EPA/451/R-93/001
AEPA
States
nmental Protection
y
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-451/R-93-001
March 1993
jerfund
AIR/SUPERFUND
NATIONAL TECHNICAL
GUIDANCE STUDY SERIES
Models for Estimating
Air Emission Rates from
Superfund Remedial Actions
-------
EPA 451/R-93-001
AIR/SUPERFUND NATIONAL TECHNICAL
GUIDANCE STUDY SERIES
REPORT ASF-30
Models for Estimating
Air Emission Rates from
Superfund Remedial Actions
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
March 8, 1993
-------
DISCLAIMER
This report has been reviewed by the Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency, and has been
approved for publication. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
-------
TABLE OF CONTENTS
List of Symbols iii
1.0 INTRODUCTION 1-1
1.1 Background 1-1
1.2 Objective 1-2
1.3 Approach 1-2
1.4 Uses and Limitation of Document 1-2
2.0 DISCUSSION OF EMISSION ESTIMATION PROCEDURES 2-1
2.1 Types of Emission Estimation Procedures 2-1
2.2 General Considerations of Models 2-2
2.3 General Sources of Input Data 2-5
3.0 VOC POINT SOURCES .3-1
3.1 Air Strippers 3-1
3.2 Soil Vapor Extraction 3-12
3.3 Thermal Incineration 3-23
3.4 Thermal Desorption 3-30
4.0 VOC AREA SOURCES 4-1
4.1 Excavation 4-2
4.2 Dredging 4-13
4.3 Solidification/Stabilization 4-21
4.4 Bioremediation 4-28
5.0 NON-PROCESS VOC AREA SOURCES 5-1
5.1 Covered Landfills 5-1
5.2 Lagoons 5-11
5.3 Spills, Leaks, and Open-Waste Pits 5-19
6.0 PARTICULATE MATTER, METAL, ACID GAS, AND PRIORITY
POLLUTANT EMISSIONS FORM POINT SOURCES 6-1
6.1 Thermal Destruction 6-1
6.2 Thermal Desorption 6-10
7.0 PARTICULATE MATTER AND METAL EMISSIONS FROM AREA
SOURCES 7-1
7.1 Materials Handling 7-1
7.2 Other Area Sources of PM and Metal Emissions 7-10
APPENDIX A - PHYSICAL PROPERTY DATA
APPENDIX B - PROCEDURES FOR CALCULATING EROSION POTENTIAL
11
-------
List of Symbols
Symbol
ft
ri
ri
r^
*"g
^1
M
>-i
'-'s
r-i
*-*v
/ig/m3
D
Da
De
de
Dw
E
E.
EF
ER
B,
EW
F
H
K
k
Kd
k,
ko
Iq.
K
L
1
M
Definition
Bulk density
Concentration in waste/soil
Concentration in water
Saturation vapor concentration
Liquid-phase cone, in soil
Cone, (loading) in soil
Vapor cone, in soil pore spaces
Phase transfer coefficient
Diffusivity in air
Effective diffusivity
Effective diameter
Diffusivity in water
Emissions
Air-filled porosity
Emission factor
Emission rate
Total porosity
Water-filled porosity
Feed rate
Henry's Law constant
Henry's Law constant
Overall mass transfer coefficient
Particle size multiplier
Volatilization constant
Distribution coefficient
Gas-phase mass transfer coefficient
Liquid-phase mass transfer coefficient
Octanol/water partition coefficient
Flow rate
Depth
Depth
Mass of waste
Mass of waste
Units
g/cm3
ug/g
mg/1
ug/m3
g/cm3
g/cm3
g/cm3
(Section 4.1 only)
cm2/sec
cm2/sec
cm2/sec
m
cm2/sec
g
-
g/VKT
g/sec
-
-
kg/hr
atm-m3/mol
dimensionless (Section 4.2
only)
cm/sec
-
I/sec
cm3/g
cm/sec
cm/sec
-
1/min
m
cm (Sections 4.2 and 5.3
only)
g
kg (Sections 4.4, 7.1, and 7
.2 only)
111
-------
Symbol
MW
P
P
PF
P.
Q
Q
R
r
P
S
s
SA
Scg
SL
Sv
T
t
U
Ug
V
Vy
w
W
X
XIEO
x;
XDJO!
X«
z
Definition
Volume of waste
Molecular weight
Vapor pressure
Days with precipitation
Metal partitioning factor
Erosion potential
Gas exit rate
Gas exit rate
Soil excavation rate
Gas constant
Gas constant
Gas constant
Stoichiometric ratio
Density
Vehicle speed
Silt fraction
Surface area
Schmidt number, gas-side
Silt loading
Volume of soil
Temperature
Time
Windspeed
Viscosity of air
Percent volatilized
Velocity of gas
Number of wheels
Vehicle weight
Fraction emitted
Percent moisture in soil
Fraction of waste in soil
Mole fraction
Organic carbon fraction
Enrichment factor
Units
1 (Section 4.4 only)
g/mole
mm Hg
-
-
g/m2
m3/min
m3/sec (Section 6.1 only)
m3/sec
atm-m3mmHg/mol-K (Sections 5.2 and 5.3)
cm3mmHg/mole-K (Section 5.1)
1-mm Hg/mole-K (Section 3.2)
-
g/cm3
km/hr
-
m2
-
g/m2
m3
K
sec
days (Section 7.2 only)
m/sec
g/cm-sec
%
cm/sec
-
Mg
-
-
-
-
-
g metal/g soil
IV
-------
SECTION 1
INTRODUCTION
This manual is a compendium of models (equations) for estimating air
emissions from Superfund sites undergoing remediation. These models predict emission rates
of volatile organic compounds (VOCs) and paniculate matter (PM) from both area and point
sources. The following treatment processes are covered: air stripping, soil vapor extraction,
thermal desorption, thermal destruction (incineration), excavation, dredging,
solidification/stabilization, and bioremediation. Emission estimation methods are also
presented for landfills, lagoons, and spills/leaks/open waste pits.
1.1 BACKGROUND
The U. S. Environmental Protection Agency (EPA) Air Program Office
(Office of Air Quality Planning and Standards) and the Regional Air Offices have been given
the responsibility and resources, beginning in 1987, to evaluate air impacts from Superfund
sites and to advise Superfund Regional Offices on appropriate clean-up actions. The
Air/Superfund Coordination Program was initiated to facilitate this effort and the EPA Air
Program Office is responsible for its overall direction.
Assessing the air impacts of Superfund remedial actions is a significant part of
the Air/Superfund Coordination program. These assessments are frequently required for
planning purposes prior to actual remediation. They are, therefore, dependent on the ability
to estimate emissions, rather than on site measurement approaches. Emission estimation can
be complex, so the need was recognized for simple screening procedures to consistently
analyze potential emissions from various remedial action alternatives.
1.2 OBJECTIVE
This report was prepared to meet the specific emission modeling needs of the
Regional Offices and the Superfund Program. The objectives were to: 1) Identify emission
1-1
-------
modeling needs of the Air/Superfund Program, 2) Select emission screening models for
commonly used remedial activities, and 3) Prepare a manual which clearly demonstrates the
use of these models.
1.3 APPROACH
As part of the preparation for this document, the ten regional Air/Superfund
Coordinators were surveyed regarding their screening model needs (as of 10/31/91). The
technical material for this manual was compiled from the results of an evaluation of the
existing literature; no original models are presented.
1.4 USES AND LIMITATIONS OF DOCUMENT
The simple screening models contained in this compendium will not accurately
predict emissions for all possible scenarios. In some cases, the existing field and process
data are too incomplete to adequately assess the validity of certain model assumptions. In
addition, the selection criteria for models included simplicity and ease of use in addition to
accuracy. Where uncertainty exists, these models and the default inputs have been designed
to err on the side of conservatism; i.e. to overpredict emissions. The models are screening
tools. They should be used to answer the question whether: 1) no emission problem is
likely, or 2) further evaluation of the emissions is needed.
Each section of this document contains a discussion of the assumptions and
sensitivities of that section's model; these sensitivities should be understood before the model
is applied. Default values have been provided for every model, in the event that field data
for these values are not available. Of course, greater accuracy will be obtained if site-
specific and process-specific measurements are used whenever possible. The key variable
that must be obtained from site data is the concentration or total mass of the contaminant in
the material to be treated.
1-2
-------
SECTION 2
DISCUSSION OF EMISSION ESTIMATION PROCEDURES
This section provides general background information about the models
presented in this document and their use.
2.1 TYPES OF EMISSION ESTIMATION PROCEDURES
There are many approaches that can be used to estimate emissions from
remediation processes. These approaches include:
Use of an emission model with default values;
Extrapolation of emissions data from laboratory-scale experiments;
Use of an emission model with site-specific and process-specific input
data;
Extrapolation of emission rate measurements made during pilot-scale
operation at the site of interest;
Emission rate measurements during full-scale operation of the process
unit of interest at a similar site; and
Emission rate measurements during full-scale operations at the site of
interest.
Field measurement data are generally preferable to model estimates. The results from
models such as those ^presented in this document should be considered to have a large degree
of uncertainty unless confirmatory field measurement data are available.
2.2 GENERAL CONSIDERATIONS OF MODELS
Various types of models are described below along with a discussion of the
advantages of modeling, the limitations of modeling, and the need for calibration and
validation of models.
2-1
-------
2.2.1
General Types of Models
The term "model" implies a simplified or miniaturized copy of the real world.
Models may be conceptual, graphical, physical, or mathematical. For example:
Conceptual: The short-term emission rate from a landfill is
proportional to the concentration of the contaminant in
the waste.
Graphical:
Mathematical:
Physical:
ER
Cone.
ER = C * k
Pilot-scale landfill
The emission models presented in this document are all mathematical models. Mathematical
models can be divided into several categories according to their underlying bases: theoretical
(based on physical laws), mass balance, empirical, and heuristic. Models frequently are a
mix of one or more of these types.
Theoretical models are based on fundamental physical laws. For example, an
emission model for landfills may be based on Pick's 2nd Law of Diffusion, or an emission
model for surface impoundments may be based on Henry's Law describing the equilibrium
partitioning between the vapor and liquid concentrations of a given compound. Emission
models that use a mass balance approach are based on the conservation of mass and are
therefore a subset of models based on physical laws. Empirically-based models estimate
emissions based on relationships or factors developed from field measurement data. In their
simplest form, these types of models are emission factors, e.g. mass of emissions per unit
time or unit operation. Heuristic models may be useful but can not be supported by either
theory or empirical bases. This type of model is typically derived from intuition or
generalized field experience. In their simplest form, heuristic models are "rules-of-thumb".
2-2
-------
A further distinction can be made between models based on whether they
describe steady-state or unsteady-state (transient) conditions. Steady state implies that under
certain circumstances, variables will reach some characteristic equilibrium or constant value.
For example, vapors migrating through the soil will adsorb onto soil particles or dissolve into
any liquid that is present until equilibrium is achieved and all the active sites in the soil and
associated liquid are saturated with the vapors. Until that point is reached, diffusion is at an
unsteady state. Once all the active sites are saturated, they can typically be ignored and
diffusion is considered to be at steady state conditions. In practice, most environmental
problems are modeled under steady-state conditions since this greatly simplifies the number
of variables to be considered and the chemical and physical processes to be modeled. In
reality, however, many environmental problems exist at unsteady-state conditions, so the
assumption of steady state introduces some bias.
2.2.2 Advantages of modeling
The advantages of a modeling approach as opposed to other options are speed
and convenience. Emission estimates can be made in a matter of minutes using a
mathematical model, whereas it may take several months to obtain field measurement data.
The cost differential between modeling approaches and field measurement approaches is
obvious. Most importantly, emission models can be used as a planning tool to make
estimates for remedial actions that have not yet taken place. In such cases, field
measurement data are not an option. Models may also be useful as a predictive tool for
estimating the effect of changes in waste properties or changes in operating parameters for an
on-going remedial action. In such cases, if field measurement data are available for one or
more set of conditions, then the accuracy of the model can be determined and empirical
correction factors added if necessary.
2-3
-------
2.2.3 Limitations of Models
The estimates provided by any model are limited by the validity of the input
data. For Superfund applications, this is a key concern since even fundamental information
such as the types of contaminants present at the site, their concentration, and their frequency
of occurrence and distribution is not usually known with much certainty. Another limitation
of the models is that they usually only consider a few key variables. This simplification of
reality introduces a variable amount of bias in the results. The greater the situation being
modeled differs from the assumptions inherent in the model, the greater the likely bias.
Models may be misused in several ways. One, an incorrect or inappropriate
model may be selected for use for a given situation. Two, incorrect input values may be
used. Three, the user may fail to utilize existing site-specific or process-specific data. Four,
the user may fail to recognize or consider the limitation of the model.
2.2.4 Calibration and Validation of Models
Model results should be compared, whenever possible, to real-world data.
Calibration of a model involves using these comparisons to adjust model parameters or to
add a correction factor. Model calibration is typically either specific for a given site or for a
given process, and the calibration results may not be applicable to other situations.
Validation of a model involves a systematic comparison of model results to field
measurements under various conditions as part of a sensitivity analysis of the model. The
degree to which a given model has been validated should be determined by the model user
before the data are used for any air pathway assessment.
2-4
-------
this
2.3 GENERAL SOURCES OF INPUT DATA
There are essentially three categories of input data for the emissions models in
is document: site-specific information (e.g., contaminated area), process information (e.g.,
feed rate), and physical properties of contaminants (e.g., diffusivity in air). Not every model
requires inputs from all three categories.
Clearly, the site-specific inputs should come from field measurements. In
some cases, typical or default values are presented for cases when field measurement .data are
not available or are suspect. The use of default values will affect the accuracy of the
emission estimates. Values for process variables should come from field observations,
design documents, or vendors. Default values are presented for cases where valid data are
not available, but these too will affect the accuracy of the emission estimates. Physical
property data for 168 contaminants are included in Appendix A. In many cases (e.g., vapor
pressure), these data are only valid for 25°C and 1 atmosphere pressure.
2-5
-------
SECTION 3
VOC POINT SOURCES
Simple air emission estimation procedures are presented in this section for point
sources of volatile organic compounds (VOCs), including: air strippers, soil vapor extraction
systems, thermal desorption units, and thermal destruction units (incinerators). The same
format is followed for each source. A brief description of the emission process is given,
followed by a discussion of available air emission models. The model selected for inclusion
in this manual is then presented along with sources of input data and default values for each
of the input variables of the selected model. The model assumptions are then briefly
discussed. Finally, an example calculation is shown and references are listed. In all cases,
the models estimate uncontrolled VOC emissions.
3.1 AIR STRIPPERS
3.1.1 Description of Emission Process
Air stripping, or packed-tower air stripping, is widely used to remove chlorinated
solvents and other VOCs from contaminated ground water. Air stripping is currently in use
or is the proposed remedy for Superfund sites in all ten U.S. EPA regions. It is often
chosen because of its cost-effectiveness and the high removal efficiencies that can be
achieved.
Air stripping is a mass transfer process in which volatile contaminants are evaporated
(stripped) into air. The contaminated water is introduced at the top of a packed-tower
through spray nozzles and allowed to slowly flow down through the column or tower. The
packing media acts to retard the water flow (increase liquid hold-up) and increase the
effective surface area of the system. Air is introduced countercurrent to the direction of
water flow. The saturated air containing the VOCs is emitted from the top of the column or
routed to a control device. The treatment system may also contain wells, separators, and
vessels for treating inorganic contaminants.
3-1
-------
A site-specific air stripping system is typically designed and constructed for each
Superfund application as opposed to using an existing design. The system design is based
upon meeting a specific performance goal: e.g., some minimum percent removal of a given
VOC from water at a specified flow rate. The level of uncontrolled air emissions from a
given system will thus depend on the performance goal and the effectiveness of the design.
The primary source of emissions from air stripping is the stripper exhaust, and VOCs
are the major pollutant of concern. For systems without control devices, the exhaust is
vented through a short stack, typically a (3-6 ft) pipe, at the top of the column. For systems
with control devices, the airflow from the column is usually vented down to the control
device at ground level. A short stack (15-20 ft) is then used after the control device.
In addition to the exhaust stack, other emission sources may exist. Any place
upstream of the air stripping tower where water is in direct contact with the atmosphere,
such as separators, holding tanks, treatment tanks, or conduits, is an emission source.
Fugitive losses from pumps, valves, and flanges are usually not significant due to the dilute
nature of the water contamination.
The important parameters affecting the emission rate for a given compound from an
air stripping unit include: the concentration of the contaminant in the influent to the stripper,
the influent flowrate, the stripping efficiency of the tower, and the effectiveness of any
control technologies that are in place. The stripping efficiency will depend on a number of
factors including: the compound's volatility and water solubility, the type of packing material
in the tower, and the gas and liquid flow rates within the tower.
3-2
-------
3.1.2 Model Selection
For a given liquid treatment rate, the magnitude of the uncontrolled air emissions
from an air stripper are governed by the effectiveness of the liquid-to-air mass transfer in the
stripper. A number of equations and associated computer models are available to aid the
system designer in selecting the appropriate tower height, gas to liquid ratio, packing
material, etc. to optimize the mass transfer and meet the performance goal in a cost-effective
manner1'7. These estimation procedures tend to be similar, and any of these design models
could be used to predict the levels of uncontrolled air emissions. One design model for
Superfund sites that is given in EPA's Air Stripper Design Manual1 has been validated8 by
comparing the model outputs to data from multiple field sites. Air emissions are estimated
from the influent mass loading and the Henry's Law constant of the compounds. This
approach also served as the basis for the simple screening model developed by Eklund, et al.9
that uses a mass balance approach and presents typical (default) inputs. This screening
model was chosen for inclusion in this document because of its usefulness, simplicity, and
connection to a field-validated model.
3.1.3 Emission Model Equation
A simple mass balance equation is given below for estimating the uncontrolled
emission rate for VOCs from air strippers. It is assumed that all contaminants removed enter
the atmosphere.
ER = C L (SE/100)(1.67 x lO'5) (Eq. 3-1)
where: ER = emission rate of species i [g/sec];
C = concentration of species i in influent water [mg/L];
L = Influent flow rate [L/min];
SE = stripping efficiency [%]; and
1.67 x 10'5 = conversion factor from [mg/min] to [g/sec].
3-3
-------
3.1.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS) to provide the input
data necessary to generate reasonably accurate estimates of air emissions. The minimum
field data required are:
Specific contaminants present in the water to be treated;
Average contaminant concentration in the water; and
Maximum contaminant concentration in the water.
3.1.5 Sources of Input Data
The preferred source of input data for Equation 3-1 is field measurements for the air
stripping system of interest. As previously mentioned, field data should be obtained
regarding the specific contaminants present in the water to be treated and their average and
maximum concentration in the water. Influent water concentrations will generally be lower
than the static water concentrations obtained from monitoring wells. The approximate total
volume of water to be treated will be of interest to estimate the duration of the cleanup.
Values for the influent flow rate and stripping efficiency may be obtained from design
specification documents and blueprints or from field measurements. Once the air stripper is
in operation, a mass balance or stack sampling of emissions from the system can be
performed to confirm the emission estimates. Any field measurements should be performed
under steady-state conditions and during typical or average operating conditions. Worst-case,
or reasonable maximum, values may also be of interest for assessing maximum, short-term
air impacts.
Default values are given in the following subsection for all input variables necessary
for Equation 3-1. The default values are intended to be used only if adequate site-specific
and system-specific data are not available.
3-4
-------
3.1.6 Default Values for Input Variables
Table 3-1 gives default values to be used in Equation 3-1 for a typical air stripper
used at a Superfund site along with a default value for the gas/liquid ratio in the stripping
tower. Figures 3-1 and 3-2 give a means of estimating the stripping efficiency of a
compound based on its Henry's law constant and the gas/liquid (G/L) ratio. Henry's Law
constants and their logs for 168 compounds are given in Appendix A. As previously
discussed, the efficiency will also be a function of this G/L ratio as well as temperature,
tower height, air/water contact time, and pollutant concentration. A worst-case scenario if
the influent contaminant concentrations are not well known is to assume that slightly soluble
organic contaminants are present in the water at their maximum solubility (see Appendix A
for solubilities of some common VOCs).
If the approximate size of the air stripper is known, the values given in Table 3-2 can
be used to estimate emissions using Equation 3-1. Table 3-2 also includes information
regarding stack parameters to allow prediction of downwind ambient air concentrations using
an EPA-approved air dispersion model.
3.1.7 Model Assumptions/Sensitivity Analysis
Equation 3-1 is valid under certain conditions. Most importantly, it is a steady-state
equation; that is, if used to predict emissions over a period of days or weeks, it requires that
the concentration of pollutant "i" in the water stays a constant, and that the stripping
efficiency is also constant. To account for the change in the contaminant concentration over
time, it would be necessary to monitor its concentration in the process water on a periodic
basis and use these data in the emissions estimation.
3-5
-------
Table 3-1.
Default Values for Estimating Air Stripper Emissions
Parameter
Concentration of species i
in influent water
Influent flow rate
Stripping Efficiency
Gas to Liquid Ratio*
Symbol
C
L
SE
G/L
Units
mg/L
L/min
%
L/L
Default Value
maximum water solubility of
species i (see Appendix A)
5,700
100
50
Expected Range
570 - 5,700
90- 100
20-200
Reference
~
9
b
10
"For use with Figures 3-1 and 3-2.
"Expected range of design criteria for VOC removal.
3-6
-------
100
90
80
2 70
g 60
O
£ 50
I«
ex
£ 30
w ^
20
10
0
G/L-50
G/L>100
G/L-200
G/L-400
-5 -4 -3 -2 -1
Log (Henry's Law Constant, atm-nf/gmole)
Figure 3-1. Stripper efficiency vs. Henry's Law constant, parameter = G/L (vol/vol),
low efficiency range.
39.99
G/L.10
G/L »20
^^ it
GA.^100
G/L-200
GA.-400
Log (Henry's Law Constant, atm-nf/gmok)
Figure 3-2. Stripper efficiency vs. Henry's Law constant, parameter = G/L (vol/vol),
high efficiency range.
Source: Reference 1
3-7
-------
Table 3-2.
Example Scenarios for Air Stripping
Parameter
Total influent liquid flowrate
Column Height
Column Diameter
Exhaust Gas Flowrate
Stack Height
Stack Diameter
Structure Dimensions
Exit Gas Velocity
Exit Gas Temperature
Ambient Temperature
Air/Liquid Ratio (G/L)
Units
L/min
gpm
m
m
nf/rnin
cfrn
m
m
m
m/sec
°C
°C
vol/vol
Typical Value
Small
570
150
7.6
1.2
29
1,020
8.5
0.31
7.6 x 1.2 x 1.2
6.4
20
20
50
Medium
2,840
750
9
3.6
140
5,000
10
0.61
9.0 x 3.6 x 3.6
8.0
20
20
50
; Large
5,700
1,500
14
3.6
285
10,000
15
0.91
13.0 x 3.6 x 3.6
7.3
20
20
50
SOURCE: Reference 9
3-8
-------
The biggest uncertainty in estimating air stripper emissions is in the values for the
contaminant concentration in the influent water. The water composition is likely to
change over time due to variability in the extent and degree of subsurface contamination
across the site. Before operation of the air stripping system is initiated, only a limited
set of data taken from monitoring well samples will typically be available for estimating
an average contaminant concentration in the influent water. Furthermore, the
measurement process itself for various VOCs in water can bias the data. Substantial
losses of VOCs can occur during sampling and in the first few days of sample storage.
All these factors combine to increase the uncertainty in the Q term in Equation 3-1.
A recent survey of air strippers10 found that systems vary widely in their design
and capacity, and that the performance between units will also vary. For a given system,
the mass of VOCs stripped may vary widely over a period of several months, most likely
due to changes in the composition of the water to be treated. A slight seasonal variation
in performance was also found to be typical, most likely due to changes in air and water
temperatures and resultant changes in the stripping efficiency. Any error in assuming a
constant removal for Equation 3-1 is slight if a reasonable worst-case water concentration
value is used.
3.1.8 Example Calculation
A contaminated aquifer to be remediated contained:
Compound
Trichloroethylene (TCE)
1,1,2-Trichloroethane (1,1,2-TCA)
Phenol
Concentration in Water (ppmw)
10
20
20
The air stripper chosen for the task had a water flow rate of 2200 gpm.
3-9
-------
To calculate the potential emissions from this project using Equation 3-1, a
stripper efficiency is needed. A conservative estimation may be reached using the
default value of 100%. However, it is more accurate to calculate the efficiency of each
compound using Appendix A and Figures 3-1 and 3-2. Use of these figures to estimate
efficiency requires a G/L value for the stripper. Since the G/L was not specified,
assume the G/L default value of 50.
From Appendix A, the logarithms of the Henry's Law constants are -3.13, -2.04,
and -6.34 for 1,1,2-TCA, TCE, and phenol, respectively. Figure 3-2 then gives stripper
efficiencies of about 90% for 1,1,2-TCA and 99.9% for TCE.
For phenol, Figure 3-1 indicates that the stripper efficiency would be near zero.
This is because phenol is relatively hydrophilic and essentially non-volatile, and one may
assume that no emissions will occur during the air stripping. If the concentration in the
discharge water exceeds applicable regulations, then some other means for its control
would need to be considered.
All that remains before using Equation 3-1 is the conversion of units from gpm to
L/min and from ppm or ppb to g/L. Using 1 ppm = 1 mg/L for dilutely contaminated
water:
10 ppm TCE = 10 mg/L TCE
20 ppm TCA = 20 mg/L TCA
Making use of the fact that 1 gal = 3.7854 L, the following emissions are found via
Equation 3-1:
= 10 x 2200 x 3.7854 x (99.9/100) x 1.67 x ia5 = 1.4 g/sec
ERTCA = 20 x 2200 x 3.7854 x (90/100) x 1.67 x 10s = 2.5 g/sec.
3-10
-------
3.1.9 References
1. U.S. EPA. Air/Superfund National Technical Guidance Study Series: Air
Stripper Design Manual. EPA-450/1-90-003. May 1990.
2. Hand, D.W., J.C. Crittenden, J.L. Gehin, and B.W. Lykins. Design and
Evaluation of an Air-Stripping Tower for Removing VOCs From
Groundwater. J. of AWWA, pp87-97, September 1986.
3. Speece, R.E., N. Mirmalakhandan, and Y.H. Lee. Nomograph For Air
Stripping of VOC From Water. J. of Env. Engineering, Vol. 113, No. 2,
pp434-443, April 1987.
4. Amy, G.L. and WJ. Cooper. Air Stripping of Volatile Organic
Compounds Using Structured Media. J. of Env. Engineering, Vol. 112, No.
4, pp730-743, August 1986.
5. Cummins, M.D., J.J. Westrick. Feasibility of Air Stripping for Controlling
Moderately Volatile Synthetic Organic Chemicals. U.S. EPA, Office of
Drinking Water, Cincinnati, Ohio, 1986.
6. Cummins, M.D. and J.J. Westrick. Packed Column Air Stripping
Preliminary Design Procedure. Presented at 1986 WPCF Conf. on
Hazardous Waster, Los Angeles, October 9-10, 1986.
