Commercial SNF Accident Release Fractions Rev 00, ICN 00

000-00C-MGR0-01700-000-000

November 2004


1. PURPOSE
The purpose of this analysis is to specify and document the total and respirable fractions for
radioactive materials that could be potentially released from an accident at the repository
involving commercial spent nuclear fuel (SNF) in a dry environment. The total and respirable
release fractions are used to support the preclosure licensing basis for the repository. The total
release fraction is defined as the fraction of total commercial SNF assembly inventory, typically
expressed as an activity inventory (e.g., curies), of a given radionuclide that is released to the
environment from a waste form. Radionuclides are released from the inside of breached fuel
rods (or pins) and from the detachment of radioactive material (crud) from the outside surfaces
of fuel rods and other components of fuel assemblies.
The total release fraction accounts for several mechanisms that tend to retain, retard, or diminish
the amount of radionuclides that are available for transport to dose receptors or otherwise can be
shown to reduce exposure of receptors to radiological releases. The total release fraction
includes a fraction of airborne material that is respirable and could result in inhalation doses;
this subset of the total release fraction is referred to as the respirable release fraction.
Accidents may involve waste forms characterized as: (1) bare unconfined intact fuel assemblies,
(2) confined intact fuel assemblies, or (3) canistered failed commercial SNF. Confined intact
commercial SNF assemblies at the repository are contained in shipping casks, canisters, or waste
packages. Four categories of failed commercial SNF are identified: (I) mechanically and
cladding-penetration damaged commercial SNF, (2)consolidated~reconstituted assemblies,
(3) fuel rods, pieces, and debris, and (4) nonfuel components. It is assumed that failed
commercial SNF is placed into waste packages with a mesh screen at each end (CRWMS M&O
1999). In contrast to bar& unconfined fuel assemblies, the container that confines the fuel
assemblies could provide an additional barrier for diminishing the total release fraction should
the fuel rod cladding breach during an accident. This analysis, however, does not take credit for
the additional barrier and establishes only the total release fractions for bare unconfined intact
1 commercial SNF assemblies, which may be conservatively applied to confined intact commercial
! SNF assemblies.
i
This analysis does not include the development of the damage ratio (DR), leak path factor,
airborne release fraction, and respirable fraction for accidents other than a droplslapdown of a
shipping cask, spent fuel assembly, or waste package. Values recommended for these
parameters (Section 6), in general, are only applicable to crushlimpact events with an impact
energy of 1.2 .T/cm3 (DOE 1994, p. 4-52) or less. A severe fire event resulting in a potential
radionuclide release is a beyond Category 2 event sequence (BSC 2004a) and, therefore, is not
considered in this analysis.
2. QUALITY ASSURANCE
This analysis is subject to the requirements of DOEN-0333P, Quality Assurance
Requirements and Description (DOE 2004). This analysis is performed in accordance with
AP-3.12Q, Design Calculations and Analyses and provides input to the design of structures,
systems, and components on the Q-List (BSC 2004b). Unverified design inputs are identified
and tracked in accordance with AP-3.15Q, Managing Technical Product Inputs. This analysis is
also subject to the requirements of AP-SV.lQ, Control of the Electronic Management of
Information.
000-00C-MGRO-01700-000-000 7 of 84 October 2004

Commercial SNF Accident Release Fractions
3. COMPUTER SOlTWARE AND MODEL USAGE
No computational support software or other computer software requiring qualification was used
in this analysis. The analyses of the respirable fractions for commercial SNF and crud were
performed using Microsoft Excel 97 spreadsheets. The use of this software is considered exempt
from the requirements of LP-SI.1 IQ-BSC, Software Management.
The variable input parameters used to calculate respirable fractions (i.e., mean geometric
diameter, geometric standard deviation, particle density, dynamic shape factor and maximum
respirable or cut-off particle diameter) are described in Section 6.2.2. Attachments A and B
provide spreadsheets for the calculation of the respirable fraction for commercial SNF and crud.
4. INPUTS
Design inputs and requirements used in this analysis include requirements developed by the U.S.
Nuclear Regulatory Commission (NRC), information developed by, and for, the nuclear
industry, design codes and standards, and information developed by the management and
operating contractor regarding design requirements. Inputs, and their sources and qualification
status, are identified and documented in this section in accordance with AP-3.15Q.
4.1 DESIGN PARAMETERS
The following design parameters are used to determine the respirable fraction of commercial
SNF in this analysis:
UOz Theoretical Density = 10.96 g/cm3 (Section 5, Assumption 5.6)
Dynamic Shape Factor = 1.3 (Section 5, Assumption 5.5)
The following design parameters are used to determine the respirable fraction of crud in this
analysis:
Crud Density (Hematite) = 5.2 g/cm3 (CRC Handbook of Chemistry and Physics [Weast
1972, p. B-991)
Dynamic Shape Factor = 1.3 (SAND88-1358 [Sandoval et al. 1991, p. 11-51) ~
Table 1 provides melting and boiling points for specific elements considered to be released from
breached fuel pins in this analysis. Melting and boiling points are also provided for some
compounds that may be expected to form with these elements through common reactions
(e.g., oxidation) as they are released from the fuel. These temperatures are used in Section 5 to
establish which radionuclides released from commercial SNF may be treated as gases, volatiles,
or particulates (e.g., fuel fines). Commercial SNF cladding surface temperatures under accidents
considered in this analysis are assumed to be less than 670°C (Section 5, Assumption 5.10).
The conversion factor of 3.7 x 101° dps = 1 Ci, from U.S. Environmental Protection Agency
(EPA) Federal Guidance Report No. 11 (Eckerman et al. 1988, Table 2.1), is used to convert
disintegrations per second (dps) to curies (Ci) in Section 6.2.2.2.
1 000-OOC-MGRO-01700-000-000 8 of 84 October 2004

Commercial SNF Accident Release Fractions
Table 1. Temperature Characteristics of Various Elements/Compounds in Commercial SNF
NOTES: a The states (e.g., gas, volatile, particulate) of the elements and compounds in this table are
established in Assumptions 5.11 to 5.14 (Section 5).
This column refers to the page number in Weast (1972) where the melting and boiling
temperatures are located.
' Present only in a steam environment.
4.2 CRITERIA
No design criteria are applicable to this analysis.
4.3 CODES AND STANDARDS
10 CFR Part 71. 2004. Energy: Packaging and Transportation of Radioactive Material.
10 CFR Part 72. 2004. Energy: Licensing Requirements for the Independent Storage of
Spent Nuclear Fuel and High-Level Radioactive Waste.
10 CFR Part 961. 2004. Energy: Standard Contract for Disposal of Spent Nuclear Fuel
andlor High-Level Radioactive Waste.
ANSI N13.1-1969. Guide to Sampling Airborne Radioactive Materials in Nuclear
Facilities.
ANSIIANS-5.10-1998. Airborne Release Fractions at Non-Reactor Nuclear Facilities.
October 2004

Commercial SNF Accident Release Fractions
5. ASSUMPTIONS
5.1 The DR and the leak path factor considered in NLTREGICR-6410 (SAIC 1998, pp. 3-30
and 3-31) and in DOE-HDBK-3010-94, Analysis of Experimental Data, Volume 1 of
Airborne Release Fractions/Rates and Respirable Fractions for Nonreactor Nuclear
Facilities (DOE 1994, pp. 1-2 and 1-3), are both conservatively assumed to be equal to 1.
Rationale: These two parameters are conservatively set to 1 in this analysis because of the
lack of applicable experimental data. It is recognized that for confined fuel assemblies,
the DR and the leak path factor will likely be much less than 1, depending on the severity
of the accident. In addition, the leak path factor will also depend on the particle size
distribution of the initially aerosolized particulate. A value of 1 is conservative and
bounds the expected values for these two parameters.
Usage: This assumption is used throughout this analysis wherever the total and respirable
release &actions are calculated.
5.2 A guillotine break or a longitudinal split of a fuel pin(s) is not considered credible in
this analysis.
Rationale: This analysis conservatively considers failure of 100 percent of the fuel rods
of a fuel assembly &om a credible accident. The mechanism of fuel failure in this
analysis is consistent with the failure mechanisms considered by Sandia National
Laboratories in SAND80-2124 (Wilmot 1981, p. 1 I), which defines an impact rupture as
a rupture of the cladding produced by bending or other deformation of a fuel rod.
No mention is made of a guillotine break or longitudinal split. Intact fuel is also
considered rugged and capable of sustaining severe impact environments (Wilmot 1981,
p. 18; Chun et al. 1987, Section 4.0).' A cask breach of greater than 1 in2 (6.4 cm2) is not
considered credible for fuel loaded in a transportation cask (Wilmot 1981, pp. 8 and 13).
Cask breaches smaller than 1 in2 (6.4cm2) are considered unlikely and were not
evidenced in severe tests reported in SAND80-2124 (Wilmot 1981, p. 13). Thus, only
small breaches of casks are deemed credible for impact events, because the structural
integrity of transportation casks mitigates any potential of a guillotine break or
longitudinal split of the fuel rods contained therein.
The previously cited documents provide the bases for release &actions in guidance
documents used by the engineering community, such as NUREG-1536 (NRC 1997,
Table 7.1) and NUREGICR-6487 (Anderson et al. 1996). This input is considered a
generally accepted engineering practice.
' Chun et al. (1987, Section 4.0) states that the weakest fuel assembly, that being the
Westinghouse 17 x 17, can sustain a static load in bending equivalent to 63g at 380°C without
exceeding the cladding yield strength.
000-00C-MGRO-01700-000-000 10 of84 October 2004

Commercial SNF Accident Release Fractions
Usage: This assumption is used in Section 6.2.1.
The bounding crud surface activity for a pressurized water reactor (PWR) or boiling
water reactor (BWR) fuel assembly is conservatively equal to the maximum spot activity
given in NUREG-1617 (NRC 2000a, Table 4-1).
Rationale: Sandoval et al. (1991, Figures 3 and 4 on pp. 21 and 22) show that over 60
percent of the PWR fuel rods and over 80 percent of the BWR fuel rods examined had
negligible surface activity. Thus, a bounding crud surface activity may be considered
equal to the maximum spot surface activity.
Usage: This assumption is used in Section 6.2.1.3.
PWR crud will have approximately the same respirable fraction as determined for BWR
crud.
Rationale: This assumption is based on the experimental data cited in Sandoval et al.
(1991, pp. 23 to 26). These experimental data indicate a similarity between the sizes of
crud particles on the surface of PWR fuel rods (as measured by a scanning electron
microscope) with the sizes of crud particles scraped off of BWR fuel rods (as measured
with filter paper). Although the PWR crud particles are smaller than those considered in
the BWR crud distribution used to determine the respirable fraction, the fraction of
respirable PWR crud is not expected to significantly increase. This is because of the
smaller particles contributing only a small amount of mass relative to the total. This is
based on the previous technical argument and the conservative manner in which PWR
crud is treated in this analysis (i.e., PWR and BWR crud are treated identically with
respect to release fraction even though PWR crud is tightly bound to fuel rods).
Usage: This assumption is used in Section 6.2.2.4.
The value for the dynamic shape factor (K) for commercial SNF is assumed to be 1.3.
No applicable data exist for this shape factor for commercial SNF aerosols.
Rationale: This assumption is based on the value for crud from a sample of Quad Cities
fuel in Sandoval et al. (1991, p. 11-5). The crud sample obtained from Quad Cities had a
lognormal distribution, which is the type of distribution used to model the commercial
SNF in this analysis. A value of 1.3 implies compact, angular shaped particles.
This value is close enough to unity that it does not play a decisive role (Sandoval et al.
1991, p. 11-5), based on a comparison between the scanning electron microscope pictures
in NUREGICR-0722 (Lorenz et al. 1980, Appendix C) and Sandoval et al. (1991,
pp. 1-37 and 1-38) and the argument that a value close to unity does not play a
decisive role. Pictures produced by the scanning electron microscope show that
commercial SNF particles and crud particles do not significantly differ in shape.
Usage: This assumption is used in Sections 4.1 and 6.2.2.2.
October 2004

I Commercial SNF Accident Release Fractions
The particle density of the aerosol in this analysis will be set equal to the theoretical
density of UOz, which is 10.96 &m3 (Weast 1972, p. B-151).
Rationale: Fuel fines released from accident events at the repository are not likely to
contain many voids commonly found in the fuel because of their small size. Hence, the
theoretical density of UO2 is justifiably applied in the calculations of this analysis.
These data are taken from Weast (1972, p. B-151), which is generally accepted by the
scientific and engineering community and is technically defensible.
Usage: This assumption is used throughout Sections 4.1, 6.2.1.2, and 6.2.2.2 and in
Attachment C.
The "~e-based crud surface activity for PWR fuel is assumed to be equal to
5,902-~~ilcm~.
Rationale: This value is cited from Jones (1992, Table I), however no basis or reference
is provided therein. Even though no value of the crud surface activity for this isotope has
been cited in other literature on crud, the results of dose calculations are not sensitive to
this parameter.
Usage: This assumption is used in Section 6.2.1.3.
The mass median diameter (MMD) of initially aerosolized commercial SNF is assumed
to be equal to 150 pm.
Rationale: This assumption is based on fuel fines collected from burst rupture tests in
NWUZGICR-0722 (Lorenz et al. 1980, p. 105 and Appendix C). The fines were
measured with a scanning electron microscope and determined to be "typically 150 pm"
in the furnace tube near the point of the fuel pin rupture. This value may be considered
conservative when applied to drop or impact events because larger particulates are likely
to be initially aerosolized in these events because of the brittle nature of the fuel
(assuming reasonable drop heights). This is supported by the MMDs measured in impact
tests on unclad, depleted, ceramic UO2 pellets in ANL-81-27 (Mecham et al. 1981, pp.
26, 34 and 35). These tests involved the impaction of two separate samples of depleted
UO2 pellets. The resulting particle-size distributions had measured MMDs of 18 rnrn and
32 mm (not pm). These MMDs are significantly larger than the 150-pm MMD assumed
in this analysis for the initially aerosolized commercial SNF. Because larger particles are
essentially irrespirable (based on the rationale presented in Section 6.2.1) and carry a
large portion of the total mass, larger MMDs equate to smaller respirable fractions (RFs)
with other parameters being equal (e.g., the standard deviation). Thus, the selection of
the 150-pm diameter to represent the MMD of the initially aerosolized commercial SNF
from a drop or impact event is a conservative assumption. Another set of test data
performed on single pellets of UOz by Alvarez is cited in SAND90-2406 (Sanders et al.
1992, Section IV-4). Alvarez performed a series of tests on pellets of clad UOz that were
both depleted and irradiated. These tests involved the detonation of explosive charges
near the fuel. This resulted in a significantly greater amount of energy imparted to the
fuel than occurred in the burst rupture tests or a drop or impact event considered in this
000-OOC-MGRO-01700-000-000 12 of 84 October 2004

Commercial SNF Accident Release Fractions
analysis. Hence, the measured MMDs from these tests, which ranged from
approximately 30 to 100 pm, are not considered applicable to this analysis.
The relatively small difference, however, between the 100-pm MMD from Alvarez's
explosive tests and the 150-pm diameter considered in this analysis, reinforces the
conservative arguments made in this analysis for selecting an MMD of 150 pm. This is
based on the preceding technical bases and on NUREGICR-0722 (Lorenz et al. 1980, p.
105 and Appendix C) providing the release fractions used in guidance documents that are
used by the engineering community, such as NUREG-1536 (NRC 1997, Table 7.1) and
NUREGICR-6487 (Anderson et al. 1996).
Usage: This assumption is used throughout Sections 6.2.2 and 7.
5.9 Three percent of the total mass of initially aerosolized commercial SNF fuel fines is
assumed to have geometric diameters of less than 12 pm.
Rationale: This assumption is based on fuel fines collected from burst rupture tests in
NUREGICR-0722 (Lorenz et al. 1980, p. 105). These tests provide the best currently
available data for the release of fuel fines from commercial SNF. It is recognized that
pellet fragmentation increases with fuel bumup. Because the commercial SNF,
potentially accepted at the repository, may have burnups that are higher than the fuel
tested in NWZEGICR-0722 (Lorenz et al. 1980) (approximately 30 GWdIMTU), then
there may be more fuel fines with diameters less than 12-pm available for release.
In these burst rupture tests, it was determined that only a small fraction, 0.8
to 2.9 percent, of the he1 mass ejected from the fuel was canied out of the furnace tube
into the thermal gradient tube and filter pack (Lorenz et al. 1980, Table 42).2 Considering
the deposition of the released fuel particles from gravity, fuel particles of diameters
greater than 12 to 15 pm are considered to have settled out before reaching the thermal
gradient tube. The most conservative interpretation of these data with respect to the
respirable fraction is to select the highest release fraction (i.e., 3 percent) and the smallest
diameter (i.e., 12 pm). This will result in the calculation of a conservative respirable
fraction. This is based on the preceding technical arguments, the confirmatory analysis in
Attachment C, and the fact that these data come from a source (Lorenz et al. 1980,
p. 105) that is commonly cited for establishing release fractions for commercial SNF
(Assumption 5.8).
Usage: This assumption is used throughout Sections 6.2.2 and 7.
Small fractions (0.8 to 2.9 percent) in Lorenz et a. (1980, Table 42) were determined by
dividing the mass in the thermal gradient tube and the filter packs by the total mass released.
October 2004

I Commercial SNF Accident Release Fractions
Accidents at the repository, involving dropping or impacting of commercial SNF
assemblies or containers loaded with commercial SNF assemblies, will occur at
temperatures that ensure that elemental forms of Sr, Ru, and Cs are below their boiling
points at the fuel cladding surface (i.e., cladding surface temperature below
approximately 670°C).
Rationale: The boiling temperatures of Sr, Ru, and Cs are listed in Table 1 (Section 4).
These temperatures establish a cladding surface limiting temperature of 670°C. Because
cladding surface temperatures are limited to approximately 400°C for fuel discharged at
least 5 years from a reactor under normal conditions in shipping casks (Levy et al. 1987),
this assumption disallows these drop event release fractions to be used unless a
significant fire occurs in the immediate vicinity of the drop. This assumption is generally
accepted and commonly applied to accident conditions for licensed cask systems as
recommended in NIJREG-1536 (NRC 1997, p. 4-3).
Usage: This assumption is used in Sections 4.1 and 6.2.1.2 to establish the release
fractions from the fuel for 90~r, Io6~u, 1 3 4 ~ ~ , and 1 3 7 ~ ~ and in Assumptions 5.11, 5.12,
5.13, and 5.14.
In this analysis cesium and its compounds are conservatively treated as volatiles because
of the low melting temperature of elemental cesium.
Rationale: The boiling temperature of elemental cesium (Section 4, Table 1) is above the
maximum fuel cladding surface temperature of 400°C assumed in this analysis
(Assumption 5.10); hence, it is probable that cesium released from the commercial SNF
would exist in the vapor phase because of the higher temperatures within a fuel pin with
this surface temperature. This is confirmed by each of the burst rupture and diffusion
tests performed in NUREGICR-0722 (Lorenz et al. 1980). In each test, the cesium
purged kom a breached fuel pin was in the form of either condensed CsI or CSZUO~ (fuel
fine) or gaseous elemental cesium, CsI, CszO or CsOH (the latter, only in the presence of
a flowing steam environment). When released to the cooler environment outside of the
fuel, however, the elemental cesium and gaseous cesium compounds: (I) quickly
condensed and were removed by fuel fines, (2) condensed in a thermal gradient tube,
(3) reacted with some nearby quartz to form a cesium silicate (particle), (4) the remainder
of the released cesium was found deposited on the HEPA filters as particulates.
Usage: This assumption is used in Section 6.2.1.2.
Iodine and its oxides (if present) are conservatively treated as gases because of iodine's
low boiling temperature.
Rationale: The boiling temperature of elemental iodine in Table 1 (Section 4) is clearly
below the assumed maximum fuel cladding surface temperature of 400°C (Assumption
5.10). No boiling temperatures were found for iodine oxides; the melting temperatures of
these compounds, however, are comparable to that of elemental iodine. Hence, these
oxides are considered to be in a gaseous state. In the commercial SNF burst rupture tests
performed in NUREGICR-0722 (Lorenz et al. 1980, pp. 117 to 119, Tests HBU-7 to
000-OOC-MGRO-01700-000-000 . 14 of 84 October 2004

Commercial SNF Accident Release Fractions
HBU-lo), it was determined that iodine was released from breached fuel in either
elemental form or as CsI. Although CsI is unlikely to be in a gaseous form when released
from the fuel matrix, because of its high boiling point, the treatment of the iodine in this
compound as a gas is as conservative as the treatment of the Cs in this compound as a
volatile (Assumption 5.1 1). This is based on the preceding technical argument, which is
based on established melting and boiling temperatures.
Usage: This assumption is used in Section 6.2.1.1.
In this analysis, ruthenium and its compound are treated as volatiles.
Rationale: The melting and boiling temperatures of ruthenium (Section 4, Table 1) are
clearly above the maximum fuel cladding surface temperature of 670°C assumed in this
analysis (Assumption 5.10). In addition, the burst rupture tests performed in
NUREGICR-0722 (Lorenz et al. 1980, pp. 117 to 119, Tests HBU-7 to HBU-10)
determined that the vaporized ruthenium, in the form of Ru02 and Ru04, was negligible
compared to the ruthenium captured in the fuel fines. The ruthenium in the fuel fines is
considered to be in an elemental form and in a condensed state because of the existing
temperatures. The formation and vaporization of RuOz and Ru04 begins at
approximately 500°C and 600°C (Lorenz et al. 1980, pp. 116 and 117). These
temperatures are under the maximum fuel cladding surface temperature of 670°C
assumed in this analysis (Assumption 5.10). Ru04 is volatile. After these compounds
have been purged from the fuel pin, they cool to temperatures where the RuO~ reverts to
RuOz and the RuOz decomposes to its elemental components. The elemental ruthenium
resulting from this decomposition is in a solid/particulate form because of its high
melting and boiling temperatures. Because of the presence of Ru04, ruthenium and its
compounds are conservatively treated as volatiles in accordance with ISG-5 (NRC 2003,
Table 7.1).
Usage: This assumption is used in Section 6.2.1.2.
In this analysis, strontium and its oxides (if present) are treated as particulate (e.g., fuel
fine) because of their high melting and boiling temperatures.
Rationale: The melting and boiling temperatures of strontium and SrO (Section 4,
Table 1) are clearly above the maximum fuel cladding surface temperature of 670°C
assumed in this analysis (Assumption 5.10). Any formation of SrO2 decomposes back to
SrO (Weast 1972 p. B-143); thus, the state of this compound is not considered.
Similarly, in the commercial SNF burst rupture tests performed in NUREGICR-0722
(Lorenz et al. 1980, pp. 117 to 119, Tests HBU-7 to HBU-lo), iodine was released from
breached fuel in either an elemental form or as CsI; no SrIz was mentioned. Thus, there
is no consideration of the volatility of this compound, based on the preceding technical
argument, which is based on established melting and boiling temperatures.
Usage: This assumption is used in Section 6.2.1.2.
Mechanically damaged commercial SNF and cladding-penetration damaged commercial
SNF are assumed contained in a canister with some assembly-like structure inside the
000-00C-MGRO-01700-000-000 15 of 84 October 2004

Commercial SNF Accident Release Fractions
canister. Fuel rods, pieces, and debris, and non-fuel components are only assumed
contained in a canister.
Rationale: The results are not sensitive to the assumption.
Usage: This assumption is used in Section 6.3.
Failed fuel that has some assembly-like structure is assumed not to pulverize under a drop
or impact event. In this assumption, credit is taken for the canister structure, cladding,
and gridslspacers providing sufficient structural support to prevent the fuel from
undergoing any significant amount of pulverization.
Rationale: The results are not sensitive to the assumption.
Usage: This assumption is used in Section 6.3.
The contents of a failed fuel canister without any assembly-like structure is assumed to be
up to 20 percent pulverized following a drop or impact event. In this assumption, credit
is taken only for the structural support provided by the canister to minimize the amount
of fuel to be pulverized. The value of 20 percent is selected based on the same
conservative assumption made in SAND84-2641 (MacDougall et al. 1987, pp. 5-15
to 5-26).
Rationale: The results are not sensitive to the assumption.
Usage: This assumption is used in Section 6.3.
6. ANALYSIS
Commercial SNF assemblies or confinement systems (e.g., casks, canisters, or waste packages)
that contain assemblies may become involved in an accident at the repository that could
potentially compromise the confinement boundaries that prevent or reduce the amount of
radioactive material released from fuel assemblies. Confinement boundaries include fuel
structure, fuel cladding, container confinement boundaries, and facility confinement boundaries.
Radioactive material may be released in the form of gases, volatiles, or particulates.
In addition to the confinement boundaries, the physical properties of commercial SNF fuel fines
and crud (surface deposits) can have an effect on the calculated dose consequences resulting
from event sequences involving commercial SNF assemblies. In particular, the distribution of
particle sizes (diameters and masses) affect the fractions of material that are locally deposited
versus the fraction that remains airborne (as an aerosol) long enough to reach potential offsite or
onsite receptors. Further, only the respirable fraction of the aerosolized fuel fines and crud that
reach a receptor significantly contributes to internal organ doses.
Section 6.1 discusses the method for determining the airborne release fractions (ARFs) and
respirable kaction (RFs) for the intact and failed commercial SNF.
October 2004

