EPA 402/R-95-16
ESTIMATES OF HEALTH RISKS ASSOCIATED
WITH RADIONUCLIDE EMISSIONS FROM
FOSSIL-FUELED STEAM-ELECTRIC
GENERATING PLANTS
August 1995
U.S. Environmental Protection Agency
Office of Radiation and Indoor Air
Washington. D.C. 20460
-------
LIST OF PREPARERS
Various staff members from EPA'suffice of Radiation and Indoor Air contributed to the
development and preparation of this report.
Chris Nelson
Byron Hunger
Gerome Puskin
Al Colli
John Karhnak
Author Reviewer
Reviewer
Reviewer
Reviewer
Reviewer
An EPA contractor, S. Cohen & Associates, Inc., McLean, VA, provided significant
technical support in the preparation of the report.
-------
CONTENTS
Page
Executive Summary ES-1
1. Background Information 1-1
1.1 Electric Power Generation: Past, Present, and Future 1-2
1.2 The Natural Radionuclide Content in Fossil Fuels 1-6
1.3 Structure of the Report 1-6
.2. Radionuclide Content of Coal Used by U.S. Electric Utilities 2-1
2.1 Introduction 2-1
2.2 Coal Facts and Background Information 2-2
2.2.1 Coal Classification 2-2
2.2.2 Coal Reserves 2-4
2.2.3 Coal Production , 2-4
2.2.4 Coal Processing .' 2-7
2.2.5 Coal Users 2-7
2.2.6 Coal Suppliers 2-9
2.3 Coal Sample Analyses of the USCHEM Data Base 2-9
2.3.1 Chemical Occurrence of Thorium and Uranium in Coal 2-11
2.3.2 The Uranium and Thorium Content of Coal 2-12
2.4 Other Coal Data 2-17
2.5 Activity Concentration Equivalents 2-18
2.6 Summary 2-20
3. Radionuclides in Natural Gas 3-1
3.1 Production, Sources, and Consumption of Natural Gas 3-1
3.2 Radon Concentrations at the Wellhead 3-4
3.3 Radon Reduction Following Gas Processing 3-9
3.4 Radon Reduction in the Distribution System 3-12
3.5 Conclusion 3-19
4. The Radionuclide Content of Residual Fuel Oil 4-1
4.1 Background Information 4-2
4.1.1 Oil Extraction 4-2
4.1.2 Radioactive Source Terms 4-5
4.1.3 Petroleum Refining Processes and the Fate of Trace Metals 4-8
4.2 A Review of Past Studies Relevant to Estimating the Radionuclide
Content in Residual Fuel Oil 4-13
4.2.1 Residual Fuel Oil Studies 4-14
4.2.2 Crude Oil Study Data 4-18
4.2.3 Residual Fuel Oil-Ash Data 4-20
11
-------
CONTENTS (Continued)
Page
4.2.4 Radionuclide Emission Data from Oil-Fired SGUs 4-25
4.2.5 Summary of Past Study Data 4-26
4.3 Results of a Recent EPA Study 4-26
4.3.1 Utility Participation and Sample Data Information 4-29
4.3.2 Sample Receipt and Sample Preparation 4-29
4.3.3 Sample Analyses 4-32
4.3.4 Sample Results 4.34
4.3.5 Quality Assurance and Composite Sample Results 4-37
4.3.6 Interpretation and Use of Experimental Data 4-39
4.3.7 The Need for Extrapolation of Data for Radionuclides Not
Measured 4-40
4.3.8 Summary . 4_41
5. Boiler Design Features that Affect Radionuclide Emissions 5-1
5.1 General Description 5-1
5.2 Basic Power Plant Components 5-2
5.3 Coal-Fired Boilers 5.3
5.3.1 Pulverized Coal Boilers 5-3
5.3.2 Cyclone Boilers .• 5.7
5.3.3 Stoker Boilers 5-8
5.3.4 Fluidized Bed Boilers 5-9
5.4 Natural Gas-Fired Boilers 5.9
5.5 Oil-Fired Boilers 5-10
5.6 Fossil Fuel Combustion 5-10
5.7 Temperature Profile 5-11
5.8 Ash Formation 5-12
5.8.1 Ash Characteristics 5-12
5.8.2 Slagging and Fouling . 5-14
5.8.3 Bottom Ash . 5-15
5,8.4 Fly Ash 5-16
5.9 Emission Control Systems 5-18
5.9.1 Paniculate Removal Efficiency 5-19
5.9.2 Electrostatic Precipitators 5-20
5.9.3 Baghouses 5-21
5.9.4 Mechanical Collectors 5-22
5.9.5 Stack Gas Treatment 5-23
5.10 Radionuclide Enrichment in Coal Fly Ash 5-24
111
-------
CONTENTS (Continued)
Page
6. Estimates of Radionuclide Plant Emissions and Health Risks to Surrounding
Populations 6-1
6.1 Statistical Profile of Relevant Plant Parameters 6-1
6.1.1 Coal 6-2
6.1.2 Gas-Fired Plants 6-4
6.1.3 Oil-Fired Plants 6-7
6.2 Emission Estimates 6-8
6.3 Estimates of Health Risks 6-13
6.3.1 Summary of CAP-93 Model 6-14
6.3.2 Summary Findings of Population Risks 6-16
6.3.3 Limitations 6-20
References R-l
Appendix A Trends in Refinery Technology and Their Impact on the Radionuclide
Content of Residual Fuel Oil A-l
Appendix B Primary Data Extracted from Radionuclide Analysis of Oil Samples
(Westinghouse Study) B-l
Appendix C Health Risks Associated with Low Doses of Radiation C-l
Appendix D An Analysis of Uncertainties in Risks for Two Selected Sites D-l
IV
-------
TABLES
Number Page
ES-1 Profile of Fossil-Fueled Steam-Electric Production for 1990 ES-2
ES-2 Utilization and Radionuclide Content by Coal Rank ES-3
ES-3 Estimates of Average Radionuclide Concentrations in 42 Residual Fuel Oil
Samples ES-5
ES-4 Average Annual Radionuclide Emissions per Operating Boiler Unit and
per Billion kWe-hr Electricity Generated ES-7
ES-5 Frequency Distribution of Lifetime Fatal Cancer Risks for All PlantO ES-10
ES-6 Plants with the Highest Estimated Maximum Individual Risks ES-11
1-1 Fossil Fueled Steam-Electric Operable Capacity 1-3
1-2 Net Utility Generation by Energy Source 1-3
1-3 Fossil Fuel Consumption for Electricity Generation in 1990 1-4
1-4 Fossil Fueled Steam-Generating Units: Planned Additions for 1991
through 2000 1-5
1-5 Fossil Fuel Consumption 1990, 2000, and 2010 1-5
2-1 Rank and Characteristics of U.S. Coal 2-2
2-2 U.S. Coal Reserves by State and Mining Method 2-5
2-3 U.S. Coal Production and Coalbed Thickness by Major Coalbeds and Type of
Mining - 1990 , 2-6
2-4 Utility Coal Supplies by Rank 2-8
2-5 Utility Coal Use by State of Origin, 1990 2-8
2-6 Coal Suppliers by Quantity Supplied 2-9
2-7 Representation of Selected States in USCHEM 2-12
2-8 Distribution of Uranium Concentration in Coal 2-13
2-9 Distribution of Uranium Concentration by Coal Rank and Origin 2-14
2-10 Thorium Analyses from USCHEM Data Base 2-16
2-11 Sample Analysis for Herrin and Springfield Coalbeds, Illinois 2-17
2-12 Uranium Sample Analysis for Wilcox and Jackson Lignite Coalbeds 2-17
2-13 Major Decay Products of Uranium-238 2-19
2-14 Major Decay Products of Thorium-232 2-19
2-15 Summary Values of USCHEM Data Base 2-20
3-1 United States Production Wells, 1989 3-2
3-2 Imports of Natural Gas to the United States 3-4
3-3 Natural Gas Consumption in 1990 3-4
3-4 Gas Consumed for Electric Generation, by State, 1975-1990 3-5
3-5 Summary of Radon-222 Concentrations in Natural Gas at Production Wells .... 3-7
3-6 Liquefied Petroleum Gas v . . 3-10
3-7 Radon-222 Concentration Measured in Gas Plant Processing Streams 3-12
-------
TABLES (Continued)
Number Page
3-8 Radon-222 Concentrations of Natural Gas in the Houston, Texas Area 3-17
3-9 Radon-222 Concentrations Found in Gas Supplies of Southern and Western
Cities 3-18
3-10 Radon-222 in Natural Gas Supplied to PG&E Facilities 3-19
3-11 Radon-222 Concentrations in Natural Gas Distribution Lines 3-20
4-1 Crude Oil Production by State 4-3
4-2 U.S. Imports of Crude Oil by Country of Origin 4-4
4-3 Producing Wells by State and Type 4-6
4-4 Twenty Leading Companies-U.S. Refining Capacity 4-8
4-5 Number and Capacity of United States Petroleum Refineries 4-11
4-6 Percentage Yields of Refined Petroleum Products from Crude Oil in the
United States, 1990 4-13
4-7 Trace Element Concentrations in Residual Oil 4-14
4-8 Average Radionuclide Content for 28 Residual Oil Fuel Samples 4-15
4-9 EPA Analysis of 6 Residual Fuel Oil Samples 4-17
4-10 Uranium Content in 24 Domestic Crude Oil Samples 4-19
4-11 Trace Element Contents of Some Crude Oils 4-20
4-12 Radionuclide Activity in Crude Oil Samples Analyzed by EPA 4-21
4-13 Derived Uranium Values in Fuel Oil from Analyses of Fuel Oil-Ash Samples . . 4-22
4-14 Radioactivity in Fuel Oil and Coal Ash 4-23
4-15* Derived Radioactivity Values for Coal and Fuel Oil from Data Contained
in Table 4-14 -. 4-23
4-16 Comparison of Radioactivity in Oil and Coal Ash 4-24
4-17 Derived Average Radionuclide Concentrations in Oil and Coal from SERHL
Fly-Ash Study 4-25
4-18 Summary of Past Study Data for Deriving the Radionuclide Content for
Residual Fuel Oil 4-27
4-19 Demographic Data for Oil Samples 4-30
4-20 Physical Properties of Residual Oil Samples 4-31
4-21 Radionuciide Analysis of Fuel Oil Samples Reported by Contractor Laboratory . 4-35
4-22 Comparison of Composite Sample Analysis Reported by Contractor Laboratory,
NAREL, and NIST 4-38
4-23 Estimates of Average Radionuclide Concentrations in 42 Residual Fuel Oil
Samples 4-42
VI
-------
TABLES (Continued)
Number Page
5-1 Effect of Burner Adjustment on Gas Temperature 5-6
5-2 Coal Ash Distribution by Boiler Type 5-14
5-3 Distribution of Bottom Ash Collection by Coal Types 5-15
5-4 Percent Distribution of Paniculate Emission Control Systems 5-18
5-5 Effect of Various Coals on Electrostatic Precipitation Efficiency 5-21
5-6 Enrichment Factors for Radionuclides 5-26
6-1 Summary of Electrical Power Generation for 1990 6-2
6-2 Statistical Profile of Coal-Fired Units 6-3
6-3 Distribution of Generating Capacity of U.S. Coal-Fired Plants for 1990 6-4
6-4 Distribution of Generating Capacity of U.S. Gas-Fired Plants for 1990 6-6
6-5 Distribution of Gas-Fired Plants 6-6
6-6 Statistical Data of Oil-Fired Units for 1990 6-8
6-7 Distribution of Oil-Fired Plants 6-8
6-8 Estimates of Collective Annual Emissions by Designated Primary Fuel Source . 6-11
6-9 Average Annual Radionuclide Emissions per Operating Boiler Unit and per
Billion kWe-hr Electricity Generated 6-13
6-10 Frequency Distribution of Lifetime Fatal Cancer Risks for All Plants 6-17
6-11 Frequency Distribution of Lifetime Fatal Cancer Risks for Coal-Fired Units . . . 6-18
6-12 Frequency Distribution of Lifetime Fatal Cancer Risks for Oil-Fired Units .... 6-18
6-13 Frequency Distribution of Lifetime Fatal Cancer Risks for Gas-Fired Units ... 6-18
6-14' Plants with the Highest Estimated Maximum Individual Risk 6-20
vn
-------
FIGURES
Number Page
1-1 Percent Net Utility Generation by Energy Source during 1990 1-4
1-2 Uranium-238 Decay Series. 1-7
1-3 Thorium-232 Decay Series 1-8
2-1 U.S. Coal-Producing Regions 2-3
2-2 Distribution of USCHEM Data 2-10
3-1 General Flow Diagram for Production and Distribution of Natural Gas and LPG . 3-9
3-2 Distribution of Gas Processing Plants in the Contiguous U.S 3-11
3-3 Least Squares Fit of the Radon-222 Concentration in Propane Versus the
Radon-222 Concentration in the Inlet Gas 3-13
3-4 Gas Utility Industry Sales 3-14
3-5 Monthly Average Concentrations of Radon-222 in Five Distribution Systems . . 3-16
4-1 Petroleum Administration for Defense (PAD) Districts 4-12
5-1 Flame Adjustment in a Pulverized Coal Boiler 5-5
5-2 Temperature Profile through a Pulverized Coal System 5-13
5-3 Inlet Particle-Size Distribution for Ash from Four Coals 5-17
5-4 Enrichment Factors of Pb-210, U-238, Ra-226, Ra-228 and Th-228 Versus
Size in Stack Fly Ash Collected Downstream from Electrostatic Precipitator . . . 5-25
6-1 Sorted Histogram of Coal-Fired Power Plants by Megawatts 6-5
6-2 Sorted Histogram of Gas-Fired Power Plants by Megawatts 6-7
6-3 Sorted Histogram of Oil-Fired Power Plants by Megawatts 6-9
Vlll
-------
EXECUTIVE SUMMARY
Shortly after the discovery of radioactivity at the turn of the century, investigators became
aware that nearly all natural materials contained trace quantities of radioactivity. Natural
radioactivity is derived from two sources. A small percentage of natural radioactivity is
derived from the interaction of cosmic radiation with specific elements (e.g., carbon-14,
tritium, etc.). The majority of naturally occurring radionuclides are classified as primordial
radioisotopes or their radioactive decay products. Primordial radionuclides are believed to
have been formed along with all other terrestrial elements except hydrogen, by nuclear fusion
reactions, neutron absorption, and beta decay in a former star, which exploded as a super-
nova (Reeves 1968).
The behavior of primordial radionuclides and their decay products in the environment is
complex. Pathways leading to significant human exposures include external radiation from .
the emiflsion of gamma rays from the ground and building materials. Internal exposure may
result from the transfer of radioactivity through root uptake by plants that serve as food to
domestic animals or are directly ingested by humans. Internal exposure may also result from
the inhalation of airborne radioactivity.
The three major fossil fuels—cpal, oil, and natural gas—contain varying quantities of the
natural occurring radionuclides of the uranium-238 and thorium-232 series and potassium-40.
When these fuels are burned to produce steam in the production of electricity, radionuclides
are entrained in the combustion gases and may be emitted into the environment. As early as
1954, Anderson, Mayneord, and Turner suggested that man's activities, in particular the
burning of coal, might significantly perturb the natural radiation environment by transferring
additional radioactivity into the air, where it is more readily available for human intake by
inhalation (Anderson 54).
Under Title HI, Section 112 of the 1990 Clean Air Act Amendment, Congress directed the
EPA to perform a study of the hazards to public health resulting from pollutants emitted by
electric utility steam-generating units (SGUs). Radionuclides are among the groups of
pollutants listed in the amendment. Over the years, the EPA has reviewed available
information and provided estimates regarding the radionuclide content of fossil fuels,
environmental emissions, human exposure, and health risks. This information has been
reported by the EPA in several reports, including the Background Information Document
ES-1
-------
supporting the decision ndt to regulate radionuclide emissions from coal-fired boilers issued
in 1989 (EPA 89). This report updates previously published data and estimates with more
recently available information regarding the radionuclide contents of fossil fuels, associated
emissions by steam-electric power plants, and potential health effects to exposed population
groups.
In 1990, a total of 2,298 generating units fired with fossil fuels produced 1,939 billion
kilowatt-hours of electricity. Fossil-fueled plants produced nearly 69% of the total electricity
generated. The remaining 31 % of total electric generation came from nuclear power,
hydropower, geothermal and other sources. Table ES-1 provides a profile of fossil-fueled
electric production for 1990, including quantities of coal, oil, and natural gas consumed.
Projections of planned additions of fuel-specific fossil plants and fuel consumptions are
largely based on economic growth and the relative costs to produce electricity among the
major sources of fuels, including nuclear power.
Table ES-1. Profile of Fossil-Fueled Steam-Electric Production for 1990
Primary Energy.
Source
U.S. Total
Coal
Petroleum
Natural Gas
Number of
Units
2,298
1,250
329
719
Generator
Namtpiate
(,,,.- __.. ...ft »\
megawaiu)
475,421
322,429
52.519
100.473
Electric
f*jm»rrifiniii
(billion kWc-br)*
1,939
1,558
117
264
Fad
Consumption
N/A
786 MST"
200.3 Mbbl*'
2.77 Tcfe>
* billion kWe-hr is billion kilowatt hours of electricity.
M MST is millions short tons (1 short ton equals 2000 Ibs.).
w Mbbl is million barrels (1 barrel equals 42 gallons).
Tcf is trillion cubic feet at standard temperature and pressure.
The Natural Radionuclide Content in Fossil Fuels
Coal. The decay series of uranium and thorium constitute the major radionuclides contained
in coal. Uranium-238 has thirteen major radioactive decay products while thorium-232 has
nine. For coal, it is generally assumed that primary members within each of the two decay
series are in secular equilibrium. Secular equilibrium denotes that the radioactivity
ES-2
-------
concentrations among primary decay chain members are constant. A national data base of
nearly seven thousand coal samples was analyzed with regard to uranium and thorium content
of the major ranks of coal used by utilities. Concentrations spanned a wide range of values
that were log-normally distributed. Table ES-2 summarizes the data by providing the
geometric mean concentration values expressed in units of parts per million and identifies the
relative percent utility consumption of coal types.
Table ES-2. Utilization and Radionuclide Content by Coal Rank
Coal Rank
Bituminous
Sub-bituminous
Lignite
Percent
Utilization
69.0
24.7
6.3
Average Uranium
(ppm)
1.24
1.07
1.41
Average Thorium
(ppm)
2.18
2.28
2.38 .
Concentration values expressed in parts per million (ppm) are readily converted to
radioactivity concentrations by means of the specific activity values for uranium-238 and
thorium-232. For U-238, 1 ppm is equal to 0.33 picocuries per gram (pCi/g) of coal; for
Th-232, 1 ppm is equal to 0.11 pCi/g of coal. For example, with an average content of
1.24 ppm uranium and 2.18 ppm thorium in bituminous coal, there is a corresponding
activity of 0.41 pCi/g for each member of the U-238 series and 0.24 pCi/g for each member
of the Th-232 series.
The radionuclide content of coal is not unique when compared to other natural materials. In
fact, it is generally assumed that the average radioactivity of the earth's crust (i.e., soil and
rocks) is about twice that of coal (UNSCEAR 1982).
Natural Gas. Radioactivity in natural gas is almost exclusively radon-222, which migrates
from proximal geologic formations into gas reservoirs. In 1989, the American Gas
Association identified 262,482 production wells that yielded more than 18 trillion cubic feet
(Tcf) of natural gas. An additional 1 .S3 Tcf of gas were imported primarily from Canada.
About 2.77 Tcf of gas were consumed by utilities to produce electricity.
The radon content of natural gas at the wellhead has been measured in thousands of wells
over several decades. However, these measurements are of limited use for estimating radon
ES-3
-------
concentrations at the point of consumption for several reasons. Radon concentrations vary
by geographic location and over time. Also, radon content is markedly reduced when natural
gas is processed to remove commercially valuable heavier hydrocarbons (ethane, propane,
butane). Further reductions in radon concentrations reflect the natural decay that occurs
during the gathering, processing, and distribution/storage of gas prior to consumption.
A more meaningful approach is to assess the radon content in gas distribution lines.
Analyses of gas in the distribution lines eliminates well-to-well variations and accounts for
radon reduction from processing and natural decay. Radon measurements of natural gas in
distribution lines near the point of consumption suggest an average value of 20 pCi per liter.
In this report, therefore, radon emissions from gas-fired boilers are based on a radon
concentration of 20 pCi per liter of processed gas.
Oil. Residual fuel oil denotes a general classification of fuel obtained as liquid still bottoms
from the distillation of chide oil. Non-radiometric analyses show crude oil and various
petroleum products may contain as many as 60 different metals in measurable quantities.
Uranium and thorium are among the trace metals commonly found in crude oil and
petroleum products. The presence of these two radioactive trace metals also implies the
presence of their radioactive decay products.
A comprehensive literature search, however, revealed that data specific to the radionuclide
content of residual fuel oil are not only sparse but are considerably more difficult to interpret
than those for coal or gas. Contributing to die difficulty in data interpretation is the absence
of secular equilibrium among primary members of the U-238 and Th-232 decay chains. Due
to the paucity of data, die EPA concluded that there was a need for additional data and '
conducted its own study.
The EPA enlisted the help of the Utility Air Regulatory Group (UARG) and the Electric
Power Research Institute (EPRI) to solicit the voluntary participation of individual utilities in
providing samples of residual oils for radioanalysis. The selection of a utility was based on
the utility's geographical location, along with its generator nameplate capacity, capacity
factor, and/or annual fuel-oil consumption. Selection, therefore, favored larger facilities with
the highest capacity factors/fuel consumption and accounted for radionuclide variability based
on origin of crude oil.
ES-4
-------
In total, 12 utilities provided 42 samples of residual fuel oil for analysis. Participating
utilities represented major regions of the United States where fuel oil serves as a primary fuel
source. Quantitatively, the 12 utilities had an annual consumption of about 2,000,000,000
gallons, which was estimated to be about 24% of the fuel oil consumed by all U.S. oil-fired
units.
Radionuclide analysis, data interpretation, and data verification involved the efforts of a
major commercial analytical laboratory, the EPA's National Air and Radiation Environmental
Laboratory (NAREL), and the National Institute of Standards and Technology (NIST).
Table ES-3 provides estimates of the average radionuclide values of the 42 residual fuel oil
samples evaluated in the EPA study. Values are well within the range of the limited study
data reported by others and support the conclusion that the radionuclide content of residual
fuel oil is low relative to coal.
Table ES-3. Estimates of Average Radionuclide Concentrations
in 42 Residual Fuel Oil Samples
U-238 Series
U-238
Th-234
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pt>-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232 Series
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Po-212
Coacenttatioa (pCi/g)
0.0018
0.0018
0.0018
0.0034
0.0068
0.0043
0.0043
0.0043
0.0043
0.0043
0.0043
0.44
0.44
0.44
v-OBcenuauon (pu/g}
0.0030
0.068
0.068
0.068
0.068
0.068
0.068
0.068
0.068
ES-5
-------
Radionuclide Emissions from Fossil-Fueled Plants
Radionuclide emissions from electric utility units are affected by the radionuclide content in
fossil fuel, by plant design features, and by operating parameters. Important design features
involve the size of the plant, type of furnace used, and the emission control systems designed
to remove pollutants from the flue gas. The most significant operational factors, which
dictate the rate of fuel consumption, involve the percentage of time a plant is operating, the
power level, and the efficiency by which a plant converts thermal energy to electric energy.
In this study, estimates of radionuclide emissions and associated human health risks are based
on fossil-fired boiler units with generating capacities of 25 MWe or more. The 25 MWe
selection criterion reflects the low probability of significant emissions for small plants
regardless of unit-specific operating parameters. Of the nation's 2,298 boiler units (Table
ES-1), 1,748 units have a generating capacity of 25 MWe or more.
From data reported to the Edison Electric Institute that includes annual fuel consumption and
paniculate removal efficiencies, emissions were estimated for each of the 1,748 boiler units
and aggregated by plant affiliation. (The 1,748 fossil-fired boiler units represent a total of
684 utility plants.) These unit- and plant-specific emission data are contained in a separate
addendum to the report. Table ES-4 provides average annual emissions per operating boiler
unit, as well as per billion kWe-hr of electricity generated. For coal-fired units, the average
annual emissions for particulates range from a fraction of a millicurie (mCi) to several
millicuries among primary radionuclides.
Although the average radionuclide content of residual fuel oil is two to three orders of
magnitude lower than that of coal, Table ES-4 reveals that average emission rates are nearly
comparable. This is explained by the fact that, unlike coal-fired units, the majority of oil-
fired units lack paniculate emission control systems that remove radionuclides from the flue
gas with efficiencies of 95% or more. Due to the fact that coal-fired units on the average
have a higher capacity factor, the degree of comparability between coal-fired and oil-fired
units is further enhanced when emissions are defined per unit of billion kilowatt-hours.
Paniculate emissions for units designated as gas-fired are generally small when compared to
either coal- or gas-fired units. Moreover, radionuclide emissions other than radon from units
designated as gas-fired principally result from the combustion of a secondary fuel.
ES-6
-------
Table ES-4. Average Annual Radionuclide Emissions per Operating Boiler Unit
and per Billion kWe-hr Electricity Generated
Radionuclide
U-238
Th-234
Pa-234tn
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
Tl-208
K-40
Emission Rates
Per Operating Unit (mCi/y)
Coal
2.3E+00
1.2E+00
1.2E+00
1.2E+00
2.3E+00
1.2E+00
1.7E+00
3.0E+02
5.6E+00
5.6E+00
1.2E+00
5.6E+00
5.6E+00
1.2E+00
5.6E+00
7.1E-01
l.OE+00
7.1E-01
7.1E-01
l.OE+00
1.6E+02
3.5E+00
3.5E+00
7.1E-01
2.1E-01
7.8E+00
Gas
1.3E-02
1.3E-02
1.3E-02
1.3E-02
2.5E-02
4.9E-02
2.9E-02
2.5E+03
3.1E-02
3.1E-02
3.1E-02
3.1E-02
3.1E+00
3.1E+00
3.1E+00
2.1E-02
4.7E-01
4.7E-01
4.7E-01
4.7E-01
5.7E-01
4.7E-01
4.7E-01
4.7E-01
1.4E-01
6.2E-03
Oil
1.1E-01
1.1E-01
1.1E-01
1.1E-01
2.1E-01
4.1E-01
2.6E-01
3.8E+02
2.7E-01
2.7E-01
2.7E-01
2.7E-01
2.7E+01
2.7E+01
2.7E+01
1.8E+01
4. IE +00
4.1E+00
4.1E+00
4.1E+00
8.4E+00
4.1E+00
4. IE +00
4.1E+00
1.2E+00
5.2E-03
Per Billion kWe-hr Generated (mCi/y)
Coal
l.SE+00
7.7E-01
7.7E-01
7.7E-01
l.SE+00
7.7E-01 .
1.2E+00
2.0E+02
3.8E+00
3.8E+00
7.7E-01
3.8E+00
3.8E+00
7.7E-01
3.8E+00
4.7E-01
7.1E-01
4.7E-01
4.7E-01
7.1E-01
1.1E+02
2.4E+00
2.4E+00
4.7E-01
1.4E-01
S.3E+00
Gas
2.6E-02
2.6E-02
2.6E-02
2.6E-02
4.9E-02
9.5E-02
5.7E-02
4.9E+03
6.0E-02
6.0E-02
6.0E-02
6.0E-02
6.0E+00
6.0E+00
6.0E+00
4.1E-02
9.1E-01
9.1E-01
9.1E-01
9.1E-01
1.1E+00
9.1E-01
9.1E-01
9.1E-01
2.7E-01
1.2E-02
Oil
1.8E-01
1.8E-01
1.8E-01
1.8E-01
3.4E-01
6.7E-01
4.3E-01
6.2E+02
4.4E-01
4.4E-01
4.4E-01
4.4E-01
4.4E+01
4.4E+01
4.4E+01
3.0E-01
6.7E+00
6.7E+00
6.7E+00
6.7E+00
1.4E+01
6.7E+00
6.7E+00
6.7E+00
1.9E+00
8.5E-03
Estimates of Population Health Risks
Radionuclide emissions from utility boilers may result in public exposure from multiple
pathways that include (1) external radiation from activity suspended in air or deposited on the
ground and (2) internal exposure from the inhalation of airborne contaminants or ingestion of
ES-7
-------
contaminated food products. Although the potential health risks are essentially independent
of whether a dose was internal or external, the assessment of internal exposures are
considerably more complex. For ingested or inhaled radionuclides, dose assessment requires
biokinetic information that describes the distribution and retention of individual nuclides, the
type of radiation emitted, and the amount of energy absorbed by individual target
tissues/organs.
Estimates of population doses from chronic atmospheric releases require the use of a
computer code that accounts for atmospheric dispersion, radionuclide concentrations in
environmental media, and radionuclide intakes by inhalation and ingestion. In support of
National Emission Standards for Hazardous Air Pollutants, the EPA, with support from Oak
Ridge National Laboratory, developed the CAP-88 computer model. The CAP-88 (which
stands for Clean Air Act Assessment Package-1988) computer model is a composite of
computer programs, data bases, and associated utility programs.
CAP-88 programs are considered among the best available verified models for population
dose and risk assessment for radionuclide air emissions. For a given facility, atmospheric
releases and dose assessment may be modeled for up to six independent sources that take into
account plant- and site-specific model parameters.
Since it was first introduced, CAP-88 has been revised periodically to reflect changes in data
base information and improved risk methodologies. For this study, the most recent version
of the code, designated as CAP-93, was used. CAP-93 contains a correction to the
procedure used to calculate wet deposition of radionuclides from the plume.
For low doses of radiation, potential health effects may not appear for years of even decades
following exposure. Such delayed effects are termed "stochastic" and are thought to result
from highly selective molecular changes in individual cell(s). Although these highly selective
changes occur rarely, when they do, the altered cell may develop into cancer. Among the
stochastic effects that have been associated with radiation exposure, medical scientists
consider cancer induction the primary health effect of concern.
A key characteristic of a stochastic effect is that the severity of the effect is not dose
dependent. However, the probability that a stochastic event (i.e., cancer) may occur is
dictated by the radiation dose. The stochastic nature of low dose radiation is not unique but
is universal to all carcinogenic agents.
ES-8
-------
The current method for estimating radiation risks relies on select.human studies where cancer
rates were observed at a higher incidence among exposed individuals than one would
normally expect to occur spontaneously. The most intensely studied human population is the
Japanese atomic bomb survivors ofHiroshima and Nagasaki. Data through 1985 show that
among the 76,000 individuals studied, 5,935 survivors have died of cancer from all causes.
It is estimated that about 340 of these cancers (i.e., 80 leukemias and 260 non-leukemias)
were the result of radiation exposure.
The data also define a dose response in which increasing doses yielded an increased
percentage of excess cancers, especially for leukemia. However, some numerical estimates
embody substantial statistical uncertainties as to the number of cancer deaths induced by
radiation. Thus, for doses less than 50,000 mrem (50 rem), the small number of excess
cancers above normal expected levels may reflect random fluctuations that are not linked to
radiation exposure. When doses exceeded 50,000 mrem (50 rem), the number of excess
cancers is sufficient to support a causal link to human cancers.
For low-dose exposures, a causal link and a quantitative relationship between radiation dose
and cancer has not been established. Yet, scientists conservatively assume that any dose of
radiation, no matter how small, may pose a risk to human health. Estimates of health risks
from low-level radiation are, therefore, derived by extrapolating risks from high doses to
lower doses using a linear non-threshold dose response model contained in the CAP-88 and
CAP-93 computer codes.
CAP-93 assesses risk for a circular grid that is defined by sixteen sectors and a radial
distance of 50 kilometers around a facility. Risk to the population is determined by summing
individual risks by distance and sector for the 0-50 km grid around each assessed facility.
Risk to the maximally exposed individual(s) corresponds to that location (i.e., distance and
sector) of highest'exposure where individuals are believed to reside.
The population risk frequency distribution identifies the number of people at various levels of
risk. The risk categories are divided into powers of ten, in which the individual lifetime
cancer risk ranges from one chance in ten to less than one chance in a million. Risk data for
k
each of the '684 assessed plants are provided in a separate addendum. Only a summary of
these data is provided below.
ES-9
-------
Summary Findings. Table ES-5 gives the distribution of fatal cancer risks to the combined
populations residing within the 50-km radii of the 684 fossil-fueled electric utility plants.
The aggregate of assessed populations living within a 50-km (35 mile) radius of a plant is
.**•
estimated to be 196.1 million, which represents approximately 75% of the U.S. population.
The individual lifetime risk of fatal cancer to more than 99.9% of the assessed population
(i.e., 196.1 million) is less than one chance in a million. The data further suggest that under
current operating conditions, there are no instances in which the release of radioactivity is
likely to result in a lifetime fatal cancer risk to any one person that is equal to or greater than
one chance in ten thousand. It is estimated that about 1,027 individuals residing within a
50-km distance of a plant may receive radiation exposures for which the lifetime risk is
between 1 in 10,000 and 1 in 100,000 (i.e., 1E-04 to 1E-05).
Table ES-5. Frequency Distribution of Lifetime Fatal Cancer Risks for All Plants
Lifetime Cancer
Risk Range
l.OE+OOto l.OE-01
l.OE-01 to .OE-02
1.0E-02to .OE-03
1.0E-03to -OE-04
1.0E-04to .OE-05
l.OE-OSto .OE-06
Less Than .OE-06
Number of
People'
0
0
0
0
1.027
^5.743
196,000.000
Average Individual
Lifetime Risk
O.OE+00
O.OE+00
O.OE+00
O.OE+00
1.3E-05
2.2E-06
1.2E-07
Deaths per Year
in this Risk Range
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.92E-04
3.06E-03
3.32E-01
Death per Year in this
Risk Range or Higher
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.92E-04
3.26E-03
3.36E-01
It must be also pointed out that the distribution of individual risks within each risk range is
heavily skewed toward the lower value. This is evidenced by the fact that the average
individual lifetime risk is a small fraction of the midpoint value within each of the risk
ranges. Correspondingly, the probability of a single fatal cancer occurrence within, the
highest risk group of 1,027 individuals is less than 2 chances in 10,000 per year. For the
entire assessed population of 196,100,000 within 50 km of these plants, the estimated cancer
risk attributable to radionuclide emissions from electric utility SGUs is less than 1 cancer
death per year (i.e., 3.36E-01 deaths/year is the risk equivalent of about 1 in 3 chances that
a single cancer death will occur in a year). Exposures and risks to individuals residing
beyond 50 km are not explicitly evaluated. However, maximum individual risks (MIRs) at
such distances would be 1E-7 or less, and the overall cancer incidence would not, based on
removal of about 1/3 of the emitted radionuclides within 50 km, exceed about 1 death/year.
ES-10
-------
Based on radionuclide emissions and plant-specific/site-specific data, CAP-93 also calculates
the MIR for each of the 684 plants. Table ES-6 characterizes those plants with the highest
estimated MER values expressed in lifetime fatal cancer risk. There were a total of 17 plants
for which the lifetime risk of fatal'cancer to the MIR is estimated to exceed l.OE-05. The
highest MIR value of 3.0E-5 corresponds to a 5-unit coal-fired facility that generated 3,340
megawatts of electricity in 1990. Of the 17 plants with the highest MIR values, 11 are
exclusively designated as coal-boilers. Only 2 facilities are identified as exclusively oil-fired
plants. The remaining 4 plants are represented by a combination of boilers, where coal is at
least one of the designated primary fuels.
Table ES-6. Plants with the Highest Estimated Maximum Individual Risk (MIR)'
Plant Name
Plant #222
Plant #247
Plant #60
Plant #301
Plant #251
Plant #406
Plant #256
Plant #17
Plant #133
Plant #318
Plant #672
Plant #668
Plant #82
Plant #207
Plant #253
Plant #489
Plant #651
MIR*
3E-S
3E-5
2E-5
2E-5
2E-5
2E-5
2E-5
2E-5
2E-5
1E-5
1E-5
1E-5
1E-5
1E-5
1E-5
I-E5
1E-5
Coal-Fired
Units
5
4
4
2
4
4
3
2
6
8
7
3
4
MWe
3.340
900
3.160
750
1.540
• 2,777
1.728
1.135
1,100
1,965
2,304
2,052
1.872
Gas-Fired
Units
3
MWe
262
Oil-Fued
Units
2
2
2
2
6
MWe
1.112
66
804
558
372
MIR is the marimuiti individual risk expressed as lifetime fatal cancer risk.
ES-11
-------
Because of limitations in the GENPOP computer code used for identifying locations of
individuals, the maximum individual risks (MIRs) shown for each plant should be viewed
with caution; errors of a few hundred meters in the location of individuals can result in an
over or under estimate of risk by factors of 2 or more. UARG re-estimated the risks for the
17 plants with the highest MIRs using refined population grids. Their results show lower
MIRs for the majority of these plants, but their highest MIR of 1E-5 is consistent with the
EPA's estimates. Thus, the EPA believes the GENPOP methodology is sufficiently accurate
to-establish die magnitude of MIRs for all SGUs.
ES-12
-------
1. BACKGROUND INFORMATION
Coal, natural gas, and residual fuel oil contain significant amounts of impurities. When
these fossil fuels are burned as fuels to produce electricity, a wide variety of gases and
particulates are produced that have the potential to be released into the atmosphere as
pollutants. Prominent gases include sulfur dioxide (SO;) and nitrogen oxides (NOJ. Non-
combustible impurities encompass a large number of metals including some which are
radioactive. Although SO2 and NO, emissions have been regulated by the Environmental
Protection Agency (EPA) for more than twenty years, some impurities contained in fossil
fuels have only recently been included as hazardous air pollutants.
Under Title HI, Section 112 of the 1990 Clean Air Act Amendments, Congress requires the
EPA "... to perform a study of the hazards to public health reasonably anticipated to occur
as a result of emissions by electric utility steam-generating units of pollutants listed under
subsection (b) after imposition of the Act."
Radionuclides, as a group, represent one of the pollutants listed. The EPA has studied the
hazards of radionuclide emissions to the public health from electric utility steam-generating
units (SGUs) in the past (EPA 89, EPA 1984, EPA 1980, EPA 1977, EPA 1973). Since the
issuance of these reports, new information has been made available, which provides an
improved understanding of potential environmental emissions of pollutants. This report
summarizes previously published data and supplements those data with the most currently
available information regarding emissions of radionuclides by electric utility SGUs and
estimates of population health risks.
A comprehensive assessment of radionuclide emissions to the environment by SGUs is
complex. Primary parameters affecting emissions include (1) electric power generation
quantities, (2) the amounts of radionuclides present in each of the three fossil fuels at time of
combustion, (3) the percentage of radionuclides entering the flue gas, and (4) operating
parameters, inclusive of emission control systems, aimed at removing pollutants from the
flue gas.
These primary factors, in turn, are affected by a host of secondary factors that may be
described as "controllable" and "non-controllable." For example, non-controllable secondary
factors affecting radionuclide content of fossil fuels include the natural geological formations
1-1
-------
containing uranium and thorium in proximity to fossil-fuel reservoirs, the chemical state(s) of
uranium, thorium, and their decay products, the quality of the fuel, and site-specific
meteorological and geological factors affecting the migration of radionuclides into or out of
fossil-fuel reservoirs.
Controllable secondary factors potentially affecting the radionuclide content in fuel include
the method(s) by which fuels are extracted from their natural reservoirs, processed, and
refined for commercial use. In the case of natural gas in which radon-222 is the dominant
radionuclide, elapsed time between gas extraction and combustion plays a critical role in
affecting emissions. Other important factors affecting emissions are conditions of fuel
combustion and furnace design.
1.1 ELECTRIC POWER GENERATION: PAST, PRESENT, AND FUTURE
Potential radionuclide emissions from electric utility SGUs are essentially proportional to the
amount of electricity generated by these units. Since electricity is used in the production and
utilization of goods and services, the demand for electricity is not a constant but varies with
economic conditions. During times of economic growth as production and consumption of
goods and services increase, the need for generating electricity increases proportionately.
For 1990, there were 2,298 operational fossil-fueled steam-generating units in the United
States (Table 1-1). Among fossil plants, coal-fired SGUs constituted 1,250 and represented
nearly 68% of the operable capacity. The 719 gas-fired SGUs and 329 oil-fired SGUs
represented 21% and 11%, respectively, of the total capacity of fossil-fueled SGUs.
Capacity values cited in Table 1-1 represent the full-load continuous rating of a generator.
Plants, however, cannot be operated at full-load continuously. Most modern plants have a
capacity factor of about 0.65. This value represents the ratio of the average load on the
plant to the nameplate capacity.
In 1990, the collective electricity generated using fossil fuel, nuclear power, hydropower,
and geothermal sources was 2,807 billion kilowatt-hours (Table 1-2). Of the total electricity
generated in 1990, fossil fuel provided 69% or 1,939 billion kilowatt-hours, corresponding to
55.5%, 9.4%, and 4.1% contributions by coal, natural gas, and oil, respectively (Figure
1-1).
1-2
-------
Table 1-1. Fossil-Fueled Steam-Electric Operable Capacity
as of December 31, 1990
Primary Energy
Source
U.S. Total
Coal
Petroleum
Gas
Operable Capacity
Number
of Units
2.298
1.250
329
719
Generator
Nameplate
(megawatts)
475.421
322,429
52.519
100.473
Summer
Capability
(megawatts)
445.559
299,876
49,437
96,246
Winter
Capability
(megawatts)
448.274
301,872
49.896
96.506
Source: DOE 1991a
Table 1-2. Net Utility Generation by Energy Source (Billion kilowatt-hours)
SGU Energy Source
Coal
Natural Gas
Oil
Nuclear Power
Hydropower
Geothermal/Other
Total Net Utility Generation
1970
704
373
184
22
248
1
1.532
1980
1,162
346
246
251
276
6
2.287
1989
1.554
267
158
529
265
11
2,784
1990
1.558
264
117
• 577
280
11
2,807
Source: DOE 1991a
Table 1-3 provides a. breakdown of the quantities and energy equivalents of coal, oil, and
natural gas consumed by the electric utilities. The consumption of fossil fuels per unit of
electricity generated can be standardized by means of their respective equivalent heat values.
The 1990 unit heat values are: anthracite coal—11,122 Btu per pound; bituminous and
lignite coal—10,632 Btu per pound; oil—5,800,000 Btu per barrel; natural gas—1031 Btu per
cubic foot. The collective fossil fuel energy of 2.02 x 1015 Btu consumed by utilities for the
generation of electricity is about 25% of the 1990 U.S. gross consumption (8.13 x 10" Btu)
from all energy sources (DOE 1991a).
1-3
-------
Natural Gas
9.4%
Geothermal/Other
0.4%
Hydropower
10%
Nuclear
20.5%
Figure 1-1. Percent Net Utility Generation by Energy Source for 1990
Table 1-3. Fossil Fuel Consumption for Electricity Generation in 1990(I)
SOU Energy Source
Coal
Natural Gas
Oil
Quantity
786 million ST*>
2.78 Tcfe>
200,3 Mbbl(d)
Btu Equivalence
(trillion)
16,189
2,876
1,161
'" Source: U.S. Energy Information Administration, Monthly Energy Review,
March 1991.
"". ST is short ton (1 short ton equals 2000 Ibs.).
(c) Tcf is Trillion cubic feet at standard temperature and pressure.
<
-------
in the United States. DOE's Energy Information Administration (EIA) expects that fuel
switching from natural gas to residual oil in dual-fired SGUs will occur in some regions as
the price of natural gas climbs and residual fuel oil becomes a more attractive economic
alternative. Projections of planned additions of fossil-fuel-fired SGUs and consumption are
provided in Table 1-4 and Table 1-5. These projections take into account uncertainties
regarding economic factors.
Table 1-4. Fossil-Fueled Steam-Generating Units: Planned Additions
for 1991 through 2000
Primary Energy
Source
U.S. Total
Coal
Petroleum
Gas
Planned Additions
Number
of Units
40
35
1
4
Generator
Nameplate
(megawatts)
17,389
16.380
259
750
Summer
Capability
(megawatts)
16.506
15.574
244
688
Winter
Capability
(megawatts)
16.566
15,609
249
708
Source: DOE 1991a
Table 1-5. Fossil Fuel Consumption 1990, 2000, and 2010
Sector
Fuel Consumption ~
(quadrillion Btu)
Coal
Natural Gas
Oil
1990
16.2
2.9
1.3
Projections
2000 2010
High
High Oil Economic
Price Reference Growth
17.9 17.9 18.1
4.2 4.8 S.5
1.6 2.1 2.6
High
High Oil Economic
Price Reference Growth
20.4 22.6 24.3
5.3 5.7 5.4
1.8 1.7 2.8
Source: AEO 1991
1-5
-------
1.2 THE NATURAL RADIONUCUDE CONTENT IN FOSSIL FUELS
A number of naturally occurring radionuclides are primordial elements, i.e., associated with
the formation of the earth. Most primordial radionuclides are isotopes of the heavy elements
of the three radioactive series headed by uranium-238, thorium-232, and uranium-235. Each
of these radionuclides undergoes a series of radioactive decays before terminating with a
stable non-radioactive element. Because of the low isotopic abundance of U-23S, this
radionuclide and its daughter products contribute negligibly to the total radionuclide content
of fossil fuels and will, therefore, not be considered in this report.
Another independent primordial radionuclide that exists in significant concentrations is
potassium-40 (K-40). K-40 can be estimated from the total potassium content in fossil fuels
since the element has a constant isotopic ratio in nature. Human exposure to K-40 emitted
from SGUs is limited to the external gamma exposure resulting from ground deposition (its
contribution to potential human internal exposure above normal levels is insignificant due to
the homeostatic regulation of potassium in the body).
The decay series and daughter products of U-238 and Th-232 are provided in Figures 1-2
and 1-3. The successive relationships among radionuclides in the chain are directed by the
modes of decay and the physical half-lives of each species. In the absence of chemical or
physical separation, the members of a series attain a state of secular equilibrium, wherein the
rate of decay of each nuclide is essentially equal to that of the first nuclide that heads the
series. Secular equilibrium is always the case on a global basis for each series, but localized
concentrations resulting from removal processes may introduce a state of disequilibrium.
Natural removal mechanisms inducing disequilibrium within the decay chain may involve the
disproportionaf leaching by ground water of a decay-chain member or the partial removal of
radon gas by diffusion.
1.3 STRUCTURE OF THE REPORT
Chapters 2, 3, and 4 provide data for estimating the radionuclide content in coal, gas, and oil
and identify specific factors which influence the radionuclide concentrations, as well as
variables that contribute to the uncertainty of such estimates. Chapter 5 characterizes plant
design specifications and operational factors of fossil fuel plants that have a direct effect on
stack emissions. Lastly, Chapter 6 provides estimates of radionuclide emissions from fossil
fuel plants and, by means of computer models, provides estimates of population health risks.
1-6
-------
u
92
Pi
91
Th
90
Ac
89
Ri
88
Fr
87
Ra
86
At
85
Po
84
Bi
83
Pb
82
n
81
Hf
80
u*».u,
(vrinium I)
451 x 10*
ytira
~U»
. U..
(uranium Q)
248x10*
/ jrtira
/PiB4-WXtV8S*)
a n 18 mmuln. J
' /
Th1*. UX,
(uranium X,)
24.1 dtyt
•
-
-
P^.UZ
U.T(0.15*)«
67 noun"'
Th»
r
»!„
(ionium)
7.52 x 104
yun
,
R."
a
(ridhim)
1622 yun
Ra8
a
'.Ra
(radon)
3.825 dip
a
(radium A)
3.05
itwiutvs
(99.98*)
Pb»4. ma'
(radium 8)
264
mmulM
r
A,™
Uiwondi
;0.02X)
a
.
Bi*4. R.C'
(radium C)
197
mlnutot
6
a
(OX>4*)
n^.iuc-
(radium CO
1J2
mmutn
PoM4.R.C
(radium C*)
1.6 xlO'4
Heond
e
(99.96*)
a
PbM.lUD
(radium 0)
22y«iri
0
(1.8 x
10-**)
U
fiunutM
_
Blno.IUE
(radium E)
lOldtyi
8
a
(5x
10'**)
TI"*. lur"
(radium E*)
4J
minuttt
9
Pono. R*F
(polonium)
138.4 dayi
0
a
n
(SUM* lltd
itottpt)
0
Figure 1-2. Uranium-238 Decay Series
1-7
-------
Th
90
Ac
89
IU
88
Pr
87
Rn
86
At
85
Po
84
Bi
83
Pb
82
Tl
81
n»Tk
(thonum)
139xl010
years
a
i
R.»M.Thl '
(mesothonum 1)
6.7 years
Ac™. MsTh, '
(mesothonum 2)
6.13 hours
B~
^a,
(ridwthonum)
1.90 years
ji
P
a
Ra°*ThX
(thorium X)
3.64 days
a
.
R«»T*
(thoron)
54.5 seconds
a
,
Po"'.ThA
(thonum A)
0.158 second
a
,
Pb^.ThB
(thonum 8)
10.6 hours
BiM2.ThC
(thonum C)
60.6 minutes
9
a
(33.7*)
•n-Thc-7
(thonum C*)
3.1 minutes-
Po*u.ThC-
(thonum C*)
-3.0x10"'
second
^(66.3»)
a
Pb^.ThD
(stable lead
isotope)
P
Figure 1-3. Thorium-232 Decay Series
1-8
-------
2. RADIONUCLIDE CONTENT OF COAL USED BY U.S. ELECTRIC UTILITIES
2.1 INTRODUCTION
More than 50% of electrical energy produced by U.S. utilities comes from coal-fired plants.
