥«/ TbrMt:5Uof Wmiek of EJnnric
Timue (o manufacturing varUtkxu,
ibe syringe should be calibrated before uk.
4 6 frthumidifKOJio* E^+ipr+eni — A vug-
gc*icd Mrtup ts given in ft icbcmAtk dUgrA/n,
r.| v
Son J CUk of ihu dchumfciirKtlioii
• ppjiiiui rc^ui'c* llul u(ci) Mfccticci relative to the
¦ or h
Y
t
cKKhin^t/t fKfT*menc |cn u
¦ «»ilihk il m«y be utrd dircclly tn*lc»d o( kxl utd
ihui climinatr |S« d«bgmidirtc*(*o« c^uipmctii TH*
wk of nurogtn don not aIict ttx rwipcvttioA r*ic.
4 1 Htx'on.cirr or Oikrr Humijilr Stx'ing
Dr*«r. cip»blc of indicitinj low humidnia.
4 * T\rrme>neters, oI luiuhU »cour»cy tuch
aSTM Bomb C»lorimtttr Tfcrrmomtier
ShC hivmj i nnje from 19 10 )S"C. lubdivi-
jioni 0 02*t" or H>'F (66 lo 9J*F with 0.05'F
tubdivisions). ind c>,-''onning lo the rc^uirc-
menu of SpcciTiciiion E I
5. Pfffifitkm of Ett^orMdcf
VI Plicj tKt ftlvcf o*f*r duk on ihe "ire
suppuri. ihrridmg Ihc hoo* (hrough I vm*ll
1 056
Wot* Uk tJU om)«m of lb* p*p*<. A lv»di Uk* hook
lo lh« *—t >(>ria| below tit tlf 4kk Md
»lkrw Okt f*f*r lad (Im p*pcr rapport to ku>|
ltw*r«¦¦ lily. Ivlt rat**-. t.i Itir,
AppfOximAlcty 3 h *.re required for tlx humid-
ity to drop lo leu tliui 5 V
5.} Adjim (1m u/ flow lo 21 L/mla (ccaur
of bftti f\o*t oppowu comet nut o« lh« r»-
toiDctcr Kab).
4. Cod(U|
6.1 Brin| the m»pt« or • portion of it lo u>
equilibrium lempenturt of 77 ± I.O'F (25 t
O.i'C) u • cootUat-temperature bath. Deter-
mine (he tpoci/k jnvity of ihe umpfc kt tKit
leraper«tureiii*cconliftc
from FWWf iormubc Ct M<: ifmi tyr*m U
Tfft LNR. II * m la*f Mm k>«W «»4. m W^( -
•vstUhlc «• ar4n ha >»r«n«. Orknint m4 Cft.
lkrc^*S FaJbrr limc^t Ca_ 711 fei%m A•*-, KakkwalL
fa 1)11*
X!
€^>p.
f&r 77'
"SOU rC^-./ « . c.-v^a/'W
D2-2
-------
OHM
I
I
8
5
¦ia* towdi Um tip of (1m k)rpo4armk imdU ko
Lkx fUtcr «r to diapaoa* llk« Urt drop of
(OfrMl.
7.6 lmax*li*i«ly k/wac Um win
b*-»cici mj froa tlM dkk Mffxxt. CX>ulU Um
ftr« rca-diaj of tb« poutioa of lha dkk
ft! 40 i ud lia tvtty 20 l Record Um tijM
t»d Ux K*k rtadiAg o« ifc* diU itirt, A
Mjnpt* d*l» iV »«i it tbova la Ajuma A3.
Son 4 —Wkfc rtrj derm mforvti^ »oh wu, I)
k M i i [.m if| to U^« r»« rW«^i m oAea H mry 30
I. TlM opmiMiM iiiif lui imk+bit limt lummt
tAet Um Am TOO t
7.7 Stop Ux timer wbca the n(fcu&g dhk
tuLi returned 10 it* od^iiuj aaloxUd pocttioo.
Non 5- TW fj'er p«p* eir b« rwewd provided
tW Itfo bo ippncublt raudxM l* cipo-
I CW-r«U(iM
S.I C*kruL»t« the cv«por»ti-^ r*i« u fol-
low*
~ \X ( B — X)
»mi« af II ¦ ar m ilim Iwaar *aa M
«y td
N - no-load icalc reading (100 % evaporatad
reading)
12 P\o< ibe percent cvipcn-ltcd agaioat
tUpitd tirtvc in teooodi aod dnw i wooodl
curvt through ih« poiou From (he curve, 6a-
IfnniK tt (0 w«i^b( % iocretDcoU to 90 *, i»d
for 9} tod 100 % rvtpoftuos the tin* is
tccondj to tlx Dttreft vtliM u foikrwt;
a* t ><¦« i«
ioo % r.«po>
rtird ^n#(. »
Um tkAR W0
>00 i o 400
*00 »o i«ro
i *nr> iu wo
WX) its M(J0
Mo»« ihjft
Nnit 6- The evrve dfivn ifcrowfjl iW ranoui
poind should pt-u ibrou^)i trro o* (!Lc on^ta. If ii
p4v*<* i»t iK< '"^bi of tS« orijin. ihr delivery tuM v«$
Rrpjn id Hiirwt tedecawi
V tJlM. (
I
5
10
M
*0
M*r«« 2 % *
t, I |Upo ¦
tkx m« k <*»>!-» Uurl fro* ik* 90 ttifth % j '
mfor»«*d ttaw for tlM im( toK«a( *»d to* »¦ J
b*tyt *c*UM (99 * «*tar).
Nk
10.1 Oa A* fcxk ot ** Unrt>bor»c»y «tWy
of Um m«%o4 la vMck opanton la kko-
ruori« 4*t«rai»*4 (U 90 * trtfiontfaia poimt
of «U K>tv«Ua oov»rUj • broad rtafi ia «vap-
Of»xioa {W bin in Wboroofy r-iffiri—t
of ww fomd m b« 4J * nh»ln al
24 dtyiW of ftwtoa tAa dtrariW^ two
drrtrjcal vmhm. Oa tk« badta of (ki MaalM
okuiatd by tlu-M UtoniorW oa tW«* of
•ofrcata kavkf 90 % rrtfomti thM of 300
lo <00 t, lk« witJua Uboemtofj nnflHrUai of
vuUdoa «a fbvad lo ba at} * rdatht at II
doff** ot fr—domL >M»d oa tWw roWVtoalt,
foOowte( criMk ritoaid b« «m4 (or ]«4f>
iof lb* tcctfUbOkj of rraahl M tk* fi %
10.1.1 6r»>n drriea. TW ipacaaaa li
dtsuibaled m cvealy u poirtli tioaf * paa-
*P«a
tw i ¦yriiMg t^« ¦mm n mim
1 A5TM laa
1«>?
A- 3
D2-3
-------
~ DNN
ciled liM draw oa tl* fUter paper. A it at 77
1 0.5*F (11 * 0.2VC) and 0 to 5 * ralalivt
humidity u admitted lo tlx evapof»[mg ctia».
b«r tl I riM of 21 L/nun (0.75 ftVaun). A at/ip
chart recorder ¦ coaaacted lo the output of the
weight leriinjdcvvc* a&d Inriifitea th« «*">(«
ia weight of the filter paper dt*k and apccimen
as a function of time. From the plot obtUMd,
the liroe for each weight peroral evaporated
can be taken directly.
II. Apparata*
13 I £«y»rrr». Auiomjuk Tkix-FUm
Ctaf«fvmntif u shown ia Ft(, 4.
12 2 FUirr fafKt CXsi--Stit 4 4.
12.3 SyriHgr—See 4.i.
12.4 MmvHUtfWatkxi E<[vlpmrnt—Se* 4.6.
12. J Sir if Chart Recorder—Any atrip chart
recorder capable of ixcofdinj ihe outpal u^aal
(0 lo 15 raA) from tbc electronic optical wrijhi.
woiing device. The recorder should provide *
range of chart ipecdj Including V< to 2 in. (6.3
lo SO mm)/min It is also desirable for the
recorder lo accommodate 2 or more mA ranges
in order to regulate the sensitivity of measure-
ment.
I J. Prrpi/itkxi of E'a^oroaartCT
13 1 Place the filter paper disk on the wire
frame threading the hook Ihroujh a small bole
L0 (he cenier of the paper. Attach Ibe wire
frame lo ihe mppon hook in tbe evaporometer.
13 2 Cloae Ihe evaporometer and cabinet
doon and equilibrate both chamben u in iX
13 3 Adjust tbe air (low lo 21 L/min.
14. CcAMai of Saja^fe
14 1 See 6 I
15. Procedure
15 1 When all component* (including ibe
filler paper in place) arc at equilibrium. adjutt
ihe recording pen to a prominent "rero" poai-
lion near I be edge of tbc chart oa the recorder,
ihcn iurn the switch for the chart root or to Ihe
"OFF" poiilioe. This constitutes the ~*ero~
load and time position for the teat.
S'riTi 7 ~ 7~he milUuapcrr ranft and ctsart speed
should be Klecled. if poutblc. ui swefe a wiaftn 'JkM
S< dimension of tha wesgbt and lime am of tto
pl'Mted curve are spprolimalely iW u« leaglfc.
^ T Measure 0 70 mL of teal sample into
tlx kypodanafc rjTMSft (refcr to 7.J).
13 J Opaa Um ms%S M* toot oa the riffcl
haad aid* of the UwlatUa cabioat tad hurt
tha kypcxWrok aiifli tLren^h the rafcbar
porthole mil the aaadh dp alfiat inarh— th«
dkk aad W jaat over the paarllad Mm.
Hon l—C»n mm W uhldroMa bnadlag of iW
Um tracea of tolvcal witk MJitkia* M«rs of tk«
filter paper aad (7) a gradual daaiavtioa of iJm araa
of tW finer paper wet aotvaat (tkal ia, la tta fiaal
itaje of evaponuoa. drjina of Um papar nrrgraaaaa
trom tW outer •dgeto^t* ocaier of tk« 4m*).
T^aa. it ia twaaca pradka lor tW evaporaiioa cyeU
lo b< coasidered "ooa^Wu*' oin the raoordiag p«a
"r»co»tri" 99J % of the orifiaal JuplacanaL
li. Caiodarioaaa aad >afarrta|
16.1 Determine the mpontioo lir&c in se<-
oods ai 10 «rci(h( % lncrtasaiu lo 90 V and
for 95 and 100 % from the cvaporatioa curve
as folkrwv
16.1.1 Divide Ibe theomical recordlnj p«a
diajplacemeot for the total yaciawi bio 10
equal ufliu a)oa| tka wcigkt aiii of i)m rvap-
ocatioo currr. Lhea prq)act (1m «tab4*ked 10
* diviiiooj to corre*po*dia( iatcnactlaf potato
oa tKe cvaporatioa cvrvc. TW 0 * rraporatad
or full-load point at mo tnae caa be ntrniaad
either by umpoUtioa of Lhe cvaporatioa
carve back to tcro cvaporatioa time or calat-
laboa using the weight of lb* aampfc aad the
'Tlw TVia-nia E if¦ ¦ ¦ mi. ami
K ¦ al II (anas 19*. k milaMi feaa fta r»i*» La'
Cor^. 1035 C iMf lliii^ Orha. Aarar*. 0L tOM.
D2-4
-------
8
C
• 0 MM
CfciitxiUo* diii for tk« buruiMl (faa Kmmm
AJ). TW* rovtiM ctViiWini of Um 0 % rrap-
oraaad, fell-toad poiM b weo»»ii»dad M I
cktck for eorr*a Um*.
14.1 J MaJrifty tfca 4i*mmat (aloag tba tfcM
axk) from tke mm auntac Km Wjr Um tfcart
tfd (toot la mco*<1« (M*( tfca total »l if«i rt
fat for aac* dton r»u, K - x 10*
wbert:
K -
C -
jram» per K^viart centimetre per accoud
Un>«« 10",
— 0.0043ft mL/cm' obtained
4 factor
from.
0.70 nL
0 80
tpccimca iux
iacrcmeol batwaaa tbe 10
* and 90 % evaporation
poaflta (the Km tad U« 10
% ujcretaeal art durv-
girticd)
D -
U1 oa' - toul awaforatfcf Mrftoa at
90 mm toM fU»af pa-
£*>xo.»
—ra—
Q404M mL/cm*,
•pacific (rarity of dta nirwM at 77*F
Cll'C)
A - 90 % tvaporuioa tk»«, a, aad
M — 10 * evaporation tkaa, a.
TIm nakipiicMd, 10*. k ^airud la Um aqaa-
tioa to avoid darfaal fraction*.
