Chapter 10

ISOTOPIC SEPARATION AND ENRICHMENT

© M. Ragheb
3/22/2018

10.1 INTRODUCTION
There is a need in nuclear applications to separate the light isotopes of the elements such
as deuterium 1D2 from hydrogen, Li6 from lithium, B10 from boron as well as the heavy isotopes
such as U235 from natural uranium and Pu239 from a mixture of plutonium isotopes. Large
commercial enrichment plants are in operation in France, Germany, The Netherlands, UK, USA,
and Russia, with smaller plants in Japan, South Africa, Brazil, Argentina, Pakistan, Democratic
People Republic of Korea (DPRK) and Iran. Four major suppliers of enrichment services are:
the Unites States Department of Energy (USDOE), Eurodif in France, Urenco, a British, Dutch,
and German group, and Techsnabexport in Russia.
Canada and India, which both rely on heavy water nuclear power plants for electricity,
make the most heavy water. Other countries with heavy water production facilities include
Argentina, Iran, Romania, and Russia.
The world has used just three very inefficient nuclear enrichment processes: gaseous
diffusion, centrifugation, and Electro Magnetic Isotope Separation (EMIS.) EMIS enrichment is
an inefficient process: It would cost $81,000 to enrich a single pound of uranium. Gaseous
diffusion is not all that much better. This 1940s vintage energy intensive enrichment process
currently accounts for 25 percent of all enriched uranium and takes more than 1,400 stages to
enrich uranium to a useful level. The centrifuge technology is taking over the enrichment
industry. The United States Enrichment Company (USEC) is building the American Centrifuge
plant in Ohio. This project is facing constant delays and budget overruns and is expected to
reach full production in 2012. There is a new enrichment facility under construction in New
Mexico by the European firm Urenco, which will not be running at full capacity until 2017.
The separation of the light isotopes occurs in nature in closed bodies of water such as the
Great Salt Lake in Utah, USA, and the Dead Sea in the Middle East. Evaporation of water leads
into an enrichment into heavy water: D2O through the process of fractional distillation, because it
has a higher boiling point than ordinary water H2O. Electrolysis can also be used to produce
heavy water since the dissociation of light water proceeds faster than that for heavy water
leading to its enrichment in the electrolytic solution. Heavy water is currently used as coolant
and moderator for a reactor design using natural uranium as a fuel: the Heavy Water Reactor
(HWR) design.
The need for separating the two isotopes of uranium: U235 and U238 arose initially as part
of the Manhattan project for the manufacture of the atomic bomb. Later, it became the basis for
using light water as a coolant and moderator for propulsion naval reactors as well as land-based
reactors. In propulsion reactors there was a need to increase the enrichment from the natural
abundance of 0.72 percent in U235 to 97.3 percent as a way of providing enough reactivity to
overcome the xenon poisoning and the reactor dead time arising in these systems, increasing the
time between refueling to about ten years, and even for the whole operational time of the core.
For nuclear warheads, the enrichment is increased to 93.5 percent. For land-based reactors, the
U235 is raised to the 3-5 percent range to achieve criticality with light water. Interestingly, this is
the range of enrichment that occurred earlier in the history of the Earth at the natural reactors at
the Oklo uranium mine site. For space reactors an enrichment of 97.3 percent is envisioned in
view of reducing the launch weight.
Isotopic separation is also encountered in the separation of the Li6 isotope from lithium
for thermonuclear weapons and future peaceful fusion applications. For the reactors using
natural uranium as a fuel, separating heavy water from ordinary water becomes a replacement for
the enrichment of uranium. However, it is easier to isotopically separate the light elements
compared with the heavy ones.
The most important separation methods for the heavy elements as well as the light
elements are described.

10.2 STABLE ISOTOPES APPLICATIONS

The stable isotopes produced by isotopic separation are used worldwide in both medical
and industrial applications.
Three segments of the medical field are served:

1. DIAGNOSTICS AND IMAGING

Nuclear diagnostic imaging has an important role in the identification and management of
conditions such as heart disease, brain disorder, lung and kidney functions and a broad range of
cancers. The high sensitivity and specificity of nuclear diagnostic imaging techniques offer the
important advantages of being able to identify diseases at an early stage, to track disease
progression, to allow for accurate disease staging and to provide predictive information about
likely success of alternative therapy options.
The two most important techniques are Gamma Imaging and Positron Emission
Tomography (PET).

Gamma Imaging

There are some 8,500 nuclear medicine departments in the world using gamma cameras
to detect diseases of various organs including heart, brain, bone, lung and the thyroid. A total of
some 20,000 gamma cameras are in use. Gamma imaging uses the stable precursors for
Gallium67 such as Zinc68 and Zinc67, Indium111 (Cadmium112) and Iodine123 (Xenon124).

Positron Emission Tomography (PET)

There are about 200 PET centers in the world operating a total of some 300 PET cameras.
They are used mainly for the diagnosis and staging of cancer. Use of PET is growing as a result
of the recognition of clinical benefits from PET.

2. RADIATION THERAPY
 Brachytherapy

Brachytherapy is the procedure of using temporary irradiation very close to the area of
disease, in particular cancer and stenosis. Iridium191 is produced for Iridium192 sources used in
remotely controlled after loaders. A modern development pursued by many companies is Seeds
Implantation: implanting radioactive sources ("seeds") in tumors, in particular prostate cancer.
This minimizes the side effects known from surgery and/or external radiation. More than
300,000 new cases of prostate cancer are diagnosed every year in the USA and the EU alone, and
patients are increasingly being treated with seeds implantation. The radioactive source most
often used in the seeds is Iodine125 (made from Xenon124).
Restenosis is the re-narrowing of a coronary artery after angioplasty treatment. This
happens in about 30 percent of the patients within a few months after treatment, a percentage
which can be dramatically reduced by restenosis treatment: the irradiation through brachytherapy
of the area unblocked by angioplasty treatment. The treatment is yet far from standard
procedure. Many tests are being conducted, including tests to determine what radioactive
isotopes and thus what precursors are most suitable such as Tungsten186, which is a precursor for
Rhenium188, a possible candidate for restenosis treatment.
A new development is to deplete Titanium46. Natural Titanium is a suitable
biocompatible material, which is being used to encapsulate radioactive seeds for brachytherapy,
in particular for prostate cancer. Instead of the complicated process of sealing radioactive
sources into Titanium capsules, it would be much easier to seal the stable precursors into
Titanium capsules before irradiating the precursors. However, this would make it necessary to
deplete the Titanium46 isotope in order to eliminate the adverse radioactivity of activated
Titanium46 (Scandium46).

3. PAIN RELIEF

Palliative care of pain arising from secondary metastasis derived from spread of breast,
prostate and lung cancers is under development. Growth is rather slow due to the problem of
acceptability and the conservatism amongst doctors preferring traditional medicines like
morphine.

10.3 INDUSTRIAL ISOTOPES

DEPLETED Zinc64 (DZO)

Depleted zinc64 is used in the nuclear industry. The addition of natural zinc to nuclear
reactor cooling water inhibits corrosion and the subsequent generation of radioactive Cobalt60,
thus acting to reduce worker radiation exposure. Natural zinc contains 48 percent zinc64 and as
the cooling water is subjected to continual neutron bombardment this isotope is activated to
radioactive zinc65. This isotope is a strong radiation emitter with a long half life and thus
contributes greatly to the storage time and hazard of waste cooling water.
The use of depleted zinc64 compounds enable the maximum benefits of zinc injection to
be reaped without the attendant radio nuclides consequences.
 Nuclear plants use Depleted Zinc Oxide (DZO) pellets or powder for BWR's, and
Depleted Zinc Acetate (DZA) for PWR's. Using centrifuge technology concentrations can be
enriched to exceed 99 percent or depleted below 1 percent.
The USA’s EPRI (Electric Power Research Institute) recommends the use of depleted
zinc. The added zinc reduces the amount of radioactive cobalt60 formed because of the
irradiation of natural cobalt in the construction materials of the reactor. 60Co is a major
contributor to radiation build up in the cooling systems and therefore causes elevated dose rates
to personnel.

TRANSMUTATION OF LONG-LIVED RADIO-ISOTOPES IN SPENT

NUCLEAR FUEL

Depleted molybdenum95 is used in an experimental process of transmutation of the long

lived radio isotopes, which are present in spent nuclear fuel. Americium in the form of AmO2 is
embedded in a matrix of molybdenum metal, processed to cermet pellets and irradiated in a
nuclear reactor. The americium is then transmuted into other, much shorter living radio isotopes.
The molybdenum has to be depleted in the isotope Mo95 in order to reduce its radiative capture
cross section for neutrons.

NON-DESTRUCTIVE TESTING (NDT) SOURCES

NDT using gamma cameras is an important and growing application of radioactivity.

The majority of the cameras use an iridium192 sources, although selenium75 sources are a
promising growth sector.
Iridium192 and selenium75 use the stable precursors iridium191 and selenium74
respectively.

SEMICONDUCTOR APPLICATIONS

Increased miniaturization of chips might require ultra pure silicon, germanium and
gallium material to build new generation semiconductors. Research in this field is ongoing.

10.4 CHEMICAL EXCHANGE

This technique makes use of different chemical reaction rates between isotopes. It works
best for light elements where the reaction rate differences are large. Practical plants use
reactions that allow the two reactants to be in different phases such as gas/liquid, solid/liquid,
and immiscible liquids. This permits convenient separation of the enriched and depleted
materials, and allows continuous countercurrent operation. By also including temperature
differences between the phases, the separation factor can be considerably enhanced. This is the
most important process today for producing heavy water, for which it is by far the most energy-
efficient method. Chemical exchange techniques for uranium have also been developed by Japan
and France, but have not been used for production work. Chemical exchange is also used for Li6
enrichment.
 HEAVY WATER, HDO AND D2O

Heavy water HDO and D2O, differs from the widely abundant light water H2O in the
presence of the deuteron nuclide D, rather than the hydrogen nuclide H, in the water molecule.
Ordinary water is a mixture of its heavy and light forms. Heavy water has a density of 1.1
[gm/cm3], compared with 1.0 [gm/cm3] for light water at room temperature, which explain its
name. In nature, deuterium occurs with an abundance that varies between 130-160 parts per
million or ppm in hydrogen, with an average of 0.015 in atomic percent or 150 ppm. The ratio
varies according to location, seasonal temperatures and meteorological factors. Because of its
rarity, it occurs in the form of HDO, or the mono-deuterated molecule.
Heavy water is used as coolant and neutron moderator in nuclear reactors using natural
uranium as fuel, such as the Canadian Uranium Deuterium CANDU type (Fig. 1).

Figure 1. Darlington 935 MWe CANDU Heavy Water Reactor (top), and reactor complex at
Ontario, Canada.

One cubic meter or one metric tonne is needed per MWe of installed capacity in the
HWR design. Its excellent moderating properties, high scattering cross section for neutrons, and
low absorption cross section makes it a unique moderator in this case. Heavy water must be
replaced in these systems because of leakage, in the form of steam, from the pressurized water
circuits. Building volumes are sealed around the coolant circuit, and the D2O vapor, which leaks
into them, is condensed and extracted for both economic reasons, and to avoid leakage of the
tritium gas.
Tritium, an isotope of hydrogen that is beta radioactive with a half-life of 12.34 years, is
produced from deuterium through neutron capture in the reaction:

1 D 2 + 0 n1  T 3 +
1 (1)

where  is a gamma photon.

