9
Nuclear Fuel Cycle
9.1 Characteristics of the Nuclear Fuel Cycle . . . . . . . . . . . . . . . . . 193
9.2 Front End of the Fuel Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
9.3 Fuel Utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
9.4 Back End of Fuel Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
9.5 Uranium Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
9.1 Characteristics of the Nuclear Fuel Cycle
9.1.1 Types of Fuel Cycle
The nuclear fuel cycle is the progression of steps in the utilization of fissile
materials, from the initial mining of the uranium (or thorium) through the
final disposition of the material removed from the reactor. It is called a cycle
because in the general case, some of the material taken from a reactor may
be used again, or “recycled.”
Fuel cycles differ in the nature of the fuel used, the fuel’s history in the
reactor, and the manner of handling the fuel that is removed from the reactor
at the end of the fuel’s useful life (known as the spent fuel). For uranium-fueled
reactors—which means virtually all commercial reactors—a key difference is
in the disposition of the plutonium and other actinides that are produced in
a chain of neutron captures and beta decays that starts with neutron capture
in 238
U to produce 239
Pu (see Section 7.4).1
The actinides are important
because (1) some, especially 239
Pu, are fissile and can be used as nuclear fuel
in other reactors or in bombs, and (2) many of the actinides have long halflives,
complicating the problems of nuclear waste disposal. The three broad
fuel cycle categories are as follows:
1
The actinides are the elements with atomic numbers Z greater than or equal
to that of actinium (Z = 89). (The terminology is not uniform and, sometimes,
actinium is not included among the “actinides.”) Neptunium (Z = 93), americium
(Z = 95), and curium (Z = 96) are referred to as minor actinides in view of their
low abundance in spent fuel compared to uranium (Z = 92) and plutonium
(Z = 94). Elements with atomic numbers greater than 92 are termed transuranic
elements.
193
194 9 Nuclear Fuel Cycle
N Once-through fuel cycle. This is sometimes called an open fuel cycle or a
“throw-away” cycle. It is not really a cycle, in that the spent fuel is treated
as waste when it is removed from the reactor and is not used further. The
239
Pu and other actinides are part of these wastes.
N Reprocessing fuel cycle. In the present standard reprocessing fuel cycle,
plutonium and uranium are chemically extracted from the spent fuel. The
plutonium is used to make additional fuel, often by mixing it with uranium
oxides to produce mixed-oxide fuel (MOX) for use in thermal reactors.
This provides additional energy and changes the nature of the wastes. In
potential variants of the reprocessing fuel cycle, the minor actinides would
also be extracted, and they and the plutonium would be incorporated in
fresh fuel for fast reactors (see Section 9.4.3).
N Breeding cycle. For this cycle, the reactor is designed so that there is more
fissile material (mostly 239
Pu) in the spent fuel than there was in the fuel
put into the reactor (see Section 8.3). As in the reprocessing fuel cycle, the
plutonium can be removed and be used in another reactor. With a sequence
of such steps, fission energy is in effect extracted from a substantial fraction
of the 238
U in uranium, not just from the small 235
U component, increasing
the energy output from a given amount of uranium by a factor that could,
in principle, approach 100.
It may be noted that uranium accounts for most of the mass of the nuclear
wastes in the once-through cycle. It is separated out in the reprocessing and
breeding cycles for possible reuse in reactor fuel.
At present, all U.S. commercial reactors and the majority of reactors worldwide
are operating with a once-through fuel cycle, although some countries,
particularly France, have large-scale reprocessing programs with use of plutonium
in the form of MOX fuel. It should be noted, of course, that even in the
once-through fuel cycle, the potential for eventually using the fuel in a reprocessing
cycle remains until the fuel is disposed of irretrievably. No country is
employing a breeder cycle at this time, although France appeared on the verge
of attempting such a program with its Phenix and Superphenix reactors—but
this effort has been abandoned, at least for the time being (see Section 8.3.3).
Although virtually all of the world’s commercial reactors have used uranium
fuel, there is continuing interest in the use of thorium fuel.2
In a thorium
fuel cycle, the thorium (all 232
Th in nature) serves as the fertile fuel. Neutron
capture and beta decay result in the production of 233
U, which has favorable
properties as a fissile fuel. To start the thorium cycle, a fissile material such as
235
U or 239
Pu is needed, but once begun, it can be sustained if enough 233
U
is produced to at least replace the initial fissile material. It is often argued
that a thorium cycle is preferable to a uranium cycle, because if 233
U is ex-
2
The Fort St. Vrain high-temperature, gas-cooled, graphite-moderated reactor in
Colorado, which was shut down in 1989, is one of several exceptions to the exclusive
use of uranium, having used thorium for part of its fuel [1, p. 41].
9.2 Front End of the Fuel Cycle 195
tracted from the spent fuel, it can be “denatured” by mixing it with natural
uranium to make a fuel that cannot be used in a bomb. Bomb material could
be obtained only after the isotopic separation of 233
U. In contrast, bomb material
can be obtained from a uranium-fueled reactor by chemical separation
of the plutonium (see Chapter 17). Isotopic separation is technically more
difficult than chemical separation; thus, a thorium fuel cycle could be more
proliferation resistant than a uranium fuel cycle unless, in the latter case, the
plutonium is well protected from diversion or theft.
9.1.2 Steps in the Nuclear Fuel Cycle
A schematic picture of the fuel cycle is shown in Figure 9.1, which indicates
alternative paths, with and without reprocessing [2]. The steps in the fuel
cycle that precede the introduction of the fuel into the reactor are referred to
as the front end of the fuel cycle. Those that follow the removal of the fuel
from the reactor comprise the back end of the fuel cycle. At present, there is
only a truncated back end to the fuel cycle in the United States, as virtually
all commercial spent fuel is accumulating in cooling pools or storage casks at
the reactor sites.
Implementation of a spent fuel disposal plan, or of a reprocessing and waste
disposal plan, would represent the “closing” of the fuel cycle. This closing is
viewed by many to be an essential condition for the increased use of nuclear
power in the United States—and perhaps even for its continued use beyond
the next several decades.
Key aspects of the fuel cycle will be surveyed in the remainder of this
chapter. The fuel cycle will be discussed particularly in the context of light
water reactors, in view of their dominance among world nuclear reactors. The
main aspects are relevant to other types of reactor as well. A more extensive
treatment of the crucial step of waste disposal will be given in Chapters 10–13.
9.2 Front End of the Fuel Cycle
9.2.1 Uranium Mining and Milling
Uranium Deposits in the Earth’s Crust
The concentration of uranium varies greatly among geological formations.
The average concentration in the Earth’s crust is about 3 parts per million
(ppm) by weight, but extremes extend from under 1 ppm to something in the
neighborhood of 500,000 ppm.3
Uranium resources are widely distributed, with substantial uranium production
in many countries, including Australia, Canada, Kazakhstan, Namibia,
3
For example, one deposit in Canada is identified as having zones of “over 50%
U3O8,” which translates to over 42% uranium [3].
196 9 Nuclear Fuel Cycle
Fig. 9.1. Schematic of the nuclear fuel cycle. (From Ref. [2, p. 45]).
9.2 Front End of the Fuel Cycle 197
Niger, the Russian Federation, South Africa, the United States, and Uzbekistan
[4, p. 36]. Most of the uranium now used in the United States is imported,
with Canada being the largest supplier. Through 2000, the United States had
been the world’s leader in cumulative production of uranium, with Canada
a close second. However, the U.S. share of production declined by the 1990s,
and now Canada and Australia are the world’s main suppliers of uranium.
Together, they accounted for about 50% of world production in 2000 [4].
Uranium in rocks is mostly in the form of a uranium oxide, U3O8. In
“conventional” mining, the rock is extracted from open pit or underground
mines, the U3O8 is then extracted in the milling process by crushing the rock
and leaching with acid, and the U3O8 is then recovered from the liquid and
dried. The concentrated U3O8 is known as yellowcake.4
In an “unconventional”
method, appropriate for only certain types of uranium deposits, U3O8
is extracted by in situ leaching (i.e., by pumping a leaching agent through the
ore without physical removal of the rock).5
At low concentrations, the uranium content is expressed in terms of the
uranium grade, given in percent by weight of either uranium or U3O8. Thus,
ore which is 1000 ppm of uranium corresponds to grades of 0.100% U or 0.118%
U3O8.6
In an extensive 1983 study of U.S. uranium ores, deposits were listed
with U3O8 grades ranging from under 0.01% to over 1.8%, with a median of
about 0.1% [6, p. 39]. The higher the grade, the less the amount of ore that
must be extracted, which, in general, leads to lower costs. Ores below a grade
of 0.05% are considered low-grade ores and have not been widely needed. Of
course, the ultimate criterion is overall cost, not grade per se, and at one time
open-pit mining utilized ores down to 0.04% [7, p. 411].