In 1990, 1,250 coal-fired units consumed about 786 million tons of coal. Of these, 1,045
units had a generating capacity of 25 MWe or greater. This chapter assesses coal use by
utilities by coal source, rank, and quantity; summarizes the current data regarding the
radionuclide content of coal; and derives statistical average and bounding estimates of
specific activities of radionuclides.
Much of the data cited in this chapter reflects information collected and published by a host
of Federal agencies. Comprehensive information concerning types, quality, quantity, and
cost of fossil fuels used to generate electricity in the United States is collected on an ongoing
basis by the Department of Energy's Energy Information Agency (EIA) pursuant to the
Federal Energy Administration Act of 1974 (Public Law 93-275). Data from steam-electric
plants with a total generator nameplate capacity of 50 megawatts or more are collected
monthly through the Federal Energy Regulatory Commission (FERC) Form 423 (Monthly
Report of Cost a"d 0"*Uty of Fuels for Electric Plants). Monthly information includes
plant-specific data; fuel consumption; Btu, ash, and sulfur content of fuel; and the fuel
source. This report is filed by approximately 225 electric utilities, representing about 90%
of all the fossil-fuel generating capacity in the United States.
Form EIA 767, "Steam-Electric Plant Operation and Design Report," is required annually
from all operators of power plants with a nameplate capacity of 10 megawatts or more. This
report is filed by approximately 320 electric utilities, representing about 95% of all fossil-
fuel generating capacity in the United States.
Data concerning the types, quality, and sources of fossil fuels used by the various individual
generating stations or plants were obtained from a compilation of the 1990 FERC Form 423
submissions, provided by DOE/EIA. This data base identified the specific fuels used and
quantities consumed by the electric utility industry (DOE 1991b).
For assessing the radionuclide content of coal, a nationwide compendium of coal data was
analyzed. Since 1975, the U.S. Geological Survey has maintained the National Coal
2-1
-------
Resources Data System (NCRDS). The NCRDS includes a USCHEM data base of trace
element and quality analyses for over 14,000 samples of coals and associated rocks. In
addition, each of the principal coal-producing States collects and archives data for the State's
coal resources.
2.2 COAL FACTS AND BACKGROUND INFORMATION
Power plants are generally designed to burn a specific fuel and, in some cases, a specific
type of coal. Coal properties impact various aspects of power plant operations, inclusive of
handling, storage, combustion, and emission control systems. Among the primary coal
characteristics are the heating value, moisture content, sulfur content, and ash content.
2.2.1 Coal Classification
Coal is classified by rank and grade. Rank is based on fixed carbon and heat content, as
calculated on a mineral-matter-free basis. Percentages of fixed carbon and heat content
generally increase from the lowest rank through the highest rank. Grade is based on the
content of ash, sulfur, and other undesirable minerals (Averitt 1975). The four major ranks
of coal, with their principal characteristics and percentage of the U.S.-coal resources, are
given in Table 2-1. U.S. coals generally decrease in rank from the eastern shore westward
with anthracite found in Massachusetts, Rhode Island, and eastern Pennsylvania; bituminous
in the Appalachian Mountains, Indiana, Illinois, and western Kentucky; low sulfur, sub-
bituminous coals in the Powder River basin of Wyoming and Montana; and lignite in Texas,
North and South Dakota, and Montana (see Figure 2-1).
Table 2-1. Rank and Characteristics of U.S. Coal
Rank
Anthracite
Bituminous
Sub-Bituminous
Lignite
Characteristics
Heat Value
(BTU/Ib)
13.000 - 15.000
10.500 - 14,000
8.300- 11,500
4.000 - 8.300
Carbon Content
(*)
90-97
85-90
75-86
50- 75-
Percent of
U.S. Reserves
2%
51%
38%
9%
Source: AGA 1991a
2-2
-------
FIELD
Bitumnious Coal
Subbitumnious Coal
Ugnite
Mole: Alaska not 10 seal*.
i Principal anthraat* deposits an in Pennsylvania. Smalt deposits occur m Alaska. Arkansas.
Colorado. Mauacftusetts-Rhod* Island. New Mexico. Ulan. Virginia. Washington.
and Wett Virginia.
Supply Regions
Northern Appalachia
Pennsylvania
dhio
Maryland
West Virginia, north
Central Appalachia
West Virginia, south
Virginia
Kentucky, east
Tennessee
Southern Appalachia
Alabama
Illinois Basin
Illinois
Indiana
Kentucky, west
West Interior
Iowa
Missouri
Kansas
Arkansas
Oklahoma
Texas
Louisiana
Northern Great Plains
North Dakota
South Dakota
Montana
Wyoming
Colorado, north
Rocky Mountains and Southwest
Colorado, south
Utah
Arizona
New Mexico
Northwest
Washington
Alaska
Source: Energy Information Administration. Office of Coal. Nuclear. Electric and Alternate Fuels.
Figure 2-1. U.S. Coal-Producing Regions
2-3
-------
2.2.2 rnal Reserves
United States coal reserves total nearly 4 trillion short tons, located principally in the
Appalachian, Mid-Western, and Rocky Mountain regions of the United States. Reserves of
bituminous, sub-bituminous, and lignite, the categories used as steam coal, equal 747 billion
tons, 486 billion tons, and 478 billion tons, respectively. However, only about 400 billion
tons of these coal reserves are considered minable by current technology, as indicated in
Table 2-2 (Averitt 1975).
The U.S. coal resources occur primarily in two contrasting structural settings represented by
(1) many large shallow basins and (2) a few very deep basins. Examples of the broad,
shallow basins with typical maximum depths for the bulk of the coal reserves are as follows:
Appalachian Basin - 3.000 feet; Eastern and Western Interior Basins - 2,000 feet; Northern
Great Plains region - 1,500 feet; Powder River Basin - 2,000 feet; San Juan Basin - 4,000
feet; and Raton Mesa - 2,000 feet. In the deep and narrower basins located in the Rocky
Mountains and the Pacific Northwest, coal beds are characterized by steep dips and narrow
marginal belts of accessible coal and coal resources 6,000 to 20,000 feet below the surface
(Averitt 1975). The coal regions are shown in Figure 2-1.
2.2.3 Coal Production
Production data for all active U.S. coal mines are compiled on an annual basis by the DOE
by company, mine, State, county, type operation, and tonnage produced. Coal production
for 1990 totaled about one billion tons from 4,056 mines in 28 States, with over half the
production coming from 3 States - Wyoming, Kentucky, and West Virginia. Approximately
75% of this was used by electric utilities for fueling steam boilers.
Coal production in the United States is approximately 60% from surface operations and 40%
from underground mining, with small amounts recovered by augering, dredging, and
reworking of old culm or waste banks. Most of the production is from a few dozen well-
known beds (Table 2-3), the characteristics and propenies of which have been extensively
studied and reported in the literature.
EIA coal production reports for 1990 listed a total of 4,056 production facilities; however,
this included many processing plants, some of which may draw from as many as several
2-4
-------
Table 2-2. U.S. Coal Reserves by State and Mining Method
(Millions of Tons)
Stale
Alabama
Alaska
Arizona
Arkansas
Colorado
Georgia
Illinois
Indiana
Iowa
Kansas
Kentucky, eastern
Kentucky, western
Maryland
Michigan
Missouri
Montana
New Mexico
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
South Dakota
Tennessee
Texas
Utah
Virginia
Washington
West Virginia
Wyoming'
Total
._ Potential Mining Method
Underground .
1,798
4.246
(a)
402
14.000
1
53.442
8.949
2.885
(a)
9.467
8.720
902
118
6.074
65.165
2.136
31
—
17.423
860
1
29.819
—
667
»
3.780
2,971
1.446
34.378
27.554
297.235
Surface
1.184
7.399
350
263
870
—
12.223
1.674
(a)
1.388
3.450
3.904
146
1
3.414
42.562
2.258
(b)
16.003
3.654
434
(b)
1.181
428
320
3.272
262
679
508
5.212
23.674
136.713
Total
2.982
11.645
350
665
14.870
1
65.665
10.623
2.885
1.388
12,917
12.624
1.048
119
9.488
107.727
4.394
31
16.003
21.077
1.294
1
31.000
428
987
3,272
4,042
3,650
1.954
39.590
51.228
433.948
(» Data insufficient to establish reserve base.
o» Less than 1 million tons.
Source: Averitt 1975
2-5
-------
Table 2-3. U.S. Coal Production and Coalbed Thickness by Major Coalbeds
and Type of Mining - 1990
-------
2.2.4 Coal Processing
Coal preparation prior to delivery toJhe utility may include one or more of the following
processes: crushing/screening to reduce the larger lumps and remove partings (inert rock,
dirt, or other non-coal material); washing.in water or solvents to reduce ash/sulfur content
which may include dense media separation, jigging, or flotation; and dewatering. Most coal
is crushed and sized to reduce oversized lumps and facilitate shipping and use by the
consumer. As much as 50% of the coal used by utilities receives some additional processing
to remove rock and mineral impurities such as shale, pyrite, etc. However, few trace
element analyses of processed or cleaned (washed) coal were found. This is to be expected
since the purpose of cleaning is to facilitate handling/combustion and reduce the ash (and
sometimes sulfur) content. This processing, whether physical or chemical, will generally not
affect the radionuclide content of the coal.
Some of the clean coal technologies being investigated and tested by industry and the DOE
involve more extensive preparation or processing of the coal prior to combustion. Processes
being tested include very fine grinding of the coal to facilitate removal of impurities and
various chemical or biological processes to leach sulfur and other mineral matter from the
coal. While these processes are directed primarily toward reduction of SO, and NO,, it is
possible that there could also be a reduction in the radionuclide content of the coal. A
potential reduction may involve radionuclides that are water soluble and are leached out by
coal washing. The pulverization of coal may also enhance the escape of radon gas, which
results in a state of disequilibrium with the short-lived daughters. However, without further
study, even an "informed guess" is outside the scope of most of the clean coal investigations.
2.2.5 Coal Users
A total of 408 coal-fired electric utility plants provided data to the DOE/EIA (DOE 91)
regarding coal use in 1990. A breakdown of coal consumption by coal rank, plants, and
tonnage is shown in the Table 2-4. Coal sources by States with tonnages, Btu, sulfur, and
ash content are given in Table 2-5.
Four plants used a total of 1,366,170 tons of bituminous coal imported from Columbia
(81%), Venezuela (16%) and Canada (3%).
2-7
-------
Table 2-4. Utility Coal Supplies by Rank
Rank*
Anthracite
Bituminous
Sub-Bituminous
Lignite
Plants**
-'3
328
110
21
Total
Tons Used
753.000
477.782,232
232.660.289
75.431,635
% Tons Used
0.1%
60.7%
29.6%
9.6%
786.627.156
(>> Reported rank of coal used may be only approximate.
*> Note: since some plants use more than one type of coal, the total
number exceeds the 408 coal-fired plants referred to in Section 2.2.5.
Table 2-5. Utility Coal Use by State of Origin, 1990
State of Origin
Alabama
Arizona
Colorado
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maryland
Missouri
Montana
New Mexico
North Dakota
Ohio
Oklahoma
Pennsylvania
Tennessee
Texas
Utah
Virginia
Washington
West Virginia
Wyoming
Subtotal
Imported
Total
Quantity
(thousand
short tons)
16.383
11.447
15.382
54.232
30.899
66
650
128.818
3.186
3.004
2.343.
35,627 '
22,644
22,983
30,103
917
50,488
4,618
49,086
15.237
17,366
4,696
88.608
176.478
785,261
1.366
786,627
Average Quantity
Btu Sulfur Ash
(per Ib.) (% by weight) (% by weight)
12.190 1.34 12.39
11,015 .50 9.17
10,694 .41 8.67
11,179 2.71 9.54
11.104 2.55 9.26
10,128 3.33 11.76
12,074 3.14 12.02
12.126 1.80 10.19
6,881 .55 12.50
12,612 1.61 12.62
10,570 4.18 10.05
9.012 .53 7.09
9,349 .70 19.47
6.586 .82 9.19
11,717 3.34 11.28
12.417 1.77 9.63
12,366 1.82 12.75
12.623 1.44 9.87
6.332 1.00 16.39
11.592 .50 10.17
12,951 1.14 9.42
8,104 .73 15.33
12.524' 1.64 10.68
8.669 .38 5.61
10,462 1.35 9.86
12,155 .72 6.57
10,465 1.35 9.85
Average Delivered Cost
(cents per (dollars per
million Btu) short ton)
203.5 49.61
107.9 23.76
143.1 30.60
158.2 35.37
127.8 28.38
163.7 33.17
123.0 29.70
154.3 37.43
133.5 18.37
152.9 38.57
150.8 31.87
136.8 24.66
150.4 28.13
72.4 9.54
149.3 34.98
139.2 34.57
154.8 38.28
144.3 36.42
108.4 13.73
118.5 27.46
168.7 43.70
160.0 25.93
157.7 39.50
132.7 23.01
145.4 30.43
175.2 42.58
145.5 30.45
Source: DOE 1991b
2-8
-------
2.2.6 Coal Suppliers
The individual power plants are designed to use coal meeting certain specifications regarding
coal size, Btu, ash, sulfur, volatility, and moisture content. Most of the electric utilities
obtain their coal supplies through long-term contracts with a few large suppliers, augmented
in some instances by spot purchases on the open market. A few utilities operate from the
committed reserves of a single deposit, with the power plant at or near the mine. This is
particularly true in the lignite areas of Texas and North Dakota where the low Btu value of
the lignite makes it uneconomical to ship more than a few hundred miles.
An analysis of utility coal suppliers indicates that over SO plants receive their coal from a
single source. Other plants obtain coal from as few as two to as many as 90 separate
suppliers. The 1990 utility reports to DOE/EIA identified a total of 1,616 suppliers,
providing from four tons to nearly 26 million tons each, as shown in Table 2-6 below
(DOE 91).
Table 2-6. Coal Suppliers by Quantity Supplied
Tnmiagg
under 1,000
1.000-9,999
10.000 - 99.999
100.000 - 999.999
1,000.000-9.999.999
over 10,000.000
Suppliers
26
275
635
485
188
7
2.3 COAL SAMPLE ANALYSES OF THE USCHEM DATA BASE
The National Coal Resources Data System (NCRDS) is a nationwide compendium of coal
data that was initiated by the U.S. Geological Survey in 1975. The NCRDS includes the
USCHEM data base and contains information and analyses for samples obtained from various
locations of the United States (Figure 2-2).
2-9
-------
USCHEM Data
Figure 2-2. Distribution of USCHEM Data (USGS) in which dots represent
sample origin
2-10
-------
Sample data in the USCHEM data base include both areal and stratigraphic sample locations,
sample methodology, date and collector, and complete analytical data. Detailed analytical
data include proximate and ultimate coal analyses as well as major and minor trace element
determinations, which identify the laboratories performing the analyses and the methods
used.
In support of this report, the USGS provided all uranium and thorium analyses available in
the USCHEM data base for coal samples representing minable beds with ash content of less
than 35%. Data fields obtained included State, County, Coal Region, Field, Group, Bed,
Formation, Sample Thickness, Sample Identifier, Estimated Rank, ASTMash, USGSash,
Moisture Content, Thorium, and Uranium. The data base, as received, totaled 6,892
records.
Although the USCHEM data base comprises the most comprehensive coal data base in the
U.S., some caveats must be noted. It does not represent a planned or systematic sampling of
the U.S. coal resources. Rather, it consists largely of samples taken during reconnaissance,
exploration, and special studies of the coal resources by Federal and State agencies. Much
of the sampling is done by State agencies as part of their coal resource evaluation and is
dependent on budget priorities. The disparities in sample coverage are evident in Table 2-7,
which shows a comparison of sample densities with coal area, reserves, and production for
selected States (Finkelman 1989).
2.3.1 Chemical Occurrence of Thorium and Uranium in Coal
Data regarding the chemical form of thorium in coals are scarce; the little research that has
been reported indicates that most of the thorium found in coal is inorganically bound,
probably in very fine-grained accessory minerals such as monazite (Finkelman 1981). On
the other hand, there have been extensive investigations of the occurrence of uranium in coal,
due in large pan to studies on behalf of the U.S. Atomic Energy Commission in the early
1950s. Generally, most of the uranium found in coal appears to be organically bound.
However, in some coals,, significant amounts of uranium jnay occur in accessory minerals
and as secondary mineralization (Finkelman 1981).
Anomalously high radioactivity was found associated with shallow coal deposits at several
localities in Wyoming and these were investigated as possible sources of uranium by the
2-11
-------
Table 2-7. Representation of Selected States in USCHEM
Coal-bearing Area
State (Sq. Mifcs/S^nple)
Arkansas
Georgia
Illinois
Kentucky
Louisiana
Michigan
Montana
North Dakota
Pennsylvania
Texas
Wyoming
U.S. Average
20
5
425
IS
1,360
4,000
100
145
1.400
200
40
50
Productivity
(short tons/sample)
1,000
N/A
700.000
160.000
2,900.000
N/A
80.000
1,260,000
750.000
620.000
150.000
100.000
Identified Resources
(short tons/sample)
30,000,000
1.000,000
1,800,000.000
75,000.000
1,000,000.000
33,000.000
560.000.000
1.600.000.000
80,000.000
200,000,000
130.000.000
200.000.000
Source: Finkelman 1989
USAEC in the early 1950s. However, it was found that the uranium mineralization was low
grade (0.01-0.02%) and very limited in occurrence, usually at coal outcrops or at the upper
contact of shallow coal seams (Pipiringos 1950, Masursky 1951, Denson 1950). A 13,000
square-mile area of shallow uraniferous lignite beds has been delimited at the common comer
of North Dakota, South Dakota and Montana. Some beds average 0.18% uranium and were
mined for uranium in the early 1960s. These beds are high in ash and contaminating trace
elements of As, Ge. Se. Co, and Zi (Averitt 1975). It should be noted that the individual
uraniferous lignite beds are typically only a few feet thick, limited in areal extent, high in
ash, and are not minable for steam coal, considering current technology.
2.3.2 The Uranium and Thorium Content of Coal
Protocols described in U.S. Geological Survey Bulletin 1823 (Golightly 1989) were used in
the collection, preparation, and analysis of the coal samples. Whole coal analyses were
carried out using instrumental activation analysis (INAA), which is accurate within 3% of
2-12
-------
reported results (Leventhal 1987). The high sensitivity of INAA allowed for thorium and
uranium measurements as low as 0.010 parts per million in coal samples.
Uranium Content. Distribution of the uranium content in behalf of the 6,881 coal samples
contained in the USCHEM data base is provided in Table 2-8. For comparison, Table 2-8
also contains summary data from a comprehensive DOE study published in 1979 (Facer
1979). The two data sets show a high degree of consistency. Samples within the USCHEM
data base indicated uranium concentration values that ranged from a low of 0.010 to a high
of 75.337 ppm. The distribution, however, was clearly skewed. Nearly three-fourths of all
samples analyzed showed a uranium concentration of less that 2 ppm.
The geometric mean of this log-normal distribution was 1.22 ppm with a 95th percentile
value of 4.75 ppm.
Table 2-8. Distribution of Uranium Concentration in Coal
Uranium Concentration
(ppm)
Less that 2
2-4
4-6
6-8
.8- 10
10-12
12- 14
14- 16
16- 18
18-20
20-30
30-60
>6Q
TOTAL
USCHEM-92
No. of Samples
5.123
1.310
239
89
38
26
18
8
6
3
12
7
2
6,881
Percent of Total
74.4
19.0
3.3
1.3
0.6
0.4
0.3
0.1
0.1
0.1
0.2
0.1
0.03
100%
DOE-79
No. of Samples
2,669
666
207
67
39
26
17
12
7
5
9
5
2
3,731
Percent of Total
71.5
17.9
5.5
1.8
1.0
0.7
0.5
0.3
0.2
0.1
0.2
0.1
0.05
100%
A more definitive analysis of the USCHEM coal sample data base is provided in Table 2-9.
This table identifies the range, geometric mean, and the 95th percentile uranium
concentration values by coal rank and source of coal. In spite of significant differences that
characterize and define coal, the uranium content only shows modest variation among the
four major coal ranks. For the three ranks of coal used by electric utilities, bituminous
2-13
-------
Table 2-9. Distribution of Uranium Concentration by Coal Rank and Origin
Field
Total
Anthracite
Bituminous
Sub-bituminous
Lignite
Bituminous:"'
Appalachia'1"
W. Interior*1
Alabama
Colorado
Indiana
Kentucky
New Mexico
Pennsylvania
Utah
W. Virginia
Wyoming
Sub-bituminous : (>>
PRB«O
Colorado
New Mexico
1 MT. WY
Lignite:"
Arkansas
Montana
North Dakota
Texas
No. of '
Sample
6.881
48
5,211
1.210
412
4.266
192
931
283
144
980
32
544
142
646
35
564
146
141
862
48
75
204
61
Uranium Concentration
Range
0.010-75.377
0.749-25.235
0.020-63.400
0.010-75.377
0.049-13.360
0.020-63.400
0.170-59.450
0.020-63.400
0.090-34.100
0.330-16.614
0.180-21.600
0.397-4.900
0.170-28.400
0.150-41.920
0.160-7.270
0.400-12.200
0.010-15.324
0.060-12.990
0.470-6.880
0.010-75.377
0.510-8.150
0.370-13.360
0.049-13.100
0.330-4.000
Geom. Mean
1.219
1.572
1.240
1.069
1.412
.224
.962
.216
.107
.381
.367
.900
.246
.296
.206
.376
0.806
0.939
2.347
1.002
2.227
1.916
1.071
1.695
(ppm)
95th Percentile"
4.750
4.533
4.556
5.256
5.058
4.285
11.338
4.211
4.136
5.133
5.065
5299«')
4.089
5.412
3.457
4.699
3.880
3.588
5.632
5.229
6.008
5.832
3.972
4.074W
w The numbers of samples are not strictly additive due to the fact that some samples are counted more
than once. For example, some samples represented by "Alabama" are also included in "Appalachia"
(see footnote b).
*' Includes Alabama, Kentucky. Maryland, Ohio. Pennsylvania, Tennessee, Virginia,, and West Virginia.
(c) Includes Iowa. Kansas, Missouri, and Oklahoma.
M> Powder River Basin; samples from Campbell. Converse, Johnson, and Sheridan counties of Wyoming
and Big Horn, Powder River, and Rosebud counties of Montana.
"' For a few data subsets, the 95th percentile values calculated based on a log-normal distribution exceed
the sample range. This is attributable to either non-log-normal sample distribution or to inadequate
number of samples.
2-14
-------
(69%), sub-bituminous (24.7%), and lignite (6.3%), mean uranium concentration values of
1.240 ppm, 1.069 ppm, and 1.412 ppm, respectively, were determined.
Thorium Content. Of the 6,881 coal samples, a total of 6,223 samples were quantitatively
assessed for thorium content. The range of thorium concentrations of 0.010-79.415 ppm was
comparable to that of uranium (Table 2-10). However, the overall geometric mean
concentration of 2.212 ppm thorium is nearly a factor of two higher than for uranium. When
data were segregated by coal rank, the geometric mean values of the three major ranks of
coal used by electric utilities (bituminous 69%; sub-bituminous 24.7%; lignite 6.3%) were
nearly identical with values corresponding to 2.175 ppm, 2.278 ppm, and 2.376 ppm,
respectively. Because anthracite is rarely used, the higher value of 5.581 ppm has limited
significance.
As observed with the uranium data, thorium variations are apparent among geographic
origins of the coal. In some instances (e.g., New Mexico), the elevated concentrations of
thorium in bituminous and sub-bituminous coal paralleled elevated uranium concentrations.
Arkansas lignite also had a higher mean (5.35 ppm), but this source is not used by coal-fired
electric utilities.
Wyoming and Montana together supply 27% of the fuel for coal-fired utilities, mostly sub-
bituminous coal from the Wyodak bed in the Powder River Basin. The coal from this basin
is low in both uranium and thorium, with means of 0.806 ppm and 1.746 ppm, respectively.
The Appalachian region supplies 43% of the electric utility coal. Bituminous coal from this
region, like that of the PRB, is low in both uranium and thorium, with means of 1.224 ppm
and 2.167 ppm and a 95% probability that levels will not exceed 4.285 ppm and 6.886 ppm,
respectively.
Bituminous coal from the Eastern Interior Basin, the other major source of utility steam coal
(11%), is slightly higher in uranium and comparable in thorium to that from the Appalachian
or PRB areas. Corresponding means for the Indiana coal are 1.381 ppm and 1.51 ppm. The
Illinois coals indicate uranium and thorium means of 1.5 ppm and 2.2 ppm, respectively
(Harvey 1983). These three regions are the major sources of coal used by the electric utility
industry, together supplying 81% of the coal in 1990. Other sources, each supplying over
1 million tons annually, are Texas, North Dakota, Colorado, Utah, Arizona, Louisiana, and
2-15
-------
Table 2-10. Thorium Analyses from USCHEM Data Base
Field
Total
Anthracite
Bituminous
Sub-Bituminous
Lignite
Bituminous:1*'
Appalachia^
W. Interior*'
Alabama
Colorado
Indiana
Kentucky
New Mexico
Pennsylvania
Utah
W. Virginia
Wyoming
Sub-bituminous:'"
PRB<«
Colorado
New Mexico
. MT, WY
Lignite:11'
Arkansas
Montana
North Dakota
Texas
No. of Samples
6,223
43
4,904
951
325
4,174
113
930
246
130
974
28
515
110
640
21
395
•143
131
811
46
70
132
56
Thorium Concentration (ppm)
Range
0.010-79.415
2.841-14.360
0.010-79.415
0.100-54.405
0.280-24.268
0.010-79.415
0.401-79.415
0.100-23.700
0.010-17.700
0.300-17.870
0.100-22.300
0.910-15.456
0.300-13.600
0.460-16.758
0.100-12.230
0.100-9.1500
0.210-15.830
0.200-8.000
0.100-17.300
0.100-54.405
1.740-24.268
0.530-11.800
0.200-12.240
0.800-9.300
Geom. Mean 95th
2.212
5.581
2.175
2.278
2.376
2.167
1.977
2.325
2.679
1.510
2.051
3.589
2.632
2.203
2.173
2.204
1.746
1.906
4.796
2.094
5.35
2.392
1.64
2.883
Percentile1"
7.869
11.686
7.185
10.818
8.898
6.886
9.646
7.666
10.871
3.986
7.253
11.894
6.898
7.975
6.803
10.506'"
7.991
6.752
14.849
9.749
15.588
6.571
5.903
7.573
(I)
(W
(c)
(e)
The numbers of samples are not strictly additive due to some samples are counted more than once. For
example, some samples represented by "Alabama" are also included in "Appalachia" (see footnote b).
Includes Alabama, Kentucky. Maryland. Ohio. Pennsylvania. Tennessee, Virginia, and West Virginia.
Includes Iowa, Kansas. Missouri, and Oklahoma.
Powder River Basin; samples from Campbell, Converse, Johnson, and Sheridan counties of Wyoming
and Big Horn. Powder River, and Rosebud counties of Montana.
For a few data subsets, the 95th percemile values calculated based on a log-normal distribution exceed
the sample range. This is attributable to either non-log-nonnal sample distribution or to inadequate
number of samples.
2-16
-------
Missouri. Minor sources, each supplying less than 1 million tons in 1990, were Iowa,
Kansas, Oklahoma, and anthracite coal.
2.4 OTHER COAL DATA
A literature search of radionuclide studies not previously identified by the EPA (EPA 1989)
yielded two additional studies. A total of 1S9 coal samples taken from the Herrin and
Springfield coalbeds of Illinois were analyzed for their uranium and thorium content (Harvey
1983). Uranium analysis was also performed on 184 lignite samples taken from the Wilcox
and Jackson beds of Texas (Tewalt 1989). The results of these, two studies are summarized
in Tables 2-11 and 2-12. Values cited by these two studies are in good agreement with those
of the USCHEM data base summarized previously in Tables 2-9 and 2-10.
Table 2-11. Sample Analysis for Herrin and Springfield Coalbeds, Illinois (Harvey 1983)
Field
Herrin and Springfield
Herrin Bed
Element
Thorium
Uranium
Thorium
Uranium
No. of
Samples
102
102
57
57
Concentration (ppm)
ComOXOD Mftjyimum
Range Mean Single Value
1.7-2.7 2.2 7.6
0.5-2.9 1.5 9.3
2.5-3.1 2.33 6.1
0.2-3.1 1.5 9.3
Table 2-12. Uranium Sample Analysis for Wilcox and Jackson Lignite
Coalbeds (Tewalt 1989)
Field
Total
Wilcox bed:
East-central
Northeast
East
Total
Jackson bed:
East
South
Total
No. of
Sample*
184
82
47
42
171
11
2
13
Concentration (ppm)
Geometric
Mean
1.7
2.1
2.0
1.2
1.7
1.9
2.1
1.9
Median
Value
1.9
2.0
2.4
1.2
1.9
1.8
3.3
1.8
Maximum
Value
7.7
7.7
5.5
4.9
7.6
3.2
4.9
5.0
2-17
-------
2.5 ACTIVITY CONCENTRATION EQUIVALENTS
Although uranium and thorium exist ^as isotopic mixtures (i.e., U-238, U-23S,
U-234, Th-234, Th-232, Th-230, Th-228), more than 99% of the mass for either element is
contributed by U-238 and Th-232, respectively. For the selected reference values, this
implies that one kilogram of bituminous coal contains on die average 1.24 mg of U-238 and
2.18 mg of Th-232. These values may be converted to activity concentrations using die
following specific activity coefficients, which are radionuclide specific:
Specific Activity u.238 = 3.3 x 10"7 Ci/g
Specific Activity^jjz = 1.09 x 10"7 Ci/g
Thus, 1 ppm U-238 is equal to 0.33 pCi/g, and 1 ppm Th-232 is equal to 0.11 pCi/g coal.
Uranium-238 and thorium-232 represent the primary radionuclides of the long and complex
decay-chain series depicted in Chapter 1, Figures 1-2 and 1-3. Principal characteristics of
the major decay series are presented in Tables 2-13 and 2-14 below.' A justifiable
assumption for coal is that for either decay series, the major radioactive members of that
series are in secular equilibrium. When secular equilibrium among decay-chain members
exists, die activity concentration is the same for all members of that series. Thus, die 1.24
ppm U-238 bituminous reference value implies an activity concentration of 0.41 pCi/g not
only for U-238, but for all 14 members of the decay series and yield a total decay chain of
activity concentration of about 5.74 pCi/g of coal. Similarly, the 2. 18 ppm Th-232 reference
value converts to an activity concentration of 0.24 pCi/g for each of the 11 major Th-232
decay-chain members.
In the case of coal within its natural geologic formation, the assumption of secular
equilibrium is generally considered reasonable. Disequilibrium within the decay chains has
been reported in a few instances (Papastefanau 1984; DeSantis 1984). Presumably,
disequilibrium results from radon migration either into or out of the coal seam.
For coal that has been mined and processed, the potential opportunity for Rn-222 or Rn-220
gas to migrate from the coal while it is awaiting use may introduce an inconsequential state
of disequilibrium at die time of combustion among the short-lived radon daughters. Thus,
the assumption of unconditional secular equilibrium at the time of coal combustion is not
considered unreasonable or overly conservative.
2-18
-------
Table 2-13. Major Decay Products of Uranium-238
Radionuclide
Uranium-238
Thorium-234
Prouciinium-234m
Uranium-234
Thorium-230
Radium-226
Radon-222
Polonium-218
Lead-214
Bismutb-214
Polonium-214
Lead-210
Bismuth-210
Polonium-210
Half-biS
4.5 x 10* y
24 d
1.2m
2.5 x 10s y
8.0 x 10* y
l.dxlO'y
3.8 d
3.1 m
27m
20m
1.6 x HPs
22y
5.0 d
138 d
Principal Radiation (MeV)
Alpha Max. Beta Gamma
4.20
0.191 0.093
2.29 1.001
4.77
4.68
4.78 0.186
5.49
6.00
0.65 0.352
1.51 0.609
7.69
0.015 0.047
1.160
5.31 0.803
Gamma
Emissions
Yield (%)
4
0.6
4
36
47
4
0.0011
y = years, d = days, h = hours, m = minutes, s = seconds
Source: Lederer 1967
Table 2-14. Major Decay Products of Thorium-232
Radionuclide
Thorium-232
Radium-228
Actinium-228
Thorium-228
Radium-224
Radon-220
Polonium-216
Lead-212
Bismuth-212
Polonium-212
Thallium-208
Half-life
1.4x 10l°y
6.7 y
6.1 h
1.9 y
3.6 d
55s
0.15s
10 h
60m
3.1x 10'7s
3.1 m
Principal Radiation (MeV)
Alpha Max. Beta Gamma
4.01
0.055
1.11 0.908; 0.966
5.43 0.084
5.68 0.241
6.29 0.550
6.78
0.589 0.239
2.25 0.727
8.78
1.8 2.614
Gamma
Emission
Yield (%)
25;20
1.6
3.7
0.07
47
7
100
y = years, d = days, h = hours, m = minutes, s = seconds
Source: Lederer 1967
2-19
-------
Based on a comprehensive literature review, the United Nations Scientific Committee on the
Effects of Atomic Radiation estimates that the typical or average potassium content in coal
corresponds to 2.7 pCi/g for K-40 (UNSCEAR 1982).
2.6 SUMMARY
The USCHEM data base of nearly seven thousand coal samples indicates that the average
uranium and thorium content exhibit modest variations for bituminous, sub-bituminous, and
lignite coal used by electric utilities. Values cited in Table 2-15 are used to estimate plant
emissions, as described in Chapter 6 of this report.
Table 2-15. Summary Values of USCHEM Data Base
Coal Rank
Bituminous
Sub-bituminous
Lignite
Percent
Utilization (%)
69.0
24.7
6.3
Average
Uranium (pom)
1.24
1.07
1.41
Average
Thorium (ppm)
2.18
2.28
2.38
Perhaps more significantly, however, are physical/chemical differences among the three coal
ranks that affect the potential emission of radionuclides by coal-fired SGUs. For example,
furnace designs are dictated by the coal rank to be used by a plant. In turn, -furnace design
features affect combustion and flue-gas temperature, percent fly ash entrained in combustion
flue gases, particle-size distribution in flue gas, and the selection of emission control
systems. The significance of coal-rank specific design factors that affect radionuclide
emissions are discussed in Chapter 5 of this report.
2-20
-------
3. RADIONUCLIDES IN NATURAL GAS'
For nearly a century, natural gas has been known to contain radioactivity. The early
discovery that radon-222 was the source of radioactivity in natural gas and natural gas
products, however, did not stimulate interest in evaluating the potential impact of radiation
exposure associated with its production and use for several decades. It was not until the late
1960s and 1970s that investigators assessed the radon concentrations and associated human
exposures. Since then, radon concentration measurements have been made at the point of
extraction from underground reservoirs and at points of gas processing, distribution, and
consumption.
In this report, concern for population exposure to radon contained in natural gas is limited to
electric utilities that use natural gas as the primary fuel for the operation of gas-fired steam-
electric generating units. The combustion of natural gas by electric utilities can safely
assume the complete discharge of radon-222 within the flue gas to the environment.
3.1 PRODUCTION, SOURCES, AND CONSUMPTION OF NATURAL GAS
Natural gas reservoirs occur either as non-associated (free) gas, gas dissolved in brine, gas
dissolved within crude oil, or free gas overlying crude oil (i.e., gas caps). Dissolved gas
must be separated from either brine or oil before it can be sent for further processing.
Natural gas, even when it is non-associated, contains varying amounts of water and other
impurities such as hydrogen sulfide (H2S), which must be removed during processing to
convert natural gas to natural gas products (Leggett 1979).
Although gas readily flows to the production wellhead under natural pressure established by
the reservoir, enhanced gas production methods (i.e., gas stimulation) are commonly
employed to accelerate gas flow to the wellhead. Fracturing techniques of geologic
formations to increase the reservoir area, integrate deposits, and increase flow rates include
the application of hydraulic pressure, in situ explosions, and directionally controlled drilling.
In 1989, there were 262,483 production wells in the United States (Table 3-1). During
1990, there were 7,170 new well completions (AGA 1991a). The relative numbers of gas
producing wells by regions and states are also shown in Table 3-1. The states with the
largest number of production wells include Pennsylvania (30,000), Ohio (34,450), West
Virginia (36,240), Louisiana (16,309), Oklahoma (27,443), Texas (48,609), and New
3-1
-------
Table 3-1. United States Production Wells, 1989
Division and State
United States
New England
Middle Atlantic
New Jersey
New York
Pennsylvania
East North Central
Illinois
Indiana
Michigan
Ohio
Wisconsin
West North Central
Iowa
Kansas
Minnesota
Missouri
Nebraska
North Dakota
South Dakota
South Atlantic
Delaware .
District of Columbia
Florida
Georgia
Maryland
North Carolina
South Carolina
Virginia
West Virginia
East South Central
Alabama
Kentucky
Mississippi
Tennessee
West South Central
Arkansas
Louisiana
Oklahoma
Texas
No. of Producing Gas
Wells in 1989
262,483
0
35,304
0
5,304
30,000
37,208
241
1,310
1,207
34,450
0
14,068
0
13.935
0
4
15
61
53
37,000
0
0
0
0
8
0
0
752
36,240
14,192
1,701
11.248
543
700
95,191
2,830
16,309
27,443
48,609
Net Wellhead Production
of Natural Gas in 1989
(millions of cubic feet)
18,547,823
0
217,243
0
25,469
191,774
321,681
1.477
416
160,058
159.730
0
651,652
0
587,740
0
4
878
58,326
4,704
224,942
0
0
8,773
0
34
0
0
17,935
198,200
395,827
152.044
72.417
168.466
1,900
13,883,835
168,300
5.100.023
2,185,204
6.430.308
3-2
-------
Table 3-1 (Continued)
Division and State
Mountain
Arizona
Colorado
Idaho
Montana
Nevada
New Mexico
Utah
Wyoming
Pacific
Alaska
California
Hawaii
Oregon
Washington
Other States
No. of Producing Gas
Wells in 1989
28,180
3
5.125
0
2.700
0
17.087
834
2.431
1,340
108
1.214
0
18
0
0
Net Wellhead Production
of Natural Gas in 1989
(millions of cubic feet)
2.079,934
1.455
230.669
0
52,406
0
860.976
121.250
813,178
772,709
403,515
366,694
0
2.500
0
0
Mexico (17,087). The net wellhead production of natural gas in the United States totaled
18.'5 trillion cubic feet (Tcf).
Although the 92,361 wells located in Louisiana, Oklahoma, and Texas constitute only 35.2%
of the total number of production wells in the United States, they produced 13.7 Tcf or
nearly 74% of the total net production.
In addition to domestic production of natural gas, the United States imports a small
percentage of its needs (Table 3-2). In recent years, Canadian gas represents the
overwhelming bulk of imports.
The single largest use of natural gas by consumer category is residential. Residential use of
gas in 1990 represented nearly 46% of the total gas consumed (Table 3-3). Electric
generation consumed less than one-seventh of the total. The distribution of gas consumed for
electric generation by state is shown in Table 3-4. States with the largest consumption
include New York, Florida, Louisiana, Texas, and California.
3-3
-------
Table 3-2. Imports of Natural Gas to the United States'"
(Millions of Cubic Feet)
Year
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
Total-""
903.949
933.336
918.407
843.060
949.715
750.449
992.532
1.293.812
1.318.520
1,532.259
Canada
762.113
783.407
711.923
755.368
926.056
748,780
992,532
1.276.322
1.339,357
1.448.065
Mexico
105.013
94,794
75.361
51.502
0
0
0
0
0
0
Algeria8"
36.824
55.136
131.124
36.191
23.659
1,669
0
17.490
42.163
84.193
w Source: U.S. Department of Energy, Energy Information Administration,
Natural Cos Monthly
(*) Quantities represent total LNG import from overseas into the U.S.
Table 3-3. Natural Gas Consumption in 1990
Consumer Category
Total
Residential
Commercial
Industrial
Electric Generation
All other
Other
% of Total
100
45.8
22.1
13.7
16.9
1.5
3.2 RADON CONCENTRATIONS AT THE
-LHEAD
Depending on the geologic formation associated with gas bearing reservoirs, varying amounts
of uranium may be present. Granger (1978) found that, on the average, sandstone contains
1.2 ppm uranium; a graywacke 2.6 ppm uranium; and marine black shale 20 ppm uranium.
Although uranium and the initial decay products are likely to remain immobilized within
3-4
-------
Table 3-4. Gas Consumed for Electric Generation, by State, 1975-1990
(Millions of Cubic Feet)
Civilian and State
United States
He* England
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
Middle Atlantic
New Jersey
New York
Pennsylvania
Bast Bortb Central
Illinois
Indiana
Michigan
Ohio
Wisconsin
•est Btorth Central
Iowa
Kansas
Minnesota
Missouri
Nebraska
North Dakota
South Dakota
South Atlantic
Delaware
District of Columbia
Florida
Georgia
Maryland
North Carolina
South Carolina
Virginia
Hest Virginia
Basic South Central
Alabama
Kentucky
Mississippi
Tennessee
•eat South Central
Arkansas
Louisiana
Oklahoma
Texas
Mountain
Arizona
Colorado
Idaho
Montana
Nevada
New Mexico
Utah
Wyoming
Pacific
Alaska
California
Hawaii
Oregon
Washington
1975
3, 146,874
2.523
344
0
1,437
177
11
555
23,43*
8, SOI
13,611
1.213
118.3*4
14.176
10,994
47.151
6,139
19,935
2C4.82S
46,92*
127,818
22,709
26.320
37,659
157
1.232
1*9, 10S
1,697
0
141,153
40,282
451
101
14.56C
496
358
'17.774
S.994
272
31.507
0
2,042.015
31.818
356.130
300.848
1.353.290
1C4.C1C
17,693
52,910
18
1.059
25.152
64,239
2.725
819
294,105
19,619
274.483
0
3
0
19SO
3.C81.S95
7.001
0
0
5,083
0
1,679
239
306,769
79,513
124,391
2. 865
65,883
19.165
1,904
26.368
4.706
13,740
143.317
6.835
101., 022
8,077
15,329
11.850
IS
259
1*1.409
6,999
0
166,047
,66C
.230
.702
.418
,285
62
99.461
1,382
1,902
95.073
1,104
2,243,820
58,863
424,991
329,940
1.430,026
175.032
49,699
11,928
46
4,182
27.541
56.280
5.133
222
541.133
28.763
518.808
0
121
941
1*85
3,044.083
49.34C
1.SS9
0
45. 161
67
2,532
95
235.563
£1.362
172,631
1.570
19.141
5.881
1.117
10. 126
699
1,317
27,333
2.100
21, 181
1,275
1,466
1.284
2
26
177.934
7,266
0
165,683
885
1,372
595
483
1,528
122
55.77*
1.078
1.121
53.579
0
1.C94.72C
11.411
285.039
200,686
1.197.590
83.716
41,731
4, 913
24
468
8.080
28,133
235
132
700.545
34,194
666,274
0
0
77
1987
2.844,051
52,884
7,350
0
40.183
IS
5,336
0
250,458
75.138
173,328
1.993
18,288
3.172
1,315
10.754
871
2,176
28,241
3.265
16,074
5,700
1.379
1.743
3
77
199,769
7,847
0
175.675
826
11.757
1.162
538
1.724
239
42.930
1,474
336
41,119
0
1.51C.6C*
32,057
246,912
187,895
1.049.804
CO. 855
26,665
7.826
3
478
7.076
18,454
263
90
C71.958
30,530
643.335
0
0
93
1988
2,635.613
21.374
1,260
0
19,874
55
185
0
202.208
51,066
148.495
2.649
27.901
5,706
1,455
15.027
974
2,739
33.459
5. 459
18.890
5.217
1.623
2.046
2
223
1(8.8*4
2,824
0
154.550
1.569
5.336
1,068
2.378
1,096
73
36.512
2,574
452
33.261
225
1,4*1.530
22.075
250.323
177,222
1,043,910
CC.203
25.358
8.488
0
286
10.658
21.064
196
183
585.512
30.841
552.938
0
0
1.753
1989
2,787,012
53,949
3.294
0
48.448
23
2,147
37
241.435
55.412
182. 000
4.022
32.883
6. 967
4.075
18.782
983
2.076
29.948
2.402
19. 152
4.427
1.242
2.593
1
132
222.979
7.999
0
186.814
684
19,184
1,673
2. 70S
3,796
124
47.031
1,760
328
44.927
IB
1.476.260
29.462
244,984
178,021
1.023.791
110. BIS
50,807
8.375
0
336
23.210
27.365
636
85
517.708
32,746
S17."700
0
12,942
8.320
1*40
2,776,190
66.235
4.845
0
55.221
0
5.484
685
273.208
47.591
223.253
2.364
42.217
9.19S
6.612
22.774
1.254
2,381
43,101
3,419
26.951
5.175
1.512
1.766
2
235
233, 767
10.776
0
187. 994
1.932
17.475
2.461
6.975
6.014
139
C9.91C
3,958
283
65.111
565
1.468,204
32.112
263.341
168,960
1.003.872
BO. 927
24,278
5.485
0
418
24.351
25.420
907
69
498,516
34,366
456.571
0
7.386
191
3-5
-------
their crystal structure, physio-chemical processes may dislodge radium-226 from the crystal
structure. Once dislodged, radium-226 may be leached from the rock formation overlying a
reservoir and concentrate near the reservoir boundary (Bloch 1980). Under these conditions,
enhanced gaseous radon-222 is ableto diffuse into the natural gas deposits. Since radon-222
is a chemically stable inert gas, it permeates porous geologic formation and is collected along
with natural gas at the wellhead. Radon-220, of the Th-232 series, may also be present in
natural gas deposits. But owing to its short half-life of about 55 seconds, its concentration in
extracted natural gas diminishes to insignificant levels within minutes and will, therefore, not
be discussed in this report.
Radioactivity in natural gas was first observed in Canadian gas in 1904 but not reported in
the literature until 1918 (Satterly 1918). At the time, it was suggested that helium might be
associated with the observed radioactivity. The radioactivity was reported as due to "radium
emanation," now known to be radon-222, which decays by alpha emissions as do its daughter
products of polonium-214 and -218 (i.e., helium is an alpha particle restored to a neutral
charge).
Between 1904 and 1918, Satterly and McLennan conducted a limited survey of Canadian gas
wells in an effort to find a relation between helium concentrations and radioactivity. No
significant correlation was established. Radon concentrations varied from 4 to about
800 pCi/L and showed regional variations. Average radon concentration values for gas wells
in Ontario, Alberta, and British Columbia were 169 pCi/L, 62 pCi/L, and 473 pCi/L,
respectively (Table 3-5).
Bunce and Sattler (1966) conducted an extensive study in 1965 to determine the radon-222
concentrations in natural gas production wells in the San Juan Basin area of Colorado and
New Mexico. They sampled over 300 wells and found an average radon level of 25 pCi/L.
Individual wells'were sampled with levels as high as 160 pCi/L and as low as 0.2 pCi/L. A
review by Barton (1971) showed that 1,250 wells in Texas, Kansas, and Oklahoma had
average radon concentrations of 100 pCi/L or less. Concentrations in these wells varied
from 5 to 1,450 pCi/L.
Faul and coworkers (1952), with the support of the United States Geological Survey,
determined the radon content of about 500 producing gas wells in the Texas panhandle area.
They observed levels from about 10 to 520 pCi/L at standard temperature and pressure.
They also noted that the radon content is nearly constant for a given well under normal
3-6
-------
Table 3-5. Summary of Radon-222 Concentrations in Natural Gas at Production Wells
Area —
Ontario, Canada
Alberta, Canada
British C&Iumbia
Colorado, New Mexico
Texas, Kansas, Oklahoma
Texas Panhandle
Colorado
Project Gasbuggy Area
Project Gasbuggy Ana
California
Gulf Coast (Louisiana, Texas)
Kansas
Wyoming
Radon-222 Level (pCi/L)
Average
169
62
473
25
<100
...
25.4
15.8 - 19.4
29.4
_
5
100
10
Range
4-800
10 - 205
390-540
0.2 - 160
5 - 1,450
10 - 520
11 -45
—
12-59
1 - 100
—
—
—
Reference
Satterly (1918)
Barton (197 1)
Satterly (1918)
Barton (1971)
Satterly (1918)
Barton (1971)
Bunce (1966)
Pierce (1964)
Barton (1971)
Paul (19S2)
McBride (1969)
Gotchy (1972)
NERC-LV (1970)
McBride (1969)
NERC-LV (1973)
Fries (1972)
Kaye (1973)
Bernhardt (1973)
Bernhardt (1973)
production conditions. A significant change in radon concentration was measured in several
wells on restarting after being shut down during the summer.
The radon values rose sharply with initial production and leveled off after removal of about
20 to SO thousand cubic feet of gas (less than one hour's production for wells normally
producing two to three million cubic feet daily). They interpreted this behavior, along with
an analysis of transient gas flow and steady state conditions, as an indication that radon
originates in the immediate vicinity of the bore in most wells.
