Home Energy Nuclear Electricity Climate Change Lighting Control Contacts Links



By Charles Rhodes, P.Eng., Ph.D.

TRU is an acronym for TRans Uranium actinides, a group if heavy element isotopes that arise from exposing U-238 to a neutron flux. Typically about 2 / 3 of the TRU is plutonium.

The motive for TRU Concentration is to reduce the cost of reprocessing of used CANDU and LWR fuel almost 10X by reducing the mass flow through the subsequent pyroprocess about 10X while maintaining the same TRU discharge mass flow.

About 2010 Peter Ottensmeyer realized that there is a way of converting used CANDU reactor fuel into FNR fuel with no spurious waste streams. The Ottensmeyer Plan has two major steps. The first step involves use of a recrystallization cascade to extract pure uranium oxide from used CANDU fuel. The second step involves electrolytic reprocessing of the remaining CANDU fuel residue to make FNR core fuel. An important non-obvious aspect of the UO2 extraction is simultaneous Cs extraction, which simplifies the subsequent electrolytic reprocessing. This web page focuses on practical implementation of the recrystallization process.

The recrystallization cascade described herein was originally conceived as a component of the Ottensmeyer Plan but this cascade potentially has more general application for TRU concentration of other types of used reactor fuel. Sometimes fuel can be used in a LWR (light water reactor), then in a CANDU reactor and then run through a TRU Concentrator and a pyroprocess to convert it into FNR core fuel.

The first step of the Ottensmeyer Plan, which should be performed on water moderated reactor sites, is to concentrate the used fuel by highly selective extraction of uranium oxide. This extraction process also releases radioactive Cs-137 which must be captured, as well as radioactive inert gas fission products, particularly krypton, xenon and argon. A small amount of Np follows the U.

For public safety reasons the resulting fuel concentrates should then be transported in shielded containers to a shared remote fuel reprocessing site.

The second step of the Ottensmeyer Plan is to electrolytically separate the CANDU fuel concentrates into its major components, to send fission products to 300 year isolated safe dry storage for natural decay and to fabricate FNR fuel rods from the remaining separated components.

The fission products are not waste. The fission products include rare earths that are in high demand in the electrical industry. After 300 years in storage the fission products will need further chemical processing to extract the valuable elements.

This web page describes an apparatus for concentrating used nuclear power reactor TRU about 10X by extracting nearly pure UO2 from used fuel. This apparatus uses a closed recrystalization process to separate the nuclear fuel bundle atoms into five components: UO2, Zr, Cs, inert gases and everything else (TRU oxides, mixed fission products).

The strategy is to reduce the used fuel shipping, storage and reprocessing costs by dividing the used fuel atoms into one large minimally radioactive portion and four smaller radioactive portions.

The largest portion, which comprises about 90% of the used nuclear fuel weight, consists of nearly pure uranium oxide. The ratios of the uranium isotopes in this portion are determined by the neutron irradiation history of the nuclear fuel. This portion has a very low radioactivity permitting relatively easy and inexpensive handling, transportation, storage and reprocessing with minimal gamma ray shielding requirements. Nearly pure UO2 extracted from used CANDU fuel has very low radioactivity whereas UO2 extracted from used Light Water Reactor (LWR) fuel has a higher radioactivity due to the presence of a larger fraction of U-232.

The required amount of recovered UO2 shielding is usually set by the U-232, Cs-137 and Np-237 concentrations in the uranium oxide. These isotopes may be present at low concentrations in the nearly pure UO2 due to non-idealities in the UO2 extraction process.

The second portion consists of the neutron activated zirconium hulls that were originally used for enclosing the uranium oxide pellets used to fabricate nuclear power reactor fuel bundles. This irradiated zirconium can be reused in the FNR core fuel alloy to prevent the plutonium fraction of FNR core fuel from forming a low melting point Pu-Fe eutectic with the Fe fraction of the fuel tube alloy.

The third portion consists of oxides of cesium (Cs) which are released when the recovered UO2 is heated. The Cs isotope Cs-137 is strongly radioactive and must be trapped and sent to 300 year storage.

The fourth portion consists of inert gas fission product atoms such as krypton, xenon and argon that were trapped within the used nuclear fuel. These inert gases are released when used nuclear fuel is dissolved in nitric acid. Since some of the inert gas isotopes are radioactive these inert gases must be safely captured and vented.

These gases are caught in an atmospheric pressure cold trap cooled by liquid nitrogen. The relevant boiling points are:
Ar87.3 K
Kr119.1 K
Xe165.1 K
CO2194.7 K sublimation
O290.2 K
AN277.3 K

The fifth portion contains the balance of the used nuclear fuel weight. For used CANDU fuel this portion is typically 76% U, 3% Pu-239, 20% fission products. This portion is intensely radioactive and must be handled and shipped in suitable shielded containers that have walls that have a gamma ray absorption thickness the equivalent of a 30 cm thickness of lead and must be stored in dry shielded vaults. The cost of transporting this portion is dominated by the cost of transporting the weight of the required shielded containers. Hence from a transportation and storage cost perspective this portion is with respect to the larger portion the better. A major safety concern with respect to transportation of this portion is ensuring that the used fuel will not go critical if water penetrates the transportation container.

The motivations for selective UO2 extraction from used CANDU fuel are to:
1) Increase the Pu/U concentration ratio in the fuel concentrates as required for making FNR core fuel;
2) Minimize the mass of CANDU fuel concentrates that must be transported to a remote fuel reprocessing site and then electrolytically reprocessed to produce FNR fuel;
3) Reduce the gamma emission from the extracted U3O8 sufficiently to make the U3O8 economic to transport and store.

One of the radioisotopes of concern is U-232 which can potentially occur in used CANDU fuel as a result of alpha particle capture and 4 n emission by impurity Th-232 atoms. Being an isotope of uranium it is not removed by the uranium selective recrystalization methodology used in this process. The U-232 has a half life of 72 years and its decay path involves a hard gamma emission. Since the Th-232 impurity content in the uranium used to produce CANDU fuel can potentially vary it may be necessary to measure the gamma ray output from the separated U3O8 and to provide sufficient shielding to ensure regulatory safety compliance (Ref: Monica Regalbuto, monica.regalbuto@inl.gov, Purex expert at INL).

This used reactor fuel concentration process will likely be of interest to parties who:
1) Have an inventory of used CANDU reactor fuel or light water power reactor fuel and who would like to use that used fuel to make new core and/or new blanket fuel for use in fast neutron reactors;
2) Have an inventory of used uranium oxide blanket fuel from fast neutron reactors and who would like to convert that used blanket fuel into new core and new blanket fuel for fast neutron reactors;
3) Have an inventory of used nuclear fuel from light or heavy water cooled reactors and who would like to mitigate the costs of transporting, storing and/or reprocessing of that used nuclear fuel;
4) Have an inventory of used nuclear fuel from light or heavy water cooled reactors and who would like to rapidly convert that used fuel into stable elements that pose no risk to future human generations.
5) Seek to prevent nuclear weapon proliferation or further nuclear waste formation via use of closed system electrolytic fuel reprocessing but who want to minimize the overall cost.

For an overview of the Ottensmeyer Plan please review OTTENSMEYER PLAN.

For an overview of nuclear fuel waste processing see the paper:
Radioactive Waste Partitioning and Transmutation.

Reference: Japanese 2002 patent

For an overview of Uranyl nitrate hexahydrate [UO2(NO3)2.6H2O] solubility see: Uranyl Nitrate Solubility
Uranyl nitrate hexahydrate solubility in nitric acid and its crystallization selectivity in the presence of nitrate salts
Uranyl nitrate hexahydrate solubility

A paper relevant to the U-232 issue is: Uranium-232 Production In Current Design LWRs.