Tritium can be extracted for radioactive isotopes applications such as activation of Liquid
Crystal Displays (LCDs), and future use in fusion reactors or thermonuclear weapons through the
DT fusion reaction:

1 D 2 + 1T 3  0 n1 + 2 He 4 + 17.6 MeV. (2)

 In future fusion reactors, deuterium can fuse with other deuterium nuclei producing
energy in the process through the reaction:

1 D2 + 1 D2  0 n1 + 2 He 4 , (neutron branch) (3)

D + 1 D2  1 H1 + 1T 3 ,
2

1 (proton branch) (4)

This reaction, which branches with equal probability, is the ultimate use of hydrogen as
deuterium in the oceans, and constitutes a practically unlimited future supply of fusion energy on
Earth.

Figure 2. Heavy water ice sinks in light water because of its higher density. Light water ice
would float because of its lower density. Source: Nova.

ELECTROLYSIS

To separate heavy water from light water, electrolysis, distillation of water or of liquid
hydrogen and a chemical exchange method, have been used commercially. The concentration of
heavy water in ordinary water is about 0.01 percent. Its concentration must be increased to 99.76
percent, which is essentially pure heavy water.
A multistage electrolysis process can concentrate heavy water, since light water can be
dissociated preferentially by an electric current gradually concentrating the heavy water. To
make the water conduct electricity potash lye, which is actually potassium hydroxide is added
which makes the water very caustic with a pH value of 14 compared with the neutral value of 7.
Germany used this process in the Second World War in an aborted attempt under the
leadership of their leading scientist Werner Heizenberg at building a heavy water nuclear reactor.
The process used abundant electrical power in a plant at Vemork, Norway, operated by Norsk
Hydro.
The Norsk Hydro company used hydroelectric power to manufacture ammonia, the basis
of fertilizer and explosives, both vital to the German war effort. To make chemical fertilizer, the
company needed hydrogen, and hydrogen could be made from water electrolysis.
The Vemork plant is shown in Fig. 3. Norwegian commandos blew up the plant, but the
Germans had it working again within three months. Then the USA Eighth Air Force bombed
Vemork, but the heavy water plant survived unscathed. Eventually, in 1944, members of the
resistance learned that the entire production plant and 15 tons of partially purified heavy water
were to be shipped to Germany. A commando attack destroyed the plant and sunk the Hydro
ferry shipment of heavy water destined for Germany at the cost of 14 local Norwegian and 4
German victims in the 400 meters deep icy waters of Lake Tinn. Some half filled barrels floated
and eventually found their way to Germany. The Germans needed about five tonnes of heavy
water for a single reactor; when at least ten reactors would have been needed for weapon effort.
but only half a ton of very diluted heavy water was shipped. The heavy water was destined for a
civilian experimental program, not a large scale military one. This explains why it was not
heavily guarded. Nevertheless, the raid slowed down and eventually stopped any German effort
at building any self-sustained fission chain reaction, even though a subcritical system using
natural uranium and heavy water was eventually built. The German nuclear bomb had inspired
much fear but was no more than a mirage. There was no German equivalent of the massive USA
Manhattan Project.
In nature, successive evaporation leads to concentration of heavy water in isolated water
bodies such as the Dead Sea in the Middle East and the Great Salt Lake in Utah, USA. The
concentration process is due to the lower boiling point of light water relative to Heavy Water.
Water containing normal hydrogen is more easily disassociated into hydrogen and
oxygen gases by an electric current than water containing deuterium. This allows the isotopes to
be separated. In the USA, at the Savannah River Site (SRS) a now dismantled heavy water plant
used the hydrogen sulfide-water exchange process to partially enrich heavy water. Deuterium
was further concentrated by fractional distillation, and then by electrolysis. The moderator
rework unit at SRS used fractional distillation to re-enrich reactor moderator that had become
depleted in deuterium.
 Figure 3. Norsk Hydro heavy water electrolysis production plant at Vemork, Norway.

Figure 4. Norsk Hydro 18 electrolysis cells at Vemork. Source: Nova.

FRACTIONAL DISTILLATION

Water molecules containing deuterium atoms vaporize at a higher temperature than those
without deuterium, so the boiling point of heavy water is slightly higher than that of normal
water. Water vapor above a mixture of normal and heavy water will be slightly depleted in
deuterium as a result, while the liquid will be slightly enriched. Enrichment results from
successively boiling off and removing vapor containing normal hydrogen.
 Figure 5. Savannah River heavy water plant’s 130 foot tall distillation towers.

Industrially, a distillation process called the GS Dual-temperature Hydrogen Sulfide

exchange method is used to produce heavy water. K. Geib and J. S. Spevack invented the
process. However GS here stands for Girdler-Sulfide, where Girdler is the name of the
company, which designed the first such plant in the USA. It uses a process whose rate depends
on the different masses of the molecules involved.

Figure 6. Production of Heavy Water using distillation by the Hydrogen sulfide process.
 Since they possess the same number of electrons, the hydrogen and the deuterium
isotopes are thought to be chemically identical. However, they do undergo chemical reactions at
different rates according to their atomic masses.
The difference in the reaction rates increases with the difference between the isotopic
masses. This is most pronounced in the case of hydrogen and deuterium, since the latter has
twice the mass of the former. Deuterium is exchanged between streams of water and hydrogen
sulfide gas, as shown in Fig. 6. An Ontario, Canada heavy water production plant is shown in
Fig. 7. The Arak, Iran heavy water plant is shown under construction in Figs. 8, 9.

Figure 7. Ontario, Canada, Heavy Water production plant.

Figure 8. Arak, Iran heavy water plant under construction.

 Figure 9. Arak, Iran heavy water production plant showing distillation towers, February 2005.

The water trickles through a series of perforated plates in a tower, while the gas flows
upward through the perforations.
In this case a reversible reaction occurs:

H 2 O +HDS  HDO + H 2S (5)

The equilibrium constant K for this reaction in terms of the concentrations, can be written
as:

[HDO]
[HDO][H 2S] [H 2 O]
K=  (6)
[H 2 O][HDS] [HDS]
[[H 2S]

If H and D were the same chemically, the equilibrium constant for the reaction would be
equal to unity. In fact K is not equal to unity, and is furthermore temperature dependent:

K = 2.37 at 25 oC,
K = 1.84 at 128 oC.

It becomes clear from the above equation that the concentration of HDO in H2O is greater
than the concentration of HDS in H2S. Furthermore, the relative concentration of HDS in H2S
increases with increasing temperature. This makes it possible to separate D from H.
At 32 0C, equilibrium favors the concentration of deuterium in water. However, at
around 128 0C, the equilibrium favors the concentration of deuterium in the hydrogen sulfide.
The separation tower is divided into an upper cold section, where the concentration of deuterium
in water is increased. This is then used as feed for the lower hot section, where the exchange
leads to a further enrichment, this time in the hydrogen sulfide stream. The enriched gas from
the hot section is fed to a second stage for further concentration. In the third stage, deuterium
from the return stream of water is transferred to hydrogen sulfide in a cold tower, and enriched
gas is transferred to water in the hot tower. Only a few stages are needed to produce heavy water
at about 20 to 30 w/o D2O. Water is then fed to a vacuum distillation system to almost pure
deuterium at a concentration of 99.75 percent, as needed by the CANDU reactor concept.
At the Bruce plant in Canada, the first stage towers are large structures. The first stage
consists of 3 towers in parallel with a height of 90 meters and a diameter of 9 meters. Large
amounts of power are needed: 300 MWth to produce 400 metric tonnes of heavy water per year.
Other possible processes would involve exchange between hydrogen and ammonia as
gases, or amino methane and hydrogen gas. These processes would reduce the power
requirements, with smaller size plants, and avoid corrosion problems typical of the hydrogen
sulfide process. Exchange between water and hydrogen gas is possible if a suitable catalyst can
be identified.
A common feature of both heavy water and light water reactors is the radiolysis or the
decomposition of water under the effect of gamma radiation into hydrogen ions and free
hydroxyl radicals:

    + D 2 O  D + OD (7)

these can recombine back into heavy water in the first way as:

D + OD  D 2 O (8)

or in another way in pairs into deuterium gas and deuterium peroxide, when the specific
ionization is high, as:

D+D  D2 (9)

OD + OD  D 2 O 2 (10)

The deuterium peroxide would then decompose leading to the release of substantial
amounts of deuterium gas D2, and oxygen O2. Arrangements must be made for the removal of
the deuterium gas since it constitutes an explosion hazard in the presence of oxygen. Oxygen
itself must also be removed since it constitutes a corrosion hazard, by injection of hydrogen gas
into the coolant to reduce the amount of oxygen, in both heavy water and light water cooled
reactors.
Deuterium can also be produced using distillation from liquid hydrogen.

10.5 LITHIUM ENRICHMENT

Deuterium separated from heavy water is combined with enriched Li6 to make ceramic-
like lithium6 deuteride, Li6D parts for the secondary stages of thermonuclear weapons. A
mixture of deuterium and tritium gases is injected into the pit of the primary fission stage of
modern USA thermo nuclear weapons to boost the nuclear explosive yield by the effect of the
14.06 MeV neutrons from the fusion DT reaction.
The USA, UK, France, China and Israel and believed of having facilities for Li6
separation in quantities sufficient for the manufacture and maintenance of boosted fission and
thermonuclear weapons. Russia exploded a device using lithium deuteride powder Li6D before
the USA did, even though the so-called sloika or layer cake design was not a true thermonuclear
device scalable to any desired yield. The USA is reported to have stopped stockpiling Li6 around
1963.

Figure 10. The Y-12 plant at Oak Ridge, Tennessee.

The Oak Ridge Y-12 Plant, which also included the magnetic separation process of U235
using Calutrons, began the initial effort to develop lithium isotope separation processes in 1950.
Three processes were explored: the COLEX, ELEX, and OREX processes.
 Figure 11. Lithium Column Exchange, COLEX process enrichment equipment. Isotopes of
lithium are partially separated when transferring between an aqueous solution of lithium
hydroxide and a lithium-mercury amalgam. The isotope Li6 has a greater affinity for mercury
than does the isotope Li7. A lithium and mercury amalgam is first prepared using natural
lithium. The amalgam is then agitated in a natural lithium hydroxide solution. The rare Li6
isotope concentrates in the amalgam and the more common Li7 isotope migrates to the hydroxide
solution. A counter-current flow of amalgam and hydroxide passes through a cascade of stages
until the desired enrichment in Li6 is reached.

Figure 12. Lithium solution tank used in the COLEX process. The Li6 can be separated from the
amalgam and the tailings fraction of Li7 is electrolyzed from the aqueous solution of lithium
hydroxide for further use.
 Figure 13. Mercury flasks used in the Li6 enrichment process.