Many radon measurements in gas have also been made in conjunction with tests for nuclear
stimulation of natural gas. Data from Gotchy (1972) and NERC-LV (1970) on studies of the
3-7
-------
Rulison and Rio Blanco gas stimulation projects indicate the average radon level for wells in
the Rulison area during 1969-1970 was 25.4 pCi/L (range 11 to 45 pCi/L). McBride and
Hill (1969) reported that levels of radon-222 in pre-shot samples for project Gasbuggy had an
average value of 19.4 pCi/L. Post-shot samples indicated that nuclear stimulation did not
raise radon-222 concentrations in neighboring wells above the naturally occurring levels.
The NERC-LV Technical Support Section (1973) also sampled natural gas from two trunk
lines serving all 28 producing gas wells within 5 miles of Project Gasbuggy from November
1969 to November 1970. The average radon level was 29.4 pCi/L, with a range of 12 to
59 pCi/L. These results were essentially the same as before Project Gasbuggy and confirmed
that nuclear stimulation does not increase radon levels. Some seasonal variation was
apparent with the higher levels occurring between March 31 and September 15, 1970.
Data summarized in Table 3-5 represent a sizeable number of wells that were operational
over a period of several decades. The Gulf Coast region of Louisiana and Texas had the
lowest average radon concentration at about 5 pCi/L. Upper Texas, Kansas, Oklahoma, and
California had average levels up to 100 pCi/L. .When all United States sample data are
averaged, an average level of 37 pCi/L is obtained. Canadian gas had considerably higher
values with a combined average value of about 200 pCi/L. The potential impact of the
higher values of Canadian gas "is mitigated by the fact that Canadian imports make up less
than-10% of the total gas consumed in the United States. Based on average sampling values
cited in Table 3-5, a range of 10 to 200 pCi/L is likely to encompass radon concentrations in*
natural gas sources in the United States.
It is important to point out, however, that the above cited measurement values and derived
estimates of average values represent radon concentrations at the wellhead. Substantial
reductions of these values must be anticipated for radon concentrations within the distribution
systems and near points of consumer use. Factors which influence/reduce radon
concentrations include (Jacobs 1972):
• gas processing
• pipeline transmission time
• storage time
• pipeline absorption
• production rate
3-8
-------
3.3 RADON REDUCTION FOLLOWING GAS PROCESSING
Natural gas at the wellhead contains from 55 to 98% methane and complementary
percentages of heavier hydrocarbon»-(ethane, propane, butane). Gas processing plants,
usually located near the gas fields that supply them, remove the commercially valuable
heavier hydrocarbons, which are bottled under pressure and sold as liquefied petroleum gases
(LPG) (Cannon 1972). The "residue gas," as the processed gas is called, consists principally
of methane, which is sold as a commercial heating fuel through a vast distribution system
(Figure 3-1).
The processing of natural gas has significance with regard to radon content. While the
details of the processing method may vary, the physical principle behind the separation of the
heavier hydrocarbons from the inlet stream mixture is thermal fractionation. Table 3-6 gives
the boiling points for the major components of natural gas along with that of radon. Due to
the fact that the boiling point of radon is bracketed by those of ethane and propane, radon
will separate and concentrate in these hydrocarbon fractions.
Clcle*
Distribution
Center
Rita* Stpuilira
•rah IK
Croii country cruraiccloo. line .
Otte Ihaii Fncdui
IKUI. EUMM. GtMfiM. riC
n*tin M
Track Oifwr. R«u3
Undcrtround re»ervolr§
pre««uriied tank*
natur«l e*> CtHC)
Figure 3-1. General Flow Diagram for Production and Distribution of Natural Gas and LPG
3-9
-------
Table 3-6. Liquefied Petroleum Gas(a)
Component
Methane
Ethane
Radon
Propane
Butane
Percent of LPG^
0.2 - 2.0
2.4 - 9.5
88-96
0.5 - 1.5
Boiling Point (°Q
-161.5
-88.3
-61.8
-42.2
- 0.5
(>> Source: Gesell 1974
"" Measured in survey of six retail sources of LPG in the
Houston. Texas area.
Gesell (1974) has surveyed nine gas processing plants in the United States. The locations of
these surveys are shown by the circles in Figure 3-2. Also shown in Figure 3-2 are the
locations of gas processing plants in the United States. The survey included measurement of
the radon-222 content of the inlet stream(s), fractionated products, and the residue (outlet)
gas. The results of this survey are given in Table 3-7. It can be seen that radon-222
concentrations in the inlet gases are in the expected range for wellhead gases (see Table 3-5).
Furthermore, the radon-222 concentrations in the propane are much higher than those in the
inlet natural gas, ranging up to 1,119 pCi/L. Figure 3-3 shows a plot of the radon-222
concentration in propane versus the radon-222 concentration in the inlet gas for eight of the
nine plants (the propane sample was lost from one plant). A least squares fit to the data gave
a slope of 8.2 and a linear correlation coefficient of 0.94. Thus for a variety of processing
plants, the average concentration of radon-222 in propane is about eight times the
concentration in the inlet gas.
In another study, Bernhardt (1973) reported that a radon concentration of 56 pCi/L in the
inlet natural gas was increased to 1,100 pCi/L in the propane fraction at a San Juan gas
processing plant, which indicates an increase in radon concentration by a factor of nearly 20.
In addition, Fries and Kilgren (1972) report data for a plant in California that indicate a
concentration factor for radon of about 14. The differences in these factors may be
attributed to differences in hydrocarbon makeup of the inlet natural gas (i.e., the relative
quantities of ethane, propane, and butane). The factor of 20 reported by Bernhardt was for a
basic propane stream. The butane stream for this facility had a radon concentration factor of
7. The factor of 14 determined by Fries and Kilgren appears to be for a mixed propane-
butane stream. When these factors are combined with the data from Gesell, the average
concentration of radon in propane is about 10 ±5 times that in natural gas.
3-10
-------
Figure 3-2. Distribution of Gas Processing Plants in the Contiguous U.S.
-------
Table 3-7. Radon-222 Concentration Measured in Gas Plant Processing Streams
(pCi/liter)
Source
Inlet gas
Ethane
Propane
Butane
Sales gas
(methane)
•'-" Plant Code
1
10.S
368.2
(35.1)
177.6
(16.9)
1.5
(0.1)
0.6
(0.06)
2
26.3
—
(b)
—
13.6
(0.5)
3
78.4
—
386.1
(4.9)
—
3.2
(0.04)
4
118.5
—
1119.0
(9.4)
—
93.6
(0.8)
5
35.4
97.7(1>
(2.8)
237.7
(6.7)
...
3.1
(0.09)
6
18.4
306.7
(16.7)
56.8
(3.1)
...
0.9
(0-05)
7
4.3
—
61.5
(14.3)
...
1.1
(0.3)
8
1.3
...
9.0
(6.9)
...
0.7
(0.5)
'" Mixture of ethane and propane.
*> Sample lost.
Note: Parenthetical values represent ratio of radon-222 concentration in the processed gas component
to inlet radon-222 concentration.
Source: Gesell 1974
The increased radon concentrations found in propane and other natural gas fractions are
offset by corresponding reductions in radon concentrations found in the outlet gas. The
outlet gas, consisting primarily of methane, is subsequently piped into distribution systems
where it is metered and sold to customers, inclusive of electric utilities.
3.4 RADON REDUCTION IN THE DISTRIBUTION SYSTEM
Beyond the 89,500 miles of field and gathering pipelines that bring natural gas from the
wellhead to the processing facility, there were 280,100 miles of transmission and 836,000
miles of distribution pipelines in 1990 (AGA 1991a). A major impact is that as gas moves
from individual production wells through regional processing plants and subsequently through
trunk line systems, it becomes thoroughly mixed (Jacobs 1972). Thus, differences in
wellhead concentrations among individual wells of a given region, and even regional
differences, are either lost or moderated. Beyond the effect of averaging radon
concentrations, the transmission of gas through collection and distribution lines allows radon
gas to undergo natural decay. Most domestic gas is transported by buried pipelines with
3-12
-------
- 1200
o>
O
tu
z
<
a.
O
cc
a.
800
a
z
O
O
c
DC
CNJ
CM
CNJ
400
40
80
120
222Rn CONC. IN INLET GAS (pCi/liter)
Figure 3-3. Least Squares Fit of the Radon-222 Concentration in Propane Versus
Versus the Radon-222 Concentration in the Inlet Gas
diameters of up to four feet. The gas is compressed to 600 to 1,000 pounds of pressure per
square inch, which drives the gas through the pipe. Valves installed at 10- to 30-mile
intervals allow for stable pressures and isolation in the event of a large pressure drop.
Pipeline transit velocity varies from about 10 to 12 miles per hour (GEH 1966). Thus, a
transit distance of about 1,500 miles to a gas-fired SGU along the east coast would require
125 to 150 hours and reduce radon concentrations of residue gas to about one-third of their
3-13
-------
original value. This distance is typical of many major transmission lines from Texas and
Louisiana production-well fields to east- and west-coast distribution centers.
Radon in residue gas may also undergo natural decay prior to consumption while in storage.
Transmission pipeline systems are designed to be operated at full capacity, which provides a
delivery rate roughly equal to the average demand rate. While production rates are relatively
constant, demands, however, vary over a 24-hour period and follow a seasonal cycle.
Figure 3-4, which presents monthly average values, shows that consumption during summer
months is less than one-half that of the peak months of December, January, and February
(AGA 199la). Storage facilities provide assurance of an adequate supply when the demand
exceeds production rates.
% of Annual Sales
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Figure 3-4. Gas Utility Industry Sales (10-Year Average Values, 1981-1990)
3-14
-------
Storage reservoirs are commonly former oil or gas wells that have been depleted. For 1990,
the underground storage of residue gas consisted of 397 pools located in 27 states (AGA
1991b). Their collective storage capacity of 7.8 Tcf represents about 40% of the annual
production of gas during 1990.
The impact of storage time on radon concentration has not been thoroughly investigated.
However, the expected reduction in radon with storage time may, in part, be offset by newly
introduced radon into the storage reservpir. Recently, Wojcik (1989) evaluated radon
concentrations in natural gas at the point of consumption over an eighteen-month period.
Peak concentrations during winter months were about twice as high as lower values during
the summer months, when production exceeds use and gas is stored. From the limited data,
it can be assumed that the net effect of underground storage reduces the radon concentration
in residue gas.
A final, but difficult to quantify impact on radon gas concentration is the adsorption of radon
by piping surfaces. Wojcik (1989) showed that the ratio of adsorbed radon in polyethylene
piping to radon in flowing gas was equal to 0.96. For PVC piping, the ratio was reduced to
0.33.
To date there exists no comprehensive data base involving radon measurements at consumer
use points. Barton et al. (1973) concluded that a nationwide gas sampling program would be
impractical and excessively costly and that radon gas measurements in the distribution system
could serve as surrogate measurements. Barton and coworkers sampled the gas supplied to
several large metropolitan areas including Chicago, New York City, and Denver over a
period of several months (Figure 3-5). When all sample measurements were averaged, a
value of 20 pCL/L was observed. This average value was heavily affected by the high
pressure line from Kansas to Denver, which has a level of about 95 pCi/L. Excluding this
set of values, the'average value was reduced to about 10 pCi/L.
A similarly comprehensive measurement program was undertaken by Gesell (1974). One of
the two Houston area natural gas suppliers was sampled at two locations on a regular basis
for eight months. A summary of his data is provided in Table 3-8. The data neither indicate
a significant difference between the two sources nor a seasonal fluctuation. The absence of
seasonal fluctuation is not surprising and can presumably be attributed to the general mild
winter temperatures of the region. Combined radon concentrations from both sources yielded
an average value of 5.4 pCi/L. A comparison of Gesell's average value with those of Barton
shows that the Houston data are bracketed by data from other parts of the country.
3-15
-------
200
100
50
u
v
T5 20
10
u
0.
£
*•
0
••
e
«i
w
§
(J
r«
I
e
o
•o
.So.s
0.
/
N
«
Pt
N
n
K
n
n
K
n
K
) o r« « *
Date (month/year)
Figure 3-5. Monthly Average Concentrations of Radon-222 in Five Distribution Systems
The identities of the samples are:
A) Colorado Interstate Gas Co. - Kansas Line
B) Natural Gas Pipeline Co. of America - Amarillo Line
C) Colorado Interstate Gas Co. - Wyoming Line
D) Natural Gas Pipeline Co. of America - Gulf Coast Line
E) Texas Eastern Transmission Co. and Transcontinental Gas Pipeline Corp.
average of all samples.
3-16
-------
Table 3-8. Radon-222 Concentrations of Natural Gas (pCi/L) in the
Houston, Texas Area
m\f | X
Week of
Sampling
Nov. 12. 1972
Nov. 19
Dec. 3
Dec. 10
Dec. 17
Dec. 31
Jan. 7. 1973
Jan. 14
Jan. 21
Jan. 28
Feb. 4
Feb. 18
Feb. 25
Mar. 4
Mar. 11
Mar. 18
Mar. 25
Apr. I
Apr. 8
Apr. 22
May 6
May 27
June 17
AVERAGE
._- Identified Major Source
East Texas Gas Fields
10.8
9.7
1.6
7.5
11.8
10.3
4.3
4.4
4.8
13.5
2.7
13.4
1.3
11.4
3.5
<0.5
2.2
4.0
15.5
6.2
3.6
3.1
4.5
5.0
West Texas Gas Fields
13.5
1.7
1.6
12.8
1.4
11.7
14.2
1.2
2.2
11.3
..
2.6
4.6
1.6
4.2
1.4
2.9
2.1
10.5
1.3
1.9
1.6
2.9
5.8
Average
12.2
5.7
1.6
10.2
6.6
11.0
9.2
2.8
3.5
12.4
2.7
8.0
3.0
6.5
3.8
1.0
2.6
3.1
13.0
3.8
2.8
2.4
3.7
5.4
In addition to the Houston data, Gesell also analyzed 23 grab samples obtained from 12
municipal gas companies in the southern and western United States (Table 3-9).
Grab-sample values fall within the expected range except for the 270 pCi/L value for gas
serving the Amarillo, Texas area. Inquiry revealed that the gas supplied to the company
serving Amarillo comes directly from nearby wells which the company owns without
significant processing. Thus, the closeness of wells, the absence of gas processing, and the
reported high wellhead concentrations collectively contribute to this high value. The
23 pCi/L average value for all 23 grab samples is reduced to 11.5 pCi/L when this high
value is excluded.
3-17
-------
Table 3-9. Radon-222 Concentrations Found in Gas Supplies
of Southern and Western Cities
State
Alabama
Arizona
California
Louisiana
New Mexico
Oklahoma
Texas
^ Metropolitan
*~ Area Served
Mobile
Mobile
Tucson
Tucson
Los Angeles
Los Angeles
San Diego
San Francisco
San Francisco
San Francisco
San Francisco
New Orleans
New Orleans
New Orleans
Albuquerque
Oklahoma City
Tulsa
Tulsa
Amarillo
Beaumont
Beaumont
Dallas
Dallas
Radon
(pCi/L)
2.8
2.4
13.5
12.8
11.4
8.2
6.8
68.0
36.0
10.2
2.1
2.6
1.9
1.6
10.6
4.9
15.4
13.3
270.1
17.6
5.2
3.3
1.3
Values within the above-cited ranges have also been reported by others:
• McBride and Hill (1969) reported radon levels of about 8 pCi/L at two
metering stations on the way to Las Vegas and Los Angeles. Levels were also
measured in natural gas at the Farmington Laboratory in New Mexico, where
the average radon content was about 45 pCi/L.
• The Rocky Mountain Natural Gas Company (RMNG) measured radon in its
main city pipelines in the Colorado towns of Aspen, Glenwood Springs, and
Delta. Average levels were about 25 pCi/L. The RMNG distribution system
is closed, i.e., it neither supplies nor obtains gas from other systems
(Bemhardt 1973).
3-18
-------
Gesell measured radon in a distribution main in Houston weekly from
November 1972 to June 1973. The average radon level was 5.4 pCi/L.
More recently, the primary source of natural gas used by plants owned and operated by
Pacific Gas and Electric (PG&E) were sampled for radon content (Personal Communications:
Jim Satranek, Santa Cruz County Environmental Health Services). Table 3-10 summarizes
sample data and identifies the primary source of natural gas for each of the seven PG&E
facilities.
Table 3-10. Radon-222 in Natural Gas Supplied to PG&E Facilities
Plant
Contra Costa
Humboldt Bay
Hunters Point
Mono Bay
Moss Landing
Pittsburgh
Potrero
Primary Source of Gas
Canada (Antioch)
Canada (Antioch)
Canada/El Paso
(Antioch/Milpitas)
El Paso (Milpitas)
El Paso/Canada
(Antioch/Milpitas)
Canada (Antioch)
El Paso/Canada
(Antioch/Milpitas)
Radon-222 (pCi/L)
10 (average)
10 (average)
8 (average)
< 10
8 (average)
10 (average)
8 (average)
Radon-222 Emission
Factor (Ci/1000 ft.3)
0.283- x 10*
0.283 x 10-*
0.227 x 10*
0.142x 10*
0.227 x 10*
0.283 x 10*
0.227 x 10*
Notes:
- The Antioch sample location represents Canadian gas.
- The Milpitas sample location represents El Paso gas.
- Canadian and El Paso results are averaged for those plants that receive gas from both sources.
- Results from California gas fields were not used since this gas is a very small fraction of the
total.
3.5 CONCLUSION
A summary of available data on radon in natural gas distribution lines of consumption areas
is shown in Table 3-11. These data indicate that overall average radon levels at points of use
are about 20 pCi/L. Highest levels were observed in the Colorado and New Mexico areas.
Regions farthest away from natural gas sources tend to have lower radon levels. This can be
attributed to pipeline transmission time and storage, which allows significant radon decay.
Also, gas may be mixed and diluted with that from several supply systems while in transit to
3-19
-------
Table 3-11. Radon-222 Concentrations in Natural Gas Distribution Lines
Area
Chicago
New York City
Denver
West Coast
West Coast
Colorado
Nevada
New Mexico
Houston
Overall Average
Radon-222
Average
14.4
1.5
50.5
15
10
25
8
45
5.4
Level (pCi/L)
Range
2.3-31.3
0.5 - 3.8
1.2- 119
1 - 100
< 8-41
6.5 - 43
5.8 - 10.4
10-53
1.4 - 14.3
= 20
Reference
Barton (1973)
Bunce (1966)
Barton (1973)
Fries (1972); Barton (1973)
Satranek (1990)
Bemhardt (1973)
McBride (1969)
Bunce (1966)
Gesell (1974)
areas such as New York City or Chicago. It should also be noted that variations in radon
levels at consumer use points could also be attributed to production rates as a function of
seasonal gas use. For example, the highest levels occurred in the winter during peak use
periods, presumably due to shorter transport/storage times.
For dose calculations, Barton et al. (1973) selected a value of 20 pCi/L for radon
concentrations at consumer use points. This value coincides with the overall average value
derived in this analysis involving multiple data sets of major regions in the United States. In
this report, estimates of radon-222 emission from natural gas-fired steam-electric generating
units are based on a value of 20 pCi/L.
3-20
-------
4. THE RADIONUCLIDE CONTENT OF RESIDUAL FUEL OIL
Residual fuel oil is the primary energy source for a total of 329 oil-fired steam-electric
generating units operated in 1990. The production of about 120 billion kilowatt hours of
electricity by these plants required the combustion of about 200 million barrels of residual
fuel oil (DOE 1991a). Residual fuel oil represents a general classification of fuel obtained as
liquid still bottoms from the distillation (i.e., refinement) of crude oil. The still bottoms may
be used alone or in blends with heavy residual liquids from other refinery process operations.
Residual fuel oil includes Grades No. 5 (light and heavy). No. 6 (Heavy grade called Bunker
C oil), and Navy special fuel oil.
An intensive search of the open literature and verbal communication with the American
Petroleum Institute, Electric Power Research Institute, and the Department of Energy
revealed that the data specific to the radionuclide content of residual fuel oils are sparse at
best.
For this reason, the literature search for data was expanded to include secondary methods for
estimating concentrations or establishing bounding values. Secondary methods for estimating
the radionuclide content in residual fuel oil came from several studies that analyzed the metal
or radionuclide content in (1) crude oil, (2) fuel oil ash, and (3) environmental samples taken
around oil-fired operating steam-generating units.
Even when secondary methods were included, the data remain sparse. The EPA concluded
that there was a need for additional data, and an independent study (see Section 4.3) was
conducted in which residual fuel oil samples were analyzed for their radionuclide content.
This chapter is divided into three major sections. The first section (Section 4.1) provides
background information that explains the source, distribution, and fate of radionuclides in
petroleum products and helps the reader to understand the methodology employed in the
interpretation of study data. The second section (Section 4.2) summarizes and interprets past
analytical studies involving residual fuel oil, crude oil, oil ash, and environmental samples.
The third section (Section 4.3) contains the findings of a recently completed study by the
EPA in which 42 residual fuel oil samples obtained from operating utilities were analyzed.
4-1
-------
4.1 BACKGROUND INFORMATION
In 1990, there were 587,762 producing oil wells in the United States (API 1992). The
average well produced 12.5 barrels per day, which yielded a total of 2,684.7 million barrels
of crude oil that same year (Table 4-1). Leading states, in descending order, were Texas,
Alaska, Louisiana, California, and Oklahoma, which produced 2,206.3 million barrels or
82.1% of the nation's total. In order to meet domestic demands, the United States imported
an additional 2,151.4 million barrels of crude in 1990. Table 4-2 identifies United States
imports of crude oil by country of origin.
4.1.1 Oil Extraction
Crude oil is extracted from underground rock formations, called reservoirs. These reservoirs
may be composed of sandstone, limestone, or dolomite structures (API 1976). Within an oil
reservoir, there may be varying amounts of dissolved gases, free gases, brine, and saltwater
deposits. The utilization of naturally existing gases and brine to force crude oil from the
reservoir to the earth's surface via the production well is termed primary production.
Crude oil at the wellhead is, therefore, usually a mixture of oil and water (brine) that occurs
as an emulsion in varying ratios (Case 1970). During the primary production period of a
reservoir, emulsion consists primarily of oil with some water. As the reservoir is depleted,
the ratio of oil to water changes. In a declining production situation, up to 80% or even
90% of the wellhead fluid is water with colloidal oil droplets (Case, 1970).
As a primary contaminant, brine is commonly removed on-site before the crude petroleum is
shipped to a refinery for processing. Free water/brine is separated from the emulsion by
simple gravimetric methods.
Water and oil emulsions from a near depleted well require more elaborate treatment
techniques to achieve separation of the two components. Heaters and electric dehydrators are
commonly used to physically remove water from the emulsion. The separation is not
complete, leaving 'small amounts of brine in the crude oil that is subsequently shipped to a
refinery.
4-2
-------
Table 4-1. Crude Oil Production by State
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Florida
Illinois
Indiana
Kansas
Kentucky
Louisiana
Michigan
Mississippi
Missouri
Montana
Nebraska
Nevada
New Mexico
New York
North Dakota
Ohio
Oklahoma
Pennsylvania
South Dakota
Tennessee
Texas
Utah
Virginia
Washington
West Virginia
Wyoming
Other
Total
Grade Oil Production
(x 1,000 Barrels)
18.538
647.310
122
10.387
350.900
30.454
5.674
19.954
3.000
55,427
5.411
393,706
19.675
27.033
146
19.810
5.890
4.012
67.247
416
36.716
10,008
112.274
2,643
1,648
508
702,160
27.604
15
—
2.143
103.855
-
2.684.700
No. of
Producing Wells
872
1,466
22
7.265
43.375
6.596
83
31,874
7.506
45.470
22.741
23,812
4.570
2.169
854
3.846
1.440
46
18,546
4,043
3.546
30.089
81.667
22.338
149
736
188,829
1.972
25
15.950
11.397
0
4,468
587.762
Average Oil Production per
Well (Barrels/day)
58.2
1,209.7
13.4
3.9
22.2
12.6
187.3
1.7
1.1
3.3
0.7
45.2
11.8
34.1
0.5
14.1
11.2
239.0
9.9
0.3
28.4
0.9
3.8
0.3
30.3
1.9
10.2
38.4
1.6
—
0.4
25.0
—
—
Source: U.S. Bureau of Mines, Annual Petroleum Statements 1976-1990
4-3
-------
Table 4-2. U.S. Imports of Crude Oil by Country of Origin
(thousand of barrels)
Country
Bolivia
Canada
Colombia
Ecuador
Mexico
Trinidad
Venezuela
Total Western Hemisphere
United Kingdom
Norway
Total Europe
Algeria
Angola
Congo
Egypt
Gabon
Nigeria
Tunisia
Zaire
Total Africa
Iraq
Kuwait
Oman
Qatar
Saudi Arabia
United Arab Emirates
Total Middle East
Australia
Borneo
Indonesia
Malaysia
Total Far East
U.S.S.R.
Total Eastern Hemisphere
ANNUAL TOTAL
1990
•••
234,516
51.041
13,886
251,345
27,803
242,910
56.497
34.878
92,044
23,035
86,095
1,387
9,005
23,349
286.126
645
8.468
451,171
187.485
28,942
13,037
1,293
436,193
3.300
678.056
17.307
—
35.912
14.749
103.092
297
826,727
1,324,660
2,151,387
4-4
-------
Secondary and Tertiary Recovery Methods. When naturally existing gases and aqueous
fluids within the reservoir are no longer able to drive crude oil to the wellhead, extrinsic
methods are applied for "enhanced oil recovery." Enhanced oil recovery employs techniques
that are categorized as either secondary or tertiary methods. In 1990, only 29,486 wells in
the United States were naturally flowing wells, representing about 5% of the total production
wells (Table 4-3).
Secondary recovery techniques include the injection, under pressure, of either treated water
or immiscible gas into an oil formation to increase the petroleum flow to the production well.
The injection of water by a series of injection wells surrounding a reservoir is termed water
flooding or field flooding and is the more common technique employed.
The injection of treated water under pressure forces trapped petroleum within the reservoir
toward the production well and up to the surface. Production fluid at the wellhead is a
mixture of oil and water. This mixture is separated by the same process described
previously (National Petroleum Council 1976, Schumacher 1978). Because natural gas is in
itself a valuable fuel, its use in injection wells has greatly diminished over the years.
There are several tertiary recovery techniques that may be employed. These techniques
either reduce the surface tension between oil and the driving fluid or reduce the viscosity of
the oil. Surface tension may be reduced by injecting gaseous hydrocarbons (Schumacher
1978) or detergent-like surfactants (Rahme 1978). Whatever their nature, their effect is to
reduce the amount of oil-water tension and to solubilize/emulsify the oil for improved flow.
Another means of reducing the viscosity of oil for tertiary recovery involves thermal
methods. Steam or in-situ combustion techniques effectively reduce.the viscosity of oil, thus
allowing it to flow more easily (Schumacher 1978).
4.1.2 Radioactive Source Terms
It is well established that crude oil may contain more than 60 different metals in measurable
quantities, including uranium and thorium. Although the natural deposits of uranium and
thorium in subsurface rock formations are clearly the source for radioactivity, a simple
relationship between radionuclide concentrations in subsurface rock and crude oil does not
exist. It has been noted that some oil-field brines contain anomalously high radium-226
4-5
-------
Table 4-3. Producing Wells by State and Type
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Honda
Illinois
Indiana
Kansas
Kentucky
Louisiana
Michigan
Mississippi
Missouri
Montana
Nebraska
Nevada
New Mexico
New York
North Dakota
Ohio
Oklahoma
Pennsylvania
South Dakota
Tennessee
Texas
Utah
Virginia
Washington
West Virginia
Wyoming
Other
U.S. Federal Waters
U.S. Total
1989
Flowing
85
698
0
ISO
1.534
541
59
0
0
58
0
2.008
1.338
363
0
363
0
10
3,170
1
543
400
4,817
0
0
0
12,548
112
0
0
0
863
0
2,292
31,953
Artificial Lift
846
639
25
4.190
42,211
5.821
42
32,441
7.543
44.911
22,859
20.864
4.219
3.226
807
3,638
1,787
36
14.617
4.349
2.899
29,794
91.527
27.218
156
713
173.678
• 2.122
20
0
15,940
10.676
0
1,589
571.403
1990
Flowing
66
767
0
100
993
666
41
0
0
59
0
2,001
605
221
0
349
0
8
3,110
1
536
375
4,080
0
0
0
11,742
99
0
0
0
850
0
2,817
29,486
Artificial Lift
806
699
22
7.165
42.382
5,930
42
31,874
7,504
45.411
22,741
21,811
3,965
1,948
854
3,497
1,440
38
15.436
4,042
3,010
29.714
77.587
22,338
149
736
177,087
1,873
25
0
15,950
10,547
0
1,651
558,276
4-6
-------
concentrations (Armbrust 1956, Gott 1953). Petroleum brine samples taken from oil fields
of Kansas, Oklahoma, and northwestern Arkansas showed Ra-226 activities as high as 3510
dpra/1. At equilibrium, this concentration would be equivalent to 4.5 ppm uranium-238.
Chemical and radiometric analyses showed that uranium was not present in significant
amounts and that the radium concentrations did not reflect equilibrium values. In 24 oil
samples of another study, the uranium concentrations ranged from 0.00001 to 0.064 ppm
(Erickson 1954). The enrichment of petroleum brine with Ra-226 is thought to involve the
removal of radium from inside rock grains to the petroleum brine containing pore spaces by
(1) partial or complete dissolution (leaching) and (2) direct alpha emission recoil (Bloch
1979, 1980).
Alpha recoil is thought to be the dominant process by which radium is concentrated in
petroleum brine. When an alpha particle is ejected from the nucleus of an atom undergoing
radioactive decay, the recoil energy imparted to the residual nucleus exceeds the magnitude
of bond energies. The recoil atom may travel a distance of 10'7 to 10'8 cm inside the
mineral, which is one to two orders of magnitude larger than the size of the crystal lattice.
Thus, enhanced brine concentrations of Ra-226 (and to a lesser extent Ra-228) are assumed
to be due to alpha recoil from Th-230 and associated breaking of chemical bonds and crystal
damage. Any thorium temporarily brought into the pore solution is quickly precipitated and
accumulated as surface-grain coatings, thus facilitating the rock-to-fluid transfer of its radium
decay product.
In summary, the disequilibrium in the U-238 to Ra-226 portion of the decay chain is thought
to involve the Th-230 coating of grains by alpha recoil which facilitates the dissolution and
leaching of Ra-226 by pore fluids. The efficiency by which radium is leached and
incorporated into extracted crude oil is further affected by the oil recovery method. High
pressurization, increase in temperature, and use of emulsifying agents/surfactants in
secondary and tertiary extraction methods are thought to (1) greatly affect the circulation of
fluids through pore spaces, (2) enhance the transfer of radium to the pore fluids, and (3)
affect the partitioning of radium and other radionuclides between the aqueous and the organic
phase of harvested crude oil.
Other causes of disequilibrium among members of the decay chain are described in Section
4.3.6 of this Chapter.
4-7
-------
4.1.3
jm Refining
;ate of Trace Metals
Petroleum refineries fractionate crude oil into various distillates and residual fractions having
a wide variety of commercial valuesT The twenty-leading companies and their refining
capacities are listed in descending order in Table 4-4.
Table 4-4. Twenty Leading Companies-U.S. Refining Capacity
(Thousand of Barrels per Day)
1989
Company
Chevron
Exxon
Shell
Amoco
Mobil
BP America
USX
Texaco
Sun
ARCO
DuPont
Ashland
Petroleos de Venezuela
Koch
Phillips
Coastal
Unocal
Aramco
Solomon
Total Petroleum
U.S. TOTAL
Capacity
1,621.0
1,147.0
1,078.6
956.0
838.0
756.6
603.0
595.0
595.0
559.0
406.5
346.5
. 344.5
310.0
305.0
302.8
295.6
295.0
277.1
190.1
15.183.6
Prior to fractionation of crude oil, inorganic metallic salts, water, suspended solids, and
certain water soluble compounds must be. removed from feedstock. This processing step is
essentially a solvent-extraction method, using water as the extracting medium. Crude
desalting may also be achieved by chemical de-emulsifiers, application of electrical fields, or
other refinery technologies. Appendix A provides a summary discussion of refinery
technologies and their impact on metal contaminants.
4-8
-------
The need for demetallization of refinery feedstocks is prompted by several concerns. First,
many of the trace metals in petroleum act as catalytic poisons and interfere with the catalytic
cracking process of refinery feedstocks. Failure to remove metals from feedstock reduces
the yield of gasoline and other petroleum distillates. Second, failure to remove metals from
feedstocks yield metal-rich fuel oils. Combustion of such fuel oils is destructive to
refractories in. furnaces. Contributing to the destructiveness of fuel oil ash are the alkali and
alkaline earth metals.
The fate of metals in crude oil is also affected by hydrodesulfurization (hydrotreating). As
crude oil reserves are depleted, more sour crudes are used as refinery stocks containing
elevated concentrations of sulfur and organo-metallics. Thus, the petroleum refinery process
of hydrodesulfurization has become essential in the utilization of poorer quality (sour) crudes
to enhance the yield of gasoline and reduce the level of these contaminants in residual fuel
oil. Various catalytic hydrotreatments are being employed in the desulfurization and
demetallization of refinery feedstock. The contaminant removal is accompanied by partial
conversion of nondistillables. Most hydrotreating processes are capable of greater than 90%
removal of sulfur and metals.
Because hydrotreating is a relatively expensive process, not all refinery feedstocks are
subjected to this treatment. The need for hydrotreating is primarily dictated by the sulfur
content. Thus, poorer quality (sour) feedstocks, which must be subjected to hydrotreating in
order to meet regulatory standards, are also likely to yield products, inclusive of residual fuel
oil, that are low in organo-metallics.
Crude oil fractionation is a variable and complex process. Petroleum refineries may be
classified on the basis of their primary products, associated processing, and resultant process
effluents. The U.S. EPA recognizes the following five basic refinery types:
1) A topping refinery employs the simplest distillation process, fractionating crude oil
into straight run fractions such as naphtha, middle distillates, and fuel oils by thermal
or vacuum distillation.
2) A cracking refinery, in addition to basic topping refinery operations, uses a variety of
catalytic or thermal "cracking" techniques that increase the proportion of lighter,
more volatile aromatic fractions. Lighter hydrocarbons with higher octane ratings are
blended into gasoline, kerosene, jet fuel, and domestic fuel oils.
4-9
-------
3) A petrochemical refinery, in addition to topping and cracking processes, incorporates
petrochemical operations. Petrochemical feed-stocks such as olefins, benzene,
toluene, and xylene are included among major products. (Petrochemical processing is
not considered a refinery operation and is not related to fuel/energy production.)
4) The lube refinery adds lubricating oil manufacturing processes as a major product to
those performed by topping and cracking refineries.
5) The integrated refinery includes all of the major refinery operations pertaining to
topping, cracking, lube oil manufacturing, and petrochemical processing.
There are approximately 200 refineries in the United States capable of processing IS million
barrels of crude oil per day (Table 4-5). The distribution of refineries is aggregated for each
of the five Petroleum Administration for Defense (PAD) Districts (see Figure 4-1). PADs
are an outgrowth of the World War II Petroleum Administration for War Districts. The 10
States with the largest total refining capacity in decreasing order are: Texas, Louisiana,
California, Illinois, Pennsylvania, Washington, Ohio, Oklahoma, Mississippi, and New
Jersey.
The yields of the major petroleum products vary with the quality of crude oil being
processed. For the total crude oil refined in the United States (i.e., domestic and imported
crude oil), the percentage yield of residual fuel oil in 1990 was 6.8%, which is about one-
fifteenth of the initial starting volume. Table 4-6 provides a breakdown of percentage yields
of the major petroleum products.
The significance of these percentage yields is that any residual radionuclides present in
refinery feedstocks are not likely to distribute themselves uniformly within the various
petroleum products. Although little is known about the chemical nature of the radioactive
trace metals, it is reasonable to assume that they exist as metal oxides, metal salts, and
organo-metallics and that the more volatile petroleum distillates will contain relatively small
amounts in comparison to the non-volatile bottom fractions. Available information suggests
that about 90% of the trace metals are retained in the residue fraction of which residual fuel
oil is a principal constituent (Smith 1975). As previously discussed, a second factor, which
obviates a rigid one-to-one relationship between radionuclide content in crude oil versus
residual fuel oil, is that refineries, as part of processing, demetallize feedstocks, inclusive of
crude oil, by various physical and chemical processes that are known to reduce the metal
content. The extent of application of these processes and their removal efficiencies are not
adequately documented.
4-10
-------
Table 4-5. Number and Capacity of United States Petroleum Refineries
District and SUM
Delaware
Georgia
New Jersey
New York
North Carolina
Pennsylvania
Virginia
Wea Virginia
PAD Diana 1 Tool
Illinois
Indiana
Kaiuai
Kentucky
Michigan
Minnesota
Nonh Dakota
Ohio
Oklahoma
Tennessee
Wisconsin
PAD Dtant O Tout
Alabama
Arkansas
Louisiana
Mississippi
New Mexico
Teias
P AD DiatnaD] Total
Colorado
Montana
Utah
Wyoming
PAD Dean* IV Tool
Alaska
Arizona
California
Hawaii
Nevada
Oiegon
Washington
PAD Diana V Total
Uuiaa Sou* Total
Number of
Operable Refineries
Toul Operating Idle
1 1 0
2 1 1
6 42
1 1 0
1 1 0
8 80
2 20
1 1 0
22 19 J
7 70
5 4 1
8 80
2 20
1 30
'2 20
1 1 0
4 40
6 60
1 1 0
1 1 0
4O 39 1
3 2 1
3 30
22 19 3
6 5 1
4 3 1
34 32 2
72 64 t
3 2 1
4 40
6 60
i 50
It 17 1
6 60
1 1 0
32 29 3
2 20
1 0 1
1 1 0
7 6 1
M 43 S
2OZ 114 18
Crude Capacity
Barrels Per
Calendar Day
Operating Idle
140.000 0
' 5.540 28.000
334.300 124.400
41 .8)0 0
3.000 0
744.313 0
56.700 0
12.500 0
I.3U.405 152.400
937.600 0
429.900 1.250
351.700 0
218.900 0
118.600 0
267.100 0
58.000 0
457.100 0
395.500 0
60.000 0
33.200 0
3.327.400 1.250
113.500 26.600
51.930 5.800
2.286.707 340.100
362.400 6.000
74.800 4.000
3.875.500 57.250
6.766.807 440.110
76.000 15.200
139.650 0
154.500 0
169.725 0
339.175 15.200
239.540 0
IO.OOQ 0
2.094.150 91.450
146.300 0
0 4.500
0 0
496.100 11.900
2.986.090 107.850
14.958.777 716.850
Barrels Per
Stream Day
Operating Idle
152.000 0
6.000 30.000
3)2.000 132.400
45.000 0
3.000 0
772.500 0
60.000 0
12.500 0
1.403.000 162.400
994.000 0
443.100 2.500
374.483 0
226.300 0
129.000 0
279.220 0
60.000 0
477.000 0
416,200 0
62.000 0
35,000 0
3.496.303 2.500
1 16.300 28.000
57,000 6.000
2.388.900 395.000
383.000 7.900
79,107 6.100
4.097.000 64.100
7.I2IJ07 607.500
85,000 16.000
145.500 0
160.000 0
175.750 0
566.250 16.000
254.700 0
12,000 0
2.217.400 98.700
150.000 0
0 4.700
0 0
531.000 I2.7S4
3.165.100 116.154
15.751.960 804.554
Charge Capacity (Barrels Per Stream Day)
Catalytic Catalytic Catalytic Catalytic
Vacuum Thermal Cracking Cracking Catalytic Hydro- Hydro-
DiiullaiKin Operation (Fresh) (Recycled) Reforming cracking treating
-95.000 49.000 67.000 5.000 54.000 18.000 123.000
000 000 2.940
258.900 21.000 256.000 37.000 77.500 17.000 298.000
28.000 00 0000
000 0000
320.250 0 247.000 6,800 200.400 51.000 471.820
29.000 14.000 27.500 2.000 10.200 0 26.500
6.000 000 3.700 4.440 4.000
737.150 80.000 597.500 50.800 345.800 90.440 917.260
383.900 126.300 378.000 10.000 302.300 67.500 617.600
235.450 28.000 173.000 4.200 99.800 0 267.800
124.650 52.500 123.800 9.000 93.800 .,3.190 211.000
92.000 57.600 100.000 0 46.000 ' 0 172.300
38.000 0 47.000 1.000 31.000 0 61.800
182.000 60.000 83.000 1.000 S9.500 0 227.000
0 0 26.000 3.600 12.100 0 19.100
172.000 31.700 174.000 17.500 162.600 87.200 196.500
147.000 26.500 149.000 5.000 101.300 5.000 177.500
12.000 0 30.000 0 10.000 0 30.000
20.500 0 11.000 1.000 8.000 0 14.800
1.407.500 382.600 1 .294.800 52.300 928.600 162.890 1.995.400
45.000 12.000 0 0 26.000 0 59.300
23.300 0 19.100 775 11.200 0 30.000
1.132.200 580.700 895.500 11.500 530.300 172.000 1.267.500
274.775 83.500 80.000 7.000 96.000 68.000 254.000
13.900 0 33.800 4.500 21.050 1.000 29.800
1.718.900 373.800 1.642.500 125.500 1.147.600 313.500 3.176.650
3,208.075 1.050.000 2.670.900 149.275 1.837.150 554.500 4.817.250
43.000 4.200 27.500 1.000 22.900 5.000 35.700
53.450 7.700 55.900 6.250 37.730 4.900 119.340
46.980 8.500 54.800 10.600 30.500 2.400 41.200
73.000 9.000 62.000 12.500 32.350 0 52.500
216.430 29.400 200.200 30.350 123.480 12.300 248.740
6.000 000 12.000 9.000 . 0
7.000 00 0000
1.347.600 530,900 656.700 14.000 542.300 408.800 1.475.180
74.250 13.000 20.000 0 13.000 18.000 3.500
000 0000
16.000 00 0000
255.500 72.000 118.500 7.000 128.500 52.000 219.000
1.706.350 615.900 795.200 21.000 695.800 487.800 1.697.680
7.275,505 2.157.900 5.558.600 303.725 3.925.830 1.307.930 9.676.330
-------
^\
^to ALASKA \
Figure 4-1. Petroleum Administration for Defense (PAD) Districts
-------
Table 4-6. Percentage Yields of Refined Petroleum Products
from Crude Oil in the United States, 1990
Petroleum Product
Gasoline
Jet Fuel
Liquefied Gases
Kerosine
Distillate Fuel Oil
Residual Fuel Oil
Petrochemical Feedstock
Special Naphthas
Lubricants
Wax
Coke
Asphalt/Road Oil
Still Gas
Miscellaneous
Shortage*
TOTAL
Yield (%)
45.8
10.7
3.6
0.3
20.9
6.8
2.9
0.4
1.2
0.1
3.9
3.2
4.8
0.5
^.9
100.0
Processing Gain (-) or Loss (+).
In summary, the fate of metals (which include the radioactive members of the U-238 and
Th-232 decay series) in crude petroleum undergoing refinement has not been thoroughly
studied. Thus, any assumptions or estimates regarding the radionuclide content of residual
fuel oil which are based on the radionuclide content of crude oil samples are subject to
considerable uncertainty.
4.2 A REVIEW OF PAST STUDIES RELEVANT TO ESTIMATING THE RADIO-
NUCLIDE CONTENT IN RESIDUAL FUEL OIL
Past interest and concern for the presence of trace metals that include radionuclides in
residual fuel oil and other petroleum products have not only been limited but, in most
instances where such studies were conducted, the primary motive was not concern for human
health. As previously noted, concern for the presence of trace metals is primarily due to
their interference in the catalytic cracking processes of refinery feedstocks, and the
destructive fouling effect of fuel-oil ash in industrial/utility furnaces, boilers, and associated
components. Not surprisingly, therefore, quantitative data for all but a few studies involved
the chemical (non-radiometric) analyses of metallic elements that at best identify uranium and
thorium, but not their radioactive daughter products.
4-13
-------
4.2.1 Residual Fuel Oil Studies
Chemical Analyses. Chemical analyses for quantifying metallic species most commonly
employ atomic absorption spectrometry. In a 1976 study aimed at identifying and
quantifying trace elements in residual fuel oil used by various utilization technologies, a
composite sample was developed for trace metal analysis (Vitez 1976). The composite
residual oil sample was based on a weighted average of U.S. crudes and represented a typical
composite of the U.S. residual oil feedstock used to supply the electric utilities and other
utilization technologies. Table 4-7 identifies 65 elements and their concentrations. The
concentration of uranium was measured to be 0.002 ppm. Although this study represents
measurements of a single sample, it is, nevertheless, significant since the sample corresponds
to a weighted composite of the U.S. residual oil feedstock used to supply the electric
utilities.
Table 4-7. Trace Element Concentrations in Residual Oil
Element
Sulfur
Silicon
Vanadium
Iron
Nickel
Calcium
Potassium
Aluminum
Magnesium
Sodium
Chlorine
Barium
Zinc
Phosphorous
Chromium
Cobalt
Manganese
Copper
Lead
Zirconium
Rubidium
Neodynium
ppfll jyiMiiaii
12.000.00
300.00
180.00
150.00
120.00
120.00
80.00
75.00
75.00
35.00
12.00
4.00
4.00
4.00
3.00
3.00
2.50
2.50
2.00
2.00
2.00
1.00
Element
Yttrium
Selenium
Titanium
Tin
Silver
Cerium
Thulium
Boron
Ruthenium
Bromine
Molybdenum
Cadmium
Platinum
Iodine
Strontium
Praseodymium
Tellurium
Germanium
Gadolinium
Hafnium
Tantalum
Terbium
PPQl TEFMHMl
1.00
1.00
0.40
0.30
0.20
0.20
0.20
0.30
0.20
0.28
0.25
0.20
0.20
0.15
0.10
0.10
0.10
0.10
0.10
0.05
0.05
0.05
Element
Osmium
Arsenic
Antimony
I anthamim
Mercury
Erbium
Scandium
Ytterbium
Cesium
Tungsten
Uranium
Dysprosium
Samarium
Lutetium
Holmium
Indium
Palladium
Europium
Gold
Rhenium
Indium
ppm median
0.01
0.02
0.02
0.02
0.01
0.01
0.01
0.01
0.005
0.005
0.002
0.002
0.002
0.001
0.001
0.001
0.001
0.0003
0.002 '
0.002
0.0002
Source: Vitez 1976
4-14
-------
Radiochemical Analysis. The Westinghouse Research and Development Center, under
contract to the EPA, conducted a comprehensive analysis of residual fuel oils for the purpose
of assessing mutagenic properties (Kaczmarck 1981). Among chemical constituents analyzed
were ten radioactive isotopes belonging to the uranium-23.8 and thorium-232 decay chains.
In total, 41 residual fuel oil samples were obtained from various oil company sources over an
extended period of time. However, due to the tardy arrival of some samples, only the first
28 samples received were analyzed. Table 4-8 presents the average activity concentrations of
the 28 samples as derived from the original study data. The original study data are provided
in Appendix B.
Table 4-8. Average Radionuclide Content for
28 Residual Oil Fuel Samples'
Radionuclide
U338
U-235
U-234
Th-230
Ra-226
Rn-222
Pb-210
Po-210
Th-232
Ra-228
Average Activity (pCi/g)
0.13
0.02
0.17
0.02
0.27
0.17
0.82
0.34
0.01
0.21
Average values were derived from the data of 28
samples. Values reported as less than the lower
limit of detection (LLD) were assumed one-half
the LLD value from computing the average.
Source: Kaczmarck 1981 (Westinghouse Study)
There is reason, however, to question the sensitivity and accuracy of this data set. The basis
for this concern involves the results of an EPA validation study described below.
The EPA Validation Study. Of the 28 samples analyzed by the Westinghouse Research and
Development Center, six samples (i.e., samples #10, 19, 23, 25, 27, and 28) were forwarded
to the EPA for independent and confirmatory analysis. Thus, samples #10, 19, 23, 25, 27,
and 28 were analyzed by both the Westinghouse and EPA laboratories and, therefore,
provide a basis for data comparison.
4-15
-------
Residual fuel oil samples were analyzed by EPA's Eastern Environmental Radiation Facility
in Montgomery, Alabama. Analytical and radiochemical procedures employed by the EPA
are defined in EPA 520/5-84-006, Eastern Environmental Radiation Facility Radiochemistry
Procedures Manual. EPA procedures are considered to be most appropriate and effective for
low-activity, ambient, and environmental levels and have been exposed to rigorous intra- and
inter-laboratory analytical controls. Standard sample count time is 1,000 minutes or longer
using state-of-the-art counting systems.
EPA's analysis of the six fuel oil samples is summarized in Table 4-9. Of the 10
radionuclides analyzed, eight belong to the U-238 decay chain and two to the Th-232 decay
chain. Based on individual and average sample activity concentrations, the following
statements apply:
• Decay-chain members do not exist in secular equilibrium.
• The activity concentrations for most of the radionuclides in residual fuel oil
are between two and three orders of magnitude lower than those for coal.
• The significantly highest average radionuclide concentrations correspond to
Pb-210 and Po-210. This suggests that the presence of Pb-210, with a half-life
of 22 years, is not dependent on the presence of its short-lived parent Bi-214.
More likely, its presence in fuel oil is dictated by physical/chemical properties
that affect the leaching and transfer from geologic formations to the brine.
• Individual and average values for samples #10, 19, 23, 25, 27, and 28 indicate
significant discrepancies between the Westinghouse and EPA studies.
The observed measurements for the six samples analyzed by the EPA are considerably lower
than the corresponding values reported by the Westinghouse study. A serious limitation for
the Westinghouse study is the threshold value of detectable quantities of radioactivity.