Selective extraction of U3O8 should be done at existing CANDU reactor sites to realize about 90% of the spent CANDU fuel weight as pure UO2 and the remainder of about 2.45% of spent CANDU fuel weight as a mixture of (fission products + TRUs) + (remaining 7.55% of the CANDU fuel weight is uranium oxide). The uranium content of this mixture is used to meet the uranium metal content requirement of the FNR core fuel. The extracted UO2 must be sufficiently pure to reduce its radioactivity sufficiently to enable low cost transportation and storage.

1) The process equipment must not be so large that a critical mass can accumulate anywhere in the dissolver or in tank T1.

2) Nitric acid (4.5 M) acting on uranium oxide produces uranyl nitrate hexahydrate. The process of recrystalization of uranyl nitrate hexahydrate [UO2(NO3)2.6H2O] is used to selectively extract nearly pure UO2 from spent CANDU fuel. This process is ineffective at rejecting the elements Np and Cs. The Np simply stays mixed with the uranium. It is not a radioactivity problem unless it contains the isotope Np-237. Np-237 arises from fast neutron n > 2n reactions in U-238. In CANDU reactors the fast neutron flux is very low, so the Np-237 production is low so the contribution of Np-237 to spent fuel radioactivity is very low. This statement is not true for spent fuel from fast neutron reactors. Hence recrystalization of [UO2(NO3)2.6H2O] will not achieve comparable radioactivity reduction in spent fuel from fast neutron reactors. However, as compared to electrolytic methods recrystalization of uranyl nitrate hexahydrate [UO2(NO3)2.6H2O] may still be a valid process for concentration of fissile elements contained in a FNR perimeter blanket.

3) The process of recrystalization of [UO2(NO3)2.6H2O] is also ineffective at rejection of CsNO3.2H2O. After the recrystalization steps Cs is rejected by heating the residue to over 650 degrees C at which point the isotopes Cs-133, Cs-135 and Cs-137 vaporize as the oxides Cs2O, Cs2O2, Cs2O3 leaving behind U3O8 with some contaminant Np. These vapors can be condensed in a cold trap. As indicated above the contaminant Np is not a problem unless it contains Np-237 from fast neutrons interacting with U-238.

4) During the initial dissolving of used CANDU fuel in nitric acid most of the trapped radioactive Kr-81, Kr-85 and Ar-39 inert gas atoms are released in the dissolver tank. This trapped inert gas mixture must be captured, adequately mixed with the atmosphere and safely vented. There must be a mechanism to prevent uncontrolled release of these inert gases while new UO2 is being added to the dissolver tank.

U3O8 = UO2 + 2 UO3

UO2 + 2 HNO3 + 6 H2O = UO2(NO3)2.6H2O + H2 + heat
UO3 + 2 HNO3 + 5 H2O = UO2(NO3)2.6H2O

UO2(NO3)2.6H2O + heat = UO3 + 2 HNO3 + 5 H2O

According to Wiki UO2(NO3)2.nH2O exists as a dihydrate, a trihydrate and a hexahydrate. It appears that unless concentrated nitric acid is used the result will be the hexahydrate. However, when the hexahydrate is gradually heated during UO2 recovery the hexahydrate may liberate water molecules to form the trihydrate and then dihydrate forms.

0 deg C = 98 g / 100 g H2O
20 deg C = 122 g / 100 g H2O
100 deg C = 474 g / 100 g H2O



The main apparatus used for selective uranium oxide extraction is a nitric acid dissolver followed by a five stage uranyl nitrate hexahydrate recrystalization cascade fitted with a nitric acid recovery system.

The cascade operates by taking advantage of the temperature dependence of UO2(NO3)2 solubility in 4.5 M HNO3. Lowering the temperature of a saturated solution triggers recrystallization. During slow recrystallization impurity atoms are excluded from the UO2(NO3)2.6H2O crystals and concentrate in the surrounding liquid solvent. In order to realize high purity UO2(NO3)2.6H2O a cascade containing five successive recrystalization steps is used.

The dissolver temperature is kept at 100 degrees C, the dissolver solution is saturated by maintaining an excess of used CANDU fuel and at steady state the impurity fraction in the dissolver solution is about 24.5% or 10X the impurity fraction in used CANDU fuel

Tanks T1 to T5 are the shell sides of vertical axis shell and tube heat exchangers. Tanks T1 to T5 oscillate in temperature. During a cooling period crystals form. During a warming period crystals dissolve. Wrapped around the outside of the tube bundle is a perforated metal sheet that prevents the crystals flowing into the shell side drain, which is located at the very bottom of the shell side.

In the crystals grown in tank T1 the impurity fraction is comparable to the impurity fraction in used CANDU fuel;
In the crystals grown in tank T2 the impurity fraction is about 1 / 10 of the impurity fraction in used CANDU fuel;
In the crystals grown in tank T3 the impurity fraction is about 1 / 100 of the impurity fraction in used CANDU fuel;
In the crystals grown in tank T4 the impurity fraction is about 1 / 1000 of the impurity fraction in used CANDU fuel;
In the crystals grown in tank T5 the impurity fraction is about 1 / 10,000 of the impurity fraction in used CANDU fuel;
Tank T6 operates at a constant 20 degrees C and contains pure 4.5 M HNO3

The cascade is fed by a closed dissolver. The dissolver has a heating coil and a removable fuel basket to enable removal of undissolved fuel and zirconium hulls. In normal operation the dissolver is always hot at 100 degrees C.

The cascade consists of a series of nitric acid resistant tanks designated by T1, T2, T3, T4 and T5. The detail of tank T5 operation differs from the other tanks to enable a portion of the pure solution produced in tank T5 to feed the UO2/HNO3 recovery apparatus. Enough UO2(NO3)2.6H2O crystals are left in tank T5 to realize a saturated solution when clean 20 degree C HNO3 from tank T6 is injected into tank T5. Note that tank T6 is smaller than the other tanks and in normal operation is always cool at 20 degrees C.

There are pipes and pumps arranged such that tank Tn can pump backwards into tank Tn-1 and can pump forward into tank Tn+1. The pump inlets draw from the very bottom of the shell side tanks. The pump discharge connections to the adjacent tanks are to the shell side above the liquid levels in those tanks to ensure that there are liquid breaks between adjacent tanks. The cooling water enters the top manifold at its back side and exits the bottom manifold at its back side, to allow removal of the manifold top and bottom caps for service access. Each tank shell is vented to a common overhead vent pipe fitted with check valves to permit easy liquid transfer between tank shell sides while safely containing hot HNO3 and related gases.

During normal oscillating operation the gases move backwards from tank #5 to tank #1 and then are expelled to a cold trap to catch argon, krypton and xenon fission products, some of which are radioactive. The CO2 fed into the vent pipe freezes in the cold trap nd is recycled.

When the liquid level in tank Tn drops gas is sucked into the Tn head space from tank Tn+1. When the liquid level in tank Tn rises gas in the Tn head space is expelled into tank Tn-1. Excess gas in the dissolver flows into the cold trap.

Tanks T1 to T5 all have ultrasonic liquid level sensors mounted on the top of outboard parallel vertical tubes. These tubes also serve as shell side drain tubes.

During an intertank solution transfer each transfer pump normally runs until the liquid level in the shell side drain tube is below the bottom of the shell side. The exceptions to this rule are for strong hot solution transfers from the dissolver into tank T1 and for pure acid transfers from tank T6 into tank T5. In these cases the transfer pump stops when the desired liquid level in the destination tank is attained.