17. huMia1
17.1 Oa
-------
-------
D2-7
-------
<$!r» o i»»
' *• on;
iwnM f •
a.nnkxes
Dt» dun him cvikfotoMtTr.K. manual hicohding
H t r> , v, (,;<«( f«.-m «u •«<« *'»ppor1
• »• f, » p.f.f >~ UmI iNOvt
m it. ;*<*« *|i.n t*» *«pjx>n (turn
Hr., >a ,»l ih* \Jtfhi v»n»t»om »ft mtnufactur*. 0
t >0 * ur>4«f tt« ptr«cr>bed («•( (MdKioni Jf tlx
fy»po«ntoo um« a ouu»4< (fet** bmiu. « tl'ijH ui*
CfCAM or (foCTCAM of ll* hit Ac* *4>Ovl4 b< WfWCflt
to ibewt th« 6curt4 mylu
ll/T
-------
O 0 UM
*1 *Awn i i YArvxurxx *att data ihitt
Alt ^ u»rif»k nif«« tuna < «t« d*t* Uk»n « t« T a Mr A J I
(iUi Ul I. 0 >¦«¦¦ tmrn Bm WLmm
M+ I
\f* >•{ Vtl.
Ifn^ w—imi. ( IfaruMk rV*«*jrf*. 0,
H+ Ljm U** Ift <*. lynaw Wm*. J • P « 6. .
7*na * U *m / - i * - lift lfr%«4 • J/C .
f t»if»*vwwi Twymtw __ .
C tbvWiu* M.j*i fimM I •>( - - • i00 i # - /1 IiUum H*i
C »K«i Ti«y«rv«n
* «< '1 i ' u-% \:\ i ft vi ,
VM i IMI N SnClMtM
VMf VAII K" All SCAll tVAfO*A-
»i tioH kLading * Lading tion
TIM( i <•> •• l\ ^i*l IIMl i (|) ||.Z) % .>( l ta< ptfn >A
k/ wa».!«»)!> til *r>f»4'|iui |in»i*mpW ifiKf
IO(4) ptf> J* f\r\ im»o >i 6 »ilu< iK«n I* «*> nuMnh
Ifef 10b 1 t|i(um(n 11 .Ad «i "hid »i*oa i.mt
U» ilkol»i< (f»* r • « 1 0 ^0 ml
»t l'«< iltfU ».*;j Jk fVo 1K4
ma U ^ i>hti >» o ¦ o '<> » 1 ?0 •
fJJS) l*tb 1(1 \ M U»
wvrtip^ 10 7.3*5 cftun divotocu.
A3 ; •/ iltf Er+porom*fft—Sifta
(Ncrr may b« vUvOtdjbll vIAAIkMU U «4ly
f»rwic. iK« EviporomWt «i' 10c pom w»ihin the idixt clumber
o( itw kpoB 4 110s ftAll S or VOLATILf. MATERIALS
Ai I The € T»M* A4 I
¦ re lypifil *«!«<* '«•' m*t«-
rtftU Ui*tn >h# »*!wrj »ho»n he ' 1 *
prvirU di»« 10 norm*) vtruiiom ia tot uxvdiiions
aftd pwnl)». CD#npo44tiOA. vault*, cU. of lc*t
02-9
-------
I
s
*
AAjtwfUNia riMldl)(Ui«M«|
A*myl msum. frtmrj,{mu+4 baMn)(V) *)
A*my< »ieo4
MO**)* KMH (*0 %\
Wo^ryl ilmfcil
boWyl i»<4wtfyT»M
hmTfi Ucvmm
C^dekuMl
DtOWIE'
DCOMBE I
DCOMLE'
DCOMME*
DiceMOM >ko4o4
cwuyn
DwuUrjl U»o—
Ow«*yl Inr»>»n1i
LAy* MUN (C *)
Ufcyl mum (~* %)
Ukji i«m< fW *)
Ukyi ¦Iwfcrffrt*)
tlliyi &9kti{\00%)
EikjH tmji Lmom
Eikfl Ltkmc
Eikyl aWt
EAfcyW** ftJycoJ
EOM »EA
EGMftE*
ECMU1
EOMEX* wi**
OotI Am-
MM - 14
12
$.7
75
11
HO
044
410
M4
1 >00
0.W
1 no
Ml
2 500
0.20
OJ
>05
0*)
00
10
I))
) 5
>05
13
xn
IJ
*40
10
00
14
470
14
*0
I.I
241
It
740
0.44
440
0.70
1 Ott
044
1 010
0.45
545
OH
4*0
00
0,02
0.01
) M0
0.12
4 520
010
Jtt
U
in
u
2 4)0
0.19
2 440
ill
2 2«0
0JI
2 400
tit
II)
4 1
no
41
in
40
10c
O
in
4.0
100
43
yy>
M
im
1.7
1 no
0J7
2 040
O.I2
542
0*4
I 000
044
1 030
044
40
11.1
001
* 7*0
047
• 100
0.04
H yoo
0 0)
1) 500
040
\ 210
0.M
1 M0
0.50
2 700
0.17
2 00
0.14
2 5)0
0.10
2 J»
0.11
1*0
05)
25 700
0.0a
0 01
1) 400
004
000
2 540
Oil
2 4*0
an
2 5«0
Oil
2 400
Olt
001
40
7.1
20 000
042
n *oo
043
5)5
Ott
955
111
tJ
50
9}
u
210
2.1
IX
X4
1 000
0.47
100
0J0
121
J.O
111
44
1 0)0
0.44
1 040
00
1 710
0J7
1 710
0J3
210
1.7
m
U
1+4
2.0
1 >10
0>4
1 230
OH
440
CM
200
14
m
14
A-
10
1864
10
D2-10
17.9
I
I
8
b
H
-------
£ o
*4.1 C.
«r< *¦*»¦ k * c~«™- "T
14
*4J I , | j
XO |) JJO l4
"¦"rt™ MJ 0 7) too o 75
) N*' *i IX )» ijj j4
* **n' *¦""« uo »i i«i j)
)J0 I J MO I t
*»*¦ »» on 4M io
*ani"' t'r"* ooi
i) u
n 4i «o so
'•j— j)s io :n io
j«3 OH
*»w*< <10 on tx o?)
TWm 4*11 •'« k«iL (^wruuj j'»^V miMt 10 * mt
'Krj-ICMUI (it)lra« -™n«»n>)l iiki () ankti) nfcuxrf) tGMEC t4k;lra> (lycst xiHiik^ 4i tCMII. ttkjU— tly**J nkn t] Wtsi; ciUmII DtGMME. Dwltibw llyml ¦nm n> i1
tiW DIC.MI f D»nk)W«a jljcoi n.«ocU)lnWi DEC UK. Dotk^KM gjy«x>< BouMfliiki
'*• ^ —-W— lm~r, /„ T,m-1 «•* XsxWi ***'"•- fSm* rrtfrnmf It* •afwfrt *( mr /»» rrj4*i auarW *
• -* ~i ifc» Mniri C»~ I •/ i**j n—i HfmUj W^WlW irwwettM ^ /W wAAr
^ », •* • «w <»- "mI Y V-f*"" Ml >•*> »¦ |A> li»tMrWf»n.no«U««« ImmWim /lw rrwi
W 4 "• ~W» —^I** IW /»»av«4i«rr m*t4f*kn/— inWa^Mu ~J— "r
'' *• •**"•'''1 ' i TM M rmif* nm fMniwi M0 mHw fmrrf*/ rem mjrrmi»« m I wni*i| ^rW
l— W 1 «¦»—¦•' -*•• r«. -Ml !/"•/>¦rflWjMiMxa, trnMirniWi/vUmi^ .<..!/
—•»" I •— ll'*( ~t II^A /»/« J4. rt4U<«V*«. /»/0J
PROBLEM HARD COPY
©
1)2-11
-------
AEROiDJULE
-------
Ml I u11U I A L
/7r
IN! M f\ I W I \ I. I I VC, | II m:
A ll( I *() I I I M ION C( )N' f KOI.
INDUSTRIAL HEALTH ENGINEERING ASSOCIATES, INC.
>.\.\n it«; ftiioui %V. i' O inn 'Mi.1 K'" nil m'Oi is '."'i i hi ¦! H ".in)
March 24, 100G
Mr. Jainos A. Gideon
Chief, tXTB, DPSE
NI OS II
Ma list op R5
•5 6 7 6 Columbia Parkway
Cincinnati, Ohio 45226
Re: Solvent Evaporation Rates
File P-03-14-86
Comment on Paper by D. C. Gray
Dear Mr. Gideon:
Following is my commentary on the paper "Solvent Evaporation Rates"
by Douglas C. Gray appearing in the AIHA Journal 35, 11, 695-710,
November, 1974. It is interesting to note that his reference No. 12,
his Master's thesis, is the basis for much of this article. It was
the existence of that Master's thesis which led some peer review
person to recommend against a research project proposal that I made
12 or 15 years ago on the sane subject; the rationale was that the
necessary work had already been done.
It -s my opinion that although this published paper is interesting,
and the Master's thesis was excellent for the purpose intended, this
information is by no means adequate to cope with the practical problem.
The following commentary is keyed to marginal numbered sections of the
paper .
1. The assumption is made here that heat transfer to the liquid
to furnish the latent heat evaporization occurs only from the
air. This obviously would not be the case for many practical
instances of liquid spills. Heat transfer from a concrete slab
to a liquid spill could easily be a significant fraction of the
heat flux represented by the data in Table IV.
2. Starting on page 696, and through page 697 and 698, Gray recites
various theories of the film thickness over the liquid. It is
not clear whether in all circumstances he is referring to a vapor
film or the classical stagnant layer or film represented by
classical fluid dynamics theory.
In the marked section on page 698, he recites that one researcher
determined that changes in airflow rate did not affect the surface
temperature (the evaporation rate?) as long as that rate exceeded
3000 cm/sec. That information is not very helpful. 3000 cm/sec
translates to 5900 fpm or 67 mph. Again, the data in Table IV
El
MAirj nrnr.F Af WASuiuITOM avPM ><¦ c.'"ii tin K'tri'ir srvM i'. >«••;>(<' .1
-------
Mc . James A. GUI con
Horcli 2.1. I'JOG
Pago 2
shown an approximate 10-fold Lncruar.o in evaporation r.;tc
with a 30-fold increase in air velocity, other parameters
remaining reasonably constant.
3. In order to perforin the computations shown here, one must
know the value of N , moles of liquid evaporated per unit time.
This seems to be n circular" solution; one must know the answer
to calculate the answer. On casual inspection, it does not appear
that a solution would converge upon iteration.
4. In Table II, the values for D range from 2.8 to 3.61. Since the
evaporation rate is postulated to be directly proportional to D,
it seems inappropriate to use the average in further computations.
It could introduce a possible 65% error in evaporation rate for
the liquids shown.
5. Figure 1 through 6 are disappointing ir. that they portray a
correlation, not of computed versus measured evaporation rates,
but oi: subsidiary parameters such as film thickness, Schmidt
number, and Reynolds number. Further, visual observation of the
graphs and the closeness of fit or lack thereof may easily be
obscured by the logarithmic nature of the plots.
6. The ratios of expected to observed evaporation rates portrayed
here is not, in my opinion, adequately accurate.
7. The experimental apparatus used by Gray was admittedly not
very sophisticated. This was no doubt caused by inadequate
funding and a basic theoretical rather than empirical approach
of his work.
The 19 references given by Gray were not studied in making this
commentary. They will of course be studied before any proposal is
made, to glean whatever guidance may be available for experimental
design.
Best regards.
Very truly yours,
KJC/a jm
cc: Mr. William Burch
E2
13?
-------
Aiun J" ,"oo !icd to prediction* (or fpt)U both Indoors and outdoors.
I nfroducfioa
o OME GENERAL OBSERVATIONS re-
•J Iniino (0 1 he rale at which liquids lend
(o evaporate were lined by Mcllan:'
(1) Evaporation rates arc not inversely pro-
portional to the boiling points, but liq-
uids within a single homologous series
of compounds do evaporate more rap-
idly if their boiling points arc lower.
(2) Liquid:, from separate homologous sc-
ries with equal boiling points have en-
tirely different evaporation rates.
(3) Hydroxy! groups greatly retard the
evaporation rate, so that compounds
such as alcohols and water evaporate
much slower than one would otherwise
expect.
(4) I( two compounds have identical boiling
points, in general the one with higher
molecular weight will tend to evaporate
more rapidly.
(i) Vaporiiaiion results in a temperature
drop in the liquid, unless heal is sup-
plied from the surroundings.
Evnporaiton studies found in the litcra-
•Preterit addrm: Health Di*iuon. Lot Alamo* Scien*
lific Laboratory, »K California, Lot Alamot,
NM 1^544.