During 2001, most of the uranium extraction in the United States was done
by in situ leaching, using ores ranging in grade from 0.09% to about 0.20%
U3O8 [8]. Worldwide, conventional mining dominates but is now economically
practical only at higher uranium grades (“above a few tenths of a percent”) [1,
p. 25].
Radon Exposures from Uranium Mining and Mill Tailings
In the early days of uranium mining, little attention was paid to radiation
safety. In the Middle Ages, long before uranium had been identified as an element,
metal miners in southern Germany and Czechoslovakia contracted lung
ailments, called Bergkrankheit (“mountain sickness”). Modern scientists have
attributed the ailment to lung cancer caused by a high uranium concentration
4
U3O8 is not yellow in its pure form. Yellowcake is about 85% U3O8 [5, p. 241],
and the yellow color results from another uranium compound in the ore.
5
The designations “conventional” and “unconventional” correspond to those, for
example, of Ref. [1, p. 25].
6
In international usage, “grade” usually refers to U content, whereas in U.S. DOE
documents it refers to U3O8 content. Note: U3O8 is 84.8% uranium and 15.2%
oxygen, by weight.
198 9 Nuclear Fuel Cycle
that, by chance, was in the rock formations being mined. The decay of the
radionuclides in the uranium series proceeds from 238
U through several steps
to 226
Ra and then to radon gas (222
Rn) and its radioactive progeny. Inhalation
of these “radon daughters” can lead to lung cancer (see Section 3.5.1).
As one would expect, the problem of radon exposure is more extreme in
uranium mines than in other sorts of mine. It became a particularly serious
problem in a number of countries—for example, in the United States,
Czechoslovakia, and Canada—when large-scale uranium mining was begun in
the 1940s to meet the demands of nuclear weapon and nuclear power programs.
By the late 1950s, steps were initiated in the United States to reduce
radon exposures, mainly through better ventilation, and by the 1970s, the
average exposures of uranium miners had become quite low (lower than that
from indoor radon in many homes). However, a good deal of damage had already
been done, and there is unambiguous evidence of increased lung cancer
fatalities among uranium miners.
The residues of the milling operation, representing the remainder of the ore
after extraction of the U3O8, are the mill tailings. All of the uranium progeny,
starting with 230
Th, are present in the tailings.7
The radionuclide 230
Th has a
half-life of 75,400 years and thus sustains the remainder of the uranium series
for a long period of time. This results in the continuous production of radon,
some escaping to the atmosphere. Of course, these steps do not increase the
rate of radon production above what it would have been without mining, but
the radon in the tailings can more readily reach the atmosphere than can radon
in underground ore. At one time, this was viewed by some as constituting
an important environmental hazard, and it is still deemed necessary to take
remedial measures to limit radon emissions from the tailings (using overlying
layers of material to impede radon escape). However, interest in the issue has
diminished as it has become obvious that exposures from “normal” indoor
radon pose a much more serious problem, in terms both of the number of
people impacted and the magnitudes of the radon concentrations to which
they are exposed.8
9.2.2 Enrichment of Uranium
Preparation for Enrichment: Conversion
There are a variety of approaches to the enrichment of uranium, each taking
advantage of the small mass difference between 235
U and 238
U. In the most
used of these processes, it is necessary to have the uranium in gaseous form.
For that purpose, the U3O8 is chemically converted to gaseous uranium hexafluoride,
UF6. This is the compound of choice, because UF6 is a gas at lower
7
The 234
U remains in the yellowcake and the radionuclides between 238
U and 234
U
in the uranium series are short-lived.
8
For a comparison of the hazards from mill tailings and indoor radon, see, for
example, Ref. [9].
9.2 Front End of the Fuel Cycle 199
temperatures than can be reached by any other uranium compound in gaseous
form [10, p. 589].
Degrees of Enrichment
Natural uranium has an isotopic abundance by number of atoms of 0.0055%
234
U, 0.720% 235
U and 99.275% 238
U.9
In the remainder of the discussion of
uranium isotopic enrichment, we will follow the standard practice of describing
the 235
U fraction in terms of mass rather than, as is common in many other
scientific applications, of number of atoms.10
For natural uranium, the 235
U
abundance by mass is 0.711%. The presence of the small amount of 234
U is
often ignored, because corrections on the order of 10−4
or less are irrelevant.
The fissile nuclide in thermal reactors is 235
U. For reactors that require
uranium with a higher fraction of 235
U than is found in natural uranium,
enrichment is necessary. This is, of course, the case for light water reactors
(LWRs). Fuel used in LWRs in past years has been enriched to 235
U concentrations
ranging from under 2% to over 4%. The anticipated average for
the United States, for cumulative production up until about 2010, is 3.0% for
BWR fuel and 3.75% for PWR fuel.11
The material used in LWRs is known as slightly enriched uranium, in contrast
to the highly enriched uranium used for nuclear weapons and submarine
reactors. Within the core of a given reactor, enrichments vary with the location
of the fuel assemblies. As discussed later in the context of the burnup of
fuel, there is a general trend toward using fuel with higher initial enrichments.
The products of the enrichment process are the enriched material itself
and the depleted uranium, sometimes called enrichment tails. Typically, enrichment
tails have in the neighborhood of 0.2% to 0.35% 235
U remaining [12,
p. 7]. As one goes to lower concentrations of 235
U in the tails, the consumption
of uranium ore is reduced, but the cost of enrichment is increased. Thus, there
is a trade-off.
The depleted uranium is sometimes used in special applications. Its use
in armor-piercing shells, where the high density of uranium is advantageous
(ρ ≈ 19 g/cm3
), has led to some public concern about the resulting environmental
risks. However, depleted uranium has a lower specific activity than
9
The 234
U arises as a member of the 238
U series, with an abundance relative to 238
U
that is inversely proportional to the half-lives of the two isotopes (2.45 × 105
yr
and 4.468 × 109
yr, respectively).
10
These descriptions of isotopic abundance are related by the expression w = [(1 −
δ)/(1 − xδ)] x, where, specialized to the case of uranium, w is the ratio of 235
U
mass to total uranium mass, x is the ratio of the number of 235
U atoms to the
total number of uranium atoms, and δ is the ratio of the difference between the
238
U and 235
U atomic masses to the 238
U atomic mass. For low enrichments (with
δ = 0.0126 for uranium), w
.
= 0.987x, and there is little difference between the
two formulations. For natural uranium, x = 0.00720 and w = 0.00711.
11
This is the planning basis for the Yucca Mountain nuclear waste repository [11,
p. 3–13].
200 9 Nuclear Fuel Cycle
does natural uranium, and there is no evidence of appreciable radiation hazards
except for occupants of a closed vehicle that has been struck by a shell
that partially vaporizes within it.12
Methods for Enrichment
The leading enrichment methods in terms of past or anticipated future use
are as follows:13
N Gaseous diffusion. The average kinetic energy of the molecules in a gas is
independent of the molecular weight M of the gas and depends only on the
temperature. At the same temperature, the average velocities are therefore
inversely proportional to
√
M. For uranium in the form of UF6, the ratio
of the velocities of the two isotopic species is 1.0043.14
If a gas sample
streams past a barrier with small apertures, a few more 235
U molecules
than 238
U molecules pass through the barrier, slightly increasing the 235
U
fraction in the gas. The ratio of 235
U/238
U before and after passing the
barrier is the enrichment ratio α. Its ideal or maximum value is given by
the velocity ratio α = 1.0043. However, one cannot calculate the number of
stages of diffusion needed to achieve a given enrichment merely in terms
of powers of α, because the ideal value is not achieved in practice and
because it is necessary to continually recycle the less enriched part of the
stream. Typically, if one starts with natural uranium (0.71%) and with
tails depleted to 0.3%, it is found that about 1200 enrichment stages are
required to achieve an enrichment of 4% [15, p. 36].