There is control logic which will stop the pumping sequence and alarm if any tank liquid level exceeds its design maximum.

The floor under the tanks is covered with an acid resistant material and is sloped to a common drain. The drain goes to the basement level dirty acid drain down tank. Thus any acid leak anywhere in the system gravity drains into this dirty acid drain down tank.

Tanks T1, T2, T3, T4, T5 all contain vertical heating/cooling tube bundles surrounded by a perforated metal sheet cylinder to confine the crystals. The purpose of the perforated cylinder is to prevent UO2(NO3)2.6H2O crystals that form on the surface of the heat exchange tubes during the cooling period from falling to the bottom drain of the tank and being lost during the weak solution transfers at the end of the cooling period. The perforated cylinders must be physically configured so as to not trap liquid acid.

The tubes in tanks T1, T2, T3, T4 and T5 are formed from 430A SS and isolate heating/cooling water flowing through them from the UO2(NO3)2 solution. In each of T1, T2, T3, T4 and T5 there are both top and bottom manifolds which interconnect the tubes in a single pass configuration. In the diagram these heating/cooling pipes are largely concealed by secondary pipes in front of them.

Each tank can be either heated or cooled at a controlled rate by circulating a temperature controlled water + ethylene glycol solution through the tubes. The temperature of the water is regulated by conventional heating/cooling equipment. On heating the maximum temperature difference between the heating water-glycol and the UO2(NO3)2 solution is set at 20 degrees C. On cooling the maximum temperature difference betwen the cooling water-glycol and the UO2(NO3)2 solution is set at 5 degrees C.

During the heating period the temperature of the heat exchange tubes is about 20 degrees C warmer than the temperatureof the tank shell. During the cooling period the temperature of the heat exchange tubes is about 5 derees C cooler than the tank shell. Hence there are oscillating length differences between the tubes and the shell. These oscillating length differences are absorbed by a thin flexible top tube sheet. The bottom tube sheet is more rigid because it must support a head of almost 5 m of solution.

Tanks T1 to T5 are weakly agitated during crystal growth by circulating cooling water plus glycol downwards so that there is weak natural coonvection within the UO2(NO3)2 solution. The purpose of the agitation is to keep the shell side solution temperature in the tank uniform and to minimize impurity accumulation on the crystal faces during crystal growth.

Each tank has a temperature sensor which is used to regulate the heating/cooling rates and to indicate when a heating or cooling cycle is complete.

At the end of the heating period the very strong solution in tank T5 is transferrred into the UO2 / HNO3 acid recovery unit which heats it to separate and recover HNO3, cesium oxides and pure UO2.

Absent U-232 in the fuel this cascade should reduce the uranium oxide gamma emission per kg down to the level of new CANDU fuel formed from natural uranium. The amount of U-232 gamma emission will be a strong function of the original Th-232 impurity concentration in CANDU fuel and a weak function of the age of the used CANDU fuel.

To enable system maintenance the dissolver basket containing remaining fuel and zirconium hulls is removed and placed behind a shielded barrier, all the tanks are heated to 100 degrees C to fully dissolve the remaining solids and then the entire system solution volume is drained down into the shielded below grade acid drain down tanks. Below the drain down tanks are pumps which can be used to transfer the acid solution back into the cascade tanks after the service work is complete.

The term "forward" refers to a solution flow forward from the dissolver toward tank T5. The term "backward" refers to a solution flow backward from tank T5 toward the dissolver. The tanks operate most of the time with similar liquid levels. The liquid transfer pumps are located sufficiently below the tanks to ensure adequate pump suction head.

The term "weak solution" refers to a UO2(NO3)2 4.5 M nitric acid solution which is saturated at 20 degrees C. The term "strong solution" refers to a UO2(NO3)2 4.5 M nitric acid solution which is saturated at 100 degrees C.

In the cascade UO2(NO3)2 net flows from the dissolver toward tank T5 while impurities including Pu and FP on average flow backwards from tank T5 toward the dissolver. Head space gases flow backwards toward the dissolver and then the cold trap.

The cascade is designed so that the various tanks operate in parallel at progressively higher purities during the slow cooling and heating periods when crystal growing and crystal dissolving occur. At the end of each heating or cooling period acid solution is transferred from tank to tank nearly sequentially. While acid transfers are occurring the liquid volume in the individual tanks fluctuates.

The T1,T2, T3, T4 and T5 tank temperatures are programmed to oscillate together, typically between 20 degrees C and 100 degrees C. The circulated heating water reaches up to 120 degrees C for fast warming and reaches down to 15 degrees C for slower cooling. Colder low end temperatures are possible with a suitable mechanical cooling equipment. The heating can be relatively fast (~ 1.0 degree C / minute) but the cooling must be slow (~ 0.25 deg C / minute) and carefully controlled. The required fine temperature control is achieved by controlling the tube side circulated water temperature.

During a tank cooling periood UO2(NO3)2.6H2O crystals tend to grow on the exposed cooling tube surfaces. At the end of a cooling period:
a) The remaining surrounding cool weak solution in tank T1 is transferred into the dissolver;
b) Then cleaner weak solution from tank T2 is transferred into tank T1;
c) Then cleaner weak solution from tank T3 is transferred into tank T2
d) Then cleaner weak solution from tank T4 is transferred into tank T3;
e) Then cleaner weak solution from tank T5 is transferred into tank T4;
f) Then a controlled amount of HNO3 is transferred from tank T6 into tank T5;
g) Then the cool weak solution in tank T1 is transferred into the dissolver;
h) Then cleaner weak solution from tank T2 is transferred into tank T1;
i) Then cleaner weak solution from tank T3 is transferred into tank T2
j) Then cleaner weak solution from tank T4 is transferred into tank T3;
k) Then cleaner weak solution from tank T5 is transferred into tank T4;

During the afore described cool weak solution backflow the UO2(NO3)2.6H2O crystals remain confined by the perforated metal cylinders.

Then the tanks are all heated. The UO2(NO3)2.6H2O crystals in tanks T1, T2, T3 and T4 re-dissolve in the cleaner surrounding warm acid solution. The solution resulting from melting the remaining crystals in T5 is transferred into the UO2 / HNO3 recovery system.

The equipment consists of a dissolver, tanks T1, T2,T3,T4,T5; acid supply tank T6 and a UO2 / acid recovery unit. The dissolver contains impurities at an equilibrium concentration about 10X the impurity concentration in the used CANDU fuel. The system goes through a long series of temperature oscillations each consisting of a slow cooling period from 100 degrees C down to 20 degrees C (320 minutes) followed by a more rapid heating period from 20 degrees C back up to 100 degrees C (80 minutes). At the end of each heating period hot strong solution is transferred forward one tank. During the subsequent cooling and crystal growth Pu + TRUs + FP are rejected to the weak solution. At the end of each cooling period the cool weak solution flows backward two tanks towards the dissolver. This backwards weak solution flow which contains a high impurity concentration moves impurities toward the dissolver faster than the forward strong solution flow which has a much lower impurity concentration moves impurities away from the dissolver. We can refer to each complete temperature oscillation as a temperature cycle. The net effect is that about (3 / 4) of the solute moves forward one tank per temperature cycle while the liquid acid component containing about (1 / 4) of the solute moves backward two tanks per temperature cycle. The strong solution that moves impurities forward is cleaner than the weak solution that moves impurities backwards so there is a net flow of impurities towards the dissolver.