Mom of ihc material vwd in ihii inkle wat contained
in 1 t*»eiJa entitled 'Evaporation Ratet of Simple Uqwidi"
wutnruncd to the Ur»»»tr>5iy of Cincinnati In partial fvl*
filWneni of iKc requirement* for tbe decree of Mauer
of S
-------
Norcnitjcf,
IV1 r ck,mi|ik , 11U.1% transfer liming i:\MpoM-
imiu ss ¦¦•'.¦if by iht .wn^nn" l%c .11
iKiOWy M v.-ipnnVc l''C li([i'»tl. A« Ix'wii1
pun is iiai, '.li: |in>L.*Ct^ irf 11 v:¦ |>i)i sliiiu it
li ii'icJ ulii n itcly Uy tin: constraint that
irtiinict u! JkmI I/ohi i)ic :i'h must
in cncigy, the transfer of i«nH lioni iltc
liquid. Hither npprotich must £ivc Uvc same
.inswiv since heal and mass ir.nisfcr arc in-
separable.
A *ecy useful formula, which GUiiUrtd'
siaics w.is originally derived front lite the ry
of molecular diffusion by M'.uwcll and talcr
modi lied by Suihcttatui, is presented and
discussed at length by Drown el ai} and by
Cif'ilnrsd and Siiecwood.11 This formula is
only iiiicily applicable to prediction ot
cvzporn'ton rales irt duels, but may be mod-
ified lo malic it suitable for rales ir» the
open air with additir V. assumptions:
E D PM Pd
S
where
(1)
R T F Pm
D = the diffusion coefficient, in cm^/scc
E = the liquid evaporation rate, in
grams/sec
F = (he air-vjpor film thickness, in cm
M a the molecular weight of the liquid,
in grams/mole
P « ihc tofaJ atmospheric prcssur*:, in
mm Hj
Pet « ihc driving force pressure, in mm
»e
Cm = the mean partial pressure of gas
above (he liquid, in mm Hfj
/? = (he 535 constant ** 6.235S (If)
cm' (ram Hg)/(motc ®K)
S « (he aica of the evaporating surface,
in cm1
T = the gas temperature, in #K
The Wa*w«ll equation is only strictly true
tinder steady state condition}, over a liquid
surface of negligible downwind length. In
addition, some gas, such as air, must move
across this surface to carry away the evolved
vapors and prevent their accumulation
above the surface. Any teal surface, of
course, docs hive Ic-ngib, and aecofiju'atior
of v.ipnr occur. Tlilv .lccuniiibtum
wtm l>c allivwsj Su? Ijjr smug the log:irit'iiHi.'
Urer.i£C« of l'*t and !'<« nbtatucd 'Jt tlt\: up-
wind and tluwnwjnij cd^;^ stt the .turf.icc.
These currucituits arc not important for
sliorf surfaces, but become vital as tlic sur-
face length grows lout; in relation to the
wind 5peed over the surface. _____
Derivation of both /V and Cm depends
upon an accurate estimate of the vapor con-
centration in live ail above il:c surface. The
assumption of complete mining of vapor
and gas is usually made for ducts, bui in
the open air sonic method of estimating mi
"equivalent duel cross sectional area" will
be required. The assumptions made in die
derivation of Equation 1 were sum mariict!
by Brutsaert:'
()) The vapor concentration Ln ijie air
above ihc liquid is constant with eleva-
tion. This assumption is obviously ncn
erjti/cJy true, bu! it enormously simpli-
fies calculations. One would expect va-
por concentration to decrease wiib
height abcre ihc JiquJd jurfjee, bui with
low concentrations in moving air the
error due to this very convenient as-
sumption is quite smili. as shown by
studies in ducts.
{2) A sta^nam air-vapor fiim exists between
the liquid surface and (he moving air
stream. This assumption is well sup-
ported by numerous studies. Thomas,*
in his studies of water losses from cool-
ing ponds, found (hat radiation, cfi'cctx
under conditions of natural convection
ace not very apparent due to iht forma-
tion of & vapor film, which he compared
lo a foj, above Ihc liquid surface. Sut-
ton* notes that the air {low over small
liquid surfaces near the. ground is af-
fected by a laminar flow region formed
by ihc liquid's surface. He further
states, however, that extension of this
concept to larger surfaces, such as
oceans or large lakes, may be unsound.
The film coiwept is also supported by
iSe findings of tot tr-asis'tr undies by
£4
-------
fih/iMi'Hil tt\£ifn /4 /out tut/
Scfjjcycv. lie found that a *t
/.one of hot air surrounds :i lie;itec) body
suspended iii an ;iir stream in duel tests
of internally l>e:ilcd :iirftiits. In the c:i
-------
Nnvrniliff, !<)}{
t,( I null' iliiiUu'is, .hiiI vciy lliin--in(i.ill)'
v,tily .v lew molecule*, ;il nio%t.
When .1 liquid surt.icc >s mere used in
.-we a. tUc new Mirf.icc l.iycr must li.'tve lite
same density of molecules pcr unit arc.i :i\
[he old one. 'I'd form this larger surl.'tcc.
llicn, molecule* uuim. be ili.iwu Okit ot (Ik
liquid toward the Surface, ,i^:iinsl I lie unbal-
anced inward force. Work done .tg.iinst tJiis
(orcc per unit .vert of new surface formed is
defined to be the surface tension of tlie
liquid.
Wyllic10 has put forward the concept of
an additional, very thin, perhaps itionomo-
Iccular, surface layer covering evaporating
liquids. After extensive lesis of the evap-
oration of liquids in a vacuum, lie reached
the conclusion that the known behavior of
these liquids could be explained by a three
layer concept. The first, or bottom layer in
this concept, is the liquid surface layer, and
is an ordered layer with a definite structure
in the cast of polar liquids like water. The
second layer consists o( mobile liquid mole-
cules attracted strongly to, but not bound lo,
the liquid surface. The Ihird layer consists
of vapot, continuous lo lite walls. It is lite
second layer in this concept lhat exhibits
the lowest liquid temperature, acting as a
type of lid that limits the vapor pressure
exerted by tJie liquid. In a vacuum, the
third l.iycr is the last, but in a moving air
stream this layer is reduced to a thin film
overlain by a fourth layer, the moving air-
vapor mixture. Most reactions, including
changes of slate, proceed by a mechanism
where a single electron is transferred to a
new energy level at any one time. These
transfers occur most readily when all of the
reacting species arc absorbed on a surface,
where the molecular geometry and atomic
charge distribution expedite (he redistribu-
tion of the electrons. Molecules may enter
or leave all (our laycis. one layer at a time,
only by the process of molecular diffusion.
This accounts for die extreme stability of
the evaporation rate under equilibrium con-
ditions.
E6
Tin* Jcntjjcr.11c known in order to esti-
mate !'i! in equation I. In most studies, the
lcm|K'i.ttutc of ilus layer lui* been taken lo
l>c equal to lhat ol the bulk liquid, resulting
tn ;iu estimate for I'tl that is slightly too
tuv.U l\n estimation of evaporation rates in
the field, this procedure is probably best.
The errors to be expected arc only slightly
larger than the errors in an estimate arrived
at by a great deal more effort.
Scrgcycv conducted numerous tests where
the aciujl surface temperature was meas-
ured with thermocouples. He found that
changes in the air How rate over the surface
were not effective in changing the surface
temperature as long as that rale exceeded
3000 cm/scc. At lower air flow rates, the
surface temperature increased. Under nor-
mal conditions, then, the surface tempera-
ture will lie somewhere between a minimum
value and the bulk liquid value, depending
upon the air flow rate. Lewis demonstrated
the mechanism responsible for this effect.
He showed lhat the reading of a wet bulb
thermometer is not influenced by the air
flow rate as long as that flow rate remains
high enough so lhat the heat loss from the
bulb due to radiation is small compared lo
(he heat absorbed by conduction from the
air. This just happens to occur in air at a
velocity of about 3000 cm/see, but would
be somewhat different for a gas with a
different coefficient of thermal conductivity.
Matsafc" derived an equation for finding
the minimum value of the surface tempera-
ture from his studies with an aspiration psy-
chrometer. Those interested are refected to
that document or to the original thesis from
which this report is derived11 foe the for-
mula and suggested methods to make use
of it. In humid air, as Matsak points out,
the surface temperature can not be far
below the dew point. Moisture condensing
on the cool liquid surface releases its heat
of condensation, which impedes any fur-
ther temperature reduction. This moisture 4
effect may alter the evaporation rate either
ill
-------
I, '
6V9
vcy 11111c ur very ideally, depeiuliiij; upm
the 111|iikI iKine cv;ip»w.iie«l ami .my icae-
1 itt(in takuM', place.
IMIiiiiiImpii tit (lie I>rivirij; 1'itrcc
After the su r f;icc ic m pc r ;i t u r c has I wen
csi
-------
Nmrnihri. I9H
1.4 7 (•! I't 7V),/!.
C " "t'v)1
where
Tc critic.il lompor;tllire nf the lii|util, in
"K
•/-- cniic.'il (oin|k:i.iitiro of the y:it, in "K
7V«KI"K for »if
The molecular volume of the vapor of a
pure substance may he calculated by the
summation of (lie atomic volumes of llic
component atoms, i( the molecular structure
is known. Atomic volumes arc given for
several atomic species in Tabic 1. This
summation can become tedious if the mo-
lecular structure is complex.
When the molecular structure is un-
known, the molecular volume may be found,
in another tedious calculation, from the Van
der Waal's equation:
(P + afVs1) (Vs - b) = R Ts I0J
where (12)
a and b arc constants, listed for various
liquids in reference texts.
To avoid the disadvantages of Equation
11 for 0, a correlation analysis was made
of an entire series of physical constants
with the experimental values of D given for
a variety of materials by Gilliland, Arnold,
and the Handbook of Chemistry and Phys-
ics." Only evaporation into air was con-
sidered. The best fit, at standard conditions
of / * 2'W"K .mil /' - 760 mm lip, was
uht.iiiK'il fur the croup of physical con-
stiinu: ,U- "'-.Vh.mil The for-
mill.i fui I) w.is therefore j kin ml .iteil to lie of
llic fnllmviMjj form:
I) =-
760 (M Ib/G)"«
(13)
K
f (sYru/c.y'1''
where
C ™ liquid surface tension at 2CC, in
ilyncs/cm
Tb ~ liquid boiling point 760 mm Hg,
in "K
Therefore:
D (M Th/G)'r- (5170/760)
= 6.80 K
See Table II, where it can be seen that:
D (M Tb/G)u: = 2.89 average for
26 values, and
K = 2.29/6.80 = 0.425
Creater accuracy could probably have been
obtained by using the surface tension value
at the actual liquid surface temperature, if
known. However, surface tension docs not
vary in a uniform manner, as docs viscosity
and vapor pressure, so that interpolation
over small temperature ranges is not feasi-
ble. Unfortunately, most standard reference
texts only give the surface tension at a stand-
ard set of temperatures, usually 0 and 20°C.
(9
TABLE I
Vapor Atomic Volumes in Liters/Mole
(D.n.i from Gilliland,' Drown ei nl.i and Arnold'J)
Jlydrojcn
J.7
0«yjen:
Carbon
14.8
O "as in aldehydes, ketones
7.4
Chlorine
24.6
—O—as in methyl esters
9.1
Sulfur
25.6
—O—as in methyl ethers
9.9
Bromine
27.0
—O—as !n higher cslcrs. ethers
11.0
Iodine
J 7.0
HO—as in acids and alcohols
12.0
Nitrogen:
Double bonded
15.6
Primary amines
I0.S
Secondary amines
12.0
Additional 1
nforntalion:
Deduct IJ for bentcne ring
Deduct 30 for naphthaline ring
H. {at equals I4.)
Air equals 29.9
E8
•JT
-------
hull"III,it IIhui-ik- ia11'm /iiiiunif
101
fvlnriuali/.cil cx|icririicnl.it I) v;ilucs re-
tried in various references ami those tie -
,;veil from liquations 11 ami 1.1 above arc
i'oiii|iaree those for ihe
cv»|«\c.>iiort of a liquid into air ,U an air
temperature of 299°K and an air pressure
of 760 mm f{g. They arc computed as fol-
lows:
On - D (/77f>0) (299/7 )5/5
-- 6.SO I) P/nn . where (14)
Da = normalized D value, in cm3/scc
As can be seen. in. Tabic Iff, experimental
values of Dn for water varied 17% among
five investigators, valjcs for chlorobenzcne
varied 19% among .Virec investigators, and
those for the remaining 11 multiple citations
v.nu'il 111% or less amnil); |wo investigators.
I?w ihe thicc calculations of /), die one best
fitting the revolted value m values is under-
lined in Tabic III It can Iseen that;
(t) Arnold's estimate produces the best
(it lor K/23 solvents
(2) liquation 13 anj Arnold's estimate
produce equally good fits for 6/23
solvents.
(3) Equation 13 produces the best fit
lor 6/23 solvents.
(4) Gilliland's estimate produces the best
fit for 3/23 solvents.