N Centrifuge separation. Any fluid—liquid or gaseous—can be separated in
a high-speed centrifuge. The centrifugal action causes the heavier component
to become more highly concentrated at large radii. As in gaseous
diffusion, only a small gain is made in any one stage, and high enrichments
of the UF6 are reached using multiple centrifuge stages, with the slightly
enriched output of one stage serving as the input to the next one. The
centrifuges used for uranium enrichment are rotating cylinders. Uranium
that is slightly enriched in 238
U (and depleted in 235
U) can be extracted
from the outer region of the cylinder and returned to an earlier stage in
the centrifuge cascade. Uranium slightly enriched in 235
U can be extracted
from regions near the center and used as input to the next higher stage
in the array of centrifuge units. High enrichments of the UF6 are reached
using multiple centrifuge stages. The power requirement for a given degree
of enrichment is much less for centrifuge separation than for diffusion
separation.
N Aerodynamic processes. These processes exploit the effects of centrifugal
forces, but without a rotating centrifuge. Gas—typically UF6 mixed with
12
In this case, direct damage from the shell is a still greater concern.
13
Detailed discussions of these methods are given in, for example, Refs. [13] and [14].
14
The atomic mass of fluorine (F) is 19.00 u.
9.2 Front End of the Fuel Cycle 201
hydrogen—expands through an aperture, and the flow of the resulting gas
stream is diverted by a barrier, causing it to move in a curved path. The
more massive molecules on average have a higher radius of curvature than
do the lighter molecules, and a component enriched in 235
U is preferentially
selected by a physical partition. The process is repeated to obtain
successively greater enrichments. The gas nozzle process was developed in
Germany as the Becker or jet nozzle process. A variant with a different
geometry for the motion of the gas stream, the so-called Helikon process,
has been developed and used in South Africa.
N Electromagnetic separation. When ions in the same charge state are accelerated
through the same potential difference, the energy is the same and
the radius of curvature in a magnetic field is proportional to
√
M. Thus, it
is possible to separate the different species magnetically. This separation
can be done with ions of uranium and, so, conversion to UF6 is not, in principle,
necessary. Overall, this approach gives a low yield at a high cost in
energy, but it has the advantage of employing a relatively straightforward
technology.
N Laser enrichment. The atomic energy levels of different isotopes differ
slightly.15
This effect can be exploited to separate 235
U from 238
U, starting
with uranium in either atomic or molecular form. For example, in
the atomic vapor laser isotope separation (AVLIS) method, the uranium
is in the form of a hot vapor. Lasers precisely tuned to the appropriate
wavelength are used to excite 235
U atoms, but not 238
U atoms, to energy
levels that lie several electron volts above the ground state. An additional
laser is used to ionize the excited 235
U atoms.16
The ionized 235
U atoms
can be separated from the un-ionized 238
U atoms by electric and magnetic
fields. An alternative to the AVLIS method is the SILEX process
(separation of isotopes by laser excitation). It is based on the selective
dissociation of UF6 (a gas) into UF5 (a solid) [16]. The costs in energy of
laser enrichment are lower than those of other enrichment methods, but
a sophisticated laser technology is required, and, to date, there are no
commercial facilities for laser enrichment of uranium. Once mastered, the
laser technique is expected to be relatively inexpensive. On the negative
side, there have been fears that if the technique develops sufficiently, laser
separation may make it easy for small countries or well-organized terrorist
groups to enrich uranium for nuclear weapons.
15
This “isotope effect” was responsible for the discovery of 2
H. It arises for two
reasons: (1) the atomic energy levels depend on the reduced mass of the electrons,
which differs from the mass of a free electron by an amount proportional to me/M,
where me is the electron mass and M the atomic mass, and (2) the energy levels
of heavy atoms depend in a small measure on the overlap between the wave
functions of the innermost electrons and the nucleus, with differences between
isotopes due to differences in their nuclear radii.
16
The ionization energy to remove an electron from uranium in its unexcited
(ground) state is 6.2 eV.
202 9 Nuclear Fuel Cycle
Adopted Enrichment Practices
During World War II, not knowing which method would be the most effective,
the United States embarked on both diffusion and electromagnetic separation,
as well as still another method that was later discarded (namely thermal diffusion,
which exploits temperature gradients). The electromagnetic separation
technique was abandoned in the United States after World War II and was
widely considered to be obsolete. However, it was found in 1991, after the
Gulf War, that Iraq had been secretly using this approach in an attempt to
obtain enriched uranium for nuclear weapons.
In the United States since World War II, the enrichment program has relied
on gaseous diffusion, as did early European programs. Since the 1950s,
the DOE (and its predecessor agencies, starting with the AEC) operated two
large gaseous diffusion enrichment facilities—one in Paducah, Kentucky and
one in Portsmouth, Ohio. In 1999, operation of these facilities was privatized
under the management of the United States Enrichment Corporation (USEC),
and in 2001, the Portsmouth plant—the smaller of the two—was shut down.
Outside the United States, there are major enrichment facilities in France
and Russia and smaller ones in a number of other countries, including the
United Kingdom, Netherlands, Germany, Japan, and China [1, p. 33]. In recent
years (1999–2001), most of the enrichment for fuel used in U.S. reactors
has been carried out abroad, particularly in Russia [17, p. 28]. The gaseous diffusion
method is used in France and the United States, whereas the centrifuge
method is used almost everywhere else.
A 2001 OECD report anticipated that most new plants will use the centrifuge
method [1, p. 85]. The USEC, which has been the sole U.S. company
pursuing enrichment, plans to replace its diffusion plant with a “secondgeneration”
centrifuge plant [18]. To that end, it submitted to the NRC in
February 2003 a license application for a preliminary demonstration facility
scheduled to be on-line in 2005. A larger, full-scale plant—the so-called
American Centrifuge—is planned for later in the decade [19]. In addition, the
Louisiana Energy Services Partnership—an organization that includes, among
others, Urenco (a major European enrichment company), the Westinghouse
Electric Company, and several U.S. utilities is seeking to build a centrifuge
enrichment facility in New Mexico [20].
USEC has also worked on developing laser enrichment technology as a
“third-generation” option. It originally focused on the AVLIS method and
later on the SILEX process, but as of Spring 2003, USEC decided to concentrate
on its centrifuge projects to the exclusion of laser options [21].
Separative Work
In a 235
U enrichment process, there are three streams of material: the input
or feed, the output or product, and the residue or tails. The system operates
9.2 Front End of the Fuel Cycle 203
with a cascade of steps, with the enrichment of the product increasing successively
in each step.17
As the enrichment cascade progresses, the tails from
an intermediate stage have a higher 235
U concentration than the original feed
material, and these tails can profitably be returned to the cascade. There are
different strategies for reusing the tails of successive steps to maximize the
efficiency of the process, including an “ideal cascade” (see, e.g., Ref. [5]).
The difficulty of carrying out uranium enrichment, as measured, for example,
by the relative energy required in the diffusion process, is described by
a quantity known as the separative work.18
Separative work has the dimensions
of mass and is specified in separative work units (SWU), as kg-SWU or
tonne-SWU. Figure 9.2 shows the separative work required to produce 1 kg of
enriched uranium product and 1 kg of 235
U in the form of enriched uranium,
for different degrees of final enrichment.
As seen in Figure 9.2, it requires more separative work per kilogram of
235
U to enrich uranium from 0.7% to 5% than to carry it the rest of the way
to 95% enrichment. Even 3% enriched uranium fuel is more than “halfway” to
95% enrichment. This could make the slightly enriched uranium produced for
reactors a somewhat attractive initial material for the production of uranium
for weapons (see Chapter 17).
Although the formalism was developed in the context of gaseous diffusion
enrichment methods, it is used for other processes as well. Separative work
serves as a general measure of what is achieved in the enrichment. For an
individual process, it serves also as a measure of relative energy consumption.
Different processes vary greatly in their energy consumption. For example,
gaseous diffusion uses 2.5 MWh/kg-SWU, while the gas centrifuge uses about
1/50th as much energy and laser isotope separation methods still less [12,
p. 530].
It is of interest, in the spirit of what is known as “net energy analysis,”
to compare the energy required to enrich uranium with the energy obtained
from it. The production by the diffusion process of 1 kg of 235
U in the form of
17
The logical structure of the system is similar to that of fractional distillation, and
some of the formalism was developed in the 19th century by Lord Rayleigh [5,
p. 649].