The first successful laboratory separation was achieved with the ELEX process-an
electrically driven chemical exchange process similar to that used in chlor alkali plants for the
manufacture of chlorine gas and sodium hydroxide. The ELEX pilot plant was built at Y-12 in
1951. This plant was cleaned out and dismantled by 1959.
The OREX pilot plant was built by the Chemical Technology Division of Oak Ridge
National Laboratory (ORNL) in 1952 and was operated in a sort of inter-plant competition with
the Y-12 COLEX process to determine the better process. Both plants were operated by the
Union Carbide Nuclear Corporation.
In the OREX process, an organic solution of lithium was exchanged with a solution of
lithium in mercury or an amalgam in pulse columns using Propylene-di-amine (PDA) as the
organic phase. The COLEX process won the competition, and the OREX pilot plant was
subsequently dismantled between 1957 and 1959.
As a side note, an amalgam of mercury and silver was used for dental filling of decayed
teeth. Due to the toxicity of the leaching of mercury, possibly leading to immune system
responses, this approach has been abandoned in favor of ceramic fillings and existing amalgam
fillings are being regularly replaced by the dentists at the request of their patients.
The COLEX process, an acronym for Column Exchange, is based on the fact that
isotopes of lithium are partially separated when transferring between an aqueous solution of
lithium hydroxide and a lithium-mercury amalgam. The isotope Li6 has a greater affinity for
mercury than does the isotope Li7. A lithium and mercury amalgam is first prepared using
natural lithium. The amalgam is then agitated in a natural lithium hydroxide solution. The rare
Li6 isotope concentrates in the amalgam and the more common Li7 isotope migrates to the
hydroxide solution. A counter current flow of amalgam and hydroxide passes through a cascade
of stages until the desired enrichment in Li6 is reached. The Li6 can be separated from the
amalgam and the tailings fraction of Li7 is electrolyzed from the aqueous solution of lithium
hydroxide for further use. The mercury is recovered and is mixed with a fresh supply for reuse.
Some anecdotal accounts are that the lithium medication on pharmacies shelves used for
the treatment of schizophrenia is primarily of the purified Li7 isotope.
 A hoax may have been perpetrated about what was for a while called “red mercury”
during the tenure of President Boris Yeltsin in Russia. Red mercury could have been the lithium
mercury amalgam. It was attributed some unfounded capability of building so called
“ballotechnic” weapons, possibly pure fusion weapons. A ballotechnic would refer to an
explosive material that, when ignited, would generate enough energy, temperatures and pressures
without exploding to compress a DT fusion charge without the need for a primary fission device.
Claims were made that it is the Hg2Sb2O7 compound.
The material was offered for sale to black market operatives attracted to its purported
characteristics. Some who fell to the hoax were bilked out of their money, including some
governments as well as the USA government who was curious about the claims and purchased a
sample for analysis in government laboratories. The other face of the hoax, it is reported, was
that it was a sting operation by the Russian authorities used to catch nuclear materials black
market operatives and dispatch them to linger in Siberian prisons.
The COLEX process supplied most of the enriched lithium needed for the USA weapons
complex. The Atomic Energy Commission, AEC built two large COLEX facilities, called
Alpha-4 and Alpha-5 at the Y-12 Plant. Alpha-4 operated from January 1955 until 1963. The
unit was placed on standby until it was dismantled in the late 1980s. Alpha-5 began operating in
1955. It was shut down in 1959 and restarted in 1963 for a six-month campaign. The Y-12 Plant
engineers dismantled and disposed of the Alpha-5 COLEX process equipment in 1965 and 1966.
Site contractors operated an open-air mercury receiving operation, where mercury flasks were
emptied into a pipe leading to the COLEX plants. They used a furnace in a shed to roast sludges,
wastes and other materials for mercury recovery.
Lithium enrichment has created a considerable amount of materials in inventory. The
Department of Energy, DOE stores the lithium enrichment tailings, depleted in the Li6 isotope, at
the Portsmouth, Ohio enrichment plant and the K-25 Site at Oak Ridge, Tennessee. K-25 also
stores a stockpile of unprocessed lithium. The Y-12 and K-25 sites both store the DOE's
stockpile of enriched lithium.
The COLEX process employed approximately 24 million pounds of mercury. Most of
the mercury used in the COLEX and ELEX processes was returned to the General Services
Administration, GSA once it was no longer needed. However, a great deal of mercury was lost
in wastes, spills, and through evaporation. A mercury-nitric acid purification system utilized in
the COLEX process between 1955 and 1960 was the source of the major mercury bearing waste
stream at Y-12. This system discharged a diluted, neutralized acid waste containing mercuric
nitrate to the East Fork Poplar Creek. Mercury vapor from the plant was exhausted to the
environment by the building ventilation system. Mercury from spills also contaminated
basement sumps which were pumped through three concrete sedimentation tanks into the storm
sewer and from there were pumped directly into the East Fork Poplar Creek. The DOE believes
that small amounts of residual mercury are still present in the Y-12 plant sewers. Inorganic
mercury compounds of the type released at Y-12 plant were not initially believed to be toxic
unless inhaled. It was not until 1970 that scientists discovered the biological methylation of
inorganic mercury in the environment, which raised concerns over mercury discharges to surface
water.
Approximately two million pounds of mercury used in the lithium enrichment processes
have not been accounted for. Approximately 730,000 pounds or about 4,000 gallons of this
material is believed to have been lost in waste streams, evaporation, and spills. A study done in
1983 estimated that evaporation during maintenance operations, seepage from pumps and other
equipment, the venting of mercury vapors, and the smelting of mercury-contaminated scrap
released 51,300 pounds of mercury into the air. The COLEX process discharged 239,000
pounds of mercury to the East Fork Poplar Creek in the process waste stream, some of which
exists in sediments at the bottom of the New Hope Pond. The DOE believes that these waste
discharges are also the source of some of the mercury contamination in Watts Barr Lake, Poplar
Creek and the Clinch River. However, these bodies of water are also downstream from a
commercial chlor-alkali plant. Residual mercury contamination at Y-12 includes sludges and
mercury residue in building sewers and drain systems. A 1983 study also estimated that
approximately 425,000 pounds of mercury were lost to the soil in eight accidental spills at the Y-
12 Plant.

10.6 BORON ISOTOPE SEPARATION

Boron10 is a powerful neutron absorber with many uses such as control rods, neutron
detectors, boron cancer therapy and autocatalytic nuclear devices applications. The B10
production process uses a dimethyl ether-boron trifluoride complex. The complex is fed into a
distillation system. When the complex is boiled, part of the vapor phase breaks down into boron
trifluoride and dimethyl ether. Boron trifluoride vapor molecules containing lighter B10 atoms
recombine into the liquid phase more rapidly than molecules containing the heavier B11 isotope.
As a result, the heavier isotope is concentrated in the vapor phase and the lighter isotope in the
liquid phase.
To supply B10, the AEC built a plant in Model City, New York, near Niagara Falls. The
plant operated from September 1954 until 1958, when the AEC placed it on standby. The Model
City plant was rehabilitated in mid 1964 and restarted. First, the restarted plant converted the
remaining inventory of B10 from potassium fluoborate, KBF4 to elemental boron to meet
immediate weapon and reactor program demands. The plant continued to produce B10, until it
was placed on standby again in March 1971. Since that time, the USA government has relied on
commercial nuclear industry suppliers to convert its inventory of enriched B10 to a powder form,
and to supply additional B10.

10.7 FISSILE AND FISSIONABLE FUELS

Nuclides that fission upon capture of a neutron in such a way that the added neutron itself
is sufficient to overcome the fission barrier, even if it had zero kinetic energy, and induce fission
are designated as fissile nuclei. Examples of fissile nuclei are:

U233, U235, Np237, Pu239 and Pu241.

These are nuclei of even atomic number Z and odd mass number A.
Nuclides that fission with neutrons possessing a significant amount of kinetic energy or
fast neutrons, are designated as fissionable nuclei. Examples of these even A isotopes are:

U232, U234, U236, U238, Th232,

as well as the fissile nuclei such as U235.
It is possible to "breed" fissile nuclei from otherwise fissionable nuclei though the
process of neutron capture. In this case a nucleus that is incapable of fissioning with "slow"
neutrons can instead be transmuted into one that can.
Breeding can be achieved with fast neutrons with the fissionable isotope U238 to produce
the fissile isotope Pu239 through the reactions:

0 n1 + 92 U 238  92 U 239 + 
92 U 239  93 Np 239 + -1 e0 +  * +
93 Np 239  94 Pu 239 + -1 e0 +  * + (11)
______________________________________
0 n1 + 92 U 238  94 Pu 239 +2 -1 e0 +2  * + 3

Breeding can also be achieved with slow neutrons with the fissionable Th232 isotope to
produce the fissile U233 isotope through the reactions:

0 n1 + 90Th 232  90Th 233 + 

90 Th 233  91 Pa 233 + -1 e0 +  *+
91 Pa 233  92 U 233 + -1 e0 + * + (12)
__________________________________
0 n1 + 90Th 232  92 U 233 + 2 -1 e0 + 2 *+ 3

10.8 URANIUM MINING

Uranium is mined in different locations in the USA and Canada in open pit or
underground mines. Two particularly rich mining districts exist in New Mexico and in
Colorado. Underground mining poses a health hazard to the miners in the process of inhalation
of the radon222 gas, and its solid daughters in mine dust. Radon222 is itself a daughter nuclide in
the uranium238 decay chain. Adequate mine ventilation is necessary to ensure the safety of the
mining process.
The Grants Mineral Belt uranium was discovered at Poison Canyon by sheepherder
Paddy Martinez in 1951. His sheep were dying and he thought it was caused by radioactivity in
uranium. The fact is that the area had abundant selenium-bearing loco weed in the area, hence
the name: Poison Canyon. Selenium occurs in conjunction of uranium, and loco weed absorbed
it, and becomes a marker for the location of uranium ore.
 Figure 14. Haysrack Mountain in the uranium mining district of New Mexico.

Figure 15. McArthur River Cameco underground uranium mine in Northern Saskatchewan,
Canada.
Figure 16. Jackpile open pit uranium mine, Grants, New Mexico. The Grants Mineral Belt is the
world’s second largest historic uranium producer despite production cessation after the collapse
in uranium prices and the collapse of the USA’s uranium industry in the 1980’s following the
1979 Three Mile Island accident.

Figure 17. Rossing open pit mine, Namibia.

Figure 18. Underground uranium mine, Grants, New Mexico.

 Figure 19. Radon gas ventilation shaft at the Ambrosia Lake uranium mine, Grants, New
Mexico.

Figure 20. Reclaimed open pit and underground Poison Canyon uranium mine site, Ambrosia
Lake area north of Grants, New Mexico. Host rock is sandstone from the late Jurassic Morrison
Formation, 140-150 million years before present. Source: Michael S. Fulp.

Figure 21. Uranium mineralization in petrified tree logs exposed by a flash flood is sought by
geologists and prospectors to locate uranium ore bodies. Petrified log at Poison Canyon,
Ambrosia Lake, New Mexico. The yellow-green secondary uranium minerals appear on the
weathered surface. Source: Michael S. Fulp.

10.9 URANIUM CONVERSION AND REFINING

Uranium leaves the mine as the concentrate of a stable oxide known as U3O8 or as a
peroxide. A typical commercial ore contains approximately 0.3 percent in U3O8. The uranium
undergoes a milling and extraction process solvent extraction leaching process concentrating
U3O8 which is now designated as the yellow cake with a 65 to 85 percent content of U3O8.
The yellow cake is dissolved in nitric acid yielding a uranyl nitrate UO2(NO3)2 liquid
solution. This is purified with a selective solvent extraction process. The purified solution is
treated with ammonium hydroxide yielding a precipitate of ammonium diuranate (NH4)2U2O7.
The different forms of these compounds are shown in Fig. 24.
Upon calcinations and reduction with hydrogen, the UO2 brown oxide is obtained. Ingots
of the uranium metal can now be obtained by reduction with magnesium or calcium in electrical
induction furnaces (Fig. 25). These can be alloyed with different elements, extruded or
machined for use as a natural uranium metallic reactor fuel (Fig. 26).
For the production of enriched fuel, hydro fluorination with hydrofluoric acid HF of the
brown oxide yields uranium tetrafluoride UF4 powder. In a fluorination step treatment in a
fluidized bed with fluorine gas F2, one obtains uranium hexafluoride UF6 as a feed to the
enrichment process.

Figure 22. The uranium mining and conversion cycle.

 Yellow Cake
65 – 85 %
U3O8

Dissolution: Nitric Acid

Impure Uranyl
Nitrate Solution
UO2(NO3)2

Selective solvent extraction

Purified Uranyl
Nitrate Solution
UO2(NO3)2

Precipitation: Ammonium Hydroxide

Ammonium
Diuranate (ADU)
(NH4)2U2O7

Calcination and Precipitation: H2

Brown Oxide: UO2

Hydrofluorination: HF

Uranium tetrafluoride powder: UF4

Fluorination: F2

Uranium
Hexafluoride: UF6

Figure 23. The uranium refining process.

Figure 24. Uranium mill at Key Lake, Northern Saskatchewan, Canada. Source: Cameco.