In fairness to Gilliland, it must be point-
ed out that a better fit to the data would be
obtained for a value of K larger than he
recommended. Gilliland's formula only
TAIJLE II
Physical Constants ami Diffusion Ci
-------
702
Nn i rinl.ft, !•)) 4
ovcrcsiunalcs 1/23 liquids, while Arnold'*
(wnuil.i, ;tnd »if cotmo liquation 11, tend
(d nvcr. .ittd uiulor-cslim.ilc an cqtirtl por-
lion of ihc v:ilucs.
The icclmii|uc »- -lit tUnvi from small
tli.nncler. k'iiipcr:tiurc controlled lubes ol
¦fMit.l- III
Niwin.ili/cU I*1 a.pcrinicot^l V;iluci ol Diffmntn Coefficient
for Various (.iqutds
O V;
-------
Jtifi'fte A ioutHfit
70 A
Iu|ukI. Tin's method is ilcseribed :it length
by Kmtptoit :uul Wall.'"' According Ui Giili-
land. i!ii> is tlic must accurate and reliable
ineifioif uf obtaining /). I.esS accurate luetti-
mis arc used for solids ami jjascs. Arnold
asscris tliat I) values for liquids (C[KMtcd in
Uic tiicr.iuiro may bo in error by as murh
¦is 20%. Hquaiion 13 give* cslimntcs with-
in these limits of observed cxpcrirncrn.il
error, and will be used to compute P in
all following calculations.
I'-stinrjiion of (he Film Thickness
theories. attempting to derive a formula
for /¦' usually assume the model of a gas
under pressure diffusing through a perforat-
ed phic of a specified porosity. The gas
pressure in this model is Pd, the driving
force. The plate thickness is F, the film
thickness. This film is usually conceived of
as a stagnant layer which the evolving va-
pors must diffuse through in order to reach
the moving air stream and be swept away.
Brown ei ai, Gilliland and Sherwood,
and O'Brien and Stutzmao14 have presented
data and theories for ducts showing that the
ratio of duct diameter to film thickness is a
function of the Schmidt and Reynold num-
bers. O'Brien and Stuizman also correlated
their work with' much o£ the appropriate
literature reported prior to their work. They
nrrivcd at the following-relationship:
4 = 0.20 Sc~9 i Ke0" Sl"
F
, where (15)
-------
?0<
Hovtnthrr, 1974
of litis roinincrci.ll gr.iile solvent w.h niOU
iypic.il o( -in uuluilri.il situation, and Ilinl
:i crucial test of the pri'iliclion equation.
Data was Collectcil as iloi.ribei.1 in the
Appendix Results of (he 20 tests with
xylene ate flu>wn in Table IV.
Data on nitrogen tctroxidc (NjO.) evap-
oration in Ihc open air given by McNcrncy,
O
y
c
> .2
ij ®
a S-
<-ii
X
r.
O
c
n
0
e
%
5
*S^
f.
£
UJ
•t t ft — "t « ^ O ^ r« o
o <0 't o
O — — rt r* rN rt c* r* r* *•* '"*} r"J •* f *» f *» *o
OOOOOOOOOOOOOOOOOOOO
N
Ej
Vp
< I e
> V
£
a *
li'
f?
E
E a
W M
.t: o
<
~ e
5 ~ *
c-cf
oooooooo*->o
•SO O O
¦ O « N '
fk.
...1
o\ *-» ^ r- so _ - .
- n "O *o o V O n'X i
<— r^c^Twvc-r-OOr*
"T " " O v) 4
OOOOOOOOOO
00*0000 o o' o d
%•
u
c
\j
0
0 0 0 0 0*0
*000*^^*0
— rx r< C* «
*-> —t -- -- m cv — r* r— —
o vo «o r*>* no «o o K r<>' r«* *o *c «o r^' ft r>' «o «o
ft ft ft ft ft Ov ?*\ftOtftftftftftftftOtftft
MNMNN(SNnMr(N(M(Nfi^(NnNr(r4
, ^ ^ *, ^ o « - ^ rs so
ft ft ft ft ft osftrvfto^osftftftftovftftftft
§881
8
*0 *0 ft Ok I
o5o 5 o 5o5 6 o o o o o o o o
00000000000000000000
ft « - O N t
»fl n v>
r* r» n» r»
-------
Amtritfi'i ltuintlt>< 1/ IfyCH'tte A t tttfiitttnn /ouenttl
I lenilerSon .illil Towsoil" W.1S •*< I SO CX.1II1-
11n.il .11 tins time. Only slton ictm (S min-
uie>) t>c.iV; evaporation t.itos front each
e*.|>c* 1111c 11 ( were eonsideieil. Oilier informa-
1 iitit listcil by ihese authors was disregarded.
Nn special allowances were made for those
runs where it mined or where (he N1O1
svnf;tcc (to/.o clvirinii lite evaporation period.
The average wind speed over the five-
minute period w;is used in calculations.
The equivalent duct cross scciional area,
into which complete mixing was assumed to
occur, was calculated by liquation 2. Wind
speed was given at the poinl of measure-
ment, assumed to be a standard 2000 cm.
The wind speed desired was (hat a( the (op
of the mixing layer, so that wind speed was
scaled as recommended by Sutton:
U = U. (L0"V£)om' . where (17)
U. = initial wind speed given, in cm/scc
Z = height of weather mast, assumed to
be 2000 cm
Three estimates of the term (1 /F) can now
be made:
Figure I. Did o( O'Brien and Stutcman'C (or
1-propanol, actonc, toluene, benzene and water
fitted 10 Cqu.nlion 20a.
70J
(1) Ity rearrangement of Equation I,
front experiments:
I//•• « H T i: I',,,J (/) /' M S I'd) (I 8)
(2) The empirical estimate of O'llricn
and SluUman:
!//• = (0.2/r/) Sc"11 He0" "" (19)
(3) The empirical estimate of liquation
16: where the terms arc rc-arranged and
the constants found by regresisoa analysts.
In order to plot the data, O'Brien and
Stutsman's Equation 19 was plotted on the
log-log scales as follows:
logu(l/F) « log.. (0.2/rf) - 1.2 log,0
(Sc) + 0.63 log,. (Re)
(20)
logt. (1 /F) + 1.2 log,, (Sc) = log,. (0.2
/d) + 0.63 ic05 log,, (Re)
(20a)
Similarly, Equation 16 was plotted as fol-
lows:
log,« (I/F) = logio (Di) — 12 logio (Sc
+ Bi Sc°-J logic (Sc U)
(21)
Figure 2. Xylene data (rom Table IV fitted to
Equation 20a,
E13
-------
loo (1/1") I.2l0gi« (He) "» lO(*¦• (Hi)
I), .Scal loijn (Sc U)
(21a)
The al>o*e ilc(ivc«J ci|ii:i|ioid; 20.i ami 2 In
arc actually those "f straight linos, Willi
Y - a I- b X. where:
Y = logn (I/'") * 1.2 logi« (Sc) for
both tiocs
X = .VcJI) h)g,o {He) for Equation 20a.
X — i'cA 1 lo£i« (Sc U) (or liquation 21a.
Actual values of a and b (or these values
of X ant\ each o( the plots for
the ihice groups.
J'hc u'com) group (Figures 2 and 5) is
iI.k.i on xylene from iaWe IV 'tlicse arc
< i r 11; 111 .i 1 experiments with a commercial
grade solvent.
I he third group is data on NiO« (Figures
3 and C) open air evaporation. This group
is die least piecise by far, and is probably
even more typical o( an industrial accident.
Figures 1-3 contain lines filled, to Equa-
i'um 20a, from O'Brien and Sluuman. Fig-
ures 4-6 contain lines fitted to Equation 22.
This modification of Equation 21a was
necessary in order to adequatety (it the dau
from the first group on water. Without this
change, the water data would plot consid-
erably below the best fit line to the remain-
ing four solvcnls.
logto {iff) + 0.9 iogio (5c) = a + b
Jc05 log,, (Sc U)
\ /
And, replacing Equation 16;
l/F = B\Sc~*f (Sc U) (23)
The fit to the five solvent data rccom-
£14
ttf
-------
/tnirtiom tmliiihiiil //ii-idii' A i lufiulum /aumtll
707
mended by O'llricn and Stui/.iuati gave
parameters as follows: (See Figures 1-3)
n » 1.960-1 with $ t .i nd ,i rit error of
0.0291
I) a 0.64817 wiih standard error of
0.00661 (versus 0.63 according to
O'llricn and StuU.man)
/;, sa |0« = 0.C 10954 with standard er-
ror of about 6.9%
Let Hi •* k/d , where d = 15.24 cm
then:
k =0.167 with standard error o( about
7% (versus 0.20 according to
O'Brien and Stutzman)
Standard error of ihc line was 0.037365.
The (it to the live solvent data using
Equation 23 gave the following parameters:
(Sec Figures 4-6)
a = —0.66322 with standard error of
0.01755
6 = 0.62502 with standard error; of
0.00721
Bi = 10* = 0.21716 with standard er-
ror o£ about 4.1%
Standard error of the line wa* O.0373I1.
As can be seen, the two estimates of
Figure J. Xylene data from Table IV fitted to
Equation 22.
(I//'*) I'.ivv very iimilar precision with I he
five solvent*. They ;ilv> provide very nearly
(lie same necui.iey in ilie prediction of the
data on cxylcnc (see figures 2 and 5), but
liquation 23 gives a much better fit to the
NjO( data (see Figures 3 and 6).
When the ratio o( (:upccicil/Qbscrvcd
evaporation rates is examined for Equation
23, the following results arc obtained: /
xylene, average of 23
runs: E/O •=> 1.37
Range = 0.73 to 1.84
N;Oi, average of 13
runs. E/0 = 1.04
Range = 0.38 to 2.2L
These ratios arc thought to be typical of
those to be expected in most industrial'hy-
giene applications, particularly the data for
NtO« evaporation rates out of doors. V/hca
predicting the maximum vapor concentra-
tion downwind, it would probably be pru-
dent to multiply the expected evaporation
rate by a factor of 3 for safety. If only the
average downwind concentration is re-
quired, no correction would be necessary.
Figure 6. Data oa open air evaporation of oi
trofen tetroxSde fitted to Equation 21.
£15
Ht
-------
50*
These estimates arc well within the limit*
of accuracy customarily experienced when
uiing the diffusion equations 10. predict the
(Jis|>c(s.il of pollutants whose source strength
is known.
Summary and Conclusions
Three formulas have been presented from
which evaporation rales of simple liquids
may be estimated in duels:
e
S
PPM I'd
KTi- I'm
0.425 7™
p,. W>'"' {n)
U~t>(MTb/Gy" v '
\/F - 0.217 Sc-°> (5c U) «*» *«*• (23)
To predict evaporation rates in the open
air, two additional formulas were offered:
A = W Lyn (2)
U=*U. (L°«VZ)4'<» (17)
Orly a limited number of solvents were
examined over a limited range of temper-
ature, pressure, and air velocities. More
solvents must be investigated, over longer
ranges, before the full utility of these for-
mulas can be determined. Tentative re-
sults, based upon the data from O'Brien
and Stutzman, arc very encouraging. Their
data correlated well with most similar data
in the literature. Tbcir experiments were
performed with putt solvents, fa'triy dry air,
and carefully calibrated equipment. Test
data derived for this work utilized commer-
cial grade solvents, ambient air of high hu-
midity, and manufacturer's calibrations ot
most equipment. These test data are known
to be of lower quality than the data from
O'Brien and Stutzman, so they were not,
used directly. They were used, however, to
indicate the reliability of the .ecommended
formulas to be expected in typical industrial
hygiene applications.
The value of these formulas lies in their
relative simplicity, which allows solution by
slide pile in the field. AU data required may
be readily found in standard reference texts,
such as the Handbook of Chemistry and
Physics. The estimates obtained provide
acceptable accuracy for average concert tra-
Nnvtmhtt, I >7*
lions, but should be multiplied by a (actor
of 3 for safety in prediction of ceiling con-
centrations.
Appendix ^
tCspCrhiiCiilfll /f i>i>iirt/lns (inrl Techiiit/uv
A. Equipment List
Anfimothcrm :iir meter, Ancmosiut Prod-
ucts Division, Dynamics Corporation of
America, Scranton, Pennsylvania.
Burette, 50 ml TD, Pyrcx 2130. I r.il
ruled.
Blast gate, 6x6 inch sheet of V* inch
Lucile, taped to lite end of the sampling
duct.
Duct, sampling, 6x6 inch I.D., ol 'A inch
Lucile, with 6x6x'/l inch sampling well.
Overall length of 100 inches; 60 inches
from inlet to well, 34 inches from well to
blast gate.
Duct, existing, 12x12 inch I.D., with
AcoustaFoil V* horsepower, 18-inch diam-
eter blower, by Hew YorV Blower Com-
pany, LaPorte, Indiana.
Incline manometer, Model 400, F. \V.