18
In a separation process, the masses are in the ratios:
MF
MP
=
wP − wT
wF − wT
and
MT
MP
=
wP − wF
wF − wT
,
where MF , MP , and MT are the masses and wF , wP , and wT are the 235
U
concentrations (by weight) of the feed, product, and tails, respectively. This result
follows from the conservation of total mass and 235
U mass: MF = MP + MT and
wF MF = wP MP +wT MT . The separative work in an isotopic enrichment process
is
∆V = MP VP + MT VT − MF VF ,
where V is the value function, defined as V = (1 − 2w) ln[(1 − w)/w]. (For the
derivation leading to this result see, e.g., Ref. [5, Chapter 12].)
204 9 Nuclear Fuel Cycle
Fig. 9.2. Separative work (in kg-SWU) as a function of 235
U enrichment for the
production of 1 kg of U and 1 kg of 235
U. (The initial feed concentration is wF =
0.00711 and the assumed tails concentration is wT = 0.0025.)
uranium enriched to 3.75% in 235
U, with a 235
U concentration in the tails of
0.25%, requires 142 kg-SWU (see Figure 9.2) or about 355 MWh. The reactor
output from 1 kg of 235
U is about 1 MWyr or 8760 MWh (see Section 9.3.2).
Therefore, the relatively energy-intensive diffusion process requires roughly
4% of the energy output of a reactor. Although this is the largest single energy
input to the production of nuclear power, it does not significantly reduce the
net positive energy balance from nuclear power.
9.2.3 Fuel Fabrication
Most nuclear fuel used in light water reactors is in the form of uranium dioxide
(UO2), also sometimes called “uranium oxide.” This is not a single compound,
but a mixture of oxides (UOn), where n typically ranges from 1.9 to 2.1 [22,
p. 226]. The UO2 is produced by chemical conversion of the enriched UF6. It is
9.3 Fuel Utilization 205
then processed into a fine powder and compacted and sintered to form rugged
pellets. During sintering, the oxygen content of the fuel can be adjusted. The
pellets are corrected in size, to close tolerances, by grinding.19
The pellets are
loaded into zircaloy fuel pins that are arranged in a matrix to form the fuel
assembly. As discussed in Section 8.2.3, the reactor core consists of a large
number of such assemblies.
9.2.4 Other Fuel Types
The focus here has been on UO2, which is the usual fuel for LWRs. Other fuel
types are of interest, however, even if not widely used at present (see, e.g.,
Ref. [1]). Possibilities include the following:
N Mixed-oxide fuel (MOX). MOX fuel, a mixture of uranium and plutonium
oxides, uses plutonium in order to exploit its energy content, reduce the
stocks of potential weapons materials, or both (see Section 9.4.2).
N Metal alloy fuels. Metallic fuel, in the form of alloys of uranium, provide
an alternative to oxide fuels that is easier to reprocess.
N Microsphere fuel particles. High-temperature gas-cooled reactors utilize
uranium or thorium oxide in the form of very small spheres, with multiple
layers of outer protection. These are the so-called TRISO fuel particles
(see Section 16.4.3).
N Thorium fuels. A fuel cycle based on thorium-232 as the fertile fuel and
uranium-233 as the fissile fuel could be used to supplement uranium resources.
It has the advantage of producing little plutonium and thereby
lessening waste disposal and proliferation problems.
N Molten salt. Although all operating commercial reactors presently use fuel
in solid form, it is also possible to have the fuel as a liquid uranium fluoride,
mixed with other liquid fluorides, as would be done in the proposed molten
salt reactor (see Section 16.6.1).
9.3 Fuel Utilization
9.3.1 Burnup as a Measure of Fuel Utilization
Thermal Efficiency of U.S. Reactors
The thermal efficiency of a reactor is the ratio of the electrical energy produced
to the total heat energy produced. Since 1973, the average thermal efficiency
of U.S. reactors has ranged between 30.6% and 32.1%, according to DOE
compilations [23, Table A6]. There has been a gradual improvement with
time, and since 1985 it has been above 31.5%, reaching 32.1% for the years
1996–2002. We will use the approximate figure of 32% as the nominal average
efficiency of LWRs.
19
For a discussion of the details of these processes, see Ref. [22, Section 7.5].
206 9 Nuclear Fuel Cycle
Basic Unit for Burnup: GWDT per MTHM
A useful measure of the performance in the nuclear fuel cycle is the energy
obtained per unit mass of fuel, known as the fuel’s burnup. The burnup is
commonly specified in megawatt-days or gigawatt-days of thermal output per
metric tonne of heavy metal (MWDT/MTHM or GWDT/MTHM). This is a
cumbersome notation for repeated use, and we represent GWDT/MTHM in a
more compact form as GWd/t (gigawatt-days per tonne). In standard energy
units, 1 GWd = 8.64 × 1013
joules (J).
For U.S. reactors, as well as most reactors elsewhere, the “heavy metal”
in the original fuel is uranium.20
The fuel is in the form of uranium oxide
(UO2). About 12% of the mass of the fuel is oxygen and, therefore, there
is a distinction between the mass of heavy metal and the mass of the fuel.
The heavy metal in the spent fuel removed from the reactor is still primarily
uranium, but it also includes isotopes of plutonium and—to a small extent—
other transuranic elements. Typically, the mass of heavy metal is about 3%
or 4% less in the spent fuel than in the initial fuel due to fission (including
fission of plutonium isotopes).21
A 1000-MWe reactor, operating at a typical thermal efficiency of 32%,
produces energy at the rate of 3125 MWt (where it is explicitly indicated
that this is the thermal output). One gigawatt-year of electric power therefore
represents a thermal output of 1141 GWd(t). If, for example, the average
burnup in a reactor is 40 GWd/t, the fuel consumption is 28.5 tonnes of
enriched uranium per gigawatt-year.
Trends in Burnup of LWR Fuel
Average burnup values for past years are shown in Table 9.1 along with the
average projected for the fuel to be deposited at the Yucca Mountain waste
repository. Overall, there has been a trend with time toward higher burnup,
on average roughly doubling in the 25 years from 1973 to 1998 and projected
to continue to rise. Thus, in a 1993 DOE projection, it was expected that
the median PWR fuel burnup for standard assemblies would be about 43
GWd/t in the year 2000—a value actually achieved in 1998—and 51 GWd/t
20
The main exception is for reactors that use a mixture of uranium and plutonium
in mixed-oxide fuels (see Section 9.4.2).
21
The designation “metric tons of heavy metal” (MTHM) commonly appears in
discussions of the utilization and disposal of nuclear fuel. Here, MTHM refers to
the heavy metal mass of the initial fuel. Alternatively, this can be made explicit
by using the designation “metric tonnes of initial heavy metal” (MTIHM). In
effect, “MTHM” and “MTIHM” are used interchangeably and mean the same
thing. Thus, the U.S. spent fuel inventory at the end of 1995 is given as 31926
MTHM in the 2002 Yucca Mountain EIS [24, Table A-7] and as 31952 MTIHM in
a 1996 report [25, Table 1.2]. (The 0.08% difference is insignificant compared to
the difference of several percent in the actual heavy metal contents of the initial
and spent fuel.)
9.3 Fuel Utilization 207
Table 9.1. Average burnup of U.S. spent fuel,
for BWRs and PWRs.
Year
Burnup (GWd/t)
Discharged BWRs PWRs
Annual
1973 12.4 23.7
1978 19.8 26.4
1983 26.7 30.1
1988 24.1 33.4
1993 30.3 38.9
1998 36.4 43.3
Average
1968–1998 25.1 33.4
Yucca Mountaina
33.6 41.2
a
This includes all commercial spent fuel slated for
Yucca Mountain (i.e., all of the fuel discharged
through about 2010).
Source: Refs. [26, Table 2], and [11, Table 3–6].
in 2010 [27, p. 144]. It should be noted that Table 9.1 gives only averages. In
any year, there is a wide disparity among reactors. Thus, in 1993, when the
average PWR burnup was 38.9 GWd/t, 12% of the spent fuel had a burnup
in excess of 45 GWd/t [28, p. 25].