In operation hot strong acid solution saturated with UO2(NO3)2 at 100 degrees C is fed from the dissolver into tank T1 and cool clean acid solution saturated with UO2(NO3)2.6H2O at 20 degrees C is generated in tank T5. As the dissolved used CANDU fuel moves from tank T1 to tank T5 it converts from being 97.55% UO2(NO3)2.6H2O to being nearly pure UO2(NO3)2.6H2O. The initial impurity fraction in used CANDU fuel is:
2.45%. The initial impurity fraction in the dissolver is 24.5%. If we assume an impurity fraction improvement of 10 at each tank after 5 temperature oscillations hopefully the impurity fraction will be 0.000245% corresponding to a UO2 purity of 99.999755%.

In order to form crystals a near saturation solution concentration must be maintained. Weak acid solution that flows backwards is saturated with UO2(NO3)2 at 20 degrees C. The strong acid solution that flows forward is saturated with UO2(NO3)2 at 100 degrees C.

1) It is desired to operate the system such that the impurity concentration in the dissolver is about 10X the impurity concentration in the crystals which grow in tank T1.
2) Similarly the impurity concentration in the crystals which grow in tank T1 should be about 10X the impurity concentration in the crystals which grow in tank T2.
3) Similarly the impurity concentration in the crystals which grow in tank T2 should be about 10X the impurity concentration in the crystals which grow in tank T3.
4) Similarly the impurity concentration in the crystals which grow in tank T3 should be about 10X the impurity concentration in the crystals which grow in tank T4.
5) Similarly the impurity concentration in the crystals which grow in tank T4 should be about 10X the impurity concentration in the crystals which grow in tank T5.
6) Hence there must be an ongoing back flow of impurities from Tank T5, through tank T4, through tank T3, through tank T2, and through tank T1 so that the impurity concentration is highest in the dissolver and decreases by about an order of magnitude through each successive tank.
7) Hence each cooling/heating temperature cycle must cause UO2(NO3)2 to move forward one tank and must cause the rejected impurities to move backward two tanks.
8) During successive temperature cycles additional fuel is added to the dissolver to keep its 100 degree C UO2(NO3)2 solution saturated.
9) After the end of each cooling period clean cool HNO3 is injected into tank T5 to combine with the remaining crystals to provide the required clean weak solution backflow.

The system goes through ongoing temperature oscillations. Assume an initial state where a cooling period has just ended. Tanks T1, T2, T3, T4 and T5 all contain UO2(NO3)2.6H2O crystals and saturated weak solution at 20 degrees C.

1) Then the remaining weak solution in tank T1 which contains a high concentration of impurities is transferred backwards into the dissolver. Allow sufficient time for the crystals in tank T1 to fully drain.
2) Then the remaining weak solution in tank T2 is transferred into backward into tank T1. Allow sufficient time for the crystals in tank T2 to fully drain.
3) Then the remaining weak solution in tank T3 is transferred backward into tank T2. Allow sufficient time for the crystals in tank T3 to fully drain.
4) Then the remaining weak solution in tank T4 is transferred backward into tank T3. Allow sufficient time for the crystals in tank T4 to fully drain.
5) Then the remaining weak solution in tank T5 is transferred backward into tank T4. Allow sufficient time for the crystals in tank T5 to fully drain.
6) Then high purity acid from tank T6 into tank T5 to form a synthetic solution in tank T5.
7) Then the remaining weak solution in tank T1 is transferred backwards into the dissolver. Allow sufficient time for the crystals in tank T1 to fully drain.
8) Then the remaining weak solution in tank T2 is transferred backward into tank T1. Allow sufficient time for the crystals in tank T2 to fully drain.
9) Then the remaining weak solution in tank T3 is transferred backward into tank T2. Allow sufficient time for the crystals in tank T3 to fully drain.
10) Then the remaining weak solution in tank T4 is transferred backward into tank T3. Allow sufficient time for the crystals in tank T4 to fully drain.
11) Then the remaining weak solution in tank T5 is transferred backward into tank T4. Allow sufficient time for the crystals in tank T5 to fully drain.
12) Then tanks T1, T2, T3, T4 and T5 are all simultaneously warmed up so that all the contained crystals in tanks T1, T2, T3 and T4 are dissolved forming strong solutions and the remaining crystals in tank T5 form a lower volume very strong solution.
13) Then the very strong solution in tank T5 is transferred into the UO2 / HNO3 recovery system. The recovered acid is required to refill tank T6 which must be done to enable implementation of step #6.
14) Then the strong solution contained in tank T4 is transferred forward into tank T5.
15) Then the strong solution contained in tank T3 is transferred forward into tank T4
16) Then the strong solution contained in tank T2 is transferred forward into tank T3.
17) Then the strong solution contained in tank T1 is transferred forward into tank T2.
18) Then a controlled volume of strong solution from the dissolver is transferred forward into tank T1.
19) Then tanks T1, T2, T3, T4 and T5 are all simultaneously gradually cooled to grow crystals in tanks T1, T2, T3, T4 and T5. This cooling step raises the impurity concentration in the liquid solution in each tank and lowers the impurity concentration in the crystals in each tank. The liquids containing impurities are later shuffled backwards in steps #1 to #11 inclusive.
20) After the impurity concentration in the dissolver has risen to 10X the impurity concentration in the used CANDU fuel, impurities can be removed from the dissolver in proportion to the amount of used CANDU fuel added or in proportion to the amount of pure UO2 produced, so as to maintain a steady state impurity concentration in the dissolver.
Note that when dissolver solution is removed for Pu + FP recovery it must be dried and the resulting HNO3 acid must be recovered and recycled.
21) Acid recovered at step #13 and step #20 is required to make the clean HNO3 required by tank #6.

The U + Pu + Fp impurities are harvested from the liquid solution in the dissolver after the dissolver has reached its equilibrium impurity concentration. The extracted dissolver solution is heated to evaporate the nitric acid which is recovered and recycled. A significant amount of energy is required. The resulting dry residue is the feedstock for successive operations to make FNR core fuel.

Very strong UO2(NO3)2.nH2O from tank T5 is transferred to the UO2 /acid recovery unit. Heating this very strong solution to recover the UO2 and the contained acid and the cesium oxide compounds requires a substantial amount of energy.

The very strong solution in the UO2 /acid recovery unit is mildly heated to drive off the nitric acid which is recycled. The dry residue is then heated to 650 degrees C to drive off radioactive cesium oxides which are caught in a cold trap and sent to 300 year storage. The remaining high purity depleted uranium oxide has a very low radioactivity and should be stored for future use as FNR blanket rod material.

Note that over time the quantity of Zr hulls in the dissolver will gradually increase. From time to time the process must be stopped to allow Zr extraction.

0 deg C = 98 g / 100 g H2O
20 deg C = 122 g / 100 g H2O
100 deg C = 474 g / 100 g H2O

Solute mass per 100 g H2O that moves one tank forward with each thermal cycle:
(474 g - 122 g) / 100 g = 352 g UO2(NO3)2 / 100 g H2O

Solute mass per 100 g H2O that moves two tanks backward with each thermal cycle:
= 122 g / 100 g H2O

Net solute that moves forward one tank in one thermal cycle is:
352 g / 100 g H2O - 2(122 g / 100 g H2O) = 108 g / 100 g H2O

The amount of CANDU fuel processed per thermal cycle is:
[UO2 / UO2(NO3)2] X 108 g / 100 g H2O
= [(238 + 32) / (238 + 32 + 2 (14 + 48))] X 108 g / 100 g H2O
= [(270) / (270 + 124)] X 108 g / 100 g H2O
= 74.01 g / 100 g H2O

Mass per tank per 100 g H2O that must be heated and cooled in each thermal cycle is:
100 g H2O + 474 g UO2(NO3)2

Note that these volumes include both the shell and tube sides and the manifolds. Dissolver, T1, T2, T3, T4, T5 Tanks:
Inside Diameter = 2.6 m
Liquid height = 4.25 m
Liquid volume plus liquid displacing tube volume:
Pi (1.3 m)^2 (4.25 m) = Pi (7.1825) m^3
= 22.56 m^3

The volume of a basement drain down tank sufficient to absorb the entire volume in the dissolver, T1 and T2 is:
Pi [3 (7.1825 m^3)] = Pi [21.5475 m^3]

Hence a sufficient 3 m diameter basement tank should be:
Pi (21.5475) / Pi (1.5 m)^2 = 9.5766 m long.