Dwyer Manufacturing Company. Michigan
City, Indiana, using 0.826 specific cavity
red gage oil.
Pipette, I ml TD, Pyccx 70-\\ ^ 1 ml
ruled, used as an incline to
maintain liquid level in the sampl. veH.
Psychrometer, blower aspiraicJ, r 2
Exax nitrogen filled, 12 inch s*ale, -26
to 124°F range, l*F ruled, thermometers.
Shroud, hood entry, local manufacture,
sheet metal and welded construction, 6x6
inch small end O.D.
Thermometer, solvent temperature, by
Fisher, 57 mm scale, 4 to 40*C range,
0.S°C ruled, 6 mm diameter.
0. Experimental Technique
All solvents used were left in the room
over night for temperature Adjustment prior
to use. Room temperatures varied from 21
to 30#C on different days during the tests.
The sampling duct (see Figure 7) was
caefully leveled along its length and across
E16
It?
-------
hi ,1n I U ml If yi'iftir . ( I irtKUlhlln /otintttl
iy>
u\ width |iinii to UsC. l.cvct W.IS Checked
il.uly pr.or 10 mc, and adjuued .it rctpiircil.
Tin; |npciic used ,iv .1 fluid level rclcreiu'c
m .n permanently to the side of the
vimphng ilcllc division, change in
ilic pipette reading. Uurciic volume changes
of I ml were detectable in this way,
Individual samples were run as follows:
1. White ;iir moved down the. duct at the
selected flow rale, solvcnl was drained from
ilic burette into (he sample well until the
liquid level was even with the top of the
well, and (lit meniscus in pipette rested di-
rectly on a division mark.
2. The burette was topped off to the
zero level.
3. The room barometer, incline manom-
cicr, well thermometer, psychrometer, and
Anemotherm were read and their readings
recorded. At regular intervals, these read-
ings were taken and recorded again. Aver-
ages of the interval readings were used as
input to a FORTRAN computer program,
with all necessary unit conversions, etc.,
being made within the program.
4. An attempt was made to adjust the
burettte stopcock so that the solvent Cow
w,is continuous at just the rate requited to
overcome cv.i|>oration tosses, hut was never
very successful, Some adjustment w;is made
in the sample flow rate approximately every
ten minutes.
5. After three hours, or 50 nil of solvent,
whichever came firs!, final readings were
taken and recorded, air flow rate was ad-
justed to the next higher rate, and the above
procedure repeated,
6. None of the devices used Cor meas-
urements made in these experiment were
calibrated. Factory calibrations were used
for all.
References
1, Mellan, 1.: Industrial Sotrtnu, p. 42-63, Rein-
hold Publiihinc Corp., New York, N.Y.,
2. Scrgcycv, C. T.: Heit and Miss Trmsfcr
Durinc Evaporation o{ A Liquid In a Forced
Gal Flow. InOtcncrno Filichcjkiy Zhu/nai
V.-77 (1?6I). (Translated by Air Force Systems
Command, Wrijht-Pallcrvoa AF8, Ohio,
FTD-TT-62-327/1 +2 + 4, 8 June 1962.)
J. Lewii, W. K.: The Evaporation of A Liquid
Into A Cm. Mtch. Enf. 44:44J (1922).
4. Cillitind. E. ft: DUtuiioo CoefGcient In Gas-
eous SyJttmj. Ini. Eng. Chem. 16:b%\ (1934).
5. Drown. G. G., el at." Unit Operations, p.
514-7. John Wiley & Sooj, New York. N.Y.,
11950).
6. Gilliland. E. R.. and T. K, Sherwood: Diffu-
L.
J.
3.
t.
7.
1.
*ir fie*
Key to Hwb«f«d Ite«ii
Anc«oth-er» i
tureit*. V0
BI * » t tC
«. Duct, lnch«»
S. Duet. 1JxlJ Incht*
]n;llnt maomt
-------
710
Ntn t/til/er, I 4
lion i>f V.ijmii Inm Air S'rc.vm /"¦'
ilfi (I v \ J |
7 W • A fur li,v j^v\r aivOw i\
A M «»'•> / J'yoc^H Info A Tiff-
fmlrnl Ahninphcu* i ('ifophys A'rf '70 50I?
(lUM.
ti fho*niv H. I. Mow U) OKuhic IIc-jI. \Va*
1 tt Cht>>* '•'>£. 12*?t Augut( i (i960),
9. Sufftin, O C.: AftCroniercrofofy. pp. )3),
Uitl. New Vr>ft, N.V.. (19S1J-
10. Wyllie, G.: ilvnpor.nion .^nd Surface Stive
lure of t'roc. Roy a) Soc. /4/P7:)IJ
(19*9)
/!, Macsakt V. G : Vapor Crcuurc aoij Mvapora*
lion, oTVSi^'.M:niCfs in Movable Air. Gifiyftia
I Sit'iitiifiyo (nn volume Of Jaic £lwcr») pp.
35-41, {Translated by Air Porct Systems
ComctiAnd, Wr if h( ¦ l';i[| t r jon Al'fl, 0h'O%
mj.rr-1044/1 + 3, H November 194).)
12. Ofjf, D. C ' fci'ttpototioH Rates of ShttpJe
tJiquuh, pp. 6^. Univcriily of Cincinnati MS
Thiols, Cincinnati, Ohio, 1970.
13. Aronlrf, J. U.: Studies in Diffusion. t**d, £.n£,
Chrm/22:1091 MQt'W
|4 • ft,in,11^ >,t I ,tj Ct'fiil'ry t\ n,1
f'ttysm, JOlU rMiliim, pp. JlM, OtemicM
KuhW-f f'uhti^hmc Co., Ocv«tifui. Ofrtift
(1V5JJ-
IJ KiwpMri, 0, H.v and (r, T. WaU: Determine*
Jn>n of Oiffuvion OvlTicicni j from Kaici *>f
/. f/iy f. 56,'7', 5
1f». O'fificn, I.. »ruJ U, I*, Siui/.i^An; M.t
Tranifcr oi Vutc l.iqutds l?rf>m A Wane.
l:rcc Surface. Ind. 0»e»i. ^2*11 ft 1 (1950).
17. Pasquilt, P.: Evaporation prorn A i'Janc.
Free Liquid Surface 1nlo A Turbulent Air
Stream. Proc, Royal Sue. s
-------
APPENQ-DLF.
£XEBei_F.BDHi_ GUI DILI WS_IOR JJS£jQ£
VA£.Qfl_CLm DJSttRS IQJLUQDf IS
fiY_JL_R. HANNA AND P. J. DRIVAS
-------
GUIDELINES FOR USE OF
Vapor Cloud Dispersion Models
By
STEVEN R. HANNA
and
PETER J. DRIVAS
APPENDIX F
For
CENTER FOR CHEMICAL PROCESS SAFETY
of the
American Institute of Chemical Engineers
345 East 47th Street, New York, NY 10017 1^9 7
F1
IS1
-------
&u o^r-aessa^
4, Source Knu««i<»n ¦ Iii. ln /
-------
ATMOSPHERIC
CONVECTION
SOLAR/
ATMOSPHERIC
RADIATION
EMITTED
RADIATION
MEAT LOSS
DUE TO
r.VAPORATION
Figure 4-3. The Heat Budget of an Evaporating Pool, from Shaw and
Briscoe (1978) .
-------
/\.X CJ\ IL <£uim&s l/a poj^ fet s.^f-'^._,,
-------
GuMollnrt for Um uf V«(>or Claud OU|wr*lnn Mixltl. A & / ^ 1
during a controlled nitrogen tetroxido (N^O^) spill. Ttw cold ground
temperatures and especially icc formation in wot soil can sim" f(NRe'NSc) (4-11)
where ¦ Sherwood nun\ber
d ¦ effective pool diameter (m)
0ffl - molecular diffusivity of the compound (units s~*),
NRe « Reynolds number, and
NSc " Schmidt number (» kinematic viscosity, u, divided by the molecular
diffusivity, C^J.
30
F5
>c<
-------
ASC/\ 7/9.S
4. Source Kmi»«ion Model#
Typically, standard chemical itng i nocr i oq empirical correlations, baMxJ on
l«\nunar or turbulent flow over flat plates, are used to calculate k^ in
equation (4-11), and Fleischer (1980) gives several examples (see eq. 4-21) in
his discussion of tho SPILLS model.
The other main method to calculate the mass transfer coefficient, k^, for
slowly evaporating pools is a meteorological approach based primarily on the
work of Sutton (1953), who solved the steady-state atmospheric diffusion
equation over a liquid pool with a power-law velocity profile and a
correspond ir>g powr-law oddy diffusivity profile. An expression for kg as a
function of windspeed, atmospheric stability, and liquid pool dimension has
boen used by Kackay and Matsugu (1973), who formal a correlation based on
experimental data on evaporation for neutral atmospheric stability:
kg - 0.00482 NSc-0*67u0-78d"0*U (4-12)
where u is windspeed in ms~*, d is depth of pool (m) , and kg is in m s-''. This
expression has boen conmonly used in source emission models to calculate kg for
the evaporation rate from single component pools.
If the liquid boiling point is above the ambient temperature, then
starriard formulas for evaporation fluxes can also be used, such as the mass
transfer equation (Hanna et al., 1982):
°a " Co" (pgs "V «-»>
where pgS is the saturation density of the gas at ambient conditions and Pg is
the actual density of the gas at a height of 10m above the pool. The drag
coefficient, Cq, is about IB"3 if u and Og are measured at a height of 10m.
Liquid spills on water will generally increase in area with time, as will
unconfined spills on land, caused primarily by gravity spreading. Shaw and
Briscoe (1978) present a good summary of the equations used to calculate how
unconfined spills increase in area on both land and water. However, storage
tanks on land normally have some form of confinement around them in the event
31
F6
-------
Guideline* fur U»f of V»por Cloud l)l«|>*nton Mod*l« /\_L (31 EL1 / *3 ?" ~J
o( .in <»:c idont. This irvjy bo a dike, a dirt embankment around the tank, or in
com: casus, the acca surrourdincj the tank is dug out to hold the tank contents
in the event of a failure.
This con(inc/Tont can be extremely important in limiting the emission rate
from a storage tank failure. As shown by equation (4-10), total evaporative
emission rates are directly proportional to the area of the spill, and thus
limiting the pool area is a very effective method to limit source emission
rate.
Figure 4-5, reproduced from Shaw and Briscoe (1978), presents some
calculations of source emissions as a function of time for both confined artl
urconfined cases of a 1000 LNG spill on land. For both instantaneous and
continuous (finite rate) spills, the confined source emissions at a given time
are roughly a factor of five less than the unconfirmed emissions.
4.1.6 Liquid Pool Evaporation (Multi-Component)
Many potentially-hazardous liquids are multicomponent, ard hence cannot be
treated by the single equations given in the last subsection. Drivas (1982)
has used these equations for single components to derive equations for
rnulticomponents. Assume that Qa is the total evaporative emission rate (kg ra~2
s_1) of all components. Then tl>e following empirical equation is derived:
Qa " (kgH-rVfry^T)) ^.{xi°Pisraie-kPisC)/j^i^i (4-14)
where kg is the mass transfer coefficient for the gas phase defined by
equation (4-12),
Hp0 is the total initial mass of evaporable liquid (kg)
nT is the total moles of liquid
Xj° is the initial liquid rrole fraction of component 1
pis is the saturation vapor pressure of component i (newtons m"^)
m- is the molecular W:lght of component i (kg mole"*)
k - k9 Ap/fry^T)
R is the universal gas constant (8.31436 joules mole"* °k-1)
32
F7
(V7
-------
APPENDIX G
P.EER-REVIEH COMMENTS AND AUTHOR RESPONSES
filAl PR, HILLIAhl-E
GIB; AUIHQFLR£,SPQMS£5-IQP
G.2A; S£UROY/[>ELL'QRCQ PEEOEVIEH CQHKiNIi.
G2B: AUTHOR RESPONSES TO SCHROY/DELL'ORCO COMMENTSL
-------
CIA: DR. HILLTAH PQPENDOPp'^ PFFR RFVTFH COMMENTS.
»f1
-------
Review of
"Evaporation Race of Volatile Liquids
Final Report
(EPA contract no £6-020'»2
-------
adhere to the wire as the liquid level drops. A wire procrudlng from below ^
(or possibly cwo wires procrudlng trom below) provides a much more adequate
indicator of liquid level. A more precise description of the location and use
of the wire should b« explained; because It was used as a ca Iibra.ion device
for the automated liquid level control, Its valldlcy should be made very
The anchors n,ake consents on page 15 suggesting alternative ways of
"present ing the data." The fibres Included in Append!* , describe change, ln
evaporation rare as a function of air temperature and liquid temperature
Clearly liquid temperature will determine the vapor pressure for ,ny chemical
tested. Mr temperature will only indirectly affect evaporation rate by
varying its contribution to the heat of evaporation (versus ch„t fro„ che
liquid) and thereby affect the liquid temperature and its associated vapor
pressure. As later statistical analysis shoved, air te.p.ratura uas not .
significant predictor of evaporation rat., u would be of .ore Interest to (j
show the effect of liquid temperature and velocity on evaporation rate. Such
a trend in evaporation versus velocity would be useful in justifying the
authors separation of 100 fp. date from the 500 and 1000 fp. data, UMIe suct,
a separation .ay In fact be justified. It ls ,t th, auchoc,s
rather than convincing the readers that • sharp differential occurs somewhere
between 100 and 500 fprn.