An estimated average for the fuel expected to be deposited in the Yucca
Mountain repository (i.e., all of the fuel generated in the United States from
the start of nuclear power until about 2010) is given in Yucca Mountain planning
documents. The indicated average burnups for BWRs and PWRs are
33.6 GWd/t and 41.2 GWd/t, respectively, with average initial 235
U enrichments
of 3.03% and 3.75% [11, p. 3–13]. The corresponding weighted average
for all LWR fuel is 38.6 GWd/t, with an enrichment of 3.5%.
The burnup depends on the power density in the fuel and the length of
time the fuel is kept in the reactor. Unless the conversion ratio is very large,
a high burnup requires a high initial enrichment in 235
U. For example, model
calculations carried out at Oak Ridge National Laboratory included a fuel
cycle that gives a burnup of 60 GWd/t, with a 4.7% 235
U concentration and
the fuel kept in the reactor for three 600-day cycles (i.e., 1800 days) [29, p. 2.4–
3]. Looking to the future, the term “high burnup” is sometimes used to refer
to a burnup in excess of 70 GWd/t [1, p. 88].
High enrichment alone is not sufficient for high burnup. In addition, the
fuel and cladding must be able to withstand the added neutron bombardment
and the buildup of fission gases. This depends both on the fuel itself and on
the composition of the alloy used for the cladding. Further, to compensate
for the initial high reactivity with high enrichment, burnable poisons are used
with the fuel.
208 9 Nuclear Fuel Cycle
With higher burnup, the mass and volume of spent fuel required for a
given energy output are reduced. The consumption of uranium is also reduced,
because a larger fraction of the 235
U is consumed and more 239
Pu is produced
and consumed. In addition, the time intervals between refueling operations can
be longer, meaning that less reactor time is lost for refueling. High burnup
fuel is also less proliferation-prone than fuel with lesser burnup because the
concentrations of 240
Pu and 242
Pu are higher and the fuel is harder to handle
due to its high level of radioactivity (see Chapter 17).
9.3.2 Uranium Consumption and Plutonium Production
In uranium-fueled reactors, there is a continual destruction of 235
U, through
fission and neutron capture, and buildup of plutonium isotopes through neutron
capture and beta decay. The plutonium sequence starts with 239
Pu, following
neutron capture in 238
U, and continues to include plutonium isotopes
up to 242
Pu, as well other heavy radionuclides—for example, 241
Am (atomic
number Z = 95), which is produced primarily from the beta decay of 241
Pu
(T = 14.39 yrs). Similarly, radionuclides of atomic mass numbers 236 and 237
(primarily, 236
U and 237
Np), are formed through neutron capture in 235
U and,
for 237
Np, subsequent neutron capture and beta decay.
Most of the reactor’s energy output comes from the fission of 235
U. However,
as discussed earlier, production and fission of 239
Pu also play a significant
part in the energy economy of the reactor. Thus, the total energy output is
the sum of the energy from the fissile plutonium isotopes (239
Pu and 241
Pu)
plus the energy from 235
U and, to a much lesser extent, 238
U.
Table 9.2 gives the masses of the main actinide isotopes in the spent fuel
when it is discharged from the reactor, for a representative PWR case, namely
UO2 fuel with a 3.75% 235
U enrichment, a burnup of 40 GWd/t, and a residence
time in the reactor of a little over 3 years. The results in Table 9.2 were
calculated with the ORIGEN program, developed at the Oak Ridge National
Laboratory, which traces the production and consumption of the nuclides as
a function of time, taking into account the simultaneous nuclear processes for
all the nuclides, including fission, neutron capture, and radioactive decay.22
Important qualitative features of the results include the following:
N Fission products. The fission product production is about 40 kg, as seen
from the 40.4-kg reduction in the heavy metal mass.23
It may seem sur-
22
I am indebted for the calculation of the activities to Dr. Edwin Kolbe, Project
Manager for Radioactive Materials at the Swiss National Cooperative for the
Disposal of Radioactive Waste (NAGRA) [31]. The mass of each radionuclide
was calculated from the ratio of its activity to its specific activity [see Eq. (3.4)].
These ORIGEN results are also used in Chapter 10.
23
The precise number is slightly less than 40.4 kg due, in part, to neutron escape
from the reactor. The mass equivalent of the energy produced in the reactor is
0.04 kg, representing a very small additional “correction” to the calculated fission
product mass.
9.3 Fuel Utilization 209
Table 9.2. Activity and mass of actinides in PWR spent fuel (per MTHM), for
burnup of 40 GWd/t and 235
U enrichment of 3.75%.
Half-life Activity Mass Isotopic
Nuclide (years) (Bq) (kg) Percent
Input fuel
234
U 2.46 × 105
0.05 0.005
235
U 7.04 × 108
37.5 3.75
238
U 4.47 × 109
962.4 96.24
Total HMa
1000
Spent fuel
234
U 2.46 × 105
4.70 × 1010
0.2 0.02
235
U 7.04 × 108
6.85 × 108
8.6 0.90
236
U 2.34 × 107
1.21 × 1010
5.1 0.53
238
U 4.47 × 109
1.16 × 1010
934.4 98.54
Total U 948.2 100
237
Np 2.14 × 106
1.49 × 1010
0.57 87
239
Np 0.0065 7.18 × 1017
0.08 13
Total Np 0.65 100
238
Pu 87.7 1.38 × 1014
0.22 2.1
239
Pu 2.41 × 104
1.28 × 1013
5.56 53.2
240
Pu 6564 2.07 × 1013
2.46 23.6
241
Pu 14.29 5.93 × 1015
1.54 14.8
242
Pu 3.73 × 105
9.66 × 1010
0.66 6.3
Total Pu 10.4 100
241
Am 432.2 6.41 × 1012
0.05 24
243
Am 7370 1.18 × 1012
0.16 76
Total Am 0.21 100
244
Cm 18.1 1.54 × 1014
0.05
Total HM 959.6
Decrease in HM 40.4
a
The initial total heavy metal (HM) mass is the uranium mass.
Source: Half-lives from Ref. [30]; activities are based on ORIGEN calculation (see
text) [31]. The mass is calculated from the activity and half-life.
prising that the fission product mass exceeds the mass of the destroyed
235
U. However, in addition to fission in 235
U, there is fission of 239
Pu and
241
Pu.
N 236
U. Neutron absorption in 235
U leads to capture in 14% of the events
(see Table 7.1). Therefore, the destruction of 235
U is accompanied by the
production of a significant amount of 236
U. The isotope 236
U has a modest
capture cross section at thermal energies (σγ = 5.1 b) and a negligible
fission cross section. Its capture product, 237
U, decays with a rather short
half-life (6.75 days) to 237
Np (T = 2.14 × 106
years), which has a large
210 9 Nuclear Fuel Cycle
capture cross section and a considerably smaller fission cross section. This
produces 238
Np, followed by beta decay to 238
Pu [32]. Overall, 236
U is more
a weak poison rather than a fissile or fertile fuel.
N 238
Pu. With a half-life of 88 years, 238
Pu is useful as an energy source
for use in “radioisotope thermoelectric generators” (RTGs), in which heat
from the radionuclide is used to produce electricity. Its intermediate halflife
means that the output from 238
Pu changes little in a decade or two,
yet its specific activity is relatively high. RTGs based on 238
Pu were at
one time used in heart pacemakers, but, more recently, their main use has
been in the space program (e.g., in the Cassini mission to Saturn launched
in 1997).
N 239
Pu. There is continuous production of 239
Pu following neutron capture
in 238
U, and also continuous destruction of 239
Pu, primarily by fission
but also by neutron capture. The final 239
Pu abundance reflects the net
effect of production and destruction. The relatively low value of its isotopic
abundance (53%) is characteristic of “reactor-grade” plutonium (see
Section 17.4.1).
N 240
Pu. The capture cross section is relatively high in 239
Pu, so that neutron
absorption results in the production of 240
Pu as well as in fission. 240
Pu
acts as a fertile fuel for the production of 241
Pu, but the absorption cross
section for 240
Pu is less than that for 239
Pu, and 240
Pu is not consumed
as rapidly as is 239
Pu.
N Other Pu isotopes. Neutron capture on 240
Pu produces 241
Pu, which is
fissile but which also has an appreciable branch for neutron capture to
242
Pu. The buildup of plutonium isotopes stops here because 243
Pu, the
next in the series, decays with a half-life of 5.0 h to 243
Am and further
neutron captures and beta decays moves the chain to atomic numbers
higher than that of plutonium.