Average radionuclide concentrations cited by the Westinghouse study are about two orders of
magnitude higher than those reported by the EPA. Beyond sensitivity, a critical assessment
of analytical methods that were employed by the Westinghouse study leads to the conclusion
that there are other inaccuracies as well.
Due to the more rigorous and focused protocol employed by the EPA, we conclude that
values cited in Table 4-9 represent the more reliable data set. A notable limitation of the
EPA study, however, is the small number of samples analyzed.
4-16
-------
Table 4-9. EPA Analysis of 6 Residual Fuel Oil Samples*
Radio-
nuclide
U-238
U-234
Th-230
Ra-226
Pb-214
Bi-214
Pb-210
Po-210
Th-232
Th-228
Sample ID
10-Flna
TX, AK
0.004 ± 0.0004
0.004 ± 0.0004
0.0008 ± 0.0001
0.009 ± 0.0002
0.016 ± 0.004
0.018 ± 0.005
1.55 ±0.99
0.13 ±0.079
0.002 ± 0.0002
0.006 ± 0.0003
19-Exxon
Persian Gulf
0.0001 ± 0.00004
0.0002 ± 0.00007
0.0007 ± 0.00004
0.0002 ± 0.00003
ND
NO
0.807 ±1.12
0.18 ±0.073
0.00005 ± 0.00003
0.0002 ± 0.00007
23-Shell
CA, other
couniries
0.004 ± 0.0005
0.007 ± 0.0004
0.013 ± 0.0005
0.0004 ± 0.00003
0.006 ± 0.002
0.005 ± 0.002
0.46 ± 0.65
0.29 ± 0.085
0.005 ± 0.0003
0.041 ± 0.001
25-Cltgo
U.S., Persian
Gulf, Africa.
S. America
0.0004 ± 0.0005
0.004 ± 0.0005
0.001 ± 0.0001
0.004 ± 0.00008
0.032 ±0.012
0.027 ± 0.007
0.62 ± 0.71
0.31 ± 0.11
0.002 ± 0.0002
0.019 ± 0.0008
27-Shell
Mid-East. TX,
other countries
0.0004 ± 0.0001
0.0005 ± 0.0001
0.071 ± 0.025
0.003 ± 0.00009
0.017 ± 0.004
0.013 ± 0.006
0.164 ±0.455
0.44 ± 0.13
0.056 ± 0.022
0.20 ± 0.042
28-Shell
Africa. LA, other
countries
0.0005 ± 0.0001
0.0005 ± 0.0001
0.0005 ± 0.00009
0.009 ± 0.0002
0.022 ± 0.005
0.020 ± 0.008
0.081 ± 1.66
0.15 ± 0.06
0.0003 ± 0.00008
0.007 ± 0.0003
Average Value
0.0016 ± 0.0001
0.0027 ± 0.0001
0.0014 ± 0.004
0.0043 ± 0.0001
0.019 ± 0.003
0.017 ± 0.002
0.614 ± 0.401
0.25 ± 0.03
0.011 ± 0.0037
0.045 ± 0.007
f-
fr-*
-4
* Values are expressed in pCi/g ± 2 standard errors
ND = Not delectable or below LLD; LLD values for Pb-214 and Bi-214 are 0.001.
-------
4.2.2 Crude Oil Study Data
Chemical Analysis. It is well established that crude oil may contain as many as 60 different
metals in measurable quantities. The-collective metal concentrations in crude oils vary
widely from 1 to more than 10,000 ppm. Vanadium and nickel are usually the most
prominent, and their concentrations generally increase with the asphaltic content or specific
gravity of the crude oil. Trace metal analyses of crude oils and residual fuel oils have also
shown the presence of uranium. Such data, however, lack information regarding the
presence of uranium decay products, which would not be detectable by the non-radiometric
techniques commonly used in trace metal analyses. Thus, the use of data that are limited to
uranium measurements of oil samples requires that unsupported assumptions be made
regarding the presence/concentrations of other decay-chain products. The following
' summarizes existing data and identifies the limitations for their use in estimating the
radionuclide content in residual fuel oil.
In the past, the U.S. Geological Survey has studied domestic petroleum from the standpoint
of geochemistry. Specifically, the interest focused on the potential mining of geological
formations and their relationship to trace metals found in crude oil. In a study published in
1960 (Ball I960), 24 samples of domestic crude oil were quantitatively analyzed for
vanadium, nickel, copper, and uranium. Table 4-10 identifies the concentrations of uranium
as well as the corresponding ash content in each of the 24 samples of crude oil.
Uranium concentrations among the 24 crude oil samples ranged from 0.00004 ppm to 0.013
ppm, with an average value of 0.00186 ppm.
These values are nearly identical to an earlier study which also involved 24 samples of
domestic crude oil (Erickson 1954). In this study, the range in uranium concentrations
varied from 0.00001 to 0.064 ppm, with an average value of 0.0008 ppm.
In a more comprehensive study, an analysis of composite crude oil samples obtained from
California, Libya, Venezuela, and Alberta showed that the concentration of some trace
metals varied by as much as three orders of magnitude (Klein 1975). Uranium was found in
detectable quantities only in Libyan crude at a concentration of 0.015 ppm (Table 4-11).
Radiometric Analysis. In 1979, the EPA's Eastern Environmental Radiation Facility also
analyzed the radionuclide concentrations in 13 crude oil samples. Crude oil samples were
4-18
-------
Table 4-10. Uranium Content in 24 Domestic Crude Oil Samples
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
17
18
19
20
21
22
23
24
Mean ± SD
Ash Content
<%)
0.041
0.004
0.022
0.001
0.001
0.002
0.002
0.130
0.012
0.003
0.001
0.022
0.004
0.010
0.038
O.OS8
0.023
0.038
0.012
0.038
0.001
0.015
0.03S
0.018
0.022 ± 0.028
Uranium Concentration
(ppm)
0.0025
0.00048
0.00088
0.00017
0.00050
0.00028
0.00024
0.00260
0.00012
0.00054
0.00032
0.0020
0.00004
0.00030
0.00076
0.0017
0.00069
0.0015
0.00036
0.0011
0.00024
0.0065
0.013
0.0077
0.00186 ± 0.00305
Souice: Ball 1960
obtained from two domestic and eleven foreign sources. Table 4-12 summarizes the findings
for the 13 crude oil samples. Of the 10 radionuclides analyzed, Th-232 was found in
detectable quantities in only two samples and Pb-214 and Bi-214 in only four samples. All
other radionuclides were below the lower limits of detection.
Limitations of Crude Oil Data. It is axiomatic that all radioactivity found in residual fuel oil
must have also been present in crude oil. However, a simple and constant ratio of
concentrations cannot be assumed for several reasons. In Section 4.1 of this chapter and in
Appendix A, various processes are described (i.e., deasphalting, demetallization, and
hydrotreating) which effectively reduce metals/radionuclides in virgin and cracked
feedstocks. Even if the removal efficiencies of these processes were known, not all
feedstock undergo such treatments.
4-19
-------
Table 4-11. Trace Element Contents of Some Crude Oils
Concentration in Crude Oil (ppm)
Element
V
Cl
I
S
Na
K
Mn
Cu
Ga
As
Br
Mo
Cr
Fe
Hg
Se
Sb
Ni
Co
Zn
Sc
•U
California-"
(Tertiary)
7.500
1.470
--
9.900
13.200
—
1.200
0.930
0.300
0.655
0.290
—
0.640
68.900
23.100
0.364
0.056
98.400
13.500
9.760
8800.000
-
Libya
8.2000
1.8100
-
4694.0000
13.0000
4.3900
0.7900
0.1900
0.0100
0.0770
1.3300
—
0.0023
4.9400
-
1.1000
0.0550
49.1000
0.0320
62.9000
282.0000
0.0150
Venezuela
(Boscan)
1100.000
—
—
—
20.300
—
0.210
0.210
..
0.284
..
7.850
0.430
4.770
0.027
0.369
0.303
117.000
0.178
0.692
4400.000
—
Alberta
(Cretaceous)
0.6820
25.5000
1.3600
1450.0000
2.9200
..
0.0480
..
..
0.0024
0.0720
..
..
0.6960
0.0840
0.0094
..
0.6090
0.0027
0.6700
—
—
Source: Klein 1975
Opposing the various treatments that reduce the metal content in refinery feedstocks, the
subsequent fractionation process is likely to concentrate the remaining metals/radionuclides in
the non-volatile still-bottoms that include residual fuel oil. Thus, there exists the potential
for reducing, as "well as increasing, the initial radionuclide concentration present in crude oil
in the production of residual fuel oil. A prudent and cautionary use of crude oil data is to
regard values as nominal surrogate measurements that have the potential of over-estimating
and underestimating the true metal/radionuclide content in residual fuel oil.
4.2.3 Residual Fuel Oil-Ash Data
The burning of fuel oil creates an ash residue that consists primarily of organically-bound
metals and metallic oxides and salts. The average ash content in fuel oil is estimated to be
4-20
-------
Table 4-12. Radionuclide Activity in Crude Oil Samples Analyzed by EPA
Sample Number
90896
90897
90898
90899
90900
90901
90902
90903
90904
90905
90906
90907
90908
Source
A-TJxxon off shore LA
Agah-Iran
Mesa-Venezuela
Escravos-Nigena
Sarir-Libya
Arabian Heavy Saudi
USSR
UAE-Dubai
North Slope
Kern River
Indonesia
Arabian Light Saudi
B-l Texas
Activity (pCi/g)
LMDL*
LMDL*
Pb-214 0.001
Bi-214 0.001
LMDL*
LMDL*
LMDL'
LMDL*
Th-232 0.022
Pb-214 0.02
Bi-214 0.006
Pb-214 0.017
Bi-214 0.011
Th-232 0.026
LMDL*
Pb-214 0.022
Bi-214 0.011
LMDL*
LMDL - less than measurable detection limit. LMDL is a function of
several factors including the type, energy, and percent yield of the
radiation emission, background activity, counting system employed, and
sample counting time.
Source: EPA unpublished data
about 0.1 percent by weight. Thus, information involving oil-ash analysis provides a useful
means for deriving the metal/radionuclide concentration for residual fuel oil. The
significance of ash sample analysis rests with the fact that an individual ash sample must be
considered the equivalent of a composite sample (i.e., ash samples are representative of the
combustion of a large quantity of fuel consumed over a period of time). A review of the
literature revealed several studies that evaluated the radionuclide content in oil ash.
Chemical Data of Oil Ash. In the first study, oil ash was analyzed for trace metals by
atomic absorption spectrometry (Ball 1960). Ash concentrations for uranium ranged between
0.0001 and 0.01 percent (i.e., 1-100 ppm). Thorium was not present in detectable
concentrations. Corresponding U-238 concentrations and specific activity values in the
parent fuel oil are given in Table 4-13 and were derived by assuming a fuel oil-ash content
of 0.1%.
4-21
-------
Table 4-13. Derived Uranium Values in Fuel Oil from Analyses
of Fuel Oil-Ash Samples
Measured Uranium
Concentration in
Study Oil Ash (ppm)
Ball (1960) 1 - 100
Derived U-238*
Activity in Fuel
Oil (pCi/g)
0.0003 - 0.03"
Geometric Mean of
U-238 Activity in Fuel
Oil (pCi/g)
0.003
Ash Content of 0.1% is assumed for residual fuel oil; and 1 ppm U corresponds to an activity
of 0.33 pCi/g for U-238.
Radiometric Data. Two studies have been identified in which comparative radiometric
measurements were reported for naturally occurring radionuclides in the residues of residual
fuel oil and the fly ash of coal. (The inclusion of coal-ash data in this section is explained
below.)
The first study (Eisenbud 1964) analyzed (1) six samples of fly ash obtained from the
electrostatic precipitators of boilers burning six different batches of pulverized semi-
bituminous Appalachian coal; and (2) three samples of the solid residues formed when fuel
oil is burned from the brcachings of smoke stacks carrying the combustion products from
boilers fired with Venezuelan oil.
The ash samples were analyzed radiochemically for the radionuclides Ra-226, Ra-228, and
Th-228. Results of the analysis are shown in Table 4-14. For coal ash, the near equality of
the Ra-228 and Th-228 content indicates that secular equilibrium exists. It can also be
inferred that the levels of activity for both Ra-228 and Th-228 must be supported by an
equivalent activity of Th-232.
For petroleum ash, however, the ratio of Th-228 to Ra-228 is 1.4, which is near the value of
1.5 that would be expected for transient equilibrium supported by the Ra-228 parent. These
data support the previously stated observations that for coal the whole decay-chain series is
contained within the matrix in secular equilibrium and for petroleum, radium may
preferentially be leached from the geologic substrate.
4-22
-------
Table 4-14. Radioactivity in Fuel Oil and Coal Ash
Sample Data
Oil-Ash Sample
ffl
n
«
Mean ± SD
foal-Ash Sample
#1
n
#3
#4
#5
#6
Mean ± SD
Ra-226 (pCi/g)
V
0.14
0.18
0.30
0.21 ± 0.07
5.0
4.2
3.8
2.8
4.8
2.1
3.8 ± 0.4
Ra-228 (pCi/g)
0.45
0.32
0.70
0.49 ± 0.16
1.9
2.5
2.6
1.8
2.7
3.1
2.4 ± 0.4
Th-228 (pCi/g)
0.67
0.52
0.83
0.68 ±0.13
2.5
2.7
2.4
2.1
2.4
3.3
2.6 ± 0.4
Source: Eisenbud 1964
On the assumption that the average ash content of coal is 10% and of oil 0.1 %, activity
concentrations may be estimated for Ra-226, Ra-228, and Th-228 in behalf of coal and oil
that were the source of the ash samples (Table 4-15). Estimates for U-238 and Th-232 are
based on secondary assumptions that these two parent nuclides (1) exist in secular
equilibrium with their corresponding progeny in coal and (2) exist in concentrations equal to
or less than their corresponding progeny in oil.
Table 4-15. Derived Radioactivity Values for Coal and Fuel Oil
from Data Contained in Table 4-14
Coal
Sample
Average
QU
Sample
Average
Coal/Oil
Ra-226
(pCi/g)
0.38
0.0002
Ra-228
(pCi/g)
0.24
0.0005
Th-228
(pCi/g)
0.26
0.0007
U-238
(pCi/g)
0.38
20.0002
Th-232
(pCi/g)
0.26
2 0.0007
Activity Ratio Values
1900
480
371
2: 1900
£ 371
4-23
-------
Table 4-15 reveals that the derived radionuclide concentrations for oil are several orders of
magnitude lower than those for coal. The credibility of the ash data is based on the
observation that the average derived uranium and thorium concentrations for coal in the
Eisenbud Study are virtually identical to values obtained from the USCHEM coal data base
of nearly 7,000 coal samples (see Table 2-15).
In 1969, a Congressional hearing was held that addressed concerns about the environmental
effects of producing electricity (Hearings Before the Joint Committee on Atomic Energy of
the United States, First Session on Environmental Effects of Producing Electric Power).
Information was presented at the hearing that included a study conducted by the Bureau of
Radiological Health's Southeastern Radiological Health Laboratory (SERHL) (Oakley 1968).
As part of a comprehensive study that included an environmental field survey (summarized
below in Section 4.2.3), samples of coal ash and oil ash were analyzed for Ra-226, Ra-228,
and Th-228. In total, six oil-ash samples from a single power plant were compared to three
sets of coal-ash samples obtained at three different coal-fired plants. The results of the
SERHL study are summarized in Table 4-16. Their equivalent derived values for estimating
the radionuclide content of oil and coal from which ash samples were derived are cited in
Table 4-17.
The results of the SERHL Study (Oakley 1968) closely parallel and support the earlier study
data reported by Eisenbud (1964).
Table 4-16. Comparison of Radioactivity in Oil and Coal Ash
Sample
Type
Oil Ash
Coal Ash
Source
(Plant)
Turkey Point
Hartsville
Colbert-TVA
Widows Creek-TVA
Number of
Samples
6
5
12
26
Ash Concentration (pCi/g)
Ra-226
0.18
2.3
3.1
1.6
Ra-228
0.17
3.1
6.9
2.7
Th-228
0.82
NP'
1.6
2.8
* NP signifies that analysis was not performed.
Source: Oakley 1968
4-24
-------
Table 4-17. Derived Average Radionuclide Concentrations in Oil and Coal
from SERHL Fly-Ash Study
Fuel
Resid Oil
Coal
Coal/Oil Ratio
„;- Derived Activity Concentrations (pCi/g)
Ra-226
0.0002
0.23
1,150
Ra-228
0.0002
0.42
2,100
Th-228
0.0008
0.22
275
U-238
1,150
Th-232
< 0.0008
0.22
> 275
Conclusion. Data from three independent studies that analyzed fuel oil ash show highly
consistent values. By simple extrapolation, ash data suggest that the average radionuclide
content of the parent fuel oil is more than two orders of magnitude lower than the
radionuclide content normally found in coal. Moreover, the strength of these studies is based
on the fact that individual ash sample(s) are representative of a large volume of combusted
fuel and, therefore, may be viewed as "composite samples."
4.2.4 Radionuclide Emission Data from Oil-Fired SGUs
In the 1960s, the Department of Health, Education, and Welfare's Bureau of Radiological
Health published data that had been collected in behalf of a comparative study which assessed
radionuclide emissions from fossil-fueled and nuclear-powered steam-electric generating
plants (Gordon 1968, Bedrosian 1969). Environmental sampling data for oil-fired plants
were obtained from a two-year study of the Florida Power and Light Company's Turkey
Point Site following start-up in 1967. The facility has two oil-fired SGUs, each having a
capacity of 430 MWe.
Sampling at the Turkey Point site included six continuously operating air samplers evaluated
monthly. Soil, vegetation, silt, and water samples were analyzed to determine if
radionuclides released in fly ash, during routine operations, resulted in detectable
radioactivity in these media. The study concluded that radionuclide emissions from the oil-
fired SGUs were not high enough in the presence of their natural concentrations in air and
environmental samples to be accurately assessed.
4-25
-------
4.2.5 Summary of Past Study Data
A comprehensive literature survey indicates that there are limited data on radionuclide
concentrations in residual fuel oil, or from which such values may be inferred. In this
section, available information was presented of past studies regarding trace metal and/or
radionuclide analyses of (1) residual fuel oil, (2) crude oil, (3) residual fuel oil ash, and (4)
environmental samples. Table 4-18 summarizes salient data as reported in these studies.
In a 1973 report, the AEC estimated that the total airborne releases of radionuclides from an
oil-fired plant were about one-hundredth of those from a coal-fired plant (U.S. AEC 1973).
On the basis of the fly-ash data reviewed in this report, the National Council on Radiation
Protection and Measurements (NCRP) concluded in a 1987 report that a typical 1,000 MWe
coal-fired plant could be expected to release 885 times more radioactive material than a
typical 1,000 MWe oil-fired plant (NCRP 1987a).
With the exception of the Westinghouse study,1 the aggregate of studies conducted to date
and summarized in Table 4-18 support past conclusions that the radionuclide content in
residual fuel oil is low relative to coal. The total number of studies that support this
conclusion, however, are few. Moreover, most of these studies are incomplete since values
for all but a few decay-chain members must be derived through extrapolation and/or
unconfirmed assumptions.
For these reasons, the EPA concluded that there was a need for additional data. Early in
1993, an independent study was initiated to obtain additional information. The general
protocol, analytical laboratory methods, data results, and data analysis of the EPA-sponsored
study are discussed in Section 4.3 below.
4.3 RESULTS OF A RECENT EPA STUDY
Early in 1993, the EPA informed the Electric Power Research Institute (EPR1) and the
Utility Air Regulatory Group (UARG) of its intent to conduct a limited survey study in
which samples of residual fuel oil would be analyzed for radionuclides. During round-table
discussions with EPRI and UARG, two options were considered as potential sampling
strategies.
1 Due to limitations imposed by analytical protocols, the Westinghouse study is insufficiently sensitive and
accurate to be useful.
4-26
-------
Table 4-18. Summary of Past Study Data for Deriving the Radionuclide
Content for Residual Fuel Oil
Sample (Reference)
Analytical
Method*
Summary of Reported Study Data
(Concentrations in Sample Medium)
Residual Fuel Oil
Vitez(1976)
Kaczmarck (1981)'
(Westingbouse Study)
EPA (unpublished)
chemical
radiochemical
radiochemical
0.002 ppm Ub
0.13pCi/g U-238
0.39pCi/g Ra-226
0.85pCi/g Pb-210
0.02pCi/g Tli-232
0.30pCi/g Ra-228
0.0016 pCi/g U-238
0.0043 pCi/g Ra-226
0.614 pCi/g Pb-210
0.011 pCi/g Th-232
0.045 pCi/g Th-228
Crude Oil
Erickson (19S4)
Ball (1960)
Klein (1975)
EPA (unpublished)
chemical
chemical
chemical
radiochemical
0.00001 - 0.064 ppm U
0.00186 ppm U
0.015 ppm U
£0.005 pCi/g Pb-214; Bi-214
£0.011 pCi/g Th-232
Fuel Oil Ash
Ball (1960)
Eisenbud (1964)
Oakley (1968)
(SERHL Study)
chemical
radiochemical
radiochemical
1 - 100 ppm U"
0.21 pCi/g Ra-226
0.49 pCi/g Ra-228
0.68 pCi/g Th-228
0.18 pCi/g Ra-226
0.17 pCi/g Ra-228
0.82 pCi/g Th-228
Gordon (1968);
Bedrosian (1969)
radiochemical
-------
The first would obtain samples from refineries that produced residual fuel oil from domestic
and imported crude oil. The criteria for selecting major oil companies/refineries (see Tables
4-4 and 4-5) for sample procurement would be that their combined annual production/sale of
residual fuel oil represented at least 25% of the total annual United States production/sales.
Another selection criterion would consider the geographic location of refineries, which
impacts the source of crude oil for the following reasons: due to the economic cost of
shipping crude oil, refineries in States with significant crude oil production may differentially
rely on local/domestic supplies; conversely, refineries that are more suitably located to the
port of entry of foreign oil may differentially rely on foreign crude oil for their production of
residual fuel oil.
A second study option was to request fuel oil samples directly from utility users. Under this
approach, the voluntary participation of individual utilities would be solicited. The selection
of individual utilities for submitting samples would be based on generator nameplate
capacity, capacity factor, and/or annual consumption. Selection would, therefore, favor
larger facilities with the highest capacity factors and/or fuel consumption. An attempt would
also be made to select utilities by geographical location for reasons cited above (i.e., a
geographic selection parameter accounts for radionuclide variability based on origin and
refinement of crude oil).
After analyzing both options, the second approach was considered preferable. The distinct
advantage of this option is that samples obtained from utilities represent fuel oil actually
consumed by the utilities under study by the EPA. (Note: under Option #1, it is possible to
target refineries that do not supply any fuel oil to utilities or supply utilities in quantities that
are not in proportion to their total production quantities.)
UARG and EPRI were asked to solicit the voluntary participation of individual utilities and
were urged to select utilities on the basis of generator nameplate capacity, capacity factor,
and geographic location.
To ensure that samples submitted by individual oil-fired plants are representative of their
station inventory of residual fuel oil, utilities were given written instructions to obtain
sample(s) from a burner feed-line (or equivalent) during plant operation and only after the
plant had been at power for a substantial period of time. Participating utilities were also
requested to supply sample demographic data.
4-28
-------
4.3.1 Utility Participation and Sample Data Information
In total, 12 utilities identified as Utility #1 through #12 provided 42 samples of residual fuel
oil. Table 4-19 summarizes information extracted from data sheets submitted in behalf of
individual samples. Utilities and corresponding sources are listed in the order in which fuel
oil samples were received and subsequently identified (Table 4-19). In all but three cases,
the utility source is represented by the on-site fuel inventory for a single power plant. For
Utility #2 and Utility #3, the source was a common storage facility that supplied multiple
plants operated by the respective utility. No information regarding sample source was
provided in behalf of samples #27 and #28 submitted by Utility #8. In most instances, the
sample was obtained from a burner feed-line or from an autosampler during fuel delivery
(Column 4). Samples derived from a storage facility represented composite samples that
were obtained by simple random sampling as defined in Chapter 9 of EPA's (OSWER) Test
Methods for Evaluating Solid Waste. Volume II: Physical/Chemical Methods. Report No.
PB-88-239224, November 1986.
Columns 5 and 6 identify the utility supplier and the geographic origin of the fuel oil or the
crude oil from which it was derived. In only a few instances, the plant operator of a utility
and/or the utility's supplier was able to identify the geographic origin. The lack of
information is likely due to the fact that most refineries receive and combine crude oil from
multiple sources that vary with time.
Column 7 provides data regarding the utilities' annual consumption of residual fuel oil. For
the previous year, 1992, plants that had provided samples collectively consumed about
2,000,000,000 gallons. The annual consumption by these participating utilities is estimated
to be about 24% of the fuel oil consumed by all oil-fired electric utility steam generating
units (U.S. Energy Information Administration, Monthly Energy Review, March 1992).
Residual fuel oil is commonly assessed for a variety of properties, such as specific gravity,
viscosity, asphalt content, etc. Table 4-20 summarizes the physical properties for the oil
samples submitted by utilities.
4.3.2 Sample Receipt and Sample Preparation
Samples submitted by each of the 22 oil-fired power plants were forwarded directly to EPA's
National Air and Radiation Environmental Laboratory (NAREL) in Montgomery, Alabama.
4-29
-------
Table 4-19. Demographic Data for Oil Samples
Utfflty
n>
Number
#1
n
n
#3
»4
#4
#5
#5
16
n
in
m
n
#8
#9
#10
#11
#11
#12
#12
#12
#12
Simple Source
(Power Plant
n»
1-A
2-A
2-B
3-A
4-A
4-B
5-A
5-B
6-A
7-A
7-B
7-C
7-D
8-A
9-A
10-A
11-A
11-B
12-A -
12-B
12-C
12-D
Sample
n>
1.2
3.4
16: 17
5
6: 7; 8
9; 10;
1!
12: 13
14; IS
18
19:20
21:22
23:24
25:26
27; 28
29:30
31
32:33
42
34; 35
36:37
38:39
40:41
Smuullny Method
Tank Sample Composite
Barge B-10 composite sample of
12 compartments
Weighted avenge of 12
compartments on Barge
Tank sample composite
Feedline
Feedline
Feedline to Burner
Feedline to Burner
Grab sample from fuel oil pump
strainer
Feedline to Burner #1
Feedline m Burner #3
Feedline to Unit #4
Feedline to Unit U
Not Specified
Feedline to Unit #5
Feedline to Boiler #3
Burner Pump Room
Fuel oil pump discharge header
Autosampler during fuel
delivery
Autosampler during fuel
delivery
Autosampler during fuel
delivery
Autosampler during fuel
delivery
Supplier
New England Petroleum Corp.
Coastal-Eagle Point
Coastal Oil. NY
Amanda Hess
Variable
Variable
Coastal Fuels Marketing
Coastal Fuels Marketing
Chevron
Coastal Supply
Housnn Fuel Oil Terminal
Coastal Supply
International Matw Tank
Not Specified
Clark Oil; Sun Oil; Coastal
Oil: Global Oil; Hess 00
Sun RAM Northeast
Petroleum:: Spot Suppliers
Coastal Supply
Spot Market
Not Specified
Not Specified
Not Specified
Not Specified
Origin
Unknown
China/Norway
West Africa
Unknown
Unknown
Unknown
Unknown
Unknown
Indonesia
Unknown
Unknown
Unknown
Unknown
Not Specified
Variable (Spot
Market)
Unknown
Not Specified
Not Specified
Gulf Coast
Gulf Coast
Gulf Coast
Tnmdad
Gallons Used
U199Z
28,540.000
74.000.000
Not Specified
29.000.000
35.500.000
81.400.000
151.000.000
243.000.000
44.560
87.600.000
109.200.000
53.800.000
1.700.000
Not Specified
92.000.000
1.900.000
92.300.000 '
182.400.000
294.000.000
168.000.000
100.800.000
159.000.000
Total: 1,956,644.560
4-30
-------
Table 4-20. Physical Properties of Residua] Oil Samples
Sample
ID
i;2
3; 4
5
6; 7; 8
9; 10; 11
12; 13
14; IS
16; 17
18
19; 20
21; 22
23;24
25;26
27; 28
29: 30
31
32; 33
34;3S
36; 37
38:39
40;41
42
Mean
SD
Specific
Gravity
CAPO
NS-
16.4
16.5
10.1
11.0
10.5
10.2
16.8
20.4
10.5
11.6
11.0
18.1
NS
11.0
13.25
NS
8.5
11.0
8.9
10.9
NS
12.7
3.5
Pour
Point
(°F)
NS
75
48
28
21
NS
NS
50
117
37.5
37.5
37.5
37.5
NS
65
60
NS
NS
NS
NS
NS
NS
51.2
25.9
Viscosity
"at 100°F
NS
290"
543
186"
114"
NS
NS
108"
Solid
6,182
5.670
6.496
269
NS
500
2,450
NS
185"
189"
49"
293
NS
2.800.4
160.1"
2.841.9
77.7"
Carbon
(*)
NS
NS
86.4
NS
NS
NS
NS
NS
4.49
87.4
87.8
87.7
86.7
NS
NS
NS
NS
NS
NS
NS
NS
NS
73.4
33.8
Water &
Sediment
(%)
NS
0.16
0
0.1
0.5
NS
NS
0.03
3.6
0.4
< 0.1
0.75
0.75
NS
< 0.1
< 1
NS
0.13
0.64
0.10
0.13
NS
0.53
0.88
Asphalt
(%)
NS '
1.0
3.6
NS
2.55
NS
NS
0.9
0.94
7.46
8.19
8.86
1.67
NS
6.4
NS
NS
3.1
4.4
1.6
3.1
NS
3.84
2.80
Heating
Value
(BTU/lb)
NS
18.886
18.726
18.194
18.420
18.141
18.286
18.846
18.942
18.258
18.389
18.293
18,680
NS
18,349
15.100
NS
18.222
18.209
18.271
18.205
NS
18.245.4
827.9
' NS = Not Specified.
" Viscosity at 122°F instead of 100°F.
4-31
-------
S. Cohen and Associates, Inc., the contractor to the EPA for the conduct of this study,
maintained custody and control over all samples received. Samples were logged in and given
an identification number that coincided with the chronological sequence in which samples
were received.
Because of the possibility that radionuclides in residual fuel oil may exist as insoluble
particulates that are distributed non-unifonnly, primary samples were first heated in a water
bath for one hour until thoroughly liquified at a temperature of 75 °C. The heated samples
were then subjected to mechanical agitation for S minutes to ensure sample homogeneity.
Multiple aliquots of 100 ml were immediately drawn into polyethylene bottles and allowed to
cool to room temperature. All sample aliquots were tightly capped and stored in a secured
cabinet with controlled access.
It is important to note that samples 1 through 30 were received between April 21 and the
latter part of May, 1993. A single composite sample was derived by combining an aliquot of
constant volume from each of the 30 samples received. In order to meet the original
timetable that had been established for the completion of the study, aliquots of samples 1
through 30 and the composite sample were forwarded without delay to a commercial
analytical laboratory for analysis. (Note: The 30-sample composite was given the ID
Sample Number 43.)
In response to a renewed request on the part of EPRI/UARG for utility participation in this
study, three utilities, representing six power stations, furnished samples 31 through 42 in
July/August 1994. Aliquots of samples 31 through 42 were also forwarded to the contractor
laboratory for analysis in September. Due to time constraints, a revised composite sample
for all 42 samples, however, was not forwarded for analysis.
4.3.3 Sample Analyses
The laboratory selected to analyze the oil samples is one of the major commercial analytical
laboratories. The selection process evaluated laboratories based on their experience and the
existence of established analytical protocols involving petroleum sample media. Competitive
bids were solicited from three commercial laboratories. On the basis of radionuclide-specific
detection limits, a 45-day delivery schedule, and costs, TMA/Eberline was selected. All
samples were analyzed for (1) radium-226/228, (2) isotopic uranium, (3) isotopic thorium,
and (4) lead-210 by means of the following methods:
4-32
-------
• Ra-226 - Modified EPA Method #903.0
• Ra-228 - Modified EPA Method #904.0
• Isotopic Uranium - EML-HASL U-04 Method
• Isotopic Thorium - EML-HASL Th-01 Method
• Pb-210 - Modified EML-HASL Pb-01 Method
A summary description of these methods is provided below.
(1) Approximately 10 grams of each oil sample aliquot was weighed in a glass beaker
containing a few grams of ashless filter paper pulp. Radioisotope tracers and carriers
were added to all samples at this time. Laboratory control samples (LCS) were
prepared by adding a measured aliquot of a NIST traceable standard solution of the
appropriate radioisotope to a beaker containing filter paper pulp. These beakers were
placed on a gas burner on low flame and heated until there was only a charred residue
left. Samples were then muffled at 600°C for approximately 8 hours.
(2) Radium-226/228 samples were digested in acid and the barium carrier and radium
precipitated out as the sulfate. The sample was redissolved in an EDTA solution and
the sulfate precipitated out again. Chemical recovery was determined by gamma
counting of a barium-133 tracer and the determination of the radium-226 was carried
out by alpha spectrometry. After alpha counting, the precipitate was redissolved and
an yttrium carrier added. After allowing for ingrowth, the actinium-228 daughter of
radium-228 was precipitated out using the yttrium oxalate as a standardized carrier
both to determine recovery and carry the actinium. The sample was beta counted and
corrected for mass absorption and decay to determine the radium-228 present in the
sample.
(3) Lead-210 was determined by transferring the sample to a hydrochloric acid solution
and extracting it into a 0.1 % solution of diethylammonium diethyldithiocarbamate in
chloroform. The first decay product of lead-210, bismuth-210, was extracted. The
chloroform fraction was then evaporated off, and the dissolved residue was analyzed
by atomic absorption. The bismuth was then precipitated as BiOCL and counted by
beta proportional counting.
Note: When data submitted by the contract laboratory hi behalf of Pb-210 was
subsequently compared to the composite sample analysis performed independently by
NIST and NAREL, it was concluded that the low temperature muffling at 600°C for
8 hours in step #1 above most likely volatized the lead content and, therefore,
invalidated the Pb-210 results. The results of the composite sample analysis by NIST
and NAREL and a reanalysis by the contract laboratory are discussed in Section 4.3.5
below.
4-33
-------
(4) Isotopic uranium and thorium separations were achieved by ion chromatography and
mounted by microprecipitation using neodymium fluoride coprecipitation onto a
membrane filter. The chemical recovery was traced using uranium-232 and thorium-
229.
(5) Alpha spectroscopy was performed using the Nuclear Data Genie alpha management
software.
(6) Calculation of the beta activity for radium-228 and lead-210 was carried out using
spreadsheets created by TMA/Eberline using equations found in ANSI 13.30.
(7) Samples were grouped into eight subsets and analyzed at discrete intervals. A spike
and blank sample were included for analysis of each subset of fuel oil samples.
4.3.4 Sample Results
The original data sheets submitted by the contract laboratory for the 42 primary oil samples
and one composite sample are summarized in Table 4-21. Inspection of the table reveals that
fully two-thirds of all measurements were reported as "less than a measurable value," that
corresponds to a lower limit of detection (LLD). (Furthermore, and as noted above, activity
concentrations for Pb-210 were considered invalid.)
Determination of the LLD value, which describes the sensitivity of a detection system under
specific counting conditions, is complex. The most significant variables that affect counting
conditions are (1) duration of sample counting time, (2) background activity and duration of
background counting time, and (3) detector system counting efficiency for radionuclide(s)
under investigation. In effect, a given LLD specifies the minimum true source activity that
would yield a significant sample activity in a specified percentage of measurements. The
minimum value depends on the confidence interval used in accepting a count/activity as
significant and the acceptable chance of not finding a true difference. For example, for an
LLD that is based on a 95% confidence interval, we accept an error frequency of 5% in
concluding the existence of a source (in addition to the background activity) when none
exists.
Samples for which significant activity measurements were observed are identified in the
shaded boxes along with their 95% confidence interval. This commonly used confidence
interval identifies the range of values for the reported observation which accounts for random
variation in such a way that the probability that this range contains the true value is 95%. A
4-34
-------
Table 4-21. Radionuclide Analysis of Fuel ^Samples Reported by Contract Laboratory
Sample
l.D.
#1
#2
#3
#4
#3
#6
#7
#8
#9
#10
#11
#12
#13
#14
#13
#16
#17
#18
#19
#20
#21
#22
#23
Sample Activity (pCi/g) ± 2 Sigma Error
U-238 Decay Series
U-238
< 0.003
< 0.01
< 0.01
< 001
< 0.006
< 0.01
< 0.01
< 0.01
< 0.007
< 0.02
< 0.02
< 0.01
< 0.007
< 0.02
< 0.008
< 0.01
< 0.003
< 0.01
< 0.01
< 0.01
< 0.01
< 0.01
< 0.01
U-234
< 0.02
< 0.02
< 0.03
< 0.04
< 0.02
< 0.02
< 0.01
< 002
< 0.02
< 0.03
< 0.03
< 0.01
< 0.02
< 0.01
< 0.01
< 0.02
< 0.008
< 0.02
< 0006
< 0.01
0.002 * '0.00ft
< 0.01
< 002
Th-230
< 0.04
< 0.03
< 0.033
< 0.02
< 0.06
< 0.02
-------
Table 4-
ntinued)
Sample
I.D.
#24
#25
#26
#27
#28
#29
#30
#31
#32
#33
#34
#35
#36
#37
#38
#39
#40
#41
#42
Sample Activity (pCi/g) ± 2 Sigma Error
U-238 Decay Series
U-238
< 0.01
< 0.008
< 0.007
$.02 ±0.11
< 0.01
< 0.007
< 0.01
< 0.01
0,006 * 0.009
< 0.01
< 0.01
< 0.008
< 0.01
< 0.01
< 0.006
< 0.006
< 0.01
< 0.01
< 0.02
U-234
< 0.02
< 0.01
< 0.007
< 0.02
< 0.008
< 0.007
< 0.03
< 0.02
< 0.03
< 0.02
< 0.02
< 0.02
< 0.01
< 0.02
< 0.01
< 0.006
< 0.02
< 0.006
<0.03
Th-230
0.01 ± 0.01
< 0.02
0,00? £ 0.00*
0.08 * 0.03
< 0.009
< 0.02
< 0.02
0.007* 0.008 '
WHtf*' 0,00*'
,Q.O|'*' 0.009
O.01 ± 041
0.00? Jfe'0.01
< 0.01
fed* ±0,01
< 0.02
'Ob**ftDi '
< 0.01
JMHiMt
< 0.02
Ra-226
< 0.03
< 0.02
0,01 * 041
0.02 ± 0,02
< 0.01
< 0.02
< 0.02
< 0.03
0.000$ ±0,007
< 0.02
< 0.03
0.01 * 0.01
- 0.006 ±0.01
0.01*0.01
'o.oofri&oi'
:04>0| * 0X»*'
04)05*0.009
< 0.02
< 0.03
Pb-210
0.09 ± 0.06
om * o.o?
< 0.02
< 0.01
0,04 * 0,06
0,01*0.08
< 0.02
9.20 * 0JO
< 0.01
< 001
< 0.01
< 0.01
0.02 ± 0.07
< 0.01
0.03 * 0.07
< 0.01
< 0.01
0.04*0.08
< 0.01
Th-232 Decay Series
Th-232
0,002 ±'0.004
< 0.01
< 0.01
< 0.01
< 0.009
< 0.01
<0.02
< 0.005
< 0.01
< 0008
< 0.01
0.007 ± 0.009
< 0.005
< 0.007
< 0.02
< 0.03
-------
wide confidence interval (relative to the reported value) indicates a low degree of precision.
An inspection of Table 4-21 reveals that for values contained in the lighter shaded boxes the
lower range of reported values approaches or includes zero activity. Samples for which the
lower end of the 95% confidence interval identifies a value greater than zero are identified in
shaded boxes in bold print.
The magnitude of reported LLD values by the contract laboratory were principally dictated
by the three-hour sample counting duration. For enhanced sensitivities and significantly
improved (i.e., lower) LLD values, sample counting durations would have been required that
are prohibitive to commercial laboratories.
4.3.5 Quality Assurance and Composite Sample Results
The contract laboratory used to analyze the fuel-oil samples is an accredited laboratory. Its
QA/QC standards include the employment of Federally-approved analytical protocols,
National Institute of Standards and Technology (NIST) traceable standards for instrument
calibration, NIST traceable standards for spike samples, and biennial performance evaluation
for National Voluntary Laboratory Accreditation Program (NVLAP) accreditation.
In addition to the QA/QC standards under which the contract laboratory operates, the EPA
submitted aliquots of the composite sample for independent analysis to the National Institute
of Standards and Technology and EPA's National Air and Radiation Environmental
Laboratory (NAREL).
NAREL was instructed to use equivalent radiochemical protocols used by the contract
laboratory. A major difference between the contract laboratory and NAREL, however, was
the sample count time, which has a profound effect on the lower limits of detection. Due to
time and cost considerations, all samples, inclusive of the composite sample, were counted
for 170 minutes by the contract laboratory. NAREL extended the counting times for the
composite sample analytes to no less than 1000 minutes.
NIST, under interagency contract, analyzed the composite sample by non-radiometric
method(s) (i.e., isotope dilution—inductive coupled plasma spectrometry and mass
spectrometry). State-of-the-art non-radiometric chemical methods are more accurate and
precise due to the fact that they are not subject to the random nature of radioactive decay and
4-37
-------
the interference created by background radioactivity. However, due to the high cost, time
required for analysis, and other limitations, chemical analyses were limited to elemental
uranium and thorium.
Lastly, the contract laboratory was requested to reanalyze the composite sample. Its revised
protocol substituted acid digestion for low temperature ashing in order to avoid the potential
loss of volatile radionuclides. Among the most volatile are radionuclides of lead.
Table 4-22 summarizes values for the composite sample (representing samples 1 through 30),
as reported by the three laboratories. Given the low concentrations of analytes, the data are
highly consistent. All values reported as "less than the specified lower limits of detection"
by the contract laboratory comply with the more definitive measurements provided by
NAREL and NIST. In most instances, LLD values reported by the contract laboratory were
one to two orders of magnitude higher than measured values reported by NAREL and/or
NIST.
Table 4-22. Comparison of Composite Sample Analysis Reported by Contract
Laboratory, NAREL, and NIST
Radionuclide
U-238
U-234
Th-230
Ra-226
Pb-210
Th-232
Ra-228
Th-228
U-235
Contract Laboratory
(pCi/g ± 2SD)
< 0.008*
< 0.02'
< 0.02'
0.04 ± 0.03
0.65 ± 0.50
< 0.02*
< o.or
0.10 ± 0.04
< 0.01'
NAREL
(pCi/g ± 2SD)
0.0005 ± 0.0009
0.0011 ±0.0013
0.0055 ± 0.0013
0.0024 ± 0.0001
0.605 ± 0.537
0.0005 ± 0.0004
0.0070 ± 0.0042
0.009 ± 0.002
0.0001 ± 0.0004
NIST
(pCi/g ± I SD)
0.00055 ± 0.00006
0.00056 ± 0.00003
' These values represent LLD values.
U-238 and Th-232 values reported by NAREL are virtually identical with those reported by
NIST. The high degree of agreement provides reasonable assurance that activity
concentrations for other radionuclides reported by NAREL are credible.
4-38
-------
4.3.6 Interpretation and Use of Experimental Data
The initial design and intent of this study was to assess the radionuclide concentrations for as
many samples as utilities could provide within a timely fashion. From the analysis of
individual samples, radionuclide-specific average values and their distribution were to have
been determined. The independent analysis of a composite sample by (1) the contract
laboratory (i.e., TMA Eberline), (2) NAREL, and (3) NIST was primarily intended to
provide an "internal" and "external" quality control check for the 42 individual samples
analyzed exclusively analyzed by the contract laboratory. The intended approach for
deriving radionuclide-specific activity concentrations for the 42 samples (Table 4-21),
however, was made difficult by the fact that, for a majority of measurements, the activity
was reported as "less than a specified lower limit of detection" (i.e., < specified LLD).
There are several options for the interpretation and assimilation of such pseudo-quantitative
measurements:
• Option #1 - A highly conservative, but questionable, treatment of such
measurements assumes that the concentration is just below the LLD and may,
therefore, be approximated by the LLD.
• Option #2 - A highly non-conservative. but equally questionable, approach is
to assume that those measurements which fail statistical significance indicate a
complete absence of the radionuclide.
• Option #3 - A more reasonable, but nevertheless unsupported, treatment of the
data assumes that, on the average, the actual concentrations lie midway
between zero and the reported TJ.n values.
• Option #4 - The most defensible approach is to (1) substitute the composite
sample data reported by NAREL (and confirmed by NIST) for samples 1
through 30 and (2) integrate the composite sample data with the remaining
individual measurements for samples 31 through 42 by appropriate weighting.
All measurements reported as "less than a specified lower limit of detection"
for samples 31 through 42 (except for Pb-210) are treated as if the actual .
values lie midway between zero and the reported un values. For Pb-210,
only the composite sample value reported by NAREL (and confirmed by
contract laboratory reanalysis) has credibility.
4-39
-------
4.3.7 The Need for Extrapolation of Data for Radionuclides Not Measured
The U-238 and Th-232 decay-chain series contain 13 and 9 major radioactive decay products,
respectively (see Tables 2-13 and 2-14). In the present study, the radiochemical analyses
provided data for only select members of these decay chains. Nevertheless, information
obtained in behalf of these select radionuclides can be extrapolated to radionuclides not
measured based on the physical and chemical relationships as described in NCRP Report No.
94 (NCRP 1987b):
For non-equilibrium conditions, the U-238 series can be separated into a primary group and
four sub-series because the intermediate decay products have different chemical properties
and have half-lives long enough to permit significant separation from their precursors. In
each of the following sub-series, the activities of the later members tend to follow the activity
of the first member.
• U-238. Th-234. Pa-234. and U-234: The first three decay-chain members may
be assumed to be in secular equilibrium. Thereafter, U-234 represents the
first sub-series. The 2.5 x 10*-year half-life of U-234 allows time for an
apparent isotopic fractionation of uranium from alpha-emission recoil of
Th-234 and chemical processes in the Th-234, Pa-234, U-234 chain.
Uranium-234 may be removed from earth materials at a greater rate than
U-238, and builds up an excess activity over U-238 in groundwater (Osmond
and Cowan 1976). Rosholt et al. (1966) have measured uranium-series
disequilibria in some soils and have found U-234/U-238 activity ratios as low
as 0.58. Data of Smith and Jackson (1969) indicate activity ratios of 0.914 to
0.985 in natural uranium from 16 widely distributed sources.
• Second Sub-Series - Th-230: Th-230, with a half-life of 8.0 x 10* years and
with the chemical inertness characteristic of tetravalent thorium compounds,
tends not to migrate with either of its uranium isotope precursors. Sorption is
considered to be the dominant means of thorium immobilization (Langmuir and
Herman 1980). Among the samples measured by Rosholt et al. (1966),
disequilibrium between Th-230 and either U-238 or U-234 ranged up to a
factor of about two, as either an excess or a deficiency. Th-230 is strongly
depleted from its precursors in sea water and enhanced in bottom sediments.
• Third Sub-Series - Ra-226. Rn-222. Po-218. Pb-214. Bi-214. and Po-214:
Ra-226, which is frequently separated from its precursors, is rarely found in
such mass concentration as to precipitate in the presence of anions for which it
has a strong affinity, particularly sulfate. Once released into natural waters by
4-40
-------
radioactive recoil or by breakdown of entrapping solids, Ra-226 and radium
isotopes of the other series are mobile until scavenged or coprecipitated with
major reactants in the water. Under conditions in which the radon gas is not
removed, Ra-226 canJx assumed to exist in secular equilibrium with radon
and its short-lived decay products.
Radon-222, a noble gas with a 3.8-day half-life, includes among its decay
products Po-218, Pb-214, Bi-214, and Po-214, which have half-lives ranging
from 26.8 minutes to 164 microseconds. The gamma rays of Bi-214 have the
highest yield and are the most energetic of the uranium series, which makes
this sub-series important with respect to external radiation. Moreover, about
45 percent of the alpha energy of the uranium series is accounted for by this
sub-series, making it important with respect to internal emitters.
• Fourth Sub-Series - Pb-210. Bi-210. and Po-21Q: The longer-lived radon
decay products, Pb-210, Bi-210, and Po-210, make up the final sub-series.
The last decay product is stable Pb-206.
The thorium series is characterized by the long-lived Th-232 at the head of the series, and
decay products that are relatively short-lived. If no migration of the series members takes
place, radioactive equilibrium may be assumed. In soils, natural waters, natural gas, and
petroleum, the disparate chemical or physical properties of the series members may,
however, cause disequilibrium. For conditions that define this study, the thorium series may
be considered to exist as two sub-series:
• Thorium-232 itself, which exists naturally in the tetravalent state as a very
stable oxide or in relatively inert silicate minerals. It is strongly adsorbed on
silicates (Langmuir 1980). Thorium-232 is the least mobile of the series
radionuclides.
• The sequence Ra-228, Ac-228, Th-228, and Ra-224, whose relative
equilibrium is determined by processes such as radioactive recoil, adsorption,
and changes in carrier compounds with which the radionuclides become
associated. Ra-224 can be assumed to be in secular equilibrium with the inert
noble gas Rn-220 and its decay products down to stable Pb-208.