7. Wood, D.F., L.L. Locicero, K.T. Valsaraj, D.P. Harrison, and LJ.
Thibodeaux. Air Stripping of Volatile Hydrophobic Compounds Using
Packed Crisscross Flow Cascades. Env. Progress Vol. 9, No. 1, pp24-29,
February 1990.
8. U.S. EPA Air/Superfund National Technical Guidance Study Series:
Comparisons of Air Stripper Simulations and Field Data. EPA-450/1-90-
002. March 1990.
9. Eklund, B., S. Smith, and M. Hunt. Estimation of Air Impacts For Air
Stripping of Contaminated Water. EPA-450/1-91-002, August 1991.
10. Vancit, M., R. Howie, D. Herndon, and S. Shareef. Air Stripping of
Contaminated Water Sources- Air Emissions and Controls. EPA-450/3-87-
017, August 1987.
3-11
-------
32 SOIL VAPOR EXTRACTION
32.1 Description of Emission Process
Soil vapor extraction (SVE) is frequently used for the treatment of soil
contaminated with volatile hydrocarbons. The process is also referred to as soil venting,
vacuum extraction, aeration, or in-situ volatilization. In general terms, soil vapor
extraction removes volatile organic constituents from contaminated soil by creating
sufficient subsurface air flow to strip contaminants from the vadose (unsaturated) zone
by volatilization. As the contaminant vapors are removed, they may be vented directly to
the atmosphere or controlled in a number of ways. Among the relative advantages of
SVE over other remediation approaches is that the air emissions are released from a
point source and thus can readily be controlled.
Soil vapor extraction has been widely used to remediate sites contaminated with
gasoline or chlorinated solvents (e.g. TCE). It is also sometimes used to minimize
migration of vapors into structures or residential areas during other types of remediation.
By its nature, SVE is an on-site, in-situ treatment method. It is often used in conjunction
with or following other remedial measures such as excavation of subsurface waste bodies,
removal (pumping) of any hydrocarbon lens that is present, or air stripping of
contaminated ground water.
Typical SVE systems include extraction wells, monitoring wells, air inlet wells,
vacuum pumps, vapor treatment devices, vapor/liquid separators and liquid phase
treatment devices (if contaminated water is extracted in the process). Various design
and operating options may be employed such as sparging, steam/heat injection, and
pulsed operation. A site-specific SVE system is typically designed and constructed for
each Superfund application as opposed to using an existing design. The system design is
based upon meeting a specific performance goal: e.g., some minimum rate of VOC
removal. The level of uncontrolled air emissions from a given system will thus depend
on the performance goal and the effectiveness of the design.
3-12
-------
The primary source of emissions from SVE systems is the exhaust gas stack, and
VOCs are the major pollutant of concern. Stack heights are typically 12-30 feet and
usually only one stack is used. Additional releases of volatile organics may occur from
any entrained contaminated water that is extracted with the vapor. Entrained water is
typically collected in a knock-out chamber or drum. Fugitive emissions are considered
negligible due to the negative pressure throughout most of the system.
The contaminants removed from the soil by SVE systems and hence present in
the off-gas generally have vapor pressures greater than 1.0 mm Hg at 20° C. The
tendency of the organic contaminants to partition into water or to be adsorbed onto soil
particles also affects the off-gas composition, as do the compound's water solubility,
Henry's Law constant, and soil sorption coefficient. The soil temperature affects each of
these variables and hence the rate of vapor diffusion and transport. Bulk soil
temperatures are typically constant unless steam or large volumes of heated make-up air
are introduced into the soil. The concentration of contaminants that are initially present
will also affect their relative partitioning between vapor and liquid phases, and the
amount that is solubilized or adsorbed. The time that the contamination has been
present is also an important factor, as mixtures of contaminants will generally become
depleted of their more volatile components over time through volatilization. This
process, referred to as weathering, will tend to cause SVE to become progressively less
applicable as the site ages. It also affects the operation of the SVE system, as the more
volatile components are typically removed first and the composition of the vapors
collected and treated varies over time.
The emission rate of VOC compounds over time from continuously operated SVE
systems tends to show an exponential-type decay curve. If the system is stopped and
then restarted, however, the VOC emission rate returns to near the original rate unless
the remediation is nearing completion. Apparently, shutting off the vacuum allows the
soil-gas equilibrium to become re-established. Due to this behavior, the most efficient
method of operation is to run the SVE system only for a part of each day or week, i.e.
operate in a "pulsed" mode.
3-13
-------
Bioventing can be considered a subset of soil vapor extraction. In bioventing,
however, the goal is to provide oxygen to subsurface microorganisms and thereby
optimize conditions for biodegradation rather than to physically transfer pollutants from
soil pore spaces to the atmosphere. Air is withdrawn from the soil and oxygen-rich
ambient air is introduced via air inlet wells. The exhaust gas from a bioventing system
will be less concentrated than SVE exhaust gas, and the total flowrate may be only 10-
20% of a typical vapor extraction rate for a SVE system.
122 Model Selection
While attempts have been made to model SVE systems as vapor transport
through a porous media1"3, the authors have generally conceded that accurate prediction
is not possible due to the complex nature of subsurface vapor transport and the large
variations in soil permeability to air flow across most sites. Therefore, the SVE system
design is typically based on pilot-scale feasibility tests at the site or other empirical
determinations of the flow rate of vapors (i.e., air permeability) through the soil1-4'5.
Various equations to assist in the design of SVE systems have been presented*
and incorporated into a software package called Hyperventilate. A remediation
company, VAPEX, has published similar equations6 and incorporated them into their 3-
D AIR model. For estimates from either of these models to be meaningful, however,
detailed information about the air permeability of the soil at the site must be known.
A simple screening model has been developed by Eklund, et al.7 based on
historical vapor extraction rates at sites where SVE systems have been used. The
screening model document encourages the user to provide site-specific extraction rate
and vapor concentration data, but also provides conservative default values. This model
was selected for inclusion in this manual (a similar screening model document is
currently being prepared for bioventing).
3-14
-------
3.2.3 Emission Model Equation
There are several approaches for estimating the emissions from a SVE system.
The best method is to directly measure the emissions from the site during full-scale or
pilot operation. This approach would not entail the use of an emission model equation.
A simple check of the total emissions potential for the site may be performed using
Equation 3-2, which divides the total mass of contaminants in the soil by the expected
duration of the site remediation:
Average Emission Potential
ERAVG = M/t = $, C (1) 0/t (Eq. 3-2)
where: E^AVG = average emission rate of species i [g/sec];
M = total mass of contaminant in soil [g];
Sy = volume of contaminated soil to be treated [m3];
C = average contaminant concentration of species i in soil [/xg/g];
18 = bulk density of soil [g/cm3];
1 = constant [g/ltf/zg x lOfcm3 /m3]; and
t = duration of remediation [sec].
Because Equation 3-2 assumes a 100% recovery of VOCs at a uniform rate throughout
the remediation process, it should be used only as a gross estimate of the average
emission rate from the SVE system over reasonably long time periods (e.g., weeks or
months). It may serve as a preliminary check of the site's emission potential if either
direct field measurements or a predictive model such as Equation 3-3 are to be used.
To estimate emissions over relatively short time periods or to estimate maximum
emission rates, Equation 3-3 should be used. Equation 3-3 is a mass balance model for
SVE emissions that requires site-specific inputs:
3-15
-------
Emission Rate
ER = q (ia6) Q/60 (Eq. 3-3)
where: ER = emission rate [g/sec];
Q = concentration of pollutant in extracted vapor j/xg/nf ];
Q = vapor extraction rate [nf/min];
106 = conversion factor [g/^g]; and
1/60 = conversion factor [min/sec].
If these inputs are not available, a conservative but less accurate estimation may be
made using the default values given below.
3.2.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS)to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
minimum field data required are:
Specific contaminants present in the soil to be treated; and
Average contaminant concentration in the soil.
To use Equation 3-2, field data are necessary also to estimate the total volume of
contaminated soil to be treated.
3.2.5 Sources of Input Data
Although provided as a very simple screening model, Equation 3-2 nonetheless
requires some knowledge of site characteristics: the volume of contaminated soil to be
treated, Sy, the contaminant concentration, Q, and the duration of the remediation
project, t.
3-16
-------
Pilot-scale field tests are typically performed to assist in SVE system design. If no
site data are available for the extraction rate, Q, an approximation may be made from
soil-air permeability measurements, if these are available. In the absence of any field
data, default values found in the next subsection may be used.
Measuring the concentrations in headspace vapors above the contaminated soil is
the next best approach to estimating C^. If these data are not available, a very
conservative value may be reached with the use of Equation 3-4, which assumes the soil
is saturated with the contaminant:
q, = P MW (Itf) / (R T) (Eq. 3-4)
where: Q = saturated value of contaminant vapor concentration l/ig/m3];
P = contaminant vapor pressure at soil temperature [mm Hg];
MW = molecular weight of contaminant [g/g-mole];
l(f = conversion factor l/ig-L/g-m3];
R = ideal gas constant = 62.4 [L-mm Hg/ mole-°Kj; and
T = absolute temperature of soil [°K].
Values of q, at 25° C (298° K), Pvap, and MW for various VOCs may be found in
Appendix A. The use of Equation 3-4 may result in a significant overestimation of the
emission rate (see Section 3.2.7).
Once the SVE system is in operation, stack sampling of emissions from the system
can be performed to confirm the emission estimates. The guidance given in Section 3.1.4
for field measurements of air strippers applies to all remediation technologies, including
SVE systems.
3.2.6 Default Values for Input Variables
Table 3-3 gives default values to be used in Equation 3-3 for a typical soil vapor
extraction system used at a Superfund site. A worst-case scenario if the soil-gas
concentration is not known is to assume that the soil-gas is saturated with the VOC of
interest as shown in the preceding subsection.
3-17
-------
If the approximate size of the SVE system is known, the values given in Table 3-4
can be used to estimate emissions using Equation 3-3. Table 3-4 also includes
information regarding stack parameters to allow prediction of downwind ambient air
concentrations using an EPA-approved air dispersion model.
Model Assumptions/Sensitivity Analysis
The largest uncertainty in estimating SVE emissions is in the values for the
concentration of the pollutant in the extracted vapors. The vapor composition may
change rapidly over relatively short periods of time. As previously noted, the
composition and concentration of vapors will depend on whether or not the operation of
the SVE system is continuous in nature or is intermittent (pulsed).
Before operation of the SVE system is initiated, only a limited set of pilot scale
data will typically be available for estimating achievable extraction rates as well as for
estimating soil-gas concentrations. The representativeness of this pilot-scale data will be
difficult to assess until full-scale operations are underway.
As mentioned above, Equation 3-2 is intended for estimating total emissions over
the course of the clean-up. It will be less accurate than Equation 3-3 for estimating
short-term emission rates. This is because the former assumes a 100% removal
efficiency and a constant removal rate, whereas this removal efficiency is not achieved in
practice, and further, the removal rate drops over time. This drop may be less important
if the operation is "pulsed", allowing soil-gas concentration to be periodically re-
established at levels near that of the initial concentration.
3-18
-------
Table 3-3.
Default Values for Estimating SVE System Emissions
Parameter
Concentration of
pollutant in extracted
vapor
Vapor extraction rate
Bulk density of soil
Duration of
remediation
Symbol
<;
Q
P
t
Units
Mg/m3
nf/min
g/cm3
sec
Default Value
Maximum: saturated vapor cone.
(see Appendix A)
Typical: 3 x itf
85
1.5
1.58 x 1CF
(6 months)
Expected Range
1.4 - 425
1.0 - 2.0
0.79X107 - 9.46X107
Reference
-
81
7
9
Author's
estimate
"Based on case studies of SVE that indicate that the exhaust gas concentration is typically less than 1,000
ppm per given VOC compound.
Table 3-4.
Example Scenarios for SVE with No Controls Based on Size of System
Parameter
Exhaust gas flowrate
Exhaust gas velocity
Exit gas temperature
Stack height
Stack diameter
Units
nfmin
cfm
m/sec
°C
m
m
Scenario ; M
Very Small
1.4
50
3.0
50
3.0
0.10
Small ;
14
500
7.4
50
4.6
0.20
Medium
85
3,000
12.5
50
7.6
0.38
; Large
425
15,000
14.2s
50
9.1
0.46
'Assume three adjacent stacks each handling 5,000 cfm. The flow is split to lower the
velocity of the exiting gas to typical design levels to minimize damage to the stack.
Source: Reference 7
3-19
-------
It is important to note that Equation 3-4 gives the theoretical maximum value of
Q, and will significantly overpredict the concentration for any contaminants present at
relatively low levels. Equation 3-4 will also overpredict the long-term average value of
Cj,, as the concentration of contaminants in the soil tend to decrease over time. This
decrease is a function of the soil type, and in general cannot be accurately modeled,
although in principle the decrease is exponential.
3.2.8 Example Calculation
A contaminated site to be remediated contains soil contaminated to the following
extent:
Benzene:
Toluene:
Carbon Tetrachloride:
Naphthalene:
100 ppm
300 ppm
50ppb
800 ppb
(100Mg/g)
(300/xg/g)
(0.050 ^g/g)
(0.800 ^g/g)
The site is a 200 nf field behind a factory. The water table is 30 m below the surface at
this location. The entire volume of soil down to the water table is assumed to be
contaminated. A vendor has quoted an estimate of five months to complete the clean-
up. No physical data on the type of soil is known.
Using Equation 3-2, a rough estimate of the long-term average emission rate may
-be obtained, assuming continual operation for 150 days (1.3 x 107 seconds), and.using the
default bulk density of 1.5 g/cm3. The uncontrolled emissions are:
= (200 x 30) x (100) x (1.5) / (1.3 x id7) = 0.069 g/sec
= (200 x 30) x (300) x (1.5) / (1.3 x iff) = 0.21 g/sec
= (200 x 30) x (0.05) x (1.5) / (1.3 x 107) = 3.5 x 105 g/sec
= (200 x 30) x (0.8) x (1.5) / (1.3 x 107) = 5.5 x 104 g/sec.
3-20
-------
The short-term emission rates are calculated with Equation 3-3. Use of this
equation requires knowledge of the vapor extraction rate; for this scenario, a medium-
sized SVE system of 85 nf/min may be assumed. One further needs the concentration
of the extracted vapors. The saturated vapor concentrations obtained from Appendix A
(in ug/nf) are:
benzene
toluene
carbon tetrachloride
naphthalene
4.00 x 10s
1.49 x 10s
9.34 x 10s
1.58 x 10?
These values all assume that the soil is saturated with each contaminant. Given the low
concentrations present in the soil, the extracted vapor will actually be well below
saturation. This can be checked by using the partial pressure, P; (see Equation 4-10), in
place of P in Equation 3-4. Therefore, the "typical" default value for VOCs of 3x10*
ug/nf for the exhaust gas concentration will be used for each compound except
naphthalene (where the saturated value is actually lower since naphthalene is a semi-
volatile compound rather than a VOC).
Putting these values into Equation 3-3 yields:
= (Sxltf) x (10*) x (85)/(60) = 4.2 g/sec
= (3xltf) x (10*) x (85)/(60) = 4.2 g/sec
= (3x10) x (10*) x (85)/(60) = 4.2 g/sec
= (1.58x10?) x (10*) x (85)/(60) = 022 g/sec.
In this case, the two sets of emission rates vary by at least an order of magnitude. If the
air pathway assessment using the higher values from Equation 3-3 indicates that the air
emissions are at unacceptable levels, then better estimates of the extracted vapor
concentration should be obtained via field measurements, so that more accurate emission
rate estimates can be determined.
3-21
-------
3.2.9 References
1. Pedersen, T.A. and J.T. Curtis. Handbook of Soil Vapor Extraction (SVE).
EPA/540/2-91/003. 1991.
2. Stephanatos, B.N. Modeling the Soil Venting Process for the Cleanup of
Soils Containing Volatile Organics. In: Proceedings of 4th National
Outdoor Action Conference, Las Vegas, NWWA, Dublin, Ohio. pp619-632.
1990.
3. Walton, J.C., R.G. Baca, J.B. Sisson, and T.R. Wood. Application of Soil
Venting at a Large Scale: A Data and Modeling Analysis. In: Proceedings
of 4th National Outdoor Action Conference, Las Vegas, NWWA, Dublin,
Ohio. pp633-650. 1990.
4. Johnson, P.C., M.W. Kemblowski, J.D. Colthart, D.L. Byers, and C.C.
Stanley. A Practical Approach to the Design, Operation, and Monitoring
of In Situ Soil-Venting Systems. Ground Water Monitoring Review.
Spring 1990.
5. Sellers, K.L., T.A. Pedersen, and C.Y. Fan. Soil Vapor Extraction: Air
Permeability Testing and Estimation Methods. In: Proceedings of the 17th
Annual Hazardous Waste Research Symposium, EPA/600/9-91/002, pp28-
42. April. 1991.
6. Marley, M.C., S.D. Richter, R.J. Cody, and B.L. Cliff. Modeling For In-
Situ Evaluation of Soil Properties and Engineered Vapor Extraction System
Design. In: Proceedings of the NWWA/API Conference on Petroleum
Hydrocarbons and Organic Chemicals in Ground Water, Houston,
November 1990.
7. Eklund, B., S. Smith, P. Thompson, and S. Malik. Estimation of Air
Impacts for Soil Vapor Extraction Systems. EPA 450/1-92-001. January
1992.
8. Pacific Environmental Services. Soil Vapor Extraction VOC Control
Technology Assessment. EPA-450-4-89-017. September 1989.
9. Schultz, H.L., et al. Superfund Exposure Assessment Manual.
EPA/540/1-88/001. April 1988.
3-22
-------
3.3 THERMAL INCINERATION
3.3.1 Description of Emission Process
Thermal incineration, also known as thermal destruction, high-temperature
thermal treatment, thermal oxidation, or incineration, is a commonly proposed remedy
for Superfund sites. Its primary advantage is that it results in the destruction of toxic
organic compounds. Add-on control devices for VOCs are not necessary.
Several types of incinerators are now in use at Superfund sites, but rotary-kiln
designs are the most common. The emission estimation methods presented in this
section are valid for any design, but the default values given are only valid for rotary-kiln
incinerators.
Incineration technology is primarily used for the remediation of organic
compounds. The emissions of organic compounds, therefore, depend on the destruction
and removal efficiency (DRE) of the incinerator. The DRE is a function of the
incinerator's operating conditions (residence time, temperature, etc.) and can not be
accurately predicted, but it can be determined by a trial burn or treatability test.
The emission rates for organic compounds can be estimated by assuming that the
DREs exactly meet the regulated standards. For example, RCRA standards require a
DRE of 99.99 percent for each principal organic hazardous constituent (POHC) in the
waste. (POHCs are compounds that are chosen for the trial burn as an indication of the
DRE of total hydrocarbons). Other DREs can be assumed for dioxins, PCBs, etc.
Products of incomplete combustion (PIC) are formed by the reactions of organic
compounds in the intense heat of the combustion chamber. PICs such as dioxins, furans,
and other polynuclear aromatic hydrocarbons (PAH) will be emitted even if not present
in the waste stream. As yet, there is no known accurate method of predicting PIC
emissions.
3-23
-------
3.3.2 Model Selection
Models have been reported1 for direct-fired high temperature rotary kiln systems
that predict the temperature of the solid bed and the kiln exit gas as a function of
measurable physical parameters such as kiln rotational speed, burner firing rate, soil feed
rate, soil moisture content, and whether the operation is co-current or counter-current.
Ho and Ding2 have presented the results of a similar model (but not the equations
themselves) for an oxygen combustion system. These models could be combined with
thermal stability data3'4 to predict the destruction efficiency of incinerators. The oxygen
content of the kiln gas is also an important variable3 and would need to be considered.
Much work would be required before this type of model was ready for field validation.
Several computerized models are available for permit writers and incinerator
operators to use in predicting the performance of incineration systems5'6. These
computer programs have not been purchased or examined for this study. The marketing
literature for these programs, however, implies that these models use an assumed
destruction and removal efficiency (DRE) and can then be used to predict such factors
as fuel requirements, gas flows, stack temperature and velocity, pressure drops, etc.
These models do not appear to be appropriate for use as screening models given their
cost and requirement for access to a personal computer. They also do not appear to
offer improved accuracy over the simple mass balance equations discussed below.
Simple mass balance equations for estimating incineration emissions with an
assumed DRE have been previously developed by Eklund, et al.7. These same general
equations were slightly modified and used by IT in another EPA study to estimate the
air impacts from incinerators8. Both of these documents summarize typical operating
rates, control efficiencies, etc. The equations from Reference 8 meet the selection
criteria and are presented in Section 3.3.3.
3-24
-------
Air emissions from materials handling operations upstream of the incinerator
should also be evaluated. Emissions from excavation and storage piles are addressed in
Sections 4.1 and 5.3, respectively. No good models exist for waste mixing and waste feed
operations. Whenever possible, such operations should take place in an enclosure that is
vented to the incinerator or to another control device.
3.3.3 Emission Model Equation
Only emissions of organic compounds are addressed in this section; emissions of
paniculate matter and metals are covered in Section 6. Equation 3-5 provides the
emission rate estimation:
ER = 0.278 F (1 - ORE/ 100) (Eq. 3-5)
where: ER = emission rate of organic contaminant of interest [g/sec];
F = feed rate of organic contaminant of interest [kg/hr];
DRE = destruction and removal efficiency of organic contaminant of
interest [%]; and
0.278 = conversion factor [kg/hr] to [g/sec] .
The feed rate of organic may be found using Equation 3-6:
F = 106 FT C (Eq. 3-6)
where: FT = total feed rate of waste into incinerator [kg/hr];
106 = conversion factor [g/ug]; and
C = concentration of the contaminant of interest
The concentration of a group of organic compounds is found by summing the
concentrations of each compound within the group.
3.3.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS) to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
minimum field data required are:
3-25
-------
Specific contaminants present in the soil or waste to be treated;
Average contaminant concentration in the soil or waste; and
Maximum contaminant concentration in the soil or waste.
3.3.5 Sources of Input Data
The primary source of input data for Equations 3-5 and 3-6 are field
measurements and design specifications. Accurate values for the feed rate and
contaminant concentration should be available from the RI/FS and from any trial burns.
Otherwise, the incinerator vendors may be able to provide a typical feed rate for the
system, although this will depend upon waste type.
3.3.6 Default Values for Input Variables
Table 3-5 gives default values to be used in Equations 3-5 and 3-6 for a typical
incineration system used on-site at a Superfund site.
If the approximate capacity of the incinerator is known, the values given in Table
3-6 can be used to estimate emissions using Equation 3-5. Table 3-6 also includes
information regarding stack parameters to allow prediction of downwind ambient air
concentrations using an EPA-approved air dispersion model.
33.7 Model Assumptions/Sensitivity Analysis
Equations 3-5 and 3-6 assume a uniform feed rate and a uniform concentration of
contaminant in the waste. If the waste is homogenous and if the concentrations of
contaminants do not vary much, then the first assumption will be valid. However, if the
concentrations do vary significantly, then the feed rate may need to be altered to
maintain the same DREs, depending on the system. Generally, if the concentrations
vary, then operating conditions (residence time, excess oxygen, etc.) are altered to ensure
the minimum DRE standards are met.
3-26
-------
3.3.9 References
1. Troxler, W.L., J.J. Cudahy, and R. P. Zink. Guidance Document for the
Application of Low Temperature Thermal Desorption for Treating
Petroleum Contaminated Soils. EPA Draft Report to James Yezzi, U.S.
EPA, Edison, NJ. January 1992.
2. Ho, M. and M.G. Ding. Field Testing and Computer Modeling of an
Oxygen Combustion System at the EPA Mobile Incinerator. In:
Proceedings of the 13th Nat. Waste Processing Conf. and Exhibit, May 1-4,
1988. Supplement, pp91-102.
3. Taylor, P.H., B. Dellinger, and C.C. Lee. Development of a Thermal
Stability Based Ranking of Hazardous Organic Compound Incinerability.
ES&T, Vol. 24, No. 3, pp316-328, 1990.
4. Mournighan, R.E., M.K. Richards, and H.O. Wall. Incinerability Ranking
of Hazardous Organic Compounds. In: Proceedings of the 15th Annual
Hazardous Waste Research Symposium, EPA/600/9-90/006, pp32-42, Feb
1990.
5. Incinerator Consultants Incorporated. Incinerator System
Design/Operation, Heat Balance Program Package (Cal Brunner model) -
Brochure. 1989.
6. Rowe Research & Engineering Associates, Inc. TOXIC Software for
Incinerator Risk Assessment - Brochure. 1986.
7. U.S. EPA. Air/Superfund National Technical Guidance Study Series,
Volume III: Estimation of Air Emissions from Cleanup Activities at
Superfund Sites. Report No. EPA-450/1-89-003. NTTS PB89 180061/AS.
January 1989.
8. IT Corp. Screening Procedures For Estimating the Air Impacts of
Incineration at Superfund Sites. EPA Contract No. 68-02-4466, WA 91-77.
February 1992.
9. Donnelly, J. Air Pollution Controls for Hazardous Waste Incinerators. In:
Proceedings of the 12th National Conference on Hazardous Materials
Control/Superfund 1991. HMCRI. 1991.
3-29
-------
3.4 THERMAL DESORPTION
3.4.1 Description of Emission Process
In the thermal desorption process, volatile and semi-volatile contaminants are
removed from soils, sediments, slurries, and filter cakes. This process typically operates
at temperatures of 200° -1000° F but is often referred to as low temperature thermal
desorption to differentiate it from incineration. At these lower temperatures, the process
promotes physical separation of the components rather than combustion.
In thermal desorption processes, contaminated soil is removed from the ground
and transferred to treatment units, making this an ex situ process. After it is excavated,
the waste material is screened to remove objects greater than 1.5" in diameter before
being introduced to the desorber. In general, three desorber designs are used: an
indirectly fired rotary dryer, internally heated screw augers, or a fluidized bed. The
treatment systems include both mobile process units designed specifically for treating soil
and asphalt kilns, which can be adapted to treat soils. Direct or indirect heat exchange
vaporizes the volatile compounds producing an off-gas that is typically treated before
being vented to the atmosphere.
Because thermal desorbers may operate near or above 1000° F, some pyrolysis and
oxidation may occur in addition to the vaporization of water and organic compounds.
Collection and control equipment such as afterburners, fabric filters, activated carbon, or
condensers prevent the release of the contaminants to the atmosphere.
Thermal desorbers effectively treat soils, sludges and filter cakes and remove
volatile and semi-volatile organic compounds. Some higher boiling point substances such
as polychlorinated biphenyls (PCB's) and dioxins may also be removed and thus be
present in the off-gas (usually associated with the particulate matter). VOC removal is
enhanced if the soil contains 10-15 percent moisture prior to treatment since water vapor
carries out some VOCs.