The plutonium isotopes through atomic mass number A = 242 have halflives
that are long compared to normal exposure periods in the reactor (the
shortest is 14.3 years for 241
Pu). Considering those with A > 238 as a group,
they are fed by neutron capture in 238
U (quickly followed by beta decay to
239
Np and 239
Pu) and are depleted primarily by the fission of 239
Pu and 241
Pu
and the beta decay of 243
Pu.
9.3.3 Energy from Consumption of Fuel
Energy per Unit Mass from Fission of 235
U
The fission of a 235
U nucleus corresponds on average to the release of 200 MeV
(3.20 × 10−11
J), including the associated contributions from neutron capture
and the decay of fission fragments (see Section 6.4.2). The number of 235
U
atoms per gram of uranium is wNA/M, where w is the fraction of 235
U in the
uranium (by weight), NA is Avogadro’s number, and M is the atomic mass
of 235
U. Thus, 1 kg of natural uranium (w = 0.00711) has 1.822 × 1022
nuclei
9.3 Fuel Utilization 211
Table 9.3. Energy per unit mass from fission of uranium, at different degrees of
enrichment.
Enrichment, w Energy per Unit Massa
Category (%) J/kg GWd/t Tonne/GWyr(e)b
Natural U 0.711 5.84 × 1011
6.8 169
Enriched U 3.2 2.63 × 1012
30.4 38
Enriched U 3.5 2.87 × 1012
33.3 34
Enriched U 3.75 3.08 × 1012
35.6 32
Enriched U 5.0 4.10 × 1012
47.5 24
Pure 235
U 100 8.21 × 1013
950 1.2
a
Thermal energy, assuming fission of all 235
U, a release of 200 MeV per 235
U fission,
and no fission of other nuclides.
b
Assuming a thermal conversion efficiency of 32%.
of 235
U and their complete fission would release 5.8 × 1011
J. The available
fission energy per kilogram of uranium, for different degrees of enrichment,
is given in Table 9.3 for a few illustrative cases, assuming fission of all 235
U
and ignoring losses due to capture in 235
U and gains from fission in 239
Pu and
other nuclides.
As indicated earlier, 1 GWyr of electric power corresponds to a thermal
output of 1141 GWd(t) or 9.86 × 1016
J. Assuming complete fission of all the
235
U and ignoring fission in plutonium, this corresponds to a fuel requirement
of 1.20 tonnes of 235
U per GWyr(e). For the specific case of a 3.75% enrichment
in 235
U (as in Table 9.2), the requirement translates to 32 tonnes of uranium.
The corresponding thermal burnup is 35.6 GWd/t, as indicated in Table 9.3.
Energy per Unit Mass of Fuel
The discussion in the preceding subsection is incomplete, because it omits
many crucial factors that significantly modify the amount of 235
U required by
a reactor. These include the following:
1. Not all of the 235
U is consumed in the reactor. For example, for the case
described in Table 9.2, the 235
U content per MTHM is 37.5 kg in the fresh
fuel and 8.6 kg in the spent fuel (i.e., a consumption of only 77% of the
235
U).
2. About 14% of the thermal neutron-absorption reactions in 235
U result in
capture rather than fission.
3. Fission in 239
Pu (and, to a lesser extent, in 241
Pu) provides a substantial
additional energy source. This reduces the 235
U required for a given energy
production.
4. There is a small contribution from fast-neutron fission in 238
U.
212 9 Nuclear Fuel Cycle
These effects all change the number of fission events, reducing the number
of fissions in 235
U and adding fission in plutonium isotopes and even 238
U.
The overall consequence can be crudely estimated by comparing the decrease
in the total heavy metal mass—which results almost entirely from fission of
uranium and plutonium isotopes—to the original 235
U mass. The ratio of
these quantities is 1.081, which would suggest that there are about 8.1% more
fission events than would be given by complete fission of 235
U. Applying this
8.1% “correction” raises the 235
U burnup of 35.6 GWd/t to a revised value
of 38.5 GWd/t. This estimate still ignores the differences in fission energy
yields and atomic masses between 235
U and the heavier actinides. Roughly
40% of the fissions are in 239
Pu and 241
Pu, not 235
U. Taking this into account
adds roughly 0.6 GWd/t. Gamma-ray emission following neutron capture in
the actinides adds further. Together, these approximate corrections bring the
total close to the burnup of 40 GWd/t indicated in Table 9.2.
A value of 40 GWd/t is a good representation of recent LWR performance,
although average PWR burnups are now higher and future ones are expected
to be still higher (see Table 9.1).24
The burnup in GWd/t can be translated
into the fuel requirement per year. For example, for a burnup of 40 GWd/t
the total uranium requirement is 28.5 tonnes per GWyr, or 1.07 tonnes of
235
U for an enrichment of 3.75%.
The precise uranium requirements for a given energy output depend on details
of the fuel cycle and reactor operation. Nonetheless, the following equivalence,
as found for the above example, is useful for approximate estimates of
the general magnitudes for a once-through LWR fuel cycle:
1 tonne of 235
U → 1 GWyr(e) (approximate).
9.3.4 Uranium Ore Requirement
The amount of uranium ore required to operate a reactor depends on the
burnup achieved, the initial enrichment, and the amount of 235
U lost in the
enrichment process. The mass MF of natural uranium used as feed input to the
enrichment facility is related to the mass MP of enriched uranium produced in
the fuel by the expression MF /MP = (wP − wT )/(wF − wT ), where wP , wT ,
and wF are the enrichments of the fuel, the tailings, and natural uranium,
respectively. For wP = 3.75%, wT = 0.2%, wF = 0.711%, and MP = 28.5
tonnes/GWyr, the natural uranium requirement is 198 tonnes/GWyr. Thus,
in round numbers, a once-through LWR fuel cycle requires about 200 tonnes
24
The average burnup is systematically less in BWRs than in PWRs because burnup
in the former is less uniform along the length of the fuel rods. The water at the
bottom of the BWR tank has a high density and is a better moderator than
the steam–water mixture at the top. This means that the maximum burnup is
achieved at the bottom of the rod, with a smaller burnup higher up on the rod.
The maximum acceptable burnups are about the same for the PWR and BWR,
but the PWR has a more uniform profile along the length of the fuel rod and
therefore a greater average burnup.
9.4 Back End of Fuel Cycle 213
of natural uranium per gigawatt-year. Present world demand for uranium is
60,000 tonnes/yr, corresponding to the requirements for the present annual
generation by nuclear power plants of about 300 GWyr [1, p. 17].
9.4 Back End of Fuel Cycle
9.4.1 Handling of Spent Fuel
Initial Handling of Reactor Fuel
Periodically, a portion of the fuel in the reactor is removed and replaced by
fresh fuel. In typical past practice, an average sample of fuel remained in the
reactor for 3 years, and approximately one-third of the fuel was removed each
year, with a shutdown time for refueling and maintenance of up to about 2
months. The trend is now to extend the interval between refueling operations
and to reduce the time for refueling.25
Currently, time intervals of 18 months
and shutdowns of 1 month are typical.
When the spent fuel is first removed from the reactor, the level of radioactivity
is very high, due to the accumulation of radioactive fission products
and radioactive nuclei formed by neutron capture. Each radioactive decay involves
the release of energy, which immediately appears as heat, so the fuel is
thermally hot as well as radioactively “hot.” Independent of the reprocessing
question, the first stage is the same, namely allowing the fuel to cool both
thermally and radioactively. The cooling of the fuel normally takes place in
water-filled cooling pools at the reactor site.
Originally, it was planned to keep the spent fuel at the reactor for roughly
150 days and then to transfer it to handling facilities at other locations. The
nature of the next step, in principle, depends on whether the fuel is to be
disposed of as waste or reprocessed. However, as yet, this “next step” has
been much delayed in the United States because no off-site facilities have
been developed. Instead, almost all of the fuel has remained at the reactor
sites—in many cases for more than 20 years.
In the absence of alternatives, some U.S. utilities are transferring older
fuel rods from cooling pools to air-cooled (dry storage) casks at the reactor
site. This may provide a workable temporary solution to the long delay in
implementing a national waste disposal program. However, it is only a stopgap
because the reactor operator cannot be counted on to be willing and able to
supervise the spent fuel for prolonged periods of time (see Section 11.1.3).