4.3.8 Summary
Table 4-23 provides the estimates for the average radionuclide concentrations in the 42
residual fuel oil samples evaluated in this study. These values were obtained by employing
4-41
-------
Table 4-23. Estimates of Average Radionuclide Concentrations
in 42 Residual Fuel Oil Samples
U-238_Series
U-238
Th-234
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232 Series
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Po-212
Concentration (pCi/g)
0.0018
0.0018
0.0018
0.0034
0.0068
0.0043
0.0043
0.0043
0.0043
0.0043
0.0043
0.44
0.44
0.44
Concentration (pCi/g)
0.0030
0.068
0.068
0.068
0.068
0.068
0.068
0.068
0.068
the methodology defined by Option #4 and assigning the derived value of the precursor
radionuclide of each subseries to corresponding radionuclides that were not assessed. Values
in Table 4-23 are well within the range of previously reported study data summarized in
Section 4.2 and provide additional support for the conclusion that the radionuclide content of
residual fuel oil is significantly lower than that of coal.
In Chapter 6 of this report, the assessment of radionuclide emissions from oil-fired SGUs
assumes radionuclide concentrations contained in Table 4-23.
4-42
-------
5. BOILER DESIGN FEATURES THAT AFFECT RADIONUCUDE EMISSIONS
5.1 GENERAL DESCRIPTION
Radionuclide emissions from electric utility units are directly affected by the fossil fuel
composition, boiler design and operation, and flue gas treatment. Elements common to fossil
fuel-fired plants include a boiler, steam turbine, generator, water cooling mechanism, fuel
preparation/handling equipment, and assorted auxiliary systems. System components that are
fuel-specific include furnace design, the ash handling systems, and flue gas emission
controls. .Thus design features and operating parameters of a coal furnace are considerably
different from those using oil or natural gas. The relatively high ash content of coal requires
specific systems for handling ash that remains in the furnace or gets entrained in the
combustion gases. This chapter provides basic information regarding boiler design features
that have a direct bearing on the emissions of radionuclide particulates and radon-222 gas.
In previous chapters, the radionuclide content of coal, gas, and oil were discussed. Upon
combustion of coal or oil, U-238, Th-232, and their respective decay-chain products attach to
particulates associated with ash or ash residue. The radionuclide content of fly ash is,
therefore, the source term for emissions. For natural gas-fired plants, the principal
radionuclide released during combustion is radon-222 gas.
Power plants are designed and operated to serve three load classes. Terms commonly used in
discussions of plant designs and operation include the following:
• Base-load plants operate near full capacity most of the time (or are dispatched
to operate in the most efficient region of the heat rate curve).
• Intermediate-load (or cycling) plants operate at varying levels of capacity each
day (about 40 percent utilization on an average annual basis).
• Peaking plants operate only a few hours per day (about 700-800 hours per
year).
• Capacity factors (often referred to as "capacity") are the ratio of energy
produced in a given period to the energy that would have been produced in the
same period had the unit been operated continuously at its rated power. The
unit of time is generally a one-year period.
5-1
-------
• Availability refers to the fraction of a year during which a unit is capable of
providing electricity to the utility grid at its rated power after planned and
forced outages have been accounted for.
• Nameplate capacity is the amount of energy a unit is capable of producing if
run continuously.
• Capability is the percentage of naraeplate capacity that is needed to meet an
average seasonal demand; this term is beginning to replace "capacity factor" as
a hallmark of plant operation.
Fossil-fueled steam-electric generating plants dominate base-load and intermediate-load
service. Coal is rarely the primary fuel for a peaking plant. Historically, new units are used
for base-load generation and cycling capacity is obtained by downgrading the older, less
efficient base-load equipment as more replacement capacity comes on line.
In 1990, the average capacity factor for coal-fired units operating in the base-load mode was
65%, while units operating in a cycling mode averaged about 40% (DOE 199la). The
availability of a coal-fired unit generally declines with increasing generating capacity.
Generating units with capacities of less than 400 MWe have average availabilities of more
than 85%; those with capacities of more than 500 MWe have only 74% to 76% availabilities.
5.2 BASIC POWER PLANT COMPONENTS
Fossil-fuel-fired boilers use a furnace in which the combustion zone includes a feeder, air
injector, and firing apparatus that are often combined in a single mechanism. Coal-fired
boilers are fired by several types of furnaces including cyclone, fluidized bed, pulverized
coal, and stoker designs. Natural gas-fired and oil-fired boilers commonly use injection
burners similar to pulverized coal boilers. For pulverized coal boilers, the burners may be
directed vertically, horizontally, in opposition, or tangentially. These features are all
designed for efficient combustion of specific coals, handling of bottom ash, removal of fly
ash. and the treatment of stack gases. Each furnace type includes specialized auxiliary
equipment for handling the coal and ash products. Heat exchange components downstream
of the furnace include a superheater, economizer, and air preheater.
A superheater is a system of tubes at the top of the boiler through which saturated steam is
superheated by combustion gases under high pressure (as high as 1,200 psi or more). The
5-2
-------
heat exchange is highly efficient, with the flue gas temperature dropping as much as
1,000° F.
An economizer is an additional system of tubes located after the superheater which heats
boiler feedwater.
An air preheater extracts heat from flue gases leaving a boiler and transfers the heat to the
combustion air being fed into the furnace. Heated air improves fuel combustion, dries the
powdered coal particles when a pulverizer is used, and lowers the viscosity of oil for
improved flowability.
Heated gases exiting the power plant are conveyed to a height that promotes dispersion in the
atmosphere. Stacks are sized by considering the gas velocity in feet per second and the
pressure drop caused by friction of the gas stream against the inside the stack. For
maximum dispersion, the stack gas pressure at the exit end of the stack is high enough to
push the gas well above the stack. Each stack design (diameter and height) considers plant
elevation, gas weight, gas density inside the stack, and gas density outside the stack. Further
considerations are given to seasonal weather temperatures (for a stack gas temperature of
300° F, the ambient temperature may vary from -10° to 110° F) and atmospheric pressure
variations (a 3-inch mercury variation in a barometer has the equivalent effect of a change in
2,500 feet of elevation).
In addition to these components, an electrical power plant consists of steam separators, fans,
pumps, fuel handling equipment, and combustion byproduct handling equipment.
5.3 COAL-FIRED BOILERS
5.3.1 Pulverized Coal Boilers
Cleaned coal is broken, crushed, and pulverized prior to burning in the furnace. The
pulverizer is located adjacent to the boiler unit and often combines powdered coal with
heated air from the air preheater. The pulverized coal and heated air are injected into the
furnace by a forced air fan.
5-3
-------
Pulverized coal panicles burn in a three-phase process of heating to ignition (incandescence
to ignition), distillation and combustion of the volatile matter, and combustion of the coke
residue while suspended in the boiler, atmosphere. Ignition occurs within a few hundredths
of a second with the total three phase process occurring in approximately one second. The
primary variables during combustion are panicle size, coal type, boiler temperature, and
excess air flow. The burners are designed to: (1) burn coal of varying quality, (2)
thoroughly mix coal and air, (3) minimize boiler pulsations with stable ignition, (4)
completely burn the coal panicles within the furnace, (5) distribute the temperature and
composition of the exiting gases uniformly, and (6) minimize scaling and fouling of boiler
interior surfaces. Pulverized coal boilers commonly use tangential firing, though boilers with
horizontal and vertical firing mechanisms are still in use.
Tangential Firing. Tangential firing (also termed corner firing) has burners located in each
of four corners of the boiler and close to the boiler floor. Burner assemblies are constructed
with a four-tier burner nozzle arrangement which includes coal nozzles, secondary air
nozzles, overfire air nozzles, oil gun for initial lighting, air dampers, and a wind box.
Flame temperature may be raised or lowered with.adjustment of the furnace mechanism.
In tangential firing, the burner nozzles are pointed so that streams of coal and air are
projected along a line tangent to a small circle, lying in a horizontal plane, at the center of
the boiler. Intensive turbulence and mixing occur where these streams meet. The turbulence
in the center of the furnace is sufficient to combine all coal particles and air. A flame front
propagating along each feeder stream combines with the other flame fronts and forms a
rotating flame body similar to a cyclone flow pattern. Air is injected through an annulus
surrounding each burner and through overfire air nozzles near the upper level of the
combustion zone. Coal particles and air come into full contact for extremely rapid and
thorough combustion. Combustion rates may exceed 35,000 Btu per cubic foot per hour.
Combustion temperatures to 3,400° F are possible. The combustion zone can be adjusted by
pointing the burner nozzles up or down with the aid of electric motors or hydraulic cylinders.
The flame body is controlled and can be either localized and intensified or spread over a
greater wall surface. Vertical adjustment on the burners can raise and lower the combustion
zone about twelve feet, as shown in Figure 5-1.
The nozzles may be pointed upward either during light load periods or when the inner boiler
walls are clean. The nozzles can be directed downward for heavy loads and when the walls
5-4
-------
Figure 5-1. Flame Adjustment in a Pulverized Coal Boiler
-------
are coated with dust or scale. The burner adjustment directly affects gas temperature at the
boiler outlet, as shown in Table 5-1.
•»
Table 5-1. Effect of Burner Adjustment on Gas Temperature
Burner Tilt
(degrees)
+ 20 up
+ 10 up
0
- 10 down
- 20 down
fan Tmnnum
\Ma I CUUJJClal
urc
Leaving Boiler
(°F)
2.210
2.160
2.120
2.090
2.075
Source: de Lorenzi 1948
The air dampers provide further control of the outlet gas temperature. A temperature range
of 160° F may be achieved by adjusting the dampers from tally closed to fully opened.
Tangentially-fired boilers are relatively clean above the combustion 'zone as carbon burnoff is
essentially complete due to fuel-rich combustion promoted by the overfire air nozzles.
Formation of nitrogen-oxygen compounds is also suppressed.
Horizontal Firing. Burners are usually placed in the front or rear furnace wall for horizontal
firing. Coal and air are distributed uniformly into the boiler combustion zone through a
circular nozzle surrounded by a housing with adjustable vanes. The nozzle is rifled and
conically shaped toward the boiler end. A centrally located tube provides an opening in
which an ignition torch can be inserted or where an oil burner nozzle is located. An
adjustable deflector is located in the burner inlet, to obtain equal fuel distribution around the
periphery of the nozzle inlet. Coal and air enter the nozzle tangentially at the rear of the
deflector. Air is injected under high pressure. Coal remains uniformly distributed by the
rifling through the main body of the nozzle. Coal and air are spun by radially adjustable
vanes, with high turbulence resulting. The flame shape in the combustion zone can be
controlled by adjusting the angularity of the air vanes. Either a short or long flame can be
produced. Opposition firing is a type of horizontal firing. Instead of a single set of burners,
burners are located in opposite walls of the boiler. A zone of intense turbulence is set up at
the point where the flames impinge. Other fuels such as natural gas and oil are fired using
horizontal burners.
5-6
-------
Vertical Firine. Burners may be placed in the furnace arch and fired vertically downward.
The fuel nozzle is housed in a box-like structure, which include controllable dampers for
regulating tertiary air injected under pressure. Primary air enters with the pulverized coal
and is used to extend the length of the flame body deeper into the combustion space.
Secondary air is injected through front wall ports along the path of die flame propagation.
Vertical firing provides a uniform distribution of temperature throughout the furnace. The
fuel is in the furnace for a long time so that combustion is complete. Unbumed fuel ranges
0.1 to 1.5% of the fuel supplied to the furnace. Coals low in volatility perform best in the
vertical burners. Natural gas and oil do not adapt easily to vertical firing.
Ash Collection. Bottom ash collection equipment in pulverized coal boilers is either of the
dry-bottom or wet-bottom type. A dry-bottom collector may use either hydraulic,
mechanical, or pneumatic equipment to remove the clinker or dry ash material. In a
hydraulic system, the dry material drops into the bottom ash bins and is removed by water
jets that suspend the ash. The bottom ash is pumped to a holding tank and either transferred
from the plant area in a slurry form or dewatered and removed as a solid. In a mechanical
system, the bottom ash drops from the collection bin through a transition chute to a water
tank. A drag-chain conveyor carries the bottom ash up an inclined chute and dumps the ash
onto a belt conveyor for discharge directly to a discharge area or to a temporary storage
facility. In a pneumatic system, hot ash is suspended in a vacuum and transported to a dry
storage area. Wet-bottom collection occurs as slag drops to the bottom of the boiler in a
molten pool. A slag drip opening is located at the hottest spot in the bottom center of the
boiler. A constant thickness of molten slag is maintained in the boiler by a water-cooled
ring, which forms a slag dam and drip ledge. The slag is kept molten by dropping the flame
in a tangentially-fired furnace to heat the slag above its ash fusion point.
5.3.2 Cvclone Boilers
The cyclone furnace operates similarly to that of a tangentially-fired pulverized coal furnace
and may be oriented either horizontally or vertically. However, larger sized coal panicles
than used in pulverized coal furnaces are fed near one end of a horizontal cylinder forming
the combustion chamber. Air is injected at high velocities tangentially along the cylinder
periphery. A spiraling outer vortex is formed which moves away from the constriction while
an inner spiraling core moves toward the constriction. Upon ignition, the flame body takes
the shape of a cyclone. Flue gases are carried through the constriction to the heat
5-7
-------
exchangers. The walls of the boiler downstream of the constriction are refractory-lined to
handle the slagging that usually occurs. Slag is removed in typical wet-bottom manner.
Advantages of a cyclone furnace arejhat.they are smaller than a pulverized furnace, produce
less fly ash in the flue gas, and require only partial sizing, saving a particle sizing step.
5.3.3 Stoker Boilers
Stoker firing is the oldest method of steam generation from the burning of coal.
Run-of-mine coal is broken and crushed to a lump-sized material and fed into a furnace by
either an overfeed, underfeed, or crossfeed method.
Lump coal is fed to the furnace to form a moving bed of 4 to 6 inches in depth over a grate.
Moisture and volatile matter are driven off when coal enters the furnace. Combustibles in
the flue gas support a flame body that hovers over the coal bed and is continually supplied by-
carbon released from the moving and ignited coal lumps.
In the overfeed method, coal is dumped onto either a stationary or continuously moving
grate. The tumbling coal tends to segregate by size, with fines being thrown to the surface
of the bed for rapid ignition and coarse particles for later burning.
In the underfeed method, coal wells up through a port in the bottom of the furnace and
spreads out over an inclined grate. Combustion air is forced with the coal and through the
grate.
In the crossfeed method, coal enters the furnace from the front and is dumped onto a moving
grate. Combustion air enters with the coal and upward through the grate. The bed tends to
be thinner than in overfeed stoker furnaces and has a greater air-to-surface ratio than in
either overfeed or underfeed stoker furnaces. The result is a higher combustion rate.
Stoker furnaces are not being used in new coal-fired electrical utility power plants and are
only found in a few older power plants still in service. Stoker furnaces are more commonly
used in older industrial boilers for small and isolated locations. Ash disposal is a concern in
all stoker boilers with excess fly ash and clinkers formed. Also, the boiler size limitations,
dictated by economics, do not allow stoker boilers to directly compete in heat energy output
with pulverized coal boilers.
5-8
-------
5.3.4 Fluidized Bed Boilers
At present, fluidized bed boilers are not a commercially prominent boiler type used for utility
w
steam generation. This boiler type was previously used and is currently being considered for
future use with a wide variety of fuels, including coal or combustible waste materials.
This boiler type employs a bed of granular, noncombustible materials such as coal ash or
lime layered atop a supporting grid plate to a depth of several feet. Heat transfer surfaces or
boiler tubes are embedded in the material bed. Fuel such as pulverized coal, residual oil, or
combustible waste is injected into an airstream passing upward through the grid plate and
through the bedded material. The granular particles in the bed are fluidized. The
combustion temperature is relatively low at about 1,500° F.
The fluidized bed boiler has two advantages: the ability to burn high-sulfur coal with low
sulfur dioxide output and high heat release, which reduces the boiler size. Particulates and
nitrogen-oxygen compounds are also largely reduced.
5.4 NATURAL GAS-FIRFJ) BOILERS
Many older natural gas boilers are similar to the pulverized coal design with a system of heat
exchangers and downstream handling equipment. Newer power plants generally use gas
turbines and are described as single cycle power plants.
Natural gas is delivered to a conventional, single cycle, or double cycle boiler by a pipeline.
The gas pipeline pressure may vary from 100 to 500 psi. Regulators are used to step down
the pressure for delivery to the burners at about 10 psi. High pressure gas burners include
the gasring type, center-diffusion type, turbine type, and tangential type.
Heat exchange by gas turbines represents the most common design in natural gas power
plants over the past 20 years and continues to grow in use as the price for raw gas continues
to be more economic than raw coal. The single cycle system includes a combustion
chamber, regenerator, compressor, turbine, and generator.
Compressed air is heated in the regenerator by exhausting flue gas and fed into the
combustion chamber. Natural gas is often injected tangentially through the side wall of the
5-9
-------
combustion chamber, forming a cyclone flame body. Hot flue gas passes from the
combustion chamber to the turbine, which drives the generator and produces electricity.
Flue gas leaving the turbine is directed to the regenerator and heats the incoming air. The
exhaust gas is released through a stack directly to the atmosphere.
Compressed air preheated in the regenerator may reach 1,000° F to produce a flame
temperature of 3,400° F. Burning is rapid and complete, producing only small amounts of
fly ash.
5.5 OIL-FIRED BOILERS
Oil-fired boilers are similar to pulverized coal boilers. The boiler includes an air preheater,
furnace, and superheater. In addition to the boiler unit, plant equipment includes a large
complement of handling apparatus, among which are: storage tanks, strainers for removing
dirt, pumps, oil heaters to raise the oil to the proper temperature for pumping and
atomization (about 150° F), piping, and burners.
Heated oil is atomized and injected through nozzles into the furnace in a mist-like spray with
heated air or steam and mixed with combustion air. The fuel oil injection rate and airflow
are held in delicate balance to burn efficiently. The oil mist must be completely consumed to
avoid drooling, fouling, and clogging. Air-to-fuel mixture must be high enough to
completely burn the oil but low enough to provide for maximum heat release. Similar to
burning other fossil fuels, excess air is injected into the furnace to insure combustion.
Generally, 10% to 30% over critical mixture is used.
5.6 FOSSIL FUEL COMBUSTION
Pulverized coal, natural gas, and oil-fuel combustion involves the optimum mix of air and
fuel to produce the most heat energy, while minimizing the effects of deleterious byproducts.
When mixed with air and injected into a furnace, fossil fuels heat to ignition and subsequent
combustion almost instantaneously. Ignition rate is regulated by the amount of volatile
matter in the fuel. A midwestern coal with 34% volatile matter ignites at 27.5 milliseconds,
while an eastern coal with 15% volatile matter ignites at 38 milliseconds (Wilbur 1982). The
coal's temperature rapidly rises from the moment it enters the boiler to the point of complete
combustion. The temperature of pulverized coal rises from 700 to 2,800° F, of natural
5-10
-------
gas—500 to 3,800° F (Haslam 1926), and of oil—150 to 2,900° F. Flame temperatures may
vary by 300 to 500° F (Haslam 1926) due to the amount of excess air provided, fuel
characteristics, and moisture content^
Excess air may be added either with the fuel or through overfire air nozzles, usually within a
range of 10% to 30% (Haslam 1926) above the minimum air-to-coal concentrations needed to
support complete combustion. Excess air provides additional oxygen to the flame within the
boiler, which lowers the temperature of the flame but provides oxygen to aid in maintaining
an oxidizing environment within the furnace (when oxygen is depleted, carbon dioxide
produced during combustion is reduced to carbon monoxide, which increases the slagging
potential of ash). Increased air pressure, independent of preheated fuel temperature, lowers
the ignition temperature of gaseous and liquid fuels and enhances combustion. Increasing the
air pressure for oil injection from 40 to 385 pounds per square inch lowers the ignition
temperature from 788 to 392° F.
Variations in ignition and combustion rates/temperatures directly affect the amount of heat,
the character and amount of the ash (which also determines the amount of radioactivity in
flue gas), and the amount of flue gas produced hi the boiler. Not all the fuel is consumed.
As much as 0.1% to 1.5% of the fuel entering the furnace is carried away with the ash (Kent
1950).
5.7 TEMPERATURE PROFILE
The temperature profile through the boiler system varies widely among the fuels. The
heating value, moisture and ash content, and volatile matter affect the furnace temperature
and temperature of downstream components. Flame temperatures are often hotter for natural
gas and oil than coal. This implies that coal boilers are larger and burn more fuel of lower
heating value to produce an equivalent amount of heat. The temperature declines rapidly
away from the flame except in the rising flue-gas column. The flue gas yields most of its
heat as it passes through the heat exchange equipment. In pulverized'coal boilers, the flue'
gas may lose more than 1,000° F just passing through the superheater. Flue gas retains
much of its heat well through the system due to the short amount of time the gas is in contact
with boiler surfaces. The residence time of the flue gas in the boiler stream is estimated to
be 4 seconds through the boiler, economizer, and air heater; 5 seconds through the
paniculate collector; and 8 seconds through the stack (Torrey 1978).
5-11
-------
In Figure 5-2, a typical temperature profile is shown for a pulverized coal boiler and
ancillary equipment for downstream gas flow. Though initial flue gas temperatures may be
plus or minus several hundred degrees from the average, heat loss is very high over the
overall system so that exhausting gas is usually in the 250 to 350° F range at the stack.
Combustion temperatures determine the degree to which radionuclide species are volatilized.
The volatilization of specific metallic species of the U-238 and Th-232 decay chains results
in an enrichment of these radionuclides in fly ash. The significance of fly-ash enrichment of
specific radionuclides is discussed in greater detail below.
5.8 ASH FORMATION
5.8.1 Ash Characteristics
Ash is created as a byproduct of combustion and is produced in varying quantities by all
three fossil fuels. Coal ash is composed almost entirely of metal oxides that have a wide
range of variability in composition and concentration. Included are 40% to 60% silicon
oxide, 20% to 60% aluminum oxide, 5% to 25% ferric oxide, 1% to 15% calcium oxide
plus small amounts of oxides of magnesium, sodium, and potassium. The concentration of
these compounds is important in the type of ash that is formed.
Combustion produces four types of ash forms: (1) fly ash, (2) molten or partially fused ash
entrained in the flue gas that may stick to high temperature surfaces (slagging), (3)
volatilized material that may condense as hard deposits at cooler temperatures (fouling), and
(4) bottom ash that may be hard and dry or soft and viscous. Coal produces the greatest
amount of ash while natural gas produces the least. Oil produces small amounts of ash and a
viscous, sticky fluid if burning is incomplete.
During coal combustion, an oxidizing atmosphere is predominant, which produces a
bituminous type of ash that characteristically contains more ferric oxide than lime and
magnesia (Rupinskas 1991). The ash is dry with low adhesion potential to boiler surfaces.
Gaseous atmospheres in the combustion zone may develop high levels of carbon monoxide
(presumably from roasting carbonates in the lime contained in a coal). Carbon dioxide, a
normal combustion product, may be reduced to carbon monoxide. A reducing atmosphere
can sometimes create a strong tendency for ash to become soft, sticky, and viscous, forming
5-12
-------
Forced Air
conomize:
775°F
Superheate
2200°F
AAAAAAAA
• • • :
.
•
Flue Gas
2500°F
Furnace
Combustion
Zone
2200°F
Powdered
and Air
Ash Hopper
Coal
J_Pulverizer
Electrostatic Precipitator
(Bag House may be used)
350°F
Air
Preheater
775°F
700°F
Scrubber
300°F
Stack
Ash Hopper
BOILER COMPONENTS
Pulverizer
Furnace
Superheater
Economizer
Air Preheater
Electrostatic Precipitator
Sulfur Dioxide Scrubber
Stack
Ash Hopper
Figure 5-2. Temperature Profile through a Pulverized Coal System
-------
a lignite type of ash. This type of ash contains more lime and magnesia than ferric oxide.
Ash, under high temperature, may remain molten. Eastern and midwestern coals generally
tend to produce bituminous type ash while western coals produce lignite type ash (Rupinskas
1991). Table 5-2 cites the distribution of fly ash to bottom ash, but does not account for
boiler slagging and fouling buildup.
Table 5-2. Coal Ash Distribution by Boiler Type
Furnace Type
Pulverized Coal:
Dry Bottom
Wet Bottom
Cyclone
Stoker
Percent Fly Ash/Percent Bottom Asb
Bituminous
80/20
65/35
13/87
60/40
Lignite
35/65
30/70
35/65
Source: Mead 1986
5.8.2 Slagging and Fouling
Slagging includes all fused deposits or resolidified molten materials that form primarily on
furnace walls and surfaces exposed to high gas temperatures. Fouling includes the buildup
on lower temperature regions of the boiler. The tendency for slagging and fouling is largely
a function of the prevailing combustion environment (oxidizing or reducing). In a reducing
combustion environment, the melting temperature of ash can be lowered by 390° F below
that of an oxidizing environment (Lawn 1987).
The potential for slagging is also based on chemical composition of coal, critical
temperatures, and ash viscosity versus ash fusion. The inner boiler walls, superheater, and
furnace bottom are typically coated with an oxide scale onto which an ash buildup occurs.
Most slags chill to a solid state at about 1,800° F but become soft and pasty above 2,000° F
(Haslam 1926, Shannon 1978). If a reducing condition occurs or an ash has a strong
tendency to slag due to chemical composition, the ash will collect on these surfaces,
increasing the thermal gradient and the buildup of slag. Particles larger than 100 microns
may be 930° F hotter than the flue gas, which may be 2,500 to 2,200° F. These large
5-14
-------
panicles tend to be sticky. Particles less than 30 microns tend to flow around rather than
impinge on boiler surfaces (Lawn 1987). Varying amounts of adhesion occur for particles
between 30 and 100 microns. The typical amount of slagging occurring in boilers'is
unavailable, but large chunks of slag several feet long and one to three feet thick are known
to break off from the walls and fall to the boiler floor, causing damage.
Fouling is related to the alkaline content of the ash in which sodium vapor in the combustion
gases promotes contact and bonding of fly-ash panicles to metal surfaces, particularly in the
boiler economizer. Ash buildups, either through slagging or fouling, occur throughout the
boiler system and are periodically removed and disposed with other precipitated ash.
Slagging and fouling represent only a minor percentage of the total ash and may be ignored
with regard to paniculate emissions. Paniculates that adhere to boiler surfaces are generally
stable and do not reenter the flue gas stream.
5.8.3 Bottom Ash
Bottom ash is the product formed in the furnace and caught in a bottom hopper. Bottom ash
may be classified as either dry or wet. Stoker boilers produce dry-bottom ash, as do most
pulverized coal boilers. In pulverized coal boilers which use coal with low potential for
slagging, a dry-bottom collection design is used. Most ash can be kept solidified by boiler
operation unless coals with a high to severe slagging potential are used. This is avoided for
dry-bottom boilers by carefully specifying favorable coal characteristics. Currently, most
coal-fired boilers use a dry-bottom collection system, as shown in Table 5-3.
Table 5-3. Distribution of Bottom Ash Collection by Coal Types
Coal Type
Low Sulfur Bituminous
High Sulfur Bituminous
Low Sulfur Sub-bituminous
High Sulfur Sub-bituminous
Low Sulfur Lignite
Bottom Ash Collection
Dry Bottom
695
18
160
-
35
Wet Bottom
90
18
22
1
10
Source: UDI 1992
5-15
-------
Slag-tab or wet-bottom boiler hearths are used for coals that become molten at relatively low
temperatures. In such boilers, it is common for the temperature in the furnace to be lower
than in dry-bottom boilers as a coal like lignite with lower heating value and greater ash
.K-
content is used.
Slagging on upper boiler surfaces and in the furnace bottom are similar processes. Chemical
elements existing in the coal form compounds in ash that may tend to slag. With proximity
to the flame, the temperature is generally high enough to keep such coal in a molten
condition. In most cyclone boilers and pulverized coal boilers that use coal known to have
strong slagging potential, a wet-bottom design is used. The ash fusion point and furnace
temperature determine the degree of slagging that will occur. In wet-bottom boilers, ash
may require additional beating to prevent solidification.
5.8.4 Flv Ash
The chemical composition and physical characteristics of fly ash are largely determined by
the fuel's content of inorganic materials. Coal and, to a lesser extent, fuel oil contains
varying concentrations of minerals, many of which exhibit melting and even boiling points
that fall within the range of furnace temperatures. During combustion, these minerals can,
therefore, be melted, deformed, and vaporized. As ambient temperature drops, solidification
occurs. For example, silicates typically form glassy spheres, while iron compounds form
jagged particulates. As much as 80% of total ash thus formed may be entrained in flue gas
as fly ash.
An important characteristic of fly ash that impacts plant emission is its panicle size
distribution.
Figure 5-3 identifies representative particle-size distribution curves for four different coals
fired in utility boilers at the inlet to a control device (Severson 1978). Typically, only about
two percent of the mass is smaller than one micron in diameter. However, it is the smaller
particles which are the more difficult to remove by emission control systems.
A related impact of particle size on plant emission involves the phenomenon of enrichment
by volatile trace elements. As the flue gas cools, smaller panicles may preferentially serve
.as condensation nuclei. Under conditions of surface condensation, smaller panicles are also
5-16
-------
09 0
98
CC 91
III
_l
< 90
s
i "
s '«
or
UJ SO
Q.
H SO
I
0 40
•a
30
•
'> 70
H
D 10
§ •
2
1
OS
0.3
0.1
0.05
0.01
.1
•
^
X
— W
S
.._
row
---"
/
._.
--
--
—
IN
. —
•7
~?
...
~~
3
,--•
x->
X
-
-
-
QC
^^
-
"1
'-
AL
I
l_ ;
1
r
-i-
i
|
-<
T
i
...
-
^
-'
-
...
^
-
-
^
x
1
-
,— *
X
1
I
I
1
1
— • • ' 1 •
t
— J--
i
JORTII DAK
I
,
^
-^—— -^
£
^
— ™
or;
•
AL
i L
^
^
A&
1C
\n
-
N:
X*
X
Ik
— 1
'T
X
c
r~
E
P
Df
..
X
>
,£
a
-
—
-
>
F— -"
-
-
X
X1
--*
...
-
X*
X"
**"
-
-
^
^*
—
• • •
..
\S
-------
subject to enrichment due to the enhanced surface area to volume ratio. Thus, smaller
particles that are less efficiently removed by emission control systems are also enriched in
volatile trace elements that include certain radionuclides. Removal efficiencies by various
emission control systems and radionuclide-specific enrichment factors are discussed in
Sections 5.9 and 5.10 below.
5.9 EMISSION CONTROL SYSTEMS
Emission control systems are installed in electric utility plants to remove particulates from
flue gas streams and prevent their release to the environment. Conventional control systems
commonly used for paniculate removal include electrostatic precipitators (ESPs), fabric
filters (baghouse), and mechanical collectors. Table 5-4 provides a distribution of paniculate
emission control systems currently employed by utilities. In some plants, multiple systems
are used in combination. Less than 1 % of all coal-fired plants lack emission control
system(s). In general, these plants are older plants with low nameplate capacity and/or
capacity factor. Flue gas particulates are most commonly removed by electrostatic
precipitation. Fabric filtration (baghouse), and mechanical collectors are used less
frequently. Because fuel oil and natural gas produce much lower quantities of flue gas
particulates, control systems are correspondingly fewer. Fully 91% of gas-fired plants and
57% of oil-fired plants are not equipped with paniculate emission control systems.
Table 5-4. Percent Distribution of Paniculate Emission Control Systems
Type of Fuel
All Coals
Bituminous Coal
Sub-bituminous Coal
Lignite Coal
Natural Gas
Oil'
Emission Control System
ESP
83%
85%
76%
79%
3%
21%
Baghouse
6%
3%
14%
15%
<1%
0%
Mechanical
<1%
1%
8%
2%
6%
19%
Scrubber
2%
<1%
7%
0%
0%
0%
Combination
9%'
11%
3%
0%
1%
0%
None
0%
0%
0%
0%
89%
56%
' The type of emission control system was not specified for 3% of the oil-fired units.
Source: UDI 1992
Beyond these designated paniculate removal systems, flue gases are frequently treated at the
lack phase for removal of sulfur and nitrogen oxides using dry and wet scrubbing.
5-18
-------
Although the primary purpose of these scrubbers, therefore, is not the removal of
particulates, scrubbers do, in fact, serve to reduce paniculate emissions.
«-r-
5.9.1 Paniculate Removal Efficiency
Factors affecting emission control efficiency include the amount of fly ash produced (loading
on the collector), sulfur content of the particulates, electrical resistivity of the particulates,
stack gas temperature, and moisture content of the fly ash and stack gas. Eastern coals
generally produce less fly-ash volume, low fly-ash sulfur content, lower panicle electrical
resistivity, higher stack-gas temperature, and lower moisture content than do western coals.
Midwestern coal properties align more closely with eastern coals than western coals.
Paniculate Size and Temperature. The location at which particle removal takes place can
affect the removal efficiency. This is important in electrostatic precipitation and a factor in
fabric filtration. A link between particle size and temperature is apparent when an
electrostatic precipitator (ESP) is placed upstream from an air preheater unit (flue gas
temperature being 700+° F at this point) or downstream (flue gas temperature being
300+° F). The collector efficiency decreases significantly for particles in the 0.1 to 2.0
micron size range through a hot-side ESP unit and for 0.2 to 5.0 microns in a cold-side ESP
unit. This decrease is attributed to the phenomena of "electric field charging" and "electric
molecular charging." In the submicron size range, "Brownian Movement" becomes
prevalent, with random velocities and directions. Above 500° F, resistivities are generally
low enough for both hot-side and cold-side ESP units, despite the gas moisture content. At
temperatures of 300 to 500° F, the resistivities change by several orders of magnitude for
different flue gas moisture contents.
Moisture and Sulfur Content. Moisture in combination with sulfur significantly changes
paniculate collection in ESP units. Low coal moisture (as found in eastern coals) tends to
raise paniculate resistivity (the degree of repulsion that a particle suspended in stack gas has
against attraction to an ESP field). Higher moisture content (as found in western coals) tends
to form more sulfur trioxide, which lowers electrical resistivity. The humidity of the
combustion air also changes the moisture content of the flue gas, thereby affecting the
interaction of sulfur trioxide with fly ash. Sulfur trioxide either combines with moisture in
flue gas and condenses as an aqueous phase, resulting in enhanced surface conductivity, or it
attaches itself to the surface of fly ash, which promotes the absorption of water and increases
5-19
-------
surface conductivity. Generally, as the sulfur content from one coal type increases over
another, the collection efficiency of an ESP unit also increases. High sulfur coals have
paniculate collection efficiencies of £.8 to 2.0 percent more than low sulfur coals.
5.9 2 Electrostatic Precipitators
Electrostatic precipitation is the most widely used method for removing fly ash from flue
gas. The method of removal consists of passing the particle-laden gas through an
electrostatic field produced by a high-voltage discharge and the paniculate matter is charged
by the interaction of the gas ions. The particles migrate to the collecting surface which has
an opposite polarity and are neutralized.
The adhesive properties of the particles and the action of the electrical field keep the panicles
on the collecting surface and inhibit reentrainment. The panicles are removed by rappers or
by other mechanical devices that vibrate the collector surface and dislodge the paniculate,
which drops by gravity to hoppers. Usually this is accomplished during normal operation;
however, in cases where severe reentrainment is a problem, sections of the precipitator may
be isolated during rapping. The paniculate matter is removed from the hoppers periodically
by either pneumatic or mechanical screw conveyors (Kinkley 1976).
The operating principle of electrostatic precipitation thus requires three basic steps:
(1) electrical charging of the suspended paniculate matter; (2) collection of the charged
paniculate matter on a grounded surface (Shannon 1971); and (3) removal of the paniculate
matter from the collecting surfaces by mechanical scrubbing or flushing with liquids (Kinkley
1976).
Electrostatic precipitators are used extensively on large volume applications where the fine
dust and paniculate is less than 10 to 20 microns in size with a predominant portion in the
submicron range. The precipitators can achieve high efficiencies (in excess of 99%),
depending on the resistivity of the paniculate matter and the characteristics of the gas stream.
The dependence of ash resistivity on fuel characteristics is very important when considering
the application of ESPs to coal-fired boilers.
5-20
-------
If the resistivity of the fly ash is great (2 x 10'° ohm-cm), the dust layer accumulating on the
collecting plates must not be allowed to build up to the thickness typical for a lower
resistivity fly ash. If this occurs, excessive sparking might result. Under these conditions,
the applied voltage would have to be reduced, and the resulting decrease in both corona
current and electric field would lower the collection efficiency (Mcllvane 1979). Particles
with low resistivity also adversely affect the ESP performance. If the resistivity of the
panicles is less than 2 x 107 ohm-cm, the electrical forces holding the dust cake onto the
collecting plates are weaker than with fly ash of a higher resistivity. As a result, more of the
particles are reentrained and the emissions are higher (Oglesby 1970).
Important fuel properties which affect the resistivity of the fly ash are the sulfur and alkali
(primarily sodium) contents of the fuel being burned. Resistivity is altered favorably
(reduced) with an increase in the sulfur content. An increase in the sodium content of the
ash also reduces the resistivity of the fly ash. The effect of coal type on electrostatic
precipitation efficiency is shown in Table 5-5.
Table 5-5. Effect of Various Coals on Electrostatic Precipitation Efficiency
ESP Elements
Fly Ash Amount
ESP Collection Size
Inlet Temperature
Panicle Resistivity
Collection Efficiency
Midwestern Coal
0.1 Ib/million Btu
270 ft2/ 1000 cfm
300°F
109 ohms/cm
99%
Western Coal
0.85 Ib/million Btu
270 ft2/ 1000 cfm
300° F
5 x 1012 ohms/cm
94%
Eastern Coal
0.37 Ib/million Btu
270 ftVlOOO cfm
300" F
5 x 10" ohms/cm
95%
Source: Kumar 1991
5.9.3 Baghouses
Baghouse filter equipment and components include baghouse compartments/hoppers, fabric
bag filters and bag hardware, flue gas inlet/outlet manifolds and isolation valves, reverse-air-
flow fans with ducts and isolation valves, control equipment/systems, insulation and lagging,
and structural support steel with access stairs and walkways to bag inlet, bag suspension, and
compartment roof levels. The fabric mesh bags are arranged in parallel modules. The fabric
mesh bags are made of Teflon-coated fibre glass material or Teflon material. Typical bags
are about 12 inches in diameter and 35- to 40- feet long. The bags are hung from top
5-21
-------
supports with metal rings mounted along the length of each bag to maintain the space interval
between bags and maintain the configuration of each bag. At the bottom end of each bag, a
12-inch long metal sleeve is installed to reduce fabric wear. A baghouse facility location is
downstream of the air preheater.
During normal filter operation, flue gas enters through the bottom of each bag, flows upward
along the inside length, and passes through the porous fabric leaving the participates on the
inside of the bag. The pressure drop across the filter medium increases as the paniculate
collects on the fabric until-a preset time limit is reached, at which time a section is isolated
and the entrapped material is dislodged and collected in hoppers located below the filtering
area. Baghouses are characterized by the methods used to clean the bags as well as the
frequency of bag cleaning. These methods are generally referred to as: (1) shaker type,
(2) reverse air, and (3) pulse jets.
Fabric filters usually provide very high collection efficiencies, exceeding 99.5%, at pressure
drops usually ranging from 4" to 6" of water. The amount of filter area required is often
based on an air-to-cloth ratio of 1.5 to 3.0 cubic feet per minute of gas per square foot of
cloth.
At an air-to-cloth ratio of 1.5 to 2.0 cubic feet per minute per square foot of fabric surface, a
coal-fired plant with 600 MWe capacity may require about 7,000 twelve-inch by forty-foot
long filter bags (Shannon 1982).
5.9.4 Mechanical Collectors
Mechanical collectors include baffle collectors and cyclonic collectors. Baffle collectors
separate the fly ash from the flue gases by passing the flue gas through parallel rows of
angled projections. The velocity of the gas is reduced by splitting the flow and creating drag
surfaces, which allow the particles to settle. Cyclonic collectors are round, conically-shaped
vessels in which the gas stream enters tangentially and follows a spiral path to the outlet.
The spiral motion produces the centrifugal forces that cause the paniculate matter to move
toward the periphery of the vessel and collect on the walls and fall to the bottom of the
vessel. The centrifugal force is the major force causing separation of the paniculate in a
cyclone separator.
5-22
-------
The cyclonic collectors are generally of two types: the large diameter, lower efficiency
cyclones, and the small diameter, multitube high-efficiency units. The larger cyclones have
lower efficiencies especially on particles less than 5 microns. The multitube cyclones are
capable of efficiencies exceeding 90%.
5.9.5 Stack Gas Treatment
The primary purpose of stack gas treatment is to reduce the emission of gaseous
contaminants in flue gas. A secondary benefit, however, is the removal of flue gas
paniculates. Major stack gas contaminants include sulfur dioxide, sulfur trioxide, and oxides
of nitrogen. These gases are most effectively removed by scrubbing the stack gas with
limestone or lime recovered from calcining limestone. Sulfur dioxide can be reduced by 95
percent through scrubbing. Three methods of scrubbing commonly used for removing sulfur
oxides include the following:
Wet Scrubbing. Limestone is ground to a fine mesh size and pulped. The slurry is
pumped directly into the scrubber. Stack gas is passed through the slurry. The sulfur
dioxide reacts very slowly with the slurry. Where scrubbing is applied for control of
fly ash, the contactor used is usually a gas-atomized spray scrubber such as venturi
and high pressure spray impingement scrubbers. The primary collection mechanism
in a wet scrubber is inertia! impaction. In typical venturi scrubber, the particle-laden
gas first contacts the liquor stream in the core and throat of the venturi section. The
gas and liquor streams then pass through the annular orifice formed by the core and
the throat, atomizing the liquor into droplets which are impacted by particles in the
gas stream. Impaction results mainly from the high differential velocity between the
gas stream and the atomized droplets. Collection efficiencies are variable and under
optimal conditions can exceed 95% efficiency.
Drv Scrubbing. Lime is introduced into the scrubber. The reactivity is higher than
wet scrubbing.
Limestone in the Boiler. Limestone is powdered and injected into the boiler furnace
with the fuel. The flue gas carries the lime that is calcined from the limestone in the
combustion zone into the scrubber. A greater potential for scaling exists with this
method.
Oxides of nitrogen may be removed with catalytic scrubbers, which are expensive, or by
modifying fuel combustion. Combustion modification may be performed by staged firing,
low excess air, and flue gas recirculation.
5-23
-------
5.10 RADIONUCLIDE ENRICHMENT IN COAL FLY ASH
Following combustion, essentially all of the original radionuclide content in coal remains in
the ash. The specific activities of certain radionuclides in slag, bottom ash, fly ash, and
post-ESP fly ash have been found to vary, indicating that some radionuclides do not
distribute themselves evenly among the ash fractions. This implies that a radionuclide may
become "enriched" in one fraction at the expense of being "depleted" in another. The
mechanism involved in this enrichment has been cited and discussed by numerous
investigators (Beck 1980, 1989; Coles 1978; Corbett 1980; Roeck 1987; UNSCEAR 1982).
Empirical measurements suggest that the smaller fly ash and post-precipitator ash tends to be
enriched with the more volatile elements, while the slag and bottom ash are depleted in these
nuclides. The maximum enrichment occurs on the smallest fly-ash particles, apparently due
to volatilized material recondensing onto particle surfaces. Since the smallest panicles have
the largest surface-to-volume ratios, it is expected that the smallest particles exhibit the
highest specific activities (i.e., enrichment).
The maximum enrichment has been found on particles with diameters less than about
1 micron. Pb-210 and Po-210 exhibit the greatest enrichment, as much as a factor of five,
while maximum enrichment for uranium isotopes is about a factor of two, and for radium
about a factor of 1.5 (Figure 5-4). The thorium isotopes tend not to show any enrichment in
fly ash. The differences in enrichment for radium and uranium isotopes appear to be related
to the fact that uranium and radium probably exist in a number of different chemical forms in
coal, each with differing volatilities. The enrichment does not appear to continue to increase
as particle sizes become smaller than about 1 micron, suggesting processes other than just
surface condensation are also taking place (Smith 1980; UNSCEAR 1982).
The phenomenon of radionuclide enrichment in fly ash needs careful interpretation. For bulk
fly ash such as that collected by precipitators or baghouse filters, the enhancement of activity
is strictly limited. For example, if a pulverized fuel furnace produces only 20% of its ash as
bottom ash, maximum enrichment of volatile elements in the 80% fly ash fraction cannot
exceed 1.25. However, the sub-fraction of small particles (i.e., < 1 micron) that may pass
through a paniculate emission control system may be significantly enriched with volatile
radionuclides.
5-24
-------
PtrltcU tilt, u
Figure 5-4. Enrichment Factors of Pb-210, U-238, Ra-226, Ra-228, and Th-228 Versus
Size in Stack Fly Ash Collected Downstream from Electrostatic Precipitator
(ESP)
For comparison enrichment factors for cerium are plotted. Cerium
concentrations taken from INAA analysis of same samples (Coles 1978).
5-25
-------
On the basis of published empirical data, the EPA has adopted enrichment factors cited in
Table 5-6.
Table 5-6. Enrichment Factors for Radionuclides
Uramum-238 Series
U-238
U-234
Th-230
Ra-226
Rn-222
Pb-210
Po-210
Thorium-232 Series
Th-232
Ra-228
Th-228
Ra-224
Rn-220
Pb-212
Po-212
Ennchment Factor
2
2
1
1.5
20
5
5
1
1.5
1
1.5
20
5
5
It is important to note, however, that these enrichment factors are based on fly ash collection
efficiencies of roughly 98% or higher. As collection efficiency drops below 90%, the
contribution of enriched submicron particles to the total mass of fly ash diminishes to an
insignificant level. Consequently, emission estimates for plants with paniculate removal
efficiencies less than 90% require no adjustments for radionuclide enrichment.
5-26
-------
6. ESTIMATES OF RADIONUCLIDE PLANT EMISSIONS AND HEALTH RISKS
TO SURROUNDING POPULATIONS
As noted in previous chapters, the principal variables affecting radionuclide emissions include
fuel type and fuel consumption, and the efficiency of emission control systems. An
additional factor unique to coal-fired plants is ash partitioning between fly and bottom ash,
which is affected by furnace design. Variables affecting exposure and risk include
meteorological and demographic factors such as annual precipitation, population distribution,
and local food production. To assure that unusual parameter values are not overlooked, risks
are evaluated on a site-specific basis.
In this final chapter of the report, the following information is presented: Section 6.1
profiles utilities on the basis of physical and operational parameters that affect emissions;
Section 6.2 summarizes unit-/plant-specific emission data; and Section 6.3 quantifies health
risks that are derived using the CAP-93 computer code that combines standardized models
and site-specific data.
6.1 STATISTICAL PROFILE OF RELEVANT PLANT PARAMETERS
The information presented in this section summarizes statistical data contained in the Edison
Electric Institute's (EEI) Power Statistics Data Base. This data base reflects information
supplied annually by utilities on EEI Form 767. The information is sorted and analyzed for
EEI by the Utility Data Institute, Inc. (1700 K Street, NW, Suite 400, Washington, D.C.
20006). The February 1991 edition of the data base reflecting information supplied by
utilities for 1990 was used. Analysis was restricted to utility units with 25 MWe capacity or
greater. This limited selection is justified since smaller units do not significantly contribute
to base-load electric power operation and their inclusion would bias emission estimates on the
low side.
In 1990, the 684 utility plants evaluated in this study had a total of 1,748 operable fossil-
fueled boilers with a combined capacity of over 460 GWe of electrical power generation
(Table 6-1). About 60% of all units used coal as their primary energy source and represent
nearly 69% of the generating capacity. In terms of generating capacity, gas-fired and oil-
fired units represent a distant second and third category, respectively, among utility boilers.
6-1
-------
Table 6-1. Summary of Electrical Power Generation for 1990
Fuel Type
Coal
Gas
Oil
TOTAL
Numbec of
Units
1.045
512
191
1,748
Generator
Capacity
(MWe)
317.635
99.573
• 43.218
460.426
Percent of
Total
Capacity
69.0
21.6
9.4
100.0
Source: UDI 1992
6.1.1 Coal
Table 6-2 identifies the number and their corresponding percent distribution of coal-fired
boilers based on coal rank consumed, furnace type, and emission control system. Based on
these three variables, 18 combinations were found to exist. The dominant combination
represents 573 units that use bituminous coal, operate a pulverized dry-bottom furnace, and
employ an electrostatic precipitator. In terms of coal use, 76% (i.e., 793 units) used
bituminous coal and represent 68% of the generating capacity among coal-fired boilers. Sub-
bituminous coal represents the second largest category and was used by 219 units with about
27% of the generating capacity. Fewer than 5% of the units use lignite.
The most prevalent furnace (85%) is the pulverized dry-bottom type. (This furnace type
produces a fly ash to bottom ash ratio of 80/20.) For paniculate emission control, 863 units,
or 83% of all coal-fired units, employ electrostatic precipitators. It is important to note that
all coal-fired units with a generating capacity of 25 MWe or greater have controlled
emissions.