3-30
-------
Point sources of air emissions from thermal desorption vary widely with each
process. The stack of an afterburner vents combustion products, as does a fuel-fired
heating system if the combustion gases are not fed into the desorber. The fuel-fired
heating system typically operates with propane, natural gas or fuel oil. If controlled, the
stack will vent small concentrations of the original VOC contaminants, as well as
products of any chemical reactions that might occur from the control devices such as
baghouse, scrubber, and vapor phase carbon adsorber. Relative to incineration, the off-
gas volume to be treated from thermal desorption may be smaller, there is less
likelihood of creating dioxins and other oxidation products, and metals are less likely to
partition to the gas-phase. As with incineration, air emission control devices are always
part of the system design (the estimates of uncontrolled emissions obtained from this
manual can be used to help estimate the required removal efficiency of an emission
control system or the size and cost of a given control system).
Fugitive emissions from area sources may contribute significantly to the total air
emissions from a remediation site. Probably the largest source is excavation of the
contaminated soil. Other sources may include the classifier, feed conveyor, and the feed
hopper. Fugitive emissions from the components of the thermal desorption system and
controls are possible as well. Emissions may also emanate from the waste streams such
as exhaust gases from the heating system, treated soil, paniculate control dust, untreated
oil from the oil/water separator, spent carbon from a liquid- or vapor-phase carbon
adsorber, treated water, and scrubber sludge.
3.4.2 Model Selection
Thermal desorption units are currently being developed by numerous vendors.
The units each have characteristic operating ranges, and typically require pilot-scale tests
of the soil at a given site to find the optimum (i.e. most cost-effective) soil feedrate,
residence time, and operating temperature for each application. In some cases, a tiered
3-31
-------
evaluation involving laboratory screening, bench-scale, and pilot-scale tests may be
preferable before selecting thermal desorption as a remedy for a given site1.
Theoretical models have been proposed to predict the evolution of volatile
compounds from soil in a laboratory-scale indirectly heated rotary dryer system2. Particle
desorption and bed desorption were examined using partial differential equations based
on mass and energy balances and on the Freundlich isotherm equation. While the
equations showed good agreement with experimental data, they are not readily useful for
modeling other types of thermal desorption processes. In addition, the model has not
been tested for full-scale systems or with a range of soil and contaminant types.
A simple screening model has been developed by Eklund, et al.3 based on a mass
balance approach. The only inputs to the screening model are the concentration of
contaminants in the soil, the mass rate of soil being treated, and the percentage of the
contaminants that are volatilized. Default or typical values for the last two terms could
be developed from existing pilot-scale and full-scale test results (summarized in
Reference 3) in an analogous manner to what has been done for SVE systems. This
model is the only screening model identified for inclusion in this compendium.
The same considerations regarding air emissions from upstream materials
handling operations that are discussed in Section 3.3.2 for thermal incineration also apply
to thermal desorption systems.
3.43 Emission Model Equation
Equation 3-7 gives a mass-balance model of uncontrolled VOC emissions from a
desorption unit. This equation does not take into account emissions from excavation or
materials handling; those may be found in Section 4.1. Fugitive emissions from the
desorber system must be calculated on an ad hoc basis. It is assumed that no control
device is present, so any combustion gases or PICs from fume incinerators would require
further modeling.
3-32
-------
- I C
-
v
FT ) I V }
- -
16QO) (100)
(Eq- 3-7)
(lOQO) \16QO
where: ER = emission rate for contaminant of interest i [g/sec];
C = concentration of the contaminant of interest frtg/g];
1000 = conversion factor [g g//tg kg];
FT = total feed rate of waste into process unit [kg/hr];
3600 = conversion factor [sec/hr]; and
V = fraction of contaminant that is volatilized [%].
Note that the units of C are ^ig/g, which is equivalent to ppmw.
3.4.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS) to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
minimum field data required are:
Specific contaminants present in the soil or waste to be treated;
Average contaminant concentration in the soil or waste; and
Maximum contaminant concentration in the soil or waste.
3.4.5 Sources of Input Data
The preferred source of input data for Equation 3-7 is field measurements for the
thermal desorption system of interest. As indicated above, field data should be obtained
regarding the specific contaminants present in the material to be treated and their
average and maximum concentration. Values for the flow rate of material to the
desorber and the volatilization efficiency may be obtained from design specification
documents and blueprints or from field measurements. Once the thermal desorption
unit is in operation, stack sampling of emissions from the system can be performed to
confirm the emission estimates.
3-33
-------
3.4.6 Default Values for Input Variables
Table 3-7 gives default values to be used in Equation 3-7 for a typical thermal
desorption system used at a Superfund site. Tables 3-8 and 3-9 contain information
regarding stack parameters to assist in the prediction of downwind ambient air
concentrations using an EPA-approved air dispersion model.
3.4.7 Model Assumptions/Sensitivity Analysis
The emission estimation equation assumes the waste material is fed into the
process unit at a constant rate and that the material is uniformly contaminated. The
former assumption is reasonable, but the waste material will certainly have a degree of
variability in the contaminants present and their concentrations. The accuracy of the
emission estimate can be improved if the distribution of soil contaminants in the waste
material can be taken into account. For example, the site may contain several areas
where the contamination in each area is consistent within an order of magnitude. The
emissions associated with the clean-up of each area could be estimated separately to
improve the accuracy of the overall emission estimate.
The removal efficiency (i.e., percent volatilized) or RE of the thermal desorption
unit for various compounds will vary. The operating temperature and residence time of
the process unit will obviously affect the RE. In addition, the moisture content of the
waste material and the concentration range will also influence the fraction of a given
organic contaminant that is volatilized.
3.4.8 Example Calculations
A site to be remediated contains soil with the following levels of contamination:
Benzene
Toluene
Xylene
Ethyl Benzene
l.
-------
Table 3-7.
Default Values for Thermal Desorption Units
Parameter
Feed rate
% Volatilized
: Symbol
FT
V
Xlnits
kg/hr
%
Default Value
27,200
Desorber Temperature
of200-600°F
VOCs/BTEX 99
SVOCs/PNAs 90
THC 95
PCBs 50
Desorber Temperature
of600-1000°F
VOCs/BTEX 99.99
SVOCs/PNAs 99
THC 99.9
PCBs 99
Expected Range
2,700 - 90,800
Reference :
3
4
4
Other Parameters of Possible Interest"
Mass of soil to be treated
Residence Time
Kg
min
~
4.5X105 - 2.3xl07
3-70
3
3
These parameters may be used to find the treatment rate, F, using:
°f s0*! treated in kg)x60
(Residence time in minutes)
Table 3-8.
Example Scenarios for Rotary Dryers and Asphalt Aggregate Dryers
.->:' y Parameter
Feed rate (soils)
Gas Volume"
Stack Height
Stack Diameter
Exit Gas Velocity*
Exit Gas Temperatureb
Units
kg/hr
nWmin
cfm
m
m
m/sec
°C
;: . .. System . j;|f _.^
i Small
7,300
110
4,000
9.1
0.4
15
jMediuin
27,200
530
18,700
7,6
1.3
6.7
'Ipargeip
59,000
7,400
26,000
6.1
1.3
9.3
*Gas volume and exit velocity assume dry standard conditions at 7% Oz (20 °C, 1 atm).
bExhaust gas temperature is highly dependent on the types of control devices used. For thermal
oxidation with no off-gas cooling assume 815 °C (1500 °F). For any configuration.with off-gas
cooling by a heat exchanger, quench chamber, or scrubber, assume 120 °C (250 °F).
3-35
-------
Table 3-9.
Example Scenarios for Thermal Screws
Parameter
Feed Rate (Soils)
Gas Volume"
Stack Height
Stack Diameter
Exhaust Gas Velocity
Exhaust Gas Temperatureb
Units i
kg/hr
mVmin
scfm
m
m
m/sec
°C
System
Small
3200
3.7
130
6.7
0.2
2.0
21
. : Large ;. . . ;;'1 4
8200
24.8
875
4.6
0.2
13.2
21
*Gas volume and exit velocity assume dry standard conditions at 7% 02 (20 °C, 1 atm).
""Assumes off-gas treatment is condensation, which is typical for thermal screws.
3-36
-------
The full-scale desorption unit has a capacity of 7.5 tons per hour, and the percent volatilized
is 99.48 for benzene, 99.98 for the other compounds of interest.
To find the emission rate using Equation 3-7, the first thing to do is to get the input
values into the proper units. The mass treatment rate of 7.5 tons/hr = 6820 kg/hr. Thus:
= 1.0/1000 x 6820/3600 x 99.48/100 = 0.0019 g/sec;
= 24/1000 x 6820/3600 x 99.98/100 = 0.045 g/sec;
yl = 110/1000 x 6820/3600 x 99.98/100 = 0.21 g/sec; and
= 20/1000 x 6820/3600 x 99.98/100 = 0.038 g/sec.
3.4.9 References
1. Troxler, W.L., JJ. Cudahy, R. P. Zink, and S.I. Rosenthal. Thermal
Desorption Guidance Document for Treating Petroleum Contaminated Soils.
EPA Draft Report to James Yezzi, U.S. EPA, Edison, NJ. January 1992.
2. Lighty, J.S., G.D. Silcox, D.W. Pershing, V.A. Cundy, and D.G. Linz.
Fundamentals for the Thermal Remediation of Contaminated Soils. Particle
and Bed Desorption Models. ES&T Vol. 24, No. 5, pp750-757, May 1990.
3. Eklund, B., P. Thompson, A. Inglis, and W. Dulaney. Air Emissions From
the Treatment of Soils Contaminated with Petroleum Fuels and Other
Substances. EPA-600/R-92-124. July 1992.
4. de Percin, Paul (EPA). Personal Communication from Paul de Percin to Bart
Eklund of Radian Corporation. August 1992.
3-37
-------
SECTION 4
VOC AREA SOURCES
Simple air emission estimation procedures are presented in this section for area
sources of volatile organic compounds (VOCs), including: excavation, dredging and
dewatering, solidification/stabilization, and bioremediation. The same format is followed for
each source. A brief description of the emission process is given, followed by a discussion
of available air emission models. The model selected for inclusion in this manual is then
presented along with sources of input data and default values for each of the input variables
of the selected model. The model assumptions are then briefly discussed. Finally, an
example calculation is shown and references are listed. In all cases, the models estimate
uncontrolled VOC emissions.
A number of other area sources of VOC emissions may be present at Superfund sites,
but no applicable estimation techniques exist. These other sources include any in-situ
treatment processes (e.g., vitrification and bioremediation) and materials handling.
Frequently used materials-handling procedures at hazardous waste sites that may result in
VOC emissions include1: excavation and removal, dredging, pumping, size and volume
reduction, separation and dewatering, conveying systems, storage containers, bulking tanks,
drum handling and removal, compaction, and equipment decontamination. Only the first two
items listed and dewatering are addressed in this section.
4.1 EXCAVATION
4.1.1 Description of Emission Process
Excavation and removal of soils contaminated with fuels is a common practice at
Superfund sites. Excavation and removal may be the selected remediation approach or it
may be a necessary step in a remediation approach involving treatment. If removal is the
'U.S. EPA. Survey of Materials-Handling Technologies Used at Hazardous Waste Sites.
EPA/540/2-91/010. June 1991.
4-1
-------
preferred approach, the excavated soil is typically transported off-site for subsequent disposal
at a landfill. If the soil contains large amounts of fuel or highly toxic contaminants, the soil
may need to be treated off-site prior to final disposal. Excavation activities are also typically
part of on-site treatment processes such as incineration, thermal desorption, batch
biotreatment, landtreatment, and certain chemical and physical treatment methods. The soil
is excavated and transported to the process unit and the treated soil is typically put back into
place on the site.
Since it is rarely feasible or efficient to dig soil and immediately transfer the soil
directly to transport vehicles or treatment systems, soil will be handled several times. In
most cases, soil will be excavated and placed into a temporary holding area and then handled
one to two more times on-site. Elevated levels of VOC (and PM) emissions are possible
each time the soil is handled.
VOC emissions from handling operations result from the exchange of contaminant-
laden soil-pore gas with the atmosphere when soil is disturbed and from diffusion of
contaminants through the soil. There are multiple potential emission points for each of the
various soils handling operations. For excavation, the main emission points of concern are
emissions from:
exposed waste in the excavation pit;
material as it is dumped from the excavation bucket; and
waste/soil in short-term storage piles.
The magnitude of VOC emissions depends on a number of factors, including the type
of compounds present in the waste, the concentration and distribution of the compounds, and
the porosity and moisture content of the soil. The key operational parameters are the
duration and vigorousness of the handling, and the size of equipment used. The longer or
more energetic the moving and handling, the greater likelihood that organic compounds will
be volatilized. The equipment size influences volatilization by affecting the mean distance a
volatilized molecule has to travel to reach the air/solid interface at the surface of the soil. In
general, the larger the volumes of material being handled per unit operation, the lower the
percentage of VOCs that are stripped from the soil. Control technologies for large area
4-2
-------
sources such as excavation are relatively difficult to apply and are often much less effective
than controls for point sources.
Relatively limited VOC emissions or emission rate data for excavation are available.
The process of measuring emission rates from dynamic processes such as excavation is
difficult and costly, and so has rarely been attempted. Further, the factors that govern
emissions from materials handling are very complex. During excavation, for example, the
physical properties of the soil that control the vapor transport rate (e.g. air-filled porosity)
are changing with time and the concentration of contaminants may be rapidly decreasing.
4.1.2 Model Selection
While numerous models are available for estimating emissions from spills and buried
wastes (see Section 5), little work has been done to estimate VOC emissions from dynamic
processes such as materials handling. Orr1 developed emission models for excavation,
dumping, transportation, storage, and grading based on the RTI Landtreatment model2. The
Orr model for excavation was tested in the field3 and served as the basis for the development
of example procedures for evaluating air impacts from soil excavation4. The model requires
several iterative calculations and is therefore of limited use as a screening model. Further
development of a screening model for excavation/dumping was recently completed by
Eklund5.
EPA's Office of Underground Storage Tanks (OUST) has also developed a model for
estimating emissions from excavation and other remediation processes6. The model is based
on Pick's Law of Diffusion and can be used to calculate an emission flux from the soil-gas
concentration, diffusivity in air, and air-filled porosity of the soil. The model is similar in
some respects to those described above (i.e., governed by the effective diffusivity of the
compound in the soil). This model, however, is essentially a simple landfill model and it is
considered to be overly simplistic to accurately model excavation. The excavation/dumping
model developed by Eklund is the one most suited for the scope of this document. (A more
rigorous excavation model is also given in Reference 5.)
4-3
-------
4.1.3 Emission Model Equation
Average Emission Potential
A simple check of the potential total emissions from a site undergoing remediation is
given by Equation 4-1, which divides the estimated total mass of contaminants by the
projected duration of activity:
Sv C j8 1.0
ER =
(Eq. 4-1)
where:
ER
Sv
c
ft
1.0
t
emission rate of compound i [g/sec];
total volume of contaminated material [m3];
concentration of compound i in soil |jug/g];
bulk density [g/cm3];
constant [g/106/ug x 106cm3/m3]; and
duration of remediation [sec].
Since it is assumed in this equation that all of the contaminant present in the soil will
eventually volatilize, it is extremely conservative.
Emission Rate
The emission rate given in Equation 4-2 is the sum of emission rates from the soil
pore space and from diffusion:
where:
ER
P
Q
0.98
iIFF
ER = ERPS + ERDIFF
= P * Q * 0.98
(CS)(10,000)(SA)
1.22 x 106 -5 + 1.79 x 109 -i 2
P P
total soil emission rate of compound i [g/sec];
soil porosity emission rate of i [g/sec];
diffusion emission rate of i [g/sec];
vapor pressure of compound i [mm Hg];
excavation rate [mVsec];
conversion factor [g/mm Hg - m3];
(Eq. 4-2)
(Eq. 4-3)
(Eq. 4-4)
4-4
-------
Cs = mass loading of compound i in soil [g/cm3];
10,000 = conversion factor [cm2/m2];
SA = area of emitting surface [m2];
1.22xl06 = conversion factor [cm2-sec-mm Hg/g]; and
1.79xl09 = conversion factor [sec2-cm-mm Hg/g].
In most cases, contaminant data will be available as a soil concentration in units of ug/g
(ppmw). Assuming a typical bulk density of undisturbed soil, the mass loading, Cs, can be
related to the soil concentration as follows:
Cs = (C) (1.5 g/cm'XIO-6) (Eq. 4-5)
where: C = Concentration of species i in soil [ug/g]; and
10"6 = Conversion Factor [g//*gl-
The emission rate obtained using Equation 4-3 should be compared to the total mass
of contaminant present in the volume of soil excavated - M. If Equation 4-3 results in a total
mass of emissions that exceeds Va of M, then the following equation should be substituted for
Equation 4-3:
033 (Eq'4-6)
ERps = M -
Vv
where: tsv = Time to excavate a given volume, Sv, of soil [sec], and
M = Sv 106 Cs (Eq. 4-7)
where: M = total mass of contaminant in soil [g]; and
106 = conversion factor [cm3/m3].
The emission rate obtained using Equation 4-4 will overpredict emissions if the partial
pressure of the contaminant in the soil is far below the published vapor pressure of the
compound. The partial pressure should be calculated as follows. The pore space
concentration of the compound (assuming that all of the compound is in the vapor-phase) is:
(Eq. 4-8)
_ (CXPK106)
v
4-5
-------
where: Cv = concentration in soil pore spaces [/xg/m3];
106 = conversion factor [cm3/m3]; and
Ea = air-filled porosity [dimensionless].
Assuming a typical air-filled porosity of 0.55 and a bulk density of 1.5 g/cm3, the pore space
concentration can be calculated as:
Cv = (CX2.7X106) (Eq. 4-9)
where: 2.7xl06 = conversion factor [g/m3].
The partial pressure of the contaminant can then be calculated as:
/ ,
m (Cy) (62,361) (298)(1Q-12)
1 ~ MW
= (^(1.86 x IP'5)
MW
where: P{ = partial pressure of compound i [mm Hg];
62,361 = gas constant R [mmHg-cm3/mol-°K];
298 = assumed temperature [°K];
10~12 = conversion factor [g-m3/ug-cm3];
1.86xlO"5 = conversion factor [mmHg-g-m3/mol-ug]; and
MW = molecular weight of compound i [g/mol].
The partial pressure for a given compound obtained using Equation 4-10 should be
compared to the published vapor pressure given in Appendix A. If Equation 4-10 results in a
partial pressure that is below the published vapor pressure, then Pt should be substituted for
P in Equation 4-4. Equation 4-4 then more closely approximates the more rigorous model is
given in Reference 5.
4.1.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS)to provide the input
data necessary to generate reasonably accurate estimates of air emissions. The minimum
field data required are:
4-6
-------
Specific contaminants present in the soil or waste to be excavated; and
Average contaminant concentration in the soil or waste.
To use Equation 4-1, field data are necessary also to estimate the total volume of waste or
contaminated soil to be excavated.
4.1.5 Sources of Input Data
Information regarding the contaminants present, their concentrations, the volume of
soil to be excavated, and the physical properties of the soil such as bulk density will
generally be developed during the RI at the site. Information regarding the excavation rate
and surface area to be exposed will generally be developed during the remedial design based
on data from the RI/FS.
Physical property data, including vapor pressure, are tabulated in Appendix A for 168
compounds. Data for other compounds of interest may be obtained from a chemical
reference handbook (e.g., References 7 and 8).
4.1.6 Default Values for Input Variables
Table 4-1 gives default values to be used with in Equations 4-1 through 4-7 for a
typical removal action at a Superfund site.
If the approximate size of the excavation job is known, the values given in Table 4-2
can be used to estimate emissions using Equation 4-2. Table 4-2 also includes information
regarding the excavation to allow prediction of downwind ambient air concentrations using an
EPA-approved air dispersion model.
The rate of materials handling operations at Superfund sites tend to be controlled by
factors such as safety concerns, storage capacity or treatment capacity, rather than being
limited by the operational capacities of the equipment that is used. For these reasons, actual
materials handling rates tend to be far below typical handling rates at construction sites9.
4-7
-------
Table 4-1.
Default Values for Emission Equations
Parameter
Vapor Pressure
Volume of soil moved (per hour)
Excavation rate
Emitting surface area
Total mass of contaminanf
Concentration in soif
Time to excavate a given volume of soif
Bulk density
Symbol
P
Sv
Q
SA
M
C
tjv
P
Units
mm Hg
m3
nf/sec
nr
g
Mg/g
sec
g/cnf
Default Value
see Appendix A
150
0.042
290
--
-
--
1.5
Expected Range
50-240
0.014-0.067
115-308
--
1.0-2.0
Reference
5
5
5
10
"Obtain from RI/FS data.
Table 4-2.
Example Scenarios for Excavation of Contaminated Soil
Parameter
Soil moved per scoop
Number of scoops per hour
Total volume of soil moved
Excavation Pit:
Dimensions
Area
Release height
Storage Pile:
Dimensions
Area
Release height
Units
m3
#/hr
nf/hr
m
m2
m
m
m2
m
Scenario
Small
1
50
50
10 x 5 x 1
50
0
5x5x2
65
1
Medium :
2
75
150
10 x 15 x 1
150
0
5 x 10 x 3
140
1.5
Large:
4
60
240
10 x 12 x 2
120
0
8 x 10 x 3
188
1.5
Source: Reference 5
4-8
-------
4.1.7 Model Assumptions/Sensitivity Analysis
The derivation of these equations is presented in the EPA report5 along with a
sensitivity analysis of the models and tabulated input parameters. A more detailed
emission estimation model is also presented in that report.
The excavation equations are conservative and may tend to overpredict VOC
emissions; there is insufficient field measurement data at this time to gauge the accuracy
of the model.
Equation 4-1 includes the conservative assumption that all VOCs present in the
material will volatilize. Furthermore, it is assumed that the emission rate will be
constant and that the contaminants are evenly spread throughout the material.
Equations 4-3 and 4-4 are based on the assumption that the soil pore gas is saturated
with the contaminant of interest. If this is not the case, the equation may overpredict the
emission rate. This is likely to be the case for contaminants present in soil at sub-ppm
levels or those with relatively high vapor pressures.
4.1.8 Example Calculations
A site has approximately 10,000 nf of soil contaminated with chloroform, 1,1,1-
trichloroethane, and trichloroethylene in concentrations of 0.1, 10, and 1.0 ppm Otg/g),
respectively. The volume of the contaminated soil is not accurately known, and neither
is the excavation rate. The soil's bulk density averages 1.5 g/cm3. Removal of all
contaminants is expected to take 20 days of continual operation (1.728X105 s).
First estimate the total emissions potential for the site using Equation 4-1:
= 10,000 x 0.1 x 1.5 x 1 / (1.728xltf) = 8.7 x 104 g/sec;
= 10,000 x 1.0 x 1.5 x 1 / (1.728x10") = 8.7 x 1Q3 g/sec;
= 10,000 x 10 x 1.5 x 1 / (1.728x1(7) = 8.7 x 102 g/sec.
4-9
-------
Next, compare these rates to those predicted by Equation 4-2. This requires
knowledge of the excavation rate and a surface area, which may be assumed to be the
default values from Table 4-1 of 0.042 nf/s and 290 m2, respectively. Also needed is the
vapor pressure of the contaminant from Appendix A, which for chloroform is 208 mm
Hg. Also, the concentrations must be converted from ppmw to g/cm3 using Equation 4-
5. The pore space emissions are estimated using Equation 4-3:
ERps = 208 x 0.042 x 0.98 = 8.6 g/sec
The value, ER^, must be compared to Vs M/t^. The mass loading in the soil is
calculated form Equation 4-5:
Q = (0.1)(1.5)(10«) = 1.5 x 1C7
The amount of soil moved in one second is 0.042 m3, so after one second M/t.v for
chloroform would be (from Equation 4-7):
= 1.5 x Iff7 x 0.042 x Itf = 6.3 x 10* g/sec
Since 8.6 exceeds (6.3 x lQ3)/3 then ER^ is clearly greater thanVs M/t^. Use Equation
4-6 instead of 4-3:
1 V6.3 x ID'3)
= 1 x M = !£/ = 2.1 x ID'3 g/sec
3 'sv 1
From Equations 4-9 and 4-10, the partial pressure of chloroform can be
calculated:
Cv = (0.1)(2.7 x 106) = 2.7 x 105 jig/m3
and
p = (2.7 x 105) (62.361) (298) (IP"12)
1 ~ 119.38
= 0.042 mmHg
4-10
-------
This partial pressure is below the vapor pressure value for chloroform of 208 mmHg
given in Appendix A. Therefore, P; should be substituted for P in Equation 4-4.
Next, calculate the emissions due to diffusion using Equation 4-4:
TO 1.5 x 1(T7 x 10,000 x 290
-------
4.1.9 References
1. Orr, D. Estimating VOC Emissions During Soil Handling Operations at
National Priority List (NPL) Sites - Technical Note. EPA Contract No. 68-
02-4392, WA54. Nov. 13, 1989.
2. U.S. EPA. Hazardous Waste Treatment, Storage, and Disposal Facilities
(TSDF) - Air Emission Models. EPA-450/3-87-026. November 1989.
3. Eklund, B., D. Ranum, and A. Hendler. Field Measurements of VOC
Emissions from Soils Handling Operations at Superfund Sites. EPA
Contract No. 68-02-4392, WA64. Sept. 14, 1990.
4. U.S. EPA. Air/Superfund National Technical Guidance Study Series -
Development of Example Procedures for Evaluating the Air Impacts of
Soil Excavation Associated With Superfund Remedial Actions. EPA-
450/4-90-014. July 1990.
5. Eklund, B., S. Smith, and A. Hendler. Screening Procedures for Estimating
Air Impacts from the Remediation of Superfund Sites. EPA-450/1-92-004.
March 1992.
6. U.S. EPA Estimating Air Emissions from Petroleum UST Cleanups.
OUST. June 1989.
7. The Merck Index, llth Ed. Merck & Co. Rahway, NJ. 1989.
8. CRC Handbook of Chemistry and Physics, 61st Edition. CRC Press. Boca
Raton, Florida. 1980.
9. Church, H. Excavation Handbook. McGraw-Hill, 1981.
10. Schultz, H.L., et al. Superfund Exposure Assessment Manual.
EPA/540/1-88/001. April 1988.
4-12
-------
42 DREDGING AND DEWATERING
4.2.1 Description of Emission Process
Dredging is defined as the removal of material from the bottom of a water-
covered structure or geologic formation. It may be used for removing sediments from
the bottom of hazardous waste surface impoundments and the remediation of
contaminated waterways. Sediment removal operations may involve hydraulic dredging
or mechanical dredging. Hydraulic dredging is more commonly employed and uses
hydraulic pressure differential (vacuum) to remove sediments and pump them in slurry
form. The slurry usually consists of 10 to 60 percent solids. Mechanical dredging uses
direct mechanical force, such as scraping, to remove sediments.
The magnitude of VOC emissions from full impoundments are dependent on the
surface area of the impoundment, the depth of water, the surface area of the suspended
sediment plume due to agitation, the type and amount of contamination present, the
physical properties of the sediment and the associated contaminants, the length of time
sediments are exposed, and the meteorological conditions. Important physical properties
in the emission process are the volatility of the contaminants (i.e., their vapor pressure),
and their diffusivities in air and water. Emissions from drained impoundments are
dependent on the same factors listed above (with the exception of depth of water), and
the air-filled porosity of the sediments also is a factor. During hydraulic dredging,
bottom sediments are broken up and loose sediment particles float in the water in a
plume of dilute slurry.