We can live with this drain down tank being only 9.0 m long because dissolver, tank T1 and tank T2 have their volumes partially reduced by heat exchange tubes.

We may need to be concerned about accumulating a critical mass in the dissolver. Hence, we may need to rethink these tank sizes.

Note that the submerged tube surface area must be consistent with the assumed heat transfer rate.


Estimate the heat required to swing all five tanks through 80 degrees C:
5 tanks X 574 g / stage/ 100 g H2O X 80 deg C X 1 cal / g-deg C X 4.18 J / cal = 959,728 J / 100 g H2O = 959,728 J / 100 g H2O X 1 W-s / J X 1 kWt / 1000 W X 1 h / 3600 s
= 0.2666 kWht / 100 g H2O

Note that this figure does not include the heat capacity of the tank metal.

Heat of formation of UO2(NO3)2.6H2O:
= -2739.5 Btu/lb of UO2(NO3)2
This is the heat that must be supplied to release UO2 from UO2(NO3)2.6H2O crystals.

1 Btu = heat required to raise 1 lb of H2O 1 deg F
= heat required to raise 454 g of H2O (1 / 1.8) deg C
= heat required to raise 252.2 g of H2O 1 deg C
= 252.2 cal

1 lb = 454 g

(1 Btu / lb) = 252.2 cal / 454 g
= 0.5555 cal / g

Thus the heat required to recover UO2 from UO2(NO3)2.6H2O is: 2739.5 Btu/lb of UO2(NO3)2
=2739.5 Btu/lb X (0.5555 cal / g) / (1 Btu / lb)
= 1522 cal / g of UO2(NO3)2.6H2O

Each thermal cycle produces 108 g of UO2(NO3)2.6H2O. Thus the heat required for UO2 recovery per thermal cycle is:
108 g X 1522 cal / g = 164,376 cal
= 164,376 cal X 4.18 J / cal = 687,092 J
= 687,092 J X 1 W-s / J X 1 kWt / 1000 W X 1 h / 3600 s
= 0.1908 kWht

Thus the energy required per thermal cycle / 100 g H2O / tank is:
0.1908 kWht + 0.2666 kWht = 0.45746 kWht / 100 g H2O / tank

From above, each thermal cycle produces:
74.01 g UO2 / 100 g H2O

Thus at a minimum a facility processing 5000 kg / day of used CANDU fuel has an average thermal power consumption of:
5000 kg / day X 1 day / 24 h X 0.47546 kWht/74.01 g X 1000 g / kg
= 1338.38 kWt
By the time fan and pump loads for heat removal are added and the heat capacity of the tanks is included this will be about 2.0 MWt. Note that this is an average heating power. In reality the system heats for 80 minutes followed by a cooling period of 320 minutes. Thus during the heating periods the peak power is 5X the average or:
5 X 2 MWt = 10 MWt.

Note that the UO2 / HNO3 recovery unit can be run continuously. Hence about 800 kWt is needed continuously and about 6 MWt are needed with a 20% duty cycle.

Each temperature cycle = 400 minutes.
Number of temperaure cycles per day = [24 h X 60 m / h] / [400 min / cycle]
= 3.6 cycles / day

In order to produce 5000 kg / day of UO2, each cycle must produce:
5000 kg / 3.6 = 1389 kg UO2

Each 100 g of acid produces:
74.01 g UO2.

Hence the required number of 100 g units of acid per tank is:
1389 kg / 74.01 g = 18.77 X 10^3

Hence the required amount of acid / tank is:
18.77 X 10^3 X 100 g
= 1877 kg
Hence it appears that the estimated tank volumes are 2 X larger than necessary. However, that extra volume allows for the tube and manifold volume in a shell and tube heat excahnge configuration with water-glycol in the tubes and volume expanding crystals on the shell side. Wrapped around the outside of the vertical tubes must be a perforated sheet to stop crystals falling into the shell side drain.

Crystal growth starts from a warm solution in tank Tn in which the various species are fully dissolved. Hence U is in the (VI) oxidation state and there will be:
UO2++ ions
NO3- ions
H+ ions
OH- ions
H2O molecules

As heat is removed from the solution in tank Tn crystals will grow on the surface of the cooling tubes. If the solution is not agitated impurities will tend to concentrate near the crystal growth faces. To minimize the incorporation of these impurities into the crystals modest continuous agitation is required. This agitation should also help keep the solution temperature uniform. Note that the tubes may need a surface texture to promote crystal adherence to the tube surfaces. This agitation is achieved by pumping cooling water iinto the topof the heat excahnge tubes to force natural circulation within the tanks during crystal growth.

In order to achieve good impurity exclusion the rate of change of temperature during the crystal growth rate must be limited to about 0.25 degrees C per minute.

The tanks should be thermally insulated from its environment so that the primary route for loss of heat is via the immersed cooling coils and cooling trays.

When the solution has cooled to its lowest temperature we must drain off and pump backward the remaining weak solution which will now have a higher concentration of impurities. The drain is at the very bottom of each tank to try to expel all possible impurities. This weak solution drains to a pump suction inlet and then flows backwards to the previous tank. There are two successive backwards flow cycles. Then the tanks are heated and strong solution moves forward starting with tank T4 transferring its entire strong solution contents into tank T5.

a) From the used CANDU fuel inventory that has been out of a CANDU reactor for at least 10 years withdraw used CANDU fuel bundles as required.

b) Mechanically shear the used CANDU fuel bundles into small pieces, each about 3 cm long.

c) Feed these pieces into the dissolver at a controlled rate sufficient to keep the dissolver solution saturated at 100 degrees C.

Over sufficient time the dissolver will fill up with Zr hulls and must be stopped to remove these hulls. The dissolver contains a large basket to expedite Zr hull recovery.

a) The dissolver is a nitric acid resistant tank with a removeable top. The dissolver tank contains a nitric acid resistant basket to allow convenient zirconium hull recovery. An empty basket is lowered into the dry dissolver tank, the dissolver tank top is replaced and the dissolver tank is evacuated. This evacuated air is exhausted to the atmosphere. The evacuation valve is then closed.

b) Then hot nitric acid temporarily stored in the drain down tank is pumped into the dissolver.

c) Due to the low overhead pressure in the dissolver tank nitric acid containing a low UO2(NO3)2 concentration flows from the drain down tank into the dissolver where it is heated to 100 degrees C. Used CANDU fuel is added to the dissolver via a feed tube. After some time in the nitric acid at 100 degrees C in the dissolver solution becomes saturated with UO2(NO3)2. Undissolved zirconium pieces collect in the dissolver's bottom basket. Radioactive inert gas fission products such as krypton bubble up through the liquid acid and collect in the sealed space above the acid.

d) Then a controlled volume of the hot nitric acid solution is pumped from the dissolver tank into tank T1 via the port at the bottom of the dissolver. Sufficient nitric acid solution remains in the dissolver tank over this port to prevent the radioactive inert gas fission products on top of the dissolver solution from exiting the dissolver via its bottom port.