The full nature of che various statistical tests of associations or
correlations within the data is quite unclear. There is a considerable amount
of existing theory that prediccs correlations with such combined parameters as
che Reynolds Number and Schmidt Number. As these dimensionless combinations
G1A-2
/W
-------
are made up of other Independent variables, the authors wisely chose to
simplify their prediction by the direct use of these other variables such as
vapor pressure and air velocity. However, ic la unclear whecher or how the
other variables such as dlffuslvity were evaluated. Were chey cested in the
appropriate manner? What test was used? l«"hac contribution would other of the
various variables and parameters listed In Appendix A have contributed to a
more accurate prediction? Parenthetically it also appears that the
diffusivicy listed in Appendix A varies as a function of liquid temperature
rather than air temperature. Air temperature should have been used since It
is diffusion in air rather than in or through liquid. It appears that
molecular weight vas used in this report instead of diffusivicy. Diffusivicy
is in fact a function of molecular weight and could if correctly calculated
have been as good or better predictor than molecular weight in accordance with
mass transfer theory.
The use of analysis of variance to refine multiple coefficients and
exponents to minimize model error nay have been a misstatement. A more
appropriate test is a stepwise regression (which may have been called step
wise analysis of variance by these authors). Stepwise regression Is a
GP
nathe-atical procedure which „U1 select th. „„c statistically ll4nlflc.nt
parameters or sets of para„.crS to „lnuiz. th. varUnce „p„llMnMl'Ij
values and the nodal . It lUo provide a of the „Uclv,
reduction In residual error and Its statistical al5nlflcanc. as each
succeeding variable is added to th. .odel. the actual test used „ay „r,ly
need co be properly stated in this report. * logistic model could also have
been used in this analysis, perhaps more successfully.
G1A-3
-------
r,W
The next paragraph In the report describes coefficients .md tables
£ h,i r ,1 c t o r I : I ng the agreemcne between the model and the d.it.n Again the
nuthors were at least misleading by stating that a line represents
perfect correlation." In fact data on any line represents perfect correlation
or a correlation coefficient of 1.0 between x and y values; other lines will
only differ in their slope or/and Intercept. Ic would be better to show
errors between the model and the measured value on the y axis as a function of
jnv measured single or combination of variables on the x axis. Also wlch
reg.ird to che figures beginning on page 17 and 18, Tables II and III refer to
"Appendix 3" when in face chey should refer to figures in Appendix C.
The implication is made on page 26 chat predictions made by the Wade
equation 16 are not a good predictor of evaporation race as measured by IHE.
It is crue there is noc good agreement becween the Uade equation and the IHE
daca, but there is a good correlation between the two. The errors are
linearly correlated with Increasing races of velocity indicating an easily
accounted for difference in che constants within each equation (a difference
likely to be related to slight differences between experimental setting
geometries) .
Along the lines of comments of a more editorial and technical presentation
nature, it would seem thac the last sentence on page 5 should be changed to
indicate thac che temperatures of open surface tanks are uncontrolled rather
chan co"Crolled. Similarly, che last sentence on page 14 under air velocity
is sorcevhac confusing and perhaps misleading; It might be beccer stated "the
flow profile was in all cases quite uniform vich less chan 10* disagreement
between the anemometer and the average velocity." Ic may in fact have been
G1A-4
-------
pf V
less chan 90* disagreement, but it Is unclear from the wording as stated "with
better than 10> agreement,"
The reference to testing the significance of "viscosity of the liquid" on
p*ge 15 seems Interesting, buc inappropriate. Differences in viscosity are
unlikely to affect evaporation from neat or pure chemicals, other than as It
n'lghc effect wave or ripple action. The face thac data presented later in
this report indicates no discernable effect from velocity (and therefore wave
action) indicate^ that viscosity did not even have the effect that might be
hypothesized by its physical importance within the system. Viscosity probably
does have some effecc on differential evaporation of components from a mixture
of solvents (which was not the case here). Rather than discard its potential
value in che future, it might have been better to clarify the potentially
inappropriate interpretation of viscosity in this system.
It was noced In Table 6 thac the average error between the Henna & Drivas
equatlon and the IHE data was quite small (average errors of approximately
9*). However, che average error Is not a very good predictor of the error for
any single measurement. It would be useful to also show in Table 6 the
c°efficlent of variation or geometric deviation of the error.
Some further clarifications would be appropriate on page 26 related to the
e Eq - 16. The text implies that this model predicts evaporation in still
lr- There appears to be two v's in Wade Eq - 16; one a large V, presumably
elocityp the other a small v of unknown designation. Further down that same
Pa&e is another equation (referred to as Wade Eq - 9) for velocities exceeding
2 f -j
ceet per minute; the coefficient preceding that equation (1.57x10*') does
G1A-5
-------
p & r
noc seem consistent with the Uade equation 16 above. Some cross checking
would clarify che appropriate v*lue.
It Is also noted on this page and a couple of ocher times chat che word
"form" has been used Instead of "from". Editorial attention to this detail
should be exercised. Some typographical errors were also noted on page 27 in
re htion co graphs, such as "dimentionally.
Finally, I prei^nc some comments with regard co future research such as che
auchors refer to on pages 2-3 and 16 and list on pages 28-29. First in
relation to some practical uses for che current daca, che proposed use for
risk assessment in spilled liquids (items (5) and (7) on pages 2-3) requires a
currencly non-existent local dilution or dispersion model. Such a model would
provide exposure assessments for people within 1-3 diameters of a spill rather
Chan from downstream plvune hazards. In a closely related practical limitation
is che discussion on page 16 regarding liquid temperatures and solar flex. I
do noc believe it would be easy Co measure che liquid cemperacure eicher on a
sPill on a floor or in an outdoor environment in which solar flux would also
play a factor as pointed ouc by the auchors. Thus che need to conduce furcher
studies co characterize Che effect of heat flow boch from che floor, che
earth, or the sun upon evaporation races would be a useful extrapolation of
the current daca. The importance of this addition would depend upon che
aPp 1 icacion of this cype of daca co EPA or other programs.
The importance of pool size was also discussed on page 16; chis effecc
c°uld use some furcher clarifications buc given che wide scale of possible J(
Slzes and che uncesced value of existing theory, is noc a high prioricy issue.
G1A-6
-------
e7
It Is doubtful whether pool size would have an effect -wer small surfaces of
loss r.h.in snveral feet. A morn dlrecc test In outdoor environments might he
needed to lec the down-wind pool length vary sufficiently The effect of low
velocity on average evaporation rate has the same effect a r flow across a
wiier f-ool. There are mathematical means to estimate the nvurage evaporation
rate through a boundary layer of changing dimensions across a pool of defined
width.
Yhe effect of free board was also discussed on page 16 and listed as item
/
^2 on page 28. To some degree this effect could be evaluated by comparing the
results of this study to those of Bishop et al, (American Industrial Hygiene
Association Journal, 43:656, 1982), in which a tank of similar dimensions was
used with a free board. From a practical perspective, the importance of free
board would cause a difference In evaporation rate from that of a spill upon
relatively flat ground as discussed above. The magnitude of this difference
would of course depend upon the size of the free board both absolutely and
relative to the downwind pool dimension. Experiments conducted with a
simulated free board (item 2) are probably the most Important and useful area
of future work suggested by the authors. Solvents in a tank with a free board
are probably Che most common source of environmental setting in which
evaporation would be important other Chan spills onto modestly flat surfaces.
Such studies of free board need to modeled as a function not only of free
board but also of source diameter (in the downwind direction) and air
velocity. The interaction between these three parameters would need to be
investigated in an apparatus similar to that tested herein. In the interim
and given the current uncertainty of the magnitude of this effect, a policy
decision would be useful defining how or under what defined set of conditions
GlA-7
-------
p8 C
or settings of EPA interest this data and report should be used.
'\
Ocher teems listed under future work have some merit but many are perhaps I
easily addressed, for Instance item 1, Improving the apparatus. Incorporating
even a small free board would nlleviace experimental problems. In particular,
with the "stilling basin" jnd the "ripple dam" while potentially (if kept
small) having a negligible effect upon evaporation (such a system might in
face be a better simulator of real world settings). I also reference my
earlier comments with regard to 2 or 3 (or more) submerged thin wire poincers
which could provide a simpler method co measure volume change with very small
toleranees.
Items 3, 4, and 7 are related co testing further chemicals. While further
chemicals might be of some interest, particularly much lower volatility
chemicals which could present problems in indoor settings might be of
interest. The likelihood thac they will obey similar laws of thermodynamics
and evaporation are quite strong. Of equal importance is the issue of
mixtures which was also addressed by the work of Bishop et al and
theorecically calculations and predictions are also available which could be
simplified and made more available Co EPA.
Similarly, the listed possibility (items 6 and 8) of studies using ocher
than isothermal conditions is noc recommended because they may not represent a
common real-world circumstance and would be difficult to standardize. For
inscance, a spill near isothermal conditions between the solvent and the
receiving ground, pavement, or what have you, would be more likely than
spilling a heated or cooled solvent (except in some already recognized
G1A-8
-------
situations such as LNG tank ruptures). One can also end up Ln a non-
Isothermal condition with high vapor pressure chemicals because they require a
large thermal flux co supporc evoporaclon and thus presenc more complicated
problems of simultaneous heac and mass transfer. The equilibrium state of
such a non-Isothermal condition in the open environment would depend upon the
rate of heat flux through various soil or pavement conditions which would be
difficult to predict in any given setting. Non-isothermal conditions provide
an interesting ac.idemlc setting, but because they are both unusual and
difficult co predicc, the results of Isothermal conditions are recommended as
a reasonable basis for many assessments.
G1A-9
-------
GIB: AUTHOR RESPONSES TO POPENDORF COMMENTS.
1*1
-------
fXICS
Olficos.
Minneapolis, Minncsoia
Tampa, Florida
I I i Coralville, Iowa
laboratories
**0*IIVONM ANAlTtlCAi CMMA1H 4 INOtNIMlMQ
inc.
1710 Douglas Drive North r; Minneapolis, MM 55422 u Phone (612) 544-5543 u FAX (61 2) 544-3974
February 27, 1989
Dr. William J. Popendorf
University of Iowa
Institute of Agricultural & Medical AMRF
Iowa CIty, IA 52242
Dear Bill:
I had submitted your name as a possible reviewer of our "Evaporation"
report because I knew you were knowledgeable on the subject. I forgot
the hazards of the reviewer knowing too much!
Kidding aside, here's a draft of our responses to your comments. Please
call If you have anything to discuss. We do appreciate the time and
effort you devoted to the review, and the report will be much Improved
due to your efforts.
Very truly yours,
Knowlton J. Capian,
Senior Consultant
IHE Division
K0C162/jb
G1B-1
an equal opportunity employer
-------
February 6, 1989
EVAPORATION RATE OF VOLATILE LIQUIDS
Response to Remarks
by HI 11 lam J. Popendorf
Dr. Popendorf's review does not have numbered pages so we have taken the
liberty of numbering the pages and also numbering the areas of comment.
Host of Dr. Popendorf's comments are quite pertinent, show a depth of
understanding of the subject, and are appreciated. Since he 1s fam111 ar
with the subject, he has suggested several possible further manipulations
of the data which, while Interesting from an academic point of view, were
not possible within the budgetary and time constraints and are not
directly related to the objective of the project.
1) Page 1, bottom, and page 2 top, Popendorf discusses problems of
using a wire point for a level Indicator. The th 1 n wire In this
experiment was suspended from above the liquid. However, the
problems that he describes were minimized In this experiment. Each
setting was for an Individual liquid at a fixed temperature, and
therefore the surface tension for each case would have been quite
constant. This was confirmed by multiple trials of "point-setting"
and comparison with the level control.
The wording In the report will be clarified to reflect this.
2) Page 2, middle of second paragraph. He do not understand this
comment, because the curves In Appendix B do show the variation with
both air and liquid temperatures. The separation of the data for
100 fpm from the 500 and 1,000 fpm data was done because the best
fit regression equation was significantly different for those two
groups. The multiple correlation coefficient, RS for 100 fpm data
using the equation for the higher velocities was lower than for the
equation presented for the 100 fpm group. The wording in the report
will be clarified to indicate this was made for that reason rather
than at the "author's discretion".