Sediment removal is only one step in the overall remediation process for
dredging. Estimates of total VOC emissions must also consider emissions from de-
watering, storage, materials handling, and treatment and/or disposal.
4-13
-------
Dewatering is performed to increase the solids content of slurries and sludges.
Processes used for dewatering include gravity separation, granular bed filters, rotary
drum filters, filter presses, centrifuges, and belt filters. Emissions occur as sediments are
exposed to the atmosphere during processing.
422 Model Selection
Dredging models have been proposed by Thibodeaux1'2 for emissions of PCBs
from harbor sediments. These equations are relatively complex and require calculation
of parameters such as air-side transport coefficients and Schmidt numbers. Radian is
currently developing somewhat simpler dredging models for the U.S. EPA3. These are
modified versions of the LAND? model4 and are considered to be more appropriate for
inclusion in this manual since they involve more straightforward calculations. The
applicability of this model for dredged materials depends on accounting for the high
moisture content of the dredged sediments.
A dewatering model3 has been developed by RTT based on field measurements of
air emissions from sludge dewatering operations in the petroleum industry.
4.2.3 Emission Model Equation
Emission Rate for Dredging
Equation 4-11 is used for the calculation of the fraction of contaminants that are
emitted for an agitated sediment with no biodegradation:
X = 0.72 (K,, if (Eq. 4-11)
where: X = fraction of pollutant that is emitted [unitless];
0.72 = empirical constant [unitless];
K,, = contaminant volatilization constant [I/sec]; and
t = time sediment is exposed [sec].
4-14
-------
In order for Equation 4-11 to be valid, the following relationship must hold:
HDe t
e < 0.25
(Eq. 4-12)
I
where:
H
1
Henry's Law constant [conc./conc.];
effective diffusivity of contaminant in sediment air pores
[cm2/sec];
depth of sediment [cm].
The pollution volatilization constant, K,j, term in Equation 4-11 is determined as follows:
(Eq.4-13)
4 I
where:
pi, 3.14159 [unitless].
The effective diffusivity, De, in Equation 4-13 is estimated as follows (a simplified version
is given in Equation 4-16):
(Eq. 4-14)
10
Ew
where:
D,
E,
Pscd
diffusivity of compound in air [cm2/sec];
air-filled porosity of sediment [unitless];
total porosity of sediment [unitless];
diffusivity of contaminant in water [cm2/sec];
volumetric water content [Er - E,], [unitless];
sediment/water bulk density [g/cm3]; and
distribution coefficient [cnf/g].
The equilibrium coefficient between liquid and air, k.,, is equal to H for dilute solutions.
The distribution coefficient, k,j, can be estimated as:
= (0.63)
(Eq. 4-15)
where:
0.63 =
weight fraction of organic carbon in the sediment
[dimensionless];
octanol-water partition coefficient [cm3 water/cm3 octanol];
and
Empirical constant [cnf/g].
4-15
-------
Using the default values given later in this section, Equation 4-14 can be simplified to:
0.45 D
e 0.55 + 0.30 Kow
Emission Rate for Dewatering
The fraction of contaminants that are emitted over the short-term from
dewatering operations can be estimated as:
(0.0068)(P)095
X =
[(0.0068)(P)0-95]
where: X = fraction of pollutant that is emitted [unitless]; and
P = vapor pressure of pollutant [mm Hg].
4.2.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS) to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
only field data required is knowledge of the specific contaminants present in the soil or
waste to be dredged or dewatered.
To convert the fraction emitted into an emission rate, it is necessary to know the
total mass of contaminant present. This would require field data to determine the
average contaminant concentration in the soil or waste and the total volume of waste or
contaminated soil to be handled.
42.5 Sources of Input Data
Appendix A contains Henry's Law coefficients and diffusivities of selected
substances in air. The Henry's Law coefficient values from Appendix A must be
converted to dimensionless units; i.e., divide by R T as shown in Section 4.2.7.
Appendix A also contains values for the log of the octanol-water partition coefficient.
The antilog of these values should be used in Equation 4-15, as shown in Section 4.2.7.
4-16
-------
42.6 Default Values for Input Variables
Table 4-3 gives default values to be used in the emission estimation equations for
a typical dredging operation. Table 4-4 also includes information regarding typical
operating rates for various dredging scenarios.
For purposes of dispersion modeling, the dredging area and any storage pile that
is present can be assumed to have combined dimensions of 10 m by 10 m and a release
height of 1 m. Site-specific values should be used if available.
4.2.7 Model Assumptions/Sensitivity Analysis
The model is limited by the diffusion of the organic contaminant through both the
air and water that fill the pore spaces of the sediment. Equations 4-11, etc. are not
appropriate for undisturbed sediment or for sediment which is not exposed to air
(although they will provide a conservative estimate of atmospheric emissions in this last
case). The model assumes that Henry's Law applies; i.e., the solution is relatively dilute.
If this is not a valid assumption, then the Henry's Law constant term would need to be
replace with a term based on Raoult's Law.
These models assume there are no emissions from the liquid surface of the
impoundments. This is because dredging will not usually agitate the liquid surface to a
great enough extent to significantly increase the baseline emissions of the
uucontaminated water. Emissions from a stagnant surface are assumed to be negligible
in comparison to emissions from dredged sludge.
The sediments are assumed to be uniformly contaminated. Water evaporation is
assumed to be small, density changes due to volatilization of contaminants are expected
to be negligible, and the concentration of the contaminants in air is neglected. These
simplifications allow a readily usable emissions equation, so if any are incorrect, the
validity of the model will become more questionable.
4-17
-------
Table 4-3.
Default Values for Estimating Dredging Emissions
Parameter
Time that sediment is exposed
Depth of sediment
Henry's Law Constant
Diffusivity in air
Diffusivity in water
Air-filled porosity
Total porosity of sediment
Bulk density
Octanol-water partition coefficient
Weight fraction of organic carbon
Symbol
t
1
H
D.
Dw
E,
Ep
P«i
KOW
Xoc
Units
sec
cm
dimensionless
cm2 /sec
cm2 /sec
dimensionless
dimensionless
g/cm3
dimensionless
dimensionless
Default Value
-
100
see Appendix A
see Appendix A
see Appendix A
0
0.55
0.50
1.05
1.10
see Appendix A
0.45
Comments
Author's estimate
hydraulic dredge
mechanical dredge
hydraulic dredge
mechanical dredge
SOURCE: Reference 3, except for depth of sediment
Table 4-4A.
Specifications and Operating Characteristics for Hydraulic Dredges
Type
Cutterhead
Suction
Dustpan
Mudcat
Width
(ft)
10-12
10-12
10-12
8-9
Length
(ft)
25-60
25-60
25-60
25-40
Dredging Depth (ft)
Minimum
3
3
3
1
Maximum
20
20
20
15
Reach
(ft)
No limit
No limit
No limit
No limit
Solids Concentration in
Dredged Material (%)
10-20
10-15
10-20
10-60
Production
Kate (ytf/hr)
30-60
30-60
30-60
20-150
Table 4-4B.
Specifications and Operating Characteristics for Mechanical Dredges
Type ;
Clamshell
Backhoe
Dragline
Bucket
Capacity (yd3)
Vi
1
1
r/2
2
3
'/a
3/j
Ifc
3
Dredging Depth (ft)
Minimum
0
0
0
0
0
0
0
0
0
0
Maximum
150
150
22
25
30
45
60
60
60
60
Reach
(ft);
100
100
100
100
100
100
68
68
68
68
Solids Concentration in
Dredged Material (%)
Up to 100
Up to 100
Up to 100
"Production Rate
$;:₯:; :(yd3/hr) :'>
20
35
75
100
130
203
30
35
65
113
SOURCE: References
4-18
-------
4.2.8 Example Calculations
A dredging operation is being performed to remediate a lagoon bed with sludge
contaminated with carbon tetrachloride (CC^). The lagoon was drained prior to the
start of remediation. The dredging will remove 0.90 m of material from the bottom of
the lagoon and will take 4 days (3.46X105 sec) to complete. To estimate emissions, all
five of the equations in this section are used. Input values are obtained from three
different sources: field data, Appendix A to this report, and the default values in Table
4-3.
The Henry's Law constant of CCd, from Appendix A, is H = S.OOxlO2 atm-nf/g-
mol. To convert this to the dimensionless units required, divide by R T, the gas
constant times the temperature:
H - <3-°° ' 10'2> = 1.227
(8.205 lO'5 298)
The diffusivity in air and in water of CC^, from Appendix A, are D, = 0.0632 cm2/sec
and Dw = S.SxlO6 cm2 /sec. The log of the octanol-water partition coefficient is log KQW
= 2.83.
The default values will be used for the other required input parameters: E^ = 0
and XQC = 0.45. The type of dredge has not been specified, so the more conservative
values of the mechanical dredge will be used: Ep = 0.50, and psed = 1.1 g/cm3.
Using Equation 4-15:
kj = 0.63 * 0.45 * 1CP3 = 191.67.
4-19
-------
This value of k^ can then be used in Equation 4-14:
1.227 0.0632
10 '
O3
[(5)2
+ 8.8 x 1Q-6
10
(.5-0) 3
(-5)2 .
1.10 x 191.67 + (0.5-0) + 0 1.227
1.65 x 10'8 cm2/sec
This value of De can then be used in Equation 4-12 to verify that Equation 4-11 may be
used:
(1.227)(1.65 x 10-8)(3.46 x lO"5) = g 65 x 1Q-7
(90)2
This value is less than 0.25, which means that the time of exposure is in the regime
where Equation 4-11 is valid.
The final step before calculating the emission fraction is the calculation of
Equation 4-13:
K = 1-227 1.65 x IP'8 (3.14159)2 = 6 1? x 1Q-i2
d" (4-(90)2)
The fraction of pollutant emitted can then be calculated from Equation 4-8:
F = 0.72 * (6.17 x 1012 3.46 x itff = 1.05 x 103 or 0.1%
The emissions estimate would be higher if a non-zero value for the air-filled
porosity were used.
The emissions due to dewatering are estimated from Equation 4-17 and the vapor
pressure value of 113 mm Hg for carbon tetrachloride from Appendix A:
(0.0068X1 13)
a95
[(0.0068)(113)095]
= 0.378 or 38%
4-20
-------
The emissions from dewatering greatly exceed those from dredging.
42.9 References
1. Thibodeaux, LJ. Modeling of VOC Emissions From Dredging at
Massachusetts Site. Report to Army COE. 1989.
2. Thibodeaux, LJ. Theoretical Chemodynamic Models For Predicting
Volatile Emissions To Air From Dredged Material Disposal. In:
Intermedia Pollutant Transport, Edited by Allen, Cohen, and Kaplan.
Plenum Publishing Corp., 1989.
3. Radian Corp. Preliminary Assessment of Potential Organic Emissions
From Dredging Operations. Draft Report to Mr. Dennis Timberlake, U.S.
EPA/ORD. September 30, 1991.
4. U.S. EPA, Hazardous Waste Treatment, Storage, and Disposal Facilities
(TSDF) - Air Emission Models. EPA-450/3-87-026. November 1989.
4.3 SOLIDIFICATION/STABILIZATION
4.3.1 Description of Emission Process
Several types of stabilization/solidification (S/S) technologies exist as alternatives
for remedial action. The goal of these processes is to immobilize the toxic and
hazardous constituents in the waste, usually contaminated soil or sludge. A few of these
processes involve in-situ treatment, however, most generally require excavation and other
soil handling activities. Nearly all the commercially available stabilization and
solidification technologies are proprietary. S/S processes may be considered to be point
sources of VOC emissions if the process is enclosed or has an air collection system.
Solidification and stabilization processes are usually batch operations, but may be
continuous and all follow the same basic steps. Wastes are first loaded into the mix bin
(wastes are sometimes dried before addition to the bin), and other materials for the
solidification or stabilization are added. The contents of the bin are then thoroughly
mixed. After a sufficient residence time, the treated waste is removed from the bin.
4-21
-------
The amount of fixative added may be equal to the mass of the contaminated material.
The solidified material is usually formed into blocks and allowed to cure for up to
several days. The blocks can then be placed in lined excavations on-site. This
description does not apply, however, to in-situ treatment methods, which use a variety of
techniques (from applied high voltage to injection of stabilizing agents) to immobilize the
contaminated waste in-place without excavation or soils handling.
Typical raw materials used in stabilization processes include fly ash, portland
cement, cement kiln dust, lime kiln dust, or hydrated lime. Other additives that may be
used to solidify or encapsulate wastes include asphalt, paraffin, polyethylene, or
polypropylene.
The primary source of air emissions from stabilization and solidification processes
is volatilization of organic contaminants in the waste. Volatilization can occur during
waste handling activities such as soil excavation and transport or during the process of
mixing the binding agents with the waste. Also, some evaporative emissions will occur
from waste even after stabilization, especially during the curing period immediately after
the blocks are formed. Lab studies, though, have shown that the largest fraction of
volatile loss occurs during the mixing phase because heat may be required to assist
mixing or generated by exothermic stabilization reactions.
In general, VOC emissions from stabilization and solidification processes will
depend on the type and concentration of the VOCs in the waste, the duration and
thoroughness of the mixing, the amount of heat generated in the process, and the
average batch size processed. The longer or more energetic the mixing and processing,
the greater likelihood that organic compounds will volatilize. The volatile losses will also
increase as the temperature of the waste/binder mixture increases. Binding agents with
high lime contents generally cause highly exothermic reactions. The batch size influences
volatilization by affecting the mean distance a volatilized molecule has to travel to reach
the air/solid interface at the surface of the stabilized waste. The larger the block of
material, the lower the rate of volatilization.
4-22
-------
During the solidification/stabilization soil remediation process, there are
numerous possible VOC (and PM) emission sources. There are fugitive emissions before
treatment, emissions during excavation and soil handling, during the preparation of the
mixing agent, during the treatment of the contaminated soil, and, finally, emissions from
the treated soil after remediation. Factors influencing the (uncontrolled) emission rate
will therefore include the soil permeability before and after remediation, the exact
treatment process and how the mixing is accomplished, and the composition of the
mixing agent. Indeed, the latter may be specifically designed to control one certain class
of contaminants, and may not be effective on any others. The impermeability of the
treated soil will also determine the amount of emissions (and leachate) that escape.
-4.3.2 Model Selection
Little information exists about the fate of volatile contaminants in wastes treated
by stabilization and solidification methods. A literature search found no available field
data on air emissions at Superfund sites using this type of remediation technology.
Laboratory studies, however, have estimated that 70-90% of the volatile contaminants in
the treated waste eventually evaporate. Experiments also show that most of the loss
occurs within 60 minutes of mixing the waste with binding agents. The only air emissions
model for solidification/stabilization is the simple mass balance equation presented by
Thompson, et al.1.
4.3.3 Emission Model Equation
VOC emissions from stabilization and solidification processes can be estimated
using a mass-balance approach. The following equation is applicable to ex-situ
solidification/stabilization processes:
4-23
-------
ER = C F (2.78 x 107)(V/100) (Eq. 4-18)
where: ER = emission rate of contaminant i [g/sec];
C = concentration of contaminant i in soil \fj.g/g]',
F = treatment (feed) rate of soil [kg/hr];
2.78 x 10"7 = conversion factor [g/kg g//*g hr/sec]; and
V = fraction of contaminant i volatilized [%].
4.3.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS) to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
only field data required is knowledge of the specific contaminants present in the soil or
waste to be stabilized or solidified and the average contaminant concentration.
4.3.5 Sources of Input Data
The inputs to Equation 4-18 are process- and site-specific. The treatment rate of
the unit can be obtained from the vendor or estimated from design documents and the
results of any feasibility study. The fraction of VOCs that will be stripped during the
process will be highly dependent on the system design and operating procedures. Field
test data should be obtained to estimate this parameter. The concentration of
contaminant in the soil or sludge to be treated should be available from remedial
investigation studies.
4.3.6 Default Values for Input Variables
Table 4-5 gives default values to be used in Equation 4-18. Feed rates vary
widely; values of 5 to 130 tons/hr have been published for ex-situ processes and 25 to
100 tons/hr for in-situ processes.
4-24
-------
Table 4-5
Default Values for Estimating Emissions From Solidification/Stabilization
Parameter
Feedrate
% Volatilized
Symbol
F
V
Units
kg/hr
%
Default Value
45,000
80"
10CP
Expected Range
4,500-120,000 (ex-situ)
23,000-91,000 (in-situ)
40-1001
10CF
Reference
6
7
5
5
"During mixing
b After 40 days of curing
4-25
-------
For purposes of dispersion modeling, an ex-situ S/S unit can be assumed to have
dimensions of 5m by 5m and a release height of 2m. This is based on a 10,000 gallon
tank set at ground level without any air emission collection hood, control devices, or
stack. An in-situ process can be assumed to have a treatment area of 10m by 10m with a
release height of 1m. For either type of process, no plume buoyancy should be assumed
(even though exothermic processes may result in plumes with some rise). As always,
site-specific or process-specific values should be used if available.
4.3.7 Model Assumptions/Sensitivity Analysis
The major limitation of this model is the lack of air emissions data available for
developing default values for the term in the model for the percent of VOCs lost from
the process. Only one field study2 and two laboratory studies3'4'5 of air emissions from
these processes have been identified, though a third study6 does provide some useful
performance data. These studies show that from 40 to 100% of the VOCs present in the
waste are lost during the mixing step of the processes. Essentially all of the VOCs are
lost to the atmosphere by the end of the curing step.
43.8 Example Calculations
Assume that S/S is to be used to clean up a site that is primarily contaminated
with heavy metals, but where benzene is also present in the soil at an average
concentration of 0.25 ug/g (ppmw). An ex-situ process unit will be used and a treatment
rate of 100 tons per hour is planned.
Using Equation 4-18 and the default value for the fraction of the benzene that
will be volatilized, the emissions can be estimated once the treatment rate has been
converted into the proper units. A rate of 100 tons/hr equals about 91,000 kg/hr. The
estimated emissions are:
4-26
-------
ER = (0.25)(91,000)(2.78 x 107)(80/100) = 0.005 g/sec
4.3.9 References
1. Thompson, P., A. Inglis, and B. Eklund. Emission Factors For Superfund
Remediation Technologies - Draft Technical Note. EPA-450/1-91-001.
March 1991.
2. Ponder, T.C. and D. Schmitt. Field Assessment of Air Emissions From
Hazardous Waste Stabilization Operations. In: Proceedings of the 17th
Annual Hazardous Waste Research Symposium, EPA/600/9-91/002.
April. 1991.
3. Weitzman, L., L. Hamel, P. dePercin, B. Blaney. Volatile Emissions from
Stabilized Waste. In: Proceedings of the Fifteenth Annual Research
Symposium. EPA-600/9-90/006. February 1990.
4. Weitzman, L., L. Hamel, and S. Cadmus. Volatile Emissions From
Stabilized Waste - Final Report. Report for EPA/RREL, Cincinnati under
EPA Contract No. 68-02-3993, WA32 and 37. May 1989.
5. Sykes, A.L., W.T. Preston, and D.A. Grosse. Volatile Emissions from
Stabilization/Solidification of Hazardous Waste. Presented at the 85th
Annual AWMA Meeting (Paper 98.07), Kansas City, MO, June 21-26,
1992.
6. Stinson, Mary K. EPA SITE Demonstration of the International Waste
Technologies/Geo-Con In Situ Stabilization/Solidification Process. Journal
of Air & Waste Management vol. 40, no. 11. November 1990.
7. U.S. EPA. HAZCON Solidification Process, Douglassville, PA -
Applications Analysis Report. EPA/540/45-89/001. May 1989.
8. U.S. EPA, International Waste Technologies/Geo-Con In-Situ
Stabilization/Solidification - Applications Analysis Report. EPA/540/45-
89/004. August 1990.
4-27
-------
4.4 BIOREMEDIATION
4.4.1 Description of Emission Process
Bioremediation of soil at Superfund sites may be either in-situ or ex-situ. Ex-situ
biodegradation is the general term for treatment processes where soil or waste is
excavated prior to treatment. In some cases, an aqueous slurry is created by combining
soil or sludge with water and then biodegraded in a self-contained reactor or in a lined
lagoon. This is an emerging technology and is often referred to as slurry biodegradation.
Ideally, the waste is decomposed into carbon dioxide and water. Ex-situ bioremediation
also may be performed as a "dry" process such as composting or landtreatment (e.g., use
of white rot fungus).
In-situ treatment employs the natural microbiological activity of soil to decompose
organic constituents. This natural biological activity may be enhanced by the use of
injection wells to provide an oxygen source (such as air, pure oxygen, or hydrogen
peroxide) to stimulate aerobic degradation or the addition of nutrients to support the
growth of waste-consuming microorganisms. In some cases, microorganisms that have
the ability to metabolize specific contaminants of interest may be added to the soil. One
specific type of in-situ bioremediation process - bioventing - is discussed briefly in
Section 3.2.
In-situ bioremediation at Superfund sites may also involve sequential isolation and
treatment of waste areas using processes that closely resemble ex-situ processes except
that it may not be necessary to excavate, pump, or otherwise transfer the waste material
prior to treatment. Ex-situ processes are more developed and demonstrated than in-situ
processes at this time.
4-28
-------
Ex-situ aqueus treatment systems have a number of components, all of which
could be emission sources: mix tank, bioreactor system (continuously stirred tank reactor
or CSTR), or lined lagoon. Since aerobic treatment is the most common mode of
operation for slurry biodegradation, aeration must be provided to the bioreactors by
either floating or submerged aerators or by compressors or spargers. Other typical
system components are a separation/dewatering system, a clarifier for gravity separation,
and wastewater storage tanks. The soils handling steps required to deliver the
contaminated soil to the treatment unit may also emit significant amounts of VOCs (and
PM).
Biodegradation is actually only one of several competing mechanisms in
biotreatment. For ex-situ processes, the contaminants may also be volatilized, undergo
chemical degradation, or be adsorbed onto the soil particles. For in-situ processes, the
same pathways exist along with leaching. The overall removal achieved by biotreatment
processes represents the combined impact of all of these mechanisms. Volatilization
may account for the disappearance of the majority of VOCs being treated.
In open lagoons, the primary environmental factors that influence air emissions
are process temperature and wind speed. Emissions tend to increase with an increase in
surface turbulence due to wind or mechanical agitation. Temperature affects emissions
through its influence on microbial growth. At temperatures outside the band for optimal
microbial activity, volatilization will increase. Emissions from self-contained reactors are
also determined by reactor design parameters such as the amount of air or oxygen used
to aerate the slurry. Higher gas flow will strip more volatiles out of solution and
increase air emissions.
4-29
-------
4.4.2 Model Selection
Many models have been proposed over the years to estimate the fate of
contaminants from bioremediation processes that include an examination of volatilization
and stripping fate mechanisms. Models have been compiled for activated sludge, surface
impoundments, batch reactors, fixed film systems, landtreatment, in-situ soil
bioremediation, etc. and the models' ability to estimate air emissions have been
evaluated1. In addition, the U.S. EPA has published models for estimating air emissions
from RCRA facilities such as wastewater treatment and landtreatment systems2.
References 1 and 2 should be used for modeling air emissions from such conventional
bioremediation processes. For in-situ processes, existing air emissions models for
landtreatment could be used as a worst-case scenario; i.e., contaminated soil in direct
contact with the atmosphere.
Many novel bioremediation processes have been proposed for use at Superfund
sites3 and air emission models for these processes have not been developed. The air
emissions model considered to be best-suited for bioremediation of contaminated soil is
the simple mass balance equation presented by Thompson, et al.4. The major limitation
of this approach is the lack of air emissions data available for developing default values
for the term in the model that accounts for the fraction of VOCs lost to the atmosphere.
In addition to the Thompson model, a component of the Namkung-Rittman model1 is
presented for biotreatment of liquids or slurries with subsurface aeration resulting in off-
gassing.
4.43 Emission Model Equation
The mass balance approach shown below can be used for estimating emissions
from a flow-through impoundment treating contaminated water or from an ex-situ
bioslurry process:
4-30
-------
(Eq. 4-19)
where: ER = emission rate of contaminant i [g/sec];
( = volume rate of water treated [L/min];
60 = conversion factor [sec/min];
C = concentration of contaminant in slurry [mg/L];
1000 = conversion factor [mg/g]; and
V = percentage of contaminant i volatilized [%].
For batch biotreatment systems such as disposal impoundments, portable covered
reactors, or landfarms, Equation 4-20 may be used to estimate air emissions:
4'20)
where: ER = emission rate of contaminant i [g/sec];
M = mass of soil treated [Kg], or volume of water [L];
t = residence time in treatment system [sec];
C = concentration of contaminant in soil [ug/g], or liquid [mg/L];
1000 = conversion factor [Kg-ug/g7] for soil, or [mg/g] for liquid; and
V = percentage of contaminant i volatilized [%].
Note that Equation 4-20 applies to contaminants in liquid or soil, with the appropriate
units chosen for mass treated and contaminant concentration.
For aqueous systems with off-gassing, the stripping rate can be estimated as
follows:
_ QHC1.0
RT
where: ER = Emission rate [g/sec];
Q = off-gas rate [of/min];
H = Henry's Law constant [atm-nf/mole];
C = concentration of contaminant in water [mg/L];
R = gas constant [8.206 x 10s irf -atm/°k-mole];
T = absolute temperature [°K]; and
1.0 = conversion factor [g/mg 1/m3].
4-31
-------
4.4.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS) to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
minimum field data required are:
Specific contaminants present in the solid or liquid waste to be treated;
Average contaminant concentration in the slurry, solid, or soil; and
Percentage of contaminant volatilized from the process.
4.4.5 Sources of Input Data
The inputs to Equations 4-19, 4-20, and 4-21 are all process- and site-specific.
The treatment rate or capacity of the biodegradation process unit and any off-gas rate
can be obtained from the vendor or estimated from design documents and the results of
any feasibility study. The fraction of VOCs that will be stripped during the process will
be highly dependent on the system design and operating procedures. Field test data
should be obtained to estimate this parameter. The concentration of contaminant in the
soil, sludge, or water to be treated should be available from remedial investigation
studies.
4.4.6 Default Values for Input Variables
The percentage of each contaminant which is volatilized in either the continuous
or batch-treatment methods will vary greatly depending on the physical properties of the
contaminant, the impoundment geometry, and the biological activity. Emission factors
have been published for biodegradation based on theoretical studies4. For mechanically-
aerated, flow-through impoundments, an assumed value of V is 80%. Field
measurement data for wastewater treatment systems is summarized in Table 4-6.
The other variables are process specific; no default values are available. For
purposes of dispersion modeling, the biodegradation unit can be assumed to have
4-32
-------
Table 4-6.
Default Values for Estimating Emissions from Bioremediation
Parameter
% Volatilized
Symbol
V
Units
%
Default Value
80 (H = la3)
10 (H = 105)
Expected Range
0.2 - 21"
11 -9?