At the end of each cooling period weak solution contining a high concentration of impurities is transferred from tank T1 into the dissolver. Later an equal volume of strong solution is transferred from the dissolver into tank T1.

e) Eventually when the inert gas pressure over the acid in the dissolver tank becomes too high or when the dissolver tank full of Zr hulls the dissolver is cooled to 20 degrees C to reduce the partial pressure of remaining HNO3 gas in the dissolver head space.

f) The inert gas plus some HNO3 gas in the dissolver head space are evacuated via a cold trap. The HNO3 vapor is caught in the cold trap. The radioactive inert gases are sent either to a high stack or to a pressure tank for later safe release to the atmosphere at a remote location. Ideally the inert gas should be stored to allow it to naturally decay. After a suitable decay period vent the residual inert gas to the atmosphere. Note that radioactive Kr-81, Kr-85 and Ar-39 must be well mixed with the atmosphere.

g) The nitric acid in the cold trap is isolated and is recycled back to the disssolver.

h) When the dissolver is full of Zr hulls the dissolver is drained into the drain down tank.

i) The dissolver tank top is removed. The basket containing zirconium pieces is removed from the dissolver and is air dried. The neutron activated zirconium is harvested from the dissolver basket for future use as a component of FNR fuel.

j) Transport the neutron activated zirconium and the CANDU fuel concentrates to the remote irradiated zirconium and fuel stores.

k) The now empty dissolver tank basket is replaced and the dissolver batch cycle repeats. Note that the dissolver tank is sufficiently large that one dissolver batch will serve many temperature cycles. Note that as the system operates the amount of UO2 in the dissolver gradually diminishes but the acid liquid level in all the tanks remains almost constant and the UO2(NO3)2 concentration in the dissolver solution remains almost constant. From time to time new used CANDU fuel is added via the feed tube.

1) The pumps that transfer acid from tank to tank must draw off the bottom of tank Tn and discharge above the top of the liquid levels in Tn-1 and Tn+1 so that there is no unintentional solute flow between adjacent tanks. There may be brief periods when the liquid flow into tank Tn from Tn+1 is equal to the liquid flow from tank Tn into Tn-1 which serve to backwash the pipes and pumps to prevnt subsequent pollution of the cleaner acid by the slug of dirtier acid that would otherwise remain in the pipes or pumps. To assist in this process the tank bottoms should be conical.

2) The pump inlets should be at the very bottom of the tanks. The pumps should discharge near the top of the adjacent tanks. The connecting pipe should be as small in volume (and hence in diameter) as practical to minimize the contained acid volume.

3) The pumps must be positioned sufficiently below the tanks to ensure sufficient suction head. All pumps should be magnetically coupled to provide a good nitric acid seals.

4) Provide perforated sheet cylinders in each tank to prevent crystals being sucked into the inlet of a backward pump. These must be perforated all the way to their bottom edges to ensure complete drainage.

5) It is necessary to carefully control the pumps to realize the optimum acid volume in each tank at a particular time in the operating cycle. If there is too much acid in a tank the system will be energy inefficient on thermal cycling and the forward solute propagation will be poor. If there is too little acid in a tank not all the crystals will be dissolved during a tank heating cycle, leading to insufficient crystal growth in the next tank during its cooling cycle. Thus the liquid levels must be carefully controlled. Each tank should have a precise liquid level sensing device and a pressure sensor at the tank discharge. The level control in tank T1 is particularly important as it sets the levels in the other tanks if the tank flows are all properly balanced. Thus the tube sheets and end manifolds must have provisions for the required shell side liquid level sensors.

1) In a suitable oven mildly heat the discharged ([UO2(NO3)2-nH2O] + [contaminant CsNO3.2H2O] strong solution from tank T5 to drive off the NO2 and H2O to realize [U3O8 + contamination CsO].

2) Condense, collect and recycle the evaporated nitric acid to tank T6.

3) Move the pure [U3O8 + contamination Cs2O,Cs2O2, Cs2O3] from #1 above to a suitable furnace.

4) Further heat the [U3O8 + contamination Cs2O,Cs2O2, Cs2O3] residue above 650 C to drive off the Cs2O, Cs2O2, Cs2O3 as a condensable gas.

5) Condense and collect the radioactive Cs2O, Cs2O2, Cs2O3.

6) Send the radioactive Cs2O, Cs2O2, Cs2O3 to fission product 300 year storage. After 30 years in storage the radioactivity should be dominated by Cs-137 (30 year half life) and Cs-135 (3.0 X 10^6 year half life). Use a segragated store in case this material contains other contaminants.

7) Collect the pure U3O8 residue and send it to supposedly pure U3O8 storage. The U3O8 may have to remain in this store for many decades until it is required for FNR blanket rod fabrication. Use a segragated store in case this material contains unspecified contaminants. Since this material is reserved for future use as reactor fuel such contaminants should have minimal import except for storage biosafety and for contamination of the future reducing agent.

1) After each cooling temperature cycle extract from the dissolver a solution volume sufficient to match the pure UO2 output (about 7 gm solid / 100 gm acid / themal cycle.

2) Heat this acid to evaporate, condense, recover and recycle the HNO3.

3) The remaining solution dry residue is CANDU fuel concentrates for further reprocessing (fission products + TRUs + uranium oxide).

4) Transport the CANDU fuel concentrates (fission products + TRUs + uranium oxide) to the remote site for electrolytic reprocessing.

1) The purification of U is not a matter of choice. The subsequent pyroprocess has to extract enough U in total first so that the cadmium cathode can start working on the transuranics. Otherwise it extracts the U into the molten Cd until the transuranics start coming out. That step is done cleaner on the iron cathode. It just happens that the iron cathode produces pure U as it extracts it without the transuranics.

2) To make the overall process as inexpensive and short as possible, the pre-extraction of U should ideally go to as high a percentage of U as possible without introducing unacceptable levels of impurities.

The envisaged cascade is orders of magnitude more mechanically simple than the earlier student design summarized below. Due to the use of many pumps from time to time the system will likely need mechanical service. To safely enable such service both the HNO3 and the radioactive species must be completely drained. To service the system heat the entire system to 100 degrees C to dissolve all of the UO2(NO3)2. Lift out the dissolver basket with a gantry crane and place it in a shielded enclosure and then drain all the fluids from the drain valves below the bottoms of the tanks to shielded below grade drain down tanks. If the radiation from residue remaining in the system is too high flush the system with clean 4.5 M nitric acid.

MP = 60.2 deg C
Dissociation at 118 deg C
UO2(NO3)2 is hygroscopic forming
6(H2O), 3(H2O), 2(H2O)

12 mg / kg (dog)

Natural Uranium:
U-238 99.27%, 4.47 X 10^9 Y
U-235 0.711% 700 X 10^6 Y
U-234 ~ 0.019%

Price of uranium oxide ~ $28 / lb


We can define the single stage separation factor S after crystal growth as:
S = (Impurity weight fraction in initial liquid before crystal growth)
/ (Impurity weight fraction in crystals after crystal growth)

Recall that from solubility data:
UO2(NO3)2 solubility in water is:
20 deg C = 122 g / 100 g H2O
100 deg C = 474 g / 100 g H2O

Immediately after crystal growth the crystal weight is: 474 g - 122 g = 352 g

Minimum weight of fluid solution is:
474 g + 100 g - 352 g = 222 g.