Commercial SNF Accident Release Fractions
Section 6.2.1 establishes the ARFs of specific radionuclides released from commercial SNF or
from the surface of commercial SNF (e.g., crud). These ARFs are based on experimental data,
analyses, or a conservative estimate accepted in license applications approved by the NRC in
cases where insufficient experimental and theoretical data are available.
Section 6.2.2 establishes the RFs of released radionuclides from commercial SNF and any
associated crud. Four distinct steps are performed to establish the RFs of these particulate
matters (Sections 6.1 and 6.2.2). Spreadsheets are provided in Attachments A and B. Different
methods for determining the RFs are also discussed in Section 6.2.2.
Section 6.3 discusses the ARFs and RFs determined for failed commercial SNF.
6.1 METHOD
The analysis first addresses the ARFs of commercial SNF following an event at the repository
that involves either a drop or an impact of a shipping cask, canister, or waste package loaded
with commercial SNF or an uncanistered, unconfined commercial SNF assembly, or canistered
failed commercial SNF. In this analysis, this event occurs in a dry environment (i.e., not in a
pool). After the ARFs have been established for commercial SNF, the fraction of the airborne
material that is respirable, denoted as the RF, is established., Finally, the ARFs and RFs are
combined to establish the respirable release fraction for commercial SNF at the repository.
The methodology applied in this analysis is consistent with those presented in NUREGICR-6410
(SAIC 1998, Section 3.2.5.2) and DOE-HDBK-3010-94 (DOE 1994, Section 1.2), for example:
Total Release Fraction (all pathways) = DR x LPF x ARF (Eq. 1)
Respirable Release Fraction (inhalation) = DR x LPF x ARF x RF (% 2)
where
DR is the damage ratio, assumed to be 1 in this analysis (Section 5, Assumption 5.1).
LPF is the leak path factor, assumed to be 1 in this analysis (Section 5, Assumption 5.1).
The ARF is the fraction of material at risk3 that can be suspended to become available for
airborne transport following a specific set of induced physical or thermal stresses. An ARF for
each radionuclide, or appropriate grouping of radionuclides (e.g., fuel fines), released from
commercial SNF is determined from experimental data, analyses, or previous precedents
established in documents approved by the NRC or conservative estimates, or both.
Attachment D includes a summary of some of the fractions used in licensing documents for other
nuclear facilities.
3 Material at risk specific to commercial SNF is typically expressed as "curies" of radionuclide
inventory associated with a unit assembly of fuel rods.
000-00C-MGRO-01700-000-000 17 of 84 October 2004

Commercial SNF Accident Release Fractions
The RF is the fraction of the initially suspended airborne material that can be inhaled and result
in inhalation doses.
The respirable release fraction calculated using Equation 2 may subsequently be used in
inhalation dose calculations. The total release fraction calculated using Equation 1 is used as
input for other dose calculations (e.g., submersion, groundshine, and ingestion) at the repository.
Equation 2 can also be used to calculate submersion, groundshine, and ingestion doses for
radionuclide releases from a dry transfer facility at the repository because it is expected that only
particles of respirable sizes (i.e., less than or equal to 10 pm aerodynamic equivalent diameter
[(AED]) are released to the environment. Non-respirable particles coming out of a ruptured fuel
rod, with a high terminal velocity of approximately 3 cm/s (Attachment C) or larger, would
deposit on the surfaces of adjacent fuel rods within a fuel assembly, deposit inside a
transportation cask or waste package, or deposit inside the ventilation ductwork.
Intact Commercial SNF
For intact commercial SNF, an analytical method was developed to quantify the RF of the
aerosolized particulate measured in various experiments (Lorenz et al. 1980; Mecham et al.
1981, pp. 26, 34, and 35, Table 2; Sandoval et al. 1991, pp. 23 to 26). In Section 6.2.2, the
method used to determine the RF for particulate from commercial SNF fuel fines and crud is
presented. Four distinct steps are followed to establish the RF of particulates:
Establish the maximum size of respirable particulates that contribute to the inhalation
doses.
Determine the relationships between the mean geometric diameter (MGD), the MMD,
and the activity median aerodynamic diameter (AMAD) for a given particulate size
distribution.
Characterize the particulate geometric size and mass distributions of commercial SNF
and crud.
Establish a method to calculate the RF based on the definition of a respirable aerosol.
Information on the size of respirable particulate is found in International Commission on
Radiological Protection (ICRP) Publication 30 (1979, Figure 5.1) and Federal Guidance Report
No. 11 (Eckerman et al. 1988, p. 14). The establishment of relationships between the MGD,
MMD, and AMAD are based on definitions and equations provided in the Handbook on
Aerosols (Dennis 1976, p. 11 1). The characteristics of the particulate size distribution for
commercial SNF are based on experimental data provided in NUREGICR-0722 (Lorenz et al.
1980) and in ANL-81-27 (Mecham et al. 1981, pp. 26, 34, and 35, and Table 2). Experimental
data provided by Sandoval et al. (1991, pp. 23 to 26) established the characteristics of the
particulate size distribution for crud. For intact commercial SNF, an iterative method is used to
calculate the RF of a particulate aerosol. This method is based on the definition of a respirable
aerosol as established in Publication 30 (ICRP 1979) and Federal Guidance Report No. 11
(Eckerman et al. 1988) and is judged to provide the most accurate values for commercial SNF
and crud. Two other methods to calculate the respirable fraction of a particulate aerosol are
000-00C-MGRO-01700-000-000 18 of 84 October 2004

Commercial SNF Accident Release Fractions
discussed in Section 6.2.2: the AMAD-10 method and the AED method. These methods are
based on different interpretations of the definition of a respirable aerosol that have been made in
some documents used in this analysis, such as SAND80-2124 (Wilrnot 1981 p. 38) and
ANL-81-27 (Mecham et al. 1981, pp. 26, 34, and 35, and Table 2). The results from these other
methods are provided, however, for comparison purposes only.
Failed Commercial SNF
There are four categories of failed commercial SNF, which include: (I) mechanically and
cladding-penetration damaged fuel, (2) consolidated/reconstituted assemblies, (3) fuel rods,
pieces, and debris, (4)non-fuel components. Non-fuel components are not important
contributors to radiological consequences and therefore are not considered in this analysis.
Experiments reported in ANL-81-27 (Mecham et al. 1981, pp. 26, 34, and 35, and Table 2) and
ANL-82-39 (Jardine et al. 1982) provide data that may be applicable to failed fuel damaged by a
drop or impact event that involves fuel pulverization. The ARFs and RFs produced from these
experiments involve unconfined (i.e., no cladding) glass and UOz ceramic specimens impacted
by a dropping weight. The applicability and details of these tests are described in Section 6.2.1.
These test data and their associated pulverization fraction (PULF), which is equal to ARF x RF,
were deemed not appropriate for application to dropped or impacted intact fuel assemblies.
Because some of the failed fuel exists, however, as small unclad fuel pieces and debris, this data
may be considered applicable.
6.2 ARFS AND RFS FOR INTACT COMMERCIAL SNF
r6.2.1 ARFs for Intact Commercial SNF
The ARFs from commercial SNF account for the fact that some of the commercial SNF
radionuclides are retained in the fuel matrix or exist ina chemical or physical form that is not
capable of release under credible accident conditions. Table 2 lists the documents containing
ARFs reviewed for this analysis and their sources for the ARFs. A review of this table reveals
two primary groups of experimental data that produced the majority of the cited ARFs for
commercial SNF:
Four experiments that burst ruptured highly irradiated commercial SNF rod segments in
a flowing steam environment. These experiments quantified and characterized fission
product release under conditions postulated for a spent-fuel transportation accident
(Lorenz et al. 1980).
Two single energy density impaction tests on three unconfined U02 pellets. These tests
characterized the size distribution and the RF of the fragments generated (Mecham et al.
1981, pp. 26,34, and 35, and Table 2; Jardine et al. 1982).
Burst Rupture Tests
NUREG-1567 (NRC 2000b, Table 9.2), NUlZEG-1536 (NRC 1997, p. 7-5), NUREG-1617
(NRC 2000a, Table 4-I), NUREGJCR-6487 (Anderson et al. 1996, pp. 31 and 32), and
SANDSO-2124 (Wilmot 1981, p. 36 ind Table XVIII) use the burst rupture data in the same


manner to produce their cited ARFs. Thus, although a burst rupture event is not necessarily
equivalent to a drop event, these documents do provide strong NRC precedents supporting use
ofthe burst rupture data as a basis for analysis of a wide range of accidents involving
commercial SNF. ~ Impact Rupture Tests
Release fractions based on unconfined (i.e., no cladding) impact tests involving glass and U02
ceramic specimens are cited in NUREGICR-6410 (SAIC 1998, Table 3-1, 3.3.4.10.d), which
provides guidance on how to calculate the characteristics of releases of radioactive materials
and/or hazardous chemicals from nomeactor nuclear facilities. Although NUREGJCR-6410
(SAX 1998) is generally applicable to repository operations, the applicability of release
fractions derived from impact tests involving unconfined test specimens to accidents involving
clad commercial SNF is questionable without hrther consideration of the potential for large
scale gross cladding damage (i.e., guillotine breaks or longitudinal splits).


Commercial SNF Accident Release Fractions
Table 2. References Containing Airborne Release Fractions
ANL-81-27 (Mecham et al.
1981. PP. 26,W and 35, Fuel fines Table 2) and ANL-82-39
Source Document
ANSIIANS-5.10-1998 Gases, fuel fines,
Radionuclides
Given ARFs
DOE-HDBK-3010-94
Gases, fuel fines
(DOE 1994)
ISG-5 Gases, fuel fines.
NUREG-1536 Gases. fuel fines,
lNRC 1997. Table 7.1) volatiles, crud
NUREG-1567 Gases, fuel fines,
NUREG-1617 Gases, fuel fines.
NUREGICR-0722
(Lorenz et al. 1980)
NUREGICR-6487 Gases. fuel fines,
(Anderson et al. 1996) volatiles. crud
Cs, I, Ru, fuel fines
NUREGICR-6410
$SAC 1998)
Regulatoly Guide 1.25 ( 3 ~ , " ~ r , lZgl
Noble gases, iodine.
tritium, fuel fines
Comments
Measured release fractions from two impact tests on bare
unclad UO2 pellets
ARFs for gases and volatiles are conservatively set equal
to 1 and ARFs for fuel fines and crud come from several
different sources (ANSIIANS-5.10-1998. Table Al)
ARFs for gases are conservatively set equal to 1, ARFs
for fuel fines are based on a linear relationship produced
from the ARFs determined in ANL-81-27 (Mecham et al.
1981, pp. 26.34, and 35, Table 2)
Values cited from NUREGICR-6487 (Anderson et al.
1aafi\
SAND80-2124
(Wilmot 1981, p. 36,
Table XVIII)
"".,,
ISG-5 (NRC 2003) indicates tnat NlrREG-1536 will be
revised to be consistent with NUREGICR-6487 (Anderson
asKr, 12g1, 1 3 4 ~ ~ I 137 CS.
%r, '06~u. 'OCO
(crud), actinides
et al. 1996)
Values cited from NUREGICR-6487 (Anderson et al.
1996)
Values cited from NUREGICR-6487 (Anderson et al.
1996)
Measured release fractions from four burst rupture tests
on I-foot fuel senments
ARFs for fuel fines from NUREGICR-0722 (Lorenz et al.
1980), other ARFs from impact tests
Does not state source of ARFs, however. ARFs for gases
appear to be from Regulatoly Guide 1.25: ARFs foriuel
fines and crud from SAND80-2124 Nilmot 1981. D. 36.
Table XVIII); ARFs for volatiles from NUREGICR-0722
(Lorenz et al. 1980)
PWR and BWR release fractions for five chemical
element classes and four types of transportation casks in
NUREGICR-6672 (Sprung et al. 2000, Table 7.31)
Presents ARFs for gas releases from fuel handling
accidents (no source of ARFs presented, regulatoly
position)
ARFs for noble gases from light water reactor fuel design
data, for fuel fines (Cs, Sr, Ru, and actinides) from
NUREGICR-0722 (Lorenz et al. 1980)
NOTES: ARFs = airborne release fractions; BWR = boiling water reactor; PWR = pressurized water reactor,
000-00C-MGRO-01700-000-000 21 of 84 October 2004

Commercial SNF Accident Release Fractions
In addition to the assumption for severe clad damage (Section 5, Assumption 5.2), application of
test data in the determination of ARF and RF for large masses of commercial SNF (e.g., bare
unconfined fuel assemblies) dropped from substantial heights (i.e., those that have an impact
energy density greater than 1.2 J/cm3) may be considered excessively conservative (DOE 1994,
p. 4-52). This is supported, albeit for a different brittle material, by a simple test showing that a
160-g glass cylinder bounces off a steel plate in a 10-m drop, rather than fracturing as would
have been predicted by a similar correlation (Jardine et al. 1982, Section 8 of Appendix D).
This reveals the largest deficiency potentially associated with the use of these test data in the
determination of ARF and RF for large masses of commercial SNF: can the physical phenomena
associated with damage produced by dropping a weight on an unclad fuel pellet be equated to the
damage produced by dropping a fuel assembly onto a potentially unyielding surface?
The PULF correlation, which is the fraction of airborne material that is respirable, is not
applicable to conditions where the surface area of the impacting component (i.e., the fuel
assemblies) is smaller than the surface impacted (i.e., the ground) (SAIC 1998, p. 3-87); a
condition that is clearly not applicable to the one considered in this analysis. Thus, unless more
experimental tests can be performed, it appears that this correlation has limited applicability to
fuel assembly drops and when used in these instances will provide grossly conservative fractions
of respirable fuel fines.
DOE-HDBK-3010-94 (DOE 1994, pp. 4-52 to 4-54) presents a numerical analysis that fits the
unconfined specimen impact test data to an equation that estimates the PULF of commercial SNF
particulate (i.e., the ARF multiplied by the RF):
where
A is an empirical correlation equal to 2 x lo-" cm-s2lg
p is the particle density (g/cm3)
g is gravitational acceleration (980 cm/s2)
h is the fall height (cm).
For U02, the value of the empirical correlation is based on two experimental data points from
single energy density (1.2 J/cm3) impaction tests on three unconfined UOz pellets (Mecham et al.
1981, Table 2 and pp. 30 to 35). The linearity of the correlation with respect to fall height is not
based on data for U02, but is based on impaction tests for Pyrex and SRL 13 1 (Savannah River
Site glass fiit) over the range of energy densities of 1.2 to 10 J/cm3 (Jardine et al. 1982, Figure 13
and pp. 28 to 31). For Pyrex, the linearity is poor at low energy densities.
In an attempt to correct some of the conservatism associated with extrapolating small specimen
impact test data to large masses of glass or ceramic materials, SAND84-2641 (MacDougall et al.
1987, pp. 5-15 to 5-26) modified the PULF correlation by including an energy partition factor
(EPF) to account for the energy absorbed by components of a fuel assembly (e.g., cladding,
spacer grids, nozzles) and the non-uniform energy density impact applied to a dropped fuel
assembly.
October 2004

Commercial SNF Accident Release Fractions
The value of the EPF can be derived from analysis or experiment. SAND84-2641 (MacDougall
et al. 1987, pp. 5-15 to 5-26), however, assumed this factor to be 0.2 for a fuel assembly without
any defensible basis, although it was considered to be conservative.
Summary
Based on the applicability of the data and previous licensing precedent, the burst rupture ARFs
are considered applicable to accidents involving commercial SNF. ARFs derived from burst
rupture experimental data referenced in technical guidance documents are summarized in
Table 3. As seen in Table 3, however, the burst rupture ARFs may be interpreted or corrected in
different manners to produce a range of ARFs for specific radionuclides released from
commercial SNF, which depend on the range of conditions, such as the temperature range.
In the following sections, the ARFs for commercial SNF radionuclides are conservatively
examined and values recommended for application to the repository.
6.2.1.1 ARFs for CSNF Gases
Iodine, Hydrogen and NobleGases
As fuel is irradiated in a nuclear power reactor, fission product atoms, of which approximately
15 percent are inert gases, are produced and buildup within the cladding of the fuel pins.
Release of these fission gases from the fuel matrix to the plenum and the gap region between the
fuel and the cladding is directly related to fuel pellet swelling which is a strong function of linear
power density. According to p. 25.2 of Regulatory Guide 1.25, all of the gap activity in the
damaged rods is released and consists of 30 percent 'kr, 10 percent of the total noble gases
other than ''~1, and 10 percent of the total radioactive iodine in the rods. These values are cited
by the NRC (1997, Table 7.1; 2000b, Table 9.2) for use in potential accident releases. Page 25.2
I inventory may be of Regulatory Guide 1.25 also states that 30 percent of the '"I and 12'
'
assumed released for the purpose of sizing filters.
The release fractions from p. 25.2 of Regulatory Guide 1.25 are assumed for oxide fuels and in
cases where the following conditions are not exceeded:
Peak linear power density of 20.5 kW/ft (67.25 kW1m) for highest power assembly
discharged
Maximum center-line operating fuel temperature less than 4,500°F (2,482OC) for this
assembly
Average bumup for the peak assembly of 25,000 MWdAMTU or less (this corresponds to
a peak local bumup of about 45,000 MWcWMTU).
23 of 84 October 2004

Table 3. Airborne Release Fractions From Fuel Retention
This value is consistent with the release fraction for volatiles.
The values are from NUREGICR-6410 (SAIC 1998. Table 3-1.3.3.4.10d and 3.3.4.12a). The crud value represents the ARF for loose surface
contamination.
90Sr is being treated as a fuel fine. See Section 6.2.1.2
'This value is consistent with the release fraction for fuel fines.
ARF for fuel fines is not provided by NUREG-1536 (NRC 1997, Table 7.1), but is assumed consistent with other listed particulates.
h This value is the ARF for actinides.
i 0.1 5 for normal and off-normal conditions, 1.0 for accident conditions.
This value is the effective crud ARF. The effective crud ARF consists of the product of the CSF with the ARF as described in Section 6.2.1.3.
106 RU
Fuel Fines
Crud
CSF = crud spallation fraction.
NOTES: a Values include cask retention and respirable fraction assumptions. Also, ISG-5 (NRC 2003, Table 7.1) indicates that this document will be revised to be
consistent with NUREG-1617 (NRC 2000a, Table 4-1) and NUREGICR-6487 (Anderson et al. 1996, pp. 30, 31, and 32).
Values are from SAND80-2124 (Wilmot 1981, Table XVIII) and are a combination of burst rupture data and oxidation data.
2.0 x lo4'
3.0 x lo+'
0.1511.00'
1.5 x
2.3 x
0.15
2.0
3.0 lo-5
0.1511.00 '
2.0
3.0 10-~8
0.1511.00
2.0
2.0 lo-46
1 .O x
2.0
3.0 IO-~Q
0.1511.00 '
8.2
2.0 lo-5h
0.25
-
-
-
2.0 x lo4
3.0 IO-~
1.5 x


Commercial SNF Accident Release Fractions
Although some of the potential commercial SNF handled at the repository exceeds some of the
values in these assumptions, the conservatism built into these release fractions allows them to be
applied to accidents involving commercial SNF handled at the repository. For example,
according to Graves (1979, p. 177 and Figure 8-9), peak power density rods of light water
reactors will typically release 5 to 10 percent of the fission-product gases to the gap, significantly
less than the 30 percent recommended on p. 25.2 of Regulatory Guide 1.25. Furthermore, the
suggested design release fraction for the linear power density equivalent to the peak power
density of 20.5 kW1ft (i.e., 53.5 Wlcm) is approximately 5 percent less than thevalue
recommended by p. 25.2 of Regulatory Guide 1.25 according to Figure 8-9 in Graves (1979).
Other results provided in NUREGICR-0722 (Lorenz et al. 1980, Table 5) indicate lower release
fractions than are provided by p. 25.2 of Regulatory Guide 1.25 for Xe and Kr based on
diffusional migration through the fuel matrix (i.e., 1.27 percent calculated fission gas gap
inventory versus 8 percent as recommended in WASH-1400 (NRC 1975, p. VII-13 and ,
Table VII-1-1). Similarly, the value in SANDSO-2124 (Wilmot 1981, Table XVIII, p. 36) for
burst ruptures shows lower release fractions for "Kr (22 percent) than are established on p. 25.2
of Regulatory Guide 1.25 (30 percent) for commercial SNF.
131 . InNUREGICR-5009 (Baker et al. 1988, Table 3.6), the release fraction of I is shown to
increase with burnup and for a fuel rod with a burnup of 60,000 MWdIMTU the release fraction
is stated to be 0.12, which illustrates the conservative nature of the 0.3 value on p. 25.2 of
Regulatory Guide 1.25. Thus, these points illustrate the conservatism associated with the release
fractions stipulated on p. 25.2 of Regulatory Guide 1.25.
Although not specifically addressed on p. 25.2 of Regulatory Guide 1.25 or in SANDSO-2124
(Wilmot 1981, Table XVIII, p. 36), the ARF for 'H is conservatively assumed to be equal to the
maximum ARF for a gas (i.e., the noble gas release fraction) as 'H will be released in a gaseous
129 . form. In addition, I is conservatively treated as a gas (Section 5, Assumption 5.12) and the
129 . ARF for I is conservatively assumed to be equal to 0.3, as noted in Regulatory Guide 1.25 and
reasoned in NUREGICR-6487 (Anderson et al. 1996, p. 30), because of its low boiling
temperature of 184°C (Weast 1972, p. B-17).
Surnmarv for Gases
The recommended ARFs for the gaseous radionuclides "Kr, 3 ~ , and ' 2 9 ~ are summarized in the
last column of Table 3. These values, deemed conservative, are based on values from p. 25.2 of
Regulatory Guide 1.25 and in NUREGICR-6487 (Anderson et al. 1996, p. 30), and are consistent
with the values used to evaluate transportation packages containing spent nuclear fuel as stated in
NUREG-1617 (NRC 2000a, Table 4-1). In addition, with the exception of the ARF for '29~,
these values are also consistent with the ARFs used to evaluate dry storage cask systems in
NUREG-1536 (NRC 1997, Table 7.1) and spent fuel dry storage facilities in NUREG-1567
(NRC 2000b, Table 9.2).
000-00C-MGRO-01700-000-000 25 of 84 October 2004

Commercial SNF Accident Release Fractions
6.2.1.2 ARFs for Commercial SNF Volatiles and Fuel Fines
Fuel fines and volatiles could also be liberatedcreated from fuel pellets because of the shaking
of the rod and grinding action between fuel pellets that occurs during handling and transport of
the fuel. Fuel fines exist as residual from the fuel manufacturing process and are produced
during irradiation from pellet cracking that is associated with thermal distortion caused while
the fuel was at high temperatures. In the latter case, the higher temperature at the center of a
fuel pellet than at the periphery produces circumferential tensile stresses that produce radial
pellet cracks.
In this analysis, the high melting and boiling temperatures of elemental Sr and of common
compounds associated with the element in commercial SNF allow 9 0 ~ r to be treated as particulate
(i.e., fuel fines) (Section 5, Assumption 5.14). Because of the presence of Ru04, '06~u is treated
as a volatile (Section 5, Assumption 5.13). The low melting point of elemental Cs means that the
radionuclides ' 3 4 ~ s and I3'cs will be treated as volatiles in this analysis in order to address
volatile Cs that may exist in partial pressure chemical equilibrium (Section 5, Assumption 5.1 1).
ARFs for Fuel Fines
For fuel fines, an average of the release fractions from the four burst rupture tests performed in
NUREGICR-0722 (Lorenz et al. 1980, p. 101, Table 40)~ is equal to 2.42 x This value is
based on release measurements from the bursting of 1-ft segments of fuel rods. Assuming that
the same amount of mass is released from a full-length fuel rod as was released from the test
segment, SAND80-2124 (Wilmot 1981, pp. 34 and 35) divided the average release fraction by a
factor of 10, because a typical spent fuel rod would have roughly ten times the mass of fission
products as would the 1-foot test section.
Identical adjustments to the burst rupture data are made in NLTREG-1536 (NRC 1997,
Table 7.1), NUREG-1617 (NRC 2000a, Table 4-I), NUREG-1567 (NRC 2000b, Table 9.2), and
NUREGICR-6487 (Anderson et al. 1996, Table 6-2) to amve at the comparable release fractions
for fuel fines cited in those references. The validity of this correction may be considered
unjustified based on the following arguments:
The released internal pressure of the full-length fuel rod is expected to entrain particles
from regions other than those directly near the burst point and carry them out of the
fuel rod. Hence, the release fractions for these particles from the full-length fuel rod
could be larger than the fractions from the 1-ft test segment if all particles were to
escape.
4 Burst rupture data was accumulated from tests performed at temperatures between 900°C and
1200°C and the internal pressure of the helium inserted into the fuel pin at the time of rupture
was approximately 2.0 MPa.
October 2004