2-73°
1 -81b
Reference
4
5
6
7
8
"For a given system, the percent volatilized should vary with the Henry's Law constant
for the compounds of interest.
b Industrial wastewater treatment system
'Bench-scale reactor
H = Henry's Law Constant in atm-nf/mol.
4-33
-------
dimensions of 5m by 5m and a release height of 2m. This is based on a 10,000 gallon
tank set at ground level without any air emission collection hood, control devices, or
stack. As always, site-specific or process-specific values should be used if available.
4.4.7 Model Assumptions/Sensitivity Analysis
The key variable/assumption in the emission equations is the fraction of a given
VOC that is lost to the atmosphere before it is degraded. Little information exists on
volatile losses from slurry biodegradation processes. Slurry processes have only recently
become commercially available and field experience to date is limited. However, data
on air emissions from wastewater biotreatment processes are available. The percentage
of each contaminant that is volatilized will vary greatly depending on the physical
properties of the contaminant and the design of the treatment system. As shown in
Table 4-6, the range of values for the percent volatilized may vary over a very wide
range. Due to the lack of data, a conservative assumption of 80% volatile losses was
made; field measurement data should be obtained to get more accurate information for
specific processes. Percentage emissions for systems without mechanical aeration would
be lower.
4.4.8 Example Calculations
Consider a site with a contaminated lagoon. The lagoon holds 500,000 L with an
area of 100 m2. The sludge beneath it is contaminated to a depth of about 3 m. The
contaminants present in the sediments are benzene and chlorobenzene. The overlying
water is considered to be uncontaminated. The concentrations are 10 ug/g benzene and
20 ug/g chlorobenzene in the sludge. The bulk density of the sediments was measured
and is 2.0 g/cm3. Therefore, the 300 m3 of contaminated sludge would weigh 600,000
Kg. A batch biotreatment system will be used with a treatment rate of 2000 Kg batches
treated for one day (86,400 sec). The Henry's Law constants for both compounds are in
the 103 range, so V is assumed to be 80%.
4-34
-------
Using Equation 4-20 and the default value for V, the emission rates are estimated
to be:
= (2,000/86,400) (10/1000) (80/100) = l.QxlO4 g/sec; and
= (2,000/86,400) (20/1000) (80/100) = S.TxlO4 g/sec.
4.4.9 References
1. Sharp-Hansen, S. Available Models for Estimating Emissions Resulting
from Bioremediation Processes: A review. EPA/600/3-90/031. March
1990.
2. U.S. EPA Hazardous Waste Treatment, Storage, and Disposal Facilities
(TSDF) - Air Emission Models. EPA-450/3-87-026. November 1989.
3. U.S. EPA The Superfund Innovative Technology Evaluation Program:
Technology Profiles - Fourth Edition. EPA/540/5-91/008. November
1991.
4. Thompson, P., A. Inglis, and B. Eklund. Emission Factors For Superfund
Remediation Technologies - Draft Technical Note. EPA-450/1-91-001.
March 1991.
5. Eklund, B., D. Green, B. Blaney, and L. Brown. Assessment of Volatile
Organic Air Emissions From an Industrial Aerated Wastewater Treatment
Tank. In: Proceedings of the 14th Annual Hazardous Waste Research
Symposium, EPA/600/9-88/021, pp468-475. July 1988.
6. Boyd, R., J. Nottoli, and C. Schmidt. Measurement of Toxic Air
Contaminants from an Industrial Wastewater Treatment Plant. Presented
at the 85th Annual AWMA Meeting (Paper 94.08), Kansas City, MO. June
21-26, 1992.
7. Sun, P., C. Meyers, and S. Rajagopalan. Competition Between
Biodegradation and Volatilization in Wastewater Biotreatment Systems.
Presented at th 85th Annual AWMA Meeting (Paper 94.07), Kansas City,
MO, June 21-26, 1992.
8. Barton, D. and J. McKeown. Field Verification of Predictive Modeling of
Organic Compound Removal by Biological Wastewater Treatment
Processes. Env. Progress, Vol 10, No.2 pp.96-103, May 1991.
4-35
-------
SECTION 5
NON-PROCESS VOC AREA SOURCES
This section presents simple estimation procedures for landfills, lagoons, and spill
sites. Spill sites are defined to mean any site where there has been a spill or leak or where
there is an open (uncovered) waste pit. Each of the estimation procedures in this section are
based on those contained in EPA's Superfund Exposure Assessment Manual.
5.1 COVERED LANDFILLS
5.1.1 Description of Emission Process
Covered Landfills Without Internal Gas Generation
For the purposes of this manual, a covered landfill is a site where hazardous waste
lies beneath a layer of soil. If the waste was deposited intentionally, a pit may have been
excavated prior to placement of the waste. In most cases, the pit will have some type of
liner such as concrete, compacted clay, or polymer sheeting. The landfill may also result
from migration of contaminated groundwater, or an accidental spill or leak. In any case, the
waste enters the atmosphere after it volatilizes and diffuses through the soil cover.
The VOC emission rate from subsurface contamination covered by clean soil is
controlled by the rate at which gas diffuses through the soil pore spaces. Any factors that
significantly affect this diffusion rate will significantly affect VOC emission rates. Important
chemical processes are the adsorption of gas molecules onto the liquid film surrounding soil
particles and subsequent reactions of the adsorbed molecules. The physical transport of
vapors through porous media such as soil has been discussed elsewhere1"4. In general,
physical transport is controlled by the diffusivity in air for the specific compound of interest
and the number and type of the air spaces that are present. Macrospaces due to cracks,
fissures, spaces between buried drums, etc. will allow for relatively rapid mass transport.
The diffusion rate through soil lacking such obvious pathways will be a function of the air-
filled porosity (i.e., permeability to air). The permeability of soil to air will vary by up to
5-1
-------
three orders of magnitude across a typical residential lot5. The air-filled porosity will also
vary over time. For example, precipitation causes water to fill some of the interstitial spaces
in the soil and thereby prevents diffusion from occurring.
The landfill may or may not have a cover in addition to a soil layer. Any polymer
cover will inhibit emissions, but this effect will be diminished if the cover contains holes or
if it has degraded over time and has become more gas permeable. The surface soils may
also act as a barrier to emissions. The operation of heavy equipment at a site may lead to
compaction of the soil and diminished emissions due to the change in soil porosity. Landfill
surfaces, however, frequently have fissures as the result of the type of soil cover and settling
of the waste over time. The fissures may be long and deep enough to extend into the waste
body. Such fissures 'may account for a significant amount of the total emissions.
Covered Landfills With Internal Gas Generation
Landfills that contain municipal or sanitary wastes in addition to hazardous waste are
called co-disposal sites. The municipal or sanitary wastes have a high organic content and
their degradation may result in the production of large volumes of methane (CH*), carbon
dioxide (CO^ and hydrogen (H2). The gas generation results in a pressure-gradient in the
soil and transport via advection of these gases through the landfill. The methane, etc. will
act as a carrier gas and will enhance emissions of other VOCs. The effect of landfill gas
generation is sufficiently large that vapor diffusion can be ignored.
The key variable affecting VOC emissions from co-disposal sites is the gas generation
rate. This in turn is dependent on the volume of municipal wastes present, the length of time
it has been in the landfill, and the levels of water, oxygen, and other nutrients in the landfill.
5-2
-------
5.1.2 Model Selection
The Superfund program has previously selected simple screening models in the
Superfund Exposure Assessment Manual (SEAMs)6. The landfill models from this reference
were adapted for this report. The model for landfills without internal gas generation is
Fanner's equation7 as modified by Shen8 and Farino, et al.9 The literature on VOC
emissions from landfills is very extensive, and recent summaries are available2'10.
5.1.3 Emission Model Equation
Emissions from covered landfills with no internal gas generation are described in
Equation 5-1:
(Eq. 5-1)
ER _ a
Xmol SA IP'12
where:
ER
Da
cg
SA
E,
io-12
0.01
1
0.01 1
emission rate of compound i [g/sec];
diffusivity of compound in air [cnvVsec];
saturation vapor concentration of compound i [/*g/m3];
exposed area [m2];
air-filled soil porosity [unitless];
mole fraction of compound i in the waste [mol/mol];
conversion factor [g/cm3 / /ig/m3];
conversion factor [m/cm]; and
depth of soil cover [m].
The saturation vapor concentration term, Cg, may be obtained from Appendix A for
many compounds of interest or calculated as follows:
(Eq. 5-2)
/« _
P MW IO12
(RT)
where:
P
MW
IO12
R
T
vapor pressure of compound i [mm Hg];
molecular weight of compound i [g/mol];
conversion factor |>g/g * cm3/m3];
ideal gas constant 62,361 [mm Hg-cm3/mol-°K]; and
absolute temperature [°K].
5-3
-------
The mole fraction term, X^,, may be calculated as follows:
C MW
^ mw
_
1001
CWMW
where: C = concentration of compound i in soil
Cw = concentration of waste in soil [/xg/g] ; and
MWW = molecular weight of waste [g/mol].
The air-filled soil porosity, Ea, can be calculated as follows:
(Eq. 5-4)
p + ^PJtAwrJ
E = 1 -
a
where: /8 = bulk density of soil [g/cm3];
XH2o '= moisture fraction in soil [wt. % moisture/100]; and
p = particle density of soil [g/cm3].
If there is internal gas generation in a covered landfill, the following equation should
be used:
(Eq. 5-5)
ER = Cv Vy SA 104
where: ER = emission rate of compound i [g/sec];
Cv = vapor concentration of compound i in soil pore spaces [g/cm3];
Vy = mean landfill gas velocity in the soil pore spaces [cm/sec];
SA = exposed area [m2]; and
104 = conversion factor [cm2/m2].
5.1.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS)to provide the input
data necessary to generate reasonably accurate estimates of air emissions. The minimum
field data required to estimate emissions from landfills without internal gas generation are:
Surface area of emitting area;
Specific contaminants present in the landfill soil or waste;
Average contaminant concentration in the soil or waste; and
Average total contaminant concentration in the soil or waste.
5-4
-------
The latter information is necessary to estimate the molecular weight of the waste term,
MWW.
For landfills with internal gas generation, the minimum field data that are required
are:
Surface area of emitting area;
Specific contaminants present in the landfill soil-gas; and
Average contaminant concentration in the landfill soil-gas.
5.1.5 Sources of Input Data
Certain inputs to Equations 5-1, 5-3, 5-4, and 5-5 should be obtained from field
measurements. The vapor concentration in the soil pore spaces should be measured in the
field or assumed to be saturated (assume Cv = Cg xlO"12). Appendix A contains values for
168 compounds for the saturation vapor concentration, vapor pressure, and diffusivity in air.
A simple method for estimating the diffusivity in air of a compound given the diffusion
coefficient of a compound of similar molecular weight and diffusion volume is:
(Eq. 5-6)
where: Da' = diffusivity in air of compound of similar volume and wt.
[cm2/s];
MW' = molecular weight of similar compound [g/mol]; and
MW = molecular weight of contaminant [g/mol].
5.1.6 Default Values for Input Variables
Table 5-1 contains default values needed for estimating landfill emissions. For
purposes of dispersion modeling, the 5 acre default landfill can be assumed to have
dimensions of 140m by 140m and a release height of 1m.
5-5
-------
5.1.7 Model Assumptions/Sensitivity Analysis
The landfill models incorporate a number of assumptions, including that the landfills
are isothermal, contain no fissures or macropores, and that waste is homogeneously
distributed. The model for landfills without internal gas generation, equation 5-1, is highly
sensitive to the air-filled porosity of the soil cover or cap as well as the depth of the cover.
The accuracy of the model for landfills with internal gas generation, equation 5-5, will
depend on the validity of the input values for the soil vapor concentration and the landfill gas
velocity. These parameters are difficult to measure accurately in the field.
It is not recommended that temperatures other than 298 °K (25 °C) be used in
equation 5-2, unless the vapor pressure term is also adjusted for temperature. Equation 5-6
is only valid for estimating a compound's diffusion coefficient from another one with a
similar molecular diffusion volume and molecular weight.
5.1.8 Example Calculations
As an illustration of these models, consider a disposal pit known to contain 1,1-
dichloroethane (DCA) and vinyl chloride. Soil core samples taken in and around the pit
were analyzed and found to contain the following levels (jtg/g) of contaminants:
Core
1
2
3
1,1-DCA
900
1100
1000
Vinyl Chloride
500
900
800
The surface area of the pit is about 250 m2. The core samples were taken ten to twenty feet
below the surface, but the cover thickness has not been measured. The total contamination
in the soil is about 1% or 10,000 /xg/g. The soil porosity and density are unknown.
5-6
-------
Table 5-1.
Default Values for Estimating VOC Emission From Landfills
Parameter
Diffusivity in air
Saturated vapor concentration
Surface Area
Air-filled porosity
Depth of soil cover
Vapor concentration in soil pore spaces
Landfill gas velocity
Vapor pressure
Molecular weight of compound
Temperature
Molecular weight of waste
Bulk density of soil
Particle density of soil
Symbol
D.
q
SA
^
1
Q,
%
p
MW
T
MWW
ft
P
Units
cnf /sec
Mg/nf
rf
unitless
m
g/cirf
cm/sec
mm Hg
g/mol
K
g/mol
g/cirf
g/cirf
Default Value
See Appendix A
See Appendix A
20,000 (5 acres)
055 (Dry, uncompacted soil)
035 (wet or compacted soil)
0.05 (sludges)
1
q xiou
0.00163
See Appendix A
See Appendix A
298
250
1.5
2.d5
Expected Range
4050-40,500
1-5
-
-
-
278-328
-
1.0 - 2.0
2.4 - 2.8
Reference
-
Author's estimate
6
Author's estimate
-
- 6
6
-
Author's estimate
Author's estimate
6
9
5-7
-------
Due to the lack of exact information, default values will be used. The pit is
therefore assumed to have a cover depth of 1 m. The site is not known to contain any
municipal or sanitary wastes, so the landfill model without internal gas generation is
applicable (calculations using the model with gas generation are also shown for
illustrative purposes). The soil density is assumed to be 1.5 g/cm3, the molecular weight
of the waste is 250 g/mol, and the porosity is 0.55.
From Appendix A, the following values are obtained:
Parameter
Diffusivity in Air
Saturated Vapor Cone.
Molecular Weight
Units
cm2 /sec
g/cm?
g/mole
1, 1-Dichloroethane
0.0919
1.24xlCP
99
Vinyl Chloride
0.0900
8.94xltf
62.5
The core samples show remarkable homogeneity, and an average may be used.
For 1,1-DCA the average is 1000 ^g/g and for vinyl chloride it is 730 /xg/g. The mole
fraction of the compounds of interest in the waste is calculated using Equation 5-3:
(1000 iig/g)(250 g/mol) _
(10,000 ng/g)(99 g/mol) "
(730 ixg/g)(250 g/mol)
(10,000 |ig/g)(62.5 g/mol)
The emission rates can now be found from Equation 5-1:
ER (1,1-DCA)
(0-0919X1.24
ER(vinyl chloride) - (0-0900)(8.94 10*)(0.55^(0.29)(250)(1Q-") _
0.01 1
'sec
5-8
-------
To calculate the emission rates from 5-5 for the case of internal gas generation,
one uses the saturated vapor concentrations and the default value for landfill gas
velocity. The emission rates are:
ER (1,1-DCA) = (1.24xlCP)(ia12)(0.00163)(250)(ltf) = 5.1 g/sec; and
ER (vinyl chloride) = (8.94xlCP)(ia12)(0.00163)(250)(10J) = 36 g/sec.
What if headspace or soil-gas information is available, but there are no
corresponding soil concentration data? The following example illustrates how to
calculate emissions for a covered landfill with no internal gas generation.
A series of soil-gas samples were collected and found to contain an average
benzene concentration of 150,000 ug/nf (i.e., about 50 ppm). From Appendix A, the
diffusivity in air of benzene is 0.0932 cm2 /sec and its molecular weight is 78.12. The soil-
gas concentration is well below the saturated vapor concentration for benzene of 4.00x10*
ug/nf. For other variables, the same default values apply as in the example calculations
given above.
The mole fraction of benzene cannot be directly calculated from Equation 5-3
since the concentration of benzene in the soil is not known. One possbile approach is to
calculate the total mass of benzene present in the pore spaces and convert this to a
concentration in the soil. For 1 nf of soil with an air-filled porosity of 0.55, the
following mass of benzene would be present:
= (1 m3)(0.55)(150,000Mg/m3) = 82,500 ug
Given a bulk density of 1.5 g/cm3, the weight of 1 m3 of soil is:
M^ = (1 nf )(1.5 g/cnf)(ltf crrf/nf) = LSxltf g
The concentration of benzene in the soil can be estimated as:
5-9
-------
C = NWMson = 82,500 ug/LSxlO5 g = 0.055 ug/g
The mole fraction can now be calculated using Equation 5-3:
= (0.055 ug/g)(250 g/mol)/(10,000 ug/g)(78.12) = 1.76xlQ5
The emission rate can now be found from Equation 5-1 using the average measured soil-
gas concentration in place of Q:
ER(Benzene) = (0.0932)(150,000)(0.55)4/3(1.76xl05)(250)(10':)/(0.01)(l) = 2.8xl09 g/sec
The emissions of benzene are negligible. This is reasonable since benzene is present at
levels well below saturation in the soil pore spaces and, therefore, the concentration
gradient (i.e., driving force) is relatively low.
5.1.9 References
1. Pedersen, T.A. and J.T. Curtis. Handbook of Soil Vapor Extraction (SVE).
EPA/540/2-91/003. 1991.
2. U.S. EPA. Hazardous Waste Treatment, Storage, and Disposal Facilities
(TSDF) - Air Emission Models. EPA-450/3-87-026. November 1989.
3. Lyman, W.F., F.W. Reehl, and D.H. Rosenblatt. Handbook of Chemical
Property Estimation Methods. McGraw-Hill, NY. 1990.
4. Devitt, D.A., R.B. Evans, W.A. Jury, T.H. Starks, B. Eklund, and A. -
Gholson. Soil Gas Sensing for Detection and Mapping of Volatile
Organics (EPA/600/S8-87/036). National Water Well Association, Dublin,
Ohio. 1987.
5. Johnson, P.C. and R. A. Ettinger. Heuristic Model for Predicting the
Intrusion Rate of Contaminant Vapors Into Buildings. EST Vol. 25, No. 8,
pp!445-1452, 1991.
6. U.S. EPA. Superfund Exposure Assessment Manual (SEAMs).
EPA/540/1-88/001. April 1988.
5-10
-------
7. U.S. EPA. Land Disposal of Hexachlorobenzene Wastes: Controlling
Vapor Movement in Soil. EPA-600/2-80-119. 1980.
8. Shen, T. Estimating Hazardous Air Emissions From Disposal Sites.
Pollution Engineering 13 (8), pp31-34. 1981.
9. Breton, M., T. Nunno, P. Spawn, W. Farino, and R. Mclnnes. Evaluation
and Selection of Models For Estimating Air Emissions From Hazardous
Waste Treatment, Storage, and Disposal Facilities. EPA-450/3-84-020.
December 1984.
10. Radian Corporation. Procedures for Conducting Air Pathway Analyses for
Superfund Activities, Interim Final Documents: Volume 2 - Estimation of
Baseline Air Emissions at Superfund Sites, EPA-450/l-89-002a (NTIS
PB90-270588), August 1990.
5.2 LAGOONS
5.2.1 Description of Emission Process
The rate of VOC emissions from quiescent liquid surfaces will depend on the
distribution of the organic species between gas and liquid phases (Henry's Law), the
concentration of the organic species in each phase, and the mass transfer characteristics
(coefficients) of the species. The overall mass transfer coefficient is the most important
term in controlling VOC emissions. The term consists of a resistance to mass tranfer in
liquid (lq) and a resistance to mass transfer in gas (k,). For most VOCs kg » Iq and
the liquid phase resistance controls the volatilization process. For mass transfer, the
chemical and physical properties of the thin film at the liquid-air interface are of more
significance than the bulk liquid and bulk gas properties. Any factor that alters the
average overall mass transfer coefficient of the surface impoundment will alter the VOC
emission rate from that source. Wind has a major effect because the liquid phase
resistance decreases in proportion to the square of the wind velocity. High winds
therefore cause low resistance to mass transfer in the liquid phase with resulting high
emissions. High winds also cause an increase in wave activity that approximates the
activity of an aerator. The emission rate is also sensitive to any factor that increases the
5-11
-------
mixing of the bulk liquid; e.g., the residence time of liquid in the surface impoundment
and the velocity of any influent streams.
The VOC emission rate from quiescent liquid surfaces with a floating organic
layer will differ from rates from liquid surfaces without such a layer. If the floating
organic is a purely volatile material, then the rate will depend on the vapor pressure of
the VOC and the mass transfer coefficient which in turn is dependent on the wind speed
and the size of the source. If the floating organic layer is primarily a heavy oil that
contains some VOCs, then the VOC emission rate will be lower than that for quiescent
lagoons. The oil layer adds an additional resistance term to the overall mass transfer
coefficient due to mass transfer in the oil phase.
Emissions from aerated liquid surfaces are generally much higher than emissions
from non-aerated liquid surfaces. This is due to the increased surface area and the
enhancement of the gas and liquid film mass transfer coefficients. The aeration air
serves to strip out VOCs from the liquid.
522 Model Selection
As previously mentioned, the models in this section were adapted from the
Superfund Exposure Assessment Manual1 was compiled. SEAMs presents a version of
the Mackay and Leinonen modef for estimating VOC emissions from lagoons, as
simplified by Farino et al.3. These models are not applicable to aerated lagoons. For
such cases, use the bioremediation model present in Section 4.4 or use EPA's model for
aerated surface impoundments4.
52.3 Emission Model Equation
A simplified emission rate from a hazardous waste lagoon is:
ER = K C SA 0.01 (Eq. 5-7)
5-12
-------
where: ER = emission rate [g/sec];
K = overall mass transfer coefficient [cm/sec];
C = compound i's liquid-phase concentration [mg/L];
SA = surface area of lagoon [m2]; and
0.01 = conversion factor [g/mg * f/cm3. * cnf/m2].
The overall mass transfer coefficient of Equation 5-7 may be calculated as follows:
11 RT -5-8>
K kL
where: V^ = liquid phase mass transfer coefficient of compound i
(cm/sec);
R = ideal gas constant, equal to 8.2xl&5 (atm-m3/mol-°K);
T = absolute temperature (°K);
H = Henry's Law constant of compound i (atm-rrf/mol); and
kb = gas phase mass transfer coefficient of compound i (cm/sec).
In many cases, the gas-phase mass transfer coefficient is much larger than the
liquid-phase mass transfer coefficient and can be ignored; (i.e.,^- =T~ ) Default
values for K are given in Section 5.2.6.
For liquid-phase mass transfer coefficients, Equations 5-9 or 5-10 may be used.
The liquid-phase mass transfer coefficient can be estimated as follows:
(Eq. 5-9)
t - { 32 }* (JL
^ I^MWJ V298,
where: 32 = molecular weight of O^ (g/mol);
MW = molecular weight of compound i (g/mol); and
1^ 02 = liquid-phase mass transfer coefficient of Oj (cm/sec).
At windspeeds below 3.25 m/sec (about 7 mph), the liquid-phase mass transfer
coefficient can be estimated as follows:
5-13
-------
k,, = 2.78X104 [Dw/8.5xlO*]0.667 (Eq. 5-10)
where: Dw = diffusivity in water of compound i [cm2/sec]; and
8.5 x 106 = empirical factor [cm2/sec].
The gas-phase mass transfer coefficient is typically described as:
. 5-11)
where: 0.482 = empirical constant [cm/sec (m/sec)"-78 (m)11];
U = windspeed [m/sec];
S
-------
Table 3-5.
Default Values for Thermal Incinerators
Parameter
Total feed rate of
waste into unit
Destruction and
removal efficiency
Symbol
FT
DRE
Units
kg/hr
%
Default Value
5,900 (soils?
1,800 (liquids/sludges)1
99.99 (VOCs, organics)
99.9999 (dioxins, furans, and PCBs)
Expected Range
900-24,000 (soils)
90-13,600 (liquids/sludges)
Reference
8
8
Not known
Table 3-6.
Example Scenarios for Thermal Incinerators9
Parameter
Feed rate (soils)
Feed rate (liquids/sludges)
Mass of soil to be treated
Gas Volume1"
Stack Height
Stack Diameter
Exit Gas Velocity
Exit Gas Temperature0
Units
kg/hr
kg/hr
kg
nf/min
cfm
m
m
m/sec
°C
System
Small
900-1,800
90-550
<4.5xltf
50-150
1,800-5,000
6
0.3
7
70
Medium
3,600-8,200
430-3,200
4.5xltf-2.7xlti7
150-280
5,000-10,000
8
0.5
10
70
Large
9,100-24,000
900-13,600
> l.SxlO7
280-710
10,000-25,000
20
1
20
70
'Assumes a rotary kiln incineration system.
bGas volume assumes dry standard conditions at 7% O^.
'Assumes a quench and wet scrubbing system at adiabatic saturation for the stack gas.
SOURCE: Reference 8
3-27
-------
The underlying principle behind Equation 3-5 is that the minimum DRE standard
will be met. Although it is possible that the DRE might be surpassed, this will only
cause the model to err on the conservative side. The required DRE may not be met
during process-upset conditions and when the waste feed composition differs significantly
from the waste used in any trial burns used to develop the standard operating conditions.
3.3.8 Example Calculations
Consider the following remediation scenario. The soil in a hypothetical site has
been tested, and it contains:
PCBs
1,2,4-Trichlorobenzene
2%
2800 ppb
The contractor will use a rotary kiln incinerator with a feed rate of 6000 kg/hr. The
device bums propane, which is assumed to not contribute measurably to the emissions of
any of the above compounds. The exit gas flow rate is not known.
First, find the mass flow rate of contaminants into the incinerator from Equation
3-6 (note that 2% = 20,000 ppm, and that 2800 ppb = 2.8 ppm):
Fpcu = 106 x 6000 x 20,000 = 120 kg/hr;
FTCB = 106 x 6000 x 2.8 = 0.017 kg/hr.
Next calculate the organic emissions using Equation 3-5 and a DRE for PCBs of
99.9999%, and a DRE for TCB of 99.99%.
= 0.278 x 120 x (1 - 99.9999/100) = 3.3 x 1Q5 g/s; and
ERTCB = 0.278 x 0.017 x (1 - 99.99/100) = 4.7 x 1Q7 g/s.
3-28
-------
The emission rate should be compared to the total mass of the contaminant of
interest to ensure that the estimated emissions over some time period do not exceed the
total mass that is present. The total mass can be calculated as follows:
M = C * SA * 1 * 1.0 (Eq. 5-15)
where: M = total mass of contaminant [g];
1 = depth of lagoon [m]; and
1.0 = conversion factor [g * L / mg * m3]
5.2.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS) to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
minimum field data required to estimate emissions from lagoons are: are:
Surface area of emitting area;
Specific contaminants present in the lagoon; and
Average contaminant concentration in the lagoon.