To obtain a distribution of the generating capacity of plants from the EEI Power Statistics
Data Base, the nameplate unit capacity from file UDESIGN.dbf was summed by plant nanie,
thus totaling the unit MWe capacity into a plant MWe capacity value. This sorting revealed
that the 1,045 units correspond to a total of 420 plants. The range of plant generating
capacity ranged from 25 MWe to a maximum of 3,564 MWe. The distribution of the
generating capacity for the 420 plants was analyzed by means of the PC program, Microsoft
Excel. This program creates a set of "bins" between the data's minimum and maximum
6-2
-------
Table 6-2. Statistical Profile of Coal-Fired Units
Coal Rank
Bituminous
Bituminous
Bituminous
Bituminous
Bituminous
Bituminous
Bituminous
Bituminous
Sub-Bituminous
Sub-Bituminous
Sub-Bituminous
Sub-Bituminous
Sub-Bituminous
Sub-Bituminous
Sub-Bituminous
Lignite
Lignite
Lignite
Lignite
Lignite
Furnace
Type
dry
dry
dry
dry
dry
wet
wet
wet
dry
dry
dry
dry
wet
wet
wet
dry
dry
dry
wet
wet
Emission
Control
ESP
Baghouse
Mechanical
Scrubber
Combination
ESP
Combination
Baghouse
ESP
Baghouse
Combination
Scrubber
ESP
Baghouse
Scrubber
ESP
Baghouse
Mechanical
ESP
( Baghouse
TOTAL
No. of
Units
573
20
5
2
85
98
5
5
137
24
8
14
29
6
1
19
4
2
7
I
1.045
% of Total
Coal Boilers
54.8
1.9
0.5
0.2
8.1
9.4
0.5
0.5
13.1
2.3
0.7
1.3
2.8
0.6
0.1
1.8
0.4
0.2
0.7
O.I
100.0
% of Total
Capacity MWe'
53.8
1 4
0 1
0.6
3.9
7.7
0.2
0.3
19.1
3.2
04
1 7
2.4
0.2
0.3
2.9
0.3
<0.1
1.3
0.1
100.0
The total generator nameplate capacity of the 1,045 coal units in 1990 was 317.635 MWe.
Not included were units with less than 25 MWe nameplate capacity.
values that is equal to the square root of the number of input data points. For the 452 plants'
generator nameplate capacity values (i.e., data points), the software program defined 21 bios.
Table 6-3 presents the capacity distribution of coal-fired plants for 1990; Figure 6-1 depicts
the distribution in graphical form. The data suggest a log-normal distribution for which the
measure of distribution is best defined by the geometric mean (Mg). From primary data, the
following statistical values were determined:
• Maximum plant generating capacity: 3,340 MWe
• Geometric mean (Mg) of generating capacity for the 420 coal-
fired plants: 451 MWe
• 95th percentile value of generating capacity: 2,300 MWe
6-3
-------
Table 6-3. Distribution of Generating Capacity of U.S.
Coal-Fired Plants for 1990
Nameplate Generating
Capacity (MWe)
25 - 199
200- 349
350- 499
500- 699
700- 849
850 - 1049
1050- 1199
1200 - 1399
1400 - 1449
1450 - 1549
1550 - 1699
1700 - 1899
1900 - 2049
2050 - 2199
2200 - 2399
2400 - 2549
2550 - 2699
2700 - 2899
2900 - 3049
3050 - 3199
3200 - 3400
Number of
Plants
97
6L
45
55
30
15
20
26
2
4
12
19
6
4
7
4
6
2
1
1
3
Cumulative
Percent (%)
23.1
37.6
48.3
61.4
68.6
72.1
76.9
83.1
83.6
84.5
87.4
91.9
93.3
94.3
96.0
96.9
98.3
98.8
99.0
99.3
100.00
TOTAL 420
6.1.2 Gas-Fired Plants
Processed natural gas served as the primary energy source for 512 boiler units, each with a
generator nameplate capacity of 25 MWe or greater (Table 6-4). Due to the lack of
substantial quantities of ash that result from the combustion of natural gas and the limited
amount of non-particulate pollutants, emission control systems are generally not found in
boilers that exclusively burn natural gas. Not surprisingly, therefore, 454 or 88.7% of gas
boilers are not equipped with an emission control system. It is not uncommon, however, for
boilers intended to use natural gas as the primary energy source to be designed to burn
residual fuel oil as a secondary energy source. Therefore, about 11% of gas-fired boilers are
equipped with paniculate emission controls that are activated during secondary fuel
operations.
6-4
-------
120
Frequency (No. of Plants)
100 -
3M 000 TOO OM 10*0 1*00 1400 I«H IMO 1700 IfiOO
Megawatts
120%
- 100%
- 80%
- 60%
40%
- 20%
Figure 6-1. Sorted Histogram of Coal-Fired Power Plants by Megawatts
6-5
-------
Table 6-4. Distribution of Generating Capacity of U.S.
Gas-Fired Plants for 1990
Paniculate
Control
None
Mechanical
ESP
Baghouse
Combination
N». of
Units
454
37
IS
1
5
% of Total
Gas Boilers
88.7
7.2
2.9
0.2
1.0
% of Total
Capacity MWE'
89.2
6.5
3.4
<0.1
<0.1
TOTAL 512
A total of 99,573 MWe was generated by boilers designated as gas-fired.
The 512 gas-fired units correspond to a total of 237 plants. The distribution of generator
nameplate capacity of the 237 plants is given in Table 6-5. As with coal-fired plants, a data
plot (Figure 6-2) suggests that gas-fired plant capacity is log-normally distributed. Analysis
of primary data reveals the following statistics:
• Maximum plant capacity: 2,314 MWe
• Geometric mean (Ms) plant capacity: 230 MWe
• 95th percentile plant capacity value: 1,609 MWe.
Table 6-5. Distribution of Gas-Fired Plants.
Plant Capacity
(MWe)
25 - 149
150 - 299
300-449
450 - 599
600-749
750 • 899
900 - 1.099
1.100- 1,599
1,600- 1.999
2.000-2.199
2,200 - 2.399
Number of
Plants
93
37
32
14
17
16
10
6
7
3
2
Cumulative
Percentage
39.2%
54.8%
68.4%
74.3%
81.4%
88.2%
92.4%
94.9%
97.7%
99.2%
100.0%
TOTAL 237
6-6
-------
130
Frequency (No. of Plants)
100 -
120%
- 100%
- 80%
- 60%
- 40%
- 20%
too soe
Megawatts
Figure 6-2. Sorted Histogram of Gas-Fired Power Plants by Megawatts
6.1.3 Oil-Fired Plants
In 1990, there were 191 operable oil-fired units, each having a generator nameplate capacity
of at least 25 MWe (Table 6-6). Due to the relatively low ash content of oil, more than one-
half (55.5%) of the .units are not equipped with paniculate emission controls. -Of those units
that have emission controls, mechanical collectors are used almost as frequently as ESPs
(19.9% versus 21.5%).
Sorting units by name and generating capacity shows that the 191 oil-fired units represent 97
plants and exhibit a log-normal distribution, as shown in Table 6-7 and Figure 6-3. Analysis
of primary data yields the following summary statistics:
• Maximum plant generating capacity: 2,126 MWe
• Geometric mean (Mg) of generating capacity for the 97 plants: 245 MWe
• 95th percentile value of plant generating capacity: 1,472 MWe
6-7
-------
Table 6-6. Statistical Data of Oil-Fired Units for 1990
Paniculate
Control
None
ESP
Mechanic
Not Specified
No. of
Units
106
41
38
6
% of Total
Oil Boilers
55.5
21.5
19.9
3.1
% of Total
Capacity MWe'
48.4
26.1
23.8
1.7
TOTAL 191
A total of 43,218 MWE electricity was generated by boilers designated as oil-
fired.
Table 6-7. Distribution of Oil-Fired Plants
Plant Capacity
(MWe)
25 - 279
280 - 549
550 - 799
800- 1.099
1.100- 1.299
1.300- 1.599
1.600- 1.849
1.850-2.124
2.125-2.399
Number of
Plants
51
17
8
11
4
2
2
1
I
Cumulative
Percentage
52.6%
70.1%
78.3%
89.6%
93.7%
95.8%
97.9%
99.0%
100.0%
TOTAL 97
6.2 EMISSION ESTIMATES
Emission estimates were made for each of the 1,748 operable fossil-fueled boiler units by
applying the previously derived average radionuclide concentrations for coal, gas, and oil and
unit-specific parameter values that directly affect emissions. Emission estimates were
determined for each unit based on average radionuclide content, amount of fuel(s)
consumption, ash partitioning (coal only), and paniculate removal efficiency. Data on fuel
consumption and paniculate removal efficiencies are from the Utility Data Institute's data
base and reflect the industry in 1990. Planned upgrades of emission control systems and/or
fuel switching are not accounted for in the emissions estimates. In addition, whenever the
6-8
-------
Frequency (No. of Plants)
120%
- 100%
- 80%
- 60%
• 40%
- 20%
25 280 550 800 1100 1300 1600 1850 2125
Megawatts
Figure 6-3. Sorted Histogram of Oil-Fired Power Plants by Megawatts
paniculate removal efficiency exceeded 90%, radionuclide-specific enrichment factors were
applied, as previously discussed.
Emission data for the 1,748 boiler units were appropriately grouped to provide estimates for
the corresponding 684 utility plants. Due to the large amount of data, this information is
provided separately in an addendum. A sample of the boiler unit-/plant-specific data is
illustrated on the succeeding page for Plant #301. This plant has three gas-fired units labeled
sources 1 through 3 and two coal-fired units labeled sources 4 and 5. Annual emissions are
estimated independently for each of the five units and combined to yield plant emissions.
Radionuclides are classified on the basis of lung clearance by the standard notation Y (year),
W (week), and D (day). All particles are assumed to have a mean diameter of 1 micron.
(The addendum also contains plant-specific estimates of cancer risks to the exposed
population. Section 6.3 that follows discusses population risks associated with plant-specific
emissions.) For the maximally exposed individual located 200 meters in sector South, the
lifetime fatal cancer risk is estimated to be 2E-OS.
Presented below is an overview of emission estimates in the form of collective emissions.
Table 6-8 cites emission estimates for 1,045 boiler units with coal designated as the primary
6-9
-------
Illustration of Data Contained in Addendum
PLANT #301 - 3 GAS-FIRED & 2 COAL-FIRED UNITS
RADIONUCLIDE EMISSIONS BASED ON 1990 FUEL CONSUMPTION DATA
Nuclide Class Size
Source Source Source Source Source
#1 82 #3 *4 85 TOTAL
Ci/y Ci/y Ci/y ci/y Ci/y Ci/y
U-23B
Th-234
Pa-234m
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232
Ra-228
AC-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
Tl-208
Y
Y
Y
Y
Y
Y
W
*
W
D
W
W
D
W
W
Y
W
Y
Y
W
*
H
D
W
D
00
00
00
00
00
00
00
0.00
.00
.00
.00
.00
.00
.00
1.00
1.00
1.00
.00
.00
.00
0.00
.00
.00
,00
1.00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
1.6E+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
1.7E+00 2.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0.
O.OE+00 0,
O.OE+00 0
O.OE+00 0,
O.OE+00 0,
O.OE+00 0.
O.OE+00 0,
O.OE+00 0,
O.OE+00 0.
O.OE+00 0,
O.OE+00 0,
OE+00 8,
OE+00 4,
OE+00 4
OE+00 0,
OE+00 8
OE+00 4
OE+00 6
3E+00 3
OE+00 2
OE+00 2
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
OE+00
1E-03
OE-03
OE-03
OE+00
1E-03
OE-03
OE-03
OE-01
OE-02
OE-02
OE-03
OE-02
OE-02
OE-03
OE-02
6.7E-03
3.4E-03
3.4E-03
O.OE+00
6.7E-03
3.4E-03
5.OE-03
4.6E-01
1.7E-02
1.7E-02
3.4E-03
1.7E-02
1.7E-02
3.4E-03
3.5E
2.4E
2.4E
. 5E
,5E
,2E
.2E
2.4E
7. IE
2.4E-03 2,
•03
•03
•03
•03
•01 2
•02 9
•02 9
•03 2
-04 5
OE-03
OE-03
OE-03
OE-03
OE-03
5E-01
9E-03
9E-03
OE-03
9E-04
1.5E-02
7.4E-03
7.4E-03
.OE+00
,5E-02
-4E-03
. 1E-02
.4E+00
.7E-02
3.7E-02
7.4E-03
3.7E-02
3.7E-02
7.4E-03
0.
1.
7.
1.
6,
3.
1.7E-02 3.7E-02
4.4E-03
6.5E-03
4.4E-03
4.4E-03
6.5E-03
4.OE-01
2.2E-02
2.2E-02
4.4E-03
1.3E-03
K-40
1.00 O.OE+00 O.OE+00 O.OE+00 2.7E-02 2.2E-02 4.9E-02
FREQUENCY DISTRIBUTION OF LIFETIME FATAL CANCER RISKS
Risk Range
Number of
People
Number of People
In This Risk
Range or Higher
Deaths/Year
In This
Risk Range
Deaths/Year
In This Risk
Range or Higher
1.OE-03 TO l.OE-04 0
l.OE-04 TO l.OE-05 6
l.OE-OS TO l.OE-06 581
LESS THAN l.OE-06 344262
0
6
587
344849
O.OOE+00
1.21E-06
1.32E-05
3.02E-04
O.OOE+00
1.21E-06
1.44E-05
3.17E-04
MAXIMUM INDIVIDUAL LIFETIME FATAL CANCER RISK = 2E-05 9 200 m S
6-10
-------
Table 6-8. Estimates of Collective Annual Emissions by Designated Primary Fuel Source
Radionuclide
U-238
Th-234
Pa-234m
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
Tl-208
K-40
All Coal-Fired Units
Ci/y
2.4E+00
1.2E-KX)
1.2E+00
1.2E+00
2.4E+00
1.2E+00
1.8E+00
3.1E+00
5.9E+02
5.9E+00
1.2E+00
5.9E+00
5.9E+00
1.2E+00
5.9E+00
7.4E-01
1.1E+00
7.4E-01
7.4E-01
l.lE-t-00
1.7E+02
3.7E+00
3.7E+00
7.4E-01
2.2E-01
8.2E+00
% of Total
- 100
- 98
- 98
- 98
- 96
- 92
- 92
- 18
- 99
- 99
- 95
- 99
- 47
- 15
- 47
- 94
- 52
- 41
- 41
- 52
- 52
- 78
- 78
- 42
- 42
- 100
All Gas-Fired Units
Ci/y
6.9E-03
6.9E-03
6.9E-03
6.9E-03
1.3E-02
2.5E-02
1.5E-02
1.3E+03
1.6E-02
1.6E-02
1.6E-02
1.6E-02
1.6E+00
I.6E+00
1.6E+00
1.1E-02
2.4E-01
2.4E-01
2.4E-01
2.4E+01
2.9E-01
2.4E-01
2.4E-01
2.4E-01
7.1E-02
3.2E-03
% of Total
< 1
< 1
< 1
< 1
< 1
- 2
< 1
- 77
< 1
< 1
- 1
< 1
- 13
- 20
- 13
- 1
- 11
-14
- 14
- 11
< 1
-5
- 5
- 14
- 14
< 1
All Oil-Fiitd Units
Ci/y
2.1E-02
2.1E-02
2.1E-02
2.1E-02
4.0E-02
7.9E-02
5.0E-02
7.3E+01
5.1E-02
5.1E-02
5.1E-02
5.1E-02
S.1E+00
5. IE +00
5.1E+00
3.5E-02
7.9E-01
7.9E-01
7.9E-01
7.9E-01
1.6E+00
7.9E-01
7.9E-01
7.9E-01
2.3E-01
l.OE-03
% of Total
< 1
- 2
- 2
- 2
- 3
- 6
- 7
- 4
- I
- 1
- 4
- 1
- 40
- 65
- 40
- 4
- 37
-45
-45
- 37
- 48
- 17
- 17
- 44
— 44
< 1
All Units
(Ci/y)
2.4E+00
1.2E+00
1.2E+00
1.2E+00
2.4E+00
UE+00
1.9E+00
1.7E+03
6.0E+00
6.0E+00
1.3E+00
6.0E+00
1.3E+01
7.9E+00
1.3E+01
7.8E-01
2.1E-KX)
1.8E+00
1.8E+00
2.1E+00
1.7E+02
4.7E+00
4.7E+00
1.8E+00
5.3E-01
8.2E+00
fuel source; 512 units with gas designated as the primary fuel source; and 191 units with oil
designated as the primary fuel source. The data reveal the following observations:
• Combined emissions for all boiler units range between 0.5 curies to 13 curies per
year for all radionuclides other than radon.
• For the majority of radionuclides, collective emissions are generally highest for units
designated as coal-fired.
6-11
-------
• Collective emissions from oil-fired units are nearly comparable to those of coal for a
limited number of radionuclides, which include Pb-210, Bi-210, Po-210, Ra-228, Ac-
228, Ra-224, Rn-220, Bi-212, and 11-208.
• More than three-quarters of all radon (Rn-222) emissions come from gas-fired units.
• Radionuclide emissions other than radon-222 among units designated as gas-fired are
due to combustion of secondary fuels, which most often involves oil.
Estimates of collective emissions by fuel type, however, are biased since they are not
adjusted for the number of boiler units or the amount of electricity generated. Table 6-9
provides a more meaningful comparison by showing average annual emissions per operating
unit as well as emissions per billion kWe-hr of electricity generated. For coal-fired units,
the average annual emissions for paniculate radionuclides ranges from a fraction of a
millicurie up to several millicuries.
The wider range of radionuclide-specific emission values for oil-fired units is principally due
to the absence of secular equilibrium among decay chain members in oil. Thus, average
emission values for some radionuclides are considerably lower among oil-fired units while
others are only marginally lower than those for coal-fired units. Furthermore, the average
annual emissions of about 27 mCi for Pb-210, Bi-210, and Po-210 is several times higher
than-corresponding estimates for the average coal-fired unit.
Because coal-fired units have, on the average, a higher generating capacity, comparative
emission values between coal- and oil-fired units are further impacted by relating emission
values to a common quantity of electrical energy production. Thus, Table 6-9 provides
average radionuclide emission estimates (in millicuries per year per billion kilowatt hours of
electricity generated for each fuel type).
Paniculate emissions for units designated as gas-fired are generally small when compared to
either coal- or oil-fired units. Moreover, paniculate emissions from units designated as gas-
fired result from the combustion of secondary fuel(s), which most commonly involves oil.
6-12
-------
Table 6-9. Average Annual Radionuclide Emissions per Operating Boiler Unit
and per Billion kWe-hr Electricity Generated
Radionuclide
U-238
Th-234
Pa-234m
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
B i-2 12
Tl-208
K-40
Emission Rates
Per Operating Unit (mCi/y)
Coal
2.3E+00
1.2E+00
1.2E+00
1.2E+00
2.3E+00
1.2E+00
1.7E+00
3.0E+02
5.6E+00
5.6E+00
1.2E+00
5.6E+00
5.6E+00
1.2E+00
S.6E+00
7.1E-01
l.OE+00
7.1E-01
7.1E-01
l.OE+00
1.6E+02
3.5E+00
3.5E+00
7.1E-01
' 2.1E-01
7.8E+00
Gas
1.3E-02
1.3E-02
1.3E-02
1.3E-02
2.5E-02
4.9E-02
2.9E-02
2.5E+03
3.1E-02
3.1E-02
3.1E-02
3.1E-02
3.1E+00
3.1E+00
3.1E+00
2.1E-02
4.7E-01
4.7E-01
4.7E-01
4.7E-01
5.7E-01
4.7E-01
4.7E-01
4.7E-01
1.4E-01
6.2E-03
Oil
1.1E-01
1.1E-01
1.1E-01
1.1E-01
2.1E-01
4.1E-01
2.6E-01
3.8E+02
2.7E-01
2.7E-01
2.7E-01
2.7E-01
2.7E+01
2.7E+01
2.7E+01
1.8E+01
4. IE +00
4.1E+00
4.1E+00
4. IE -1-00
8.4E+00
4. IE +00
4.1E+00
4.1E+00
1.2E+00
5.2E-03
Per Billion kWe-hr Generated (mCi/y)
Coal
1.5E+00
7.7E-01
7.7E-01
7.7E-01
1.5E+00
7.7E-01
1.2E+00
2.0E+02
3.8E+00
3.8E+00
7.7E-01
3.8E+00
3.8E+00
7.7E-01
3.8E+00
4.7E-01
7.1E-01
4.7E-01
4.7E-01
7.1E-01
1.1E+02
2.4E+00
2.4E+00
4.7E-01
1.4E-01
5.3E+00
Gas
2.6E-02
2.6E-02
26E-02
2.6E-02
4.9E-02
9.5E-02
5.7E-02
4.9E+03
6.0E-02
6.0E-02
6.0E-02
6.0E-02
6.0E+00
6.0E+00
6.0E+00
4.1E-02
9.1E-01
9.1E-01
9.1E-01
9. IE -01
1.1E-J-00
9.1E-01
9.1E-01
9.1E-01
2.7E-01
1.2E-02
Oil
1.8E-01
1.8E-01
1 8E-01
1.8E-01
3.4E-01
6.7E-01
4.3E-01
6.2E+02
4.4E-01
4.4E-01
4.4E-01
4.4E-01
4.4E+01
4.4E+01
4.4E+01
3.0E-01
6.7E+00
6.7E+00
6.7E+00
6.7E+00
1.4E+01
6.7E+00
6.7E+00
6.7E+00
1.9E+00
8.5E-03
6.3 ESTIMATES OF HEALTH RISKS
The atmospheric releases of radioactivity derived for each of the 684 fossil fuel plants may
contribute to population radiation exposure through external and internal exposure pathways.
External exposure pathways include direct radiation from radioactive plumes and from
6-13
-------
ground deposition. Internal exposures may result from the inhalation of airborne
radioactivity or the ingestion of contaminated water and food products.
For low doses of radiation, the primary health risk is the potential induction of cancer. The
association of cancer risk with low level exposures is complex. For this reason, a brief
overview is provided in Appendix C, which explains fundamental concepts of radiation
injury, epidemiologic data and risk models, and the scientific method for estimating cancer
risks.
Estimates of population risks that result from chronic atmospheric releases require the use of
a computer code that (1) models atmospheric dispersion, radionuclide concentrations in
environmental media, and radionuclide intakes by the exposed population and (2) calculates
whole-body and tissue doses and their associated health risks.
In support of 40 CFR 61, National Emission Standards for Hazardous Air Pollutants, EPA,
with support from Oak Ridge National Laboratory, developed the CAP-88 computer model.
The CAP-88 (which stands for Clean Air Act Assessment Package - 1988) computer model is
a composite of computer programs, data bases, and associated utility programs. CAP-88
programs are considered among the best available verified models. It produces results that
agree with experimental data as well as other confirmed models (Beal 1986, Maheras 1994).
Since it was first used, CAP-88 has been revised periodically. The most recent version of
the code, designated as CAP-93, was used to estimate radiation doses and fatal cancer risks
for radionuclide emissions from the 684 fossil fuel plants.
6.3.1 Summary of CAP-93 Model
For a given facility, atmospheric releases may be modeled for as many as six independent
sources. Plume rise can be calculated assuming either a momentum- or buoyancy-driven
plume that reflects facility-specific plant parameters. Plume dispersion is based on a
modified Gaussian plume equation and accounts for plume depletion that includes
precipitation scavenging and dry deposition. Based on availability, primary model
parameters for plume dispersion and depletion are based on site-specific meteorological data.
(A library of meteorological data that include wind data files, annual precipitation, ambient
temperatures, and lid-height for all major cities are provided by the code.)
6-14
-------
From plume dispersion and plume depletion calculations, the program computes radionuclide
concentrations in air and rates of deposition and build-up on ground surfaces and in soil.
Estimates of the radionuclide concentrations in produce, leafy vegetables, milk, and meat are
made by coupling the output of the atmospheric transport models with the terrestrial food-
chain models defined in the U.S. Nuclear Regulatory Commission's Regulatory Guide 1.109.
The quantities of foodstuff produced locally are based on the average agricultural
productivity data of the State in which the assessment area is located.
For dose and risk estimates, the population distribution at each of the 684 assessed sites were
developed by means of the GENPOP computer code and 1990 Census Bureau data. Dose
estimates reflect the exposure from external (air immersion and ground surface) and internal
(inhalation and ingestion) sources. For low-LET external radiation, CAP-93 employs the
nominal risk coefficient of 3.9E-4 fatal cancers per rem.
For internal exposures, dose and risk estimates are defined by ICRP tissue/organ weighting
factors that account for route of entry, clearance class, and transfer factors within body
compartments. In summary, dose and cancer risks can be tabulated for individual exposure
pathways, radionuclides, and tissues/organs. All risk estimates pertain to the risk of fatal
cancer and assume that exposure occurs over the lifetime of individuals within the assessed
population.
EPA's methodology for estimating risks from Rn-222 emissions is based on an extrapolation
of epidemiologic findings of underground miners (NAS 1988, NAS 1991) exposed to radon.
CAP-93 calculates working levels (WL), not concentrations of specific radon daughter
products. A WL is defined as any combination of short-lived radon decay products in one
liter of air that- will result in the emission of 1.3 x 10s MeV of alpha-panicle energy. Risk
is not derived from dose but from time-integrated exposure expressed in working level
months (WLM). Under typical residential exposure conditions, it is assumed that I WLM
corresponds to 170 hours of exposure at 200 picocunes per liter (pCi/L) of radon gas.
CAP-93 employs a risk coefficient of 3.6E-4 fatal lung cancers per WLM.2
1 Recently, the Agency has revised its estimates of radiogenic cancer risks to reflect the current
epidemiotogical data and scientific consensus on extrapolations from the available data to chronic low dose
exposures (EPA94). The revised estimates yield a nominal value of 5.1E-4 fatal cancers per rad for uniform
whole body exposure to low-LET radiation and 2.2E-4 fatal lung cancers per WLM for exposure to radon-222
and us decay products. The radon risks reported in this study can be adjusted to the new radon risk coefficient
simply by applying a correction factor of about 0.6 No simple adjustment can be made to the non-radon risks
to reflect the Agency's current values. However, since the ground surface pathway dominates the risk for
maximally exposed individuals, an upward adjustment of approximately 30 percent would bound-their risks
6-15
-------
CAP-93 assesses risk for a circular grid that is defined by sixteen sectors and up to 20 radial
distances around a specified facility. For this study radial distances of 400, 1,500, 3,500,
7,500, 10,000, 15,000, 25.000, 35,000, 45,000, and 50,000 meters were used. Risk to the
population is determined by summing individual risks by distance and sector for the 0-50 km
grid around each assessed facility. Risk to the maximally exposed individual(s) corresponds
to that location (i.e., distance and sector of highest exposure) where individuals are believed
to reside.
The population risk frequency distribution identifies the number of people at various levels of
risk. The risk categories are divided into powers of ten, in which the individual lifetime
cancer risk ranges from one chance in ten to less than one chance in a million. Risk data for
each of the 684 assessed plants are provided in the previously identified Addendum. Only a
summary of these data is provided below.
6.3.2 Summary Findings of Population Risks
Table 6-10 defines the distribution of fatal cancer risks to the combined populations residing
within the 50-km radii of the 684 fossil-fueled electric utility plants. The aggregate of
assessed populations living within a 50-km (35 mile) radius of a plant is estimated to be
196.1 million persons, representing approximately 75% of the U.S. population. The
individual lifetime risk of fatal cancer to more than 99.9% of the assessed population (i.e.,
196 million) is less than one chance in a million. The data further imply that under current
operating conditions there are no instances in which the release of radioactivity is likely to
result in a lifetime fatal cancer that is equal to or greater than one chance in ten thousand to
any one person. It is estimated that about 1,027 individuals residing in the 50-km distance of
a plant may receive radiation exposures for which the lifetime risk is between 1 in 10,000
and 1 in 100,000 (i.e., 1E-04 to IE-OS).
It must be also pointed out that the distribution of individual risk within each risk range is
heavily skewed toward the lower value. This is evidenced by the fact that the average
individual lifetime risk is a small fraction of the midpoint value within each of the risk
ranges. Correspondingly, the probability of a single fatal cancer occurrence within the
highest risk group of 1,027 individuals is less than 2 chances in 10,000 per year. For the
entire assessed population of 196.1 million within 50 km of these plants, the estimated cancer
risk attributable to radionuclide emissions from SGUs is less than 1 cancer death per year
6-16
-------
Table 6-10. Frequency Distribution of Lifetime Fatal Cancer Risks for All Plants
Lifetime Cancer
Risk Range
l.OE+OOto l.OE-01
l.OE-01 to l.OE-02
1.0E-02to .OE-03
1.0E-03to .OE-04
1.0E-04to .OE-05
l.OE-OSto .OE-06
Less Than .OE-06
Number of
People •-"
0
0
0
0
1.027
95,745
196.000,000
Average Individual
Lifetime Risk
O.OE+00
O.OE+00
O.OE+00
O.OE+00
1.3E-05
2.2E-06
1.2E-07
Deaths per Year in
this Risk Range
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.92E-04
3.06E-03
3.32E-01
Deaths per Year in
this Risk Range or
Higher
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.92E-04
3.26E-03
3.36E-01
(i.e., 3.36E-01 deaths/year is the risk equivalent of about 1 in 3 chances that a single cancer
death will occur in a year). Exposures and risks to individuals residing beyond 50 km are
not explicitly evaluated. However, MIRs at such distances would be 1E-7 or less and the
overall cancer incidence would not, based on removal of about 1/3 of the emitted
radionuclides within SO km, exceed about 1 death/year.
An overview of the relative contribution to population risks is also gained by segregating the
collective risks defined in Table 6-10 on the basis of primary fuel designation.
Correspondingly, Tables 6-11, 6-12, and 6-13 provide the frequency distribution of risks
from the combustion of coal, oil, and gas. Among the 684 plants, there were 420 plants
where at least one unit consumed coal as its primary fuel, and the combined 0-50 km
populations totalled about 141 million. It is estimated that 67 individuals fall within the risk
range of 1 in 10,000 to 1 in 100,000. Although coal-fired units represent the largest
percentage of fossil fuel units, they contribute only about 6% to the 1027 individuals
previously identified in this risk range for all plants/all fuels. For the next lower risk range,
the estimated 9,499 individuals also constitute a small percentage (i.e., -10%) to the total
pool of 95,745 individuals within the risk range of 1.OE-05 to 1.OE-06.
Inspection of Table 6-12 indicates that the oil-fired plants disproportionately contribute to the
pool of persons whose individual risks are greater are greater than 1 in 100,000. Moveover,
the disproportionate number of individuals at risk from oil-fired plants is primarily the
contribution of a single plant. Plant #651 consists of 6 oil-fired units and is estimated to
contribute 947 and 73,462 individuals to the risk ranges of l.OE-4 to l.OE-5 and 1.OE-05 to
6-17
-------
Table 6-11. Frequency Distribution of Lifetime Fatal Cancer Risks for Coal-Fired Units
Lifetime
Cancer
Risk Range
l.OE+OOio 1 OE-01
l.OE-01 to 1 OE-02
I .OE-02 to l.OE-03
l.OE-03 to l.OE-04
1.0E-04to l.OE-05
l.OE-OSto l.OE-06
Less Than 1 .OE-06
Number
of
People •-"
0
0
0
0
67
9,499
141.000.000
Average Individual
Lifetime Risk
O.OE+00
O.OE+00
O.OE+00
O.OE+00
1.5E-05
2.0E-06
. 3.4E-08
Deaths per Year
in this
Risk Range
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.43E-05
2.70E-04
6.89E-02
Deaths per Year ID
this Risk Range or
Higher
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.43E-05
2.84E-04
6.92E-02
Table 6-12. Frequency Distribution of Lifetime Fatal Cancer Risks for Oil-Fired Units
Lifetime
Cancer
Risk Range
l.OE+OOto l.OE-01
l.OE-01 to 1. OE-02
1. OE-02 to l.OE-03
l.OE-03 to l.OE-04
l.OE-04 to l.OE-05
l.OE-OSto l.OE-06
Less Than l.OE-06
Number
of
People
0
0
0
0
960
8S.62S
90.100.000
Average Individual
Lifetime Risk
' O.OE+00
O.OE+00
O.OE+00
O.OE+00
1.3E-OS
2.2E-06
I.2E-07
Deaths per Year
in this
Risk Range
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.77E-04
2.75E-03
1.57E-01
Deaths per Year in
this Risk Range or
Higher
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
1.77E-04
2.92E-03
1.60E-01
Table 6-13. Frequency Distribution of Lifetime Fatal Cancer Risks for Gas-Fired Units
Lifetime
Cancer
Risk Range
l.OE+OOto l.OE-01
l.OE-01 to 1. OE-02
1. OE-02 to l.OE-03
l.OE-03 to l.OE-04
l.OE-04 to l.OE-05
l.OE-OSto l.OE-06
Less Than 1 OE-06
Number
of
People
0
0
0
0
0
1.855
118,000.000
Average Individual
Lifetime Risk
O.OE+00
O.OE+00
O.OE+00
O.OE+00
O.OE+00
1.2E-06
6.4E-08
Deaths per Year
in this
Risk Range
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
3.2SE-05
1.07E-01
Deaths per Year in
this Risk Range or
Higher
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
3.25E-OS
1.07E-01
6-18
-------
l.OE-06. respectively (see Addendum). The unusual number of individuals at risk from this
plant is due primarily to the fact that a sizeable population resides proximally and in sectors
affected by the dominant wind direction. In effect, the general location of the 947
individuals in the highest estimated fisk range coincides with that of the maximally exposed
individual established at 2,500 meters west southwest of the plant. For Plant #651, the
estimated maximum individual lifetime fatal cancer risk of about l.OE-5 is applicable to all
of the 947 individuals.
For the 237 gas-fired plants, there are no individuals whose lifetime risk exceeds 1 in
100,000. Moreover, among the 1,855 individuals within the risk range of l.OE-05 to
l.OE-06, the contribution to risk is due primarily to emissions resulting from the combustion
of a secondary fuel.
Based on radionuclide emissions and plant-specific/site-specific data, CAP-93 also calculates
the maximum individual risk (MIR) for each of the 684 plants. Table 6-14 characterizes
those plants with the highest estimated MIR values expressed in lifetime fatal cancer risk.
There were a total of 17 plants for which the lifetime risk of fatal cancer to the MIR is
estimated to exceed l.OE-05. The highest MIR value of 3.0E-5 corresponds to a 5-unit coal-
fired facility that generated 3,340 megawatts of electricity in 1990. Of the 17 plants with the
highest MIR values, 11 are exclusively designated as coal-boilers; 3 facilities are identified as
exclusively oil-fired plants; and the remaining 3 plants are represented by a combination of
boilers, where coal is at least one of the designated primary fuels.
Because of limitations in the GENPOP computer code used for identifying locations of
individuals, the maximum individual risks (MIRs) shown for each plant should be viewed
with caution; errors of a few hundred meters in the location of individuals can result in an
over or under estimate of risk by factors of 2 or more. UARG re-estimated the risks for the
17 plants with the highest MIRs using refined population grids. Their results show lower
MIRs for the majority of these plants, but their highest MIR of 1E-5 is consistent with the
EPAs estimates. Thus, the EPA believes the GENPOP methodology is sufficiently accurate
to establish the magnitude of MIRs for all SGUs.
6-19
-------
Table 6-14. Plants with the Highest Estimated Maximum Individual Risk (MIR)"
Plant Name
Plant #222
Plant #247
Plant #60
Plant #301
Plant #251
Plant #406
Plant #256N
Plant #17
Plant #133
Plant #3 18
Plant #672
Plant #668
Plant #82
Plant #207
Plant #253
Plant #489
Plant #65 1
Annual Dose
to MIR
(mrem)
1 5
1 2
1 0
10
1.0
09
09
09
0.8
0.7
07
07
07
07
0.7
07
0.7
MIR'
3E-5
3E-5
2E-5
2E-5
2E-5
2E-5
2E-5
2E-5
2E-5
1E-5
1E-5
1E-5
1E-5
1E-5
1E-5
1-E5
1E-5
Coal-Fired
Units
5
4
4
2
4
4
3
4
6
8
7
3
4
MWe
3.340
900
3.160
750
1.540
2.777
1,728
1.135
1.100
1,965
2.304
2,052
1.872
Gas-Fired
Units
3
3
MWe
262
66
Oil-Fired
Units
2
2
2
6
MWe
1.112
804
558
372
Risk to the MIR is expressed in terms of the lifetime fatal cancer risk.
6.3.3 Limitations
Like all models, there are some limitations to the results generated by CAP-93. Risk
estimates for the 50-km populations are restricted to emissions from individual plants. These
calculated estimates, therefore, do not reflect the combined risks to individuals in cases
where individuals reside within a 50-km distance of more than 1 plant. A survey of data that
links the geographical locations of the 684 plants with the 196,000,000 individuals suggests
that (1) about one-third, or 65,000,000, of these individuals live within a 50-km radius of 5
"IT more plants and (2) about 23,000,000 live within 50 km of 10 or more plants.
6-20
-------
A cursory assessment, however, indicates that the impact of concurrent exposure from
multiple plants has a minor impact on risk estimates and risk distribution. This is due to the
fact the multiple plant exposures most frequently involve individuals identified in the "less
than l.OE-06" category. Since the average individual risk in this category is about 1E-07
(see Table 6-10), the additive risk from multiple exposures is considered to have a trivial
impact on the distribution of population risk as defined in this report.
The potential significance of exposure to the radionuclide emissions from multiple sources
(plants) is also evaluated in terms of maximum individual risks. For each of the 17 plants
where the lifetime fatal cancer risk to the most exposed individuals equals or exceeds 1E-5,
the impact of multiple source exposure has been evaluated with a bounding estimate as
follows:
1. identify all plants within 50 kilometers;
2. bound the maximum contribution from each plant identified in Step 1 by
identifying the maximum risk for that plant at the appropriate distance without
regard to direction;
3. sum the results of Step 2 for all plants within 50 kilometers to obtain the
maximum (bounding) risk; and
4. calculate the impact by computing the ratio of the maximum (bounding) risk to
the plant only risk.
The results of this evaluation are as follows:
PLANT #17 -
• Estimated MIR from Plant Emissions = 2E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 1E-7
Plant #268 @ -20 km = 4E-8
Plant #496 @ -35 km = 1E-7
Plant #277 @ -40 km = 2E-9
Plant #203 @ -40 km = 6E-9
• Maximum MIR/Plant MIR = 1.007
PLANT #60 -
• Estimated MIR from Plant Emissions = 3E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 5E-7
Plant #247 @ -40 km = 5E-8
• Maximum MIR/Plant MIR = 1.002
6-21
-------
PLANT #82 -
• Estimated MIR from Plant Emissions = 1E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 3E-7
Plant #297® -5 km =3E-7
Plant #91 @ -40 km = 5E-9
• Maximum MIR/Plant MIR =1.03
PLANT #133 -
• Estimated MIR from Plant Emissions = 2E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 0
no plants within 50 km
• Maximum MIR/Plant MIR = 1
PLANT #207 -
• Estimated MIR from Plant Emissions = 1E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 0
no plants within 50 km
• Maximum MIR/Plant MIR = 1
PLANT #222 -
• Estimated MIR from Plant Emissions = 3E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 5E-8
Plant #521 @ -50 km = IE-8
Plant #489 @ -50 km = 4E-8
• Maximum MIR/Plant MIR = 1.002
PLANT #247 -
• Estimated MIR from Plant Emissions = 3E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 9E-8
Plant #60 @ -40 km = 9E-8
• Maximum MIR/Plant MIR =1.003
PLANT #251 -
• Estimated MlR from Plant Emissions = 2E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 1E-8
Plant #573® ~50km= 1E-8
Plant #22® -50 km = 2E-9
• Maximum MIR/Plant MIR = 1.0006
PLANT #253 -
• Estimated MIR from Plant Emissions = 1E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 3E-8
Plant #546 @ -25 km = 2E-8
Plant #206 @ -50 km = 9E-9
i* Maximum MIR/Plant MIR = 1.003
6-22
-------
PLANT #256 -
• Estimated MIR from Plant Emissions = 2E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 1E-8
Plant #206 @ -15 km =-lE-8
• Maximum MIR/PIant MIR = 1.0005
PLANT #301 -
• Estimated MIR from Plant Emissions = 2E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 0
no plants within 50 km
• Maximum MIR/PIant MIR = 1
PLANT #318-
• Estimated MIR from Plant Emissions = IE-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 1E-7
Plant #581 @ -10 km = 1E-7
• Maximum MIR/PIant MIR =1.01
PLANT #406 -
• Estimated MIR from Plant Emissions = 2E-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 5E-8
Plant #227 @ -15 km = 5E-8
• Maximum MIR/PIant MIR = 1.003
PLANT #489 -
• Estimated MIR from Plant Emissions = IE-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 3E-7
Plant #521 @ -1 km = 2E-7
Plant #222 @ -50 km = 9E-8
Plant #183 @ -35 km = 2E-8
• Maximum MIR/PIant MIR =1.03
PLANT #651--
• Estimated MIR from Plant Emissions = IE-5 @ 2,500 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 4E-7
Plant #326 @ -20 km = 2E-7
Plant #276 @ -15 km = 2E-7
• Maximum MIR/PIant MIR = 1.04
6-23
-------
PLANT #668 -
• Estimated MIR from Plant Emissions = IE-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 1E-7
Plant #628® -5km = .2E-8
Plant #68 @ -10 km = 6E-8
Plant #85 @ -20km = 2E-8
Plant #451 @ -45 km = 4E-9
• Maximum MIR/Plant MIR = 1.01
PLANT #672 -
• Estimated MIR from Plant Emissions = IE-5 @ 200 m
• Estimated Contribution to MIR from Plants within 50 kilometers = 0
no plants within 50 km
• Maximum MIR/Plant MIR = 1
Based on this bounding methodology, the increases in risk to the maximally exposed
individuals at these 17 plants range from 0 to 4 percent. As the average increase is less than
1 percent, we conclude that not explicitly accounting for exposures to multiple plants does
not significantly affect the accuracy of the risk estimates presented here.
6-24
-------
REFERENCES
AEO, 1991, Energy Information Administration's Annual Energy Outlook (AEO)
Forecasting System.
AGA, 1991a, American Gas Association, 1991 Gas Facts. Department of Statistics,
Arlington, VA.
AGA, 19915, American Gas Association, "The Underground Storage of Gas in the United
States and Canada-1990 Update."
Anderson, W.t W.V. Mayneord, and R.C. Turner, 1954, "The Radon Content of the
Atmosphere," Nature 174, 424.
API, 1976, American Petroleum Institute, Primer of Oil and Gas Production. 3rd ed.,
Dallas, Texas.
API, 1992, American Petroleum Institute, Basic Petroleum Data Book. Petroleum Industry
Statistics, Vol. Xn, Number 2, Washington, D.C.
Armbrust, B.F. and P.K. Kurods, 1956, "On the Isotopic Concentration of Radium in
Petroleum Brines," Am. Geophis. Union Trans. (EOS) 37: 216.
Averitt, Paul, 1975, "Coal Resources of the United States, January 1, 1974," U.S.
Geological Survey, Geological Survey Bulletin 1412.
Ball, J.S., W.J. Winger, H.J. Hyden, C.A. Horr, and A.T. Meyers, 1960, "Metal Contents
of Twenty-Four Petroleums," Journal of Chemical and Engineering Data, 5: 4.
Barton, C.J., R.E. Moore, and P.S. Rohwer, 1973, "Contribution of Radon in Natural Gas
to the Natural Radioactivity Dose in Homes," ORNL-TM: 4154.
Barton, C.J., 1971, "Radon in air natural gas and houses," ORNL Central Files 71-5-48.
Beal, S.K. (S. Cohen & Associates), H.J. Chmelynski (Jack Faucett Associates), B.S.
Parks, and J. Hardin (U.S. Environmental Protection Agency), 1986, "Comparison of
AIRDOS-EPA Predictions of Ground-Level Airborne Radionuclide Concentrations to
Measured Values."
Beck, H., 1989, "Radiation Exposure Due to Fossil Fuel Combustion," Rad. Phvs. Chem.
34(2): 285.
R-l
-------
Beck, H.L., C.V. Gogolak, K.M. Miller, and W.M. Lowder, 1980, Proc. Natural Radiation
Environment - P. U.S. Department of Energy Report, CONF-780422, 2:1521.
Bedrosian, P.H., 1970. "Radiological Survey Around Power Plants Using Fossil Fuel,"
Southeastern Radiological Health Laboratory, Bureau of Radiological Health, Public
Health Service, Report EERL 71-3, November.
Bernhardt, D.E.. 1973, Memorandum to D.W. Hendricks, "Radon in Natural Gas," May 24.
Bloch, S., 1979, "Origin of Radium-Rich Brines - A Hypothesis," Oklahoma Geology Notes,
39: 177.
Bloch, S. and R.M. Key, 1980, "Modes of Formation of Anonymously High Radioactivity in
Oil-Fuel Brines, Geological Notes. The American Association of Petroleum
Geologists: 154.
Bunce, L.A. and F.W. Sattler, 1966, "Radon-222 in Natural Gas," Radiological Health Data
Reports, 2- 441-
Cannon, R.E., 1972, "Natural Gas Processing - 72," Oil and Gas Journal, July 10.
Case, L.C., 1970, Water Problems in Oil Production. 2nd ed., Petroleum Publishing Co.,
Tulsa, Oklahoma.
Coles, D.G., R.C., Ragaini, and J.M. Ondov, 1978, "Behaviour of Natural Radionuclides in
Western Coal-Fired Power Plants," Environ. Sci. Technol. 12(4) 442.
Corbett, J.O., 1980, "The Enhancement of Natural Radiation Dosage by Coal-Fired Power
Generation in the United Kingdom," UK Central Electricity Generating Board Report
RD/B/N4760.
de Lorenzi, 0. (Editor), 1948, Combustion Engineering. Combustion Engineering Company,
Inc., New York.
DeSantis, V. and I. Longo, 1984, "Coal Energy vs Nuclear Energy: A Comparison of the
Radiological Risks," Health Physics 46(1) 73.
DOE, 1991a, Department of Energy, Annual Outlook for U.S. Electric Power 1991. Energy
Information Administration, DOE/EIA A-0474(91).
DOE, 1991b, Department of Energy, Cost and Quality of Fuels for Electric Utility Plants
1990. Energy Information Administration, DOE/EIA-019(90).
R-2
-------
EPA, 1994, U.S. Environmental Protection Agency, "Estimating Radiogenic Cancer Risks,"
EPA-402-R-93-076, Office of Rdiation and Indoor Air, Washington, D.C.
EPA, 1989, U.S. Environmental Protection Agency, "Background Information Document for
Environmental Impact Statement on NESHAPS for Radionuclides," EPA-550/1-89-
006-1.
EPA, 1984, U.S. Environmental Protection Agency, "Radionuclides Background Information
Document for Final Rules, Volume n, EPA 520/1-84-022-2., Office of Radiation
Programs, Washington, D.C.
EPA, 1981, U.S. Environmental Protection Agency, "The Radiological Impact of Coal-Fired
Industrial Boilers," (Draft Report), Office of Radiation Programs, Washington, D.C.
EPA, 1980, U.S. Environmental Protection Agency, "Radiological Impact Caused by
Emissions of Radionuclides into Air in the United States - Preliminary Report," EPA
520/7-79-006, Office of Radiation Programs, Washington, D.C.
EPA, 1978, U.S. Environmental Protection Agency, "Advanced Oil Processing/Utilization
Environmental Engineering: EPA Program Status Report," EPA-600/7-78-007, Office
of Research and Development, Washington, D.C.
EPA, 1977, U.S. Environmental Protection Agency, "Assessment of Potential Radiological
Population Health Effects from Radon in Liquefied Petroleum Gas," EPA 520/1-75-
002, Office of Radiation Programs, Washington, D.C.
EPA, 1973, U.S. Environmental Protection Agency, "Assessment of Potential Radiological
Health Effects from Radon in Natural Gas," EPA 520/1-73-004, Office of Radiation
Programs, Washington, D.C.
Erickson, R.L., A.T. Myers, and C.A. Horr, 1954, "Association of Uranium and Other
Metals with Crude Oil, Asphalt, and Petroliferous Rock," AAPC Bull. 38: 2200.
Facer, J.F., Jr., 1979, "Uranium in Coal," U.S. Department of Energy Report, GJBX-
56(79), Washington, D.C.
Paul, H., G.B. Gott, C.E. Manger, J.W. Mython. and A.Y. Sakakura, 1952, "Radon and
Helium in Natural Gas," International Geol. Congress, 19th, Algiers, Comptes
rendus. Sec. 9, Pt. 9: 339.
Finkelman. R.B., 1981, "Modes of Occurrence of Trace Elements in Coal," U.S. Geological
Survey, Open File Report 81.
R-3
-------
Finkelman, R.B., 1989, "What We Don't Known About the Occurrence and Distribution of
the Trace Elements in Coal," presented at the 1989 International Conference on Trace
Elements in Coal, The journal of Coal Quality, Vol. 8, No 3-4, p 63.
Fries, B.A. and K.H. Kilgren, 1972. "Radon in Natural Gas and Gas Liquids," Chevron
Research Company, California, June.
Gesell, T.F., 1974, "Radiological Health Implications of Radon in Natural Gas Products -
Final Report," Institute of Environmental Health, The University of Texas Health
Science Center at Houston, April.
Gesell, T.F., C.A. Mullins, and N.D. Burgess, 1973, "Radiological Health Implication of
Radon-222 in Retail Liquified Petroleum Gas," in the 18th Annual Meeting of the
Health Physics Society, Miami, Florida, June 17-21.