Disposal or Storage of Spent Fuel
For many years, it had been assumed that all U.S. civilian nuclear waste would
be reprocessed, but U.S. reprocessing plans have been abandoned. Instead,
25
An annual refueling shutdown of 2 months would mean a maximum capacity
factor of 83%, which is well below the present U.S. average.
214 9 Nuclear Fuel Cycle
official plans now call for disposing of the spent fuel directly, while retaining for
many decades the option of retrieving it. The fuel is to remain in solid form and
the fuel assemblies eventually placed in protective containers and ultimately
moved in secure casks to either a permanent or an interim repository site.
In the latter case, the waste would be moved to a permanent repository at a
later time.
A distinction is sometimes made been “disposal” and “storage.” The former
suggest permanence, whereas the latter suggests the possibility that the
spent fuel might be later retrieved. This possibility is made explicit in retrievable
storage systems, where the permanent sealing of the repository is
deferred, allowing the spent fuel to be recovered should this be desired at a
later time.26
In this case, the reprocessing option is not foreclosed, and the
spent fuel may ultimately not be a “waste.”
There are several motivations for maintaining retrievability: (a) It allows
for remedial action in case surprises are encountered in the first decades of
waste storage that require modifying the fuel package or the repository; (b)
it keeps open the option of recovering plutonium from the fuel; and (c) it
allows the recovery of other materials deemed useful—for example, fission
products for use in medical diagnosis and therapy or in the irradiation of food
or sewage sludge. When the placement becomes irreversible, with no prospect
of retrieving the fuel, this becomes final disposal.
9.4.2 Reprocessing
Extraction of Plutonium and Uranium
The alternative to disposing of the spent fuel is to reprocess it and extract at
least the uranium and plutonium. In reprocessing, the spent fuel is dissolved in
acid and the plutonium and uranium are chemically extracted into separate
streams, for use in new fuel. The most widely used method for this is the
suggestively named PUREX process.
Most early U.S. plans for reprocessing assumed that 99.5% of the U and
Pu would be removed. The remainder constitutes the high-level waste. In
the traditional plans, the wastes include almost all of the nonvolatile fission
products, 0.5% of the uranium and plutonium, and almost all of the minor
actinides [i.e., neptunium (Z = 93), americium (Z = 95), and curium (Z =
96)]. The uranium represents most of the mass of the spent fuel, but the fission
products contain most of the radioactivity.
Extraction can be more complete than contemplated in the original U.S.
thinking. The French program has exceeded the 99.5% goal, separating out
more than 99.9% of the uranium and 99.8% of the plutonium [33, p. 28].
There is no essential reason to limit extraction to plutonium and uranium,
26
Plans for the Yucca Mountain repository call for it to remain open for perhaps
as much as several hundred years, but the preservation of the reprocessing option
does not now appear to be the major motivating factor (see Section 12.2.1).
9.4 Back End of Fuel Cycle 215
although the former represents the valuable fuel and the latter represents
the bulk of the mass. It is possible to extract other radioisotopes as well,
either because they are deemed pernicious as components of the waste or
because they are useful in other applications. The minor actinides have been
of particular interest. They include long-lived products whose removal would
decrease the long-term activity in the waste. One option is to separate them
and return them to a reactor where they would be transmuted in neutron
reactions.
Of course, if the chief goal is safety, it is necessary to balance the benefits
from decreased activity in the wastes against the increased hazards of handling
and processing them when they are still very hot. At present, this further
separation option has not been adopted in the major reprocessing programs
in France and the United Kingdom, and the minor actinides remain with the
fission products [34, p. 149].
The residue of reprocessing constitutes the wastes. They are to be put in
solid form for eventual disposal. The standard method is to mix the high-level
waste with molten borosilicate glass and contain the solidified glass in metal
canisters. Although other solid waste forms have been suggested, borosilicate
glass has been used in the French nuclear program and had figured prominently
in the original U.S. plans for reprocessing commercial wastes. It is
being used for the sequestering of already reprocessed U.S military wastes at
the Savannah River site in South Carolina and is planned for the wastes at
the Hanford reservation in Washington state.
Status of Reprocessing Programs
Until the late 1970s, reprocessing had been planned as part of the U.S. nuclear
power program. A reprocessing facility at West Valley, New York was
in operation from 1966 to 1972, with a capacity of 300 MTHM/yr. This is
enough, roughly speaking, for the output of 10 large reactors. There were
plans for further facilities at Morris (Illinois) and Barnwell (South Carolina)
which would have substantially increased the reprocessing capacity. However,
all these plans have been abandoned.27
In part, the abandonment was impelled by technical difficulties. There had
been high radiation exposures of workers at West Valley and the plant was
shut down in 1972; plans to remodel and expand it were later aborted. When
the Morris plant was first tested with nonradioactive materials, it did not
perform reliably, and the General Electric Co., which was building the plant,
decided there were serious difficulties. The Barnwell plant moved ahead until
the early 1980s, but it faced problems of meeting increasingly strict standards
on permissible radioactive releases.
These difficulties might have been surmounted had there been a belief
that reprocessing was needed. However, the fundamental motivation for re-
27
For a discussion of this history, see Ref. [35].
216 9 Nuclear Fuel Cycle
processing began to slip away. There was no prospective near-term shortage
in uranium supply, and uranium prices were low enough to remove the economic
incentive for reprocessing. Further, an important body of opinion had
developed in the United States against reprocessing, on the grounds that it
might make plutonium too readily available for diversion into destructive de-
vices.
This view was expressed in the 1977 report Nuclear Power Issues and
Choices, sponsored by the Ford Foundation and authored by an influential
group of national science policy leaders. In its conclusions on reprocessing,
the report stated:
[T]he most severe risks from reprocessing and recycle are the increased
opportunities for the proliferation of national weapons capabilities and
the terrorist danger associated with plutonium in the fuel cycle.
In these circumstances, we believe that reprocessing should be deferred
indefinitely by the United States and no effort should be made to
subsidize the completion or operation of existing facilities. The United
States should work to reduce the cost and improve the availability of
alternatives to reprocessing worldwide and seek to restrain separation
and use of plutonium. [36, p. 333]
Consistent with this thinking, the Carter administration decided in 1977 to
“defer indefinitely the commercial reprocessing and recycling of the plutonium
produced in U.S. nuclear power programs” [37, p. 54]. Work on U.S. reprocessing
plants for commercial fuel was phased out, culminating in the closing
of Barnwell at the end of 1983 [35, p. 124].
Nonetheless, reprocessing has been pursued in other countries. The largest
reprocessing programs are in France and the United Kingdom, both of which
completed major expansions of reprocessing capacity in 1994 to handle both
domestic and foreign fuel. In addition, a large facility is being built in Japan.
France has the most fully developed fuel cycle. Although much of its present
reprocessing capacity is devoted to foreign orders, it also has a program of
reprocessing and plutonium recycle of domestic fuel.
Table 9.4 lists the reprocessing plants in operation or under construction.
It omits plants that were closed down before 2002, including plants in Belgium,
France, Germany, the United Kingdom, and the United States.
Use of Mixed-Oxide Fuel
The fuel manufactured from the output of the reprocessing phase is generally
a mixture of plutonium oxides and uranium oxides, with 3% to 7% PuO2 and
the remainder UO2. It is called a mixed-oxide fuel or MOX. At the higher
239
Pu enrichments, a burnable poison would be added to the fuel to reduce
its initial reactivity. Due to differences in the nuclear properties of 239
Pu and
235
U, most LWRs are limited to using only about a one-third fraction of MOX
9.4 Back End of Fuel Cycle 217
Table 9.4. Reprocessing plants for commercial nuclear fuel in operation or under
construction, 2002.
Year of Capacity
Country Location Start-up (MTHM/yr)a
In operation
France La Hague (UP2)a
1976 800
France La Hague (UP3) 1989 800
India Tarapur 1974 100
India Kalpakkam 1998 100
Japan Tokai-mura 1977 90
Russia Chelyabinsk 1984 400
United Kingdom Sellafield (B205) 1964 1500
United Kingdom Sellafield (Thorp) 1994 1200
Under construction
China Diwopu 2002 (?) 25–50
Japan Rokkasho-Mura 2005 800
a
UP2 was upgraded and redesignated as UP2-800, with full capacity reached in 1994
[40].