Let Wia = total weight of impurities in tank "a" during crystal growth
Let Wsa = weight of impurities in solution in tank "a" after crystal growth
Let Wca = weight of impurities in crystals in tank "a" after crystal growth
Wia = Wsa + Wca
S = [(Wia / 574 g)/ (Wca / 352 g)]
= (352 / 574)(Wia / Wca)
Wca = (352 / 574)(Wia / S)
= K Wia
K = (352 / 574)(1 / S)
Wsa = Wia - Wca
= Wia [1 - K]

Wca / Wsa = [K Wia] / {Wia [1 - K]}
= K / [1 - K]
Wsa = Wca [(1 - K) / K]

This is a general formula applicable to any tank. Hence:
Wsa = Wca [(1 - K) / K]
Wsb = Wcb [(1 - K) / K]
Wsc = Wcc [(1 - K) / K]
Wsd = Wcd [(1 - K) / K]
Wse = Wce [(1 - K) / K]

In the next tank assume that each cycle is like the one before it:
Wib = Wca + Wsb

After crystal growth:
Wcb = K Wib
Wsb = Wib [1 - K]
= (Wca + Wsb)[1 - K]

Wca[1 - K] = Wsb [K]

From above:
Wsb = Wcb [1 -K] / [K]

Wca[1 - K] = Wsb [K]
= {Wcb [1 - K] / [K]} [K]
= Wcb {1 - K}

Wcb / Wca = [1 - K] / {1 - K}
= 1

Thus single stepping backwards does not give the required improvement in crystal purity.

The proposed solution is to double step backwards. Then:
Wi1 = Wco + Ws2
Wi2 = Wc1 + Ws3
Wi3 = Wc2 + Ws4
Wi4 = Wc3 + Ws5
Wi5 = Wc4 + Ws6

Wco results from a measured volume injected from the dissolver. Ws6 results from a synthetic solution made in the CE by adding clean HNO3 to part of the CE crystal output.

From above:
Ws1 = [(1 - K) / K] Wc1
Ws2 = [(1 - K) / K] Wc2
Ws3 = [(1 - K) / K] Wc3
Ws4 = [(1 - K) / K] Wc4
Ws5 = [(1 - K) / K] Wc5

Ws1 = Wi1 [1 - K]
Ws2 = Wi2 [1 - K]
Ws3 = Wi3 [1 - K]
Ws4 = Wi4 [1 - K]
Ws5 = Wi5 [1 - K]

Eliminate the Wi terms:
Ws1 = (Wco + Ws2) [1 - K]
Ws2 = (Wc1 + Ws3) [1 - K]
Ws3 = (Wc2 + Ws4) [1 - K]
Ws4 = (Wc3 + Ws5) [1 - K]
Ws5 = (Wc4 + Ws6) [1 - K]

Now express Wsi in terms of Wci:
[(1 - K) / K] Wc1 = Wco [1 - K] + [(1 - K) / K] Wc2 [1 - K]
[(1 - K) / K] Wc2 = Wc1 [1 - K] + [(1 - K) / K] Wc3 [1 - K]
[(1 - K) / K] Wc3 = Wc2 [1 - K] + [(1 - K) / K] Wc4 [1 - K]
[(1 - K) / K] Wc4 = Wc3 [1 - K] + [(1 - K) / K] Wc5 [1 - K]
[(1 - K) / K] Wc5 = Wc4 [1 - K] + [(1 - K) / K] Wc6 [1 - K]

Now factor out [1 - K] everywhere to get:
[1 / K] Wc1 = Wco + [(1 - K) / K] Wc2
[1 / K] Wc2 = Wc1 + [(1 - K) / K] Wc3
[1 / K] Wc3 = Wc2 + [(1 - K) / K] Wc4
[1 / K] Wc4 = Wc3 + [(1 - K) / K] Wc5
[1 / K] Wc5 = Wc4 + [(1 - K) / K] Wc6

Now multiply through by K to get:
Wc1 = K Wco + (1 - K) Wc2
Wc2 = K Wc1 + (1 - K) Wc3
Wc3 = K Wc2 + (1 - K) Wc4
Wc4 = K Wc3 + (1 - K) Wc5
Wc5 = K Wc4 + (1 - K) Wc6

We have 5 equations in 7 unknowns, and we have a synthetic constraint on Wc6 so we should be able to find (Wc5 / Wco)

Assume: Wc1 / Wco = Wc2 / Wc1 = Wc3 / Wc2 = Wc4 / Wc3 = Wc5 / Wc4 = Wc6 / Wc5 = R Wc1 = K Wco + (1 - K) Wc2
= K Wco + (1 - K) (Wc2 / Wc1) Wc1

Wc1 / Wco = K + (1 - K) R (Wc1 / Wco)
R = K + (1 - K) R^2

K (1 - R^2) = R - R^2 = R (1 - R)
K = R (1 - R) / (1 - R^2)

For R = 0.1:
K = (0.1)(0.9) / .99
= .09 / .99
= 9 / 99

Recall that:
K = (352 / 574)(1 / S)
= 9 / 99

Thus the minimum required value of S is:
S = (352 / 574)(1 / K)
= (352 / 574)(99 / 9)
= 6.75

Hence for this cascade to work as contemplated the impurity fraction in the crystal must be at least 7X less than the impurity fraction in the solution from which the crystals are grown.

One of the important design constraints is set by tube sheet flexing. During the heating period the heat exchange tubes are about 20 degrees C hotter than the shell wall. During the cooling period the tubes are about 5 degrees C cooler than the shell wall. Hence even if the tubes and shell wall are perfectly TCE matched the tube sheets will still flex back and forth with each temperature cycle. Hence the heat exchange tubes must be centrally located in the tube sheets and the width of the non-tubed ring around them in combination with the tube sheet material thickness must accommodate continuous tube sheet flexing. Thus the tank height, the temperature coefficient of expansion, the Young's modulus, the yield stress, the tube wall thickness and the shell wall thickness will all be important. Keeping the tubes centrally located will reduce the heat transfer area. In order to move this design forward we need the physical properties of the material, which will be determined in part by the corrosion resistance. It may prove necesssary to make the top tube sheet thinner than the bottom tube sheet which must support the fluid head.

Peter Ottensmeyer:
The only plastic that is HNO3 resistant is Teflon (PTFE).
The preferred materials are 440 stainless steel, ceramic, and titanium

John Rudesill:
The concept Charles has described has merit. Counter current separations are ideally more efficient than single stage separations and are a preferred design practice in chemical engineering when practical.

The link I sent to Charles earlier https://www.rolledalloys.com/technical-resources/environments/nitric-acid/ indicates that various SS alloys are used in contact with 60% HNO3 at temperatures well in excess of 110 C. Plastic could work, but is both structurally and thermally inferior to SS alloys, The heat conduction coefficient is also much lower than metals. I will dig a little deeper for approved metals for this service. We have to be aware that the FP's contain halides which can make HNO3 far more corrosive. I need to see the expected solution analysis of the product coming from the fuel element dissolution step to account for the halide content. I am aware than jacketed tanks are made of even carbon steel and lined with teflon to make them near impervious to corrosion until the eventual pin hole forms leading to liner failure and jacket penetration. Heat transfer is impeded, but is still workable. Teflon can be used over 200 C as long as it is not used in a structural role. Similarly metal tubing can be teflon coated. For some services polyvinyl difluoride (F analog of Saran) can be used in place of teflon and it is somewhat more structurally robust.

I will read over the procedure Charles has provided and see if I can make a hand sketch of this process to study. It makes sense to scope this concept out at the kg/d scale or even smaller before making any commitments to a detailed design for a ~5mt/d facility. Materials handling will be a dominant aspect of the design given the radio hazard content.