Figure 25. Uranium refinery at Blind River, Ontario, Canada. Source: Cameco.
 Figure 26. Yellow cake, U3O8, Uranyl Nitrate liquid UO2(NO3)2, Ammonium Diuranate,
(NH4)2U2O7, and uranium dioxide UO2.

Figure 27. Separated yellow cake, U3O8.

Uranium hexafluoride: UF6, commonly referred to as “hex” is stored in cylinders made of

the Monel alloy, nickel or steel, shown in Fig. 27. At room temperature it forms crystals that
look like rock salt (Fig. 28).
Figure 28. Induction furnaces for the magnesium reduction of uranium metal.

Figure 29. Billets of depleted uranium weighing 1,100 lbs each.

Figure 30. Inspection of UF6 storage cylinders.

 Figure 31. UF6 crystals in a glass vial.

After initial refining, which may involve the production of uranyl nitrate, uranium
trioxide is reduced in a kiln by hydrogen to uranium dioxide: UO2. This is then reacted in
another kiln with hydrogen fluoride HF to form uranium tetrafluoride UF4, a powder. The
tetrafluoride is then fed into a fluidized bed reactor with gaseous fluorine to produce UF6.
Removal of impurities takes place at each step.
An alternative wet process involves making the UF4 from UO2 by a wet process, using
aqueous HF.

Figure 32. Enriched U235 buttons.

Uranium hexafluoride, UF6, particularly if moist, is highly corrosive. When warm, it is a

gas, suitable for use in the enrichment process. At lower temperature and under moderate
pressure, the UF6 can be liquefied. The liquid is run into specially designed steel shipping
cylinders which are thick walled and weigh over 15 metric tonnes when full. As it cools, the
liquid UF6 within the cylinder becomes a white crystalline solid (Fig. 28) and is shipped in this
form.
The siting, environmental and security management of a conversion plant is subject to the
regulations that are in effect for any chemical processing plant involving fluorine-based
chemicals.
The enriched uranium undergoes a metallurgical process in which it is turned into metal
buttons (Fig. 29), which could be used as a metal or converted to the fuel form desired such as
uranium dioxide, UO2.

10.10 NUCLEAR FUEL CYCLES

Three types of fuel cycles are considered for fission power plants, depending on whether
the fuel is recycled and on the type of reactor used for electrical generation as shown in Fig. 30.

Figure 33. The once through, thermal reactor recycle and the fast reactor fuel cycles.

In the once-through fuel cycle, the spent fuel is not reprocessed but kept in storage in
water pools at the reactor sites, until it is eventually reprocessed or stored as waste with a
contingent of the ability of being able to retrieve if ever needed. This is the fuel cycle currently
adopted in most currently operating reactors. It entails a waste of resources since the fuel is
disposed for primarily structural and metallurgical reasons, not for the depletion of the fuel in it.
The abbreviated overall reactions are:
 0 n1 + 92 U 238  92 U 239 + 
92 U 239  93 Np 239 + -1 e0 +  * +
93 Np 239  94 Pu 239 + -1 e0 +  * +
(13)
94 Pu 239  92 U 235 + 2 He4  
______________________________________
0 n1 + 92 U 238  92 U 235 + 2 He 4 +2 -1 e0 +2  * + 4

This shows that the produced Pu239 decays into U235. The mined U238 is ultimately
converted into U235 in this cycle, in addition to the generated fission products and alpha particles.
In the thermal reactor recycle, the spent fuel is reprocessed and the uranium and
plutonium isotopes are separated from the fission products. These are recycled in new fuel
elements in the form of the Mixed Oxide fuel: MOX composed of UO2 and PuO2. It is also
possible to recycle only the uranium and to store the plutonium, or vice versa. This fuel cycle
entails a better use of resources, and the disposal is simplified to the volume of the fission
products, instead of the whole fuel elements.
In the fast reactor breeder cycle, the uranium and plutonium in the spent fuel are
separated. They are then refabricated into new fuel elements. In fast breeder reactors, a blanket
of depleted U238 intercepting the otherwise leaking neutrons and breeding plutonium, surrounds a
central core of uranium and plutonium fuel. Breeder reactors can produce more fuel than they
consume, hence the name: breeder.

10.11 FUEL BURNUP AND CONSUMPTION RATES

For a reactor operating at a power of Pth thermal megawatts (MWth), and assuming a
recoverable energy per fission event of 190 MeV/fission, excluding the unavailable 10 MeV of
energy carried by the antineutrinos, one can write for the fission rate dF/dt:

dF Watt (Joules/sec) 1 fission 1 MeV sec

= Pth [MWth] .106 [ ][ ]. [ ]. -13
[ ].86,400[ ]
dt MWth Watt 190 MeV 1.6 x 10 Joule day
(14)
fissions
= 2.84  10 Pth [
21
]
day

Using Avogadro's law we can express the number of fissions N in g grams of U-235:

g Av
N= (15)
M

where Av is Avogadro's number,

M is the molecular weight of U235 = 235 amu,
g is the mass in grams,
we can convert the fission rate into the burnup rate, BR, where:
 dF g
BR= .
dt N
fissions M gm U 235
= 2.84  1021 Pth [ ]. [ ]
day A v fissions
235
= 2.84  1021 Pth (16)
0.6 1024
gm U 235
= 1.112 Pth [ ]
day

Thus, a reactor operating at a power of 3,000 MWth, will have a burnup rate of about
3,336 gms U235/day.
Actually, not all the U235 nuclei present in the reactor undergo fission, some will absorb
the neutrons through the radiative capture process, where a neutron is absorbed and a gamma
photon is emitted.
If the ratio of radiative captures to fission reactions is given by:


b= (17)
f

where b = 0.169 for U235, the total number of absorptions per fission is given by:

 f    a
(1 + b) =  (18)
f f

where  is the microscopic radiative capture cross section [barns],

f is the microscopic fission cross section [barns],
a is the microscopic absorption cross section [barns].

Consequently, the consumption rate, CR is larger than the burnup rate by the factor
(1+b):

CR = 1.112 (1+b) Pth [gm U235/day]. (19)

Substituting the value of b for U235, we find that for thermal neutrons fissions, U235 is
consumed at the rate of:

CR = 1.112 (1+0.169) Pth [gm U235/day],

= 1.299 Pth [gm U235/day]. (20)

Thus 1.299 grams of U235 per day are consumed per one MWth of reactor power.
10.12 MASS BALANCE IN THE ENRICHMENT PROCESS
An important consideration in the fuel cycle is for a utility to contract for a given amount
of feed material to obtain a certain amount of enriched fuel needed for the operation of the plant
(Fig. 34). The yearly mass CR of U235 consumption from Eqn. 19 is given by:

gm U 235
CR  365 1.112 (1+b) Pth [ ] (21)
year

The amount of the natural uranium consumed per year is:

CR 365[1.112 (1+b) Pth ] gm nat U

Mf   [ ] (22)
xf xf year

where xf=0.72 percent is the enrichment of U235 in natural uranium.

This amount of natural uranium would be contracted-for with a uranium mine for a
natural uranium reactor such as the CANDU heavy-water design.
For reactors using enriched fuel such as research reactors, light water reactors, gas cooled
reactors and fast neutron breeder reactors, it is interesting to estimate the amount of natural
uranium feed-material Mf that should be contracted for with a uranium mine for power plant
operating at a thermal power Pth.

Figure 34. Mass balance in an enrichment plant.

 Notice that the product mass being Mp, the amount of U235 in the product is consequently
equal to: xpMp. The product material in this case is given by:

CR 365[1.112 (1+b) Pth ] gm enr U

Mp   [ ]
xp xp year

Assuming that there are no losses in the enrichment process, a mass balance for the feed
material as uranium entering and exiting the plant can be written as:

Mf = Mp + MT (23)

Similarly, a mass balance can be conducted for the U235 isotope as:

x f . Mf = x p . M p + x T MT (24)

where xi stands for the enrichment for the feed (f) as natural uranium, product (p) enriched fuel,
and depleted uranium tailings (T).
The two equations 23 and 24 can be solved for the two unknowns Mf and MT through the
variables elimination method.
Substituting for the tails mass from the first equation as:

MT = Mf - Mp ,

into the second equation, we get:

x f . M f = x p . M p + x T [M f - M p ]
= x p . M p + x T .M f - x T . M p

Rearranging, we get:

[x f - x T ]. M f = [x p - x T ]. M p , (25)

From this we can write for the needed mass of feed material in terms of the enrichments
of the product, the feed and the tails:

[ xp - xT ]
Mf = .M p (26)
[ xf - xT ]

The enrichment of natural uranium feed is xf = 0.72 percent. The enrichment of the tails,
xT, is normally at 0.2 percent in weight.
10.13 SEPARATIVE WORK UNIT (SWU)
The capacity of uranium enrichment plants is measured in terms of the Separative Work
Units or SWUs. The SWU is a complex unit which is a function of the amount of uranium
processed and the degree to which it is enriched, or the extent of increase in the concentration of
the U235 isotope relative to the remainder, and the level of depletion of the remainder. The unit is
specifically expressed as “Kilogram Separative Work Unit,” and it measures the quantity of
separative work performed to enrich a given amount of uranium a certain amount. It is thus
indicative of energy used in enrichment when feed and product quantities are expressed in
kilograms. The unit “Metric tonnes SWU” is also used.
As an example, to produce one kilogram of uranium enriched to 3 percent U235 requires
3.8 SWU if the plant is operated at a tails assay of 0.25 percent, or 5.0 SWU if the tails assay is
0.15 percent, thereby requiring only 5.1 kg instead of 6.0 kg of natural uranium feed.
About 100,000-120,000 SWU is required to enrich the annual fuel loading for a typical
1,000 MWe light water reactor. Enrichment costs are substantially related to electrical energy
used. The gaseous diffusion process consumes about 2,500 kWhr or 9,000 MJ per SWU, while
modern gas centrifuge plants require only about 50 kWhr or180 MJ per SWU.
Enrichment accounts for almost half of the cost of nuclear fuel and about 5 percent of the
total cost of the electricity generated. It also accounts for the main greenhouse gas impact from
the nuclear fuel cycle if the electricity used for enrichment is generated from coal. However, it
still only amounts to 0.1 percent of the carbon dioxide from equivalent coal fired electricity
generation if modern gas centrifuge plants are used, or up to 3 percent in a worst case situation.

10.14 THERMAL DIFFUSION

The kinetic theory of gases predicts the differences in the rates of diffusion of gases of
different molecular weights. In this process, a liquid compound rises as it heats, falls as it cools,
and tends to separate into its lighter and heavier components as it cycles around a column.
If there is a temperature gradient in a mixed gas, there is a tendency for one type of
molecule to concentrate in the cold region and the other in the hot region. This tendency
depends on the molecular weights as well as on the forces between the molecules. If the gas is a
mixture of two isotopes, the heavier isotope may accumulate at the hot region or the cold region
or not at all, depending on the nature of the intermolecular forces. The direction of separation
may reverse as the temperature or relative concentration is changed.
H. Clausius and G. Dickel in 1938, in Germany, built a vertical tube containing a heated
wire stretched along the axis of the tube and producing a temperature difference of about 600 oC
between the axis and the periphery. The heavy isotopes became concentrated near the cool outer
wall. In addition, the cool gas on the outside tended to sink while the hot gas at the exit tended
to rise. Thermal convection thus set up a counter current flow, and thermal diffusion caused the
preferential flow of the heavy molecules outward across the interface between the two currents.
Applied to uranium enrichment, a gaseous uranium compound such as uranium
hexafluoride is circulated in an annular region between two vertical pipes kept at different
temperatures. The lighter molecules of U235F6 and U234F6 end to get more concentrated near the
hot surface where they are carried upwards by the convection current. An exchange occurs with
the current moving downwards along the cold surface producing a fractionation process. After a
state of equilibrium is reached, the gas near the upper end contains more of the light molecules
than near the lower end.
A system of two concentric tubes of 2 mm separation, 3 cm in diameter, and 150 cms
long, would produce a difference of 40 percent in the concentration of the rare isotopes between
its ends, and about 1 gm/day could be drawn from the upper end without upsetting the
equilibrium condition.
To produce large amounts of enriched U235, a large number of these units must be used in
parallel and in series. To produce 1 gm/day of 90 percent enriched U235 would require about
1,000 tubes.
In the thermal diffusion rack shown in Fig. 32, steam circulates through an inner pipe and
cooling water through an outer pipe surrounding it. This caused U235 to diffuse inward and
circulate upward. The S-50 thermal diffusion plant shown in Fig. 33 was built during the
Manhattan project at Oak Ridge Tennessee. The resulting slightly enriched material could be
later fed to other enrichment devices such as the magnetic separation Calutrons.