GEH, 1966, Gas Engineering Handbook. The Industrial Press, 1st Edition.
Golightly, E.W. and F.O. Simon, 1989, "Methods for Sampling and Inorganic Analysis of
Coal," U.S. Geological Survey Bulletin 1893, Washington, D.C.
Gordon, J.A., 1968, "Interim Report of the Study of Public Health Aspects of Fossil Fuel
and Nuclear Power Plants," Southeastern Radiological Health Laboratory, August.
Gotchy, R.L., 1972, "Project Rulison final operational radioactivity report production tests,"
NVO-112.
Gott, G.B. and J.W. Hill, 1953. "Radioactivity in some Oil Fields of Southern Kansas,"
U.S. Geological Bulletin 988: 69.
Granger, H., 1978, The 1978 U.S.G.S. Branch of Uranium and Thorium, Course on
Uranium Geology. U.S. Geological Survey.
Harvey, R.D., et. al., 1983, "Mineral Matter and Trace Elements in the Herrin and
Springfield Coals, Illinois Basin Coal Field," Illinois State Geological Survey.
Haslam, R.T. and R.P. Russel, 1926, Fuels and Their Combustion. McGraw-Hill, Inc. New
York.
Jacobs, D.G., E.G. Struxness, M.J. Kelly, and C.R. Bowman, 1972, "Consideration of the
Radiological Impact from the Hypothetical Use of Contaminated Natural Gas from
Nuclearly Stimulated Reservoirs," Health Physics, 22: 429.
Jones, M.C. and R.L. Hardy, 1952, "Petroleum Ash Components and their Effect on
Refractories," Industrial and Engineering Chemistry. 44: 11.
R-4
-------
Kaczmarck, T.D. and A. Zervins, 1981, "Bioassay and Chemical Analysis for Hazardous
Materials in Residual Oils," Westinghouse Research and Development Center, PB82-
117078, Prepared for U.S. EPA.
Kaye, S.V., 1973, Memorandum to Distribution, Progress Report for October, Oak Ridge
National Laboratory, December 31.
Kent, 1950, Kent's Mechanical Engineer's Handbook (Power Volume). John Wiley & Sons,
Inc., New York.
Kinkley, M.L. and R.B. Neveril, 1976, "Capital and Operating Costs of Selected Air
Pollution Control Systems," Report PB 258 484, Research Triangle Park, NC,
Environmental Protection Agency.
Klein, D.H., A.W. Andren, and J.A. Carter, 1975, "Pathways of Thirty-Seven Trace
Elements Through Coal-Fired Power Plants," Environmental Science and Technology,
9: 10.
Kumar, K.S., R.E. Sommerlad, and P.L. Feldman, 1991, "Know All Impacts from
Switching Coals for CAA Compliance," Power Magazine, May 1991.
Langmuir, D. and J.S. Herman, 1980, "The Mobility of Thorium in Natural Waters at Low
Temperatures, Geochim. Cosmochim. Acta44:1753.
Lawn, C.J. (Editor), 1987, Principles of Combustion Engineering for Boilers. Academic
Press, London.
Lederer, C.M., J.M. Hollander, and I. Perlman, 1967, Table of Isotopes. Sixth Edition,
John Wiley and Sons, NY.
Legett, N. and T. Williams, 1979, National Gas Processing. The MITRE Corp., Metrek
Division', McLean, VA.
Leventhal, J.S. and R.B. Finkelman, 1987, "Uranium in the Vermillion Creek Core
Samples," U.S. Geological Survey, Geological Survey Bulletin 1059-A.
Maheras, S.J., P.D. Ritter, P.R. Leonard, and R. Moore, 1994, "Benchmarking of the CAP-
88 and GENII Computer Codes Using 1990 and 1991 Monitored Atmospheric
Releases from the Idaho National Engineering Laboratory," Health Phys, 67(5):509.
Martin. J.E.. E.D. Harward, and D.T. Oakley, 1969, "Comparison of Radioactivity from
Fossil Fuel and Nuclear Power Plants," in Hearings Before the Joint Committee on
Atomic Energy, Congress of the United States, First Session on Environmental
Effects of Producing Electric Power, Part I, U.S. Government Printing Office,
Washington, D.C.
R-5
-------
McBride, J.R. and D. Hill, 1969, "Off-site Radiological Surveillance for Project Gasbuggy,"
Radiological Health Data Reports, 10: 535.
Mcllvane Company, 1979, Mcllvane-Electrostatic Precipitator Manual. Northbrook, IL.
Mead, R.C., B.K. Post, and G.W. Brooks, 1986, Summary of Trace Emissions from, and
Recommendations of Risk Assessment Methodologies for. Coal and Oil Combustion
Sources. Radian No. 203-024-41, Radian Corporation, Research Triangle Park, NC.
NAS, 1980, National Academy of Sciences/National Research Council, "The Effects on
Populations of Exposure to Low Levels of Ionizing Radiation (BEIR III)," National
Academy Press, Washington, D.C.
National Petroleum Council, 1976, EOR, "An Analysis of the Potential for Enhanced Oil
Recovery from Known Fields in the United States; 1976-2000," National Petroleum
Council, Washington, DC.
NCRP, 1987b, National Council on Radiation Protection and Measurements, "Exposure of
the Population in he United States and Canada from Natural Background Radiation,"
NCRP Report No. 94.
NERC-LV, 1973, "Results of Sampling Natural Gas Wells in the Vicinity of Project
Gasbuggy," (NERC-LV, 539-9).
NERC-LV (S WRHL), 1970, "Off-site Radiological Safety and Resident Evacuation Program
for Project Rubison," SWRHL-94r.
Oakley, D.T. and R.I. Chissler, 1968, "Evaluation of the Study of Public Health Aspects of
Fossil Fuel and Nuclear Power Plants," Nuclear Facilities Section, ESCP, National
Center for Radiological Health.
Oglesby, S., 1970, "A Manual of Electrostatic Precipitator Technology, Part I -
Fundamentals," Southern Research Institute, Birmingham, AL, NTIS PB 196 380, p.
369.
Osmond, J.K. and J. B. Cowart, 1976, "The Theory and Uses of Natural Uranium Isotopic
Variations in Hydrology," Atomic Energy Review 14:621.
Papastefanou, C. and S. Charalambous, 1984, "On the Escaping Radioactivity from Coal
Power Plants (CPP)," Health Physics 46(2) 293.
Pasternack, B.S. and N.H. Harley, 1971, "Detection Limits for Radionuclides in die
Analysis of Multi-Component Gamma Ray Spectrometer Data," Nucl. Inst. and
Meth. 91:533.
R-6
-------
Pearson, E.S., and H.O. Hartley, 1966, Biometrika Tables for Statisticians. Vol. 1, p. 227,
Cambridge, England: Cambridge University Press.
Pierce, A.P., G.B. Gott, and J.W. Myton, 1964, "Uranium and Helium in the Panhandle
Gas Field of Texas and adjacent Areas," U.S. Geological Survey Professional Papers,
454-G.
Rahme, H., 1978, "Enhanced Oil Recovery," WP-13465, The MITRE Corp., Metrek
Division, McLean, VA.
Reeves, H., 1968, "Stellar Evolution and Nucleosynthesis," Gordon and Breach Science
Publishers, New York.
Roeck, D.R., T.C. Reavey, and J.M. Hardin, 1987, "Partitioning of Natural Radionuclides
in the Waste Streams of Coal-Fired Utilities," Health Physics 52(3) 311.
Rosholt, J.N., B.R. Doe, and M. Tatsumoto, 1966, "Evolution of the Isotopic Composition
of Uranium and Thorium in Soil Profiles," Geol. Soc. Am. Bull. 77:987.
Rosholt, J.N., Jr., 1959, National Radioactive Disequilibrium of the Uranium Series. U.S.
Geol. Survey Bull. 1084-A, Government Printing Office, Washington, D.C.
Rupinskas, R.L. and P.A. Hiller, 1991, "Switching to Low-Sulfur Coal: Facilities and
Performance," Power Engineering Magazine, December 1991.
Satranek, J., 1990, Santa Cruz County Environmental Health Services, Memo to Al Colli,
ORP, EPA: "Summary of PG&E Natural Gas Radon Test Results."
Satterly, J. and J.C. McLennan, 1918, "The Radioactivity of the Natural Gases of Canada,"
Trans. Royal Soc. Canada, HI: 153.
Schumacher, M.M., 1978, "Enhanced Oil Recovery: Secondary and Tertiary Methods,"
Noyes Data Corp., Park Ridge, New Jersey
Senftle, F.E.. L. Stieff, F. Cuttitta, and P.K. Kuroda, 1957, "Comparison of the Isotopic
Abundance of U-235 and U-238 and the Radium Activity Radios in Colorado Plateau
Uranium Ores," Geochim. Cosmochim. Acta 11:189.
Shannon, L.J., P.G. Gorman, and M. Reichel, 1971, "Paniculate Pollutant System Study, 0,
Fine Panicle Emissions," Report PB 203 521, Kansas City, MO, Midwest Research
Institute.
Shannon, R.H., 1982, Handbook of Coal-Based Electric Power Generation. Noyes
Publication, New Jersey.
R-7
-------
Smith, I.e.. T.L. Ferguson, and G.L. Carson, 1975, "Metals in New and Used Petroleum
Products and By-Products: Quantities and Consequences'" in: The Role of Trace
Metals in Petroleum, p. 123., T.F. Yen, ed., Ann Arbor Press, Ann Arbor, MI.
Smith, R.D., 1980, "The Trace Element Chemistry of Coal During Combustion and the
Emissions from Coal-Fired Plants," Prog. Energy Combust. Sci. 6, 523.
Smith, R.F. and J.M. Jackson, 1969, Variations in the U-234 Concentration of Nataural
Uranium. Report KY-581, Union Carbide Corp. Nuclear Div., Paducah Gaseous
Diffusion Plant, Paducah, Kentucky.
Sterlinski, S., 1969, "The Lower Limit of Detection for Very Short-Lived Radioisotopes
Used in Activation Analysis," Nucl. Inst. Meth. 68:341.
Tewalt, S.J., R.T. Hildebrand, and R.B. Finkelman, 1989, "Analyses of Lignites and
Associated Rocks from the Fort Union Region, North Dakota and Montana," U.S.
Geological Survey, Open File Report 89-285.
Torrey, S. (Editor), 1978, Trace Contaminants from Coal. Noyes Data Corporation, Park
Ridge, New Jersey.
UDI, 1992, Utilities Data Institute, Inc., Edison Electric Institute's (EEI) Power Statistics
Data Base, Washington, D.C.
UNSCEAR, 1982, United Nations Scientific Committee on the Effects of Atomic Radiation,
"Ionizing Radiation: Sources and Biological Effects," United Nations, NY.
U.S.AEC, 1973, U.S. Atomic Energy Commission, "Comparative Risk-Cost-Benefit Study
of Alternative Sources of Electrical Energy," WASH-1224, Division of Reactor
Research and Development.
Vitez, B, 1976, "Trace Elements in Flue Gases and Ah- Quality Criteria," Power
Engineering, January.
Wilbur, L.C. (Editor), 1982, Handbook of Energy Systems Engineering. John Wiley &
Sons, New York.
Wojcik, M., 1989, "Long-Term Measurements of Rn and Short-Lived Rn Daughter
Concentrations in Natural Gas From Distribution Line," Health Physics, 57: 989-991.
R-8
-------
APPENDIX A
TRENDS IN REFINERY TECHNOLOGY
AND THEIR IMPACT ON THE RADIONUCLIDE
CONTENT OF RESIDUAL FUEL OIL
-------
APPENDIX A
TRENDS IN REFINERY TECHNOLOGY AND THEIR IMPACT
ON THE RADIONUCLIDE CONTENT OF RESIDUAL FUEL OIL
INTRODUCTION
During the combustion of residual fuel oils by public utilities for electric power generation,
any metals contained in the fuel will be entrained in the flue gas and potentially emitted to
the environment. Among the trace metals are radionuclides belonging to the decay chains of
uranium and thorium.
The objective of this appendix is to review trends in refinery technologies that may
significantly impact the concentrations of radionuclides in residual fuel oils. In recent years,
three important factors have driven the changes in refinery technology. These changes were
the increase in the amount of heavy crude oil refined, changes in air pollution regulations,
and a lower demand for residual fuel oil. Heavier crude oils contain more metals, sulfur,
nitrogen, and oxygen than light crude oils. Because these materials poison expensive
refinery catalysts, a number of processes were developed and implemented to remove these
materials from the process stream. Air pollution regulations controlling the emission of
sulfur and nitrogen have also resulted in a lower metals emission rate because many
technologies that lower the sulfur and nitrogen content of residual oils also lower the metals
content. Because processing heavier crude oils generates a higher fraction of residual oil, the
lower demand for residual fuel oil forced refineries to implement additional upgrading
technology to create a marketable product.
BACKGROUND
Crude oil is a complex mixture of thousands of types of hydrocarbon molecules. During
refining, these molecules are separated into related groups by distillation (fractionation). The
high molecular weight molecules that are not distilled are collectively called the residual
fraction. Part of this residual fraction is sold as residual fuel oil, part is converted to solids
e.g., coke or asphalt, and part is further upgraded.
A-l
-------
Because almost all of the sulfur, nitrogen, oxygen, and metals found in crude oil are
associated with high molecular weight molecules, these elements are normally concentrated
in the residual fraction (Branthaver,.1987). Some of these materials may be entrained during
distillation, however, and appear in the higher boiling distillates from entrainment (Woodle
and Chandlev, 1952).
Crude oils contain varying amounts of metals. Vanadium and nickel are the most common
metals, with concentrations greater than 100 ppm in heavy crude oils. These metals are
almost entirely found in complex organic poiphyrine molecules (Filby and Van Berkel,
1987). Porphyrines are five-membered heterocyclic rings that are biological in origin and
are believed to have developed from the diagenesis of chlorophyll in phytoplankton.
Uranium is also found in crude and residual oils, although at much lower concentrations,
typically 1-10 ppb (Horr et al., 1961; Hyden, 1961). Uranium, however, is not thought to
be associated with the porphyrines and its concentration in crude oils is relatively
independent of the crude oil properties (Hyden, 1961). Thus, unlike other metals, the
concentrations of uranium and its decay daughters are not necessarily higher in heavy crude
oils than in light crude oils.
Even though uranium is not strongly associated with porphyrines, it will still be concentrated
in the residual fraction during refining. Nevertheless, the concentration of uranium in the
ash from combusted residual oils is of the same order of magnitude of that in the earth's
crust. This suggests that the radioactivity of ash from residual fuel may be similar to that of
windblown dust. The concentrations of nickel and vanadium in residual oil ash, however,
are several orders of magnitude higher than those in the earth's crust (Hyden, 1961).
Following the energy crisis in the 1970s, the amount of heavy crudes that were refined
increased significantly. Because heavy crude oils have a higher concentration of most
metals, sulfur, nitrogen, and oxygen compounds, this resulted in an increase in the average
metals content of the crude oils being processed and a higher metals concentration in the
residual fraction. However, because the uranium concentration in heavy crude oils is not
significantly higher than that in light crude oils, the trend toward refining heavier crude oils
probably did not significantly change the concentrations of radionuclides in the crude oils
refined.
A-2
-------
During the 1980s, there was an increased demand for gasoline and other transportation fuels
and a lower demand for residual fuel oil. As seen in Figure A-l. the total demand in the
U.S. for residual fuel oil fell by a factor of two from 1980 to 1990 (Twentieth Century
Petroleum Statistics, 1992). A major reason for the lower demand in residual fuel oil was
the increase in its cost relative to that of other energy sources. The decline in the demand
for residual fuel oil for electricity generation was offset by an increased use of coal and
nuclear power.
IB _
1.2E+9
1.0E+9
8.GE+8
1*
£ | 6.0E+8
•o o 4.0E+8
•
*
^
2.0E+8
O.OE+0
1970
1975
1980
YEAR
1985
1990
Figure A-l. Total U.S. Demand for Residual Fuel Oil
Because of the. lower demand for residual oil, refineries had to convert more of the residual
fraction to other products. As seen in Figure A-2, the fraction of the crude oil capacity
convened to residual oil decreased by about 40% during the same period (Basic Petroleum
Data Book, 1993). To increase the yield of transportation fuels and lower the fraction of
crude oil converted to residual fuel, new technologies were developed to upgrade part of the
residual fraction. As discussed below, some of these processes completely removed the
metals and others further concentrated the metals in the remaining residual fraction.
The presence of an increased metal concentration (primarily vanadium and nickel) created
significant problems with refinery processes to upgrade lower quality feedstocks to more
valuable products. Many of the upgrading processes use expensive catalysis to crack.
A-3
-------
•o
I
o
u
0.15
0.10
| 0.05
0.00
1970
1975 1980 1985
YEAR
1990
Figure A-2. Fraction of Crude Oil Converted to Residual Fuel Oil
reform, or otherwise alter the molecular structure of the hydrocarbons. Metals, sulfur,
nitrogen, and oxygen, if present, deposit on these catalysts and poison them. Thus,
refineries were forced to install processes to remove these materials from their feedstocks
before they could upgrade them.
An example 9f one upgrading processes that is easily poisoned by metals is catalytic
cracking. As seen in Figure A-3, the fraction of crude oil that is catalytically cracked
increased slightly from 1980 to 1990 (Oil and Gas Journal Annual Refinery Reports). This
increase in the fraction of crude oil that is catalytically cracked during a period when heavier
oils with higher metals contents were refined is an indication of changes in refinery
technology to remove metals from various feedstocks.
Mother factor impacting the concentration of metals in residual oil has been changes in air
pollution regulations. The most significant change in regulations affecting the metals content
of residual oil occurred with the passage of the Clean Air Act Amendments in 1970. This
act set a federal standard for the concentration of sulfur dioxide of 0.03 ppm and nitrogen
dioxide of 0.05 ppm. The objective of the act was for the air quality to meet the federal
standard by 1975 with no detrimental effects from air pollution within a reasonable time after
A-4
-------
2
u.
u
2
u
••
n
U
0.4
0.3
0.2
0.1
0.0
1970
1975
1980
YEAR
1985
1990
Figure A-3. Fraction of Crude Oil Catalytically Cracked
that time (Jorling, 1974). An extension of meeting these standards of up to two years could
be granted if the governors of .individual states indicated that the technology needed for
compliance did not exist and a good faith effort had been made to comply. These regulations
put significant pressure on refineries to lower the sulfur and nitrogen content of residual fuel
oils.
Although the Clean Air Act did not directly regulate the emission of metals or radioactive
isotopes from residual oil burned for fuel, many of the technologies that remove sulfur or
nitrogen also remove metals. Thus, complying with the air pollution regulations on the
emissions of sulfur and nitrogen also resulted in a reduction in concentration of metals
emitted.
A final factor suggesting changes in refinery technology is the total refinery capacity. As
seen in Figure A-4 (Oil and Gas Journal Annual Refinery Reports), the total refining capacity
of the U.S. has declined in recent years. The primary reason for the reduction in refinery
capacity was the increase in the amount of heavy crude oils refined in the presence of more
stringent air pollution regulations. Older refineries could not process the heavier crudes oils
and still comply with the air pollution regulations. The older refineries were closed and only
A-5
-------
•o
I
2.0E+7
1.8E+7
1.6E+7
1.4E+7
1.2E+7
« 1.0E+7
&
e 8.0E+6
•o
J 6.0E+6
4.0E+6
2.0E+6
O.OE+0
1970
I
1975
1980
YEAR
1985
.1990
Figure A-4. Refinery Capacity in United States
the more efficient refineries remained. Thus, the fuel oil samples obtained for the follow-up
study on radionuclide concentrations came from the better refineries that could meet the more
stringent environmental regulations.
DEMETALLIZATION TECHNOLOGY
A variety of processes have been developed to lower the metal content of the residual
fractions, with no single method being used by refineries in the U.S. Many petroleum
companies favor multiple methods, so no single technology can be considered typical or
standard. Further, most refineries consider their processes and operating parameters to be
Dprietary, so few data are available. Thus, few generalizations can be made about which
A-6
-------
pnethods are in use in the U.S. and what typical industry-wide levels of demetaliization might
be. Processes exist that can routinely remove greater than 90% of the metals from
feedstock. The ability of a particular process to remove metals, however, depends on the
composition of the feedstock (Dolbeatet al., 1987).
The primary processes used to lower the metals content of residual fuel oils include catalytic
hydrotreating, solvent processes, and coking (Speight, 1981; Le Page et al., 1992). Catalytic
hydrotreating can remove metals from the process stream, while solvent processes and coking
preferentially separate the heavy molecules that contain metals from those that do not, further
concentrating the metals in the metal-bearing fraction. The processing capacity in North
America for the residual fraction is about 27 million cubic meters per year for catalytic
hydrotreating, about 19 million cubic meters per year for solvent processes, and 87 million
cubic meters per year for coking (Le Page et al., 1992). Other processes are available to
lower the metals content, but are not widely used.
Catalytic Hvdrotreating
Hydrotreating is the selective hydrogenation of feedstock in the presence of excess hydrogen
over a catalyst at elevated temperatures and pressures. In this process, the hydrocarbon
molecules are selectively broken and hydrogen is added at the breaks. Depending on the
catalyst and the operating conditions used, the bond points broken can be selected to be
between sulfur, nitrogen, oxygen, or metals. The resulting products are primarily hydrogen
sulfide, ammonia, and water, respectively. Any remaining metal is bonded to the metals in
the catalyst. Other catalytic hydrotreatment processes selectively break carbon-carbon bonds
to reform the molecules. The catalysts used in hydrotreating for metal removal are different
from the more expensive ones used for other upgrading processes, e.g., catalytic cracking.
In 1965, the primary purpose of catalytic hydrotreating was to desulfurize fuel oils. In 1988,
the primary purpose was still desulfurization, but a greater emphasis was placed on
demetaliization to prevent the rapid deactivation of the downstream catalysts (Le Page et al.,
1992). This shift in emphasis has lead to changes in catalytic hydrotreating technology.
Over a dozen catalytic hydrotreating processes are available (Beaton et al., 1986; Billon
et al., 1988; Bridge et al., 1981; Eccles et al., 1982; Farjon et al., 1986; George et al.,
1985: Ito et al., 1986; Richardson et al., 1979; Shah et al., 1979; Sikonia, 1980; Takeuchi
et al., 1983; Van Zijll Langhout et al., 1980; Yanik et al., 1977). The suitability of any
A-7
-------
particular process depends on the nature of the feedstock and the amount and type of
treatment that is required.
Because the air pollution regulations are aimed primarily at reducing the emissions of sulfur,
many of the hydrotreating processes are optimized to remove sulfur, not metals. Although
those processes also remove metals, the fraction of metals removed relative to the fraction of
sulfur removed depends on the selectivity of the catalyst employed (Dautzenberg and De
Deken, 1987). The amount of metals removed by catalytic hydrotreating depends on the
composition of the feed, operating conditions, the catalyst, and the method used. Some
hydrotreating processes are more effective at removing metals than others and each process is
somewhat selective in how efficiently it removes each metal. For catalysts designed to
remove metals, up to 90% of the metals can be removed for a sulfur removal of less than
50%. On the other hand, catalysts designed to remove sulfur can remove up to 90%, while
removing less than 50% of the metals. Other catalysts are very effective for both sulfur and
metals removal. Because different refineries used different processes and an individual
refinery routinely changes catalysts, it is difficult to quantify the average reduction of metals
from feedstocks by means of catalytic hydrotreating.
One indication of changes in refinery technology is the fraction of crude oil that undergoes
some type of catalytic hydrotreating. The terminology used to describe catalytic
hydrotreating, however, varies, from author to author and company to company. The Oil and
Gas Journal describes three process involving hydrogenation of feedstocks over a catalyst:
catalytic hydrocracking, catalytic hydrorefming, and catalytic hydrotreating. A'short
description of these processes is given in Table A-l. Note that the definition of catalytic
hydrotreating used in the Oil and Gas Journal Annual Refining Reports is somewhat different
from that of most other authors and is different from that of this report. This distinction,
however, is not critical to the objectives or conclusions of this report.
Table A-l. Description of Catalytic Refinery Processes Using Hydrogen
Hydrocracking
Distillate upgrading
Residual upgrading
Lube-oil manufacturing
Hydrorefining
Residual desulfurizing
Heavy gas-oil desulfurizing
Cat-cracker feed prctreatmem
Middle distillates
Hydrotreating
Cat-reformer feed pretreaunent
Naphtha desulfurizing
Aromatic saturation
Straight-run distillates
Lube-oil polishing
A-8
-------
The fraction of residual oil catalytically treated with hydrogen (as a fraction of the crude oil
capacity of U.S. refineries) for these three processes is shown in Figure A-5. From this
figure it can be seen that the fraction of crude oil catalytically treated with hydrogen has
significantly increased since 1980, winch suggests a general reduction in the trace metal
content of petroleum products.
•o
s
I
o
•5
o
S
0.6
0.5
0.4
0.3
0.2
0.1
0.0
1970
-a Cat Hydrocracking Fraction
•*— Cat Hydrorefining Fraction
•*— Cat Hydrotreating Fraction
1975
1980
1985
1990
Figure A-5. Fraction of Crude Oil Treated with Hydrogen
Although the impact of catalytic processes is not readily quantified, steady improvements in
catalyst design have resulted in the ability to remove an ever increasing fraction of sulfur and
metals from feedstocks. A surrogate measure of the improving technology for catalytic
demetallization- of residual oils is the number of related patents issued. Patents reflect
interest by refineries in further technical developments in this area.
To determine the number of patents issued, two types of computerized patent searches were
;
conducted. The first type of patent search was to search key words in patent titles. The key
words searched were variations on "demetallization," "desulfurization," and "hydrotreating."
As seen in Figure A-6, a significant number of patents have been issued each year since 1969
(the starting date of computerized records). A sharp increase in the number of
desulfurization patents occurred in the 1970s following the passage of the Clean Air Act
Amendments of 1970. Similarly, a significant increase occurred in the 1980s in the number
A-9
-------
nua
I
50
40
30
20
10
n
uemetamzaiion raienis
Desulfurization Patents
Hydrotreating Patents
\
1 V ' /- ,
-
/ SN'
• s ' ,''• /
\4
V' I
. :v \
w
•~\
1960
1970 1980
Year
1990
Figure A-6. Number of Patents Issued in the US per Year by Topic
of demetallization and hydrotreating patents following the increase in the amount of heavy
crude oils refined.
The second type of patent search was also conducted using patent classification numbers to
estimate the number of patents that involved demetallization. It was found that patents are
classified in a variety of ways and with a variety of classification numbers. To simplify the
search, only one class and four subclasses were searched. The class searched was "Mineral
Oils: Processes and Products." The four subclasses are all associated with metals removal
during refining and are identified on Table A-2. The number of patents issued each year for
each of these subclasses is given in Figure A-7. From this figure, it is also clear that there
has been a significant increase in the number of related patents issued during the 1980s.
Table A-2. Patent Subclasses Searched
Subclass
208/25 1R
208/25 1H
208/253
208/252
Subclass
Refining:
Refining:
Refining:
Refining:
Description
Metal
Metal
Metal
Metal
Contaminant
Contaminant
Contaminant
Contaminant
Removal
Removal
Removal
Removal
Employing
with
with
Metal
Acid
Hydrogen
or Metal Compounds
A-10
-------
c
80
60
1 40
20
s.
Subclass 251R
Subclass 251H
Subclass 253
Subclass 252
Combined
1960
2000
Figure A-7. Number of Patents Issued in the US per Year by Subclass
Solvent Processes
Solvent extraction (solvent deasphalting) is commonly used to separate the asphaltic fractions
(suspended hydrocarbon solids) from the hydrocarbon liquids in residual oil. In this process,
the residual oil is mixed with a solvent, e.g., a light hydrocarbon or alcohol, and the
asphaltenes in the residual oil precipitate (Salazar, 1986; Bonilla et al., 1986; Le Page et al.,
1992). Because most of the contaminants in crude oil, e.g., metals, sulfur, nitrogen, and
oxygen, are associated with the asphaltenes, this precipitation concentrates them in the
precipitate, yielding a relatively clean hydrocarbon liquid.
Solvent extraction can be very effective at removing metals from the hydrocarbon stream.
For example, the vanadium content can be reduced by 95% by solvent precipitation with
n-pentane (Speight, 1981). The efficiency of solvent extraction, however, depends on the
solvent and the temperature used.
After separation, the asphaltic fraction, which contains most of the metals, can be used for
road construction materials, as liquid fuel oils, or as solid fuels like coke (Le Page et al.,
1992). If the asphaltic fraction from solvent extraction were used to produce residual fuel
A-ll
-------
oil, there is a potential for the metals in residual fuel to be increased. A limiting factor for
this scenario and the enhancement of metal concentrations in residual fuel oil is the
concurrent concentration of sulfur compounds and the need to meet Federal regulations.
The total volume of asphalt in U.S. refineries is shown in Figure A-8. From this figure, it
can be seen that there was not a significant difference between 1981 and 1993. Because of
the increase in the amount of heavy crude oil processed, which has a higher asphaltene
content, the figure indicates that a higher fraction of the residual oil has been upgraded to
other products.
1,000,000
800,000
600,000
"3
400,000
200,000
1970 1975 1980 1985
Year
1990
Figure A-8. Asphalt Capacity of U.S. Refineries
Coking converts heavy molecules, e.g., asphaltenes, to a solid at high temperature by
altering their molecular structure. This solidification concentrates the metals in the solid
(coke) and yields higher quality hydrocarbon liquids. About half of the coke produced
during refining is burned as a fuel and the other half is used to make calcined coke for the
petrochemical and chemical industries (Le Page et al., 1992).
A-12
-------
Coking is very effective in concentrating metals in the coke solid, with up to 95% of the
metals in the feed being included. Coking is significantly less effective in concentrating
sulfur, however, with only 30-40% ef the initial sulfur being deposited in the coke (Le Page
et al., 1992).
The amount of coke generated during refining and the fraction of the crude oil processing
capacity converted to coke is shown in Figure A-9. Both quantities increased significantly
between 1980 and 1990, indicating a potential increase in the amount of metals concentrated
in the coke solids.
1970
Coke Fraction
0.003
0.002
1980
Year
1990
0.000
u
•8
• 0.001 i.
Figure A-9. Amount of Coke Generated During Refining
SUMMARY AND CONCLUSIONS
There have been significant changes in refinery technology within the past two decades.
Two important factors that have driven the technological changes include an increase in the
amount of heavy crude oil refined and the enactment of more stringent air pollution
standards. Under EPA's Clean Air Act, pollution standards were primarily directed at
controlling the emission of sulfur and nitrogen pollutants.
A-13
-------
In context of more stringent pollution standards, the expanded use of heavy crude oils which
contain higher concentrations of sulfur, nitrogen, and metals demanded major changes in
refinery technology. Catalytic processes were introduced that not only served to remove
sulfur and nitrogen but also removed-metals which poisoned the costly catalysts. To protect
and preserve catalysts, metal-containing feedstock had to first be subjected to demetallization
processes.
In summary, recent refinery technologies aimed at meeting Federal emission standards for
sulfur and nitrogen contaminants have had the concurrent impact of reducing metallic
contaminants that include radionuclides. Barring any unforseen changes in (1) the global
supply and demand for petroleum products and (2) current Federal regulations, the
radionuclide content of residual fuel oil can be expected to remain at their current level as
estimated by the analysis of fuel oil samples described in Chapter 4 of this report.
A-14
-------
REFERENCES
Basic Petroleum Data Book: Petroleum Industry Statistics, American Petroleum Institute,
Vol. 13, No. 3, Sept. 1993.
Beaton et al., Oil and Gas Journal, p. 47, July 7, 1986.
Behling, U.H., private communication, SC&A, Inc., 1993.
Billon, A., Bousquet, J., and Rossarie, J., NPRA Annual Meeting, San Antonio, Texas,
March 1988. -
Bonilla, J.A., Feintuch, H.M., and Godino, R.L., "FW Solvent Deasphalting," in Handbook
of Petroleum Refining Processes. (Meyers, R.A., editor), McGraw Hill Book
Company > New York, 1986.
Branthaver, J.F., "Influence of Metal Complexes in Fossil Fuels on Industrial Operations,"
in Metal Complexes in Fossil Fuels: Geochemistry. Characterization, and Processing.
(Filby, R.H. and Branthaver, J.F., editors), American Chemical Society, Washington,
D.C., 1987.
Bridge, et al., Oil and Gas Journal, p. 85, January 19, 1981.
Dautzenberg, F.M. and De Deken, J.C., "Modes of Operation in Hydrodemetallization," in
Metal Complexes in Fossil Fuels: Geochemistry. Characterization, and Processing.
(Filby, R.H. and Branthaver, J.F., editors), American Chemical Society, Washington,
D.C., 1987.
Dolbear, G.E., Tang, A., and Moorehead, E.L., "Upgrading Studies with California,
Mexican, and Middle Eastern Heavy Oils," in Metal Complexes in Fossil Fuels:
Geochemistry. Characterization, and Processing. (Filby, R.H. and Branthaver, J.F.,
editors), American Chemical Society, Washington, D.C., 1987.
Eccles et al., Oil and Gas Journal, p. 121, April 12, 1982.
Farjon et al., 54th ACFAS Congress; Montreal, May 12-16, 1986.
Filby, R.H. and Van Berkel, G.J., "Geochemistry of Metal Complexes in Petroleum, Source
Rocks, and Coals: An Overview," in Metal Complexes in Fossil Fuels:
Geochemistry. CharaptgriTation. and Processing. (Filby, R.H. and Branthaver, J.F.,
editors), American Chemical Society, Washington, D. C., 1987.
George, S.E.. et al., UNITAR Third International Conference on Heavy Crude and Tar
Sands, Long Beach, California, July 23-31, 1985.
A-15
-------
Horr, C.A., Myers, A.T., and Dunton, P.J., "Uranium and Other Metals in Crude Oils: A.
Methods of Analysis for Uranium and Other Metals in Crude Oils with Data on
Reliability," U.S. Geological..Survey Bulletin No. 1100, 1961.
Hyden, H.J., "Uranium and Other Metals in Crude Oils: B. Distribution of Uranium and
Other. Metals in Crude Oils," U.S. Geological Survey Bulletin No. 1100, 1961.
Ito, Y., et al., AIChE National Meeting, New Orleans, April 1986.
Jorling, T., "The Federal Law of Air Pollution Control," in Federal Environmental Law.
Environmental Law Institute, West Publishing Co., St. Paul, MN, 1974.
Kaczmarck, T.D. and Zervins, A., "Bioassay and Chemical Analysis for Hazardous
Materials in Residual Oils," Westinghouse Research and Development Center,
PB82-117078, Prepared for the U.S. EPA, 1981.
Le Page, J.F., Chatila, S.G., and Davidson, M., Resid and Heavy Oil Processing. Editions
Technip, Paris, 1992.
Oil and Gas Journal Annual Refinery Reports, Oil and Gas Journal, Third Week each March,
1970-1989.
Richardson et al.. Oil and Gas J., p. 80, May 1979.
Salazar, J.R., "UOP Demex Process", in Handbook of Petroleum Refining Processes.
(Meyers, R.A., editor), McGraw Hill Book Company, New York, 1986.
Shah, G.N., et al., Hydrocarbon Processing, Vol. 58, No. 5, p. 103, 1979.
Sikonia, J.G., Hydrocarbon Processing, Vol. 59, No. 6, p. 73, 1980.
Speight, J.G., The Desulfiirization of Heavy Oils and Residua. Marcel Dekker, Inc., New
York, 1981.
Takeuchi, C., et al., Ind. Eng. Chem., Process Dews. Dev., Vol. 22, p. 236, 1983.
Twentieth Century Petroleum Statistics, 48th Edition, DeGolyer and MacNaughton, Dallas,
TX, 1992.
Van Zijll Langhout, W. C., et al., Oil and Gas Journal, p. 120, Dec. 1, 1980.
Wnodle, R.A. and Chandlev, W.B., Ind. Eng. Chem., Vol. 44, p. 2591, 1952.
fanik, S.J., et al., Hydrocarbon Processing, Vol. 56, No. 5, p. 97, 1977.
A-16
-------
APPENDIX B
PRIMARY DATA EXTRACTED FROM
RADIONUCLIDE ANALYSIS OF OIL SAMPLES
(WESTINGHOUSE STUDY)
-------
APPENDIX B
PRIMARY DATA EXTRACTED FROM RADIONUCLJDE ANALYSIS
OF OIL SAMPLES (WESTINGHOUSE STUDY)
Customer: Westinghouse Electnc Corporation
Address: 1310 Beulab Road
City: Pittsburgh, PA 15235
Samples Received: 12/08/78
Westinghouse Sample Number
Analysis
Radmm-226
Radium-228
Polonium-210
Lead-210
Uranium-234
Uranium-235
Uranium-238
Thorium-230
Thorium-232
Radon-222
#1
pCi/g
< 0.3
< 0.3
0.41 ± 0.29
0.33 ± 0.19
0.04 ± 0.03
0.01 ± 0.03
0.04 ± 0.03
0.04 ± 0.04
0.05 ± 0.04
< 0.3
#3
ECi/g
< 0.3
< 0.3
0.78 ± 0.34
1.62 ± 0.33
0.09 ± 0.04
< 0.02
0.07 ± 0.03
0.04 ± 0.04
0.05 ± 0.03
0.3 ±0.1
#4
pCi/g
< 0.3
0.4 ± 0.2
0.62 ± 0.32
3.07 ± 0.36
0.08 ± 0.06
< 0.04
0.05 ± 0.05
< 0.01
< 0.01
< 0.3
#5
pCi/g
< 0.3
< 0.3
0.77 ±
1.74 ±
0.04 ±
< 0.02
0.02 ±
< 0.02
0.02 ±
< 0.3
0.35
0.21
0.03
0.02
0.02
B-l
-------
Westinghouse Sample Number
Analysis
Radium-226
Radium-228
Polonium-210
Lead-210
Uranium-234
Uranium-235
Uranium-238
Thorium-230
Thorium-232
Radon-222
Analysis
Radium-226
Radium-228
Polonium-210
Lead-210
Uranium-234
Uranium-235
Uranium-238
Thorium-230
Thorium-232
Radon-222
#6
EQ/S
< 0.3
0.4 ± 0.3
0.47 ± 0.32
0.34 ± 0.27
0.08 ± 0.02
< 0.02
0.03 ± 0.02
0.02 ± 0.01
0.02 ± 0.01
< 0.3
no
pCi/g
< 0.3
0.3 ± 0.3
0.28 ± 0.26
0.84 ± 0.18
0.36 ± 0.06
< 0.01
0.24 ± 0.05
0.03 ± 0.02
0.05 ± 0.02
0.3 ± 0.2
#7
pCi/g
0.5 ± 0.4
0.3 ± 0.2
0.38 ± 0.28
1.48 ± 0.24
0.07 ± 0.03
< 0.02
0.06 ± 0.03
0.03 ± 0.01
< 0.01
< 0.3
#11
pCi/g
< 0.3
< 0.3
0.46 ± 0.30
0.59 ± 0.24
0.31 ± 0.12
< 0.07
0.18 ±0.11
< 0.01
< 0.01
< 0.3
18
pCi/g
< 0.3
0.2 ± 0.2
1.1 ±0.4
0.39 ± 0.24
0.25 ±0.12
< 0.08
0.09 ± 0.09
< 0.01
0.02 ± 0.01
< 0.3
#12
pCi/g
< 0.3
< 0.3
0.38 ± 0.27
0.78 ± 0.02
0.04 ± 0.04
< 0.02
0.02 ± 0.04
0.02 ± 0.01
< 0.01
< 0.3
#9
pCi/g
< 0.3
0.4 ± 0.2
0.87 ± 0.
0.42 ± 0.
0.06 ± 0.
< 0.02
0.03 ± 0.
< 0.01
< 0.01
< 0.3
#13
pCj/g
< 0.3
< 0.3
0.24 ± 0
< 0.18
0.14 ± 0
< 0.02
0.07 ± 0
< 0.02
0.01 ± 0
< 0.3
32
18
02
02
.29
.03
.02
.02
B-2
-------
Westinghouse Sample Number
#14 #15 #16 #17
Analysis pCi/g pCi/g pCi/g pCi/g
Radium-226 < 0.3 < 0.3 < 0.3 < 0.3
Radium-228 0.3 ±0.1 < 0.3 < 0.3 0.1 ± 0.3
Polonium-210 0.09 ± 0.28 0.11 ± 0.28 0.10 ± 0.22 0.05 ± 0.25
Lead-210 0.33 ± 0.12 0.42 ±0.11 0.76 ± 0.17 0.42 ± 0.16
Uranium-234 0.06 ± 0.03 0.04 ± 0.03 0.12 ± 0.08 0.04 ± 0.03
Uranium-235 < 0.02 < 0.02 < 0.06 < 0.02
Uranium-238 0.04 ± 0.03 0.01 ± 0.02 0.08 ± 0.08 0.01 ± 0.02
Thorium-230 < 0.02 < 0.01 < 0.02 < 0.02
Thorium-232 < 0.02 < 0.01 < 0.02 < 0.02
Radon-222 < 0.3 < 0.3 < 0.3 < 0.3
#19 #20 #21
Analysis pCi/g pCi/g PCi/g pCi/g
Radium-226 0.3 ± 0.4 < 0.3 < 0.3 < 0.3
Radium-228 < 0.3 < 0.3 0.3 ± 0.2 < 0.3
Polonium-210 0.65 ± 0-.35 0.30 ± 0.27 0.14 ± 0.24 < 0.24
Lead-210 1.30 ± 0.14 1.08 ± 0.15 0.21 ±0.13 < 0.10
Uranium-234 0.17 ± 0.09 0.25 ± 0.16 0.12 ± 0.11 0.33 ± 0.07
Uranium-235 0.05 ±D.06 < 0.12 < 0.05 0.01 ± 0.02
Uranium-238 0.17 ± 0.09 0.16 ± 0.15 0.05 ± 0.07 0.33 ± 0.07
Thorium-230 < 0.02 0.01 ± 0.04 < 0.01 < 0.06
Thorium-232 0.03 ± 0.02 0.01 ± 0.03 < 0.01 < 0.05
Radon-222 < 0.3 < 0.3 < 0.3 < 0.3
B-3
-------
Westinghouse Sample Number
Analysis
Radium-226
Radium-228
Polonium-210
Lead-210
Uranium-234
Uranium-235
Uranium-238
Thorium-230
Thorium-232
Radon-222
Analysis
Radium-226
Radium-228
Polonium-210
Lead-210
Uranium-234
Uramum-235
Uranium-238
Thorium-230
Thorium-232
Radon-222
#22
oCi/g
0.3 ± 0.3
0.5 ± 0.1
0.17 ±0.27
1.23 ± 0.21
0.11 ± 0.04
0.01 ± 0.02
0.05 ± 0.03
< 0.01
< 0.01
< 0.3
#26
oCi/e
1.2 ± 0.3
< 0.3
0.22 ± 0.22
0.36 ± 0.18
0.18 ± 0.05
< 0.02
0.08 ± 0.04
< 0.02
< 0.01
< 0.3
#23
fiCl/g
< 0.3
< 0.3
< 0.23
2.49 ± 0.23
0.13 ± 0.03
< 0.02
0.04 ± 0.02
< O.Q2
< 0.02
< 0.3
#27
pCi/g
1.3 ± 0.4
0.3 ± 0.2
0.43 ± 0.22
0.41 ± 0.10
0.05 ± 0.03
< 0.02
0.03 ± 0.03
0.07 ± 0.02
< 0.01
< 0.3
#24
pCi/g
0.7 ± 0.3
0.1 ± 0.3
0.12 ± 0.24
0.83 ±0.18
0.33 ± 0.06
0.01 ± 0.01
0.29 ± 0.05
0.02 ± 0.02
0.01 ± 0.02
0.3 ± 0.1
#28
oCi/g
< 0.3
< 0.3
0.12 ± 0.24
1.03 ± 0.20
0.02 ± 0.02
< 0.02
0.02 ± 0.02
0.04 ± 0.02
< 0.01
< 0.3
#25
ECi/g
0.3 ± 0.4
< 0.3
0.16 ±0.23
0.25 ± 0.13
1.06 ± 0.11
0.03 ± 0.02
1.07 ± 0.11
< 0.02
< 0.02
< 0.3
#29
pCi/e
< 0.3
< 0.3
0.09 ± 0.25
0.86 ± 0.15
0.27 ± 0.12
< 0.08
0.21 ± 0.12
0.02 ± 0.01
0.02 ± 0.01
< 0.3
B-4
-------
APPENDIX C
HEALTH RISKS ASSOCIATED WITH LOW DOSES
OF RADIATION
-------
APPENDIX C
HEALTH RISKS ASSOCIATED WITH LOW DOSES OF RADIATION
Medical scientists have been studying ionizing radiation and its effects on living systems for
more than eight decades. Much of our knowledge about the effects of radiation comes from
laboratory experiments in which various species of animals as well as cells/tissues taken from
animals and humans are, irradiated. Under laboratory conditions, many variables important
to the understanding of radiation health effects can be accurately controlled. Variables
important to understanding radiation health effects include (1) total radiation dose, (2) type
and quality of radiation, (3) dose rate, and (4) characteristics, such as age and sex, of the
exposed organism.
While laboratory experiments yield useful information regarding the mechanism(s) of
radiation injury and dose-response relationships, conclusions drawn from biological effects
observed in irradiated animals and in-vitro studies cannot always be applied to potential
health effects in humans. Thus, the most informative studies regarding potential health
effects involve various human population groups for which the radiation exposures were
inadvertent, intentional, or unavoidable. Epidemiologic studies are a critical tool in assessing
radiation risks, since they alone provide data directly applicable to humans. Interpreting
epidemiologic data is complex and requires an understanding of the disease process as it
affects large populations and the application of mathematical model(s) that most closely
reflects the observed data.
This Appendix provides the reader with a general understanding of how ionizing radiation
can cause biological damage and the basic units in which radiation exposure is expressed. In
support of this discussion, important health studies are identified along with the basic
scientific method for deriving estimates of health risks for low-dose, low-dose-rate
exposures.
Mechanisms and Determinants of Radiation-Induced Health Effects
When radiation passes through matter, it has sufficient energy to ionize atoms or molecules.
An atom is ionized when it gains sufficient energy for one or more of its electrons to be
C-l
-------
removed from the atom. Ionizing radiation is of two types: paniculate and electromagnetic
waves. Paniculate radiation may be either electrically charged (alpha, beta, proton) or have
no charge (neutron). Ionizing electromagnetic radiation in the form of x-rays or gamma rays
shares a limited number of propenies with other non-ionizing electromagnetic waves (e.g..
microwaves, visible light, and radiowaves) but has considerably more energy.
Both electromagnetic and paniculate radiation can be released by radioactive elements
(radionuclides). Each radioisotope is inherently unstable and undergoes radioactive decay.
The energy and type(s) of radiation emitted and the rate of decay (i.e., half-life) are unique
to each radionuclide. Some radionuclides decay directly into stable non-radioactive elements.
Others, including the natural radionuclides contained in fossil fuels, are pan of a predictable
decay chain of successive radionuclides (decay daughters), each with its own distinctive half-
life and type of radiation emission.
Linear Energy Transfer (T.F.T) Of critical importance to the understanding of radiation
health effects is the distribution of ionization events along the path of impinging radiation.
The rate of energy dissipation by single ionization event is referred to as linear energy
transfer (LET) and is affected by the mass, charge, and energy of the radiation. X-rays and
gamma rays may penetrate deeply into an absorbing medium and do not directly produce
ionization. It is only through their interaction and production of energized electrons (i.e.,
photo-electrons or compton electrons) that sparse tracks of ionization result. Thus, the LET
value of x- or gamma rays is essentially similar to those of beta particles of equivalent
energy. X-rays, gammas, and beta particles produce sparse ionization tracks and are,
therefore, classified as low-LET types of radiation. Other paniculate radiation such as alpha
particles cause interactions in which large a amount of energy is dissipated within a shon
distance. Hence, alpha panicles are considered high-LET radiation.
Units of Radiation Dose. The relative effectiveness of different types of radiation in
producing biological harm is primarily due to differences in the linear energy transfer. In
other words, a given amount of tissue-absorbed energy from gamma radiation does not
necessarily have the same biological effect as the identical amount of absorbed energy from
alpha panicles. To characterize this difference, the concept of relative biological
effectiveness (RBE) has been introduced. RBE values have been tabulated for a variety of
LET values and biological endpoints as a basis for deriving risk per unit dose of high-LET
radiation. In general, low-LET X-rays, gamma rays, and beta particles are assigned an RBE
C-2
-------
value of unity (1). Depending on the biological endpoint under investigation, RBE values for
alpha radiation of 10 to 20 are commonly used.
Because the RBE for a constant radiation type (i.e., source) may vary based on die specific
biological endpoint under investigation, another term called "quality factor" (Q) has been
used to simplify decisions regarding dose limits and radiation protection standards.
Conceptually, Q has a meaning similar to RBE but is used for a different purpose. The
application of the quality factor, Q, is explained below.
Units of Radiation Dose. In radiation protection, the absorbed dose, commonly expressed in
the unit of rad, is converted to the unit rem by means of the quality factor (Q), which
accounts for the fact that high-LET radiation produce more biological harm than low-LET
radiation for a given amount of absorbed energy (rad). Thus, the quality factor, Q, is
defined as an LET-dependent factor by which the absorbed dose is multiplied to obtain a
quantity that corresponds more closely to the degree of biological effect produced by X-rays
or low-energy gamma rays. Thus, while a 1 rad dose of gamma radiation results in a dose
equivalent of 1 rem, a 1 rad dose of high-LET alpha radiation may cause a dose equivalent
of up to 10 to 20 rem. Because these units represent relatively large quantities, for purposes
of radiation protection, they are often expressed in millirad (mrad) or millirem (mrem).