52.5 Sources of Input Data
The only site specific data required to estimate emissions of VOCs from a lagoon
are the concentration of the contaminants of interest in the lagoon, the surface area of
the lagoon, and the wind speed (at a height of 10m). The remaining variables are
generally available in Appendix A to this report. Additional information is given in
References 4 through 7 on how to calculate mass transfer rates and estimate -infra*1--~
parameters.
5.2.6 Default Values for Input Variables
Table 5-2 contains default values needed for estimating VOC emissions from
lagoons. Data for the viscosity and density of air at temperatures other than 25° C can
be found in Reference 6; estimation methods can be found in Reference 5. Table 5-3
contains default values for the overall mass transfer coefficient.
5-15
-------
Table 5-2.
Default Values for Estimating VOC Emissions From Lagoons
Parameter
Surface Area
Temperature
Henry's Law Constant
Liquid-phase mass transfer
coefficient of oxygen
Diffusivity in water
Wind speed
Viscosity of air
Density of air
Diffusivity in air
Molecular weight
Gas-phase mass transfer coefficient
of water vapor
Symbol
SA
T
H
w
Dw
U
P»
Po
Da
MW
H3.H2O
Units
m2
°K
atm-
nf/mol
cm/sec
cm2 /sec
m/sec
g/cm-sec
g/cnf
cnr/sec
g/mol
cm/sec
Default Value
4,050 (1 acre)
298
See Appendix A
0.002
See Appendix A
2.0
1.81 x 104
1.21 x 103
See Appendix A
See Appendix A
0.83
Expected
Range
4050 - 20,200
278-328
SxlQ4 -0.02
0-4.47
0.0170-0.0185
1.1034-1.2928
-
0.0047 - 1.4
Reference
Author's
estimate
Author's
estimate
Author's
estimate
Author's
estimate
4,6
4,6
8
Table 5-3.
Default Values for Overall Mass Transfer Coefficient
Parameter
Overall mass transfer coefficient
(by process type)
Overall mass transfer coefficient*
(by compound)
Symbol
K
K
Units
cm/sec
cm/sec
: :pefauft Value ;=
4.2 x 10"4 (quiescent surface)
0.077 (turbulent surface, mechanically altered)
0.34 (turbulent surface, diffused aeration)
1.9 x ID'3 (MW = 32)
1.3xlO-3(MW = 64)
9.5 x 10-4 (MW = 128)
6.8 x 1O4 (MW = 256)
A; Reference
4
4
4
Author's estimate
Author's estimate
Author's estimate
Author's estimate
'Assuming H = 1 x
5-16
-------
For purposes of dispersion modeling, the default one-acre lagoon or surface
impoundment from Table 5-2 can be assumed to have dimensions of 64m by 64m and a
release height of 1m. As always, site-specific values should be used if available.
52.1 Model Assumptions/Sensitivity Analysis
The emission rate of Equation 5-7 assumes a low-solubility contaminant at steady-
state conditions. The liquid-phase concentration of the contaminant is assumed to be
uniform throughout the lagoon, the air/water interface is assumed to be stagnant, and
the air-phase concentration of the contaminant is assumed to be negligible. If the last
assumption is invalid, then the original equations of Mackay and Lienonen2 should be
used.
It is not recommended that temperatures other than 298 ° K (25° C) be used in
equations 5-8 and 5-9, unless the Henry's Law constants and mass transfer coefficients
are also adjusted for temperature.
5.2.8 Example Calculations
A 2 acre body of water is contaminated with 3 ppm methylene chloride
(dichloromethane).
To find the emission rate from Equation 5-7, the concentration will have to be
expressed in units of mg/L. For dilute solutions, 1 ppm = 1 mg/L. The above
concentration is thus 3 mg/L. Also, the area must be converted to m2 using 1 acre =
4,046 m2, so 2 acres equals about 8,100 m2.
The overall mass transfer coefficient must be calculated using Equation 5-8. The
Henry's Law constant from Appendix A is 0.00319 atm-nf/mol. The liquid-phase and
gas-phase mass transfer coefficients are found using Equations 5-10 and 5-11,
respectively. Default values of windspeed, viscosity of air, and density of air are
5-17
-------
assumed. Additional inputs needed for these equations are obtained from the
appendices:
Dw = 1.17xlCT5 cm2/sec;
Da =0.117 cm2/sec;
The Schmidt number and effective diameter of the lagoon are thus:
Sc = (1.81 x 1Q-4) =12S
° (1.21 x 10-3)(0.117)
de = [1.27 8,100]05 = 101 m
Using these inputs, the mass transfer coefficients are:
ki. = (2.78xl04)(1.17xl05/8.5xlO-6)0667 = 3.44X104 cm/sec; and
kc = (0.482)(2°'78)(1.28-0667) (lOr0-11) = 0.423 cm/sec.
As expected, the liquid-phase mass transfer coefficient is the rate limiting step in the
overall mass transfer process. The overall mass transfer coefficient, and thus emission
rate, can now be found:
_L - 1 + (8.2 x IP"5 x 298) =
K " 3.44 x lO'4 + (0.00319)(0.423)
or K = = 3.4 x 10"4 cm/sec
2920
ER - f^L\3)(8,100)(0.01) = 0.083 g/sec
This is the estimated initial emission rate. The rate will decrease over time.
5-18
-------
52.9 References
1. U.S. EPA. Superfund Exposure Assessment Manual (SEAMs).
EPA/540/1-88/001. April 1988.
2. Mackay D. and Leinonen, PJ. Rate of Evaporation of Low-Solubility
Contaminants to Atmosphere. Environ, Sci. Tech. Vol 9(13). 1975.
3. Farino, W., Spawn, P., Jasinski, M., and Murphy, B. Evaluation and
Selection of Methods for Estimating Air Emissions From Hazardous
Waste Treatment, Storage, and Disposal Facilities. Revised Final Draft
Report to US EPA Office of Solid Waste, Land Disposal Branch. Contract
No. 68-02-3168. 1983.
4. U.S. EPA. Hazardous Waste Treatment, Storage, and Disposal Facilities
(TSDF) - Air Emission Models. EPA-450/3-87-026. November 1989.
5. Lyman, W.J., W. F. Reehl, and D.H. Rosenblatt. Handbook of Chemical
Property Estimation Methods. American Chemical Society, Washington,
D.C., 1990.
6. DeWolf, G. and R. Wetherold. Protocols for Calculating VOC Emissions
From Surface Impoundments Using Emission Models - Technical Note.
Report to EPA/OAQPS/EMB under EPA Contract No. 68-02-3850,
WA17. December 1984.
7. Sharp-Hansen, S. Available Models for Estimating Emissions Resulting
from Bioremediation Processes. A Review EPA/600/3-90/031 (NTIS
PB90-228610). March 1990.
8. Springer, C, K. Valsaraj, L. Thibodeaux, and P. Lunney. Emission of
Hazardous Chemicals from Surface and Near Surface Impoundments to
Air. Draft Report to Steve James, U.S. EPA, Cincinnati. June 1991.
5.3 SPILLS, LEAKS, AND OPEN WASTE PITS
5.3.1 Description of Emission Process
The emission processes from contaminated surface soils are intermediate in
nature between the applicable processes for landfills and for lagoons. Surface
contamination due to spills, leaks, or landtreatment results in areas of pooled waste both
on and below the soil surface. The pooled waste quickly evaporates or percolates down
5-19
-------
through the soil. The majority of the contamination becomes adsorbed onto the surface
of soil particles. The emission rate is usually assumed to be controlled by the diffusion
rate in the air pore space when the waste loading and soil particles are both small. In
this case, the emission rate is controlled in the same manner as for covered landfills and
the same considerations apply. If, however, the surface soils are tilled or otherwise
disturbed, large increases in emissions will occur. This is a result of the contaminants
being redistributed so that the depleted near-surface soil layer receives additional waste
material. Soil disturbances also expose moist subsurface layers which leads to loss of
moisture over time with a resulting increase in the air-filled porosity of the soil. -
5.3.2 Model Selection
The Superfund Exposure Assessment Manual1 uses a simplified version of the
Thibodeaux and Hwang mode? for estimating VOC emissions from spills, leaks, and
open (uncovered) waste pits. In this model, the emission rate is controlled by the rate of
diffusion through the soil. The contaminant concentration in the soil is assumed to
remain constant and the contaminant release occurs by the loss of successive
unimolecular layers of contaminant from the surface of the wet, or contaminated, zone.
Over time, this process is assumed to result in a dry zone on increasing depth at the soil
surface and a wet zone of decreasing depth below the dry zone.
5.3.3 Emission Model Equation
A "fresh" spill is a site with saturated surface soil. The emission rate is given by:
ER = ka Q, SA 104 (Eq. 5-16)
where: ER = emission rate of contaminant i (g/sec);
k~ = gas phase mass transfer coefficient of i (cm/s);
Q = vapor concentration of compound i in soil pore spaces
(g/cnf);
SA = contaminated surface area (m2); and
101 = conversion factor (crrf/nf).
5-20
-------
The emissions model for "old" spills is the same used for uncovered landfills and
open waste pits:
2 D C SA 104 (Eq. 5'17)
TTT> _ ~
where:
ER
D
Q
Q
SA
Itf
2 D
r
average emission rate of compound i over time [g/sec];
phase transfer coefficient [cm2/sec];
liquid-phase concentration of i in soil [g/cm3];
bulk contaminant concentration of i soil [g/cm3 ];
contaminated surface area [m2];
conversion factor [cnf/m2];
depth of dry zone at sampling time [cm]; and
time since sampling occurred [sec].
If the contaminant has been entirely absorbed into the soil surface, soil phase
mass transfer resistance is expected to be important and Equation 5-17 should be used.
Note that the emission rate at the time of sampling (t = 0) is given by a simpler
equation:
D q SA lo4
An expression for the phase transfer coefficient, D, of Equation 5-17 is:
(Eq. 5-19)
D =
Da E/ H
-* = -
RT
where:
D,
E.
H
R
T
diffusion coefficient of compound i in air [cm2/sec];
total soil porosity [unitless];
Henry's Law constant of compound i [atm-nf/mol];
ideal gas constant, equal to 8.2xlQ5 [atm-nf/mol-'K]; and
absolute temperature [°K].
The time required from the last measurement until this point is 1^, and may be
calculated from:
(Eq. 5-20)
5-21
-------
,-fejl.fS
2D C
where: ^ = time between measurement and total volatilization [sec];
2
lj = depth from soil surface to bottom of contaminated region [cm].
5.3.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS)to provide the
input data necessary to generate reasonably accurate estimates of air emissions. The
minimum field data required to estimate emissions from spill or leak sites are:
Surface area;
Specific contaminants present in the soil; and
Average contaminant concentration in the soil.
53.5 Sources of Input Data
Uncovered landfills, spills, and open waste pits as modeled in Equations 5-16 and
5-17 require some inputs previously described in Sections 5.1 and 5.2. Procedures for
determining the gas-phase mass-transfer coefficient were given in Equations 5-11 and
5-14 of Section 5.2.3. The vapor concentration can be measured in the field or assumed
to be equal to the saturated vapor concentration as shown in Equation 5-2 of Section
5.1.3. An equation for calculating the air-filled porosity of soil was given as Equation 5-4
in Section 5.1.3.
5.3.6 Default Values for Input Variables
Default values for equations 5-16 and 5-17 are given in Table 5-4. Contaminant
concentrations, depths, and areas of contamination all spill sites will vary greatly from
site to site, and default values are not appropriate.
5-22
-------
Table 5-4.
Default Values for Estimating VOC Emissions from Spill Sites
Parameter
Vapor concentration in soil pore spaces
Bulk density of soil
Diffusivity in air
Air-filled porosity of soil
Henry's Law Constant
Temperature
Gas-phase Mass Transfer Coefficient
Symbol
Cv
&
Da
Ea
H
T
ko
Units
g/cm3
g/cm3
cm2/sec
Unitless
atm-m3/mol
°K
cm/sec
Default Value
Cg 10'12
(see Appendix A)
1.5
(see Appendix A)
0.55 dry, uncompacted soil;
0.35 wet or compacted soil;
0.05 sludge
(see Appendix A)
298
0.15
Expected
Range
-
1.0-2.0
-
~~
-
278 - 328
0.0062-0.52
Reference
-
3
-
3
3
*
-
*
4
'Author's estimate
5-23
-------
For purposes of dispersion modeling, the emission source should be assumed to
be a square with a release height of 1m.
5.3.7 Model Assumptions/Sensitivity Analysis
Equation 5-16, the "fresh" spill model, is only valid for short term emission rates.
The assumption that the contaminant concentration remains constant until all of it has
volatilized, results in an overprediction of the emission rate since an exponential decay in
emission rate will more likely occur. The Q term in Equation 5-16 assumes a single
component spill (i.e., mole fraction equals one).
It is not recommended that temperatures other than the default temperature be
used unless the Henry's Law constant and other variables that are temperature
dependent are also adjusted.
53.8 Example Calculations
A site has just been contaminated with a spill of hexane and phenol. An area of
approximately 1/4 acre (1,000 m2) is saturated with the two contaminants. No other
information is available. The emissions can be calculated using Equation 5-16. From
Table 5-3, a default value for l^ is 0.15 cm/sec and a default value for Q is Q x 10~12.
From Appendix A, the saturated vapor concentration (Q) for hexane is 6.96xltf ug/nf
and it is 1.72X105 ug/m3 for phenol. Inserting these various values into Equation 5-16
yields:
ERHex^c = (0.15)(6.96xltf)(ia12)(1000)(10t) = 1000 g/sec
ERpheno. = (0.15)(1.72xl05)(lQ12)(1000)(10i) = 0.26 g/sec
After six months, the spill has spread and is now found to cover an area of 1/2
acre (2,000 m2). It no longer fits the criteria of a "fresh" spill- the liquid is not standing
on the surface of the soil, and no longer saturates it. Preliminary sampling indicates a
concentration of 67 /*g/g phenol and 333 /ig/g hexane underneath an average "dry zone"
5-24
-------
that is 10 cm deep. Assuming a bulk density of 1.5 g/cm3, the contaminant
concentrations can be converted to units of g/cm3 as previously shown in Equation 4-5:
67 Mg/g * 1.5 * 1Q6 = l.OOxlO4 g/cm3
333 /tg/g * 1.5 * 106 = 5-OOxlO4 g/cm3
This is assumed to equal the liquid phase concentration in the soil. The phase transfer
coefficient is calculated using values of diffusivity in air and Henry's Law constant from
Appendix A along with the default temperature (298° K) and the air-filled porosity.
Inserting these values into Equation 5-19 yields:
D (phenol) = (0.0820)(0.554/3)(4.54xia7) / (8.2xia5)(298) = 6.87xltt7
D (Hexane) = (0.2000) (0.55473) (0.200) / (8.2xia5)(298) = 0.738
To find the emission rate at the time the soil samples were collected, use Equation 5-18:
ER (phenol) = 6.87xl(T7 * l.OOxlO4 * 2000 * 10* / 10 = 1.4X104 g/sec.
ER (hexane) = 0.738 * S.OOxlO4 * 2000 * itf / 10 = 740 g/sec.
5.3.9 References
1. U.S. EPA. Superfund Exposure Assessment Manual (SEAMs).
EPA/540/1-88/001. April 1988.
2. Thibodeaux, LJ. and Hwang, S.T. Landfarming of Petroleum Waste--
Modeling the Air Emission Problem. Environ. Progress Vol. 1(1). 1982.
3. Schultz, H.L., et al. Superfund Exposure Assessment Manual.
EPA/540/1-88/001. April 1988.
4. Eklund, B.M., S. Smith, and A. Hendler. Estimation of Air Impacts for the
Excavation of Contaminated Soil. EPA-450/ 1-92-004. March 1992.
5-25
-------
SECTION 6
PARTICULATE MATTER, METAL, ACID GAS,
AND PRIORITY POLLUTANT EMISSIONS FROM POINT SOURCES
Simple air emission estimation procedures are presented in this section for two point
sources of paniculate matter (PM) and metal emissions, namely: thermal destruction units
(incinerators) and thermal desorption units. The same format is followed for each source. A
brief description of the emission process is given, followed by a discussion of available air
emission models. The model selected for inclusion in this manual is then presented along
with sources of input data and default values for each of the input variables of the selected
model. The model assumptions are then briefly discussed. Finally, an example calculation
is shown and references are listed.
6.1 THERMAL DESTRUCTION
6.1.1 Description of Emission Process
Thermal destruction, also known as thermal treatment, high-temperature thermal
treatment, thermal oxidation, or incineration, is a very commonly used remediation at
Superfund sites. Its primary advantage is that it destroys toxic organic compounds.
However, a disadvantage is the generation of acid gases from halogenated organic
compounds and the creation of toxic products of incomplete combustion (PIC) such as
dioxins. Further, toxic metals in the waste are not controlled by incineration, so some
emissions will occur.
Several types of incinerators are now in use at Superfund sites, but rotary kiln designs
are the most common. The emission estimation methods presented in this section are valid
for any design, but default values given are valid only for rotary-kiln incinerators. The
procedures provide estimates of uncontrolled emissions. In actual practice, however, control
devices for paniculate matter, etc. are always used. The procedures for estimating
uncontrolled emissions given in this section may be useful for estimating the required
removal efficiency of a proposed control device or for estimating the size and cost of a
6-1
-------
control device capable of achieving a given removal efficiency. Information on the
effectiveness of various control devices is available1-2-3'4. The primary inorganic
contaminants emitted from these types of incinerators are discussed below.
Particulate Matter
The waste feed, auxiliary fuel, and combustion air can all serve as sources for
particulate matter emissions from an incineration system. Particle emissions may result from
inorganic salts and metals that either pass through the system as solids or vaporize in the
combustion chamber and recondense as solid particles in the stack gas. High molecular
weight hydrocarbons may also contribute to particulate matter emissions if oxidation is not
complete. Particulate matter (PM) emissions are sensitive to operating conditions, and are
affected by waste composition, feed rate, PM size distribution within the waste, and
incinerator design. A conservative estimate of PM emissions may be reached by assuming
the maximum allowable under RCRA regulations, which is 0.08 grains/dscf (0.18 g per dry
standard m3) corrected to 7% oxygen in the stack gas.
Metals
Toxic metals may be present within the waste feed or the soils themselves. The 10
metals identified by the US EPA5 as hazardous to humans or the environment are: antimony,
arsenic, barium, beryllium, cadmium, hexavalent chromium, lead, mercury, silver, and
thallium. Four (arsenic, beryllium, cadmium, hexavalent chromium) are known or suspected
carcinogens. The Clean Air Act Amendments address emissions of four additional metals:
cobalt, manganese, nickel, and selenium.
The emission rate of a metal is affected by the partitioning of the metal within the
combustion chamber. Metals in the waste feed will either remain in the solids and be
discharged in the bottom ash, or they will be vaporized and carried out by combustion gases.
Conservative estimates of metals partitioning in incinerators have been developed by the
EPA, based on actual testing.
6-2
-------
Acid Gases
Halogens or sulfur in the waste stream (whether elemental or in compound form) will
result in the production of acid gases during incineration. The acid gases of primary interest
are hydrogen chloride (HC1), hydrogen fluoride (HF), hydrogen bromide (HBr), and sulfur
dioxide (SOJ. The content of halogens and sulfur in the waste and the fuel feed determine
the uncontrolled emission levels of their respective acid gases. The concentrations of these
elements range widely for different waste types; consequently, the resulting levels of acid gas
emissions will also show wide variability. Emissions of free chlorine (Cy from incinerators
may also be possible if there is insufficient hydrogen available to react with all of the-
chlorine present in the off gases.
Other Pollutants
Achieving high levels of destruction of organic wastes is directly related to
combustion chamber temperature: the higher the temperature, the greater the destruction and
removal efficiency for organic compounds. Unfortunately, the fixation of nitrogen and
oxygen to form oxides of nitrogen (NOJ also increases with combustion temperature. NOX
emissions caused by this mechanism are referred to as thermal NOX. Additional NOX
emissions, called fuel NOX, will be formed if there are bound nitrogen atoms in the waste
(e.g., amine compounds). The rate of NOX formation will depend on fuel the firing rate, the
amount excess oxygen, combustion temperature, and other operational controls.
Carbon monoxide (CO) monitoring of stack gases is a regulatory requirement for
incinerators under RCRA. If the required DREs for organic compounds are met, CO
emissions will generally be low (< 100 ppmv) due to the high operating temperatures and the
excess oxygen maintained in the process.
6.1.2 Model Selection
No applicable detailed models for estimating non-VOC emissions from incinerators
had been identified in the preparation of this report. Eklund et al.3 have developed simple
mass-balance equations to estimate incinerator PM emissions. In another EPA study, IT
6-3
-------
modified these equations slightly4, and this latter reference contains the model used here.
Both documents summarize typical operating conditions, feed rates, etc. , and also address
metal and other inorganic emissions.
6.1.3 Emission Model Equation
The models presented below for PM, metals, acid gases, and priority pollutants are
all for uncontrolled emission rates.
Equation 6-1 is a simple mass-balance formula which gives a conservative estimate of
PM emissions from an incinerator based on the assumption that the emissions will be the
maximum allowable under RCRA regulations:
ER = 0.18Q (Eq. 6-1)
where: ER = emission rate of paniculate matter [g/sec];
0.18 = maximum allowable PM emissions in stack [g/dscm]; and
Q = exit gas flow rate [dry standard mVsec].
Equation 6-2 provides an emission rate estimate for metals:
ER = 0.278 F
where: ER = emission rate of metal [g/sec];
0.278 = conversion factor [g/sec / Kg/hr];
Fm = feed rate of metal [kg/hr]; and
PF = partitioning factor of metal [%].
To calculate emission rates of acid gases, a conservative assumption may be made: if
all of the acid-forming element in the halogenated compounds reacts with hydrogen present
in the combustion chamber, the stoichiometric ratio will predict the amount of acid gas
produced. Thus:
ER = 0.278 Fa r (Eq. 6-3)
6-4
-------
where: ER = emission rate of acid gas i [g/sec];
0.278 = conversion factor [g/sec / kg/hr];
F. = feed rate of halogen or sulfur [kg/hr]; and
r = stoichiometric ratio of acid-gas-to-element [unitless].
Stoichiometric ratios of acid-gas-to-element for several common elements are given in
Section 6.1.5. The stoichiometric ratio for sulfur dioxide to sulfur is also included; the
partitioning of sulfur to sulfuric acid and sulfur dioxide is not known, so it is assumed that
all sulfur is converted to SO2, and no H2SO4 is formed and no sulfur leaves in the bottom
ash.
The feed rates of elements in the waste stream may be calculated from Equations 6-4
and 6-5 if the total feed rate and concentration of compounds containing acid-forming
elements is known. An analogous equation can be used for estimating the feed rate of metal
species.
F. = FT Ca 10-* (Eq. 6-4)
where: FT = total feed rate [kg/hr];
Ca = concentration of acid in feed rate [>g/g]; and
10"6 = conversion factor [g//tg].
The total feed rate will depend on whether the waste being treated is a solid or liquid. If
sulfur is present in the auxiliary fuel, the feed rate of sulfur from the fuel must be added to
the feed rate of sulfur from the waste to obtain the total feed rate of sulfur.
(Eq. 6-5)
where: C. = concentration of acid-forming element in the waste
Q = concentration of compound i containing above element [/*g/g];
MWa = molecular weight of acid-forming element [g/mol]; and
= molecular weight of compound i containing above element
[g/mol].
6-5
-------
If an ultimate analysis of the waste has been performed, the concentration of each element
will be known and the use of Equation 6-5 will not be necessary.
6.1.4 Minimum Requirements for Field Data
The minimum field data required to estimate emissions from thermal treatment
systems are:
Specific contaminants present in the soil or waste to be treated;
Average contaminant concentration in the soil or waste; and
Maximum contaminant concentration in the soil or waste.
6.1.5 Sources of Input Data
The preferred source of input data for Equation 6-1 through 6-4 is field measurements
for the thermal destruction system of interest. At the very least, field data should be
obtained regarding the specific contaminants present in the material to be treated and their
average and maximum concentration. Values for the flow rate of material to the incinerator
and the efficiency of any control devices may be obtained from design specification
documents and blueprints or from field measurements. Once the incineration unit is in
operation, stack sampling of emissions from the system can be performed to confirm the
emission estimates.
6.1.6 Default Values for Input Variables
Table 6-1 gives default values to be used in the emission estimation equations for a
typical rotary-kiln incinerator. Table 6-2 lists some conservative estimates of metals
partitioning in incinerators for several metals. Table 6-3 presents information regarding
stack parameters for a relatively small incinerator to assist in the prediction of downwind
ambient air concentrations using an EPA-approved air dispersion model.
6-6
-------
Table 6-1.
Default Values for Estimating Emissions from Incinerators
Parameter
Exit gas flow rate
Total feed rate
Stoichiometric ratio
Symbol
Q
FT
r
Units
dsnf/sec
kg/hr
unitless
Default Value
4.8f
5,000 solid^
1,500 liquids*
1.013 HBr/Br
1.028 HC1/C1
1.053 HF/F
1.998 SQz/S
Expected Range
0.83 - 11.8
900-24,000 solids
90-13,600 liquids/sludges
"Assumes 15,000,000 Btu/hr capacity.
"Equals 50,000 ACFM at 2,200° F
SOURCE: Reference 4
Table 6-2.
Default Values for Estimating Metal Partitioning
Parameter
Metal Partitioning
Symbol
PF
Units
%
Default Value
100 (liquids)
5 (solids)
100 (solids)
Comments
Beryllium
Chromium
Antimony
Arsenic
Barium
Cadmium
Lead
Mercury
Silver
Thallium
SOURCE: Reference 4
Table 6-3.
Stack Parameters for Rotary Kiln Incinerators
Parameter
Physical stack height
Stack diameter
Exit velocity
Exit temperature"
Range
6 -20m
0.3 - 1.0 m
1400 - 4000 ft/min
7 - 20 m/s
150° - 180° F
338° -355°K
: Default VaSlte ;;
8m
0.5m
2000 ft/min
10 m/s
160° F
344° K
"Assumes a quench and wet scrubbing system at adiabatic saturation of the stack gas.
Assumes an exit gas rate of 3700 dscf/min or 1.7 dsnf/sec).
SOURCE: Reference 4
6-7
-------
6.1.7 Model Assumptions/Sensitivity Analysis
The emission estimation equation assumes the waste material is fed into the
process unit at a constant rate and that the material is uniformly contaminated. The
former assumption is reasonable, but the waste material will certainly have a degree of
variability in the contaminants present and their concentrations.
The model's assumption of partitioning constants for metals is also conservative.
For an added level of conservativeness, all chromium in the exit gas can be assumed to
be in the hexavalent state.
For PM emissions, the assumption that the minimum regulated standard will met,
but not exceeded, is obviously a conservative assumption.
6.1.8 Example Calculations
A site is to be remediated using incineration to destroy organic compounds that
are present in soil. An ultimate analysis of the soil shows it to contain: 1.0% S, 0.5% Cl,
0.15% Ba, and 0.08% Pb. The contractor will use a rotary kiln incinerator with a feed
rate of 1000 kg/hr. The device burns propane, which is assumed to not contribute
measurably to the emissions of any of the above inorganic compounds. The exit gas flow
rate is not known.