Commercial SNF Accident Release Fractions
The larger volume of a full-length fuel rod is expected to result in a gas exhaust that is
sustained over a longer period of time, albeit a very short time, than the exhaust of the
I-foot fuel rod segment. Although the pressure in a full-length fuel rod is less than the
fuel rod pressure that causes a burst rupture, the larger volume of the full-length fuel rod
will likely sustain a gas exhaust over a longer period of time, which may allow for more
particles to be transported through the fuel pellet-clad gap and released out the break.
A guillotine break or a longitudinal split of a fuel rod would produce a significant
increase in the breach of the confinement compared to the breach caused in a burst
rupture. This has the potential to significantly increase the amount of fuel fines that may
escape a fuel rod.
The first two preceding arguments may be countered based on the trends of particle deposition in
turbulent flow through vertical tubes from Table B3 of ANSI N13.1-1969, and characteristics of
the flow paths these particles must travel through. The following arguments were considered to
validate this interpretation (Wilmot 1981, pp. 34 and 35) of the rod burst data and to ensure that
it is conservative:
The fraction of particle deposition (independent of particle size) increases as tube
diameter decreases (ANSI N13.1-1969, Table B). For a fuel rod, the flow paths for
exhausted gases consist of an annular gap (formed by the fuel pellet on the inside and
the cladding on the outside) and any penetrating cracks and/or crevices through the fuel.
The equivalent diameter of these flow paths is expected to be very small (e.g., less than
approximately 0.03 cm for typical commercial SNF). Thus, the fraction of deposition is
expected to be high for both the full-length fuel rod and the 1-ft fuel rod segment with
only particles local to the breach being released. This suggests that there will be no
difference in the mass released from the full-length and 1-ft fuel rod segments.
The fraction of particle deposition (independent of particle size) increases with the
length of the tube (ANSI N13.1-1969, Table B). Thus, the further a particle is from a
cladding breach the larger its probability is for being deposited along the tube before it
can be released. For a full-length fuel rod, this trend reduces the amount of mass
(theorized to be released from the unadjusted burst rupture data) released through a
cladding breach. The actual deposition fraction cannot be established because the
density of the fuel fines released in this analysis is assumed to be 10.96 &m3
(Section 5, Assumption 5.6), which is greater than the values presented in Table B3 of
ANSI N13.1-1969. This characteristic of particle transport in a flowing gas stream
supports the assumption that higher proportions of particulate present near a break will
be released relative to particulate present at large distances from the break location.
000-00C-MGRO-01700-000-000 27 of 84 October 2004

Commercial SNF Accident Release Fractions
The fraction of particle deposition increases with the size of the particle (ANSI
N13.1-1969, Table B). Thus, the larger fuel fines released in the burst rupture
experiments, which comprise the vast majority of the mass of the released fuel, were
likely located in the immediate vicinity of the cladding breach. For the full-length fuel
rod, the larger diameter, high-density fuel fines are not likely to escape the fuel rod and
hence, no significant change in the released mass from the burst rupture data is
expected.' This further supports the reduction in the burst rupture release fractions.
The internal gas pressure of a cool fuel rod is less than the pressure that occurred during
the burst rupture tests. Thus at the onset of an event that breaches the fuel cladding, the
1-A fuel rod segment will have a higher exhaust velocity than the full-length fuel rod.6
Because velocity and the Reynolds number are directly proportional, this higher velocity
results in a larger Reynolds number which, based on the general trends for turbulent
flow through vertical tubes in Table B3 of ANSI N13.1-1969, results in less particles
being deposited. Thus, more particles are expected to be exhausted resulting in a higher
release fraction for the 1-foot fuel rod segment.
The flow paths in a fuel rod are not likely to be smooth and continuous. Fuel pellet
irradiation induced cracking produces non-smooth flow paths for the mixture of fill
gases and fission gases and the particles entrained in those gases and pellet-to-clad
interference is likely to occur randomly over the length of a fuel rod segment.
Deposition is expected to increase per unit flow length through these paths because of
the affinity of the deposited particles and the larger particles to adhere to or plate-out on
these surfaces.
The final argument against dividing the experimental burst rupture release fractions by a factor
of 10 concerns the potential of radionuclide releases that are equal to or larger than the burst
rupture from a guillotine break or a longitudinal split of a fuel rod. In this analysis, 100 percent
of the fuel rods are assumed to fail because of a credible accident event. The failure of 100
percent of the fuel rods involved in an event by a guillotine break or a longitudinal split is not
considered credible within the scope of this analysis (Section 5, Assumption 5.2).
5 Page 36 of ANSI N13.1-1969 states that for particles larger than those given in Table B3 (i.e.,
greater than 10 pm), significant re-entrainment is expected at higher flow rates. Neither the
inner clad nor pellet surfaces, however, are expected to be smooth. In addition, the pressure
transient time history is characterized by rapidly reducing pressure and flow. Hence, any
particle that is deposited on internal surfaces is likely to remain adhered to it during the short
duration of the pressure transient, making re-entrainment an insignificant consideration.
In the worst-case condition of a large break or rupture (i.e., pinhole leaks or hairline cracks
would not produce significant flow rate to promote particulate entrainment), the pressure
difference between the full-length he1 rod and 1-fi he1 rod segment would be short-lived, as
eventually both pressures would quickly equilibrate with the environment outside of the
fuel rod.
000-00C-MGRO-01700-000-000 28 of 84 October 2004

Commercial SNF Accident Release Fractions
In SAND80-2124 (Wilmot 1981, p. 1 I), an impact rupture is defined as a rupture of the cladding
produced by bending or other deformation of a fuel rod; no mention is made of a guillotine break
or longitudinal split. Intact irradiated fuel is considered quite rugged and capable of sustaining
severe impact environments according to SAND80-2124 (Wilmot 1981, p. 18).
Based on these arguments, dividing the experimental burst rupture release fractions by a factor
of 10 is justified based on regulatory precedents, physical trends evident in Table B3 of
ANSI N13.1-1969, and Assumption 5.2 (Section 5).
Burst Rupture ARFs Avvlied to Imvact Accidents
The release fractions for burst rupture are also expected to produce conservative results when
applied to repository accidents, which are of an impact nature. In SAND80-2124 (Wilmot 1981,
p. 33), it is stated that an impact rupture is expected to produce more particles (through
pulverization and grinding between pellets) than were present in the spent fuel before a burst
rupture. There will be less pressure, however, to exhaust these particles after an impact versus a
burst rupture and it is expected that an impact rupture would have a more restricted release
pathway because of the cladding deformation (Wilmot 1981, p. 33). Indeed, as reported in
SAND80-2124 (Wilmot 1981, p. 33), burst rupture release fractions were arbitrarily reduced by a
factor of 10 to account for this physical expectation. It is apparent, however, that in more recent
NRC guidance, such as (NRC 1997, Table 7.1), release fractions associated with burst ruptures
(corrected for fuel rod length) should not be reduced if applied to impact accidents.
ISG-5 and NUREGICR-6487
JSG-5 (NRC 2003, Table 7.1), NUREG-1617 (NRC 2000a, Table 4-l), and NUREGICR-6487
(Anderson et al. 1996, pp. 30 and 31) provide release fractions to evaluate normal, off-normal,
and hypothetical accident doses for storage casks. These release fractions for fuel fines are the
average ARFs (corrected for fuel rod length) from the burst rupture tests in NUREGICR-0722
(Lorenz et al. 1980, Table 40) conservatively rounded-up (i.e., 2.4 x is rounded up to
3 x lo-'). This use of the burst rupture ARFs for both the transportation cask (NRC 2000a,
Table 4-1) and the storage cask (NRC 2003, Table 7.1) hrther demonstrates the applicability of
these data to credible accidents involving commercial SNF. Based on these arguments, an ARF
of 3 x lo-' is recommended to be conservatively applied to estimate the release of non-volatile
fuel fines during accidents at the repository.
October 2004

I Commercial SNF Accident Release Fractions
NUREGJCR-6487 (Anderson et al. 1996, p. 30) stated that Sr, Cs, and Ru are treated as volatiles
in accidents involving transportation packages. For this analysis, with the exception of Cs and
Ru, Sr is considered as fuel fines (Section 5, Assumptions 5.13 and 5.14). Event sequences
involving fuel assemblies at the repository surface facilities are not expected to involve high
enough temperatures for melting or volatilizing of these elements to be of concern, unlike some
potential transportation accidents, such as the fire described in 10 CFR 71.73(~)(4) (Section 5,
Assumption 5.10). Thus, the ARF associated with Sr a able 3), is consistent with the ARF
values associated with the fuel fines in NUREG-1617 (NRC 2000a, Table 4-1) and
NUREGICR-6487 (Anderson et al. 1996, p. 31).~
I ARF for Cs
The ARF value for the Cs radioisotopes ' 3 4 ~ s and I3'cs in NUREGJCR-6487 (Anderson et al.
1996, p. 30) and in NUREG-1617 (NRC 2000a, Table 4-1) is the burst rupture test data for Cs
from NUREGJCR-0722 (Lorenz et al. 1980, Table 40)' uncorrected for fuel rod length. Because
Cs is considered a volatile in these NUREGs, it is treated similar to the fission and fill gases
found in the fuel pin (Section 5, Assumption 5.11). These gases are considered to be fully
purged from the gap and plenum regions during an event that breaches the cladding. Thus, no
correction for fuel rod length to the measured release fraction is made for the potentially volatile
Cs radioisotopes. Because this burst rupture value is uncorrected for fuel rod length or for
impact rupture, meaning that the value is reduced by an additional 90 percent according to
SANDSO-2124 (Wilmot 1981, p. 33), its use to estimate the release of Cs during accidents at the
repository is recommended and considered conservative.
In comparison, the Cs ARF values in NUREG-1536 (NRC 1997) are derived from the burst
rupture data for fuel fines presented in NUREGICR-0722 (Lorenz et al. 1980, Table 40) and are
corrected for fuel rod length (i.e., divided by a factor of 10). These data were not used in
SANDSO-2124 (Wilmot 1981, Table XII) for burst rupture release fractions, as shown in
Table 2, but were used in SANDSO-2124 (Wilmot 1981, .Table XII) to determine the release
fractions for an impact rupture.9
' Although volatile ruthenium was detected as a volatile in some tests reported in
NUREGJCR-0722 (Lorenz 1980, pp. 116 to 119), in the burst rupture tests, the volatile
ruthenium was negligible compared to the ruthenium contained in the fuel fines (Lorenz 1980,
p. 119). Given the results from these tests, ruthenium is considered a volatile in this analysis
per Assumption 5.13 and ISG-5 (NRC 2003, Table 7.1). ' An average of the Cs released in these burst rupture tests (Lorenz et al. 1980, Table 40) is
determined to be 0.0306 percent, which is higher than the 0.02 percent used in
NUREGICR-6487 (Anderson et al. 1996, Table 6-2, pp. 30 and 31).
The release fractions for impact rupture are the release fractions from burst rupture reduced by
90 percent to 10 percent of the burst rupture release fractions. SAND80-2124 (Wilmot 1981,
Table XII) provides burst rupture experiments from NUREGJCR-0722 (Lorenz et al. 1980,
Table 40).
October 2004

Commercial SNF Accident Release Fractions
The burst release fractions for ' 3 4 ~ s and I3'cs in SANDSO-2124 are calculated using "Fission
Product Source Terms for the Light Water Reactor Loss-of-Coolant Accident" (Lorenz et al.
1979, p. 406, equation I), which has several parameters that were evaluated using data from
pressure rupture tests. The tests were performed at temperatures (700°C to 900°C) greater than
those that exist for commercial SNF discharged from a reactor core a minimum of five years
(cladding surface temperatures are limited to approximately 400°C in shipping casks under
normal handling conditions). Hence, the ARFs produced in Lorenz et al. (1979, p. 406) are
higher than those expected at the repository.
Summary for Fuel Fines and Volatiles
The recommended ARF values for fuel fines and volatiles are summarized in the last column of
Table 3. These values are consistent with the ARF values presented in NUREG-1617
(NRC 2000a, Table 4-I), which are based on the ARF values for transportation casks listed in
NUREGICR-6487 (Anderson et al. 1996, p. 31). Although Cs is treated as a volatile, the
recommended ARF values assume accidents at the repository occur at temperatures where Sr
does not melt or boil (i.e., temperatures below approximately 670°C) (Section 5, Assumptions
5.10 and 5.14).
6.2.1.3 ARFs for Commercial SNF Crud
Crud releases originate from the surface of a fuel rod. In contrast to fuel fines, gases, and
volatiles released from a fuel rod, the crud release fraction is not based on the fraction of fuel
rods that are breached. The release mechanism involves surface spallation rather than leakage
past fuel cladding barriers. Crud is primarily composed of iron-based compounds and some
-'nickel, copper, cobalt, chromium, manganese, zinc and zircalloy. The actual amount and type of
crud varies from reactor to reactor and from cycle to cycle.
Crud becomes radioactive through neutron activation. The nuclear industry recognizes that
excessive crud negatively affects fuel performance and it is making a concerted effort to control
factors that contribute to crud formation. Thus, crud accumulation on older fuel rods is expected
to be greater than the crud on fuel discharged over the last several years. The crud activity on
older fuel rods, however, is less than that on freshly discharged fuel rods because of the
relatively short half life of the radionuclides that contribute to crud activity.
In general, PWR fuel has less crud activity than BWR fuel (Sandoval et al. 1991, p. 2). Crud can
also be classified into two general categories: (1) fluffy, easily removed crud composed mostly
of hematite (Fez03) and (2) a tenacious crud that is tightly bound to the rods composed mostly of
spinel (NiFe204). PWR crud is primarily of the second category while BWR crud is composed
of both types.
ARFs for Crud
Mishima and Olson (1990, p. 1134) examined measured data on crud spalling as a result of
various mechanical forces during the fuel rod consolidation process and derived values of
4.1E-04 for the crud spallation fraction (CSF) during rod consolidation and handling, and
2.2E-07 for the fraction of crud airborne (i.e., CSF x ARF). This suggests that only a small
000-00C-MGRO-0 1700-000-000 31 of 84 October 2004

Commercial SNF Accident Release Fractions
fraction of the crud that spalls off rods becomes airborne during the rod consolidation process.
The crud ARF is equal to 5.4E-04 (= 2.2E-0714.1E-04). This value is bounded by an ARF of
1.OE-01 for venting gases in a pressurized volume in which loose powder of surface
contamination exists (DOE 1994, p. 5-22). The ARF of 1.OE-01 is also a bounding value for the
suspension of loose surface contamination by vibration shock (DOE 1994, Section 4.4.3.3.1).
Sandoval et al. (1991, pp. 21 to 22, Table 1, p. 1-50) provides measured release fractions for the
crud located on the outer surface of fuel rods. Sandoval et al. (1991, Table 4 on p. 24) shows
spallation fractions for transportation conditions. The maximum crud spallation fraction for both
PWR and BWR fuel occurs at elevated fuel temperatures (450°C) where a 0.15-m long region of
crud is assumed to bubble up and spall off, resulting in a crud spallation fraction of 15 percent.
The elevated fuel temperatures are caused by exposing fuel to a fire at a temperature of 800°C
for 30 minutes. At lower temperature conditions (i.e., fuel rod surface temperature of 300°C), a
maximum of 8 percent of the crud may spall off. Therefore, the bounding CSF value is 0.15,
which is much larger than the CSF of 4.1E-04 for the rod consolidation process.
The release fraction for crud found in SAND80-2124 (Wilmot 1981, p. 34) was an estimate that
has no experimental basis. This estimate is based on an assumption that 25 percent of the crud
that plates on spent fuel assemblies during reactor operation and contaminates the transportation
cask cavity surface is loosely adhering. The remainder adheres tightly and requires abrasion and
chemical treatment for removal. SAND80-2124 (Wilmot 1981, p. 34) assumes that this 25
percent release fraction includes not only the crud on the cladding surface, but also crud on the
assembly structure and the cask interior. This report goes on to state that the 25 percent release
fraction from the cask is comparable to the release of all the crud located on the surface of the
fuel rod cladding (Wilmot 1981, p. 34). Thus, the 25 percent crud release fraction from
SAND80-2124 (Wilmot 1981, p. 34) equates to 100 percent release of the crud on the fuel rod
cladding surfaces.
In NUREG-1617 (NRC 2000a, Table 4-I), the crud spallation fraction under hypothetical
accident conditions is unity. This value is based on assumptions made in NUREGICR-6487
(Anderson et al. 1996, p. 28). No justification is given for this assumption in either document.
Therefore, the bounding CSF value of 0.15 instead of 1.0 will be recommended for use at the
repository.
Using a CSF of 0.15, a crud ARF of 1.OE-01, the crud effective ARF for an event sequence is:
Crud effective ARF = CSF x ARF = 0.15 x 1.OE-01 = 1.5E-02 0% 4 4
Using a crud RF of 1.0, the crud respirable release fraction is 1.5E-02. This crud respirable
release fraction is much larger than crud respirable release fractions of 1.3E-06 and 4.7E-08
found in SAND99-0963 (Luna et al. 1999, p. 7) calculated for sabotage attacks of rail casks
employing high energy density devices. These devices have energy densities approximately one
thousand times higher than potential cask drops at the repository.
October 2004

Commercial SNF Accident Release Fractions
Co-60 Crud Surface Activities
A summary of the crud surface activities found in the literature is summarized in Table 4.
The crud surface activities in NUREG-1617 (NRC 2000a, Table 4-1) and NUREG-1536 (NRC
1997, Table 7.1)" are based on data in Sandoval et al. (1991, Table 1 on p. 15). These values are
in turn based on measured maximum spotAocal surface (vs. total surface average) activities of
6 0 ~ o on fuel rods freshly discharged from the reactor. These values may be considered
conservative in two manners:
At the repository, fuel must be cooled a minimum of 5 years following discharge from
the reactor core before it will be accepted as required by 10 CFR Part 961. Thus, the
surface activity should be approximately half of the cited values, based on "CO being
the principal radionuclide in the crud and its 5.271-year half life (Panington et al. 1996.
p. 25).
The average fuel rod will not have the maximum spotflocal surface activity. In fact,
Sandoval et al. (1991, Figures 3 and 4 on pp. 21 and 22) show that the majority of
examined fuel rods (greater than 60 percent) had little or no crud activity associated with
them at fuel discharge. In addition, the maximum spot/local activity density is
approximately 7 times the value of the average activity density (Sandoval et al. 1991,
p. 1-50), Thus, using the maximum spot/local surface activity for a typical fuel assembly
is conservative (Section 5, Assumption 5.3).
Thus, a reasonably conservative treatment of the amount of crud released should include an
.approximate 50 percent reduction in the maximum spot/local activity density cited in
PUREG-1617 (NRC 2000a, Table 4-1). The 50 percent reduction is justified based on an
approximately 50 percent reduction because of the fuel discharge age. The crud surface activity
density is a function of time after discharge fiom the reactor. The time-dependent crud surface
activity is based on the following radioactive decay equation (BSC 2004c, p. 23):
N(t) = N(0) exp (-t x In 2 l tp)
where, I
N(t) = crud surface activity at time t,
N(0) = crud surface activity at time 0,
lo Crud surface activity is 140 p ~ i / c r n ~ for PWRs in NRC (2000a, Table 4-1; 1997, Table 7.1).
Crud surface activity, however, is 600 pcilcm2 for BWRs in NRC (1997, Table 7.1), which
ignores data from the Tsuruga reactor and 1254 pcilcm2 in NRC (2000a, Table 4-I), which
includes data from the Tsuruga reactor.
October 2004

I Commercial SNF Accident Release Fractions
- t1'2 - radionuclide half-life in years
t = the decay time in years.
Using Equation 4b for PWR assemblies, the OOCO crud surface activity afier 5 years decay is:
N(t) = 140 p~i/cm' exp (-5 y x In 2 1 5.271 y) = 72.5 w~i/cm2
Using Equation 4b for BWR assemblies, the "CO crud surface activity after 5 years decay is:
Nit) = 1,254 p~i/cm2 exp (-5 y x in 2 15.271 y) = 649.7 wci/cm2
Fe-55 Crud Surface Activities
In the vast majority of the literature (e.g., NRC guidance documents, Sandia National
Laboratories reports, safety analysis reports [SARs] for storage casks, and SARs for independent
spent fuel storage installation), the principal radionuclide in the crud of fuel that has been
discharged for a minimum of 5-years is 6 0 ~ o . However there is some evidence that 5 5 ~ e may
also be a large component of the crud activity. An isotopic analysis of crud from the Carolina
Power & Light Brunswick nuclear power plant (BWR SNF) revealed the presence of a
significant amount of 5 5 ~ e (Jones 1992, pp. 6-7 and cited Ref. 6). The presence of this isotope is
plausible because of the significant presence of iron in crud, usually as hematite. Specifically,
the production of 5 5 ~ e is through neutron activation of 5 4 ~ e (5.9 percent natural abundance).
Thus, consideration of this isotope in the crud is justified.
At the time of discharge, the surface activity of 5 5 ~ e present was determined to be 7,415 p~i/cm2
(Jones 1992, Table 2). This value is significantly greater than the BWR values of 600 p~i/cm2
given in NUREG-1536 (NRC 1997, Table 7.1) and 1,254 p~i/cm2 given in NUREG-1617 (NRC
2000a, Table 4-1). The half life of "Fe, 2.73 years (Parrington et al. 1996, p. 25), is about half
of the half life of 6 0 ~ o (5.271 years). So approximately a reduction of 72 percent can be assumed
for 5 5 ~ e crud. This reduction is based solely on the half life argument as used previously for the
60 Co-based crud and therefore is applicable to bare unconfined fuel assemblies and confined fuel
assemblies. For PWR fuel, the "~e-based crud surface activity at the time of discharge from the
core is assumed equal to 5,902 p~ilcm2 (Section 5, Assumption 5.7).
Using Equation 4b for PWR assemblies, the 5 5 ~ e crud surface activity after 5 years decay is:
N(t) = 5,902 li~i/cm2 exp (-5 y x in 2 / 2.73 y) - 1,658 p~i/cm2
Using Equation 4b for BWR assemblies, the " ~ e crud surface activity after 5 yeam decay is:
N(t) = 7,415 y ~ i / c ~ n 2 exp (-5 y x ln 2 / 2.73 y) = 2.083 p~i/cni'
October 2004

0
0 I i
0 Table 4. crud 'surface Activities
I I Crud Surface Activity (p~ilcm2) I NUREGICR- I I I NUREG- I NUREG- I NUREG- 6487
NOTES: 'Data are taken from ISG-5 (NRC 2003, Table 7.1). ' ~ a t a are taken from SAND88-1358 (Sandoval et al. 1991. Table 1-12).
'These values have been corrected for half life of %o over 5 years.
'~hese values have been reduced using the half life of " ~ e over 5 years
'Jones 1992. Table 1. T = 0 (fresh discharge)
'Jones 1992. Table 2. T = 0 (fresh discharge)
W
BWR = boiling water reactor; PWR =pressurized water reactor
VI
%
m
P