Figure 35. Liquid Thermal Diffusion Rack.

 Figure 36. The S-50 Thermal Diffusion Plant, Oak Ridge, Tennessee.

10.15 ELECTROMAGNETIC SEPARATION

Magnetic separation depends on the principle of the mass spectrometer. A mixture of the
ions that are being separated is generated at an ion source. Upon emerging from the source, the
ions are accelerated through a potential difference V, maintained by an electric field. The
positive ions carrying a charge q, acquire a kinetic energy equal to qV:

1
qV  2
M 235v 235 , for the lighter U235 isotope, (27)
2
and:
1
qV  2
M 238 v 238 , for the heavier U238 isotope. (28)
2

The velocities acquired by the isotopes depend on their masses as:

1
 2 qV  2
v 235   (29)
 M 235 
and:

1
 2 qV  2
v 238   (30)
 M 238 

Lorentz equation governs the behavior of a charged particle in an electromagnetic field

as:

F  qE  q vxB
 v is the velocity vector
B is the magnetic field vector
where:
E is the electric field vector
x denotes the cross vector product

For a magnetic field presence only without an electric field:

F  q vxB

This suggests that a charged particle injected in a magnetic field is subjected to a force
that is perpendicular to both the magnetic field vector as well as the particle’s velocity vector,
resulting in a rotational motion around the magnetic field lines. Equating the centrifugal force to
the magnetic force, using the magnitudes of the vectors, we get:

Mv 2
F  q vB 
R

If the magnetic field B, generated by cylindrical coils in which a current is circulated, is

perpendicular to the plane of the page, the ions will be forced to move in a circular path defined
by the centrifugal force relationship:

v 2235
Bqv 235  M 235 , for the lighter U235 isotope, (31)
R235

and:
v 2238
Bqv 238  M 238 , for the heavier U238 isotope, (32)
R238

where R235 and R238 are the circular radii of the path taken.
We can write expressions for the ensuing radii as:

 2qVM 235 
1 1
M v M  2qV  2 2
R235  235 235  235    , (32)
Bq Bq  M 235  Bq
and:
 2qVM 238 
1 1
M v M  2 qV  2 2
R238  238 238  238    . (33)
Bq Bq  M 238  Bq

There is a difference in the radii:

  2qVM 238   2qVM 235 
1 1
2 2
  R238  R235  
Bq Bq
1
 2V  2
 q 
  ( M 12  M 12 )
238 235
B

The ratio of the radii can also be written as:

R235 M 235
 (34)
R238 M 238

Figure 37. An 180-degrees magnetic separation device, showing the magnets and the vacuum
system.
Figure 38. Two magnets surrounding a single collection chamber in an alpha track 255 degrees
Calutron with a non-uniform 1/r magnetic field.

This shows that the isotopes of different masses will move in trajectories with different
radii. These radii are proportional to the square root of the masses. Collectors can be placed at
these radii to collect the different nuclei. These devices can generate only gram quantities of
isotopes. That is why other methods like gaseous diffusion and centrifugation had to be
developed.
In the magnetic separators or Calutrons (for Califiornia cyclotrons), uranium chloride
UCl4 is reported to have been used as a feed material. It oxidizes when exposed to air, creating a
processing problem.
Figure 34 is a diagram of a beta Calutron separator, including the ion source, accelerating
system, receiver, vacuum system and semicircular focusing ion paths in the magnetic field. In
more advanced Calutrons, the magnets with an iron core are designed to generate a non uniform
magnetic field with a radial dependence:

1
B( r ) 
r
resulting in a uniform magnetic field in the space between them where the vacuum chambers and
collection system are emplaced.

10.16 GASEOUS DIFFUSION

Commercial uranium enrichment was first carried out by the diffusion process in the
USA. It has since been used in Russia, UK, France, China and Argentina as well. Today only
the USA and France use the process on any significant scale. The remaining large USEC plant
in the USA was originally developed for weapons programs and has a capacity of some 8 million
SWU per year. At Tricastin, in Southern France, a more modern diffusion plant with a capacity
of 10.8 million kg SWU per year has been operating since 1979. This plant can produce enough
3.7 percent enriched uranium a year to fuel some ninety 1,000 MWe nuclear reactors.
The gaseous diffusion process accounts for about 40 percent of world enrichment
capacity. However, though they have proved durable and reliable, most gaseous diffusion plants
are now nearing the end of their design life and the focus is on centrifuge enrichment technology
which seems to be replacing them.
The diffusion process involves forcing uranium hexafluoride gas under pressure through
a series of porous membranes or diaphragms. As U235F6 molecules are lighter than the U238F6
molecules they move faster and have a slightly better chance of passing through the pores in the
membrane, possibly manufactured from sintered powdered nickel. The UF6 which diffuses
through the membrane is thus slightly enriched, while the gas which did not pass through is
depleted in U235.
This process is repeated many times in a series of diffusion stages called a cascade. Each
stage consists of a compressor, a diffuser and a heat exchanger to remove the heat of
compression. The enriched UF6 product is withdrawn from one end of the cascade and the
depleted UF6 is removed at the other end. The gas must be processed through some 1,400 stages
to obtain a product with a concentration of 3 percent to 4 percent U235.
It is not useful to try to isolate the isotopes of uranium in metallic form since uranium
would rapidly oxidize. In the gaseous diffusion process for uranium enrichment, the gaseous
uranium hexafluoride of the mixture of U234, U235 and U238 is used:

U234F6 + U235F6 + U238F6.

Fluorine is selected since it occurs in nature with 100 percent atomic abundance in the
isotope F19. The lighter molecules flow slightly faster than the heavier molecules across the wall
of a porous tube barrier. The portion of the gas passing through the barrier is slightly richer in
U235. The small fraction of U234 ends up with the U235 stream.
Uranium hexafluoride has a vapor pressure of 1 atmosphere at a temperature of 56
degrees C.
The separation factor, also known as the enrichment or fractionation factor of a process,
is the ratio of the relative concentration of the desired isotope after processing to its relative
concentration before reprocessing. If before and after processing the number of atoms of
isotopes 1 and 2 are defined as:

b1 = number of atoms of isotope 1 per gram of the isotope mixture before processing,
 b2 = number of atoms of isotope 2 per gram of the isotope mixture before processing,
a1 = number of atoms of isotope 1 per gram of the isotope mixture after processing,
a2 = number of atoms of isotope 2 per gram of the isotope mixture after processing,

then, we can define the separation factor r as:

 a1 
a 
r 2. (35)
 b1 
b 
 2

This definition applies to one stage of a separation plant or to an entire plant consisting of
many stages. One is interested usually in the either the single stage separation factor, or the
overall separation factor of the whole process. For a single stage, r is usually slightly larger than
unity, and we define the enrichment factor as:

 a1 
a 
  r 1   2  1 (36)
 b1 
b 
 2

In natural uranium the mass of isotope 1 is 235 and the mass of isotope 2 is 238, and:

 b1  1
  = 0.72 percent = 0.0072 = . (37)
 b2  140

If we want to enrich the uranium to 90 percent in the isotope 235,

 a1  90
  = 90 percent = 0.90 = .
 2
a 100

Consequently, a process producing 90 percent enriched uranium would need to produce a value
of r equal to:

 90 
 
r
100 
 126.0
 1 
 
 140 

In nearly every process, a high separation factor is associated with a low yield. This
requires an optimization of the process. A separation device with a separation factor of 2,
  a1 
a 
r2r  2 ,
 1 
 
 140 

implies a value of :
 a1  1
  .
 2
a 70

A yield of 1 gram per day is one that, starting from natural uranium, produces in one day
material that is 1 gram of U235 mixed with 70 grams of U238.

Figure 39. Gaseous diffuser enrichment cell.

The holdup, or total amount of material tied up in a separation plant may be very large in
a plant consisting of many stages. In a plant with large holdup, a long time in the range of weeks
or months is needed for steady operating conditions to be attained.
 A mixture of two gases of different atomic weights can be partly separated by allowing
some of it to diffuse through a porous barrier into an evacuated space as shown in Fig. 36. The
molecules of the lighter gas will have a higher average speed than the heavier molecules for the
same kinetic energy:
1 1
E  M 235v 235
2
 M 238 v 2238 . (38)
2 2

The ratio of the velocities becomes:

1
v 235  M 238  2
  (39)
v 238  M 235 

The molecules of the lighter gas will diffuse through the barrier faster so that the gas,
which has passed through the barrier (the diffusate), is enriched in the lighter isotope. The
residual gas which has not passed through the barrier is impoverished, or depleted in the lighter
isotope. For a mixture of two gases since the diffusion rates are proportional to the molecules
velocities, the separation factor for the instantaneous diffusate, called the ideal separation factor
 is inversely proportional to the square root of the molecular weights, and is given by:

M 238
 (40)
M 235

where M235 is the molecular weight of the lighter gas, and M238 is the molecular weight of the
heavier one.
Figure 40. Gaseous diffusion cascade showing the cells, compressors, and the enriched and
depleted streams.

Using the molecular weight of uranium hexafluoride,

352
  1.00429
349
 If the fraction of the original gas that has diffused is small,

r =  = 1.00429.

If the fraction of the original gas that has diffused is appreciable, an expected diminution in
separation is to be expected. If half the gas diffuses,

r - 1 = 0.69 (.

For uranium hexafluoride, r = 1.003, a lower value than the ideal 1.004.
Because of back diffusion, of imperfect mixing on the high pressure side, and of imperfections
in the barrier, a value in the range of 1.0014 can be expected in practice. For this reason, if one uses a
cascade in which a reasonable overall enrichment factor is achieved per stage, then it turns out that
about 4,000 stages are needed to obtain 99 percent enriched uranium, as shown in Fig. 37.
Thus, the theoretical separation factor is defined as the maximum increase in percent
enrichment per stage and is given by:

1
 m(U 238 F6 )  2
rdiffusion  235  1
 m (U F6 
)
1
 352  2
  1
 349 
 0.00429

and the molecules speeds of the two types of UF6 molecules differ by only about 0.43 percent.
Since uranium hexafluoride is highly corrosive, special alloys such as stainless steel or nickel
plated components must be used. Figure 38 shows the compressors and diffuser cells.

Figure 41. Gaseous diffusion plant cells arrangement showing the compressors and diffusers for the
cascade process.
Figure 42. The Tricastin Gaseous Diffusion plant, France can provide enough enriched fuel for 90
power plants. Four nuclear reactors are used to supply it with its electrical needs.

Figure 43. The K-25 Gaseous Diffusion Plant, Oak Ridge, Tennessee; no longer operational.
 Figure 44. The Portsmouth, Ohio, gaseous diffusion enrichment plant, no longer operational.

Figure 45. The Paducah, Kentucky enrichment plant, only operating gaseous diffusion plant in the
USA as of 2008.