For PM emissions, Equation 6-2 requires an exit gas flow rate. Since this is not
known, the default value will be used but it will be adjusted for the known feed rate.
Since in Table 6-1 a feed rate of 5,000 kg/hr yields a flow rate of 4.8 nf/sec, this
incinerator will have an exit gas rate of:
(1,000/5,000) x 4.8 = 0.96 dsnf/sec
The PM emissions are then:
ER = 0.18 x 0.96 = 0.17 g/sec.
6-8
-------
The metals emission rates are found in a similar manner. The feed rate of the metals
are found by multiplying the total feed rate by the percentage of metal in the waste:
FB. = 1000 x 0.15/100 = 1.5 kg/hr; and
Fpt, = 1000 x 0.08/100 = 0.8 kg/hr.
The partitioning factor for both metals is 100%. The emission rates are then:
= 0.278 x 1.5 x 100/100 = 0.42 g/sec; and
= 0.278 x 0.8 x 100/100 = 0.22 g/sec.
The feed rates of Cl and S are:
Fa = 1000 x 0.5/100 = 5.0 kg/hr; and
Fs = 1000 x 1.0/100 = 10 kg/hr.
Their stoichiometric ratios, from Table 6-1, are 1.028 for HC1 and 1.998 for
EHQ = 0.278 * 5.0 x 1.028 = 1.4 g/sec; and
£502 = 0.278 * 10 x 1.998 = 5.5 g/sec.
6.1.9 References
1. Eklund, et al. Control of Air Toxics at Superfund Sites. Report to EPA's
Center for Environmental Research Information. June 1992.
2. Eklund, et al. Air Emissions From the Treatment of Soils Contaminated
With Petroleum Fuels and Other Substances. EPA-600/R-92-124. July
1992.
3. U.S. EPA. Air/Superfund National Technical Guidance Study Series,
Volume IE: Estimation of Air Emissions from Cleanup Activities at
Superfund Sites. Report No. EPA-450/1-89-003. NTIS PB89 180061/AS.
January 1989.
4. IT Corp. Screening Procedures For Estimating the Air Impacts of
Incineration at Superfund Sites. EPA Contract No. 68-02-4466, WA 91-77.
September 1991.
5. U.S. EPA. Technical Background Document: Control of Metals and HC1
Emissions from Hazardous Waste Incinerators. August 1989.
6-9
-------
62 THERMAL DESORPTION
Little information about emissions of non-organic compounds from thermal
desorption units is available. Therefore, it is recommended that the procedures for
incineration be used with slightly different default values in some cases.
62.1 Description of Emission Process
Thermal desorption is distinguished from thermal incineration chiefly by the
operating conditions of the equipment. The operating temperatures are much lower, so
the fraction of metals that partition to the vapor phase is lower. Given the lower
temperatures, the formation of NOX should be less of a concern. Also, the volume of
exit gas may be somewhat smaller, and if so, there will be less paniculate matter carry
over.
6.22 Model Selection
No applicable detailed models for estimating non-VOC emissions from thermal
desorption units had been identified in the preparation of this report. Eklund et al.1
have developed simple mass-balance equations to estimate incinerator PM, metal, and
inorganic gas emissions. In another EPA study, IT modified these equations slightly2,
and this latter contains the model used here. Although written for incinerators, these
mass balance equations are general in nature, and are applicable to desorption as to
incineration.
62.3 Default Values for Input Variables
Table 6-4 contains feed rates and other default parameters applicable to thermal
desorption units3. The feed rate and exit gas values are based on a single system; the
use of process-specific data is strongly recommended. For all other parameters, the
default values given in Section 6.1.6 for incinerators should be used.
6-10
-------
Table 6-4.
Default Values for Estimating Emissions from Thermal Desorption Units
Parametef
Total feed rate
Exit gas flow rate
PM Loading in
Stack Emissions
Metal partitioning
Symbol
FT
Q
PF
Units
kg/hr
dsnf/sec
g/dscm
g/dscm
%
Default Value
27,200
8.8
0.46*
0.08=
100 - Mercury
20 - Lead
10 - Beryllium
10 - Chromium
10 - Copper
10 - Iron
10 - Zinc
Expected Range
2,700 - 90,800
1.8 - 16.2
0.01 - 0.17
Reference
2
3
3
3
4
'Use default values given in Section 6.1.6 for incinerators for all other parameters.
bAsphalt plant
cRotary dryer
SOURCE: Reference 3
6-11
-------
6.2.4 References
1. U.S. EPA. Air/Superfund National Technical Guidance Study Series,
Volume III: Estimation of Air Emissions from Cleanup Activities at
Superfund Sites. Report No. EPA-450/1-89-003. NTIS PB89 180061/AS.
January 1989.
2. IT Corp. Screening Procedures For Estimating the Air Impacts of
Incineration at Superfund Sites. EPA Contract No. 68-02-4466, WA 91-77.
September 1991.
3. Eklund, et al. Air Emissions From the Treatment of Soils Contaminated
With Petroleum Fuels and Other Substances. EPA-600/R-92-124. July
1992.
4. de Percin, P. (EPA). Personal communication from Paul de Percin to
Bart Eklund of Radian Corporation. August 1992.
6-12
-------
SECTION 7
PARTICULATE MATTER AND METAL EMISSIONS FROM AREA SOURCES
Simple air emission estimation procedures are presented in this section for area
sources of particulate matter (PM) and metals, including: materials handling and other area
sources such as solidification/stabilization (S/S), storage piles, and dry surface
impoundments. The same format is followed for each source. A brief description of the
emission process is given, followed by a discussion of available air emission models. The
model selected for inclusion in this manual is then presented along with sources of input data
and default values for each of the input variables of the selected model. The model
assumptions are then briefly discussed. Finally, an example calculation is shown and
references are listed.
7.1 MATERIALS HANDLING
Emission estimation procedures are given below for transfer operations, waste
mixing, grading, and traffic on paved and unpaved roads.
7.1.1 Description of Emission Process
Materials handling is a very common source of particulate matter emissions at
Superfund sites; excavation of soils, soil transport, dumping and formation of soil storage
piles, and grading are all routinely performed. The PM emissions arising from these
operations should be evaluated, whether the material is contaminated or not since PM
emissions (less than 10 microns in diameter) are a criteria pollutant.
7.1.2 Model Selection
Few emissions models for PM from materials handling exist which meet the criteria
of this manual. Compilations1'2-3 of such models produced by the EPA have themselves been
produced by Cowherd et al.4. This latter document contains a comprehensive collection of
7-1
-------
empirically based screening models and was used as the principal source of all models in this
section as well as Section 7.2.
7.1.3 Emission Model Equation
The emissions of PM from all transfer operations - adding to or removing from piles,
conveyor belts, truck dumping - are expressed in Equation 7-1:
u V-3 (Eq" 7"1}
p _ \*"")
where: E = emissions [g];
particle size multiplier [unitless];
empirical constant [g/Kg]; and
mass of waste handled [Kg];
mean wind speed [m/sec];
empirical constant [m/sec]; and
percent moisture content [%].
E
k =
0.0016 =
M
U
2.2 =
Reference 1 provides a more detailed equation for this same activity that takes into account
the drop height, the silt content of the material, and the capacity of the dump bucket. The
particle size multiplier, k, for several sizes of particles are:
size
<50um
< 30 urn
<15 urn
< 10 urn
<5 um
<2.5 um
multiplier
1.0
0.74
0.48
0.35
0.20
0.11
7-2
-------
For emissions from the erosion of intermittently active piles, use erosion equation 7-9
from Section 7.2 for each period between activity; use the above equation during the activity
itself.
For emissions during materials handling involving mixing and tilling (waste
incorporation and cultivation), a simple model is:
E = k (0.00538) SA 10* (s)06 (Eq. 7-2)
where: E = emissions [g];
k = particle size multiplier (0.21 for PM10) [unitless];
0.00538 = empirical constant [g/hectare];
10"4 = conversion factor [hectare/m2];
SA = area treated [m2]; and
s = percent silt content [%].
If wastes or soil are being graded by a bulldozer or any other tractor with a blade, then the
following equation should be used to predict the PMj0 (paniculate matter of less than 10
microns) emissions:
where: ER = PM10 emission rate [g/sec];
0.094 = empirical constant [g/sec];
s = percent silt content [%]; and
= percent moisture content [%].
The emission rate of traffic on paved roads in grams per vehicle kilometers traveled
(VKT) is given by Equation 7-4.
(Eq- 7-4)
/sL\o.3 ^^ '
EF = 220
(12)
where: EF = PM10 emission factor [g/VKT];
220 = empirical constant [g/VKTJ;
sL = silt surface loading [g/m2];
12 = empirical constant [g/m2]; and
0.3 = empirical constant [unitless].
7-3
-------
For unpaved roads, the emission model is given by Equation 7-5:
(Eq. 7-5)
.«.mfiHirterflfi^
where: EF = emission factor [g/VKT];
610 = empirical constant [g/VKT];
s = percent silt content of road surface [%];
12 = empirical constant [unitless];
S = mean vehicle speed [km/hr];
48 = empirical constant [km/hr];
W = mean vehicle weight [Mg];
2.7 = empirical constant [Mg];
w = mean number of wheels per vehicle [unitless];
4 = empirical constant [unitless];
365 = no. of days per year [days]; and
p = number of days with < 0.01 inches precipitation [days].
The emission factors can be converted into a total mass emitted if multiplied by the number
of vehicle kilometers traveled.
If the dust is contaminated, the PM emission rates of Equations 7-1 through 7-3 may
be translated to emission rates of the contaminant as follows:
EF£ = Xi EF (Eq. 7-6)
where: EF; = emission factor of contaminant i [g/VKT]; and
X; = fraction of contaminant i in particulate matter [unitless].
In general, the dust and silt at a site will contain a higher fraction of the metal species than
the bulk soil at the site; i.e. the particulate matter is enriched with the metals. Therefore, X;
is equal to:
x; = C Z 10-6 (Eq. 7-7)
where: C = concentration of metal in the bulk soil [/tg/g];
Z = enrichment factor [unitless]; and
10"6 = conversion factor
7-4
-------
7.1.4 Minimum Requirements for Field Data
Site-specific field data must be collected (e.g., during the RI/FS)to provide the input
data necessary to generate reasonably accurate estimates of paniculate matter emissions. The
minimum field data required to estimate emissions for the various sources covered in this
section are:
Transfer operations: percent moisture content of the material;
Mixing and tilling: area treated and silt content of soil;
Grading: percent moisture content and silt content of material;
Traffic on paved roads: silt surface loading;
Traffic on unpaved roads: silt content of road surface; and
Metal emissions for any operation: average concentration of metal in bulk soil.
7.1.5 Sources of Input Data
Aerodynamic particle size multipliers for Equation 7-1 are provided in Section 7.1.3.
In general, meteorological data will be available from an on-site monitoring station. If not,
meteorological data may be obtained from a local airport or government monitoring station.
Soil data is available from the state agricultural service or the federal Soil Conservation
Service.
7.1.6 Default Values for Input Variables
Default values for equation input parameters are provided in Table 7-1. Some input
variables, such as mass of material handled and surface area;graded, are extremely site- and
operation-specific, so no default values for these variables are given. Table 7-2 contains
default values for metal enrichment of soils for use in Equation 7-7. Figure 7-1 shows a
geographic map of areas of the U.S. and the average number of days with >0.01 inch of
precipitation annually.
7-5
-------
Table 7-1.
Default Values for Estimating Emissions from Materials Handling
Parameter
Mean wind speed
Moisture content
Silt content
Silt surface loading
Mean vehicle speed
Mean vehicle weight
Mean # of wheels
Symbol
U
XHSO
s
sL
S
W
w
Units
m/sec
%
%
g/m2
km/hr
Mg
unitless
Default Value
4.4
10
8(<75/im)
5
20
3 (plant vehicle)
20 (Commercial haulers)
30 (plant haul trucks)
10
Expected
Range:
0 - 4.47
2-20
0.3 - 30
8-45
2-9
9-45
20-50
4-18
^Reference
*
*
4
4
4
.4
4
4
4
* = Author's estimate.
Table 7-2.
Metal Concentration and Enrichment Data (Z)
Arsenic (As)
Barium (Ba)
Cadmium (Cd)
Chromium (Cr)
Lead (Pb)
Mercury (Hg)
Selenium (Se)
Silver (Ag)
Median EnrichmeTit;lbaiSos^Z)
1.28
1.85
1.31
4.72
7.34
3.00
2.00
1.00
Source: Reference 6
7-6
-------
Figure 7-1. Mean Annual Number of Days With At Least 0.01 inches of Precipitation.
7-7
-------
7.1.7 Model Assumptions/Sensitivity Analysis
These models are equally applicable to a wide variety of materials handling activities.
They are based on the premise that a certain percentage of a soil's surface area has a high
"erosion potential", and that the rest of the surface will not be emitted. The equations
presented in this section are all empirically based and drawn from measurements at actual
sites; they are meant to predict the behavior of average sites. If a particular site has unusual
meteorological conditions, rubble, debris, or high silt content of soil, etc. , these model
accuracy may be affected. It is prudent to always monitor actual field emissions, at least
from some test location, to verify the model predictions.
7.1.8 Example Calculations
Assume that a Superfund site exists in Durham, NC and soil is excavated from a pit
and transported to a storage pile 500 m away. The backhoe moves 4 m3 of soil at a time,
and 10 truckloads a day are moved with each truck containing 20 m3 of soil. In addition, a
bulldozer works over the storage pile for an hour each day. The soil moisture content is
10% and the average wind speed at the site is 2 m/sec. The lead content of the soil is 100
To find the total PM10 emissions from this site, first convert the 20 m3 of soil for ten
trucks to a mass, using the default soil density of 1.5 g/cm3:
20 m3 x 1.5 g/cm3 * 106 cm3/m3 x 10 trucks = 300,000 kg/soil.
The particle size multiplier for < 10 /tm is 0.20. Use Equation 7-1 for the backhoe
emissions:
ER = 0.20 * (0.0016) * 300,000 * (2/2.2)1 3 / (10/2)1 4 = 8.9 g.
7-8
-------
This number should be multiplied by 2, because the soil will be dumped once into the trucks
and dumped a second time onto the storage pile. Thus the total emissions from dumping are
18 g and the average emission rate is 18 g/day.
To find the lead emissions from the backhoe operations, first calculate the fraction of
lead in the windblown dust using Equation 7-7 with the lead content of the soil (100 ug/g)
and the enrichment factor for lead from Table 7-2 (7.34):
X; = (100)C7.34)(10*) = 7.34X10"4 (g lead/g windblown dust)
This value is then used with the average emission rate calculated above (18 g/day) and
Equation 7-6:
EFleid = (7.34xlO-4)(18) = 0.013 g/day.
To find the PM10 emissions from transport, the silt content of the unpaved surface is
needed, as well as the number of wheels/truck. Assume that both equal the default values
from Table 7-1. The number of days with precipitation > 1" for North Carolina is found
from Figure 7-1. From Equation 7-5, the transport emissions are:
EF = 610 (8/12) (20/48) (25/2.7)0-7 (10/4)0-5 (365 - 120)/365 = 850 g/km.
A total of 10 truckloads are driven over a 1 km roundtrip, so the total emissions (ignoring
the weight difference between the empty and full truck) are 11,000 g or 11 kg. The average
emission rate is 11 kg/day.
Finally, to find the emission rate due to the bulldozing, use Equation 7-3:
ER - a094 '*>" - 0.085 g/sec
(10)"
Since the activity is underway for one hour, the total emissions are about 300 g.
7-9
-------
7.1.9 References
1. U.S. EPA. AP-42: Compilation of Air Pollutant Emission Factors, Fourth
Edition. U.S. EPA, Office of Air Quality Planning and Standards, Research
Triangle Park, NC. September 1985.
2. Cowherd, C., G. Muleski, P. Englehart, and D. Gillette. Rapi Assessment of
Exposure to Particulate Emissions from Surface Contamination Sites.
EPA/600-8-85/002. February 1985.
3. Englehart, P. and D. Wallace. Assessment of Hazardous Waste TSDF
Particulate Emissions. EPA Contract No. 68-02-3891. October 1986.
4. Cowherd, C., P. Englehart, G. Muleski, and J. Kinsey. Hazardous Waste
TSDF Fugitive Particulate Matter Air Emissions Guidance Document. EPA
450/3-89-019. May 1989.
5. Cowherd, C., et al. Development of Emission Factors for Fugitive Dust
Sources. EPA/450/3-74/037. 1974.
6. Eklund, B., et al. Air/Superfund National Technical Guidance Study Series.
Volume in. Estimation of Air Emissions from Cleanup Activities at
Superfund Sites. EPA-450/1-89-003. NTIS PB89. 1800 GI/AS. January
1989.
7.2 OTHER AREA SOURCES OF PM and METAL EMISSIONS
7.2.1 Description of Emission Process
Fugitive dust may be released from a variety of origins other than materials handling.
A remediation activity that may be a significant area sources of fugitive dust is
solidification/stabilization. Non-remediation sources include storage piles and dry
impoundments.
7.2.2 Model Selection
Equations based on fundamental physical laws have been reported for windblown
dust1'2, but the most widely accepted equations are those empirically derived by Cowherd, et
a!.3'4-5, which are still in current use by the Superfund program6. The most suitable 'equations
7-10
-------
for inclusion in this manual are those given by Cowherd, et al.7 for open waste piles and
staging areas, dry surface impoundments, and waste stabilization. These are incorporated in
the manual along with the metal enrichment factors for dust presented in Volume in of the
National Technical Guidance Series (NTGS) documents8.
7.2.3 Emission Model Equation
A simple model of erosion from level areas such as dry surface impoundments during
a time period t between disturbances is given by:
(Bq. 7-8)
ER =
k SA pt
t 86,400
where:
ER
k
SA =
Pt
t
86,400 =
emission rate from surface material during period t [g/sec];
particle size multiplier [unitless];
area of contamination [m2];
erosion potential corresponding to fastest mile of wind during
period t [g/m2];
no. of days between disturbances [day]; and
conversion factor [sec/day].
Particle size multipliers for Equation 7-8 are:
Size
< 30 urn
< 15 urn
< 10 urn
< 2.5 um
Multiplier
1.0
0.6
0.5
0.2
Total suspended particulates (TSP) from wind erosion of continuously active piles can
be estimated as:
(Bq. 7-9)
235 Us)
EF = 1.9 f J.) I365 -& ( f
7-11
-------
where: EF = emission factor (g/m2-day);
0.19 = empirical constant (g/m2-day);
s = percentage silt of aggregate (%);
1500 = empirical constant (unitless);
365 = no. of days/year (days);
p = number of days of precipitation > 0.01 inch per year (days);
235 = empirical constant (days);
f = fraction of time wind > 5.4 m/sec at mean pile height
(unitless); and
15 = empirical constant (unitless).
The fraction of TSP that is PM10 can be assumed to be 50% . Equation 7-9 is valid for piles
that are active at least once per day.
The emissions of PM from stabilization and solidification have been found to be over
1 kg/hr for full-scale operations. No equations, however, are available for estimating PM
emissions from the actual mixing process. An emission factor has been published that can be
converted into an emission estimation equation:8
ER = (0.05)(Q)(2.78xlO'4) (Eq. 7-10)
where: ER = emissions (g/sec);
0.05 = emission factor (g/kg);
Q = treatment rate (kg/hr); and
2.78x10^* = conversion factor (hr/sec).
PM emissions from the transfer of the stabilized waste can be estimated as:
0.00056 (M)
,u
where: E = emissions [g];
0.00056 = empirical factor [g/kg];
U = wind speed [m/sec];
2.2 = empirical factor [m/sec];
M = mass of material handled [kg];
= moisture fraction [%]; and
= empirical factor [%].
7-12
-------
7.2.4 Minimum Requirements for Field Data
The minimum field data required to estimate emissions for the various sources
covered in this section are:
Dry surface impoundments: surface area of contamination and the number of
days between disturbances;
Continuously active piles: percentage silt of aggregate and the fraction of time
with high winds; and
Stabilization and solidification: percent moisture content and mass of material
handled.
7.2.5 Sources of Input Data
Procedures for calculating the erosion potential are given in Appendix B. For other
variables, see the discussion in Section 7.1.5.
7.2.6 Default Values for Input Variables
Table 7-3 provides default values for the input parameters needed for Equations 7-8
and 7-9. For Equation 7-11, the fraction of TSP made up of PM10 is estimated to be 0.5.
7.2.7 Model Assumptions/Sensitivity Analysis
These emission models assume that after a disturbance, only a certain fraction of the
soil's surface will erode, regardless of the time exposed. That is why Equation 7-8 does not
depend on time, except for the length of the period between disturbances. Equation 7-9 is
for continuously active disturbances, and so it assumes that at any point in time, a
disturbance has just occurred, and the same fraction is able to erode.
For in-place contaminated soil, over-prediction of the emissions is possible as a soil
crust tends to form, reducing the erosibility of the pile or field.
7-13
-------
Table 7-3.
Default Values for Estimating PM Emissions from Other Area Sources
Parameter :
Surface area
Erosion potential
Percentage of silt
Fraction of time with high winds
Wind speed
Moisture fraction
Treatment Rate
Symbol
SA
P,
s
f
U
XHZO
Q
Units
m2
g/m2
%
unitless
m/sec
%
kg/hr
Defeult Value
2000
33
2.2
20
4.4
2
45,000
Expected Range
Site specific
0 - 525 (see Appendix B)
0.44 - 19
Site specific
0.6 - 6.7
0.25 - 4.8
4,500 - 120,000
Reference
7
7
7
7
-
7-14
-------
7.2.8 Example Calculations
A site in Raleigh, NC contains a one acre dry surface impoundment. The soil has a
silt content of 8%. The soil is excavated every other day, placed in an (active) storage pile,
and then fed into a stabilization process. The storage pile has a surface area of 2000 m2.
One thousand kilograms per day of the stabilized material is placed in a lined portion of the
dry surface impoundment.
The particulate matter emissions from the excavation and transfer (dumping)
operations can be calculated using procedures previously described. Example calculations for
particulate matter emissions of less than 10 microns from the other sources are given below.
Emissions from the surface impoundment are estimated using Equation 7-8. The
surface area of the impoundment is one acre or 4050 m2. The particle size multiplier for
< 10 nm is 0.5, the time between disturbances is two days, and the default erosion potential
is 33 g/m2. The emission rate from the surface impoundment is:
ER - (0-5X4050X33) . Q 3^
(2X86,400)
Particulate matter from wind erosion of continuously active piles can be estimated
using Equation 7-9. The number of rainy days from Figure 7-1 is 55:
1.9 s - 55)
- 1-78 g/m2-day
235
(i)
Assuming 50% of the TSP is PM10 , the emissions of PM,0 from the storage pile are:
ER = (1.78)(50/100)(2000) = 1800 g/day.
7-15
-------
The emissions of PM emissions from the transfer of the stabilized waste can be
estimated using Equation 7-11:
/44V-3
0.00056 (1,000)
E = ±-LJ. = 1.4 g (per day)
UJ
The total daily emissions from transfer are thus:
»-7
= (2.29 . IP"7 g/kg)(1000 kg) = . ^
14 *« J
day
7.2.9 References
1. Momeni, M.H., Y. Yuan, and A.J. Zielen. The Uranium Dispersion and
Dosimetry (UDAD) Code. U.S. NRC. NTIS NUREG/CR-0553. May 1979
2. RTI. A Method for Estimating Fugitive Paniculate Emissions from Hazardous
Waste Sites. EPA/600/2-87/066. NTIS PB87-232203. August 1987.
3. U.S. EPA. User's Guide - Emission Control Technologies and Emission
Factors for Unpaved Road Fugitive Emissions. EPA/625/5-87/022.
September 1987.
4. U.S. EPA. AP-42: Compilation of Air Pollutant Emission Factors, Fourth
Edition. U.S. EPA, Office of Air Quality Planning and Standards, Research
Triangle Park, NC. September 1985.
5. Cowherd, C., G. MulesM, P. Englehart, and D. Gillette. Rapi Assessment of
Exposure to Particulate Emissions from Surface Contamination Sites.
EPA/600-8-85/002. February 1985.
6. U.S. EPA. Guideline for Predictive Baseline Emissions Estimation Procedures
for Superfund Sites, Interim Final. Environmental Quality Management, Inc.
January 1992.
7. Cowherd, C., P. Englehart, G. MulesM, and J. Kinsey. Hazardous Waste
TSDF Fugitive Particulate Matter Air Emissions Guidance Document. EPA
450/3-89-019. May 1989.
8. Eklund, B., et al. Air/Superfund National Technical Guidance Study Series,
Volume HI: Estimation of Air Emissions from Cleanup Activities at
Superfund Sites. Report No. EPA-450/1-89-003. NTIS PB89 180061/AS.
January 1989.
7-16
-------
APPENDIX A
-------
ill
s
es
$
8
en
oa
o
o
o
o
en
o\
o
o
00
eb
.S
8
in
en
8
o.
ci
ts
eo
S
8
es
S
*-H
en
en
VO
\o
VO
E
E+06
25E
E+06
so-faoo
62E+04
E+05
23E+06
61E+07
09E+09
69E+09
98E
07
9E
E+0
3
50E+09
E
s
eM
en
8
8
00
§
a
04
c
8
u
en
o
es
00
VO
8
s
s
a
eN
\o
S
ts
8
S!
r-
8
O
"
o
U
cJ
0
S
-QO
0,
0
-44-
BQ
O
s
S
5892-23-6
S
5
B
S
1
rganic Co
ho
.2
cet
u
T3
C
1
1
<
i
B
u
c
«.
N
§
OQ
Bromo
a
2
D
:S
VO
Butano
ts
g
CQ
o
«
"a
fide
Carbo
(S
-------
8
§
u
1
I
'-='.' '!*
\£
I-3
f S?
" ?"* .**».
11 f|
.-5**.1
'.i" - : SC
-.ivs: ."':>*
K: -'.f^
*o c--»-
ot'Si ^
' **'s
:|il
"I!
-..**: >*^
>;jl|
3
1
U
:- -1
...:^ ' ;
;- .1
1
1
1
o
S
3
120-80-9
o
a
OE
1
'
1
Ov
^
O
es
oo
_c
o
.c
e>
00
N
2
00
oo
10
«?
1
m
7.14E+07
o
g
b
00
g
vH
Q
si
o
108-90-7
enzene
§
f
O
e'
t
1
-
8
i
1
'-'
i
^
VO
00
&N
U
X
U
3
ifluoromethane
o
O
S
t
~*
!
8
oT
Tt
<^
cn
en
1.33E+09
i
b
R
00
en
Ov
ff
U
!
£
o
e'
O
O
1
b
"3
in
s
10
H
X!