Commercial SNF Accident Release Fractions
Summarv for Crud
The recommended ARF for crud, which consider 6 0 ~ o and 5 5 ~ e as the principal radionuclides
present, is assumed to be 1.5E-02, and is summarized in the last column of Table 3. Table 4
summarizes the surface activities to be used with each crud isotope. NUREGICR-6487
(Anderson et al. 1996, p. 28) states that these surface activities should be multiplied by the
surface area associated with a rod to obtain the crud activity. Thus, the surface area associated
with the assembly hardware (e.g., fuel rod spacers, nozzles, etc.) is neglected.
6.2.2 RF for Intact Commercial SNF
The RF is the fraction of the initially suspended airborne material (material at risk multiplied by
the ARF) that can be inhaled. The actual fraction of commercial SNF that is respirable depends
on the p&iculate size distribution of the aerosolized commercial SNF. For nuclides in the form
of a gas or a volatile, 100 percent are considered respirable because they are assumed to be gases
(i.e., 100 percent aerosolized).
The extent of fuel fines and crud that are respirable is dependent on the size distribution of the
aerosolized particulate. For consistency with the dose conversion factors used in dose analyses
for the repository, the fraction of respirable particulate will be based on the fraction of a
distribution with an AMAD less than a specified value. The dose conversion factors are based
on an AMAD of 1 pm, however the RF will be conservatively based on the fraction of
particulate with an AMAD of 10 pm or less. The combination of dose conversion factors based
on an AMAD of 1 pm and an RF based on an AMAD of 10 pm will ensure that inhalation doses
are conservatively calculated (Section 6.2.2.1). Establishing the RF for commercial SNF fuel
fines is limited to analyzing the same two primary groups of experimental data that produced the
majority of the cited ARFs for commercial SNF (Table 2) as follows:
0 Four experiments that burst ruptured highly irradiated commercial SNF rod segments
in a flowing steam environment. These experiments quantified and characterized
fission product release under conditions postulated for a spent-fuel transportation
accident (Lorenz et al. 1980).
0 Two single energy density impaction tests on three unconfined U02 pellets. These
tests characterized the size distribution and the RF of the fragments generated
(Mecham et al. 1981; Jardine et al. 1982).
In this section, a methodology to produce an RF for particulate with a given lognormal sizelmass
distribution is established. First, an estimate of the respirable particulate size is established for
an aerosol (i.e., an AMAD of less than 10 pm). Based on this particulate size, relationships are
then established between a distribution's MGD, MMD, and AMAD. Finally, the distribution of a
typical aerosol of commercial SNF fuel fines is characterized to allow for the calculation of the
commercial SNF RF. A similar analysis to determine the RF for crud is performed using
experimental data in Sandoval et al. (1991, pp. 23 to 26).
000-00C-MGRO-01700-000-000 36 of 84 October 2004

Commercial SNF Accident Release Fractions
6.2.2.1 Size of Respirable Particulate
When radioactive aerosols are inhaled, parts of the respiratory system are irradiated. Other
organs and tissues of the body are also irradiated by radiation originating from the lungs and as a
result of translocation of inhaled material to body tissues kom the respiratory system.
After inhalation of radioactive aerosols, the doses received by various regions of the respiratory
system will differ widely, depending on the size distribution of the inhaled material.
ICRP, DOE Handbook, and American Nuclear Society Standard Res~irable Definitions
Figure 1 is reproduced from Publication 30 (ICRP 1979, Figure 5.1) that is applicable to aerosol
distributions with AMADs between 0.2 and 10 pm and with a geometric standard deviation of
less than 4.5." Figure 1 shows the effect aerosol size distributions have on their deposition in
the respiratory system. Aerosol distributions with an AMAD of less than 10 pm have significant
deposition in the lungs (line labeled Dp in Figure 1) and, hence, may significantly contribute to
internal doses. Aerosol distributions with an AMAD distribution greater than 10 pm do not
deposit as readily in the lungs (deposition occurs almost entirely in the nasal passage) and,
hence, do not contribute as readily to lung doses. This trend can be seen in the lung dose
conversion factors, which decrease as the AMAD increases from 1 to 10 pm (ICRP 1979, Figure
5.1). Thus, the combined use of an AMAD of 10 pm and a dose conversion factor based on an
AMAD of 1 pm represents a conservative methodology, when compared to RFs and dose
conversion factors both based on 10 pm (or both based on 1 pm). According to
DOE-HDBK-3010-94 (DOE 1994, pp.1-4 to 1-6), several other definitions of respirable particles
have been presented by various groups at different times, as follows:
- Particles with terminal velocities equal to that of a 5-pm diameter particle were
considered respirable dust by the British Medical Research Council in 1952.
Particles with a 50 percent respirable cut-size of 3.5-pm AED were considered
respirable dust by the US. Atomic Energy Commission.
Particles with a 50 percent respirable cut-size of 2 - p AED were considered respirable
dust by the American Conference of Governmental Industrial Hygienists.
I ' The only data considered in this analysis that have a standard deviation greater than 4.5 is the
pulverization data for brittle materials (Mecham et al. 1981; Jardine et al. 1982). These data
are not used in the calculation of the RFs for commercial SNF or crud aerosols.
000-00C-MGRO-01700-000-000 37 of 84 October 2004





Commercial SNF Accident Release Fractions
Particles with a 50 percent respirable cut-off at 15-pm AED were considered inhalable
dust (particles entering the upper respiratory airway and entering the thorax) by
the EPA.
Particles with a 50 percent respirable cut-size at 10-pm AED were considered inhalable
dust @articles entering the nasal or oral passages) by the International Standards
Organization-Europe.
An AED of 10 pm adequately represents the cut-off diameter for respirable particulate (DOE
1994, pp. 1-4 to 1-6). This value is further supported by Appendix B2.1.4, p. 19, of ANSUANS-
5.10-1998, which states that the respirable fraction "is commonly assumed to include particles
10 pm Aerodynamic Equivalent Diameter (AED) and less as a conservative approximation."
AMAD versus AED
If AMAD and AED were equivalent, then a diameter of 10 pm would conservatively represent
the cut-off diameter for respirable particulate. According to DOE-HDBK-3010-94 (DOE 1994,
p. xviii), however, the AED is equivalent to the diameter of a sphere of density 1 &m3 that
exhibits the same terminal velocity as the particle in question. In this case, the particle in
question represents the cut-off diameter for respirable particles (&,/,). In terms of the terminal
settling velocity (v-), this definition can be written as:
where
p,,, is the density of the particulate in the aerosol
Whereas, according to Publication 30 (ICW 1979, p. vii), the AMAD is equivalent to the
diameter of a unit sphere (1 &m3) with the same terminal velocity in air as that of the aerosol
particle whose activity is the median for the entire aeros01.'~ This activity median diameter will
be shown in the next section to be equivalent to the MMD assuming that activity and mass of the
particle are proportional. In terms of the terminal settling velocity (v,,), this definition can be
written as:
12 The definition of AMAD in DOE (1994, p. xviii) is not the same as the definition of AMAD
in ICRP (1979, p. vii). The DOE (1994, p. xviii) defines AMAD as the diameter of the
particle for which half of the activity is associated with particles larger than and half the
activity associated with particles smaller than this size particle. No mention is made of the
unit density sphere in this definition and this definition is actually equivalent to the
aerodynamic mass median diameter because activity and mass are proportional.
October 2004



Commercial SNF Accident Release Fractions
Thus, these two characteristic diameters, AED and AMAD, are similar in definition (i.e., both
are based on unit density). The only difference between AMAD and AED is the particles they
share terminal velocities with (i.e., the respirable particle cut-off diameter versus the diameter
representative of the median activity for the entire aerosol). Hence, by selecting a diameter of
10 pm for AED and AMAD, the preceding relationships indicate that the value of the respirable
particle cut-off diameter must be equivalent to the diameter of the particle representing the
median activity of the entire aerosol. However this does not necessarily indicate the RF from
these two interpretations (AMAD vs. AED) are equivalent. In fact, it will be shown that the RF
from the AED method will be equal to or less than the RF associated with the AMAD method.
Consider a lognormally distributed aerosol with a density of 10 &m3 and an MMD of 3.16 pm.
Using the expression for the terminal settling velocity in the next section, the preceding
relationships reduce to:
Hence, if AMAD and AED are set equal to the maximum respirable diameter of 10 pm, then
MMD and 41, calculated from the preceding expressions are equal to approximately 3.16 pm.
Because this value is equal to the MMD of the entire distribution, 100 percent of this aerosol is
considered respirable according to the AMAD method and 50 percent of this aerosol is
considered respirable according to the AED method (50 percent of the total mass has a diameter
less than 3.16 pm, because 3.16 pm is the mass median diameter of the distribution).
If the MMD of the whole distribution in the preceding example had been less than the calculated
values for MMD and &I,, then the AMAD method would still have an RF of 100 percent because
the AMAD of the whole distribution is less than 10 pm. The RF for the AED method would
increase in this case (i.e., greater than 50 percent) because the AED would be greater than the
MMD of the whole distribution. Obviously, however, the RF from the AED method would
never exceed the RF of 1 from the AMAD method. On the other hand, if the MMD of the whole
distribution in the preceding example had been greater than the calculated values for MMD and
bo, then the RFs for both methods would decrease. However, the RF from the AMAD method
would always be greater than the RF from the AED method, because the AMAD method will
determine the fraction of the whole distribution that has an AMAD of 10 pm and less. This
fraction of the whole distribution will almost always include particles with diameters greater than
3.16 pm (except in cases with extremely large MMDs), whereas the AED method will only
include those particles with diameters less than 3.16 pm.
October 2004



Commercial SNF Accident Release Fractions
Summary of Respirable Diameter
Therefore, for this analysis, respirable particulates are those with a particulate size distribution or
any fraction thereof, with an AMAD of less than 10 pm. This ensures consistency with the
inhalation dose conversion factors used in dose analyses for the repository, which are also based
on a maximum AMAD. These factors are calculated in Federal Guidance Report No. 11
(Eckerman et al. 1988, p. 14) and are conservatively based on an AMAD of 1 prn.l3 Thus, the
combination of dose conversion factors based on an AMAD of 1 pm and an RF based on the
mass fraction of particulate that have an AMAD of less than 10 pm are considered to produce a
conservative methodology for inhalation dose calculations.
6.2.2.2 Relationships between MGD, MMD, and AMAD
Definitions
Before relationships can be established between the MGD, the MMD, and the AMAD, these
quantities are defined as follows as they are used in this analysis:
AMAD-the diameter of a unit density sphere (1 &m3) with the same terminal settling velocity in
air as that of the aerosol particle whose activity (radioactivity) is the median for the entire aerosol
(ICRF' 1979, p. vii).
MGD-the mean or average geometric diameter of the particle number distribution for an entire
aerosol and for a lognormal distribution, it represents the size occumng with the greatest
frequency (Dennis 1976, p. 11 1).
the geometric diameter of the particles in a distribution for which half the mass is
associated with particles greater and half the mass associated with particles less than the stated
size (DOE 1994, p. xix).
AMAD and MMD Relationship
To establish a relationship between AMAD and MMD, it must first be recognized that the mass
(m, grams) and activity (A, curies) of a distribution are proportional (Lamarsh 1983, p. 22), for
example:
A a m
'' If an AMAD of 10 pm were used instead of an AMAD of 1 pm, then the inhalation dose
factors for the lung would decrease.
October 2004

I Commercial SNF Accident Release Fractions
where
h is the decay constant (sec-')
N is the number of nuclei present (nuclei)
NA is Avogadro's number (atomdmole)
M is the gram atomic weight (dmole)
3.7 x 10" is the conversion factor from disintegrations per second to curies.
Equation 9 shows that the particle activity is directly proportional to its mass when all particles
contain the same radionuclide. This relationship establishes the equivalency between the aerosol
particle whose activity is the median for the entire aerosol with the aerosol particle whose mass
is the median for the entire aerosol. Thus, the diameter of the activity median particle is equal to
the diameter of the MMD. Using this relationship and the preceding definitions, the following
relationship between the terminal settling velocities (vtm) of the particulate representing the
MMD and the AMAD of a distribution can be established:
where
p,,, is the density of the particulate in the aerosol.
The terminal settling velocity is determined from Stoke's solution for the drag on a sphere in
creeping flow with correction factors for particulate shape and slip (Sandoval et al. 1991, p. 11):
where
p is the aerosol particulate density (g/cm3)
g is gravitational acceleration (980 cm/s2)
d is the aerosol diameter (cm)
C(Kn) is the Cunningham-Knudsen-Weber slip correction factor (dimensionless)
Kn is the Knudsen number (dimensionless)
h i r is the bulk gas viscosity (g/(cm-s)
K is the dynamic shape factor (dimensionless). K = 1 for a unit density sphere.
Substituting this equation into the previous equation yields:
AMAD = MMD dm
October 2004

Commercial SNF Accident Release Fractions
I Correction Factor and Density Effects
The Cunningham-Knudsen-Weber slip correction factor has the following form (Sandoval et al.
1991, p. 33):
where
Kn is the mean free path of the particle divided by the particle volume equivalent
diameter.
The mean free path of a particle in air is (Holman 1990, p. 632 and equation 12-45):
I where
A is the mean free path in meters
T is the temperature in Kelvin
P is the pressure in pascals.
Figure 2 shows the Cunningham-Knudsen-Weber correction factor (Sandoval et al. 1991, p. 11-5)
as a function of particulate diameter at a temperature of 300 K and a pressure of lo5 Pa. Except
for particulate with diameters less than 2 pm, the Cunningham-Knudsen-Weber correction factor
is nearly equal to unity and hence, the ratio of C001,,,) to C ~ A M A D ) is approximately unity.
1 Thus the relationship between AMAD and MMD reduces to:
!
AMAD Relationship for Commercial SNF
The value for the dynamic shape factor (K) for UOz is assumed to be 1.3 (Section 5,
Assumption 5.5). The particulate density of the aerosol in this analysis will be set equal to the
theoretical density of the SNF POz) which is 10.96 g/cm3 (Section 5, Assumption 5.6).
Figure 3 illustrates the effect density has on the MMD for a fixed value of AMAD of 10 pm.
The density sensitivity of the relationship between AMAD and MMD is insignificant over the
range of densities considered acceptable for fuel particulate (i.e., between 9.5 and 11 g/cm3).
Thus this relationship can be reduced to:
AMAD = 2.9 MMD (for U02) ( ~ q . is)
000-00C-MGRO-01700-000-000 43 of 84 October 2004



I Commercial SNF Accident Release Fractions
I AMAD Relationship for Crud
For aerosols of crud, the dynamic shape factor (K) is equal to 1.3 (Sandoval et al. 1991, p. II-5).
This value is based on visual examination of two scanning electron microscope
microphotographs of Quad Cities crud and implies compact, angular shaped particles.
This value is close enough to unity that it does not have a significant effect on the results
(Sandoval et al. 1991, pp. 31 and 32). The density of crud is equal to 5.2 g/cm3 (Sandoval et al.
1991, Table 9 on p. 33). This value is equal to the density of hematite (e.g., FezO3) (Weast 1972,
p. B-99) that is commonly found on BWR he1 rods and is approximately the density of spinel
(e.g., of the form Ni,Fe3.,04) that is commonly found on PWR fuel rods. Thus, the relationship
between AMAD and MMD for crud reduces to:
AMAD = 2.0 MMD (for crud) (Eq. 19)
Standard Deviation, MMD, and MGD Relationship
The geometric standard deviation, o, is a unitless quantity defined as the ratio between the size
associated with the cumulative mass of 84.1 percent and the median size (50 percent cumulative
mass) or between the median and the 15.9 percent cumulative mass size (DOE 1994, p. xix, and
Dennis 1976, p. 113), for example:
Thus for distributions with small geometric standard deviations (approximately equal to I), the
MMD and the MGD are nearly equivalent (i.e., MMD z MGD). The relationship between the
MGD and MMD for an aerosol with a lognormal distribution can be represented as (Dennis
1976, p. 113):
ln(MMD) = In(MGD) + 3.0 ln (o) (Eq. 21)
6.2.2.3 Characterization of a Commercial SNF Aerosol Particulate Distribution
Having established the size characteristic of respirable particulate, it must now be determined
what fraction of an aerosol produced from an accident involving commercial SNF is respirable.
To perform this task, the type (e.g., lognormal, bimodal) and characteristics (e.g., MGD, standard
deviation) of a particulate distribution must be established for commercial SNF.
6.2.2.3.1 Lognormal Distribution Function
Aerosol particle sizes are usually reasonably well described by one of several mathematical
distribution functions: the normal, lognormal, bimodal and multimodal distribution functions
(Dennis 1976, p. 112). Particles of depleted U02 pellets from pulverization tests in ANL-81-27
(Mecham et al. 1981, Figures 16 and 17, p. 30) showed distributions that were characterized as
000-00C-MGRO-01700-000-000 46 of 84 October 2004

Commercial SNF Accident Release Fractions
lognormal functions with some departure from lognormal at the end of the spectrum.'4
Otherbrittle materials (similar to UOz), pulverized in ANL-81-27 (Mecham et al. 1981,
Figures 16 and 17, p. 30) and ANL-82-39 (Jardine et al. 1982), demonstrated similar lognormal
distributions, indicating that brittle materials often exhibit lognormal distributions when
pulverized.
In addition, Publication 30 (ICRP 1979, p. 24) states that a lognormal distribution of diameters is
typical for aerosols. This assumption is used to calculate the inhalation dose factors tabulated
therein and in Federal Guidance Report No. 11 (Eckerman et al. 1988, p. 14). These tabulated
values are used to calculate doses for the repository.
The lognormal distribution function often represents aerosols produced by attrition processes
(i.e., a wearing away or rubbing-off process that liberates particulate) (Dennis 1976, p. 113).
At the repository, the process of releasing and aerosolizing commercial SNF after a drop or
impact event may best be described as an attrition process. Thus, the particulate distribution of
an aerosol produced from commercial SNF will be represented by a lognormal distribution
function. The lognormal probability distribution function for particle diameters can be written
(Dennis 1976, p. 113):
where
P(d) is the probability distribution function
d is a geometric diameter
MGD is the mean geometric diameter of the distribution
o is the geometric standard deviation.
[P(d) Ad / dl is the probability that a particle has a diameter in the range (d - Adl2) to (d + Ad2).
The number distribution function is given by (Dennis 1976, p. 113):
14 The deviation at the end of the spectrum (i.e., for larger particles) is attributed to: (1) uneven
stresses in the impacted specimens with impact energy being absorbed in highly stressed
zones, leaving the low-stressed zones nearly intact and (2) gross discontinuities at the grain
boundaries allowing for a lot of energy to be absorbed and only creating microcracks
(Mecham et al. 1981, p. 30).
October 2004

Commercial SNF Accident Release Fractions
Figure 4 illustrates a typical differential [nD] and -integrated [n'] lognormal number distribution:
where
the extended trapezoidal rule (Press et al. 1992, p. 127) was used to numerically integrate
the preceding integral from zero to some diameter d, where:
1=1,2 ,..., K-1,K.
K is the total number of integration
In addition, the various diameters discussed in Section 6.2.2.2 and the count mode diameter
(i.e., the most frequent count diameter) are shown on Figure 4 as well.
IS Because of the large range of diameters considered in this analysis (i.e., lod to 10+'-~m
diameters are used for the fuel fines), a non-uniform mesh spacing was used to perform the
numerical integration. A fine mesh was used at thesmaller diameters with the meshing
becoming coarser as the diameter increased and the value of the integrand diminished (at the
tail of the particle/mass distribution, Figures 4 and 5). The use of non-uniform mesh spacing
is permitted for this form of numerical integration because of its relative simplicity
(i.e., simply summing the area of trapezoids formed by the integrand and each set of two
bounding mesh points [Atkinson 1989, p. 2531).
000-00C-MGRO-01700-000-000 48 of 84 October 2004

Because the respirable particle diameter is defined in terms of the median activity or the median
mass (Section 6.2.2.1), the probability distribution function is converted to a mass distribution
function [P,(d)] using the following equations:
m
P, (d) = m P(d) = ex.{- [ln(d)- ~(MGD)]' } JG ~n(o) 2 inZ (c)
0%. 28)
where
p is the theoretical density of UOz (10.96 g/cm3).
Figure 5 illustrates a typical normalized differential [mD] and integrated [mi] lognormal mass
distribution and shows the various diameters discussed in Section 6.2.2.2.
where again the extended trapezoidal rule (Press et al. 1992, p. 127) was used to numerically
integrate the probability integral from zero to some diameter, dl, where:
I= 1, 2,. . ., K-1, K. K is the total number of integration points.
With these relationships established, now only the standard deviation and the MGD of the
aerosol are needed to determine the RF of an aerosol distribution of commercial SNF and crud.
6.2.2.3.2 Characteristics of a Commercial SNF Aerosol
The two variables in the lognormal probability distribution function that are required to calculate
number and mass distributions for an aerosol are the standard deviation and the MGD.
In Section 6.2.2.2, relationships between MGD, MMD, and AMAD are shown in terms of the
standard deviation. Thus, if any two of these properties (i.e., MGD, MMD, AMAD, and
standard deviation) are known for a commercial SNF aerosol, then the lognormal probability
distribution function can be solved and the number and mass distributions calculated.
October 2004

~ Commercial SNF Accident Release Fractions
An alternative method of establishing a lognormal distribution is to determine any one of these
properties and a single point on any of the curves discussed in Section 6.2.2.3.1. Then an
iterative method can be applied where another variable is varied in a logical manner until the
known point is captured on the curve. For example, say the MGD of a distribution is known and
a point on the differential lognormal number distribution is known, then the standard deviation
can be calculated using the differential lognormal number equation (n') in the previous section.I6
I
The burst rupture data used to establish the release fractions for fuel fines in Section 6.2.1.3
(Lorenz et al. 1980, p. 105 and Appendix C) provide data that can be interpreted as the MMD
of the initially released particulate and also, provide a point on the integrated mass curve.
The burst rupture test data do not provide a standard deviation or the MGD for the initially
released particulate.
The impact test data on unclad UOz fuel pellets, discussed in Section 6.2.1, may be used to
establish an RF for ceramic UOz, as an MMD and standard deviation have been established for
these data. In fact, the RF for ceramic U02 is reported from these tests in ANL-81-27 (Mecham
et al. 1981, pp. 26, 34, and 35, and Table 2) and ANL-82-39 (Jardine et al. 1982). This RF is
based on the fraction of mass associated with the lognormally-distributed particles that have
diameters less than 10 pm. This is a conservative interpretation of the RF relative to the
interpretation of the RF established in Section 6.2.2.1, because the mass kaction of particles with
diameters less than 10 pm is greater than the fraction of particles with an AMAD of 10 pm.
In addition, the particulate smaller than 90 pm were separated by wet sieving which effectively
detaches smaller particles from larger fragments that would not ordinarily be considered
respirable, thereby ariificially increasing the respirable fraction. Furthermore, the quantity of
respirable particulate is based on the amount measured from the material that had settled on to
the bottom of the test apparatus. This quantity is likely inaccurate, because the actual respirable
particulate were likely still airborne and could only be measured after being collected onto a
filter, as was done for the burst rupture data. Nevertheless, the applicability of these impact data
towards drop accidents involving commercial SNF was discounted in Section 6.2.1 and hence,
the burst rupture data will be used to establish the RF for commercial SNF.
MMD of Commercial SNF Fuel Fines Aerosol
Fuel fines collected from burst rupture tests in NTJREGICR-0722 (Lorenz et al. 1980, p. 105 and
Appendix C) were measured with a scanning electron microscope and determined to be
"typically 150 pm" in the furnace tube near the rupture point on the fuel pin. This 150-pm
diameter is interpreted to be the MMD of initially aerosolized commercial SNF (Section 5,
Assumption 5.8). This value may be considered conservative when applied to drop or impact
events because larger particulates are likely to be initially aerosolized in these events because of
l6 This is a non-linear equation and hence, solving for the standard deviation will require either
applying a root finding scheme (e.g., conjugate gradient or bisection method) or simply
iterating on the standard deviation until the correct solution is obtained.
October 2004