Obviously it Is the liquid temperature that governs the liquid
vapor pressure, however, since the air temperature Is virtually
the only source of heat Input in the simultaneous heat
transfer-mass transfer phenomena In this case, the air
temperature directly effects the liquid temperature. Since air
temperature and velocity are the Independent variables
-------
2
3) Page 2, bottom of page. We are aware of the existing theory that
predicts correlations using parameters such as Reynolds Number and
Schmidt Number. Both of these parameters were tested, as described
In the 3rd paragraph of Page 15 of the report. Further, the
diffuslvlty Is part of the formulation of the Schmidt Number.
In the analysis of the data, all of the variables tabulated In
Appendix A were tested against the data to determine the best fit;
and the best fit Is as reported. There seemed little point to
tabulating and discussing at length all these negative findings or
poorer fits, but no doubt the wording of the report on this subject
could be amplified and clarified.
At this point It Is further pertinent to mention that the objective
of the research was not to refine or amplify the classic theories of
heat transfer and mass transfer; It was an attempt to find a
simplified and adequately accurate approach to the evaporation
phenomena over a range of conditions that Is quite narrow as
compared to classical chemical engineering practice.
4) Page 3, center, first paragraph. Dr. Popendorf is concerned that
"It appears that the diffuslvlty listed In Appendix A varies as a
function of liquid temperature rather than air temperature". He are
unable to understand this remark. The diffuslvlty of the vapor was
computed at the air temperature.
5> Page 3, first paragraph, end. The diffuslvlty was attempted as a
predictor and was not as good a predictor as molecular weight In
correlating our data, as stated in the 3rd paragraph, p. 15, of the
report.
6) Page 3, last paragraph. The first part of this paragraph Is not
understood by us. Perhaps the problem Is one of terminology. We
did not use the term "stepwise analysis of variance."
7) Dr. Popendorf objects to us stating that the 45 degree line
represents perfect correlation. He Is correct. The line represents
perfect agreement between data and model. We apologize fo' the
sloppy language and have corrected the text.
The graphs as presented show the formula or "model" evaporation rate
on the X axis and the measured evaporation rate on the Y axis. The
slope of the line is 45 degrees. If the formula evaporation rate
and the measured evaporation rate were exactly the same, the point
would fall on the 45 degree line; the location of the starred points
Is an Indication of the difference between measured and formula
evaporation rate.
He have clarified the language In the reports.
G1B-3
lit
-------
3
8) Page 4, end of first paragraph. Or. Popendorf Is correct, the
Tables 2 and 3 are Incorrect In their appendix references. This
will be corrected.
9) Page 4, second paragraph. Or. Popendorf obviously feels that any
orderly relationship of one value to another represents a good
correlation and prefers the use of the term "agreement". We have
corrected our wording accordingly.
10) Page 4, last paragraph. Dr. Popendorf Indicates that he believes
that the temperatures of open surface tanks are uncontrolled rather
than controlled. This Is a simple matter of fact, many such tanks
have temperatures which are controlled, and many have temperatures
which are not controlled. Hlthout having any factual data, It Is
the Impression of the Investigators that more such tanks have a
controlled temperature rather than an uncontrolled temperature, for
the case concerning tanks whose temperature Is controlled, the
sentence 1s valid. Dr. Popendorf is right; "less than 10X
disagreement" 1s a more precise statement than "better than 101
agreement".
11) Page 5, first paragraph. Or. Popendorf feels that we could have
discoursed at length on the potential for viscosity having an effect
on evaporation rate. In general, the project did not Involve a
theoretical treatise on mass transfer from liquid to gas phase arid
was not SO Intended. Certainly, viscosity did not seem Important to
evaporation rate In cur tests. It was included In the program
mainly because It Is one of the directly and easily measured
"handbook" properties of liquids.
Further, our experiments were designed to minimize heat transfer to
the liquid by any route other than the evaporating surface. Thus a
heat flu* from air to liquid exists at the evaporating surface.
There may be a temperature difference between the boundary layer of
the liquid and the body of the liquid, and viscosity would affect
the mixing In the 1Iquld.
In any event we fall to understand why Dr. Popendorf criticizes Its
Inclusion, especially since little cost was Involved.
12) Dr. Popendorf points out that we could have more thoroughly analyzed
the Hanna and Drivas data, which Is of course true, but somewhat
beside the point and certainly outside of the scope of this
project. Dr. Popendorf probably missed the point, which Is that 1n
the Hanna and Drivas equation, the vapor pressure to be used is the
vapor pressure of the liquid at the temperature of the air, as shown
In Table V, and which leads to very erroneous results as compared to
our data. In our Initial massaging of the Hanna and Drivas data, we
G1B-4
-------
4
mistakenly used the vapor pressure at the liquid temperature, which,
of course, would make more sense, and got better agreement with our
data as shown In Table VI. It Is possible that the Hanna and Or Was
publication was merely a misstatement, that they Intended to call
for the vapor pressure at the liquid temperature; It Is Impossible
to tell from the publication.
13) Page 5, bottom of page. Dr. Popendorf Is concerned that there are
two V's In an equation, a small v and a large V. We apologize for
this typo error, both of them should be the same symbol. The Wade
equation 9 Is exactly from Wade's publication.
14) He agree with Or. Popendorf's complaint about typographical errors
and will make a strong attempt to correct them.
15) Page 6. center paragraph. Or. Popendorf says that complete rls1'.
assessments cannot be made from evaporation rate alone. That Is, of
course, correct. However, If the generation rate of the contaminant
Is not known, any dispersion models fall. This research Is an
attempt to provide a better way for estimating contaminant
generation rate over the narrow range of air velocity and
temperature than currently exists.
Further, on the basis of our field experience, we would have the
following commentary on this subject.
In the Interior of a manufacturing plant, the airflow currents
are random, or highly variable, as frequently as they are
defined. The random or variable air currents will not produce a
downwind plume such as would be produced by a defined air
current. Outdoors, except for the "dead calm" situation, wind
direction Is usually definable within an arc of 20 to 30 degrees.
Therefore, we feel that knowledge of evaporation rate I.e.,
generation rate, Is essential to any dilution or dispersion model.
16) Dr. Popendorf's comments on tii? Importance of pool size seemed to be
an expression of his personal opinions on the subject with which we
only partially agree. In any event, we feel the effect of pool size
can be Important, especially for fast evaporating liquids where the
boundary layer concentration of vapor would Increase rapidly with
distance from the upwind side of the pool. There are published
equations for evaporation rate which Include correction factors for
large outdoor pools, which we will not bother to annotate.
17) Dr. Popendorf discusses the effect of freeboard and other parameters
and suggests our data be compared with that of Bishop. While we
agree with most of Dr. Popendorf's qualitative statements, we do not
find any data or equations In Bishop's work which lead directly to
computation of an evaporation rate.
GlB-5
-------
5
18) The remainder of Dr. Popendorf's comments are addressed to which
portions of the possible future work he feels are most important.
In general, we agree with his predictions of the usefulness cf
different types of testing and lines of investigation. However^ he
persists in his misunderstanding that a pointer was used
incorrectly; it was not, as explained previously. The improvements
needed are not to improve the level control result, which was not
done by pointer but by optics and controlled generally to +0.1 mm.
The improvements needed are to reduce setup and calibration time and
to remove hardware in the wind tunnel downstream of the pan.
G1B-6
-------
-------
Recently, Industrial Health Engineering Associates has completed
experimentation regarding the emissions rate of volatile organic
compounds. This work has resulted in several equations correlating
emissions to various experimental parameters for the limited system.
The purpose of this paper is to examine this experimental work,
documented in "Evaporation Rate of Volatile Liquids", and comment on its
validity and applicability.
Emissions rates collected from this experimentation have been
correlated with apparent accuracy to several chemical physical properties
and system variables. These correlations are only valid for the
experimental conditions.
Included on the following pages are a discussion of the experimental
techniques, a data comparison with the fortran program Emissions from
Spills, authored by Wu and Schroy, and final comments and
recommendations on this work and future works.
Summary
It is recommended that this work not be published, due to its
experimental procedures and inconsistencies, and its limited scope.
Emissions estimates, and emissions correlations, in this experiment arc
suspect because of the following reasons:
--too many parameters are varied per set of wind speed runs,
--important parameters, such as downstream air temperature and
pressure above the pan are not listed.
--mass loss effects do not balance with heat loss effects, no matter
what pan material is used, and given the experimental data.
--it is likely that more heat transfer occurs in the system than is
suggested.
--data correlation and analysis was not handled by an acceptable
algorithm.
and --the experiment is of such limited scope that the emissions
equations mean nothing in any semblance of a real scenario.
Data Discussion
The apparatus described in the manual has been designed to
eliminate heat transfer parameters as effects on evaporation. Efforts are
made to keep the liquid at a constant temperature, so that the on!'
effective heat transfer occurs at the liquid-air interface. Important system
1
G2A-1
-------
temperatures are constantly monitored to insure that heat transfer does
not affect the emissions rates.
Similarly, rhe liquid level in the pool and the base temperature
remain constant to eliminate these variables from emissions
considerations. In effect, the experimenters vary wind speed, wind
temperature, and liquid temperature only.
The first apparent problem with the data collected is the varying of
too many parameters per experiment. For nearly all chemicals, evaporation
rates are studied at three wind speeds, 100, 500, and 1000 fpm. Three to
four runs are made at each of these wind speeds. A problem occurs,
however, because there are two variables at each wind speed, these being
air and liquid temperature. As a result, meaningful data cannot b-t
analyzed with respect to either of these variables alone. The graphs
relating emissions rates to air temperature and liquid temperature have no
real meaning, as another variable exists in each case.
It is good practice in collecting experimental data to keep all
variables constant, save one. It is recommended first that the air
temperature be the variable, while the liquid temperature remains
constant for all runs, even at higher wind speeds. This would allow
analysis of the effect of air temperature only on emissions rates. Similarly,
for the same wind speeds, air temperatures should remain constant while
liquid temperatures vary. This way, liquid temperatures could be related
to emissions rates. These studies would involve much more data collection,
but would result in meaningful relations between emissions rates and air
temperatures, between emissions rates and liquid temperatures, and thus
between emissions rates and vapor pressures. As stands, the data analyzed
will lack in significance because of the multi- parameter variation.
Another piece of data notably missing is the absolute pressure inside
of the duct. It is quite possible, since there will be a pressure drop across
the duct, that the vapor pressure could vary from 760 mmHg by as much
as 20 mm. Atmospheric pressure will vary slightly with temperature and
humidity conditions. If the pressure is lower than 760 mmHg, an
overestimation of emissions will result.
Furthermore, the outlet air temperature and air temperature above
the pan should also ' e monitored to insure that the convection heat
transfer to the liquid is constant. Variations in air temperature will effect
the emissions rates of the constants.
Other possible data that could effect the emissions rates are ignored.
Surface velocity has been related to volatilization by other authors, but
this is not assumed to effect emissions in this experiment. Similarly, pool
depth could also be a variable, as could pool area. Granted, it would be an
impossible task to examine the singular effects of all system variables, but
2
G2A-2
-------
questions regarding other possible variables should at least be addressed
in the paper
The correct way to find equations correlating experimental data
would be to plot, on logarithmic paper, an independent variable against an
dependent variable with all other parameters constant. If the result is a
straight line, the slope will represent the exponential effect of the variable
and the y-intercept will be a constant. This will give the singular effect of
the parameter, if an effect exists. All other parameters must remain
constant. The data correlation method used by IHEA is suspect; using this
method, variables which have no effect on emissions could be included
with a high correlation coefficient.
Using a correlation technique such as ANOVA, one could easily obtain
relations that are physically meaningless, or purely coincidental. One
should also attempt in data correlation to allow for correct unit output It
would be helpful if the authors would discuss the physical meaning of
their equations, if any exists. For example, it is difficult to believe that the
molecular weight will have a direct exponential effect on emissions, as is
presented in some of the ANOVA results.
The apparatus also needs a better description in the manual. The
photographs and the figures arc difficult to read. Also, it should be noted
what areas are insulated, the construction materials of the duct and the
pan, and the conditions under which the pan was initially filled
(evaporation will result from the turbulence created by a pouring liquid).
Physical properties of the metals and plastics used, such as heat capacity
and thermal conductivity, would also be useful information.
Discussion of Collected Data
After the report was read and analyzed qualitatively, as reported
above, a quantitative comparison was done comparing results of the
experimental data contained here with data collected from the fortran
program Emissions from Spills, authored by Wu and Schroy. Interestingly
enough, trends comparing Spills data to this experimental data were
similar to those in Table VI of the IHEA report. For low wind speeds and
air temperatures, Emissions from Spills underestimate.'! the IHEA data,
while for high wind speeds and air temperatures, it overestimated the
IHEA data. This led to an investigation of why these trends held.