10.17 CENTRIFUGATION PROCESS

INTRODUCTION

The principle of centrifugation was first advanced by Linderman and Aston in 1919. At
the University of Virginia in 1936, Beams separated the isotopes of chlorine using a gas
centrifuge.
A centrifuge plant uses just 5 percent of the electricity needs of a gaseous diffusion plant
of comparable size.
The gas centrifuge process was first demonstrated using a gas centrifuge operating in a
vacuum chamber in the 1940s during the Manhattan project in the USA and in Germany but was
shelved in favor of the simpler diffusion process. It was then developed and brought on stream
in the 1960s as the second generation enrichment technology. It is economic on a smaller scale
at a performance of under 2 million Separative Work Unit per year (SWU/yr), which enables
staged development of larger plants.
It has been deployed at a commercial level in Russia and in Europe by Urenco, an
industrial group formed by British, German and Dutch companies. Russia¹s four plants at
Seversk, Zelenogorsk, Angarsk and Novouralsk account for some 40 percent of the world
capacity. Urenco operates enrichment plants in the UK, the Netherlands and Germany and is
building one in the USA.
In Japan, the companies JNC and Japan Nuclear Fuel Limited (JNFL) operate small
centrifuge plants. The capacity of JNFL's plant at Rokkasho was planned to be 1.5 million
SWU/yr. China also has a small centrifuge plant imported from Russia at Lanzhou, with a
capacity of about 0.5 million SWU/yr. Another small plant at Hanzhong is operating at 0.5
million SWU/yr. Brazil has a small plant with a capacity of 0.2 million SWU/yr. Pakistan has
developed centrifuge enrichment technology, as well as North Korea. South Africa has an
enrichment capability.
Iran became was a shareholder with French, Belgian, Spanish and Italian interests in
Eurodif the gaseous diffusion group. Following the overthrow of the Shah of Iran by the Islamic
Revolution, Iran was unceremonially thrown out of the consortium and since then it embarked on
developing its own centrifuge technology as a matter of energy independence as well as national
pride.
Both France and the USA are now considering centrifuge technology to replace their
aging diffusion plants, not least because they are more economical to operate. No action has
been taken in the USA in favor of the USA utilities purchasing highly enriched uranium from
supposedly dismantled Russian nuclear warheads, and then blending it to a low enrichment to be
used in the light water power plants under the “Megatons to Megawatts” program extending to
2009.
A license valid for 30 years for a plant in Piketon, Ohio including authorization to enrich
uranium up to an assay level of 10 percent U235 was issued to the United States Enrichment
Corporation (USEC) which began construction of the American Centrifuge Plant in late May
2007. USEC is working toward beginning commercial plant operations in late 2009 and having
approximately 11,500 machines deployed in 2012, which would provide about 3.8 million
SWU/yr of production based on current estimates of machine output and plant availability.
As noted, a centrifuge plant requires as little as 50 kWhr/SWU power. For instance the
Urenco plant at Capenhurst, UK has an input of 62.3 kWhr/SWU for the whole plant including
infrastructure and capital works.

HISTORY

In the 1940s, the USA adopted gaseous diffusion as the most developed of the known
enrichment techniques at the K-25 plant at Oak Ridge, Tennessee, which shipped enriched
uranium starting in 1945. By 1941 Jesse W. Beams at the University of Virginia conducted the
first gas centrifuge experiments using the uranium isotopes. By 1956, two more gaseous
diffusion plants were built at Portsmouth, Ohio and at Paducah, Kentucky. In 1976, responding
to the needs of the civilian power industry, Porstmouth, Ohio was chosen to host a gas centrifuge
enrichment plant using third generation machines designated as Set III with an annual
performance of 200 SWUs per machine. Later the Advanced Gas Centrifuge (AGC) Set V
design was developed with a capability of 600 SWUs.
At this time the promising Atomic Vapor Laser Enrichment Isotope Separation (AVLIS)
method was developed. The AVLIS technology was chosen in 1985 as more versatile than the
AGC and required less funding, being at an earlier stage of development. In 1992, the Energy
Policy Act created the government Corporation: United States Enrichment Corporation (USEC)
and transferred the uranium enrichment enterprise to it to be later privatized as a public
corporation in July 1998. USEC invested $100 million on AVLIS then dropped it in favor of a
centrifuge design with 300 SWU of performance as a more economical option.

COMPONENTS

The centrifuge design in the USA, Europe and Russia is based on one introduced by
Gernot Zippe from Germany. In 1946, as a prisoner of war he started working on them in the
USSR. He worked with Beams after his release in the USA from 1958 to 1960, when he
returned to Europe.
The components of the Zippe design are a rotor, motor drive, casing, vacuum system,
suspension system, and a column containing the feed, tails and product lines.
The casing surrounds the rotor and provides leak tightness to provide a vacuum and
physical protection against the spinning rotor that rotates in a vacuum to minimize friction.
In advanced design centrifuges, magnetic bearings such as used in modern jet engines
and Magnetically Levitated (Maglev) trains are used instead of ball bearings.
The rotor is a thin walled vertical cylinder that is spun by the drive motor. The
suspension system holds the rotor upright within the casing.
The radial separation factor is proportional to the absolute mass difference between the
two separated isotopes rather than the ratio of the molecular masses in the gaseous diffusion
process.
In addition to the radial separation by the centrifugal force, there is also separation in the
axial direction through thermal diffusion.

OPERATION

The UF6 gas is introduced near the center of the rotor. It is acted upon by two
simultaneous forces:

1. The centrifugal force induced by the rotation of the centrifuge.

2. An internal countercurrent circulation flow induced by an axial thermal gradient
along the length of the rotor.

The thermal gradient is established by creating thermal non uniformities of temperature at

the rotor end caps or along the rotor wall. The gas at the hotter lower end rises in the centrifugal
field moving radially inward. The gas at the colder end top cap sinks down and flows outwards.
The countercurrent circulation created by thermal convection is superimposed on the radial
centrifugal flow resulting in a relatively large assay difference between the bottom and top of the
centrifuge, where the enriched product and depleted tailings are extracted at the top and bottom
respectively.
In a gaseous diffusion plant about 1,000 stages are needed to reach the 4-5 percent U235
enrichment needed for light water reactor designs. In contrast, these levels can be reached in
about 10 stages in a centrifuge cascade. On the other hand, the throughput from a single
centrifuge cascade is small and a commercial plant requires a large number of cascades, which
are the basic building blocks of the plant.

THEORY

When a mixture of the two molecules of U235F6 and U238F6 is rotated at high speed in a
cylinder, the pressure distributions for the two molecules take slightly exponential gradients.
The relative partial pressures at the cylinder axis and at its wall lead to a radial process
separation factor. The pressure ration between the wall and the axis varies largely with the
peripheral speed of the cylinder: at 300 m/s, the pressure ratio is 550:1, whilst at 500 m/s it rises
to 3.9x107:1.
The radial separation factor at 27 oC and a peripheral speed of 300 m/s is about 1.055,
compared with the theoretical process separation factor in the gaseous diffusion process of just
1.004. In practical centrifuges, the radial separation factor is further enhanced.
In the counter current centrifuge developed at the University of Virginia by Zippe in
1958, an internal recirculating flow sweeps the light molecule to the opposite end. This allows
withdrawal of the two light and heavy molecular streams of the enriched and tails.

SEPARATION FACTOR

Defining:

 = the enrichment factor between the feed stream and the product stream,

1/ = the depletion factor between the feed stream and the depleted stream,

the maximum separation factor is described by the ratio:

  1
mv 2 1
2
rmax   e2 RT d

where : m is the difference in molecular weight between the isotopes,
R is the universal gas constant,
(41)
T is the absolute teperature in o K,

is the length to diameter ratio of the rotor,

d
v is the peripheral velocity, [m/s].

The theoretical separation factor is reduced by inefficiencies in the counter current flow
and the feed and take off systems. It gives the discrete enrichment in one step of enrichment.
To obtain an estimate of the total number of centrifuge steps needed to obtain a certain
level of enrichment, one uses the separative work concept. It depends on the enrichment and
depletion of the isotopes, as well as on the mass of the gas handled. The Separative Work Unit
(SWU) has the dimension of mass, since the enrichment and depletion factors are dimensionless.
Enrichment services are provided in kg units of separative work.

SEPARATIVE WORK POWER

The rate of performing separative work is termed separative power and has the units of
[kg/s] or flow, which should not be confused with the product flow rate in the centrifuge. Dirac
in 1941 derived the following expression for the maximum separative power of a centrifuge:

2
1 1 
 U max   D  m v 2  
2 RT  2
where : D is the back or self diffusion coefficient,
 is the gas density,
m is the molecular weight difference of the isotopes,
v is the peripheral rotor velocity, (42)
R is the universal gas constant,
T is the gas temperature in degrees Kelvin,
is the rotor length,
 is the circulation efficiency.

It must be noticed that the separative power is proportional to the centrifuge length, and
to the fourth power of the peripheral velocity. It is independent of the rotor diameter, and favors
operation at low temperatures.
A practical limit exists on lower temperatures to prevent the condensation of the UF6 gas.
Two optimization factors are the centrifuge length and the peripheral velocity. The TC12 design
used by Urenco is three times less in length than the Advanced Gas Centrifuge (AGC) Set V
developed in the USA. The TC12 has an approximate annual performance of 40 SWU per
machine, whereas the longer and faster Series V USA machine can provide 300 SWU per
machine per year.
The maximum velocity achievable in a thin cylinder rotating around its axis is
determined by the strength/density ratio of the material of construction. Each material possesses
a typical bursting speed. It is also limited by the ability to attenuate harmonic flexural vibrations.
The maximum peripheral speed of a rotor is given as:


v

where:σ is the rotor tensile strength, (43)
 is rotor density.

A typical centrifuge can reach a rotational speed of 1,500 revolutions/ sec or 90,000 rpm,
compared with 12-25 rpm in the spin drying cycle of washing machines. To reduce friction, a
vacuum is created between the rotor and the stator. A magnetic bearing holds the top of the rotor
steady in a levitated mode with just a needle contact at the bottom.
A pulsating magnetic field imparts angular momentum to the rotor and causes it to rotate.
Various techniques are used to avoid destructive vibrations such as bellows allowing the
controlled flexing of the rotor. Control of the rotational speed of the rotor should allow it quick
passing through its natural frequency resonance speed during startup and shutdown.
Maraging-steel instead of high strength aluminum allows faster rotational speeds. A
carbon fiber composite rotor as used in aircraft wings and wind turbine rotor designs stands out
in that it can spin faster than other light weight and high strength materials.
Ultra high strength and super light alloys steels such as vanadium steel are promising
candidates. Vanadium alloyed with titanium would even offer the best strength to weight ratio of
any engineered material and is worthy of investigation.

Table 1. Candidate materials for centrifuge rotor tubes.

Maximum
Tensile Tensile
Density peripheral
Strength strength to
Material [gm/cm3] speed
[Newton/mm2] density ratio
[m/s]

Carbon fiber composite 1.7 >800 2,000 1176.47

Glass fiber composite 2.0 600 700 350.00
Maraging steel, superhard alloy 8.0 525 2,250 281.25
High strength steel 8.0 455 1,700 212.50
Titanium 4.6 440 920 200.00
High strength aluminum alloy 2.8 425 520 185.71
 The length is determined by the materials choices as well as the dynamics of the rotor.
Long small diameter and thin walled rotors can be rigid at only low rotational speed. As its
length increases, the transverse bending frequency decreases up to the point of the frequency of
angular rotation designated as the critical frequency. A subcritical centrifuge design such as the
Zippe design would have a length that allows it to operate just below this critical frequency.
Beyond the critical speed, a supercritical machine experiences flexural vibrations. In this
case the rotor vibrations are countered with precision balancing that reduces the vibration
amplitude to a level within the confines of the centrifuge casing. This uses high quality
manufacturing processes and fast computers to balance the rotor.
Circulation inefficiency resulting from back diffusion along the axial concentration
gradient in the rotor limits the actual performance.