Sources: Refs. [1, p. 46]; [38, Table 4]; and [39, p. 119].
in the reactor core, with the remainder ordinary uranium-oxide fuel [41, 42].28
Some LWRs, however, have been designed to accommodate a full load of MOX
fuel.29
By 2001, about 20 PWRs in France (out of 58) were using MOX for onethird
of their fuel [34, p. 138]. In the United States, the interest in MOX
fuel has been motivated by the need to dispose of plutonium from dismantled
nuclear weapons (see Section 18.3.3). Toward this end, the DOE is planning
to build facilities for conversion of plutonium into MOX fuel at its Savannah
River site. At least one nuclear plant operator (Duke Energy) has made a
28
It is more difficult to control a thermal reactor using plutonium than one using
uranium. Contributing reasons include (a) the delayed neutron fraction, β, is
smaller for 239
Pu than for 235
U and (b) the fission cross section resonance in
239
Pu near 0.3 eV (see Figure 6.1) leads to a positive feedback if the reactor
temperature rises. In addition, with 239
Pu, the neutron and gamma-ray spectra
are more energetic than with 235
U, causing more radiation damage [42, p. 119].
As a result, it is necessary to have design changes, including more control rods, if
a full load of MOX fuel is used in place of uranium fuel. This cannot be readily
accomplished in most LWRs. However, it is possible in the so-called System-80
PWRs. Three such reactors are in operation at the Palo Verde nuclear plant in
Arizona, but, at present, no U.S. LWR is licensed by the NRC to operate with
MOX fuel.
29
See Section 18.3 for a further discussion of MOX fuel, in the context of the burning
of plutonium from dismantled nuclear weapons.
218 9 Nuclear Fuel Cycle
commitment to use MOX fuel in some of its reactors, beginning in 2007 if
plans proceed according to the initial schedule [43].
9.4.3 Alternative Reprocessing and Fuel Cycle Candidates
Advanced Aqueous Process
In the widely used PUREX process, the plutonium and uranium are extracted
and the fission products and minor actinides constitute the wastes. The advanced
aqueous process is a modification of the PUREX process in which the
minor actinides are recovered as well. Uranium is crystallized out at an early
stage to reduce the bulk of the material that must be dealt with in the further
chemical processing [44, p. 60]. The two product streams, one of uranium and
the other of plutonium and the minor actinides, are used to fabricate fuel for
use in either thermal or fast reactors.
The UREX Process
An alternative to the advanced aqueous process is the uranium extraction
process (UREX and UREX+). It differs in the means of separating out the
uranium. Several output streams are specifically identified in this process [45,
p. II-3]:
1. Uranium. The uranium is extracted in very pure form (“at purity levels
of 99.999 percent”). The leaves it free of highly radioactive contaminants
and makes it easy to handle for disposal or reuse in a reactor.
2. Plutonium and minor actinides. Neptunium, americium, and curium are
retained with the plutonium. These elements can be incorporated into the
reactor fuel.
3. Long-lived fission products. Long-lived fission products (in particular,
iodine-129 and technicium-99) are separately extracted, for destruction
in a reactor (see Section 11.3.3).
4. Other fission products. These become the wastes. The waste disposal problem
is simplified because the long-lived radionuclides have, for the most
part, been removed.
Pyroprocessing
The above-discussed chemical reprocessing processes are known as aqueous
processes. An alternative approach, under active exploration for use in conjunction
with future reactors, is the pyroprocess or electrorefining process. In
this method, the spent fuel is dissolved at very high temperatures in molten
cadmium, creating an “electrolytic bath.” Groups of chemical elements are
separately extracted on the basis of differences in the potentials at which they
dissolve and ionize. In particular, ions of the actinides, including uranium,
9.4 Back End of Fuel Cycle 219
plutonium, and the minor actinides, are attracted to cathodes and are extracted.
The actinides are then incorporated in the fabrication of new fuel
elements.
Full Actinide Recycle
A fuel cycle based on the nearly complete extraction of plutonium and minor
actinides (collectively, the transuranic elements) and their consumption by
fission in fast reactors has been sketched in the MIT report The Future of
Nuclear Power [46]. This cycle envisages a global nuclear economy in 2050
with a capacity of 1500 GWe based on a balanced combination of thermal
and fast reactors. The thermal reactors are assumed to be LWRs, fueled by
enriched uranium oxide. The fast reactors, undefined as to type, are fueled
by transuranics obtained from the LWR spent fuel. Pyroprocessing is used to
extract the transuranics from both the thermal and fast reactor spent fuel.
For each load of fresh fuel in the fast reactors, 20% of the transuranics are
consumed in fission and the remaining 80% are available for recycle.
A balanced system is one in which the spent fuel from the thermal reactors
provides the transuranics needed to make up for those consumed in the fast
reactors. With the assumptions made in the MIT analysis, this is achieved
by having slightly more capacity in the thermal reactors than in the fast
reactors (815 GWe and 685 GWe, respectively). A variety of choices exist for
the fast reactors, including some of the Generation IV reactors discussed in
Section 16.6.
The uranium requirement for the entire fuel cycle is the amount needed
to provide for the 815 GWe of LWRs, which is 54% of the amount needed
if LWRs accounted for the full 1500-GWe capacity. A further, and perhaps
even more important, benefit is the almost complete elimination of plutonium
and minor actinides from the stream of wastes that require permanent
disposal.
General Features of Reprocessing Options
Any fuel cycle that recycles the fissile components of the spent fuel (mainly the
remaining 235
U and the plutonium isotopes 239
Pu and 241
Pu), increases the
energy obtained from the existing uranium resources. If the minor actinides
are included with the uranium and plutonium in the new fuel, the wastes will
have much less long-term radioactivity than wastes in the once-through fuel
cycle. The mass of the spent fuel is greatly reduced if the uranium is either
returned to the reactor or is separated from other radionuclides to become
low-activity depleted uranium. The fission products then constitute the waste
product that requires long-term disposal. This greatly reduces the mass of the
waste product and the period during which it must be kept isolated from the
environment.
220 9 Nuclear Fuel Cycle
All of the reprocessing fuel cycles that have been described in this section
achieve these resource extension and waste reduction benefits. The PUREX
process accomplishes much of this, but in its standard form the minor actinides
are not removed.
A long-standing objection to reprocessing is based on the increased proliferation
risks if 239
Pu is in wide circulation. The above-described methods
lessen the risks in two ways: (1) The presence of the minor actinides increases
the activity of the fuel and makes it more difficult to handle, and (2) the reprocessing
and fuel fabrication facilities can be located adjacent to the reactor,
making theft or diversion of the fuel very difficult without the collaboration
of the plant operators. The collocation aspect was particularly stressed in
planning documents for the Integral Fast Reactor (see Section 16.5.1) and
combined facilities were tested on a small scale using fuel from the experimental
breeder reactor in Idaho (EBR-II).
A further objection is based on economics. Given present demand and
prices, it is more expensive to reprocess spent fuel than to obtain fuel from
newly mined uranium. This is a cogent objection at the present scale of nuclear
power use. The advantages of these reprocessing approaches become more
relevant in the context of a possible major expansion of nuclear power.
At present, the reprocessing approaches discussed here are in the development
and study stage, except for the long-used PUREX process. In general,
the pyroprocessing technique is more suitable for use with fuel in metallic form,
while the aqueous processes are more suitable for oxide fuels. Thus pyroprocessing
was originally studied for use with metallic fuel from a sodium-cooled
fast reactor. However, either class of process could be used with a wide variety
of fuel forms, given appropriate pretreatment stages.
9.4.4 Waste Disposal
All countries with announced plans for disposing of high-level radioactive
wastes are planning on eventual disposal in deep geologic repositories, typically
made by excavating caverns or holes in favorable environments. Many of
the plans for these permanent disposal facilities include a period during which
the waste could still be retrieved.
Deep geologic disposal has been the favored course in U.S. thinking since
the first attempts to formulate plans. There have been continuing efforts to
locate and design a suitable facility. A site at Yucca Mountain in Nevada was
selected in 1987 as the candidate for a U.S. repository and it has been under
intense study since. The DOE in 2002, with the subsequent concurrence
of the president and Congress, recommended going ahead with the Yucca
Mountain project. The announced goal is to have a facility ready to receive
wastes by 2010, subject to approval by the Nuclear Regulatory Commission.
(Much more extensive discussions of nuclear wastes are presented in Chapters
10–13.)