Commercial SNF Accident Release Fractions
the brittle nature of the fuel (assuming reasonable drop heights). This is supported by the MMDs
measured in impact tests on unclad, depleted, ceramic UO2 pellets in ANL-81-27 (Mecham et al.
1981, pp. 26,34 and 35). These tests involved the impaction of two separate samples of depleted
UOz pellets. The resulting particle-size distributions had measured MMDs of 18 mm and 32 mm
(not pm). These MMDs are significantly larger than the 150-pm MMD assumed in this analysis
for the initially aerosolized commercial SNF because these MMDs are based on particles resting
on the bottom of the test apparatus while 150-pm MMD is based on airborne particles. These
MMDs would become smaller when only airborne particles are used to determine MMDs.
Because larger particles are essentially irrespirable (based on the rationale presented in Section
6.2.2.1) and cany a large portion of the total mass, larger MMDs will equate to smaller RFs, with
all other parameters being equal (e.g., the standard deviation). Thus, the selection of the 1 5 0 - p
diameter to represent the MMD of the initially aerosolized commercial SNF from a drop or
impact event is believed to be a conservative assumption.
Another set of test data performed on single pellets of U02 by Alvarez is cited in Sanders et al.
(1992, Section IV-4). Alvarez performed a series of tests on pellets of clad U02 that were both
depleted and irradiated. These tests involved the detonation of explosive charges near the fuel.
This would result in a significantly greater amount of energy imparted to the fuel than occurred
in the burst rupture tests or a drop or impact event considered in this analysis. This energy
imparted to the fuel would likely to create more small particles (given the brittle nature of the
fuel) relative to the number created by a burst rupture or an impact event that would likely have
some damage dampening to the upper sections of the fuel rod. Hence, the measured MMDs
fkom these tests, which ranged from approximately 30 to 100 pm, are not considered applicable
to this analysis. However, the relatively small difference between the upper range value of
100-pm MMD from Alvarez's explosive tests and the 150-pm diameter considered in this
%nalysis, reinforces the conservative arguments made in this analysis for selecting an MMD of
150 pm.
The only potential drawback of considering the 150-pm MMD of initially aerosolized
commercial SNF, in the analysis of the RF, is that this value is based on burst ruptures at high
temperatures and not on an impact rupture caused by a drop event. Thus, it is not necessarily an
accurate value for modeling the RFs for the considered accident events. However, the use of the
burst rupture data towards impact rupture is supported in SAND80-2124 (Wilmot 1981, pp. 32
and 33). The basic tenet is that although an impact rupture is expected to produce more particles
than were present in the spent fuel before a burst rupture, there will be less pressure to exhaust
these particles after an impact than a burst rupture (Wilmot 1981, pp. 32 and 33). In addition, it
is expected that an impact rupture would have a more restricted release pathway because of the
cladding deformation (Wilmot 1981, p. 33). These arguments provide additional justification for
applying the burst rupture test results to impact ruptures and hence, the 150-pm burst rupture
MMD to the initially aerosolized commercial SNF from a drop or impact event at the repository.
October 2004


Commercial SNF Accident Release Fractions
Additional Data Point for Commercial SNF Fuel Fines
The other data point used from the burst rupture test data in NUREGJCR-0722 (Lorenz et al.
1980, p. 105) involves the fraction of the total mass less than a specified diameter. According to
the masses summarized for each of the burst rupture tests in NUREGICR-0722 (Lorenz et al.
1980, Table 42 on p. 105), the largest fraction of UO2 found in the thermal gradient tube and the
filter pack was 0.0293, conservatively 0.03.'~ Considering the deposition of the released fuel
particles from gravity, fuel particles of diameters greater than 12 to 15 pm are considered to have
settled out before reaching the thermal gradient tube (Lorenz et al. 1980, p. 105), and Attachment
C of this analysis for separate confirmation of this settling calculation). Thus, the fraction of
U02 collected in the thermal gradient tube and filter packs is considered to have diameters less
than 12 to I5 pm. This has been confirmed by sampling (albeit somewhat randomly and
sparsely) some of the particulate collected in the filters with a scanning electron microscope and
determining that these particulate had diameters of typically 10 pm (Lorenz et al. 1980, p. 105
and Appendix C). Summarily, approximately 0.03 of the total mass of commercial SNF released
from the burst rupture tests has a diameter of less than 12 to 15 pm (Section 5, Assumption 5.9).
6.2.2.4 Calculated Respirable Fractions
Calculation Methods
In Section 6.2.2.1, the RF of an aerosol that contributes to the inhalation dose is defined as a
distribution or a portion of a distribution with an AMAD of 10 pm and less in this analysis.
In instances when the AMAD of a whole distribution is greater than 10 pm, the RF is interpreted
to be the mass fraction of particles below a cut-off diameter that has an AMAD of 10 pm.
To determine this cut-off diameter, an iterative scheme must be applied with the objective of
obtaining an MMD of 3.5 pm for commercial SNF particulate and 5.0 pm for crud particulate
(each MMD is approximately equal to an AMAD of 10 pm) for the mass distribution below the
cut-off diameter. The following steps, henceforth referred to as the iterative method, were used
to determine the cut-off diameter:
1. Select a cut-off diameter greater than 3.5 pm for U02 (or 5.0 pm for crud).
2. Normalize the mass distribution less than the selected cut-off diameter.
3. Determine the MMD of the normalized distribution (the geometric diameter associated
with 50 percent of the cumulative mass).
" This value is calculated from data for test HBU-10 in Lorenz et al. (1980, Table 42 on
p. 105). According to this table, a total of 46.03 mg of UO2 was released from the burst fuel
pin and 1.35 mg of this was measured in the thermal gradient tube and filter pack. Hence, the
fraction of U02 collected in the thermal gradient tube and filter pack is 1.35l46.03 or 0.0293.
000-00C-MGRO-01700-000-000 54 of 84 October 2004



Commercial SNF Accident Release Fractions
4. If the MMD of the normalized distribution is greater than 3.5 pm (5.0 pm for crud),
then select a lower cut-off diameter and repeat steps 2 and 3.
5. If the MMD of the normalized distribution is less than 3.5 pm (5.0 pm for crud), then
select a higher cut-off diameter and repeat steps 2 and 3.
6. When the MMD of the normalized distribution is approximately equal to 3.5 pm
(5.0 pm for crud), then the respirable fraction is equal to the cumulative mass fraction
of the entire aerosol distribution less than the cut-off diameter.
There are, however, several other methods of interpreting the RF that are not as arduous and
were considered in this analysis:
A conservative method, applied in ANL-81-27 (Mecham et al. 1981, pp. 4 and 5),
henceforth referred to as the AMAD-10 Method, is to assume the mass fraction of
aerosol particulate with geometric diameters less than 10 pm are respirable.
This respirable cut-off diameter is based on equating the maximum respirable AMAD of
10 pm to a geometric diameter (essentially neglecting the AMAD and particle density
differences). However, the AMAD of the fraction of the aerosol with a geometric
diameter less than 10 pm is likely to be greater than 10 p, thereby over-predicting the
respirable fraction.
A non-conservative method (relative to the previously described method), henceforth
referred to as the AED Method, is to assume the mass fraction of commercial SNF
aerosol particulates with geometric diameters less than 3.5 pm (5.0 p for crud
.. + aerosols) are respirable. This respirable cut-off diameter is based on the relationship
between the respirable cut-off particle diameter (&/J and the AED shown in
Section 6.2.2.1 (DOE 1994, p. 1-4). According to this relationship, a commercial SNF
aerosol with an AED of 10 p has a &I, of approximately 3.5 pm (5.0 pm for crud
aerosols). The AMAD of the commercial SNF aerosol mass fraction less than 3.5 pm
(5.0 p for crud aerosols) will be less than 10 p, thereby under-predicting the
respirable fraction.
Example Problem Imlementing RF Methods
Consider a lognormal commercial SNF mass fraction distribution (Section 6.2.2.3) with an MGD
of 1 pm and a standard deviation of 2 as shown in Figures 4 and 5. From Figure 5, the MMD of
this distribution is established as 4.2 pm (the geometric diameter associated with 50 percent of
the total mass of the distribution). Using the relationship between MMD and AMAD from
Section 6.2.2.2, the AMAD of this distribution is equal to 12 pm. According to the AMAD-10
method, 89.9 percent of this aerosol is respirable and this respirable fraction of the aerosol has an
AMAD of 1 1.1 pm. According to the AED method, 40.2 percent of this aerosol is respirable and
the AMAD of this respirable fraction of the aerosol is 6.5 pm. The actual fraction of the aerosol
that has an AMAD of 10 pm (the iterative method), and would be considered respirable in this
analysis, is 80.2 percent. Table 5 summarizes the results from these different methods and notes
the cut-off diameter (the assumed maximum respirable particle diameter) in the final column.
000-00C-MGRO-01700-000-000 55 of 84 October 2004


Commercial SNF Accident Release Fractions
This example shows the large variance in the RF because of the different interpretations of the
definition of respirable particles in Section 6.2.2.1. In this analysis, the iterative method is
judged to provide the most accurate results and is used to establish the RF of the commercial
SNF aerosol.
Table 5. Respirable Fractions (Expressed as Percent) of Lognormal Distribution with an MGD of 1 pm
and a Standard Deviation of 2 (For Example Purposes Only)
Method Respirable Percent AMAD- Cut-Off Diameter
AMAD-10 89.9% 11.1 pm 10 pm
AED 40.2% 6.5 ~m 3.5 pm
Iterative 80.2% 10 pm 7.5 prn
NOTES: AED = aerodynamic equivalent diameter; AMAD = activity median aerodynamic diameter.
Verification of RF Methods
To verify the application of lognormal distributions to determine respirable fractions for
commercial SNF aerosols, the experimental pulverization data collected in ANL-81-27 (Mecham
et al. 1981, pp. 23 to 35, UOz specimen 1) is simulated. The minimum amount of input data
needed to perform an accurate calculation of the respirable fraction is the MGD or MMD of the
aerosol distribution, the standard deviation or a point on one of the number or mass distribution
cwes, the particle density, and the dynamic shape factor. For commercial SNF, Section 6.2.2.2
establishes the particle density and dynamic shape factor as 10.96 &m3 and 1.3 (Section 5,
Assumption 5.5), respectively. ANL-81-27 (Mecham et al. 1981, Table 2) determined that the
MMD is 18 rnm and the standard deviation for U02 (Specimen 1) is 8.18 mm.
By fixing the standard deviation, the MGD was varied until the calculated MMD of the
distribution equaled 18 mm. The resulting MGD of this distribution was 0.032 pm.'8 Figure 6
compares the measured cumulative mass fraction with the cumulative mass fraction calculated
using the lognormal distribution. The fit is very good. The respirable fraction for this
distribution is 2.0 x with the inherent assumption that the airborne release fraction is 1 for
these data (Mecham et al. 1981, Table 2). This RF is based on the mass fraction of particulate
with diameters less than 10 pm (Mecham et al. 1981, pp. 4 and 5), which is equivalent to using
the AMAD-10 method. Using the AMAD-10 method and the calculated lognormal distribution,
an RF of 1.84 x is obtained. This value and the measured value are very close, thus
verifying the method.
The MGD could also be calculated from ln(MMD) = In(MGD) + 31n2(o). The MGD
calculated, using this method, is equal to 0.032 pm (identical to the presented value).
October 2004

Commercial SNF Accident Release Fractions
Applying the iterative method to data from ANL-81-27 (Mecham et al. 1981, Table 2) yields an
RF of approximately 4.83 x with a cut-off diameter (maximum respirable diameter) of 4.95
pm. The AED method determines an RF of 2.41 x lo-'.
6.2.2.4.1 Respirable Fraction for Commercial SNF
The MMD of an initially released commercial SNF aerosol of fines was conservatively
established to be 150 pm (Section 6.2.2.3.2), based on burst rupture data in NUREGJCR-0722
(Lorenz et al. 1980, p. 105 and Appendix C) (Section 5, Assumption 5.8). In addition, these
burst rupture data also provided a data point on the integrated mass curve: 0.03 of the total mass
of released commercial SNF has a maximum diameter between 12 and 15 pm (Section 5,
Assumption 5.9). These two characteristics of commercial SNF fuel fines released from a burst
ruptured fuel pin allow for the RF to be calculated.
Determination of MGD
First, solving the relationship between MGD, MMD, and the geometric standard deviation
provided in Section 6.2.2.2 for MGD, allows for the calculation of MGD given MMD and the
standard deviation:
ln(MMD) = ln(MGD) + 3 In2 (a) (Eq. 31)
MGD = ~X~{~~(MMD) - 3 inZ (a)} (Eq. 32)
However a standard deviation was not established for these data. Thus another iterative process
must be established such that the selection of the standard deviation results in an MGD
that correlates to an MMD of 150 pm and 0.03 of the total mass has a maximum diameter
between 12 and 15 pm. The process steps are as follows:
Select a standard deviation, o.
Calculate the MGD using Equation 32 with an MMD of 150 pm.
Calculate the integrated lognormal mass distribution using the MGD and o.
Determine whether 0.03 of the total mass has a maximum diameter of either 12 pm or
1.5 pm.
If the fraction of the total mass is less than 0.03 at 12 or 15 pm increase the standard
deviation and repeat steps 2,3, and 4.
If the fraction of the total mass is greater than 0.03 at 12 or 15 pm decrease the
standard deviation and repeat steps 2,3, and 4.
If the fraction of the total mass is approximately 0.03 at 12 or 15 pm, record the RF
from the various methods described earlier (e.g., the iterative method).
000-00C-MGRO-01700-000-000 58 of 84 October 2004

Commercial SNF Accident Release Fractions
Figure 7 shows how the fraction of mass that has a maximum diameter of 12 and 15 pm varies
with the standard deviation using the preceding iterative scheme; these results are produced for
an MMD fixed at 150 pm. In addition, Figure 7 also shows how the iterative method RF varies
with the standard deviation. Each standard deviation has only one RF associated with it from
each of the methods previously described because the MMD of this aerosol distribution is fixed
at 150 pm, which in turn fixes the value of the MGD to the standard deviation through the
Equation 32.
Determination of RF
Figure 7 also shows how the RF from the iterative method is established for this distribution with
an MMD fixed at 150 pm. First, the standard deviation that produces a total mass fraction
of 0.03 for particles with diameters less than 12 and 15 pm must be established. This is done in
Figure 7 with the horizontal solid line leading from the left abscissa value of 0.03 to the dashed
lines representing the mass fraction for 12 and 15 pm. At the intersection of the horizontal line
with the mass fraction lines a vertical line is drawn to the ordinate. The intersection between the
vertical line and the ordinate establishes the standard deviations.
For particles with a maximum diameter of 15 pm, the standard deviation and corresponding
MGD are 3.4 and 1.678 pm, respectively. For particles with a maximum diameter of 12 pm, the
standard deviation and corresponding MGD are 3.8 and 0.715 pm, respectively. The vertical
lines drawn to the ordinate also intersect the iterative method RF line which establishes the RFs
for the diameter limits of 12 and 15 pm (two horizontal lines drawn to the right abscissa).
The iterative method RF for particles with a maximum diameter of 15 pm is 2.17 x lo-' and for
, particles with a maximum diameter of 12 pm the RF is 4.90 x lo-'.
Table 6 summarizes the RFs for these two limiting diameters using the other RF methods.
This table also summarizes the cut-off diameters and AMADs for each of the methods. The RF
results from the iterative method, which is considered to provide the most concise RFs relative to
the other methods, indicate that for commercial SNF aerosols the RF should conservatively be
0.005. This result is the rounded-up value for the case with particles with a maximum diameter
of 12 pm (an MMD of 150 pm, an MGD of 0.715 pm, and a standard deviation of 3.8).
Comparison between Burst Rupture and Impact Rupture Distributions
Figure 8 compares the lognormal distributions produced for the burst rupture data (MMD = 150
pm, MGD = 0.715 pm and o = 3.8) from NUREGICR-0722 (Lorenz et al. 1980, p. 105 and
Appendix C) and the impact rupture data (MMD = 18 mm and o = 8.18) from ANL-81-27
(Mecham et al. 1981, Table 2). The resulting distributions show that, as expected, the attrition
process for commercial SNF escaping through a hole in the clad produced by a burst rupture has
fewer large particulates relative to the commercial SNF particulates produced by the impaction
of a tup onto unclad fuel pellets. Thus, the RF for the burst rupture data (0.005) is higher than
the RF for the impact rupture (0.0002). This is expected because the RF for the impact rupture
(0.0002) is based on the entire inventory of particles resting on the bottom of the test apparatus
while the RF for the burst rupture (0.005) is based on airborne particles only. These airborne
particles were collected on filters, baths, etc. located away from the point of release.
000-00C-MGRO-01700-000-000 59 of 84 October 2004

Commercial SNF Accident Release Fractions
Table 6. Respirable Fractions of Lognormal Distributions with MMDs of 150 prn
Respirable
Data Set Method Fraction AMAD Cut-Off Diameter
AMAD-10 0.0224 20.1 pm 10 pm
Particles with Diameters <
12 pm and make-up 3% of AED 0.00253 7.6 pm 3.5 pm
total mass a
Iterative 0.00490 10 um 4.7 um
AMAD-10 0.0139 21.3 pm 10 pm
Particles with Diameters < .
15 pm and make-up 3% of AED 0.001 11 8.0 pm 3.5 pm
total mass
Iterative 0.00217 10 pm 4.5 pm
a These particles have an MGD of 0.715 pm and a standard deviation of 3.8.
These particles have an MGD of 1.678 pm and a standard deviation of 3.4.
AED = aerodynamic equivalent diameter; AMAD = activity median aerodynamic diameter: MMDs =
mass median diameters.
Sensitivity of Results
The calculated RFs that led to the conservative selection of 0.005 as the RF for accident events at
the repository involving commercial SNF are sensitive to:
Density of the particulate (10.96 &m3) through the MMD and AMAD relationship
established in Section 6.2.2.2
Dynamic shape factor (1.3) through the MMD and AMAD relationship established in
Section 6.2.2.2
MGDMMD of the lognormal distribution (0.715 pm/150 pm) through the lognormal
distribution function insection 6.2.2.3.1
Standard deviation of the lognormal distribution (3.8) through the lognormal distribution
function in Section 6.2.2.3.1
Definition of the respirable AMAD (10 pm) through the definition of a respirable
particle as established in Section 6.2.2.1.
A decrease in the density or the MGDMMD will result in an increase of the RF. An increase in
the dynamic shape factor, the standard deviation, or the respirable AMAD will also result in an
increase of the RF for commercial SNF.
000-00C-MGRO-01700-000-000 62 of 84 October 2004

Commercial SNF Accident Release Fractions
6.2.2.4.2 Respirable Fraction for Crud
The analysis to determine the RF of crud is much simpler than the analysis performed for
commercial SNF fuel fines because experimental data presented in Sandoval et al. (1991, pp. 23
to 26) provide specific characteristics of the lognormal distribution of crud. Specifically, the
MGD, MMD, and standard deviation of a crud particle distribution were determined from
of a Quad Cities BWR fuel rod using ascanning electron microscope. A plot of the
collected data illustrated a precise lognormal distribution with an MMD of 9.7 vm, an MGD of 3
pm and a standard deviation of 1.87 (sandoval et al. 1991, p. 24 and p. 26).
These data resulted in a distribution of particle diameters that are somewhat larger than those
observed for crud scraped fiom fuel rods. However, the scraping action of these tests likely
produced smaller particles than are observed intact on a fuel rod. In addition, during an accident
hypothesized for the repository, the majority of intact crud particles are likely to spall off of fuel
rods rather than be pulverized, because the crud is not in a constricting volume. Hence, larger
and likely less respirable crud particles may be produced during an accident event than were
measured using the scanning electron microscope and scraped from the surface of a fuel rod.
Using the iterative method discussed earlier, a respirable fraction of 0.30 for BWR crud was
established. The cut-off diameter or maximum respirable particle diameter size was established
to be 6.9 pm. No database for PWR crud number distributions is presently available (Sandoval
et al. 1991, p. 25). However, there are some data on particle sizes determined for crud on the
cladding surface of a H.B. Robinson PWR fuel rod in Sandoval et al. (1991, p. 24). The PWR
crud has a comparable size range with BWR crud (Sandoval et al. 1991, p. 24). Thus, assuming
the PWR crud particles also exhibits lognormal behavior, then it may be assumed that PWR crud
will have the same RF to that of BWR crud (Section 5, Assumption 5.4).
Table 7 summarizes the results from the different methods to calculate the RF that were
discussed earlier. In addition, this table gives the cut-off diameter (or the maximum respirable
particle diameter) and AMAD for each method.
Table 7. Respirable Fractions for the Lognormal Distribution of Crud (MGD = 3.0 pm and 0 = 1.87)
October 2004
Method
AMAD-10
AED
Iterative
NOTES: AED = aerodynamic equivalent diameter; AMAD = activity median aerodynamic diameter; MGD = mean
geometric diameter.
Respirable Fraction
0.54
0.15
0.30
AMAD
13.0 pm
7.88 pm
10.0 pm
Cut-Off Diameter
10.0 pm
5.0 pm
6.9 prn

Commercial SNF Accident Release Fractions
The calculated RFs are sensitive to:
The density of the particulate (5.2 &n3) through the MMD and AMAD relationship
established in Section 6.2.2.2.
The dynamic shape factor (1.3) through the MMD and AMAD relationship established
in Section 6.2.2.2.
The MGDMMD of the lognormal distribution (3 pm19.7 pm) through the lognormal
distribution fimction in Section 6.2.2.3.1.
The standard deviation of the lognormal distribution (1.87) through the lognormal
distribution function in Section 6.2.2.3.1.
The definition of the respirable AMAD (10 p ) through the definition of a respirable
particle as established in Section 6.2.2.1.
A decrease in the density or the MGD/MMD will result in an increase of the RF. An increase in
the dynamic shape factor, the standard deviation, or the respirable AMAD will also result in an
increase of the RF for crud. Because of this sensitivity, a conservative value of 1.0 was selected
as the RF for accident events at the repository involving commercial SNF crud.
6.2.2.4.3 Summary of Respirable Fractions
To be consistent with the assumptions (i.e., AMAD = 1 pm) used to calculate the inhalation dose
conversion factors in Federal Guidance Report No. 11 (Eckerman et al. 1988, Table 2.1) and
Publication 30 (ICRP 1979, Figure 5.1), the RFs from the iterative method are recommended for
use in repository accident analysis involving commercial SNF. This method establishes the RF
based on a respirable particle size distribution with an AMAD of 10 pm or less. The
combination of dose conversion factors based on an AMAD of 1 pm and an RF based on an
AMAD of 10 p or less are considered to produce a conservative methodology for inhalation
dose calculations.
For commercial SNF fuel fines, an RF of 0.005 is recommended. This value is produced for
lognormally-distributed particulates with the following characteristics: an MMD of 150 pm, an
MGD of 0.715 p , and a standard deviation of 3.8. For crud from the cladding surface of
commercial SNF, a conservative RF of 1.0 is recommended. For gases and volatiles, an RF of
1.0 is recommended.
6.2.3 Comparison of Respirable Release Fractions
The respirable release fractions have been calculated for rail casks and truck casks under
hypothetical accident conditions in NUREGICR-6672 (Sprung et al. 2000, pp. 7-74 and 7-75).
These values are compared with the values recommended in this analysis in Table 8.
000-OOC-MGRO-01700-000-000 64 of 84 October 2004

Commercial SNF Accident Release Fractions
Table 8. Respirable Release Fractions for Commercial SNF
I Fuel Fines 1 1 . 5 ~ 10.' I 1.0 x 10.' I 4.0 x 10.' I 1.3 x 10.' I 1.1 x lo-' I
Nuclide
3~
Recommended
Respirable Release
F ~ ~ ~ ~ I ~ ~ ~ (ARF xx RF)
0.30
Crud (60~0)
Table 8 shows that the respirable release fractions, ARF multiplied by RF, for gases, Cs, Sr, Ru,
and fuel fines recommended in this analysis are more conservative than the corresponding values
given in NUREGICR-6672 (Sprung et al. 2000, pp. 7-74 and 7-75) for cases with impact speeds
ranging from 30 to 60 mph and interior cask temperatures ranging from the ambient temperature
to 350°C. The crud respirable release fraction of 1.5 x lo", recommended in Table 8, is
comparable to the values given in NUREGICR-6672 (Sprung et al. 2000, pp. 7-74 and 7-75).
Truck Casksa
Crud (55~e)
6.3 ARFS AND WS FOR FAILED COMMERCIAL SNF
PWR
Release
Fractions
1.4~10-'
Rail Casksa
1.5 x
This section examines and establishes the ARFs and the RFs for failed commercial SNF. This
section first examines the applicability of the ARFs and RFs established for intact commercial
SNF in Section 6.2 to failed commercial SNF. For cases where the intact commercial SNF
values do not bound those of the failed commercial SNF, a preliminary attempt has been made to
establish a bounding value for the failed commercial SNF.
BWR
Release
Fractions
5.4x105
PWR
Release
Fractions
1 . 8 ~ 1 0 "
NOTES: a NUREGICR-6672 (Sprung et al. 2000, p. 7-74 and 7-75) Case 2 for truck casks and Case 5 for
rail casks.
ARF = airborne release fraction; BWR = boiling water reactor; PWR = pressurized water reactor;
RF = respirable fraction.
1.5 x 10.'
According to the report 1999 Design Basis Waste Input Report for Commercial Spent Nuclear
Fuel (CRWMS M&O 1999, p. C-I), there are four categories of "problematic" commercial SNF
projected to be placed into disposable canisters (defined as canisters that can be placed into
disposal containers without being repackaged). These four categories include:
BWR
Release
Fractions
1.5x10-'
1.4 x
1. Mechanically and Cladding-Penetration Damaged SNF, which includes: (1) fuel that
is mech&ically damaged such that it can not be vertically lifted or fit within a standard
dimension andlor (2) fuel that has lost containment as a result of cladding damage.
1.4 x
000-00C-MGRO-01700-000-000 65 of 84 October 2004
4.5 10.'
4.5 10.'
1.8 x lo-3 1.3 x
1.8 1.3 x