OPERATIONAL CHARACTERISTICS

Like the diffusion process, the centrifuge process uses UF6 gas as its feed and makes use
of the slight difference in mass between U235 and U238. The gas is fed into a series of tubes, each
containing a rotor, one to two meters in length and 15-20 cm in diameter. When the rotors are
spun rapidly, at 50,000 to 70,000 rpm, the heavier molecules with U238 increase in concentration
towards the cylinder's outer edge. There is a corresponding increase in concentration of U235
molecules near the center. These concentration changes are enhanced by inducing the gas to
circulate axially within the cylinder using the process of thermal diffusion. The vacuum
contributes to thermal insulation.
The enriched gas forms part of the feed for the next stages while the depleted UF6 gas
goes back to the previous stage. Eventually enriched and depleted uranium are drawn from the
cascade at the desired assay.
To obtain efficient separation of the two isotopes, centrifuges rotate at very high
supersonic speeds, with the outer wall of the spinning cylinder moving at between 400 and 500
meters/sec to give a million times the acceleration of gravity.
 P(r) U238F6

U235F6

Figure 46. Radial pressure distribution in the centrifuge process.

Although the capacity of a single centrifuge is much smaller than that of a single diffusion
stage, its capability to separate isotopes is much greater. Centrifuge stages normally consist of a
large number of centrifuges in parallel. Such stages are then arranged in cascade similarly to
those for diffusion. In the centrifuge process, however, the number of stages may only be 10 to
20, instead of a thousand or more for gaseous diffusion.
 Figure 47. Single Centrifuge showing the flow of the depleted and the enriched streams. Input gas is
released into the center. The centrifugal forces concentrate the heavier molecules near the edges.
Heating the bottom of the centrifuge further causes the lighter gas to move through convection to the
top while the heavier molecules would concentrate near the bottom. The top scoop collects the
depleted stream, whilst the bottom scoop collects the enriched stream.

The centrifuge process uses less power than the gaseous diffusion process, but requires
high strength components in the rotating parts on the centrifuge and more elaborate power
inversion equipment. Magnetic bearings are used. High strength materials that would resist the
extreme corrosive effect from the uranium hexafluoride including stainless steels and composites
such as in aircraft components have been used in the rotors construction.
The great appeal of the centrifuge in the separation of the heavy isotopes is that the
separation factor depends on the difference between the masses of the two isotopes:

 centrifugation  m  ( M 238  M 235 )

(44)

not on the square root of the ratio of the masses as in diffusion methods:
 M 238
 diffusion 
M 235
(45)

Figure 44 shows a cutout through a single centrifuge, the European and American centrifuge
designs are shown in Fig. 45, assembly of Set III American centrifuges is shown in Fig. 46, and Fig. 47
shows a cascade of centrifuges.

Figure 48. Centrifuge banks at the European Urenco plant (top), and the USA USEC demonstration
plant (bottom).
Figure 49. Assembly of third generation Set III centrifuges with annual performance of 200 SWU/yr
per machine.
 Figure 50. Serial and parallel connection in a centrifuge cascade configuration.

The theoretical separation factor is defined as the maximum increase in percent

enrichment per stage and is given by:

 m(U 238 F6 ) 
rcentrifuge   235  1
 m(U F6 ) 
 352 
  1 . (46)
 349 
 0.0086

Thus the molecules speeds of the two types of UF6 molecules differ by only about 0.86
percent, double the amount in gaseous diffusion.
Half the stages required for enrichment to the same percentage as required by diffusion,
makes the centrifuge process advantageous. However, this is associated with a large number of
moving parts requiring maintenance and lubrication, in addition to constant inspection due to the
stresses created by the high rotational speeds.
The centrifuge is a low output device measured in kg SWU/year. Many thousands of
centrifuges are needed in a plant measured in the hundred of metric tonnes of SWU/year. The
separation factor of a single centrifuge is inadequate to obtain the desired enrichment in one step,
requiring the use of cascades of centrifuge operating in series and parallel arrays. The serial
centrifuges provide the multiplication of the separation effect, while the parallel components
provide the magnitude of the separative work.

PLANT DESIGN

The number of centrifuges in a cascade is governed by the separative work of the

individual centrifuge, its cost, the cost of the cascade associated equipment, and the cost of the
protection devices. A plant consists of many cascades grouped into operational units sharing the
same process services. The whole plant must operate under high vacuum with the UF6 pressures
within the cascade pipe work of just a few hundred pascals.
Natural uranium as UF6 is fed to the plant in its storage containers. They are steam
heated to just below the triple point producing vapor through sublimation. The gas is passed
through a pressure reducing manifold to the cascade manifold.
A cascade is bounded by feed, tailings, and product valves. The pumping action of the
centrifuges causes the flow of the gas. The product and gas streams are withdrawn through
desublimation in cold traps operated at – 70 oC. These are used in a batch mode: they are
isolated when they are filled and heated to 55 oC and their contents passed to the tails and
product containers.
The cascades need a medium frequency electric drive supply and a continuous supply of
cooling water to remove the waste heat generated by inefficiencies in the motors and the
frictional drag of the UF6 gas on the static scoop arms used inside the rotor to remove the product
and tails streams.
The cooling water is also used to create a thermal gradient along the length of the
centrifuge to stimulate the counter current flow. This adds thermal diffusion to the centrifugal
separation process.
Other components include steam heating, chilled water, venting systems, effluent
scrubbing and UF6 sampling systems.

DEVELOPMENT PROGRAMS

One of the main objectives of centrifuge design is to maximize the length and peripheral
velocities. This necessitates supporting studies of stress analysis, rotor dynamics,
superconducting magnetic bearing development, separation theory, gas dynamics and heat
transfer. Advances are pursued in the motor drive, composite materials, automated
manufacturing techniques and digital instrumentation and control.
New developments include a rotating plasma centrifuge where a plasma is spun at high
speed separating the two streams. A plasma as a conducting fluid is affected by a magnetic field,
superimposing a magnetic filed configuration could lead to a centrifuge process combining the
advantages of centrifugation and electromagnetic separation without the limitations of the
mechanical moving parts of the conventional centrifuge.
 Figure 51. Plasma Centrifuge concept and experiment.

10.18 AERODYNAMIC NOZZLE AND VORTEX PROCESSES

Two aerodynamic processes were brought to demonstration stage. One is the jet nozzle process,
with a demonstration plant built in Brazil, and the other the Helikon vortex tube process developed in
South Africa. They depend on a high speed gas stream bearing the UF6 being made to turn through a
very small radius, causing a pressure gradient similar to that in a centrifuge. The light fraction can be
extracted towards the centre and the heavy fraction on the outside. Thousands of stages are required to
produce enriched product for a reactor. Both processes are energy intensive using over 3,000
kWhr/SWU.
 Figure 52. Isotopic Separation in the Nozzle Process.

The Becker nozzle separation process uses an aerodynamic process to separate the different
isotopes. A mixture of UF6 and an auxiliary gas such as hydrogen or helium, is forced to flow
along a curved wall. The heavier molecules tend to remain closer to the wall than the lighter
ones. The auxiliary gas aids in accelerating the uranium hexafluoride gas to allow separation of
the molecules with the use of a knife's edge.
The separation nozzle approach provides enrichment per stage ratio that is 3 to 4 times that of
the gaseous diffusion approach. It also has fewer moving parts than the gas centrifuge,
minimizing wear and maintenance. The knife edge is however, subject to corrosion and must be
adjusted continuously to provide the desired separation.
The separation nozzles are joined in a series of ten nozzles together in a tubular separation
element. The inner channels provide the means for injecting the UF6 and for the removal of the
depleted gas. The enriched gas leaves radially and becomes the input to the next stage.

10.19 LASER ISOTOPE SEPARATION

Existing methods of enrichment cannot economically separate more than about 65 percent of
the U235 isotope from natural uranium. The remaining 35 percent is left in the tails as depleted
uranium. Thus one third of the useful U235 cannot be economically recovered by the existing
methods. Laser isotope separation may make that fraction available from the existing one half
billion kilograms of tails.
Laser enrichment processes have been the focus of interest for some time. They are a
possible third-generation technology promising lower energy inputs, lower capital costs and
lower tails assays, hence significant economic advantages. None of these processes is yet ready
for commercial use, though one is well advanced.
Development of the Atomic Vapor Laser Isotope Separation AVLIS, and the French
SILVA, began in the 1970s. In 1985 the US Government backed it as the new technology to
replace its gaseous diffusion plants as they reached the end of their economic lives early in the
21st century. However, after some USA $2 billion in Research and Development (R&D), it was
abandoned in the USA in favor of the SILEX molecular process. French work on SILVA has
been stopped.

Figure 53. Excitation cross sections for the U235 and U238 isotopes at different laser wavelengths.

Atomic vapor processes work on the principle of photo-ionization, whereby a powerful

laser is used to ionize particular atoms present in a vapor of uranium metal. An electron can be
ejected from an atom by light of a certain frequency. The laser techniques for uranium use
frequencies which are tuned to ionize a U235 atom but not a U238 atom. The positively-charged
U235 ions are then attracted to a negatively-charged plate and collected. Atomic laser techniques
may also separate plutonium isotopes.
In photochemistry, shining light into a mixture of chemicals can trigger chemical reactions.
When a molecule absorbs light it rotates and vibrates faster, that is it is excited to a higher level.
Photochemical reactions induced by lasers can be very specific in converting light energy into
stored energy.
 The wavelength and intensity of the absorbed light determine the excitation level of the
reactants, and also determine the type of products the reaction will yield.

MOLECULAR SILEX PROCESS

The main molecular processes which have been researched work on the principle of
photo-dissociation of UF6 to solid UF5, using tuned laser radiation. Any process using UF6 fits
more readily within the conventional fuel cycle than the atomic process.
The laser process under development on the world stage is Silex, an Australian
development which is molecular and utilizes UF6. In 1996 USEC secured the rights to evaluate
and develop Silex for uranium. It is also usable for silicon and other elements, but relinquished
these in 2003. The Silex process is now at prototype stage with the Silex Company in Sydney,
Australia and applications to silicon and zirconium are also being developed.
Laser photochemistry differs from classical photochemistry in major ways related from
the extreme brightness of the laser within a narrow frequency band, or color. This brightness
allows enough energy to be deposited in specific kinds of molecules so that all those molecules
dissociate or break up into fragments, forming new products. Laser sources can be tuned to the
frequencies required by a variety of molecules.

Figure 54. An infrared laser excites the U235F6 molecule. The excited U235F6 molecule loses one of its
fluorine atoms as it absorbs an ultraviolet laser.
 Figure 55. Schematic of the laser isotope enrichment process.

The purity of the wavelength in lasers provides the option of exciting only selected
molecules. The molecules of U235F6 and U238F6 have identical chemical properties, but have
different natural vibrational frequencies
Lasers can selectively be tuned to the frequency of a specific molecule with the result that
only that molecule absorbs laser energy and undergoes a change.
Such selectivity is used to drive a chemical reaction in one isotope but not the other. An
infrared long wave length laser excites the U235F6 molecule. The excited U235F6 molecule loses
one of its fluorine atoms as it absorbs an ultraviolet or short wave length laser. The new U235F5
molecule is in the form of a powder that can be easily separated from the gas, and collected by
standard techniques.
A complete process can be devised where the removal of the U235 isotope can be made to
reach up to 100 percent. If applied to the existing tails, this would be a beneficial use of a now
wasted resource. To avoid proliferation concerns, such a plant can be designed to enrich the
uranium to a specified level suitable for nuclear reactor applications and no more. A plant could
be designed to just enrich depleted uranium to the natural abundance of uranium for use for
instance in heavy water reactors. It appears that its proliferation risk is lower than the other
methods.
 Figure 56. Krypton fluoride cable-fed rare gas halogen ultraviolet laser.