CM
Ov
i
m
o
00
Ef
5-
i
IM
1
t
1
1
£
o
U
en
en
1
'
1
VO
9
|
ex
1
rj
in
£
o
d1
m
cntafluoroethane
n.
c
JS
T
en
1
1
1
*
9
S
in
en
en
1.12E+09
b
8
m
00
oo
c
cJ
126-99-8
0
_c
e»
Jj
^
1
I
in
en
S>
3
"*
4.65E+05
I
b
b
2
2
0
5
108-39-4
Q
s
vo
en
I
en
oo'
1
3V
10
I
^
1.40E+06
I
b
b
2
eo
O
0
X
*r
oo
I
U
6
en
1
1
s
>n
en
^
1
"*
6.39E+05
1
b
b
2
oo
o
o.
d
I
i
0,
00
en
1
I
1
oT
tn
«?
ov
*
1.11E+10
!
o
en
s
S
w
u
460-19-5
e
c
c
U
Ov
en
5
en
I
Ov
8
SO
«-H
1
I
i-H
4.52E+08
Ov
b
8
VO
00
2
0
I
o
c
0
*o
U
C
q
**
00
00*
en
m
"?
t>
Tf
"*
6.57E+06
§
b
5
*
8
q.
u
108-93-0
~e
c
.c
"o
U
r
_
t
I
oo
a
oo
en
"?
en
*
2.53E+07
1
b
oo
^
Ov
°o
§
108-94-1
o
c
c
c
a
.C
"o
U
«
8
M
1
n
« i
O
« 4
0
0
-
1
i
m
S
0
5
110-83-8
t
"o
U
5
r
i
M-^
8
v
rt
Ov
OS
I
n*
90
I
|
~*
1.24E+09
8
b
a
£
00
_:
5
S
lloroethane
c
Q
_
S
s?
~
Ov
Ov
oo*
g
ex
ex
~
4.26E+08
b
o
00
VO
ec
_J
5
107-06-2
o
i
c
J
j
C
b
5!
*
~"
i
o
Ov
in
J1
Ov
in
ex
3.13E+09
b
1
Ov
Ov
-y
S
i
lloroethylenc
c
Q
K
9
O
'
1
en
*?
1
Tt
1.08E+09
I
S
VO
Ov
O
5
i
Dichloroethylene
i
0
*C
T
-------
8
8
es
m
6
in
ts
.S
.S
t-
s
8
w
S
**
VO
ts
8
5?
0\
S
ts
ts
1
1
u
I
09
E
ts
09
E
E
E
cs
E
m
79E
3
46E+07
S
C
O
8
S
8
S!
ts
3
S
ts
S!
s
0
C
o
o
==.
ts
S!
3
i
>n
in
S
in
'
VO
3
8
Q
vO
,
U
w
S
de
in
rba
S
C
-------
g
1
u
<
X
1
. -|
Maximum
Water
-.;
«3- s
|l:
so
1
£
+^
1
?
1
1 1
f'l
'i
j!
;, a
i." **
:W
O :
;H
* SS
Uf
ill
*3
g
Organic CoWpdttnd : "
o
o
o
b
*
41E+08
-*
o
b
*~*
00
00
2
0
d*
106-93-4
dibromide
o
n
CM
00
I
~4
"s
ON
ON
"f
.03e-07
rt
34E+05
*
o
S
b
CO
b
g
S
1
i
i
o
n
CO
oo
1
"s
Tl
9
v>
"*
1
1
g
3
z
as,
u
j
o
t
'I
o
J
n
0
CM
M
9
S
1
>n
00
CO
|
^^
60+396
CM
b
o
s
5
o.
°P
1
o
o
f
H
vi
oo
O
b
o
ON
n
Tf
CM
t
1
VI
65E+09
V>
1
b
8
CO
0
0
CO
q.
u
i
u
c
Ji
o
T3
.Q
NO
OO
b
o
'
£
f
1
f*"
04E+08
*"
o
1
b
"*
S
58
4;
u
64-18-6
T:
OS
£
r-
00
1
s
CM
2"
JM
o
b
1.60E-04
8
CM
ON
q
u
V,
2""
c
3
ON
OO
I
O
b
CO
CO
o
02e+00
CM
48E+08
CM
O
b
*
CM
8
|
142-82-5
o
c
e
h
a
o
ON
1
1
CO
s
9
N
b
80+396
NO
b
CO
b
oo
s
s
CO
1
e
x
o
2
o>
?
1
1
00
^
i
NO
48E+07
CM
1
Tf
S
CM
CO
2
a
S
CO
o
_e
^
a
CM
1
1
1
36E+10
NO
O
CM
CO
s?
NO
CO
U
7647-01-0
'c
0
t
o
o
a
CO
1
1
K
?!
^
i
"*
i
s
s
1
i
s
a
C
J
a
a
1
1
1
78E+10
CM
b
1
S
CO
X
7783-06-4
u
T
c
<
T
a
ON
1
M
CM
of
$
1
|
CM
98E+07
CO
b
o
CM
?
K
O
CO
oo
o
e
i
o
G
ON
1
1
O
NO"
J\
1
1
^^
I
1
NO
NO
cf
S
110-19-0
tl
s
^^
i
S;
s
9
1
c
CM
oo
L''
|
^^
38E+08
*
g
8
b
00
s
1
9
1
"c
1
e
C
B
OO
ON
1 '
1
3
^
L-'
n
46E+09
^
1
-
ON
1
O
9
n
04E+07
f-
u
©
ON
O
ON
§
jf
op
S
ON
benzene
c
G
I
1
=
n
'
|
80+396
*
8
0
^
S
CM
CO
t
1
O
.S
1
o
1
v>
^
S
'
3
36E+08
ON
b
VI
CO
CM
§
?
S
O
i
i
"e
cd
1
r^
O
1
1
8
S
^
|
i
i
i
8
NO
00
O
CO
C'
«
*
«
1
ff\
0
I
'
8
CM
«?
I
29E+09
i
S
CO
1
i
o
J
'g
e
J
O
1
1
1
S
9
1
CM
1
1
ON
S
a.
u
t~
u
t
1
f>
1
rf
o
b
9
00
00*
Tf
CO
M
n
NO
S
CO
NO"
8
-------
s
5?
cs
8
8
IS
X
00
5?
CM
VO
VO
I
.^*
+-»
a
o
U
<
X
i
M
S
B)
W
CM
0)
w
u
8
u
8
1*1
3'
s
S
en
O
CM
VO
8
8
8
8
8
VO
SO
JS
t
o
O
c
g
60-
ov
i
VT)
,2
I
O
eto
pta
pa
CM
VO
CM
CM
3
es
CM
£
.2
-------
"2
u
1
I
; 1
f 1?
R ^ -5
III?
'J?;l*
: :" «
i a8
I3 !;;s=
;;||t
^'^
53 ^
O 3 1C
^ M ^^
u ^^
"^ J*-i
ll^ «
»
I
:!;.^ : x.i:- Organic Compound: ' ..Iv: >,:;
z
'
'
s
2
£
a
^^
05
Q
§
00
§
c
e
1
$
i
i
s
|
32E+07
~
1
en
S
£
1
5
ti
>n
beta-Propiolactone
t-
l
I
J
1
S
"?
I
^H
37E+08
o>
S
d
8
00
o
8
3
oc
Propionaldehyde
01
1
1
"s
^
t
1
90
"*
98E+07
1
5
JJ
CN
r>
"
64E+09
*"
d
i
§
00
X
5
r-
Propylene oxide
-<
o
9
i
J
26E+08
en
^J
V-H
5
X
°?
1 ,2-Propylenimine
3
d
00
S
S
T
n
CM
50E+07
oo
O
s
d
ts
^
?
X
o1
1
u
c
1
(t1
s?
I
1
1
1
1
s
00
o
o,
U
106-51-4
Quinone
f
cs
00
8
00
o
^
1
04
09E+07
*
o
i
d
S
S
s
£
u°
1
o
c
S
Tf
1
1
,
S
9
5
5?
P4
!
I
n
oo
3
u"
3
1,1,1 ,2-Tetrachloro-2,2-Difluoroethane
T
0\
cn
i
8
S
?
1
n
00
5o
O
5
I
o
e
9
2
|
c-
5
|
00
n
S
*^
N
(S
80+369
**
£
d
o\
00
!8
U
T
00
i
u
c
J
J
o
1
01
1
1
"s
m
"T
|
^
79E+08
cs
0
oo
d
VH
P
i-H
S
O
X
tj
i
Tetrahydrofuran
a
n
ri
00
2
00
s»
1
NO
49E+08
I
d
o
*
2
5
00
00
i
Toluene
o
«n
i
3
s
^
1
'-'
73E+06
»-(
i
CO
d
NO
5
z.
106-49-0
P-Toluidine
V
cQ
00
00
3
o
1
r*
*""
82E+08
00
0
00
d
2
5
8
3*
?
1,1,1 -Trichloroethane
S
s
X5
00
1
*
r>
«?
|
r~
79E+08
^
S
d
S
5
rn
C^
I
1 , 1 ,2-Trichloroethane
CO
vn
o
ts
o\
8
s
o
1
d
*"
^.
92E+09
^
1
o
d
S
r-
«
G
o
1
Trichlorofluoromethanc
S
1
J\
1
S5
1
N
ts
46E+07
ts
o
i
d
3
s>
$
5
96-18-4
1 ,2,3-Trichloropropane
V
d
s
00
3
«
9
1
S
9
1
~"
t
t
2
S
f
S
m
1 ,2,4-Trimethylbenzenc
S
i
i
8
»
9
I
"
20E+07
1
«
o\
S
C
S
8-/L9-80I
1 ,3,5-Trimethylbcnzene
S
-------
«*
13
|
C
O
u
<
12
c
(D
1
1
'1
^1?
*s *s* S
<5 ^ ^5
x *s &
S -^^s
j s $ 00
fe" :'^2 B
:SJ!r
5.:^ '-J
** «
lit
Jtt
s
-f:fa?
"S < 4E
^3 ^ e
Q H^'
^
||f
* S
** >**
111
JS
I
1
«^
^
.1
«
!
o
e
~Zi
1
i
N
Ov
o"
r»
§
cs
»
32E+08
1
6
»n
^H
^ t
S
vd
00
~
cJ
i
00
o
«*
^s
^c
s
1
1
1
.15E+09
«o
o!
00
0
£
IH
CQ
§
3
m
o\
s
00
**!
"
«n
i
3
rt
^H
'
i
i
00
.94E+09
00
g
s
d
o
0)
10
D
S:
t
9
1C
8
nyl-Chlori
>
8
»
en
i
30
^
8
(S
J)
1
>
X
6
S
2
en
0
d
X
*-<
00
CM
CM
i
0
.42E+07
Wl
O
d
10
o\
i
0
ajT
i
^5
^^
o
c
Jj
x"
0.
S
g
««!
-------
-------
APPENDIX B
PROCEDURES FOR CALCULATING EROSION POTENTIAL
Source: Cowherd, C., P. Englehart, G. Muleski, and J. Kinsey. Hazardous Waste
TSDF Fugitive Paniculate Matter Air Emissions Guidance Document. EPA-
450/3-89-019 (NTIS PB90-103250). May 1989.
-------
3.2.2 Wind Erosion
Dust emissions may be generated by wind erosion of open waste piles
and exposed areas within a disposal facility. These sources typically
are characterized by nonhomogeneous surfaces Impregnated with nonerodible
elements (particles larger than approximately 1 cm in diameter). Field
testing of coal piles and other exposed materials using a portable wind
tunnel has shown that (a) threshold wind speeds exceed 5 m/s (11 mph) at
15 cm above the surface or 10 m/s (22 mph) at 7 m above the surface, and
(b) particulate emission rates tend to decay rapidly (half life of a few
minutes) during an erosion event. In other words, these aggregate
material surfaces are characterized by finite availability of erodible
material (mass/area) referred to as the erosion potential. Any natural
crusting of the surface binds the erodible material, thereby reducing the
erosion potential.
3.2.2.1 Emissions and Correction Parameters. If typical values for
threshold wind speed at 15 cm are corrected to typical wind sensor height
(7-10 m), the resulting values exceed the upper extremes of hourly mean
wind speeds observed in most areas of the country. In other words, mean
atmospheric wind speeds are not sufficient to sustain wind erosion from
flat surfaces of the type tested. However, wind gusts may quickly
deplete a substantial portion of the erosion potential. Because erosion
potential has been found to increase rapidly with increasing wind speed,
estimated emissions should be related to the gusts of highest magnitude.
The routinely measured meteorological variable which best reflects
the magnitude of wind gusts is the fastest mile. This quantity repre-
sents the wind speed corresponding to the whole mile of wind movement
which has passed by the 1-mi contact anemometer in the least amount of
3-6
-------
time. Dally measurements of the fastest mile, are presented 1n the, -~ -,
monthly Local CUmatological Data (LCD) summaries.l« The duration of the
fastest mile, typically about 2 m1n (for a fastest mile of 30 mph),
matches well with the half life of the erosion process, which ranges
between 1 and 4 min. It should be noted, however, that peak winds can
significantly exceed the daily fastest mile.
The wind speed profile 1n the surface boundary layer is found to
follow a logarithmic distribution:
u(z) - £5 In f- (z > 2Q) (3-2)
0
where: u = wind speed, cm/s
u* = friction velocity, cm/s
2 - height above test surface, cm
2Q = roughness height, cm
0.4 = von (Carman's constant, dimensionless
The friction velocity (u*) is a measure of wind shear stress on the
erodible surface, as determined from the slope of the logarithmic
velocity profile. The roughness height (z0) is a measure of the rough-
ness of the exposed surface as determined from the y-intercept of the
velocity profile, i.e., the height at which the wind speed is 2ero.
These parameters are illustrated in Figure 3-2 for a roughness height of
0.1 cm. The roughness height (20) is needed to convert the friction
velocity to the equivalent wind speed at the typical weather station
sensor height of 7 to 10 m above the surface.
Emissions generated by wind erosion are also dependent on the fre-
quency of disturbance of the erodible surface because each time that a
surface is disturbed, its erosion potential is restored. A disturbance
is defined as an action which results in the exposure of fresh surface
material. On a storage pile, this would occur whenever aggregate mate-
rial is either added to or removed from the old surface. A disturbance
of an exposed area may also result from the turning of surface material
to a depth exceeding the size of the largest pieces of material present.
3-7
-------
o
o
"a;
CD
o
o
c
o
-------
3.2.2.2 Predictive Emission Factor Equation*1. The emission factor
for wind-generated participate emissions from mixtures of erodible and
nonerodible surface material subject to disturbance may be expressed in
units of g/m2-yr as follows:
N
Emission factor » k Z P, (3~3)
1 « 1 1
where: k « particle size multiplier
N * number of disturbances per year
P.J * erosion potential corresponding to the observed (or
probable) fastest mile of wind for the 1-th period between
disturbances, g/m*
The particle size multiplier (k) for Equation 3-3 varies with
aerodynamic particle size, as follows:
Aerodynamic Particle Size Multipliers for Equation 3-3
<30 um <15 »m <10 um <2.5 um
df
1.0 0.6 0.5 0.2
This distribution of particle size within the < 30 jun fraction is
comparable to the distributions reported for other fugitive dust sources
where wind speed is a factor. This is illustrated, for example, in the
distributions for batch and continuous drop operations encompassing a
number of test aggregate materials (see AP-42 Section 11.2.3).
In calculating emission factors, each area of an erodible surface
that is subject to a different frequency of disturbance should be treated
separately. For a surface disturbed daily, N » 365/yr, and for a surface
disturbance once every 6 mo, N * 2/yr.
3-9
-------
The erosion potential function for a dry, exposed surface has the
following form:
P = 58 (u* - u*)z 4- 25 (u* - u*)
P * 0 for u* < u* (3-4)
where: u* = friction velocity (m/s)
ut « threshold friction velocity (m/s)
Table 3-2 presents the erosion potential function in matrix form.
Because of the nonlinear form of the erosion potential function, each
erosion event must be treated separately.
Equations 3-3 and 3-4 apply only to dry, exposed materials with
limited erosion potential. The resulting calculation is valid only for a
time period as long or longer than the period between disturbances.
For uncrusted surfaces, the threshold friction velocity is best
estimated from the dry aggregate structure of the soil. A simple hand
sieving test of surface soil (adapted from a laboratory procedure pub-
lished by W. S. Chepil12) can be used to determine the mode of the
surface aggregate size distribution by inspection of relative sieve catch
amounts. This procedure is specified in Section 4.2.1. The threshold
friction velocity for erosion can be determined from the mode of the
aggregate size distribution, following a relationship derived by
Gillette,13 as shown in Figure 3-3.
Threshold friction velocities for several surface types have been
determined by field measurements with a portable wind tunnel.13-1* These
values are presented in Tables 3-3 and 3-4 for industrial aggregates and
Arizona s.f£es. Figure 3-4 depicts these data graphically.
The fastest mile of wind for the periods between disturbances may be
obtained from the monthly LCD summaries for the nearest reporting weather
station that is representative of the site in question.i° These sum-
maries report actual fastest mile values for each day of a given month.
3-10
-------
o
=3
u_
UJ
O
o.
UJ
CM
1
CO
UJ
is
M
o
a.
«r
CM
O
CM
CO
II
VO
rl
^-<
CM
*
*"H
CD
» i
00
o
so
0
o
CM
o
ii
* VI
OOOOOOOOOOOO^OvvO
*-4 en u) eo
^^ n 10 oo *
<
-------
TABLE 3-3. THRESHOLD FRICTION VELOCITIESINDUSTRIAL AGGREGATES
Threshold wind
Material
Overburden*
Scoria (roadbed
material)*
Ground coal
(surrounding coal
pile)
Uncrusted coal pile
Scraper tracks on
coal pilea»D
Fine coal dust on
concrete pad
Threshold
friction
velocity,
m/s
1.02
1.33
0.55
1.12
0.62
0.54
velocity at
Roughness
height,
cm
0.3
0.3
0.01
0.3
0.06
0.2
10 m
ZP *
actual
21
27
16
23
15
11
(m/s)
2o *
0.5 cm
19
25
10
21
12
10
Ref.
9
9
9
"9
9
15
Western surface coal mine.
bLightly crusted.
cEastern power plant.
TABLE 3-4. THRESHOLD FRICTION VELOCITIESARIZONA SITES
Location
Mesa - Agricultu- " site
Glendale - Construction site
Maricopa - Agricultural site
Yuma - Disturbed desert
Yuma - Agricultural site
Algodones - Dune flats
Yuma - Scrub desert
Santa Cruz River, Tucson
Tucson - Construction site
Ajo - Mine tailings
Hayden - Mine tailings
Salt River, Mesa
Casa Grande - Abandoned
agricultural land
Threshold
friction
velocity
(m/s)
0.57
0.53
0.58
0.32
0.58
0.62
0.39
0.18
0.25
0.23
0.17
0.22
0.25
Roughness
height
(cm)
0.0331
0.0301
0.1255
0.0731
0.0224
0.0166
0.0163
0.0204
0.0181
0.0176
0.0141
0.0100
0.0067
Threshold
wind velocity
at 10 m
(m/s)
16
15
14
8
17
18
11
5
7
7
5
7
8
3-12
-------
o
o
m
£
o
S i
o> ^ «
O)
o
o
4_>
.O
i.
0)
M
I/I
o
u
o
o
O
u)
Ol
c
o
\a * n CM
O
O
i
co
a>
(oes/iuo) i*n 'A;ioo|eA uoiiouj
3-13
-------
r
o
Q>
w
3
M
O
1 1
1
I
I 1JL
iLili
O
o
c
o
.2
o
75
E
o
0)
o
tt
o -o
N O
« E
tt C
oir
5
5.
11
«»
14
I 1*1
8 g-o i a,
2 IIU
1 *i?l
i Sill
J !
1 1 1
CM
\n
c>
CM
d
d
d
VI
41
u
JO
41
0
0
vt
L
£
0)
c
O
"53
o
re
c
o
LL
§
5
o
(0 C
o m
O W
u
3
-------
Because the erosion potential 1s a highly nonlinear function of the
fastest mile, mean values of the fastest mile are Inappropriate. The
anemometer heights of reporting weather stations are found in Refer-
ence 17, and should be corrected to a 10 m reference height using
Equation 3-2.
To convert the fastest mile of wind (u*) from a reference anemometer
height of 10 m to the equivalent friction velocity (u*), the logarithmic
wind speed profile may be used to yield the following equation:
u* = 0.053 uto (3~5)
where: u* « friction velocity (m/s)
u{0 * fastest mile of reference anemometer for period between
disturbances (m/s)
This assumes a typical roughness height of 0.5 cm for open
terrain. Equation 3-5 is restricted to large relatively flat piles or
exposed areas with little penetration into the surface wind layer.
If the pile significantly penetrates the surface wind layer (i.e.,
with a height-to-base ratio exceeding 0.2), it is necessary to divide the
pile area into subareas representing different degrees of exposure to
wind. The results of physical modeling show that the frontal face of an
elevated pile is exposed to wind speeds of the same order as the approach
wind speed at the top of the pile.
For two representative pile shapes (conical and oval with flat-top,
37 degree side slope), the ratios of surface wind speed (us) to approach
wind speed (ur) have been derived from wind tunnel studies.lu The
results are shown in Figure 3-5 corresponding to an actual p-il-e-height of
11 m, a reference (upwind) anemometer height of 10 m, and a pile surface
roughness height (z0) of 0.5 cm. The measured surface'winds correspond
to a height of 25 cm above the surface. The area fraction within each
contour pair is specified in Table 3-5.
3-15
i
-------
Flow
Direction
Pile A
Pile B1
Pile B2 P"e B3
Figure 3-5. Contours of normalized surface wind soeeds, u./u .
3-16
-------
TABLE 3-5. SUBAREA DISTRIBUTION FOR REGIMES OF u$/ur
Pile subarea
0.2a
0.2b
0.2c
0.6a
0.6b
0.9
1.1
Percent
Pile A
5
35
-
48
-
12
of pile surface
Pile 81
5
2
29
26
24
14
area (Fiqure
Pile 82
3
28
_
29
22
15
3
3-3)
Pile 83
3
25
w
28
26
14
4
The profiles of us/ur in Figure 3-5 can be used to estimate the
surface friction velocity distribution around similarly shaped piles,
using the following procedure:
1. Correct the fastest mile value (u*) for the period of interest
from the anemometer height (z) to a reference height of 10 m
(u|0) using a variation of Equation 3-2, as follows:
u+ - u* In (10/0.005) (3.6)
ui° ~ u In (2/0.005)
where a typical roughness height of 0.5 cm (0.005 m) has been
assumed. If a site specific roughness height is available, it
should be used.
2. Use the appropriate part of Figure 3-5 based on the pile shape
and orientation to the fastest mile of wind, to obtain the
corresponding surface wind speed distribution (u^), i.e.,
«; - <£>«:. <3-7>
3. For any subarea of the pile surface having a narrow range of
surface wind speed, use a variation of Equation 4-2 to calcu-
late the equivalent friction velocity (u*), as follows:
0.4 u*
u* = ~i=0.10u! (3-8)
i_^' 5
lno
3-17
-------
From this point on, the procedure 1s Identical to that used for a flat
pile, as described above.
Implementation of the above procedure 1s carried out 1n the
following steps:
1. Determine threshold friction velocity for erodlble material of
Interest (see Tables 3-3 and 3-4 or use Figure 3-3 to determine
the mode of the aggregate size distribution).
2. Divide the exposed surface area Into subareas of constant fre-
quency of disturbance (N).
3. Tabulate fastest mile values (u*) for each frequency of dis-
turbance and correct them to 10 m (uto) using Equation 3-6.
4. Convert fastest mile values (ut0) to equivalent friction
velocities (u*), taking Into account (a) the uniform wind expo-
sure of nonelevated surfaces, using Equation 3-5, or (b) the
nonunlform wind exposure of elevated surfaces (piles), using
Equations 3-7 and 3-8.
5. For elevated surfaces (piles), subdivide areis of constant N
into subareas of constant u* (i.e., within the isopleth values
of u$/ur in Figure 3-5 and Table 3-5) and determine the size of
each subarea.
6. Treating each subarea (of constant N and u*) as a separate
source, calculate the erosion potential (P.,-) for each period
between disturbances using Equation 3-4 and the emission factor
using Equation 3-3.
7. Multiply the resulting emission factor for each subarea by the
size of the subarea, and add the emission contributions of all
subareas. Note that the highest 24-h emissions would be
expected to occur on the windiest day of the year. Maximum
emissions are calculated assuming a single event with the
highest fastest mile value for the annual period.
The recommended emission factor equation presented above assumes
that all of the erosion potential corresponding to the fastest mile of
wind is lost during the period between disturbances. Because the fastest
mile event typically lasts only about 2 min, which corresponds roughly to
the half-life for the decay of actual erosion potential, it could be
3-18
-------
argued that the emission factor overestimates participate emissions.
However, there are other aspects of the wind erosion process which offset
this apparent conservatism:
1. The fastest mile event contains peak winds which substantially
exceed the mean value for the event.
2. Whenever the fastest mile event occurs, there are usually a
number of periods of slightly lower mean wind speed which
contain peak gusts of the same order as the fastest mile wind
speed.
Of greater concern is the likelihood of overpredlctlon of wind
erosion emissions in the case of surfaces disturbed infrequently in
comparison to the rate of crust formation.
I
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-451/R-93-001
3. RECIPIENT'S ACCESSION NO.
. TITLE AND SUBTITLE
Air/Superfund National Technical Guidance Study Series-
Models for Estimating Air Emission Rates from Superfund
Remedial Actions
5. REPORT DATE
March 1993
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
Bart Eklund and Charles Albert
8. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
8501 Mo-Pac Boulevard
Austin, Texas 78759
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-DO-0125
2. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPOFIT AND PERIOD COVERED
U.S. Environmental Protection Agency
OAR/OAQPS (MD-13)
Research Triangle Park, North Carolina 27711
14. SPONSORING AGENCY COD'E
5. SUPPLEMENTARY NOTES
EPA Project Officer: James F. Durham - (919) 541-5672_
6. ABSTRACT
This report is a compendium of models (equations) for estimating air
emissions from Superfund sites undergoing remediation. These models "predict
emission rates of volatile organic compounds (VOC's) and particulate matter (PM)
from both area and point sources. The following remedial processes are covered:
air stripping, soil vapor extraction, thermal desorption, thermal destruction
(incineration), excavation, dredging, solidification/stabilization, and
bioremediation. Emission estimation methods are also presented for landfills,
lagoons, and spills/leaks/open waste pits.
The models contained in this compendium will not accurately predict
emissions for all possible scenarios. Where uncertainty exists, these models
and the default inputs have been designed to overpredict emissions.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Toxic Air Emissions
Superfund Cleanup
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
170
20 SECURITY CLASS (This page I
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the LPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, it used, in smaller
type or otherwise subordinate it to mam title When a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it uas selected (e.g.. date o] mi/i', date oj
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc./. List author's afliliation if it differs from the performing organi-
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the progiam element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Insert appropriate code.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with. Translation of, Presented al conference of.
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or less) factual summary of the most significant information contained in the report. It the report LontJins a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authon/ed terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI HELD GROUP - Field and group assignments are to be taken from the 1965 C'OSATl Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary I icId/Group assignments that will follow
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." C'ite any availability to
the public, with address and price.
19. &20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, it any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, it known.
EPA Form 2220-1 (Rev. 4-77) (Reverse)
------- |