Commercial SNF Accident Release Fractions
Note: It is assumed that the fuel in this category is made-up of intact fuel assemblies
and fuel pinslrods. The fuel pinslrods are stored in some grid-like structure within a
canister that provides the pinslrods structural support equivalent to an assembly
(Section 5, Assumption 5.15).
2. Consolidated/Reconstituted Assemblies, which includes fuel that was disassembled
and when reassembled has a form that is dimensionally different from the original.
Note: It is assumed that the fuel in this categorv is made-uu of intact fuel assemblies. - .
If the fuel assembly is a reassembled assembly, then its structural support is
considered equivalent to an un-reassembled assembly (Section 5, Assumption 5.15).
3. Fuel Rods, Pieces, and Debris, which includes variable-sized pieces of fuel and debris
combining fuel and nonfuel materials.
Note: It is assumed that the fuel in this category is in the form of stray fuel rods,
pieces, and debris that do not have any structural support system while in a canister
(Section 5, Assumption 5.15).
4. Nonfuel Components, which includes in-core assembly components physically
separated from the assemblies and shipped separately.
Results for the final category, Nonfuel Components, is considered out of scope for this report,
but it is expected that potential doses created from an accident event involving this material will
be bounded by those from the other categories. Thls is due principally to the lack of fuel
material included with the components of this category. The source term from this category is
likely be dominated by the surface crud source term.
The ARFs and RFs for specific isotopes identified in Section 6.2 for intact commercial SNF
involved in drop or impact events are examined to establish whether they bound the potential
ARFs for each of the three remaining categories of failed commercial SNF. The identified
groups of specific isotopes are gases ('H, and lz91), volatiles ('34~s, I3'cs, and 'O'RU), fuel
fines (''~r, etc.), and crud ( 6 0 ~ o and 55~e). Crud also includes a surface activity that will be
examined for each of these failed commercial SNF categories.
The failed fuel analyzed in this analysis is assumed always contained within an annular canister
with mesh screen caps at each end. Fuel handling at the repository will not involve removing
this fuel from these canisters. Fuel classified as failed but maintaining the form of an assembly
or group of fuel pins placed in some grid-like spacers provided within the canister (matrixed) are
considered to withstand a drop or impact event without any pulverization. This essentially
means that the structural support of the fuel assemblylfuel pins and the canister wall act
monolithically to mitigate any pulverization of the fuel.
October 2004

Commercial SNF Accident Release Fractions
6.3.1 ARFs and RFs for Gases
As fuel is irradiated in a nuclear power reactor, fission product atoms, of which approximately
15 percent are inert gases, are produced and buildup within the cladding of the fuel pins.
Release of these fission gases from the fuel matrix to the plenum and the gap region between the
fuel and the cladding is directly related to fuel pellet swelling which is a strong function of linear
power density. The primary fission gases released to this gap region for commercial SNF are
noble gases, iodines, and tritium. The recommended ARF values for these gases for intact
commercial SNF are based on p. 25.2 of Regulatory Guide 1.25, which states that the release
fractions of these fission gases from the gap region of a fuel rod are conservatively assumed to
consist of 30 percent %, 10 percent of other noble gases, and 10 percent of the radioactive
iodine. Page 25.2 of Regulatory Guide 1.25 further states that 30 percent of the '"I and '1
inventory may be assumed released for the purpose of sizing filters. These ARFs for gases are
also consistent with values cited in NUREG-1617 (NRC 2000a) and NUREGICR-6487
(Anderson et al. 1996) for use in potential accident releases involving intact fuel rods
(SAIC 1998, Table 4-1; Anderson et al. 1996, Table 6-2). These ARFs for gases also bound
ARFs cited in other literature: NUREG-1536 (NRC 1997, Table 7.1), NUREG-1567
(NRC 2000b, Table 9.2), and SANDSO-2124 (Wilmot 1981, Table XVIII). Table 3 summarizes
the ARFs for gases associated with each category of failed commercial SNF and the intact
commercial SNF. A discussion of how these values were established follows. RF for gases is
equal to 1 .O, as 100 percent are considered aerosolized.
6.3.1.1 Mechanically and Cladding-Penetration Damaged SNF
The release of fission gases from mechanically and cladding-penetration damaged SNF will only
'occur from the plenum and from the gap region between the cladding and the fuel of a fuel pin.
An impact or drop event may result in the cladding of this failed fuel category to be punctured,
penetrated, or cracked resulting in the release of the gap fission gases. Fission gases from the
fuel matrix, excluding the fraction considered released to the gap, are not expected to be released
as a result of an impact or drop event. The additional structural support provide to the
assemblies or matrixed fuel pins of this failed fuel by the canister will mitigate the potential
release of additional gases from the fuel matrix by preventing fuel pulverization potentially
caused by a drop or impact event.
Thus, for the mechanically and cladding-penetration damaged SNF category of failed fuel, the
ARFs for fission gases will be bounded by those established in Table 3. This is clearly true for
the cladding-penetrated commercial SNF where the majority of the fission gases originally in the
gap have been purged through the leak paths provided by the cladding penetrations prior to an
impact or drop event, thereby removing the motive force contributing to the expulsion of these
fission products from the irradiated fuel. These penetrations also ensure that no more fission
gases accumulate in the gap. The cladding of mechanically damaged commercial SNF prior to
an impact or drop event may be either intact or failed. For the portion of this fuel that has
previously failed cladding, the majority of the fission gases in the gap will have been purged
through the paths provided by the clad penetrations prior to an impact or drop event. These leak
paths will also ensure little or no fiuther fission gas accumulation in the gap. Thus, the gas ARFs
for this fuel are clearly bounded by those from Table 3. For the fraction of this fuel that has
intact cladding, the differences between it and the fuel analyzed in Section 6.2 are insignificant.
000-00C-MGRO-01700-000-000 67 of 84 October 2004

Commercial SNF Accident Release Fractions
Thus, the gas ARFs established in this analysis for intact fuel are considered bounding for the
mechanically damaged commercial SNF with intact cladding.
6.3.1.2 Consolidated/Reconstituted Assemblies
The release of fission gases from consolidated and reconstituted assemblies will only occur from
the gap between the cladding and the fuel of a fuel pin. An impact or drop event may result in
the cladding of this failed fuel category to be punctured, penetrated, or cracked resulting in the
release of the gap fission gases. Fission gases from the fuel matrix, excluding the fraction
considered released to the gap, are not expected to be released as a result of an impact or drop
event. The additional structural support provide to the assemblies (or matrixed fuel pins) of this
failed fuel by the canister will mitigate the potential release of additional gases from the fuel
matrix by preventing fuel pulverization potentially caused by a drop or impact event.
Furthermore, the differences between this category of failed fuel and the fuel analyzed in
Section 6.2 are insignificant with respect to the ARFs for fission gases. Thus, for the
consolidated/reconstituted assemblies category of commercial SNF failed fuel, the ARFs for the
fission gases will be bounded by those established in Table 3.
6.3.1.3 Fuel Rods, Pieces, and Debris
The release of fission gases from fuel rods, pieces, and debris will occur from the fuel matrix and
for a fuel rod with intact cladding, from the gap. An impact or drop event may result in the
cladding of a fuel rod in this failed fuel category to be punctured, penetrated, or cracked resulting
in the release of the gap fission gases. In addition, because this category of failed fuel is not
considered to have any assembly-like structure while in a canister, an impact or drop event is
assumed to pulverize up to 20 percent (Section 5, Assumption 5.17) of the fuel thereby releasing
fission gases from the fuel matrix.
Thus, for the fuel rods, pieces, and debris category of failed commercial SNF, the ARFs for the
fission gases are bounded by the values listed in Section 6.2 for intact commercial SNF. In this
report, the ARFs for the fission gases are conservatively established to be 0.3 for fuel rods with
intact cladding. Because fuel rods in this group may include rods with intact cladding that are
conservatively assumed to be pulverized in the event of a drop or impact, fission gases in the gap
(less than 0.005) (Lorenz et al. 1980, Tables 4 and 5) and the fraction of the fuel matrix
pulverized (0.2) are conservatively assumed released.
However, if the fuel rods in this failed fuel category are considered to have previously failed
cladding, then the fraction of fission gases released (i.e., the ARF) can be reduced to a value of
0.2. This ARF can also be applied to fuel pieces and debris, which are considered to be devoid
of any intact cladding and thus do not contain gap fission gases. This value is equal to the
fraction of fuel assumed pulverized (see Section 6.3.2.3). It does not include fission gases
released from a fuel rod prior to the rods placement into a canister or the fission gases that
remain in the particulate created by pulverization which equate to about 30 percent of the total
fission gas inventory. Page 25.2 of Regulatory Guide 1.25 conservatively estimates that
30 percent of the fission gases are released to the gap. Considering this fraction and also making
allowance for the fission gases that remain captured in the smaller particulate following
pulverization, the use of a fission gas ARF of 0.3 is conservative, as shown in Table 9.
000-00C-MGRO-01700-000-000 68 of 84 October 2004

Commercial SNF Accident Release Fractions
Table 9. ARFs and RFs for Gap Gases
ARF 1 RF
Fuel Category 'H 129,
1. MechanicalICladding Damaged 0.30 I 1 .O 0.30 11.0 0.30 I 1 .O
2. ConsolidatedIReconstituted 0.30 I 1 .O 0.30 11.0 0.30 1 1 .O
3a. Fuel Rods with Intact Cladding 0.30 I 1 .O 0.30 11.0 0.30 I 1 .O
3b. Other Fuel Rods, Pieces, and Debris 0.30 I 1 .O 0.30 I 1 .O 0.30 1 1 .O
Intact Commercial SNF 0.30 1 1 .O 0.30 1 1.0 0.30 1 1 .O
NOTES: ARF = airborne release fraction; RF = respirablefraction; SNF = spent nuclear fuel.
6.3.2 AFWs and RFs for Volatiles and Fuel Fines
Fuel fines and volatiles found in the gap are liberated or created from fuel pellets because of the
shaking of the rod and grinding action between fuel pellets that occurs during handling and
transport of the fuel. Fuel fines exist as residue from the fuel manufacturing process and are
produced during irradiation from pellet cracking that is associated with thermal distortion caused
while the fuel was at high temperatures. In the latter case, the higher temperature at the center of
a fuel pellet than at the periphery produces circumferential tensile stress that produce radial
pellet cracks.
Some of the primary constituents that makeup fuel fines and volatiles of commercial SNF, as
specifically noted in NUREG-1536 (NRC 1997, Table 7.1) and NUREG-1567 (NRC 2000b,
134 Table 9.2), are 90~r, I o 6 h , CS, and 1 3 7 ~ ~ . AS discussed in Section 6.2.1.2, the only volatiles
potentially present in a credible accident at the repository are Cs and Ru. Volatiles are assumed
to be 100 percent respirable.
The recommended ARF values for these nuclides, and fuel fines in Section 6.2 are based on data
collected from burst rupture tests in NUREGICR-0722 (Lorenz et al. 1980, Table 40). These
ARFs are also consistent with values cited in NUREGICR-6487 (Anderson et al. 1996, pp. 31
and 32) and NUREGICR-6410 (SAIC 1998, Table 4-1) for use in potential accident releases
from intact fuel rods. An extensive analysis of the RF in Section 6.2 for fuel fines released in the
burst rupture tests in NUREGICR-0722 (Lorenz et al. 1980, Table 40), established an RF
of 0.005.
Another set of experiments by ANL-81-27 (Mecham 1981) and ANL-82-39 (Jardine 1982) that
were not used to establish the ARFs for fuel fines in Section 6.2, provides alternative data that
may be applicable to failed fuel damaged by a drop or impact event that involves fuel
pulverization. The ARFs and RFs produced from these experiments involve unconfined (i.e., no
cladding) glass and U02 ceramic specimens impacted by a dropping weight. The applicability
and details of these tests are described in further detail in Section 6.2. In summary, these test
data and their associated PULF were deemed not appropriate for application to dropped or
impacted intact commercial SNF assemblies. These data are considered applicable, however,
because some of the failed fuel exists as small unclad fuel pieces and debris.
000-00C-MGRO-01700-000-000 69 of 84 October 2004



Commercial SNF Accident Release Fractions
Thus, the equation for the PULF fraction is rewritten here:
where
A is an empirical correlation equal to 2 x lo-" cm-~'/~
p is the U02 particle density (10.96 &m3)
g is gravitational acceleration (980 cm/s2)
h is the fall height (cm).
Table 10 summarizes the ARFs and RFs for volatiles (i.e., Cs and Ru) and fuel fines associated
with each category of failed commercial SNF and the intact commercial SNF.
Table 10. ARFs and RFs for Volatiles and Fuel Fines t
ARF 1 RF
Fuel Category
-
1. Mechanicallcladding Damaged 1 2.OE-04 I 1.0 I 3.OE-05 15.OE-03 I 3.OE-05 1 5.OE-03
2. Consolidated/Reconstituted I 2.OE-04 I 1.0 I 3.OE-05 15.OE-03 3.OE-05 15.OE-03
I 3a. Fuel Rods with Intact Claddina I 2.OE-04 11.0 1 5.9E-07 11.0 1 5.9E-07 11.0 I
I 3b. Other Fuel Rods. ~iebes. 8 Debris I 2.OE-04 11.0 1 5.8E-07 I 1.0 1 5.8E-07 I 1.0 I
I Intact Commercial SNF I 2.OE-04 11.0 I 3.OE-05 15.OE-03 I 3.OE-05 15.OE-03 I
NOTES: '~hese values assume a cask drop from 80 inches (203.2 cm).
ARF = airborne release fraction; RF = respirable fraction; SNF = commercial spent nuclear fuel
6.3.2.1 Mechanically and Cladding-Penetration Damaged SNF
The release of volatiles and fuel fines from mechanically and cladding-penetration damaged SNF
will occur from the fuel surface and gap between the cladding and the fuel of a fuel pin. An
impact or drop event may result in the cladding of this failed fuel category to be punctured,
penetrated, or cracked resulting in the release of the volatiles and fuel fines captured by the
exhausting gases in the gap. Volatiles and fuel fines in the fuel matrix are not expected to be
released as a result of an impact or drop event.
The additional structural support provide to the assemblies or matrixed fuel pins of this failed
fuel by the canister will mitigate the potential release of additional volatile and fuel fine
quantities kom the fuel matrix by preventing fuel pulverization potentially caused by a drop or
impact event (Section 5, Assumption 5.16).
October 2004

Commercial SNF Accident Release Fractions
Thus, for the mechanically and cladding-penetration damaged SNF category of failed fuel, the
ARFs for the volatiles and fuel fines on the surface of the fuel and in the gap will be bounded by
those established in Table 3. This is clearly true for the cladding-penetrated commercial SNF
where the majority of the fission gases originally in the gap have been purged through the leak
paths provided by the cladding penetrations prior to an impact or drop event, thereby removing
the motive force contributing to the expulsion of these fission products from the irradiated fuel.
For the mechanically damaged commercial SNF, the ARFs are bounded for the same reasons the
intact commercial SNF is bounded by these ARFs: high deposition through narrow, often long,
and tortuous paths to the breach point especially for larger particulate which carry a majority of
the total released fuel. Furthermore, the canister walls increase the amount of surface area where
local deposition of fuel fines/volatiles can take place, effectively increasing the tortuous paths
they pass prior to escaping to the environment. These characteristics provide sufficient
justification that the fuel fine and volatile ARFs for intact commercial SNF bound the same
ARFs for this category of failed commercial SNF.
! 6.3.2.2 Consolidated/Reeonstituted Assemblies
The release of volatiles and fuel fines f?om consolidated and reconstituted assemblies will only
occur from the fuel surface and gap between the cladding and the fuel of a fuel pin. An impact
or drop event may result in the cladding of this failed fuel category to be punctured, penetrated,
or cracked resulting in the release of the volatiles and fuel fines captured by the exhausting gases
in the gap. Volatiles and fuel fines in the fuel matrix are not expected to be released as a result
of an impact or drop event. The additional structural support provide to the assemblies (or
matrixed fuel pins) of this failed fuel by the canister will mitigate the potential release of
additional volatile and fuel fine quantities fiom the fuel matrix by preventing fuel pulverization
potentially caused by a drop or impact event (Section 5, Assumption 5.16).
Thus, for the consolidated~reconstituted assemblies category of commercial SNF failed fuel, the
ARFs for the volatiles and fuel fines in the gap will be bounded by those established in Table 3.
This is because of the same reasons the intact commercial SNF is bounded by these ARFs: high
deposition through narrow, often long, and tortuous paths to the breach point especially for larger
particulate which carry a majority of the total release mass (see Section 6.2). The canister walls
also increase the amount of surface area where local deposition of fuel fines/volatiles can take
place, effectively increasing the tortuous paths the fuel fines/volatiles pass prior to escaping to
the environment. These characteristics provide justification that the fuel fine and volatile ARFs
for intact commercial SNF bound the same ARFs for this category of failed commercial SNF.
6.3.2.3 Fuel Rods, Pieces, and Debris
The release of fuel fines and volatiles from fuel rods, pieces, and debris will occur from the fuel
matrix and for fuel rods with intact cladding, from the gap. An impact or drop event may result
in the cladding of a fuel rod in this failed fuel category to be punctured, penetrated, or cracked
resulting in the release of the volatiles and fuel fines captured by the exhausting gases in the gap.
In addition, because this category of failed fuel is not considered to have any assembly-like
structure while in a canister, an impact or drop event may pulverize some fraction of the fuel,
thereby releasing fuel fines and volatiles from the fuel matrix.
000-00C-MGRO-01700-000-000 71 of 84 October 2004



Commercial SNF Accident Release Fractions
1 Thus, for the fuel rods, pieces, and debris category of commercial SNF failed fuel, the product of
the ARF and the RF for the fuel fines are conservatively assumed equal to the P n F from
DOE-HDBK-3010-94 (DOE 1994, Section 4.3.3):
PULF = ARF x RF = (A) (p) (g) (h) (Eq. 34)
I
where
A is an empirical correlation equal to 2 x lo-'' cm-s21g
p is the U02 particle density (10.96 &n3)
g is gravitational acceleration (980 cm/s2)
h is the height the fuel falls (cm).
Because this equation was derived &om data involving small specimens (i.e., three pellets), there
may be a large degree of conservatism built into this equation when its use is extrapolated to
masses greater than a few pellets. This conservatism is due in a large part to a cushioning effect
created by the pulverization of the fuel at the bottom of a stack of fuel (e.g., pellets in a fuel rod).
The fuel at the bottom of the dropped stack is pulverized into a powder and then acts as a
cushion to the fuel dropping over it. In addition, this equation does not take into account the
structural credit provided by the presence of cladding or the canister.
In an attempt to correct for some of this conservatism, SAND84-2641 (MacDougall et al. 1987,
pp. 5-16 to 5-26) modified the PULF correlation by including an EPF. Unfortunately, values for
EPF are not available from analysis or experiment, however, SAND84-2641 (MacDougall et al.
1987, pp. 5-16 to 5-26) assumed that this factor is conservatively 0.2. This is equivalent to
assuming that approximately only 20 percent of the total fuel will be pulverized, most likely the
bottom 20 percent of the fuel.
The respirable &action of particulate is based on the mass of particles less than 10 pm in
geometric size. For uranium oxide particles with a density of 10.96 &m3, the respirable size is
about 3 pm (see Section 6.2.2.1). In a 1.2 Ucm3 drop-weight impact on a set of three U02
pellets, the &actions of particles smaller than 10 pm and 3 pm are 2 x and 2 x
(Mecham et al. 1981, Figures 16 and 17). Because the PULF equation is based on a respirable
particle size of 10 pm, the PULF value will be reduced by an additional factor of 0.1 (=2 x
10-'/2 x to correct for the fraction of respirable fuel fines.
Three rocks (1.29 kg, 1.1 7 kg, and 1.82 kg) (DOE 1994, p. 4-87) were dropped from a height of
3.7-m onto sand held in an open-lid steel quart can in a vented metal box placed on a plywood
sheet or onto unconfined sand placed on a plywood sheet. The rocks were dropped from the
same distance onto this can. Three experiments were performed. The first experiment involved
sand screened to be less than 2 mm (2000 pm) in diameter, while the second experiment used
sand screened to be less than 0.5 mm (500 pm) in diameter with 1.8 percent less than 0.025 mm
(25 pm). The third experiment was performed with the same powder used in the second
experiment, except that the sand is unconfined.
I 000-00C-MGRO-01700-000-000 72 of 84 October 2004



Commercial SNF Accident Release Fractions
The measured ARF and RF for the first experiment with confined sand were 3.OE-04 and 0.01
(DOE 1994, p. 4-87), respectively. The measured ARF and RF for the second experiment with
confined sand were 3.OE-04 and 0.07 (DOE 1994, p. 4-87), respectively. The ARF and RF for
the third experiment with unconfined sand were 8.7E-04 and 0.36 (DOE 1994, p. 4-87),
respectively. The experiments showed that the RF was significantly reduced by the presence of
an open-lid can. The ARF multiplied by the RF reduction factor (RED) because of the presence
of a can is defined as (ARF x RF)co,,fine~/(ARF x RF)unconfined:
The larger of the two calculated RED values (0.067) because of the presence of a can is chosen
to conservatively reduce the ARF x RF calculated from the PULF correlation. A failed fuel
canister is expected to provide the same degree or better confinement than an open-lid can.
Therefore, the application of an ARF x RF reduction factor of 0.067 to the failed fuel canister is
conservative.
One of the most important processes affecting the concentration of airborne particulate within a
plume is depletion by deposition of aerosols (particulates and vapors) onto the ground,
vegetation, and structures. The source term depletion fraction, DEP, is defined as Q(x)lQ(O),
where Q(x) and Q(0) are airborne particulate source strength at distance x and at the release
point, respectively. The source term depletion fraction at a distance of 5 km has been calculated
for five sample materials including UO2 and defense high level waste glass (MacDougall et al.
1987, Table 5-12). The calculated values of DEP range from 3.4 x lo-' to 5.8 x
A conservative DEP value of 1.0 instead of 0.05 was chosen for repository consequence analysis
(MacDougall et al. 1987, p. 5-49).
By including a source term depletion fraction of 1.0, a reduction factor (RED) of 0.067 because
of the presence of a confining surface (i.e., failed fuel canister), and an EPF of 0.2, the PULF
equation can be rewritten as:
Failed Fuel Release Fraction = DEP x RED x EPF x PULF
= (1.0) (0.067) (0.2) (A) (P) (g) (h)
= (1.0) (0.067) (0.2) (2 x lo-" cm-s21g) (10.96 g/cm3) (980 cm/s2) Q
= 2.88 x (h) (cm) (Eq. 37)
For a drop from 80 inches (203.2 cm), the normal handling height of a shipping cask at the
repository (CRWMS M&O 1998, p. 14), the corrected PULF is 5.8 x This value does not
include any releases of fuel fines from the gap.
For fuel rods with intact cladding, the additional fraction of respirable fuel fines located in the
gap between the cladding and the fuel must be included. Therefore, for these fuel rods, the
fraction of respirable fuel fines (ARF x RF) is 5.9 x This value includes the fraction
5.8 x lo-' for the fuel fines from the fuel matrix plus the fraction 1.0 x lo-' (=3.0 x
(ARF) x 5.0 x lo-' (RF) x 0.067) for the fuel fines from the gap.
000-00C-MGRO-01700-000-000 73 of 84 October 2004