One of the candidate lasers is the Krypton fluoride rare gas halogen ultraviolet laser.

ATOMIC VAPOR LASER ISOTOPE SEPARATION, AVLIS

In another approach to laser enrichment, molten uranium is used instead of uranium

hexaflouride, where the uranium vapor is generated by an electron gun. The uranium vapor is
expanded and irradiated with a laser light at a wavelength of 5,027.3 angstroms and a bandwidth
of less than 0.1 angstroms. The U235 molecules are ionized and drawn off to a product collector,
while the U238 passes off to the tails collector.

Figure 57. Green and yellow lasers beams dye corridor, AVLIS plant.
 Figure 58. Atomic Vapor Laser Isotope Separation (AVLIS) Configuration.

Figure 59. The AVLIS pilot plant at the Lawrence Livermore National Laboratory, LLNL.

This technology makes use of the fact that isotopes of different masses absorb slightly
different wavelengths of light; an indirect consequence of the nucleus mass difference. Precisely
tuned lasers would excite only the isotope atoms desired in a stream of atomic vapor. The
ionized atoms would then be separated from the neutral ones electromagnetically or by chemical
reaction. AVLIS has not been used on an industrial scale yet. This technique promises to allow
high efficiency production of high-purity U235 and Pu239, although its true useful is difficult to
judge without an operating plant to observe. The AVLIS technology, if available, could make it
possible for a country to produce substantial batches of weapon-grade uranium, neptunium or
plutonium from commercial reactor fuel. The energy required for separation itself is very low,
only enough to break the molecular bond or ionize the atom. Energy consumption is mostly
determined by the efficiency of the laser used, which is generally on the order of 0.1 percent.
Due to its economic advantage, laser isotope separation may become the choice
separation method for the twenty first century.

10.20 PLASMA ION CYCLOTRON FREQUENCY SEPARATION

PROCESS
The Savannah River Plant (SRP) in the USA under contract with TRW developed the
Plasma Separation Process (PSP) to recover the isotope U236 and to a lesser extent U234 and U238
from irradiated naval reactors fuel. All three build up during continued fuel recycle and
irradiation.
The concentration of U236 may reach 35 percent. Left in highly enriched fuel the
fissionable but non fissile isotopes absorb neutrons and result in significant deteriorated
performance during reactor operation. Upon the removal of the fissionable non fissile isotopes
the reactor efficiency increases and the need for additional Highly Enriched Uranium (HEU)
decreases.
A variant of the process uses a plasma source and a uniform superconducting magnetic
field. A microwave antenna excitation field feeds the ion cyclotron frequency that excites the
isotope to be separated. Upon absorption of the resonance energy, its circular orbit around the
magnetic field lines is increased leading to its collection on an array of plates.
 Figure 60. Schematic of plasma separation process using ion cyclotron frequency excitation.

Table 2. Percent isotopic content of recycled fuel.

Without Plasma Separation With Plasma Separation

Process Process
U234 1.6 1.2
U235 49.0 68.7
U236 35.0 25.8
U238 14.4 4.3

10.21 CHEMICAL PROCESS

A chemical process has been demonstrated to pilot plant stage but not used. The French
Chemex process exploited a very slight difference in the two isotopes U235 and U238 propensity to
change valency in oxidation/reduction, utilizing aqueous III valency and organic IV phases.

10.22 COUPLED LASER AND ELECTROMAGNETIC SEPARATION

Figure 61. Simulated ion trajectories in the separation of the Li7 and Li6 isotopes. Lasers alter the
magnetic moments of the different isotopes which would then follow different trajectories in
applied magnetic fields. Source: Physics World.

A new methodology under study combines two techniques of isotopic separation: the
electromagnetic separation or Calutron method and the laser isotope separation method. It is applied to
the separation of the Li6 and Li7 isotopes. Lasers alter the magnetic moments of the different isotopes
which would then follow different trajectories in applied magnetic fields.
Using lasers alone to separate isotopes with different nuclear magnetic moments has energy
requirements that are prohibitive. The methods of enrichment using lasers are primarily based on
ionization of the atoms requiring several different wavelengths and very high-power lasers. A subtler
effect called optical pumping is suggested, whereby the absorption of laser light can change the
magnetic moment of an ion, making it move in a specific direction in response to a magnetic field.
Different isotopes are excited at slightly different frequencies, and lasers emit at only one, very precise
frequency. This provides a very efficient way of changing the magnetic states of some isotopes but not
others, and thus to separate the two.
The isotope Li7 in the form of Li7OH is used as a buffering agent added to the boric acid used as
a burnable poison in the Pressurized Water Reactors (PWRs) water coolant. Li7 is also used in the resin
used to purify the water coolant and control its chemistry. The isotope Li6 is primarily thought-after in
the isotopic separation process. Li6 is used for the production, through neutron capture in a thermal
neutrons spectrum, of tritium for maintaining the thermonuclear weapons stockpile. It would be used in
future first generation fusion reactors using the DT fusion reaction. This is necessary since tritium
decays with a half-life of 12.33 years into the He3 isotope and hence about 5 percent of it is lost every
year and needs to be replenished to maintain the thermonuclear and the spiked (for yield control)
weapons stockpile. Li7 is thus a left-over or byproduct from the Li6 production process using the
COLEX process.
The introduction of the Th232-U233 fuel cycle and the use of a lithium fluoride molten salt as a
coolant would necessitate isotopic tailoring and the use of Li7 as a salt component. In a thermal neutron
spectrum, the naturally occurring Li6 would produce undesirable tritium.
Technical issues exist in the choices of lasers as well as the magnets that would be needed for
the process. Usually, electromagnets with copper wiring and iron cores would be used with the
Calutron or magnetic separation process. Instead, permanent magnets using the rare earths alloying
elements such as neodynium or lanthanum are suggested; and even superconducting magnets. A
discerned difficulty is that he altered nuclear magnetic moment of an isotope has to be maintained via
the optical pumping process during its movement in the applied magnetic field.

10.23 RE-ENRICHMENT OF RECYCLED URANIUM

In some countries spent fuel containing between 0.7-1.3 percent U235 is reprocessed to
recover its uranium and plutonium, and to reduce the final volume of high-level wastes. The
plutonium is normally recycled promptly into mixed oxide (MOX) fuel, by mixing it with
depleted uranium.
Where uranium recovered from reprocessing spent nuclear fuel is to be re-used, it needs
to be converted and re-enriched. This is complicated by the presence of impurities and two new
isotopes in particular: U232 and U236, which are formed by neutron capture in the reactor. Both
decay much more rapidly than U235 and U238, and Thallium208, one of the daughter products of
U232 emits very strong 2.6 MeV gamma rays, which means that shielding is necessary in the
plant.
U236 is a neutron absorber which impedes the chain reaction, and means that a higher
level of U235 enrichment is required in the product to compensate for its presence. Being lighter,
both the U232 and U236 isotopes tend to concentrate in the enriched rather than the depleted
output, so reprocessed uranium which is re-enriched for fuel must be segregated from enriched
fresh uranium.
Both the diffusion and centrifuge processes can be used for re-enrichment, though
contamination issues prevent commercial application of the former. A laser process would
theoretically be ideal as it would ignore all but the desired U235, but this remains to be
demonstrated with reprocessed feed.

10.24 UNITED STATES ENRICHMENT CORPORATION (USEC)

With plans underway for about 32 new reactors in the USA, a stable, domestic source of
enriched uranium is vital. The USA’s Nuclear Regulatory Commission (NRC) issued in April
2007 a construction and operating license for USEC Inc.'s American Centrifuge Plant in Piketon,
Ohio. The license, which is good for 30 years, includes authorization to enrich uranium up to an
assay level of 10 percent U235. USEC expects operates a Lead Cascade of centrifuge machines in
the American Centrifuge Demonstration Facility. The Company was working toward beginning
commercial plant operations and having approximately 11,500 machines deployed in 2012,
which would provide about 3.8 million Separative Work Units (SWU) of production. USEC
operates the only uranium enrichment facility in the USA: a gaseous diffusion plant in Paducah,
Kentucky. The American Centrifuge Plant is expected to use 95 percent less electricity than a
comparably sized gaseous diffusion plant.

EXERCISES
1. An executive at an electrical utility company needs to order uranium fuel from a mine.
The utility operates a single 1000 MWe power plant of the CANDU type using natural uranium,
and operating at an overall thermal efficiency of 33 percent. What is the yearly amount of:
a. U235 burned up by the reactor?
b. U235 consumed by the reactor?
c. Natural uranium that the executive has to contract with the mine as feed to his nuclear unit?
2. An executive at an electrical utility company needs to order uranium fuel from a mine.
The utility operates a single 1,000 MW(e) PWR power plant operating at an overall thermal
efficiency of 33 percent. The fuel needs to be enriched to the 5 w/o in U235 level. The
enrichment plant generates tailings at the 0.2 w/o in U235 level. What is the yearly amount of
natural uranium that the executive has to contract with the mine as feed to his nuclear unit?
3. Compare the ratios and the difference in the separation radii in the electromagnetic separation
method for the separation of the ions of the isotopes:
a) U235 and U238,
b) Li6 and Li7.
4. Identify the level of U235 enrichment in:
a. Natural uranium,
b. LWR: BWR and PWR, reactor fuel,
c. Depleted uranium discharge from enrichment plant,
d. Burnt-out discharged reactor fuel.

REFERENCES
1. Tim Wogan, “Isotope Separation with a Light Touch,” Physics World,” Institute of Physics,
http://physicsworld.com/cws/article/news/48830, March 2, 2012.
2. Ron Greene, “Back to the Future,” Nuclear Engineering International, Sept. 2003.
3. J. J. Duderstadt and L. J. Hamilton, "Nuclear Reactor Analysis," John Wiley and Sons, 1976.
4. I. R. Cameron, "Nuclear Fission Reactors," Plenum Press, New York and London, 1982.
5. J. R. Lamarsh, "Introduction to Nuclear Engineering," Addison-Wesley, 1983.
6. L. Horpedahl et. al., "Introduction to Laser Isotope Separation," LASL-78-13, Los Alamos National
Laboratory, 1978.
7. R. B. Kehoe, “The Uranium Enrichment Services Market,” Neville Geary, ed., Nuclear Technology
International, Sterling Publications, ltd., p. 90, 1987.
8. T. T. Edwards, “Uranium Enrichment by the gas Centrifuge Process,” Neville Geary, ed., Nuclear
Technology International, Sterling Publications, ltd., p. 95, 1987.

APPENDIX
CENTRIFUGE BEARINGS AND VIBRATION DAMPING
Figure A1. Initial centrifuge drawing in patent application by Jesse W. Beams and Leland B. Snoddy.

In a centrifuge there are certain speeds called the critical speeds at which the shaft is distorted or
flexed out of the axis of rotation causing vibration forces which are transmitted to the bearings. The
critical speed occurs when the speed of rotation of the device is equal to a natural mode of vibration of
the rotating member and the supporting structure. They depend on the stiffness of the shaft and the
inertia of the stationary and rotating parts of the device.
The vibrations must be damped through the design of the bearings, such as the use of dual
bearings that are independently supported for the limited movement relative to each other. The
bearings of the assembly are spaced axially of the shaft a distance different from the spacings of the
nodes of the vibration waves at the critical speeds.
The rotating shafts are hollow so as to allow the introduction and withdrawal of the depleted
and enriched streams. The bearings have an L shape and can be made of lead bronze with 20 percent
Pb. The vibration absorbing rings could be made of neoprene. The bearing assemblies could be
lubricated with oil or any suitable lubricant.