Another unit that is commonly used in population risk assessment is person-rem. Person-rein
is a unit that measures collective dose to a given group of individuals. It is calculated by
adding the doses received by each member of the exposed population. For example, if
10,000 individuals each receive a dose of 100 mrem (i.e., 0.1 rem), their collective dose
would be 1,000 person-rem.
To gain a sense of perspective regarding the magnitude of a rem or mrem, the average
individual living in the United States receives about 300 mrem (0.3 rem) annually from
natural background radiation.
The Role of Dose Rate. Radiation health effects are not only influenced by the dose of
radiation but also by the rate at which the dose is delivered. In general, the risk of health
effects for a constant radiation dose tends to decrease if the dose rate decreases. The
reduced risk from a dose delivered at low dose rates results from the interaction of several
factors most notably (1) the reduced concentration of ionization events in time and space
within critical sub-cellular volumes, (2) enhanced cellular repair, and (3) compensatory cell
division.
C-3
-------
Based on experiments with laboratory animals exposed to high doses at high dose rates, dose
and dose-rate effectiveness factors (DDREF) of 2 to 10 have been postulated as applicable to
estimate the reduced risk for low dose/low dose-rate exposures. Based on human data, the
range of DDREF values can not exclude 1 as a lower limit but may assume a value as high
as 5.
Standardization of Risk from Internal Versus External Radiation. The release of
radionuclides into the environment may result in exposure from multiple pathways that
include (1) external radiation from activity suspended in air or deposited on the ground and
(2) internal exposure from the inhalation of airborne contaminants or ingestion of
contaminated water and foodstuff. Although the potential health risks are essentially
independent of whether a given dose was internal or external, the assessment of dose and
associated risks are considerably more complex for internal exposures.
While exposure to an external source of penetrating radiation is generally assumed to result
in a near-uniform dose to all body tissues, the dose from an internal exposure is likely to be
non-uniform and difficult to measure. The dose resulting from the inhalation or ingestion or
adioactive materials must be calculated on the basis of (1) the quantity of radioactive
Jaterial in the body, (2) the distribution and retention of the radioactive material within the
body, (3) the type and energy of radiation emitted, and (4) the fraction of energy absorbed
by individual tissues.
The dose to individual tissues is commonly referred to as the dose equivalent (HT), which is
the product of the absorbed dose (rad or mrad) and the quality factor, Q.
For the purpose of relating internal exposure to risk, another important units is the effective
dose equivalent (EDE). EDE is the weighted sum of doses to all irradiated issues and is
defined as
EDE = EHTWT = HWB
where HT is the dose equivalent to tissue T, WT is a weighting factor', and HWB is a uniform
dose to the whole body. The weighting factor, WT, is normalized so that
EWT = 1
C-4
-------
and corresponds to the fractional contribution of tissue T to the total risk of health effects
when the entire body is uniformly irradiated. From human studies, for example, WT for the
lungs has been established at 0. 12. ..The relationship of these specialized dose units are best
illustrated by example. Let us assume an individual is exposed to airborne radioiodine
(1-131) that results in a lung dose equivalent of about 100 mrem. By multiplying this dose
equivalent with the lung weighting factor of 0.12, we obtain the EDE value of 12 mrem:
EDE - W
12 mremEDE - (0.12)^ (100 mrem) H
•W BMf
In effect, this equation defines a relationship of risk that states the following: A confined
dose of 100 mrem to the lung has the equivalent health risk as a dose of 12 mrem delivered
uniformly to the whole body.
In some instances, a dose may be due to both uniform and non-uniform irradiation based on
combined exposure to external and internal sources. The radiation risk for such conditions is
defined by the total effective dose equivalent (TEDE):
TEDE = HWB + £ HT WT
Human studies described below involve either irradiation of specific organs or tissues only
(as in many medical procedures) or near uniform irradiation of the entire body (Japanese A-
bomb survivors). The effective dose equivalent provides a means for comparing risk in these
differing circumstances.
Low-Dose Radiation Health Effects
Dose-Response Relationship. As a general rule, potential health effects associated with low
doses of radiation may not appear for years or even decades. Such delayed effects are
termed "stochastic" and result from specific changes that occur in just a few cells or a single
cell. Although these selective cellular changes occur rarely, when they do, the 'altered cell
may develop into cancer. Among the stochastic effects that have been associated with
radiation exposure, medical scientists consider cancer induction the primary health effect of
concern for low-dose ionizing radiation.
C-5
-------
A key characteristic of a stochastic effect is that the severity of the effect is not dose
dependent. However, the probability that a stochastic event may occur is directly influenced
by the absorbed dose of radiation and. modified by dose rate and type and quality of
radiation.
A familiar example of a stochastic effect is that of smoking and lung cancer. Indisputably,
cigarette smoking is a direct cause of human lung cancer, but not all smokers develop lung
cancer. Moreover, lung cancer may also be observed in some non-smokers. It is important
to note that the "severity" of a lung cancer is independent of whether the individual was a
heavy smoker, light smoker, or non-smoker. Thus, the causal relationship of cigarette
smoking and lung cancer was established when a higher incidence rate of lung cancer was
observed among smokers than among non-smokers. The level of increase was found to
increase with the amount and duration of cigarette smoking. While large differences in lung
cancer rates were readily observable when heavy smokers were compared to non-smokers,
these differences diminished to indistinguishable levels for very light smokers or individuals
who had smoked only for a very short time.
A similar relationship exists between radiation exposure and several types of stochastic
ffects, inclusive of cancer. For small doses of radiation, the likelihood that even a single
cell will undergo a selective alteration, which leads to a cancer is low. Even for very large
doses of radiation, radiation-induced cancers can be observed only as small increases above
the spontaneous incidence that is observable in the normal population. Spontaneous cancer
incidence refers to the frequency of observed cancers in the general population from all
causes.
The American Cancer Society estimates that one person in three (about 30%) can expect to
develop cancer some time during their lives and about half of these people will eventually die
of cancer. Contrary to popular belief, cancer is not a new disease brought on by
industrialization. Studies indicate the prevalence of cancer within the general population
largely involves cancer-causing agents of natural origin and are affected by major risk factors
including sex of the individual, diet, life style, personal habits, and occupation. However,
among the most important risk factors is age of the individual, particularly after the age of
40. The individual risk of cancer in the general population starts to increase significantly at
about age 40 and increases approximately 100 fold by age 80. (Thus, the general trend in
increased cancer incidence over the last few decades among industrialized nations are largely
C-6
-------
the result of a corresponding increase in average life span among individuals.) The
proportion of cancers attributed to various factors is shown in Table C-l.
•f
Table C-l. Cancer Deaths Attributed to Various Factors
Factor
Tobacco
Alcohol
Diet
Food additives
Reproductive and sexual behavior
Occupation
Pollution
Industrial products
Medicines and medical procedures
Geophysical factors (sunlight, natural radiation, etc.)
Infection
Unknown
Percent of All Cancer Deaths
Best
Estimate
30
3
35
<1
7
4
2
<1
1
3
10?
7
Range of Acceptable
Estimates
25-40
2-4
10-70
0.5-2
1 - 13
2-8
<1 -5
<1 -2
0.5-3
2 -4
1 -?
7
Source: Doll 1984
Major Epidemiologic Studies and the Derivation of Cancer Risk Values
The current method for estimating radiation risks relies on past human studies where cancers
were observed at a higher incidence rate among exposed individuals than one would normally
expect to occur spontaneously. The excess cancer incidence increases with dose and displays
a predictable time course following exposure (Figure C-l).
The time from exposure (XJ to a later time period when radiation cancers are observed to be
increased is defined by the minimal latent period, 1. The way in which incidence increases
thereafter and the length of time in which the excess cancer incidence persists depends
primarily on the type of cancer and the age of the individuals at time of exposure. For some
cancers such as leukemia, the minimum latency period may be as short as two years. The
minimum latency period for most solid cancers is generally assumed to be about ten years.
For the estimation of cancer risk, the most relevant studies involve population groups
exposed to relatively high doses of radiation. When individual doses are considerably less
than 50,000 mrem (50 rem), the potential number of excess cancers diminish to levels that
are difficult to statistically identify in the presence of naturally occurring cancers.
C-7
-------
0)
CJ
Q)
T3
o
Exposed
population
Radiogenic
increment
Spontaneous
incidence
X. + I
Time
Figure C-l. Superimposition of Radiogenic Effect on Spontaneous Incidence
(from NAS 1980, BEIR IE)
There have been many studies of humans exposed to radiation under a variety of conditions.
These studies are generally grouped by the circumstances in which radiation exposure
occurred. Principal categories include:
• atomic bomb survivors,
• medical exposures,
• fallout from experimental weapons testing,
• occupational exposures, and
• others.
Not all of these studies, however, have contributed equally to our understanding of the
cancer-causing effect of radiation. To be useful in deriving estimates of risk, a study must
first demonstrate that any observed excess cancers in the exposed population are statistically
jnificant (i.e., the excess cancers were not due to chance variation) and, second, that
radiation exposure was the probable cause of the observed excess.
C-8
-------
In order to sarisfy the first requirement, a study is likely to require a large number of people
exposed to relatively large doses and who are studied for many years following exposure.
The second requirement is met wheoTall confounding factors have been either accounted for
or eliminated by proper selection of the unexposed control group. Furthermore, it is
important that individual exposure among the exposed population is known or can be
reasonably esablished. Verification of a causal relationship between radiation and an
observed excess cancer rate is normally demonstrated by a positive dose-response relationship
in which the excess cancer rate increases proportionately to radiation dose among subsets of
the exposed population.
The most intensely studied human population are the Japanese survivors of the 1945 atomic
bombing of Hiroshima and Nagasaki. Survivors in the two cities were exposed to the
immediate external radiation produced by bomb blast and to a lesser extent subsequent
internal/external exposure from fallout. Of the approximate 76,000 survivors for whom dose
was estimated, about 34,272 were so far from the hypocenters that their radiation doses were
less than 500 mrem. The dose distribution among the remaining 41,719, along with
summary cancer data, are shown in Table C-2.
Table C-2. Observed Cancer Deaths and Number of Expected Cancer Deaths
Among A-Bomb Survivors
Dose
(Rem)
0
1 -10
10-50
50- 100
100-200
200 +
Total
Appro*.
No of
Survivors
34.272
23.321
11.500
3,500
2.000
1.000
76.000
Leukemia
Excess
Observed Expected (No.) (%)
58 88-0
38 61-0
32 20 12 38
19 6 13 68
23 3 20 87
30 2 28 93
202 122 80 40
Non-Leukemia
Excess
Observed Expected (No.) (%)
2,443 2.593 - 0
1.655 1.688 - 0
927 866 61 7
329 273 56 17
218 147 71 33
132 68 64 48
5.734 5.474 260 5
Source: Shimiie 1987
C-9
-------
Ankylosing Spondylitis. This study of 14,106 patients treated with radiotherapy to
the spine for the debilitating effects of ankylosing spondylitis (arthritis of the bone)
between 1935 and 1954 has been ongoing for 48 years (Darby 1987). Because the
therapeutic treatment was aifhed at the spine, multiple other tissues/organs received
substantial doses. Statistically significant excess cancers have been reported among
these patients for leukemia and at several anatomical sites, including the esophagus,
pancreas, breast, lung, prostate, spinal cord, bone, and lymph nodes.
Cancer of the Cervix. About 150,000 women treated for cervical cancer were
studied. An estimated 70 percent of these patients were treated with radium implants
which result in substantial exposure of a number of other tissues and organs. Among
these patients, leukemia and cancers of the urinary bladder, breast, kidney, stomach,
and rectum have been observed in excess (Boice 1987).
Postpartum Mastitis (inflammation of the breasts following child birth). This survey
covers 601 women treated with radiotherapy during the 1940s and 1950s. The
women received doses ranging from 60 rad to 1,400 rad with an average dose of 247
rad to the breasts. The 115 cases of observed breast cancer among the women was
several times higher than the expected number of cases for controls (Shore 1986).
Tuberculosis and Fluoroscopic Examinations. Women who received multiple chest
fluoroscopic examinations to monitor pneurnothorax treatment (the presence of air or
gas in the pleura! cavity) of tuberculosis in the 1930s and 1940s in the United States
have been studied. The average dose to breast tissues was estimated at 150 rad
(Boice 1981). A total of 74 breast cancers have been reported among these 1,742
patients, which is significantly higher than expected (Hrubec 1989).
Thymus Enlargement. In the 1930s and 1940s, it was common for children to
receive intense radiation therapy to shrink enlarged thymus glands. The Rochester
Thymus Study investigated 2,652 irradiated subjects, who were less than 1 year of
age at time of exposure, and compared them to 4,823 of their non-irradiated siblings
(Shore 1985). There were 37 cases of thyroid cancer among the 2,652 irradiation
subjects and only 1 observed thyroid cancer among the 4,823 non-irradiated siblings.
Tinea Capitis. A total of 10,834 subjects who were irradiated for ringworm of the
scalp during their childhood have been studied. Although the average dose to the
thyroid among these children is estimated at less than 10 rad, there were 39 cases of
thyroid cancer among the 10,834 irradiated subjects versus 16 cases among the
16,226 unirradiated control subjects (Ron 1984).
Individuals irradiated in childhood for ringworm of the scalp have also been assessed
for the development of skin cancer (Shore 1984). In a study of 2,226 patients for
whom the average absorbed dose to the scalp was estimated at 450 rad, a total of 80
basal cell carcinoma were observed among 41 exposed individuals. Among 1,387
C-ll
-------
• With exception of breast cancer, low-LET radiogenic cancer risks are assumed to be
reduced by a dose and dose-rate effectiveness factor (DDREF) of 2 at low doses and
dose rates. (EPA's DDREF for breast cancer is 1.)
• For high-LET (alpha panicle), risks are presumed to increase linearly with dose and
to be independent of dose rate (i.e., DDREF value of 1).
• For high-LET, the relative biological effectiveness (RBE) factor of 20 is applied for
all cancers except leukemia and breast cancer, which are assigned RBE factors of 1
and 10, respectively.
• For low-dose, uniform external whole body exposure conditions, the calculated
lifetime risk of a premature cancer death is 5.1 cancers per 10,000 person-rem.
• Using these parameters and the most current biokinetic models, EPA has also derived
cancer risk coefficients for individual radionuclides under constant exposure rate
conditions for each of the four major exposure pathways (EPA, 1994).
C-13
-------
Reimer, R.R., R. Hoover, J.F. Fraumeni, Jr., and R.C. Young, 1978, "Second Primary
Neoplasms following Ovarian Cancer," J. Nail. Cancer Inst. 61:1195.
Ron, E. and B. Modan, 1984, "Thyfoid and Other Neoplasms Following Childhood Scalp
Irradiation," in Radiation Carcinogenesis: Epidemiology and Biological Significance.
J.D. Boice and J.F. Fraumeni, Editors, Raven Press, New York, p 139.
Shimizu, Y., H. Kato, W.J. Schull, D.L. Preston, S. Fujita, and D.A. Pierce, 1987, "Life
Span Study Report 11. Part 1. Comparison of Risk Coefficients for Site-Specific
Cancer Mortality Based on the DS86 and T65DR Shielded Kerma and Organ Doses,"
Technical Report RERF TR 12-87, Hiroshima: Radiation Effect Research
Foundation.
Shore, R.E., E. Woodard, N. Hildreth, P. Dvoretshy, L. Hempelmann, and B. Pastemack,
1986, "Breast Cancer among Women Given X-ray Therapy for Acute Postpartum
Mastitis," J. Natl. Cancer Inst. 77:689.
Shore, R.E., E. Woodward, N. Hildreth. P. Dvoretsky, and L. Hemplemann, 1985,
"Thyroid Tumors Following Thymus Irradiation," /. Natl. Cancer Inst. 74:1177.
Shore, R.E., R.E. Albeit, M. Reed, N. Harley, and B.S. Pastemack, 1984, "Skin Cancer
Incidence among Children Irradiated for Ringworm of the Scalp," Rod. Res. 100:192.
Tucker, M.A., A.T. Meadows, J.D. Boice, Jr., R.N. Hoover, and J.F. Fraumeni, Jr., 1984,
"Cancer Risk following Treatment of Childhood Cancer," in Radiation
Carcinogenesis: Epidemiology and Biological Significance. J.D. Boice and J.F.
Fraumeni, Editors, New York: Raven Press, p 211.
C-15
-------
APPENDIX D
AN ANALYSIS OF UNCERTAINTIES IN RISKS
FOR TWO SELECTED SITES
-------
APPENDIX D
AN ANALYSIS OF UNCERTAINTIES IN RISKS FOR SOME SELECTED SITES
D.I INTRODUCTION
Chapter 6 presents estimates of the risks attributable to radionuclides released to the air from
the 1748 fossil-fuel-fired boilers located at the 684 plant sites evaluated in this study. The
risks were estimated using data characterizing airborne emissions and the models and
assumptions described in Chapters 5 and 6 and Appendix C.
Rather than using mathematical models to assess impacts, one would prefer to measure the
actual impacts directly; i.e., radionuclide concentrations and radiation fields in the
environment and radionuclide concentrations in the various organs of the exposed
populations. However, this is seldom possible because the radionuclide releases do not
generally result in detectable levels of radionuclides in the environment or in the exposed
members of the population. In addition, any additional theoretical cancers that may be
attributable to radionuclide exposures cannot be detected in the presence of the large numbers
of cancers^endemic~in~ahy population. Accordingly, the actual or potential impacts of the
emissions must be estimated using mathematical models.
The risk estimates for each facility are presented as discrete values. Each of these calculated
values is an expression of impact on an individual or small group of individuals or on a
population as a whole. These values are intended to be reasonable best estimates of risk; that
is, to not intentionally underestimate or overestimate risks and be of sufficient accuracy to
support decision making. However, while each facility is unique, the models used are
generalizations and simplifications of the processes which result in exposure and risk. In
addition, the ability to model the processes is limited by the availability of data characterizing
each site and the understanding of the processes. As a result, the estimates of risk have a
considerable degree of uncertainty.
Because of these uncertainties, the values presented are of more use to decision makers when
there is some characterization of their uncertainty. For example, a calculated risk may be
small, e.g.. 10* lifetime risk of cancer for an individual. If the uncertainties in these
numbers are several orders of magnitude, the real risks of this source of emission may in fact
be higher than another source of emission which has a calculated lifetime risk of cancer for
D-l
-------
an individual of 10'3 but a small degree of uncertainty. Alternatively, a risk of 10"2
calculated using upper bound techniques may appear to represent an unacceptable risk. This
situation may occur when, due to liguted information and uncertainty in the calculational
parameters, conservative assumptions are used throughout the calculation in order to ensure
that the risks are not underestimated. This can result in a risk estimate that is near the upper
limit of what is plausible because it is based on a very unlikely combination of conservative
assumptions. However, a central estimate of the risk may be several orders of magnitude
smaller.
Quantitative uncertainty analysis can provide results that indicate the likelihood of realizing
different risk levels across the range of uncertainty. This type of information is very useful
for incorporating acceptable and reasonable confidence levels into decisions.
Due to the large numbers of facilities it is not possible to perform a quantitative uncertainty
analysis of each facility, nor is it necessary. In this appendix, uncertainties in the risk
estimates for two of the seventeen facilities where the estimated maximum individual lifetime
risk (MIR) for a lifetime exposure is greater than 10"5 are presented. At both sites, direct
radiation exposure from deposition is the principal pathway. However, the uncertainty
analysis considers risks at both the location of maximum total risk and the location of
maximum risk due to inhalation. As a result, a discussion of uncertainties in the risks for
these two sites has broad applicability to the critical pathways of exposure for all sites
addressed in this report.
This appendix presents a discussion of the uncertainty in estimating the best estimate of the
lifetime fatal cancer risk to members of the general population that reside at locations which
tend to maximize risk. These individuals are referred to as "maximum individuals." A
detailed description of the mathematical models and calculational assumptions used in the
uncertainty analysis is provided in EPA89.
D.2 GENERAL APPROACH
D.2.1 Application of Uncertainty Analysis to Environmental Risk Assessment
The use of quantitative uncertainty analysis to address environmental risks became
widespread following the Reactor Safety Study (NRC75), and in 1984 was recommended by
the Agency in support of environmental risk assessments (EPA84). Since then, the Agency
D-2
-------
has employed quantitative and semiquantitative uncertainty analysis in support of is decision
making process.
Quantitative uncertainty analysis results in a range of values of impact rather than a single
discrete value by using a range of values for the calculational input parameters. In this way,
the impacts of a given technological activity can be bounded and different technologies can
be intercompared. In cases where probability distributions can be assigned to the set of
calculational model parameters, the model results can also be expressed as probability
distributions.
Qualitative or semiquantitative uncertainty analysis describes the key assumptions and sources
of uncertainty but does not attempt to assign distributions to the uncertain parameters or to
propagate the uncertainties through the calculations of risk. This approach is used when
distributions cannot be reliably assigned to the uncertain parameters or the processes
responsible for fate and effects cannot be explicitly modeled mathematically due to
incomplete site data or an incomplete understanding of the fate and effect processes. Under
these circumstances, time or other resource constraints may justify an analysis of sensitivity
and uncertainty that simply discloses the sources of uncertainty and demonstrates that the
analysis of risks are not significantly underestimated by the models employed.
D.2.2 Design of the Uncertainty Analysis
A review was performed of previous uncertainty analyses and guidance documents (H079,
H082, RI83, and CR88) to identify the approach that most appropriately applies to the
analyses presented in this report. The review addressed the extent of the analysis required
and the alternative analytical techniques available to support the analyses. In addition, an
evaluation was performed to determine if all facilities required an uncertainty analysis, or
whether a limited number of selected sites could be used to characterize the overall
uncertainty.
D.2.2.1 Extent of the Analysis. Uncertainty in the results of any risk assessment are the
result of the following (Cr88):
(1) Modeling uncertainties
(2) Completeness uncertainties
(3) Parameter uncertainties
D-3
-------
D.2.2.1.1 Modeling Uncertainties. Modeling uncertainties pertain to the formulation of
mathematical models used to predict risk and the degree to which they accurately represent
reality. One way to address this source of uncertainty is to perform the analysis using a set
of feasible alternative model structures.
In general, modeling uncertainty is the most difficult component to assess since it is often
impossible to justify a set of plausible alternative models in light of the available data and to
assign probabilities to these alternatives. To an extent, modeling uncertainty is incorporated
into the estimates of uncertainty. For example, the uncertainty in the dose and risk
conversion factors includes a consideration of the uncertainty in the form of the
dose-response and risk projection models. On the other hand, uncertainty in the formulation
of metabolic models is a serious problem in estimating dose conversion factors for many
radionuclides. Modeling uncertainty for dispersion and pathway calculations poses similar
problems. As a result, the estimates of uncertainty in radiological risk do not fully reflect
the contribution of modeling uncertainty.
One method that may be used to validate the models and, therefore, reduce this source of
uncertainty, is to perform field tests of the models under the conditions of interest.
However, this is rarely done due to cost and other limitations. Alternatively, additional
uncertain parameters could be included in the model or the range of the values assigned to
the uncertain parameters could be expanded to account for this source of uncertainty.
D.2.2.1.2 Completeness Uncertainties. Completeness uncertainties are applicable to all risk
assessments. The issue has to do with whether all significant radionuclides and pathways of
exposure have been addressed. For the facilities addressed in this report, the source terms
are well characterized and there is little likelihood that a significant unaccounted radionuclide
release is occurring. With regard to pathways of exposure, the analyses assume that four
pathways of exposure (ingestion of milk, meat and vegetables, inhalation, immersion hi
contaminated air, and exposure to contaminated ground) are present at all sites. The
groundwater pathway is not included because the deposited material is on the ground surface
in a physical and chemical form that minimizes its potential to leach to ground water. In
addition, the assumption that the deposited material does not readily leach from the soil
increases its potential to accumulate in the soil and serve as a source of external exposure,
which is the limiting pathway of exposure.
D-4
-------
However, even though a pathway is included, assessment of its contribution may be
incomplete. For example, the risk assessment analyses do not explicitly address the direct
ingestion of contaminated soil and the use of goat's milk (vs. cow's milk) in the ingestion
pathway. In addition, changes in land use and living habits could introduce pathways not
considered here, and some sites may have unique pathways. These types of completeness
uncertainties were not explicitly addressed in the uncertainty analysis because, though these
pathways could contribute to risk over any given year, they are unusual, and it is unlikely
that they would persist over the life of an individual. Hence, they would not contribute
significantly to risk or the uncertainty in the lifetime risk or contribute risks which are
substantively different than those explicitly addressed in this report.
One method that is sometimes used to account for this type of completeness uncertainty is to
add an additional term to the pathway model to represent unknown pathways and assign to it
a distribution based on judgement. This approach was not used because it is considered
unlikely that unusual pathways, such as goat's milk and soil ingestion, would be present at
the critical locations for prolonged periods of time.
D.2.2.1.3 Parameter Uncertainties. Uncertainties in the values of the calculational input
parameters are the major sources of uncertainty in the risk assessments when modelling and
completeness uncertainties are small. In addition, model and completeness uncertainties are
not readily amenable to explicit analysis. Accordingly, the quantitative uncertainty analysis
focuses on parameter uncertainties.
The assessment of parameter uncertainty involves the development of quantitative
characterizations of the uncertainties associated with key model parameters. These
characterizations can be probability distributions or a set of discrete values. Once key
uncertain parameters are characterized, their uncertainties are propagated through the models
using a simulation technique producing a probability distribution representing uncertainty
about the risk assessment model results.
In order to perform an uncertainty analysis, it is necessary to clearly define the risk that is
being estimated. Is the risk for a real or hypothetical person, is it the maximum or the
average risk, and is it the current or possible future risk that is of concern? Is it a
conditional risk based upon a given exposure scenario? The individuals constructing the
distributions must clearly understand the objectives of the analysis or the resulting
distributions will be incompatible.
D-5
-------
The results of the risk assessment provided in this report are expressed in terms of the
lifetime risk to the maximum individual. In addition, the cumulative population impacts are
presented in terms of the total incidence of fatal cancer in the exposed populations. Because
population risks represent the sum of individual risks, uncertainties in the individual risks
tend to cancel each other out during the summing process. As a result, the uncertainty in
estimates of population risk are smaller than the uncertainty in the estimates of the risks
associated with the individual members of the population. Because of this, the uncertainty
analysis is limited to the uncertainty in risks to individuals.
The concept of individual risk must also be clearly defined in order to develop the
appropriate distributions for use in the uncertainty analysis. In this report, the individual
lifetime risk is defined as the average lifetime risk from a lifetime exposure to a typical
member of the population currently residing at the inhabited location of greatest risk. It is
assumed that the individual resides at the same location for a lifetime. Since the risk being
estimated is the lifetime risk, year to year variabilities average out. This is an important
consideration since, over any given short period of time, a particular person could have
highly unusual living habits. But over a prolonged period of time, living habits tend to
resemble the population average, thereby reducing uncertainty. The differences in risks
among different age groups and their associated uncertainties also average out when
addressing lifetime impacts. Parameter distributions for the average individual represent
uncertainties in average values and do not represent the variations among individuals.
A final consideration important to the development of meaningful uncertainty distributions is
individual differences in metabolism and radiosensitivity. The risks provided in this report
are for "typical".members of the population, and, as a result, the uncertainties in these risks
are, in part, dependent on the uncertainty in our understanding of these parameters as they
apply to a typical member of the population. The biological behavior of radionuclides taken
into the body and the potential adverse effects of exposure radiation are reasonably well
known. As a result, the uncertainty in these parameters is relatively small. Conversely, any
one individual in the population could have biological characteristics that differ markedly
from "typical." The uncertainty distributions for the biological parameters for atypical
individuals is not addressed in this uncertainty analysis.
In summary, for the purpose of the uncertainty analysis, distributions were developed for the
best estimate of the values of the parameters as they pertain to the calculation of the lifetime
D-6
-------
fatal cancer risks to typical members of the population residing for a lifetime at
currently-occupied locations that have the maximum risk potential.
D.2.2.2 Techniques for Propagating Uncertainties. After each of the calculational
parameters are assigned probability distributions, these distributions are used as input to
models that propagate the uncertainties. Two widely used analytical and numerical
approaches for propagating uncertainties are method of moments techniques and Monte Carlo
techniques. Method of moments is the standard method for propagating error described in
fundamental texts on statistics. This method propagates errors by calculating a linear
combination of the first and second moments for each model factor. This is the simplest of
the methods for propagating error but requires that the distributions of the values of the
uncertain parameters can be approximated by their first two moments. In addition, since the
coefficients which quantify uncertainty about each parameter depend on the values of the
parameters, the method is only useful when the uncertainty in each parameter is small
enough that it will not significantly perturb the coefficients.
The alternative to the method of moments is the use of numerical techniques, primarily
Monte Carlo analysis. Numerical techniques have the advantage that they do not require
parameter values to have particular distributions or have a small degree of uncertainty
relative to the mean.
Monte Carlo techniques calculate risk in the same manner as described in Chapter 6, except
they perform the calculation many times, each time randomly selecting an input value from
each of the probability distributions representing uncertainty about each parameter. The
output is a risk' distribution. The number of repetitions determines the precision of the output
distribution.
A Monte Carlo technique for propagating uncertainty was chosen for use in this analysis.
D.2.2.3 Choice of Sites for Uncertainty Analysis. Of the 684 plant sites addressed in this
report, an uncertainty analysis was performed on two, Plants #222 and 60. These two
facilities have power capacities in excess of 3,000 MWe and are among the 17 plants where
the calculated MIR exceeds 10s. The exposure pathways that dominate their calculated risks
are direct exposure from deposited materials and inhalation of particulates. As these two
pathways are modeled differently, the parameters that contribute to their uncertainty differ.
D-7
-------
D.3 UNCERTAINTY IN PARAMETERS
The calculational parameters used to-derive the risks to the maximally exposed individuals
can be conveniently divided into the following categories:
• Source Terms
• Atmospheric Dispersion Factors
• Environmental Transport and Usage Factors
• Risk Conversion Factors
The following sections present a discussion of the distributions used to characterize
uncertainty about the values of the parameters in each of these categories. The choices of
underlying distributions and their characteristic values reflect consideration and professional
judgement of the models and nominal parameter values used for the site assessment.
To mitigate the possibility of absurdly small or large values for the parameters, the normal
and lognormal distributions were truncated by imposing limits of three standard deviations
from the mean. That is, if the Monte Carlo sampling process selected a value that was
more than three standard deviations away from the mean, it was programmed to go back and
try again until the value was within the limits. In the case of normal distributions, the
distributions were restricted so that they could not be negative (this is not a problem for
lognormal distributions). For parameters whose uncertainty spanned more than one order of
magnitude, a logarithmic distribution was used (i.e., log-uniform, lognormal, or
log-triangular). This tends to give comparable weight on a relative basis to both ends of the
distribution and makes the sampling more representative.
D.3.1 Source Term
The source terms used in the risk assessment are estimates reflecting the total fossil fuel
consumption reported for each boiler for a single year (1990), average radionuclide values
for the specific fuels, removal factors that account for ash partitioning in the boiler and
removal by effluent control systems, and enrichment factors that account for the
disequilibrium observed for the uranium and thorium series radionuclides in fly ash.
D-8
-------
For coal-fired units, the source term estimates are developed using the geometric means for
U-238, Th-232, and K-40 for the specific coal type (rank) used. The mean values for U-238
and Th-232 were calculated from data reflecting measurements on almost 7,000 individual
coal samples from all major coal-producing regions in the United States. As the
measurement protocols employed are capable of producing results within a few percent and
the variations between the means for coals from the various regions are only about a factor
of two, the estimates of radionuclide concentrations in the coal are highly certain.
Partitioning between bottom and fly ash varies with the type of boiler. For the two plants
considered here, both of which have only dry-bottom units, a 20/80 partitioning factor was
used. While this reflects the minimum value for these types of boilers, it is the typical
value; the overall range for the fly-ash fraction is only 80-85 percent. At worst, the ash
partitioning factor biases the estimates low by about 6 percent. Both plants use electrostatic
precipitators (ESPs) for removal of paniculate from the effluent, and the manufacturer's
efficiency ratings of the specific ESPs employed were applied to the estimated activity in the
fly ash to define the source term for radionuclides that are not enriched in the effluent.
Actual performance of the ESPs can be expected to fluctuate over time, but over the long
term (which the risk assessment reflects) and assuming proper maintenance, they can be
expected to achieve their rated efficiencies. Finally the enrichment factors applied at both
plants to account for observed disequilibrium in the uranium and thorium series reflect
typical measured values for sub-micron particles. The values for the enrichment parameter
are the most uncertain of the parameter values used to estimate the source terms. Only a
limited number of measurements are available and they show a variability on the order of a
factor of 2 or 3 within size fractions. Using enrichment factors observed for sub-micron
particles is appropriate given the efficiencies of the ESPs used at these plants, so the
uncertainty for'this parameter is almost entirely confined to the observed variability within
size fractions, i.e., a factor of 2 to 3.
D.3.2 Atmospheric Dispersion and Deposition
The product of the average annual source term (in Ci/s) and the location specific average
annual atmospheric dispersion factor (Chi/Q, in s/m3)1, yields the average annual airborne
1 The atmospheric dispersion factor is often referred to as Chi/Q, where Chi is the radionuclide
concentration at a particular location and Q is the source term. When like units are canceled, the resulting units
of Chi/Q are s/m3.
D-9
-------
concentration of radionuclides at specific locations (Ci/m3). This is a key parameter in the
calculation of inhalation risks. Dry deposition is also proportional to the air concentration at
the ground surface. Wet deposition.(precipitation scavenging), on the other hand, is
proportional to the vertically integrated air concentration (Ci/m2). For elevated releases from
facilities such as- these, the dominant exposure pathway at the location of maximum risk is
direct exposure due to wet deposition. Since wet deposition is independent of the elevation
of the release, this location can be near the facility. The risks due to inhalation and direct
exposure due to dry deposition occur at the location providing the maximum ground level air
concentration. Because of their elevated release heights, this location is thousands of meters
from these sites.
D.3.3 Pathway and Usage Factors
Once the airborne radionuclide concentration is determined by the product of the source term
and Chi/Q, the concentrations of radionuclides in various components of the environment,
such as in food and on the ground, are determined through the use of pathway factors. A
preliminary analysis showed that the only pathways that contributed significantly to the
individual risk were inhalation and direct exposure from deposition on the ground surface.
Therefore, for the purpose of this uncertainty analysis, the risk contributions to the maximum
individual from food ingestion and air submersion pathways and indoor radon are not
considered.
Tables D-l through D-5 give definitions of the parameters used in the risk assessment for the
maximally exposed individuals. In these tables, the distribution type is followed by two
parameters. The meaning of these parameters is as follows:
Distribution type 1st parameter 2nd parameter
normal (mean, standard deviation)
lognormal (GM, GSD)
loguniform (minimum, maximum)
where GM and GSD are the geometric mean and geometric-standard deviation, respectively.
A detailed description of the bases for the distributions is provided in "Analysis of the
Uncertainties in the Risk Assessment Performed in Support of the Proposed NESHAPs for
dionuclides" (EPA89).
D-10
-------
Table D-l. General Parameter Distributions
risk model uncertainty (unitless):
breathing rate (mVd):
removal rate (1/y):
vertical dispersion coefficient (unitless)
scavenging model coefficient (y/cm/s):
dry deposition velocity (m/s):
buoyant plume nse model coefficient (unitless):
lognormal (1.0, 1.8)
lognonnal (22.0. 1.2)
logumform (0.002, 0.2)
lognonnal (1.0. 1.2)
lognormal (l.OE-7, 2.0)
lognormal (1.8E-3. 3.0)
lognonnal (1.6, 1.4)
Table D-2. Site Dependent Parameter Distributions
fraction of time with precipitation:
precipitation rate (cm/y):
wind direction frequency (unitless):
heat release rate (cal/s):
lid (m):
wind speed (m/s):
release rate (Ci/y):
Pasquill-Gifford (P-G) stability class fractions (unitless):
lognormal (0.02, 1.4)
normal (Norn*. Norn/10)
lognonnal (Nom, 1.5)
lognormal (Nom, 1.2)
logunifonn (Nom/2, Nom*2)
lognonnal (Nom, 1.2)
lognormal (Nom. 1.2)
lognonnal (Nom. 1.4)
* Values labeled Nom are replaced with the nominal parameter value for the actual site (see
Tables D-3 and D-4).
The distributions presented in Tables D-l and D-2 are based primarily upon distributions
reported in the literature. They provide an indication of the range of possible values;
however, for a-specific site, the range may be narrowed by selecting only those studies that
are closely related to that site. Such a level of refinement was not possible for this study,
and thus the degree of dispersion of risk about the mean for specific sites may be an
overestimate.
D-ll
-------
Table D-3. Nominal Parameter Values Used for Risk Assessment of Plant #60
stack height
heat release rate
hd
precipitation rate
Radionuchde
U-238
Th-234
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
location of maximum
wind frequency
P-G stability class
wind speed (m/s)
fraction
location of maximum
inhalation risk
wind frequency
P-G stability class
wind speed (m/s)
fraction
304.8 metfrs
S.6SE7 calones/second
1 .000 meters
122.5 cm/y
Release Rate (Ci/v)
8.1E-2
4.0E-2
4.0E-2
4.0E-2
8.1E-2
4.0E-2
6.0E-2
3.0E+0
2.0E-1
2.0E-1
4.0E-2
2.0E-1
2.0E-1
2.0E-1
Radionuclide
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
Tl-208
K-40
risk 200 m, north
O.OS8
A B C D
1.344 2.057 3.592 3.846
.0119 .0684 .1276 .4866
20 km, east
0.104
A B C D
1.48S 2.29S 3.632 4.006
.0112 .0715 .1450 .3544
Release Rate (Ci/v)
2.3E-2
3.5E-2
2.3E-2
2.3E-2
3.5E-2
1.8E+0
1.2E-1
1.2E-1
2.3E-2
7.0E-3
2.6E-1
E F G
3.239 1.48S 0.000
.1662 .1392 0.000
E F G
3.481 1.605 0.000
.1995 .2183 0.000
D-12
-------
Table D-4. Nominal Parameter Values Used for Risk Assessment of Plant #222
stack height
heat release rate
lid
precipitation rate
152.0 meten
S.8E7 calories/second
1,000 meters
98.0 cm/y
Radionuclide
U-238
Tn-234
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Release Rate (Ci/v)
4.0E-2
2.0E-2
2.0E-2
4.0E-2
2.0E-2
3.0E-2
2.8E+0
l.OE-1
l.OE-1
2.0E-2
1.1E-1
l.OE-1
2.0E-2
l.OE-1
Radionuclide
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
Tl-208
K-40
Release Rate (Ci/v)
1.2E-2
1.8E-2
1.2E-2
1.2E-2
1.8E-2
1.6E+0
5.8E-2
S.8E-2
1.2E-2
3.5E-3
1.3E-1
location of maximum risk
wind frequency
P-G stability class
wind speed (m/s)
fraction
location of maximum
inhalation risk
wind frequency
P-G stability 'class
Wind speed (m/s)
fraction
200 meters, north
0.116
A B C D E F G
1.119 1.547 2.723 3.644 3.133 1.035 0.000
.0163 .0735 .1155 .4919 .1044 .1984 0.000
8750 meters, north
0.116
A B C D E F G
1.119 1.547 2.723 3.644 3.133 1.035 0.000
.0163 .0735 .1155 .4919 .1044 .1984 0.000
D-13
-------
Table D-5. Nucli'de Dependent Parameters
inhalation risk (risk/pCi): lognormal (CM. 2.2)
ground surface nsk (risk/toCi y/m2)):.'-" lognormal (GM, 1.2)
Nuclide
U-238
Th-234
Pa-234m
U-234
Th-230
Ra-226
Po-218
Rn-222
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
Tl-208
K-40
Branching
Fraction
parent
1.0
1.0
parent
parent
parent
0.0
.0
.0
.0
.0
parent
.0
.0
parent
1.0
1.0
1.0
1.0
0.0
1.0
1.0
1.0
0.36
parent
Half-life
4.5E9 y
2.4E1 d
1.2EO m
2.4E5 y
7.7E4 y
1.6E3 y
3.8EO d
3.1EO m
2.7E1 m
2.0E1 m
1.6E4 s
2.2E1 y
5.0EO d
1.4E2 d
1.4E10 y
5.8EO y
6.1EO h
1.9EO y
3.6EO d
5.6E1 s
1.5E-1 s
1.1E1 h
6.1E1 m
3.1EO m
1.3E9 y
Risk Factor
Inhalation (risk/jtCi)
2.2E-2
2.9E-5
1.5E-9
2.5E-2
2.9E-2
2.8E-3
4.7E-7
5.4E-7
2.7E-6
2.0E-6
3.E-13
1.4E-3
7.5E-5
2.4E-3
2.9E-2
5.8E-4
2.5E-5
7.2E-2
1.1E-3
l.OE-7
5.E-10
4.1E-5
6.2E-6
4.4E-9
0.
Ground Surface
[risk/(jiCi y/m2)!
1.9E-8
3.0E-7
3.8E-7
2.4E-8
2.7E-8
2.4E-7
1.3E-8
0.
8.8E-6
4.8E-5
2.8E-9
0.
0.
3.E-10
2.0E-8
2.E-14
3.1E-5
8.6E-8
3.6E-7
1.8E-8
5.E-10
S.3E-6
6.0E-6
l.OE-4
4.7E-6
D-14
-------
D.3.4 Precipitation Scavenging Coefficients and Depletion Rate of Ash from Soil and
Surfaces
For both sites, the principal pathway of exposure is direct radiation from particles deposited
by precipitation scavenging. The precipitation scavenging coefficient was calculated by the
following equation:
(D-l)
where:
X is the scavenging coefficient (s*1),
c is 0.5,
p is the rainfall rate, mm/s,
E is the collection efficiency, 0.36, and
Rm is raindrop size, 0.54 mm.
The dose rate from the deposited material is based on the assumption that the ash deposits
and accumulates over a 100 year period of plant operation. A key modeling assumption is
the rate at which the deposited material is depleted from the soil and other surfaces. A
depletion rate of 0.02 per year is assumed for all radionuclides except K-40, where a value
of 0.06 per year is used.
D.3.5 Radionuclide risk coefficients
These risk coefficients are expressions of the lifetime risk of fatal cancer per unit exposure
or intake of individual radionuclides. A detailed discussion of the sources and magnitudes of
uncertainties associated with the calculation of risk is provided in Chapters 5 and 6 of
EPA 89.
The uncertainties in the radionuclide cancer mortality risk factors are due to uncertainties in
the two steps that are used to derive them. First, organ dose rates are calculated as a
function of age for each radionuclide and exposure pathway. Then the risks attributable to
these organ doses are calculated.
D-15
-------
For radionuclides other than radon, the total risks were calculated from the following
expression:
Risk = F E E(J R,, (D-2)
'j
where:
Risk is the lifetime risk of fatal cancer from exposure to all radionuclides via all
pathways,
Ey is the intake or exposure from nuclide i via pathway j,
R,j is the risk factor for nuclide i via pathway j, and
F is a factor to account for the overall uncertainty in the risk model.
Each parameter in the equation is assigned a distribution. However, the distribution assigned
to the risk factor (R,,) only accounts for the portion of the uncertainty associated with
estimating dose from a given radionuclide intake or exposure. The contribution to overall
uncertainty in going from dose to risk is accounted for through the use of F, which is a
unitless multiplier. This approach allows the uncertainty in the risk model, which is common
to all radionuclides, to be treated separately from the uncertainty in the dose estimates.
F is assumed to be lognormally distributed with a geometric mean of 1.0 and a geometric
standard deviation of 1.8 (1.8329, or a factor of 7, would encompass about 90 percent of the
risk). The choice of 1.8 as the geometric standard deviation is based on the discussion of
uncertainty provided in Section 6.2.12 of EPA 89.
Table D-5 presents the data used to characterize R,j. For inhalation, it is assumed that the
probability distributions are lognonnal having a geometric mean equal to the values of the
risk factors and a geometric standard deviation of 2.2. For external exposures, the value of
the geometric standard deviation is assumed to be 1.2.
D-16
-------
D.4 RESULTS
D.4.1 Cumulative Frequency Distributions
Figures D-l and D-2 present the cumulative frequency distributions from the Monte Carlo
runs for Plants #60 and 222, respectively. The locations of maximum individual risk at both
sites are at 200 m N of the plant. Data for the location with the maximum inhalation risk
(20,000 m E for Plant #60, 8750 m N for Plant #222) are also shown. The straight lines on
the figures correspond to lognormal plots based on the GM and GSD of the Monte Carlo
data. It is clear from these plots that the data are well represented by lognormal
distributions.
In addition, the medians (50th percentiles) and the geometric means differ by only a few
percent, while the medians and arithmetic means differ by much larger factors. If a
distribution is lognormal, the median is equal to the geometric mean; if it is normal, the
median is equal to the arithmetic mean.
D.4.2 Comparison of the Results of the Uncertainty Analysis tn the Results Provided in
Chapter 6
Table D-6 presents the geometric means and ranges of the results of the uncertainty analysis.
The 90 percent credibility ranges of values were derived by dividing and multiplying the
geometric means by the 1.645 power of the geometric standard deviation.
Table D-6 also includes the nominal values of individual risks calculated for this report.
These values lie reasonably within the credibility intervals of the Monte Carlo calculations.
This provides a reasonable level of confidence that the nominal values represent a reasonable
and realistic estimate of risk.
The results also reveal that there is substantial uncertainty associated with the risk estimates.
In all cases, the range of uncertainty spans well over an order of magnitude. This means
that it is possible that the true risks could be several times higher or lower than the values
reported in Chapter 6.
D-17
-------
95
90-
80-
£ 70-
5 60^
I 30 L
Q.
20 -
10 -
5 -
20,000 m E
200 mN
1E4 1E-7 1E-6 1E-5 1E-4
Individual Lifetime Risk
1E-3
Figure D-l. Cumulative Distribution Function of Risk for Plant #60
95-
90-
~ 80b
d 70-
*60^
i so-
I 40-
£ 30
2
10-
5-
1E-8 1E-7 1E-6 1E-5 1E-4 1E-3
Individual Lifetime Risk
D-2. Cumulative Distribution Function of Risk for Plant #222
D-18
-------
Table D-6. Comparison of Monte-Carlo Individual Risk Estimates to Nominal Estimates
Pta
Plant #60
200m, norm
20,000 m. east
Plant #222
200m. north
8,750 m. north
Individual Lifetime Risk
Nominal Estimate
2.8E-S
1.8E-7
3.3E-5
3.5E-7
Monte-Carlo
Geometric Mean
1.7E-5
2.1E-7
2.9E-5
6.2E-7
90% Credibility Interval
1.9E-6 1.6E-4
3.8E-8 1.2E-6
3.3E-6 2.7E-4
1.1 E-7 3.6E-6
D.4.3 Principal Pathways and Major Parameters Affecting Risk
The major pathway for both Plant #60 and Plant #222 is direct radiation from paniculates
deposited by wet deposition. The next most important pathway (inhalation) is significantly
lower as can be seen in Figures D-l and D-2. The significance of this finding is that the risk
is not affected by the very complicated food pathway or the somewhat less complicated
ground exposure pathway due to dry deposition. Thus uncertainties in hard-to-determine
parameters, like the deposition velocity and food pathway parameters are not significant for
these facilities. The uncertainties in the risks from these plants are primarily due to
uncertainties in precipitation scavenging and the removal rates of deposited radionuclides.
D.5 REFERENCES
CR88 .Crick, M.J. et al., "Uncertainty Analysis of the Foodchain and Atmospheric
Dispersion Modules of MARC," NRPB-R184, National Radiological
Protection Board, May 1988.
EPA84 U.S. Environmental Protection Agency, Proposed Guidelines for Exposure
Assessment, Request for Comments, 49 FR 46304, November 23, 1984.
EPA89 U.S. Environmental Protection Agency, "Background Information Document
for Environmental Impact Statement on NESHAPs for Radionuclides," EPA-
550/1-89-00005.
D-19
-------
H079 Hoffman, P.O. and Baes III, C.F., "A Statistical Analysis of Selected
Parameters for Predicting Food Chain Transport and Internal Dose of
Radionuclides," NUREG/CR-1004. Prepared by Oak Ridge National
Laboratory for the NRC, November 1979.
H082 Hoffman, P.O. et al., "Variability in Dose Estimates Associated with the Pood
Chain Transport and Ingestion of Selected Radionuclides," NUREG/CR-2612,
prepared by Oak Ridge National Laboratory for the NRC, June 1982.
NRC75 U.S. Nuclear Regulatory Commission, "Reactor Safety Study: An Assessment
of Accident Risks in United States Commercial Nuclear Power Plants,"
WASH-1400, October 1975.
RI83 Rish, W.R., Mauro, J.M., and Schaffer, S.A., "Analyses of Uncertainties in
the EPA Ore Body Release and River Mode Exposure Pathway Models Used
as the Bases for Proposed Geologic Repository Release Limits," Final Report
to Battelle Project Manager Division (BPMD) for the Department of Energy,
June 10, 1983.
D-20
------- |