Commercial SNF Accident Release Fractions
The preceding calculations show that the fuel fine ARF x RF for failed fuel in a canister is
bounded by the ARF x RF of 1.5 x lo-' for intact commercial SNF.
For the volatile nuclides released from failed fuel in the category of fuel rods, pieces, and debris,
the use of ARF/RF of 2.OE-0411.0 for intact commercial SNF is conservative. The ARFJRF
values of 2.OE-0411.0 are bounding for the release of Cs isotopes from failed fuel.
6.3.3 ARFs, RFs, and Surface Activities for Crud
Crud releases originate from the surface of a fuel rod and associated components (e.g., grid
spacers). In contrast to the fuel fines, gases, and volatiles released from a fuel rod, the crud
release fraction is not based on the fraction of fuel rods that are breached and the release
mechanism involves surface spallation rather than leakage past fuel cladding barriers. Crud is
primarily composed of iron-based compounds and some nickel, copper, cobalt, chromium,
manganese, zinc, and zircalloy. The actual amount varies from reactor to reactor and cycle
to cycle. Crud has become radioactive through neutron activation and has a relatively short
half life. In general, pressurized water reactor fuel is found to have less crud activity than
boiling water reactor fuel (Sandoval et al. 1991, p. 2).
In Section 6.2.1.3, the crud effective ARF on intact commercial SNF fuel assemblies is
conservatively assumed equal to 1.5E-02, which is recommended for use for failed commercial
SNF categories with cladding. Similarly, the RF for crud is conservatively established in Section
6.2.4.2 to be equal to 1.0, which is also recommended for use for failed commercial SNF
categories. In addition, the surface activities established in Section 6.2.1.3 as bounding for intact
commercial SNF are considered bounding for each failed commercial SNF category. According
to NUREGICR-6487 (Anderson et al. 1996, p. 28), these surface activities should be multiplied
by the surface area associated with a rod or pin to obtain the total crud activity. Thus, the surface
area associated with the assembly hardware (e.g., fuel rod spacers, grids, nozzles) is neglected.
Table 4 lists the surface activity densities (p~i/cm2) for the constituents of crud recommended to
be applied to each category of failed commercial SNF (and intact commercial SNF) by origin of
the fuel (i.e., BWR vs. PWR).
6.3.3.1 Mechanically and Cladding-Penetration Damaged SNF
A crud effective ARF of 1.5E-02 and an RF of 1.0 applied to this failed fuel category are
considered hounding, because this fuel is unlikely to have crud characteristics significantly
different from intact commercial SNF. In addition, mechanically deformed fuel may have lost
some crud from its surface because of the deformation and potentially have been in the reactor
core for a period insufficient to build-up a crud surface density equivalent to the value
established in Section 6.2 as bounding. Hence, the bounding crud surface activities
recommended in Table 4 are deemed sufficient for application to the failed commercial SNF in
this category. The surface areas applied to this crud surface activity for this failed fuel category
are those associated with each fuel rod within an assembly.
October 2004

Commercial SNF Accident Release Fractions
6.3.3.2 Consolidated/Reconstituted Assemblies
Application of a crud effective ARF of 1.5E-02 and an RF of 1.0 is clearly conservative for the
consolidated/reconstituted assemblies because a significant amount of crud on the surfaces of the
fuel pins was removed when the pins were pulled through the grid structures. Any remaining
crud on the pin surfaces is likely to be tightly adhering and unlikely to flake off during a drop or
impact event. In addition, removal of the crud also reduces the crud surface activity to a value
significantly less than the values cited in Table 4. Thus, the crud surface activity for this failed
fuel category is bounded by the value for intact commercial SNF. The surface areas applied to
this crud surface activity for this failed fuel category are those associated with each fuel rod
within an assembly.
6.3.3.3 Fuel Rods, Pieces, and Debris
For fuel rods, pieces, and debris, a crud effective ARF of 1.5E-02 and an RF of 1.0 are likely to
be conservative. Fuel rods will have been likely pulled through grid structures, which removes
the majority of loosely adhering crud leaving only tightly adhering crud, hence a crud effective
ARF of 1.5E-02 is certainly conservative. Similarly, pieces and debris are likely to have only a
small amount of crud or no crud depending on the origin of fuel. For example, fuel pellets that
have only been exposed to a spent fuel pool environment are likely to have no crud on their
surfaces. In addition, this lack of crud will also reduce the crud surface activity to a value
significantly less than the values cited in Table 4. Thus, the crud surface activity for this failed
fuel category will be bounded by the value for intact commercial SNF. The surface areas applied
to this crud surface activity for this failed fuel category are those associated with each fuel rod
-and no crud adherence to pellets is assumed.
6.3.4 Fuel Oxidation ARF
Failed commercial SNF will be shipped to the repository in screen-end or closed-end canisters in
a transportation cask. Failed commercial SNF canisters will be transferred from the cask to a
waste package in a waste transfer cell in the Fuel Handling Facility (FHF) or the dry transfer
facilities during normal operations. Exposing the failed fuel in air while in the FHF or dry
transfer facilities could cause oxidation of the fuel OJOz) to higher oxides, such as U~OX. When
U3O8 starts to form, the fuel pellet volume expands and eventually cladding could unzip. Test
data have shown that when cladding temperatures exceed 350°C (Einziger 1991, Figure 4), the
time to start forming U308 is short as compared to the time it takes to fully unzip the cladding.
In an accident involving failed fuel, releases would be due to the impact or drop and subsequent
oxidation of the exposed fuel. The release due to oxidation is bounded by the release due to the
impact or drop event. During the fuel oxidation period, only fission product gases and volatile
species are released. Release fractions of fission product gases from oxidation have been shown
to be less than 30 percent by Einziger (1991, Figure 8) and NLJREGICR-0722 (Lorenz et al. 1980
p. 34). Therefore the values presented on Table 9 bound the release fractions from oxidation. The
volatile species release fraction data on I3'cs and Io6~u taken from NUREGJCR-6672 (Sprung et
al. 2000, p. 7-46) is 1.4 x 10.~ for I3'cs, and 7.27 x 10.~ for '06~u. These release fractions are
bounded by the values presented on Table 10. No release of U308 was observed in existing fuel
rod oxidation tests (Lorenz et al. 1980; Einziger 1991).
000-00C-MGRO-01700-000-000 75 of 84 October 2004

Commercial SNF Accident Release Fractions
7. CONCLUSIONS
In this analysis, the total and respirable fractions of radioactive materials that are released from
an accident at the repository involving commercial SNF in a dry environment are specified and
documented. These total release fractions are defined as the fraction of total inventory of a given
radionuclide that is released to the environment from a waste form. The radionuclides are
released from the inside of breached fuel rods (or pins) and from the detachment of radioactive
material (crud) from the outside surfaces of fuel rods and other components of fuel assemblies.
The total release fraction accounts for several mechanisms that tend to retain, retard, or diminish
the amount of radionuclides that are available for transport to dose receptors or otherwise can be
shown to reduce exposure of receptors to radiological releases. The total and respirable release
fractions in this analysis are calculated from the following relationships:
Total Release Fraction (all pathways) = DR x LPF x ARF (Eq. 38)
Respirable Release Fraction (inhalation) = DR x LPF x ARF x RF (Eq. 39)
where
DR is the damage ratio
LPF the leak path factor
ARF the airborne release fraction
RF the respirable fraction.
The total and respirable release fractions established for commercial SNF in this analysis may be
applied to drop or impact accidents involving either a bare unconfined fuel assembly or a
confined fuel assembly contained in a shipping cask, a canister, or a waste package.
This analysis does not take credit, however, for the container that confines the fuel assemblies,
potentially providing an additional bamer for diminishing the total release fraction should the
fuel rod cladding breach during an accident. This implies that the DR and the LPF in the
preceding relationship for the total release fraction were assumed to be equal to one (Section 5,
Assumption 5.1). Thus, applying the total and respirable release fractions from this analysis to
confined commercial SNF assemblies may be considered conservative.
Table 11 summarizes the recommended respirable release kactions associated with intact or
failed commercial SNF, confined or unconfined. The total release fractions for commercial SNF
that may be applied in calculations of other doses (e.g., submersion) is essentially equal to the
ARFs in Table 11 because the LPF and DR are assumed equal to one in this analysis (Section 5,
Assumption 5.1). In addition to these parameters, Table 12 shows the recommended values for
the surface activity per unit area of crud as established in this analysis. The unit area of crud for
a fuel assembly is stated in NLJREGICR-6487 (Anderson et al. 1996, p. 28) to be equal to the
surface area associated with the rods or pins contained in the assembly. Thus, the surface area
associated with the assembly hardware (e.g., fuel rod spacers, nozzles, etc.) is neglected.
000-OOC-MGRO-01700-000-000 76 of 84 October 2004

Commercial SNF Accident Release Fractions
Respirable release fractions, ARF x RF, for gases, Cs, Sr, Ru, fuel fines, and crud given in
Table 11 are more conservative than corresponding values given in NUREGICR-6672 (Sprung
et al. 2000, p. 7-74 and 7-75) for cases with impact speeds ranging from 30 to 60 mph and cask
temperatures ranging from ambient to 750°C.
ARFs for the volatiles Cs and Ru, and the fuel fines (e.g., Sr) established by this analysis are
based primarily on data collected from several burst rupture tests in NUREGICR-0722 (Lorenz
et al. 1980, p. 101). ARFs for the noble gases iodine and tritium are based on p. 25.2 of
Regulatory Guide 1.25. These ARFs are consistent with the ARFs used to evaluate
transportation packages containing SNF as stated in NUREG-1617 (NRC 2000a, Table 4-1).
In addition, these values are also fairly consistent with the ARFs used to evaluate dry storage
cask systems (NRC 1997, Table 7.1).
Table 11. Respirable Release Fractions for Commercial SNF
I ARF I RF
Radionuclide
3~
8 5 ~ r
12g1
'% a
137 cs (v)
NOTES: ' These values assume a cask drop from 80 inches (203.2 cm) for fuel fine and particulate (p) release
fractions.
Category I :
Mechanical1
Cladding
Damaged
0.31 1.0
0.3 I 1.0
0.31 1.0
1% and
137~s (p)
%r
106 Ru
Fuel Fines
Crud
O The crud ARFIRF values are bounding values. RF is conservatively assumed to be 1.0. The values
presented here are the crud effective ARFs. The crud effective ARF consists of the ~roduct of the CSF
2.OE-04 1 1.0
with the ARF as described in Section 6.2.1.3
Category 2:
Reconstituted
0.311.0
0.311.0
0.31 1.0
0
3.OE-0515.OE-03
2.OE-0411.0
3.OE-0515.OE-03
1.5E-02 1 1.0
ARF = airborne release fraction; CSF = crud spallation fraction: SNF = spent nuclear fuel; RF =
respirable fraction.
2.OE-04 / 1.0
October 2004
Category 3a:
Fuel Rods with
Intact Cladding
0.31 1.0
0.31 1.0
0.3 11.0
0
3.OE-0515.OE-03
2.OE-0411.0
3.OE-0515.OE-03
1.5E-02 I 1.0
2.OE-04 I 1.0
Category 3b:
Other Fuel
Rods, Pieces,
and Debris
0.31 1.0
0.31 1.0
0.3 1 1.0
0
5.9E-07 1 1.0 '
2.OE-04 1 1.0 '
5.9E-07 1 1.0 a
1 SE-02 1 1.0
Intact
Commercial
SNF
0.31 1.0
0.3 1 1.0
0.3 11.0
2.OE-04 1 1.0 2.OE-04 I 1.0
0
5.8E-07 1 1.0 a
2.OE-04 I 1.0 a
5.8E-07 I 1.0 '
1.5E-02 1 1.0
0
3.OE-0515.OE-03
2.OE-0411.0
3.OE-0515.OE-03
1.5E-02 I 1.0

Commercial SNF Accident Release Fractions
7
Table 12. Crud Surface Activities (p~ilcm2)
Crud Surface Activities
Isotope Reactor Type (p~ilcrn~"
PWR 72.5
6 0 ~ ~ Crud
BWR 649.7
I I PWR I 1656 I
"Fe Crud I
BWR 2083
NOTES: a These values are assumed for a single bare unconfined fuel assembly or multiple assemblies.
These values have been corrected for half life of %o over 5 years and for fuel assembly average
activity (only in the case of multiple commercial SNF assemblies).
These values have been reduced using the half life of 5 5 ~ e over 5 years.
BWR = boiling water reactor; PWR = pressurized water reactor
A rigorous evaluation of the RF has also been carried out in this analysis to supplement the
sparse existing data for the fraction of commercial SNF and crud that may he considered
respirable and contribute to the inhalation doses. This evaluation used measured data presented
in NUREGICR-0722 (Lorenz et al. 1980, p. 101) and Sandoval et al. (1991, pp. 23 to 26) to
determine the RFs for commercial SNF fuel fines and for crud found on the fuel surface and
other components of fuel assemblies. Alternative methods to determine the RF were also
examined and results from these methods are presented in Tables 6 and 7 (Section 6).
Key assumptions to attaining the RF for commercial SNF are that the initially aerosolized
commercial SNF fuel fines released from the burst rupture tests had a mass median diameter of
150 pm (Section 5, Assumption 5.8) and that 3 percent of the total mass of released commercial
SNF had a maximum diameter somewhere between 12 and 15 pm (Section 5, Assumption 5.9).
The combination of dose conversion factors based on an AMAD of 1 pm and an RF based on the
mass fraction of particulate that have an AMAD of less than 10 pm are considered to produce a
conservative methodology for inhalation dose calculations.
With these recommended parameters and given the potential material at risk per fuel assembly
and site meteorological conditions, doses to workers and members of the public can be
calculated for the drop and impact accidents involving commercial SNF that may occur in the
repository surface facilities.
8. REFERENCES
8.1 DOCUMENTS CITED
Anderson, B.L.; Carlson, R.W.; and Fischer, L.E. 1996. Containment Analysis for Type B
Packages Used to Transport Various Contents. NUREGICR-6487.
Washington, D.C.: US. Nuclear Regulatory Commission. TIC: 237945.
Atkinson, K.E. 1989. An Introduction to Numerical Analysis. 2nd edition. New York,
New York: John Wiley & Sons, Inc. TIC: 248684.
000-00C-MGRO-01700-000-000 78 of 84 October 2004

Commercial SNF Accident Release Fractions
B&W Fuel Company 1991. Final Design Package Babcock & Wilcox BR-100 I00 Ton
Rail/Barge Spent Fuel Shipping Cask. Volume 2. 5 1 - 1203400-01. DBABE0000-00272- 1000-
00014 REV 00. Lynchburg, Virginia: B&W Fuel Company. ACC: MOV.19960802.0083.
Baker, D.A.; Bailey, W.J., Beyer, C.E.; Bold, F.C.; and Tawil, J.J. 1988. Assessment of the Use
of Extended Burnup Fuel in Light Water Power.Reactors. NUREGICR-5009.
Washington, D.C.: U.S. Nuclear Regulatory Commission. TIC: 234496.
BSC 2004a. Categorization of Event Sequences for License Application. 000-00C-MGRO-00800-
000-00A. Las Vegas, Nevada: Bechtel SAIC Company. ACC: ENG.20040722.0001.
BSC 2004b. Q-List. 000-30R-MGRO-00500-000-000 REV 00. Las Vegas, Nevada: Bechtel
SAIC Company. ACC: ENG.20040721.0007.
BSC 2004c. PWR Source Term Generation and Evaluntion. 000-00C-MGRO-00100-000-00B.
Las Vegas, Nevada: Bechtel SAIC Company. ACC: ENG.20040524.0007.
Chun, R.; Witte, M.; and Schwartz, M. 1987. Dynamic Impact Effects on Spent Fuel Assemblies.
UCID-21246. Livermore, California: Lawrence Livermore National Laboratory.
ACC: HQX.19881020.0031.
Crane Company 1988. Flow of Fluids Through Valves, Fittings, and Pipe. Technical Paper
No. 410. Joliet, Illinois: Crane Company. TIC: 237812.
CRWMS M&O (Civilian Radioactive Waste Management System, Management & Operating
Contractor) 1998. Drop Accident - ISF TSAR Design Basis Event Analysis.
CC0000000-01717-0200-00011 REV 01. Las Vegas, Nevada: CRWMS M&O.
ACC: MOV.19981029.0008.
CRWMS M&O 1999. 1999 Design Basis Waste Input Report for Commercial Spent Nuclear
Fuel. B00000000-01717-5700-00041 REV 00. Washington, D.C.: CRWMS M&O.
ACC: MOV.19991006.0003.
Dennis, R., ed. 1976. Handbook on Aerosols. TID-26608. Oak Ridge, Tennessee: Energy
Research and Development Administration, Technical Information Center. TIC: 242478.
DOE (U.S. Department of Energy) 1994. Analysis of Experimental Data. Volume 1 ofAirborne
Release Fractions/Rates and Respirable Fractions for Nonreactor Nuclear Facilities. DOEHDBK-
3010-94. Washington, D.C.: U.S. Department of Energy. TIC: 233366.
DOE 2004. Quality Assurance Requirements and Description. DOEIRW-0333P, Rev. 16.
Washington, D.C.: U.S. Department of Energy, Office of Civilian Radioactive Waste
Management. ACC: DOC.20040907.0002.
Eckerman, K.F.; Wolbarst, A.B.; and Richardson, A.C.B. 1988. Limiting Values of Radionuclide
Intake and Air Concentration and Dose Conversion Factors for Inhalation, Submersion, and
Ingestion. EPA 52011-88-020. Federal Guidance Report No. 11. Washington, D.C.:
U.S. Environmental Protection Agency. ACC: MOL.20010726.0072.
October 2004

Commercial SNF Accident Release Fractions
Einziger, R.E. 1991. "Effects of an Oxidizing Atmosphere in a Spent Fuel Packaging Facility."
Proceedings of the Topical Meeting on Nuclear Waste Packaging, FOCUS '91, September 29-
October 2, 1991, Las Vegas, Nevada. Pages 88-99. La Grange Park, Illinois: American Nuclear
Society. TIC: 231 173.
Graves, H.W., Jr. 1979. Nuclear Fuel Management. New York, New York: John Wiley &
Sons. TIC: 242479.
Holman, J.P. 1990. Heat Transfer. 7th edition. New York, New York: McGraw-Hill.
TIC: 242477.
Holtec International 1995. 10 CFR 72 Topical Safety Analysis Report for the Holtec
International Storage, Transport and Repository Cask System (HI-STAR 100 Cask System).
HI-941 184, Rev. 3. [Marlton, New Jersey]: Holtec International. TIC: 249856.
ICRP (International Commission on Radiological Protection) 1979. Limits for Intakes of
Radionuclides by Workers. Volume 2, No. 314 of Annals of the ICRP. Sowby, F.D., ed. ICRP
Publication 30 Part 1. New York, New York: Pergamon Press. TIC: 4939.
INEL (Idaho National Engineering Laboratory) 1996. Safety Analysis Report for the INEL
TMI-2 Independent Spent Fuel Storage Installation, Revision 0. Docket 72-20. Idaho Falls,
Idaho: U.S. Department of Energy, Idaho Operations Office. TIC: 233637.
INEL 1998. Fort St. Vrain Independent Spent Fuel Storage Installation Safety Analysis Report,
Revision 2. Docket No. 72-09. Idaho Falls, Idaho: U.S. Department of Energy, Idaho
Operations Office. ACC: MOL.20010721.0047.
Jardine, L.J.; Mecharn, W.J.; Reedy, G.T.; and Steindler, M.J. 1982. Final Report of
Experimental Laboratory-Scale Brittle Fracture Studies of Glasses and Ceramics. ANL-82-39.
Argonne, Illinois: Argonne National Laboratory. TIC: 225736.
Jones, R.H. 1992. Spent Fuel Corrosion Product and Fuel Cleaning Assessment. Los Gatos,
California: Robert H. Jones, P.E., Consultant. ACC: HQX.19920825.0007.
Larnarsh, J.R. 1983. Introduction to Nuclear Engineering. 2nd Edition. Menlo Park, California:
Addison-Wesley. TIC: 244841.
Levy, I.S.; Chin, B.A.; Simonen, E.P.; Beyer, C.E.; Gilbert, E.R.; Johnson, A.B., Jr. 1987.
Recommended Temperature Limits for Dry Storage of Spent Light Water Reactor Zircaloy-Clad
Fuel Rods in Inert Gas. PNL-6189. Richland, Washington: Pacific Northwest Laboratory.
TIC: 231836.
Lorenz, R.A.; Collins, J.L.; and Malinauskas, A.P. 1979. "Fission Product Source Terms for the
Light Water Reactor Loss-of-Coolant Accident." Nuclear Technology, 46,404-410. [La Grange
Park, Illinois: American Nuclear Society]. TIC: 242480.
October 2004

Commercial SNF Accident Release Fractions
Lorenz, R.A.; Collins, J.L.; Malinauskas, A.P.; Kirkland, O.L.; and Towns, R.L. 1980. Fission
Product Release From Highly Irfadiated L WR Fuel. NUREGICR-0722. Washington, D.C.:
US. Nuclear Regulatory Commission. TIC: 21 1434.
Luna, R.E.; Neuhauser, K.S.; and Vigil, M.G. 1999. Projected Source Terms for Potential
Sabotage Events Related to Spent Fuel Shipments. SAND99-0963. Albuquerque, New Mexico:
Sandia National Laboratories. ACC: MOL.19990609.OI60.
MacDougall, H.R.; Scully, L.W.; and Tillerson, J.R., eds. 1987. Nevada Nuclear Waste Storage
Investigations Project, Site Characterization Plan Conceptual Design Report. SAND84-2641.
Volume 4, Appendices F-0. Albuquerque, New Mexico: Sandia National Laboratories.
ACC: NN1.19880902.0017.
Mecham, W.J.; Jardine, L.J.; Pelto, R.H.; Reedy, G.T.; and Steindler, M.J. 1981. Interim Report
of Brittle-Fracture Impact Studies: Development of Methodology. ANL-81-27. Argonne,
Illinois: Argonne National Laboratory, Chemical Engineering Division.
ACC: NNA.19890411.0034.
Mishima, J. and Olson, K.M. 1990. "Estimate of the Source Term for a Repository Surface
Facility kom the Routine Processing of Spent Fuel." High Level Radioactive Waste
Management, Proceedings of the International Topical Meeting, Las Vegas, Nevada, April 8-12,
1990. 2, 1132-1 137. La Grange Park, Illinois: American Nuclear Society. TIC: 202058.
NAC (Nuclear Assurance Corporation) 1995. Safety Analysis Report for the NACLegal Weight
Truck Cask. Revision 13. Docket Number 9225. T-88004. Norcross, Georgia: NAC
International. TIC: 2449.
NRC (U.S. Nuclear Regulatory Commission) 1975. Reactor Safety Study: An Assessment of
Accident Risks in US. Commercial Nuclear Power Plants. WASH-1400.
Washington, D.C.: US. Nuclear Regulatory Commission. TIC: 236923.
NRC 1997. Standard Review Plan for Dvy Cask Storage Systems. NUREG-1536. Washington,
D.C.: U.S. Nuclear Regulatory Commission. ACC: MOL.20010724.0307.
NRC 2000a. Standard Review Plan for Transportation Packages for Spent Nuclear Fuel.
NUREG-1617. Washington, D.C.: U.S. Nuclear Regulatory Commission. TIC: 249470.
NRC 2000b. Standard Review Plan for Spent Fuel Dry Storage Facilities. NUREG-1 567.
Washington, D.C.: U.S. Nuclear Regulatory Commission. TIC: 247929.
NRC 2003. "Interim Staff Guidance - 5, Revision 1. Confinement Evaluation." ISG-5, Rev 1.
Washington, D.C.: U.S. Nuclear Regulatory Commission. Accessed January 24,2003. ACC:
MOL.20030124.0247. http://www.ncr.gov/reading-rm/doc-collectionslis~spent-~el.h~l
Parrington, J.R.; b o x , H.D.; Breneman, S.L.; Baum, E.M.; and Feiner, F. 1996. Nztrlides and
Isoropes, Char? of the Nuclides. 15th Edition. San Josc, California: General Electric Company
and KAPL, Inc. TIC: 233705.
October 2004