In the IHEA data, the latent heat of the evaporateo liquid is assumed
to be offset by the convective and conductive heat input to the liquid.
These values must offset each other if the liquid temperature is to remain
constant.
A serious problem arises in the IHEA data because the convective
heat from the air, when coupled with the conductive heat from the pan,
3
G2A-3
-------
will not equal the evaporative heat from the liquid. The only possible
explanation is that the heat goes to the air, but this is impossible to know
because even if the downstream air temperature is measured, it will not
account for any of the organic that might condense on the duct walls.
Because it is unlikely that this latent heat goes to the air (most volaiiles
will exhibit evaporative cooling), it is suggested that another heat source is
present, or that the temperatures have not been correctly measured.
This is shown by Table 1. In this table, the ratio of k/B has been
calculated with the data of the IHEA. This ratio should be constant across
any wind speed or temperature range, if there are no other heat effects in
the system. However, Table 1 shows that k/B is anything but constant
across the range of wind speeds and air temperatures. Furthermore, no
trend is evident that would lead to any conclusion on the source or
functionality of the other heat source. When the pan temperature is less
than the liquid temperature, k/B is computed to be negative. This is
impossible. Therefore, another heat source or sink must be present.
It is said in the IHEA paper that the only significant heat transfer
occurs at the liquid air interface. This statement is incorrect. The heat
transfer at the base/pan interface will be important for large temperature
gradients. For'some liquids, the temperature difference between the base
and the pan is as greater than 6 degrees Fahrenheit. If the pan has a large
thermal conductivity, such as an aluminum pan would, the heat transfer
from the base to the pan will be the largest source of heat transfer. Even
for stainless steel, the conductive heat transfer will be much greater than
the convective heat transfer from the air, in some cases.
In effect, heat transfer occurs in three ways. Heat is conducted to or
from the liquid from or to the base. Heat is lost from the liquid by
vaporization, and heat is convected to or frcm the liquid from or to the air.
Unless the air is heated, the convective and the conductive heat must
balance the vaporized heat, if no other heat transfer occurs.
Table 2 shows a comparison of the emissions rates calculated by
Emissions from Spills with the experimental data from the IHEA. A low
wind speed was used to reduce convection effects caused by wind speed,
which is an uncertain functionality. Emissions from Spills runs were
organized so that all heat canceled, as is said to occur in the IHEA
experiments. The pan was assumed made of stainless steel, although
another material would be able to be input with similar output. As can be
seen from the table, Emissions from Spills underestimates the IHEA data
by approximately a factor of 4 for all cases. All the heat transfer in
Emissions from Spills is accounted for. The heat transfer in the IHEA data
is suspect, because of lack of additional experimental data.
Wind speed will eventually make values from Emissions from Spills
greater. This is so because the wind speed effects two parameters in the
4
G2A-4
-------
Table 1
Sample Calculation for Heat Balance on IHEA Data
Heat Balance, if only convective, evaporative and conductive heat:
^evap ''cond + ^conv
q =1" * (T - Tr ) ; q =h *(T . - Tt. ) ; q = AH * E
Mconv B pan liq xonv c air hq' ^evap v
k
— is unknown, but should be consiant for each chemical, as k and B will both
O
vary little with temperature changes. Therefore, if there is no other heat
k
transfer other than that described, we can solve for g for each chemical.
Below is a table for methanol. The h term can be calculated by b =.8 ~ .24*V
c ' c w
No trends are detectable in the data . This is partially due to the existence of too
many variables. Other chemicals would show similar trends.
Windspeed
fpm
Air Temp
Fahrenheit
Pan Temp
Fahrenheit
Liq Temp
Fahrenheit
k
B
Btu/ftA2-hr-F
100
51.5
42.5
35.7
34.9
100
70.5
43.5
39.2
56.0
100
99.7
46.6
45.8
308.1
100
120.8
47.8
47.6
1571.5
500
45.3
42.5
34.0
26.2
500
70.2
42.9
38.9
36.1
500
100.7
44.2
43.2
57.3
500
119.7
45.6
47.5
-18.0
1000
44.2
41.8
32.2
41.1
1000
70.7
43.6
38.3
70.2
1000
100.1
37.9
40.8
-102.4
1000
120.4
46.9
48.2
-267.3
q „ » heat flux rate due to convection. Btu/ftA2- hr.
conv
q * heat flux rate due to conduction. Btu/ftA2 -hr.
^conv
q = heat flux rate due to evaporation
6 V1 p
k = thermal conductivity of pan, Btu/ft-hr-F
B » pan thickness, ft.
T « temperature, Fahrenheit
AHy » heat of vaporization of organic. Btu/lb.
h » heat transfer coefficient for convection. Btu/ftA2-br.-F
c
G2A-5
-------
Table 2
Chemical
Air Temp. F
Liq Temo. F
Emissions,
Soills *
Emissions,
I HE A •
Methanol
70.5
39.2
0.178
0.52
n-propOH
70.5
52.9
0.046
0.24
Acetone
70.7
36.2
0.448
1.55
MEK
69.6
48.8
0.261
1.01
Hexane
70.7
37.3
0.437
1.91
Heptane
70.5
50.2
0.154
0.98
Octane
69.6
55.8
0.0558
0.32
Benzene
70.1
40.6
0.243
1.12
Toluene
69.5
51.3
0.109
0.40
Xylene
100
83.4
0.0629
0.40
* Emissions are in Biu/ftA2-hr.
Windspecd = 100 fpm (basically stagnant air)
Thermal Conductivity of Pan = 9.4 Btu/ft*2-hr. (Stainless Steel), for Spills
n-propOH » n-propanol
MEK = methyl ethyl ketone
2-octanone and l-pentanol not applicable at 100 fpm wind speed
G2A-6
-------
Emission from Spills model (wind speed and convective heat transfer
coefficient), while it only has a singular effect in the IHEA model. The
variance is such that for Emissions from Spills, a higher heat convection
rate will occur for higher wind speeds. Since the IHEA model claims no
heat effects, an increased wind speed will not effect results from heat
transfer effects (actually, both effects may be accounted for in the
exponent for wind speed that IHEA uses, but the dependences will still
differ. Wind speed has a greater effect as it increases, in Emissions from
Spills).
It would behoove the authors to include a heat balance for each case
to show that heat will balance as well as mass. They only show half of the
picture. The heat inputs and outputs are suspect numbers; it is possible
that there is a greater input heat than is implied by the data.
Heat transfer is a very difficult parameter to eliminate from
consideration in any experimental apparatus. All surfaces will absorb or
conduct heat, and the extent of the heat transfer may often be difficult to
calculate. Heat is constantly being transferred at ail interfaces, It might be
more productive to allow all heat transfer and estimate emissions. More
experimental work and a different procedure would be required, but the
result would be a more realistic modeling approach. On the same note, it
would be instructive in future studies to emulate solar radiation and heat
transfer to the ground. Both of these process simulate real situations and
would be useful for accurate emissions modeling.
Conclusions/Recommendations
Because of there experimental and data problems shown herein, it is
recommended that this report not be publish^. The results from this
experiment are of no practical use because of too much parameter
variance, and because of the inability for the system to account for the
heat loss from the liquid.
Furthermore, any results would be void of practicality because they
ignore the effects- of real spillage surfaces, such as asphalt or soil.
Additionally, no radiation heat effects are accounted for. In a real scenario,
heat effects will play a major role in emissions estimates, as evaporative
cooling will cause temperature drops for most volatiles, and the vapor
pressure driving force will be reduced by the temperature drop. To treat
evaporation as an isothermal process is errant.
In further studies, it is recommended that experiments let heat
effects take their toll, and measure emissions on different surfaces with
different solar radiations, while keeping the initial temperature and the
wind temperature constant. Pool temperatures can be measured
continuously, and meaningful emissions estimates can be made. The scope
5
G2A-7
-------
of studying emissions process from an experimental point of view is
endless; their axe literally thousands of conditions that could be studied. A
more reasonable approach is a theoretical approach, using empiricism
where needed.
G2A-8
V*
-------
AlOQR-RESEQflSii-rQ-SCllRQlZDEiLLORCOjCQtfllEJilS^
-------
laboratories, rc
OMicos:
Minneapolis, Minnesota
Tampa, Florida
Coralvillo, Iowa
HOMUOMfcV AH* 1 IMOtNlttlMQ
1710 Douglas Dfive Nodh u Minneapolis* MN 55422 u Phone (612) 544*5543 Q FAX (612) 544-3974
February 28, 1988
Jerry M. Schroy
Monsanto Chemical Company
800 North Lindbergh Blvd.
St. Louis, HO 63167
Dear Jerry:
I had suggested your name as a reviewer for the report "The Evaporation
Rates of Volatile Liquids" which actually was reviewed by Philip C.
Del 1 'Oreo.
Enclosed is a copy of our response to Dell'Oreo's comments. Of course,
we will also send Mr. Dell*Oreo a copy.
I am sorry that our paper was not sufficiently clear in disclaiming any
attempt to define or extend the classical theories of simultaneous mass
transfer/heat transfer. We certainly will correct that in the final
i ssue.
Very truly yours,
Senior Consultant
KJC/mp(CAP 129-4)
G2B-1
an equal opportunity employer
-------
laboratories
Offices
Minneapolis, Minnesota
Tampa, Florida
Cornlvillc. Iowa
, inc
O* I\V«0»H *1 CMI»
-------
February 6. 1989
EVAPORATION RATE Of VOLATILE LIQUIDS
Response to Doll'Oreo Comments
Mr. Dell'Orco commented on the copy of the report "Evaporation Rates of
Volatile Liquids" which was sent to Mr. Jerry M. Schroy for review.
It Is apparent that Dell'Orco reviewed the report on the assumption that
the author's objective was to refine and/or extend the classical theory
for solving simultaneous heat transfer/mass transfer phenomena Involved
In the evaporation of volatile liquids. That was not the objective of
the work; and the Introduction and scope portion of the report will be
modified to make this very clear, so that other readers do not make the
same mistake.
The objective of the work was to determine whether, for the narrow range
of air velocities and temperatures encountered In practice and for the
intended use of the data, some alternative and simpler method for
predicting evaporation rate could be developed.
If the apparent assumption made by Dell'Orco was correct, most of his
comments and criticisms would be pertinent. However, since the purpose
and objective of the work was not as assumed by Dell'Orco, It Is
Inappropriate to attempt a response to most of his comments.
Our response to the summary points on Dell'Orco's page 2 Is as follows:
1) "too many parameters are varied per set of wind speed runs."
In fact there were only two Independent variables, wind speed and
air temperature. The liquid temperature Is a dependent variable.
The temperature of the base was controlled to match the liquid
temperature In order to ellmlnate the parameter of heat flux through
the pan.
Z) "Important parameters, such as downstream temperature and pressure
above the pan are not listed."
(a) Downstream air temperature would have no effect on the
evaporation rate. Its usefulness would be limited to providing
a separate basis for computing heat flux; and for slow
evaporating liquids, extraordinary precision would have been
required for temperature measurements.
(b) Pressure over the pan was always slightly negative and less than
1 in. w.g. Dally changes In barometric pressure can easily be
13.5 in. w.g. <1 In. Hg). The dally variance In atmospheric
pressure Is about ± 3.31.
G2B-3
-------
3) Has? loss effects do not balance with heat loss effects...."
The analysis by Dell'Orco In his Table I Is not appropriate. If our
control of base temperature had been perfect, the a 1 across the
pan wall would have been zero and Oell'Orco's equations would be
useless. With the very small a T achieved In most of the runs, the
heat transfer coefficient In his equations would need to be
extremely ("unbelievably") high to show a non-neglIble heat flux,
and thus Oell'Orco's analysis Is reduced to an absurdity.
1) "It Is likely that more heat transfer occurs In the system than Is
suggested." We did not suggest any heat transfer. We described our
attempt to ollmlnate uny heat transfer other than at the air -
liquid interface
5) "data correlation and analysis was not handled by an acceptable
a 1gorIthm."
There was no algorithm In tne usu»l s'lse of the word. The least
squares method for analyst', of \ :r isiK.t was used, and we trust that
ii. acceptable. This Is soaWi an of Oell'Orco's assumption
as to the nature of the urojict, ho I:, seeking a theoretical model
as the basis for an algorithm r^pres^ru 5ng the process.
6) "The experiment Is of such limited scope snat the emissions
equations mean nothing In any wblance of a r .cenarlo." This
extreme statement reinforces our Impress! that Dell'Orco Is
expecting something we did not propose. This ny n Is not Intended
to Improve heat-and-mass transfer cakul alibis for chemical
processes. The range of air velocity (100-1500 fpm) and air
temperature (40-140F) obviously Includes practically all Industrial
plant environments, and a large majority of outdoor weather.
G2B-4
------- |