Commercial SNF Accident Release Fractions
PGE (Portland General Electric) n.d. Trojan Independent Spent Fuel Storage Installation, Safety
Analysis Report. PGE-1069. Portland, Oregon: Portland General Electric. TIC: 243815.
Press, W.H.; Teukolsky, S.A.; Vetterling, W.T.; and Flannery, B.P. 1992. Numerical Recipes in
Fortran 77, The Art of Scientzfic Computing. Volume 1 of Fortran Numerical Recipes. 2nd
Edition. Cambridge, United Kingdom: Cambridge University Press. TIC: 243606.
SAIC (Science Applications International Corporation) 1998. Nuclear Fuel Cycle Facility
Accident Analysis Handbook. NUREGJCR-6410. Washington, D.C.: US. Nuclear Regulatory
Commission. ACC: MOL.20010726.0069.
Sanders, T.L.; Seager, K.D.; Rashid, Y.R.; Barrett, P.R.; Malinauskas, A.P.; Einziger, R.E.;
Jordan,,H.; Duffey, T.A.; Sutherland, S.H.; and Reardon, P.C. 1992. A Method for Determining
the Spent-Fuel Contribution to Transport Cask Containment Requirements. SAND90-2406.
Albuquerque, New Mexico: Sandia National Laboratories. ACC: MOV.19960802.0116.
Sandoval, R.P.; Einziger, R.E.; Jordan, H.; Malinauskas, A.P.; and Mings, W.J. 1991. Estimate
of CRUD Contribution to Shipping Cask Containment Requirements. SAND88-1358.
Albuquerque, New Mexico: Sandia National Laboratories. ACC: MOV.19960802.0114.
Shetler, J.R. 1993. "Docket No. 72-1 1, Rancho Seco Independent Spent Fuel Storage
Installation, Revision 1 to the Rancho Seco Independent Spent Fuel Storage Installation License
Application and Safety Analysis Report." Letter kom J.R. Shetler (SMUD) to R.E. Cunningham
(NRC), October 27, 1993, DAGM/NUC 91-135, with attachment. TIC: 226550.
SNC (Sierra Nuclear Corporation) 1996. Safety Analysis Report for the TranStorTM Storage
Cask System. SNC-96-72SAR, Revision A. Scotts Valley, California: Sierra Nuclear
Corporation. TIC: 248383.
SNC 1997. Safety Analysis Report for the TranStorTM Shipping Cask System. SNC-95-71SAR,
Rev. 2. Scotts Valley, California: Sierra Nuclear Corporation. TIC: 243170.
Sprung, J.L.; Arnrnerman, D.J.; Breivik, N.L.; Dukart. R.J.; Kanipe, F.L.; Koski, J.A.; Mills,
G.S.; Neuhauser, K.S.; Radloff, H.D.; Weiner, R.F.; and Yoshimura, H.R. 2000. Reexamination
of Spent Fuel Shipment Risk Estimates. NUREGJCR-6672. Two volumes. Washington, D.C.:
U.S. Nuclear Regulatory Commission. ACC: MOL.20001010.0217.
Transnuclear 1989. lN-24 Dry Storage Cask Topical Report. E-7107. Hawthorne, New York:
Transnuclear. TIC: 23265 1.
Van Wylen, G.J. and Sonntag, R.E. 1986. Fundamentals of Classical Thermodynamics.
New York, New York: John Wiley & Sons. TIC: 245655.
Vectra Technologies 1995. Safety Analysis Report for the Standardized NUHOMSB Horizontal
Modular Storage System for Irradiated Nuclear Fuel. NLJH-0003, Rev. 3A. Volume 1.
Docket 72-1004. San Jose, California: Vectra Technologies. TIC: 104635.
October 2004

Commercial SNF Accident Release Fractions
Vectra Technologies 1996. Safety Analysis Report for the NUHOMS(BMP187 Multi-Purpose
Cask. NUH-05-151, Rev. 2. Two volumes. Docket 71-9255. San Jose, California: Vectra
Technologies. TIC: 233483.
I
Virginia Electric and Power Company 1995. North Anna Power Station Independent Spent Fuel
Storage Installation, License Application. Docket No. 72-16. Richmond, Virginia: Virginia
Electric and Power Company. TIC: 104549.
Weast, R.C. ed. 1972. CRC Handbook of Chemistry and Physics. 53rd Edition.
Cleveland, Ohio: Chemical Rubber Company. TIC: 219220.
Westinghouse 1996. Safety Analysis Report Large On-Site Transfer and On-Site Storage
Segment, CLIN 0004PC. MPC-CD-02-016, Rev. 1. Monroeville, Pennsylvania: Westinghouse
Government and Environmental Services Company. ACC: MOV.19961028.0056.
Wilmot, E.L. 1981. Transportation Accident Scenarios for Commercial Spent Fuel.
SAND80-2124. Albuquerque, New Mexico: Sandia National Laboratories.
ACC: HQ0.19871023.0215.
I 8.2 CODES, REGULATIONS, AND STANDARDS I
10 CFR 71.2004. Energy: Packaging and Transportation of Radioactive Material. Readily
available.
10 CFR 72.2004. Energy: Licensing Requirements for the Independent Storage of Spent Nuclear
Fuel and High-Level Radioactive Waste, and Reactor-Related Greater than Class C Waste.
Readily available.
10 CFR 961.2004. Energy: Standard Contract for Disposal of Spent Nuclear Fuel andlor High-
Level Radioactive Waste. Readily available.
ANSI N 13.1 - 1969. Guide to Sampling Airborne Radioactive Materials in Nuclear Facilities.
New York, New York: American National Standards Institute. TIC: 204994.
ANSIIANS-5.10-1998. Airborne Release Fractions at Non-Reactor Nuclear Facilities.
La Grange Park, Illinois: American Nuclear Society. TIC: 235073.
Regulatory Guide 1.25 Rev. 0. 1972. Assumptions Used for Evaluating the Potential
Radiological Consequences of a Fuel Handling Accident in the Fuel Handling and Storage
Facility for Boiling and Pressurized Water Reactors. Washington, D.C.: U.S. Nuclear
Regulatory Commission. Readily available.
I 8.3 PROCEDURES
I AP-3.12Q, Rev. 2, ICN 2. Design Calculations andAnalyses. ACC: DOC.20040318.0002.
1 AP-3.15Q, Rev. 4, ICN 5. Managing Technical Product Inputs. ACC: DOC.20040812.0004.
October 2004

Commercial SNF Accident Release Fractions
AP-SV.l Q, Rev. 1, ICN 1. Control of the Electronic Management of Information.
ACC: DOC.20040308.0001.
LP-SI.1 IQ-BSC, Rev. 0, ICN 1. Software Management. ACC: DOC.20041005.0008.
000-00C-MGRO-01700-000-000 84 of 84 October 2004

Commercial SNF Accident Release Fractions
ATTACHMENT A
Microsoft Excel 97 Spreadsheets for the Calculation of the Respirable Fraction of
Commercial SNF
The following Excel Spreadsheet consists of three separate worksheets: the controller worksheet,
which basically controls the inputs, the particle distribution calculations worksheet, which
performs the respirable fraction calculations, and the results worksheet.
The controller worksheet is the worksheet where the user inputs specific parameters that are used
in the calculation of the respirable fraction. Under the column with the heading "Use User Input
(Y/N)," the user may select to either use the default values in the worksheet by placing an "n" in
this column or input hisher own inputs for a specific parameter by placing a "y" in this column.
The user-supplied input would be placed under the column labeled "User Input" and will be used
in the calculation provided a "y" has been placed under the previous column. The default values
in the final column represent the values that produce the respirable fraction for commercial SNF
in this analysis. By allowing the users to supply their own input parameters, this spreadsheet can
be generically used to determine respirable fractions for other fuel types or materials.
Attachment B illustrates how user supplied inputs into this controller worksheet can be used to
establish the respirable fraction for crud using this same spreadsheet.
The following is a brief discussion of the inputs into the controller worksheet that are needed to
calculate the respirable fraction for commercial SNF and for crud:
1. Input: Multiplier to Increment
Default Value: 1.03 (good for commercial SNF particle distributions)
Description: This parameter is equal to a constant that is multiplied by the particle diameter
to create the non-uniform mesh as discussed in Section 6.2.2.3.1.
2. Input Mean
Default Value: 0.715 pm (good for commercial SNF particle distributions)
Description: This parameter is equal to the MGD of the whole particle distribution in the unit
of pm as discussed in Section 6.2.2.4.1.
3. Input: Standard Deviation
Default Value: 3.8 (good for commercial SNF particle distributions)
Description: This parameter is equal to the standard deviation of the whole particle
distribution, o as discussed in Section 6.2.2.4.1.
4. Input: Density
Default Value: 10.96 g/cm3 (good for commercial SNF particle distributions)
Description: This parameter is equal to the density of the particulate in the aerosol in the
units of g/cm3 as discussed for Assumption 5.6 (Section 5).
October 2004

Commercial SNF Accident Release Fractions
5. Input: Dynamic Shape Factor
Default Value: 1.3 (good for crud & commercial SNF particle distributions)
Description: This parameter is equal to the dynamic shape factor of the particulate in the
aerosol as discussed for Assumption 5.5 (Section 5).
6. Input: Respirable Fraction Cut-Off MMD
Default Value: 3.5 pm (good for commercial SNF particle distributions)
Description: This parameter is equivalent to the maximum MMD that provides for an AMAD
of 10 pm. For commercial SNF and crud the value of this parameter should be 3.5 pm and
5.0 pm, respectively. The unit of this input is pm as discussed in Section 6.2.2.4.
7. Input: Maximum Respirable Particle Size
Default Value: 4.7 pm (good for commercial SNF particle distributions)
Description: This parameter is used only for the iterative method described in
Section 6.2.2.4. It is equal to the diameter that is considered the maximum respirable size,
below which the distribution has an AMAD of 10 pm (or an MMD of 3.5 pm for fuel or
5.0 pm for crud). The unit of this input is pm. If this value is too high or too low, then the
results worksheet will indicate so (i.e., when the AMAD is not equal to 10 pm).
The results worksheet summarizes the inputs and presents the following results:
1. Output: Respirable Percent
Description: This is the respirable percent calculated using the iterative method. Note 1
below the results states whether this is the actual respirable percent. It is based on the "% of
Sum at Particle Diameter of Interest" being approximately equal to 50 percent. This will also
indicate whether the Maximum Respirable Particle Size must be changed, and how it should
be changed (i.e., reduced or increased). This is part of the iterative scheme and has the units
of percent as discussed in Section 6.2.2.4.
2. Output: Conservative Respirable Percent
Description: This is the respirable percent calculated using the AMAD-10 method described
in Section 6.2.2. Basically, this is the fraction of the mass of particulate with diameters less
than 10 pm and it is expressed as a percent. If the AMAD is less than 10 pm, then the
resulting respirable percent is not conservative.
3. Output: Non-Conservative Respirable Percent
Description: This is the respirable percent calculated using the MhfD method
(non-conservative compared to the AMAD-10 method) described in Section 6.2.2. Basically,
this is the fraction of the mass of particulate with diameters less than the respirable fraction
cut-off MMD previously described. It is expressed as a percent. If the AMAD is less than
10 lm, then the resulting respirable percent is very non-conservative.
000-00C-MGRO-01700-000-000 A-2 of A-30 October 2004

Commercial SNF Accident Release Fractions
4. Output: Mass Median Diameter
Description: This is the median diameter of the entire mass distribution calculated using
Equation 31. It is used as a check of the MMD calculated using the iterative process shown
in iten1 5 below. It has the unit of pm.
5. Output: MMD
Description: This is the median diameter of the entire mass distribution and it is used to
calculate the AMAD for the whole distribution. It has the unit of pm.
6. Output: AMAD
Description: This is the activity median aerodynamic diameter of the entire mass distribution.
It has the unit of pm.
7. Output: Percent of Sum at Particle Diameter of Interest
Description: For a particle distribution with an AMAD greater than 10 pm, this value is used
by the iterative method to determine the respirable fraction. It is desired that this sum should
be approximately 50 percent.
The particle distribution calculation worksheet performs the respirable fraction calculations.
The first column of this sheet contains the particle diameters that are used to determine the
lognormal particle distribution probability in column two. Column three divides this lognormal
probability by the diameter thereby creating the normalized lognormal probability function.
The fourth column multiplies the third column by the change in the diameter (Ad). The fifth
column sums the normalized probabilities providing the integrated particle distribution.
The sixth column calculates the particle volume assuming the particles are spheres. Using the
-inputted density, the mass of each particle is established in column seven by multiplying the
particle volume and density. The lognormal distribution probability and particle mass are
multiplied in column eight. The ninth column normalizes the particle mass lognormal
distribution probability with the total mass providing the differential mass particle distribution.
The tenth column re-normalizes the mass distribution that is less than maximum respirable
particle size to establish the AMAD of this fraction of the distribution. The final column sums
the differential mass particle distribution in column nine, providing the integral mass
distribution.
October 2004

Commercial SNF Accident Release Fractions
4. Output: Mass Median Diameter
Description: This is the median diameter of the entire mass distribution calculated using
Equation 31. It is used as a check of the MMD calculated using the iterative process shown
in iten1 5 below. It has the unit of pm.
5. Output: MMD
Description: This is the median diameter of the entire mass distribution and it is used to
calculate the AMAD for the whole distribution. It has the unit of pm.
6. Output: AMAD
Description: This is the activity median aerodynamic diameter of the entire mass distribution.
It has the unit of pm.
7. Output: Percent of Sum at Particle Diameter of Interest
Description: For a particle distribution with an AMAD greater than 10 pm, this value is used
by the iterative method to determine the respirable fraction. It is desired that this sum should
be approximately 50 percent.
The particle distribution calculation worksheet performs the respirable fraction calculations.
The first column of this sheet contains the particle diameters that are used to determine the
lognormal particle distribution probability in column two. Column three divides this lognormal
probability by the diameter thereby creating the normalized lognormal probability function.
The fourth column multiplies the third column by the change in the diameter (Ad). The fifth
column sums the normalized probabilities providing the integrated particle distribution.
The sixth column calculates the particle volume assuming the particles are spheres. Using the
-inputted density, the mass of each particle is established in column seven by multiplying the
particle volume and density. The lognormal distribution probability and particle mass are
multiplied in column eight. The ninth column normalizes the particle mass lognormal
distribution probability with the total mass providing the differential mass particle distribution.
The tenth column re-normalizes the mass distribution that is less than maximum respirable
particle size to establish the AMAD of this fraction of the distribution. The final column sums
the differential mass particle distribution in column nine, providing the integral mass
distribution.
October 2004

Commercial SNF Accident Release Fractions
Attachment B
Microsoft Excel 97 Spreadsheets for the Calculation of the Respirable Fraction of Crud
The following Excel Spreadsheet is nearly identical to the spreadsheet presented in
Attachment A. It consists of three separate worksheets: (1) the controller worksheet, which
basically controls the inputs, (2) the particle distribution calculations worksheet, which perfoms
the respirable fraction calculations, and the (3) results worksheet. A brief discussion of the
inputs and outputs and calculations performed in this spreadsheet are discussed in Attachment A.
The values calculated in this spreadsheet are for crud that has the characteristics noted in
Section 6.2 of this analysis. The user supplied inputs option has been used in the controller
worksheet to establish the crud respirable fraction.
October 2004



Summarv of Results -
[All Values Automatically Updated] I See
Results: 1 Nott
Res~irable Percent = 30.0517 % I 1
[AN Values Automatically Updated]
Conservative ~espirable Percent = 53.64537 %
Non-Conservative Respirable Percent = 14.92573 %
Mass Median Diameter = 9.71828 pm
Mass Median Diameter (MMD) = 9.551214 pn
AMAD = 19.10243 pn
% of Sum at Particle Diam. Of Interest = 49.66684 %
Inputs:
Geometric Mean fd) = 3 um
2
3
4
5
6
7
~,
Standard Deviation (0) = 1.87
Maximum Respirable Particle Diameter = 6.9 p
Particulate Density (p) = 5.2 glcc
Dynamic Shape Factor (K) = 1.3
lote 1: Fraction of Distribution with MMD -5 microns (i.e., 10 microns AMAD).
lote 2: Fraction of Distribution less than 10 microns (if AMAD < 10 microns value is non-conservative).
lote 3: Fraction of Distribution less than 3 microns (if AMAD c 10 microns value is non-conservative).
lote 4: Mass Median Diameter for entire distribution - used to check the MMD note 5
lote 5: MMD for entire distribution.
lote 6: AMAD for entire distribution (if < 10 microns then 100% respirable, if > 10 microns could be 0% respirable).
lote 7: Desire to be -50% which means particles with diameter < Max Respirable Particle Diam. have an MMD of 5 mm.




Commercial SNF Accident Release Fractions
Attachment C
GRAVITATIONAL DEPOSITION CONFIRMATORY ANALYSIS
NUREGICR-0722 (Lorenz et al. 1980, p. 105), states that a small fraction of fuel particles
ejected fiom a burst fuel pin was carried out of the furnace tube into the thermal gradient tube
and filter pack. At the time of rupture, the velocity of steam flowing past the rupture point,
through the furnace tube and down through the filter pack was 15 cds. Thus, for particles to
settle out before reaching the thermal gradient tube, they would have to fall at a rate of about
3 cm/s (the terminal settling velocity). NUREGICR-0722 (Lorenz et al. 1980, p. 105) states that
particles with diameters greater than 12 to 15 pm would fall at this rate. This has been
confirmed by sampling (albeit somewhat randomly and sparsely) some of the particulate
collected in the filters with a scanning electron microscope and determining that these
particulates had diameters of typically 10 pm (Lorenz et al. 1980, p. 105 and Appendix C). To
confirm that particles with diameters greater than 12 to 15-pm settle out before reaching the
thermal gradient tube, gravitational deposition methods, presented in Appendix B of ANSI
N13.1-1969, will be applied. The first step is to confirm the terminal settling velocity.
According to ANSI N13.1-1969, the length for 100 percent deposition (cm) is:
8rV
L,, = -
3%
where
Lloo is the length for 100 percent deposition (cm)
r is the radius of the tube (cm)
Vis the average velocity in the tube (cm/s)
u, is the terminal settling velocity (cm/s).
The 100 percent deposition length in this case is assumed to be approximately the length of the
furnace tube (i.e., 44 cm). The actual length will be somewhat shorter depending on where the
actual rupture point occurred. The average velocity in the tube is stated to be approximately
15 cm/s and the radius of the furnace tube is approximately 3.5 cm based on NUREGICR-0722
(Lorenz et al. 1980, Figure 4 on p. 105). Thus, the Equation C-1 can be solved for the terminal
settling velocity:
(Eq. C-2)
000-00C-MGRO-01700-000-000 C-l of C-4 October 2004





Commercial SNF Accident Release Fractions
This velocity is very close to the 3 c d s in NUREGICR-0722 (Lorenz et al. 1980, p. 105).
With the terminal settling velocity now calculated, the diameter of the particles settling at this
velocity can be calculated from Stoke's Law (ANSI N13.1-1969, p. 34):
where
g is the gravitational constant (980 cm/s2)
d, is the diameter of the particle (cm)
p, is the density of the particle (g/cm3)
p, is the density of the steam (g/cm3)
(Eq. C-3)
pis the steam viscosity (g/cm-s)
K,,, is the Cunningham correction for slip (unitless).
This equation can be solved for the diameter of the particle.
The terminal velocity has already been shown to be approximately 3 cm/s, the particle density is
assumed to be 10.96 g/cm3 (Section 5, Assumption 5.6), the Cunningham correction factor has
been shown'to be nearly unity (see Figure 2), and the density of steam and viscosity of steam at
900°C are approximately 1.8 x g/cm3 and 2.8 x g/cm-s, respectively.' Thus, the
particle diameter can be calculated from:
(Eq. C-4)
' The steam density is from: Fundamentals of Classical Thermodynamics (Van Wylen and
Sonntag 1986, p. 641). The steam viscosity is from: Flow of F'luids Through Valves, Fittings,
and Pipe (Crane Company 1988, p. A-2).
October 2004



This value is essentially equal to the smallest diameter presented in IWREGICR-0722 (Lorenz et
al. 1980, p. 105) as having settled before reaching the thermal gradient tube (i.e., 12 pm). If
shorter 100 percent deposition lengths were considered in this analysis, then the terminal settling
velocities would increase which in turn results in a larger diameter. This larger diameter is likely
near the 15-pm diameter noted in NUREGICR-0722 (Lorenz et al. 1980, p. 105).
Thus, both the terminal settling velocity of 3 cmls and the diameter of the fuel particles which
settled out before reaching the thermal gradient tube has been confirmed from NUREGICR-0722
(Lorenz et al. 1980, p. 105) using methods presented in Appendix B of ANSI N13.1-1969.
October 2004




Commercial SNF Accident Release Fractions
Attachment D
Summary of Release Fractions from Other NRC Licensed FacilitieslCasks
The following SARs for cask systems and independent spent fuel storage installations (ISFSIs)
provide valuable information on values of release fractions listed in TableD-1 used in
consequence analyses and approved by the NRC. The values of release fractions given in
Table D-1 are provided for information only and they do not impact the analysis itself. Each of
the following documents is found in the Nuclear Regulatory Commission's Public Document
Room, 2120 L Street, NW, Lower Level, Washington, DC 20555-0001, and are cross-referenced
by number in Tables D-1, D-2, and D-3.
(1) Fort St. Vrain Independent Spent Fuel Storage Installation Safety Analysis Report,
Revision 2 (INEL 1998).
(2) North Anna Power Station Independent Spent Fuel Storage Installation, License
Application (Virginia Electric and Power Company [VEPCO] 1995).
(3) "Docket No. 72-1 1, Rancho Seco Independent Spent Fuel Storage Installation, Revision 1
to the Rancho Seco Independent Spent Fuel Storage Installation License Application and
Safety Analysis Report" (Shetler 1993) for the Sacramento Municipal Utility District
(SMW.
(4) Safety Analysis Report for the ZNEL TMI-2 Independent Spent Fuel Storage Installation,
Revision 0 (INEL 1996).
(5) Trojan Independent Spent Fuel Storage Installation, Safety Analysis Report (PGE n.d.)
(6) 10 CFR 72 Topical Safety Analysis Report for the Holtec International Storage,
Transport and Repository Cask System (HI-STAR-100 Cask System) (Holtec International
1995).
(7) Safety Analysis Report Large On-Site Transfer and On-Site Storage Segment, CLIN
0004PC (Westinghouse 1996).
(8) Safety Analysis Report for the NUHOMSGMP187 Multi-Purpose Cask (Vectra
Technologies 1996).
(9) Safety Analysis Report for the Standardized NUHOMSB Horizontal Modular Storage
System for Irradiated Nuclear Fuel (Vectra Technologies 1995).
(10) 71V-24 Dry Storage Cask Topical Report (Transnuclear 1989).
(1 1) Safety Analysis Report for the TranStorTMStorage Cask System (SNC 1996).
(12) Final Design Package Babcock & Wilcox BR-100, 100 Ton RaiNBarge Spent Fuel
Shipping Cask (B&W Fuel Company 1991).
October 2004



I Commercial SNF Accident Release Fractions
(13) Safety Analysis Report for the NAC Legal Weight Truck Cask (NAC 1995).
I (14) Safety Analysis Report for the TranStorTM Shipping Cask System (SNC 1997).
I Indeuendent Spent Fuel Storage Installations (ISFSIs)
Table D-1 summarizes the release fractions used in the preceding SARs of the various ISFSIs.
For each ISFSI, a hypothetical loss of confinement accident was analyzed. In each case,
100 percent of the fuel cladding was assumed ruptured and the cask breached in such a manner
that the release fraction from the cask to the environment was equal to unity. In nearly each
case, except for the Fort St. Vrain and INEL TMI-2 ISFSIs, 85Kr is either the only analyzed
radionuclide or the only radionuclide that significantly contributes to the dose at the site
boundary. It is also noted in the INEL TMI-2 SAR that the fraction of particulate and solids that
are released from the cask to a filtration system is approximately 1 percent. This value is
expected to be high compared to releases from originally intact fuel assemblies, because most of
the TMI-2 fuel is no longer confined by cladding. The release fractions for the gases 3~ and ' 2 9 ~
span the whole spectrum (i.e., a release fraction between 0 and 1). The release fraction for the
gas 85Kr ranges from 0.25 to 1.00, with an equal number selecting 0.3 and 1.0.
Storage Cask Systems
Table D-2 summarizes the release fractions used in the preceding SARs of the various storage
cask systems. For each storage cask, a hypothetical loss of confinement accident was analyzed.
In each case, 100 percent of the fuel cladding was assumed ruptured and the cask breached in
such a manner that the release fraction from the cask to the environment was equal to unity.
In each case, 85Kr is either the only analyzed radionuclide or the only radionuclide that
significantly contributes to the dose at the site boundary. No particulate releases are considered
in any of these SARs, as the particulates are expected to locally deposit near their release point.
The release fractions for the gases 3~ and Iz91 ranges from 0 to 0.3. The release fraction for the
gas 85Kr is nearly always equal to 0.3 for the storage casks with the exception of the z 4 ~ ~ cask
system, which used a release fraction of 0.1.
Transportation Cask Systems
Table D-3 summarizes the release fractions used in the preceding SARs of various transportation
cask systems. For each transportation cask, the accident release fractions are analyzed under the
Confinement chapter of the SAR. In each case, 100 percent of the fuel cladding was assumed
ruptured and the cask breached in such a manner that the release fraction from the cask to the
environment was equal to unity. In each case, "Kr is the only radionuclide that significantly
contributes to the dose at the site boundary. No particulate releases are considered in any of
these SARs, as the particulates are expected to locally deposit near their release point. However,
the Sierra TranStor cask does consider 100 percent of the crud (60~o) to be released from the fuel
rod surfaces and 100 percent of this to be aerosolized. The release fractions for tritium ranges
from 0.1 to 0.3. The release fraction for Iz91 ranges from 0 to 0.3, but its contribution to the dose
is nearly always ignored. The release fraction for the gas 8 5 ~ r is always equal to 0.3 for these
transportation casks.
000-00C-MGRO-01700-000-000 D-2 of D-6 October 2004