A 5 mt /d process rate may seem like a lot, but is a very modest pilot plant. There is little incentive to try to save on equipment costs as they will be at large variance with usual engineering economy of scale cost calculations. The design must prioritize absolute containment and separation completeness over almost all other considerations. We want to do each step only once--no do overs at least in a continuous flow mode. In a batch mode, a step can be repeated if necessary. Also, batch mode can enable single shift 5 day week operation, initially. Continuous operation must be 24/7. As I write this, I think batch mode is the wiser choice. Once it is up and running and understood, a prudent consideration of continuous 24/7 operation is reasonable.

The Ottensmeyer Plan originated in student work at the University of Toronto circa 2010.

The following notes are an edited version of student work. There are various claims made by the students based on references that should be confirmed before large sums of money are committed to implementation of the process.

2.1.1 Fuel Bundle Shearing

CANDU nuclear fuel bundles are made from fuel pellets, inserted into zircaloy cladded fuel rods and loaded into channels of a cylindrical metal assembly (See Appendix B for diagram and composition). Thus, the first step of reprocessing begins with the removal of these long, narrow fuel rods from the bundle and liberating the spent fuel. The proposed process achieves this by mechanically shearing the rods into small segments, approximately 3cm [6], and dissolving the exposed fuel in a hot nitric acid dissolver solution.

Other alternatives for fuel liberation were considered, such as chemical decladding, mechanical decladding and perforating the cladding. Decladding proved uneconomical due to excess waste and high losses, while perforating the cladding alone does not provide enough exposed fuel surface area to dissolve it in a timely fashion [6]. Sawing as an alternative to mechanical shearing was also considered; however, it is less consistent and produces more metal fines [7].

2.1.2 Dissolution
The dissolver selected is a countercurrent flow, multistage, rotary dissolver. After shearing, pellets are fed into one end of the dissovler, while hot nitric acid is fed into the other end at 4M and 95°C [6]. The resulting solution, after dissolution of the spent fuel, is approximately 300g/L (using U as a basis). The concentration and temperature of nitric acid was found by analysis of calculated rate data and experimental data (see Appendix C). While higher concentrations of nitric acid yield higher rates of dissolution (75 minutes versus approximately 3 hours), it also introduces further hazards with regards to acidity and corrosion. The dissolver stage was not rate limiting in the process and thus the lower, less hazardous concentration could be applied.

As the dissolver drum rotates, the pellets in the initial stage are transferred along its length, while nitric acid flows counter-currently dissolving the fuel within. The drum is rotated to propel the solids forward and moves in a rocking motion to establish appropriate agitation for speedier dissolution. It also allows for better mass transfer and more efficient dissolution, as the least soluble fuel particles are contacted with the strongest acid [7]. See Appendix E-1 for schematic.

At the last stage of the dissolver, the solid cladding is ejected from the bottom, while the loaded nitric acid, containing uranium nitrate, plutonium nitrate and other dissolved fission products and actinides, flows out from the opposite end. The main dissolution reactions (Rxn. 1 and 2) are the dissolution of uranium dioxide (UO2) in nitric acid. Rxn. 2 is dominant when nitric acid concentration is less than 10M [7]. The other actinides dissolution reactions proceed similarly.

UO2+ 4HNO3 = UO2(NO3)2+ 2NO2+ 2H2O ΔHr°= -74.9 kJ/mol [Rx.1]
3UO2 + 8HNO3 = 3UO2(NO3)2 + 2NO + 4H2O ΔHr°= -367.6 kJ/mol [Rx.2]

During dissolution, NOx gases and radioactive iodine vapours from the fuel are emitted and sent for gas treatment (see section 6.2) while the concentrated solution is sent for crystallization. The undissolved cladding is rinsed, monitored for fissile material, packaged and transferred to the solid waste storage area for disposal.

This dissolver design is an update on more traditional spent fuel dissolvers that involve placing pellets into perforated baskets, followed by immersion in hot nitric acid. There are a number of operating limitations associated with this method, including the requirement of batch processing. Furthermore, the use of a highly corrosive solvent, off-gas emissions, as well as the potential for criticality, leads to challenges and limitations on the amount processed per batch. The equipment also decreases any criticality risk due to the processing of less mass in one location and its long, narrow geometry [7].

2.1.3 Crystallization and Clarification
The dissolver solution contains both long lived actinides used for FNR fuel and shorter-lived fission products. Before separation of actinides from fission products, the majority of U can be crystallized out of solution directly into Uranyl Nitrate Hexahydrate (UNH), a yellow green crystal. Due to its high concentration of U in solution (300g/L present as uranyl ions, UO22+), as well as its relative insolubility, a decrease in solution temperature to 10-20°C is capable of crystalizing approximately 70-80 wt% of the U [8].This was determined based on solubility-temperature data, as well as experimental results (see Appendix D). Co-crystallization of smaller amounts of Pu(VI) and Np(VI) is possible and favourable, as this would further diminish the amount of actinides requiring later processing. The crystallization reaction proceeds as follows:
UO22++ 2NO3- + 6H2O = UO2(NO3)2•6H2O ΔHr°= -20 kJ/mol [Rx.3]

One issue with crystallization is the contamination of U with fission products. Testing has shown that there is negligible occlusion of fission products within the crystals, and crystals can be washed to achieve a decontamination factor of about 100 for fission products (after 3 wash cycles) [8]. Washing is done with cooled water in the last stages of a multistage crystallizer.

Although crystallization prior to extraction is not carried out in conventional PUREX, this modified process uses it to reduce the amount of aqueous and organic solution being processed. This addition provides process and economic benefits (e.g. less process fluid), as well as environmental benefits as the corresponding liquid waste requiring treatment and disposal is reduced. While it may be possible to achieve greater amounts of U crystallization with lower temperatures (i.e. ~99% at -30°C [8]), liquid-liquid extraction equipment (see Section: 2.1.4 Main Liquid-Liquid Extraction and Stripping) would still be required to achieve separation of the remaining actinides. The smaller temperature reduction for approximately 70% U crystallization maximizes the benefit of crystallization, without incurring unnecessary costs and complexities.

Before proceeding to the main extraction stage the exiting dissolver solution must be clarified, also by centrifugation, to remove any suspended particles that may interfere in extraction. The collected particles are recycled for further dissolution. The dissolver solution then proceeds to a mixing tank for acidity and concentration adjustment, mainly to maintain the high concentration of nitric acid (approximately 5-6M) required for high efficiency separation in the main extraction stage [4]. The UNH crystals which exit the crystallizer as a slurry, are gravity fed into a mixing tank for further processing (see section 2.1.7 Metal Formation for more details).

The crystallizer selected is an annular, rotary, screw-type, continuous crystallization device, with a cooling jacket for temperature control. This design was selected due to its continuous operation and its similarity in functioning to the other screw-type equipment used in the overall process. The crystallizer was designed to operate at an incline to aid in nucleation and growth of crystals. This also facilitates easier removal of crystals without damage. As the internal spiral blades rotate they dislodge the crystals growing on the wall and move them upward to a centrifugal basket for washing and filtration, while the remaining dissolver solution flows downward. The rotary screw design, in a similar way to the rotary dissolver, also provides compartments or stages that allow for crystal washing to be done internally. See Appendix E-1 for schematic.

One operational disadvantage of the crystallizer design is the potential for crystal accumulation and blockage of discharge streams. This can be mitigated by appropriate operation. For example, the screw rotation, dissolution feed and coolant feed can be adjusted to restrict new crystal growth or break down crystal agglomerates. Scrub acid may also be fed to decompose any crystal blockages [9] (see Section 3. for more details).

2.1.4 Main Liquid-Liquid Extraction and Stripping
After crystallization, clarification and concentration, the dissolver solution is ready for the main liquid-liquid extraction and stripping. The solution is fed into the middle of a multistage centrifugal contactor to begin the separation of the remaining actinides from fission products.

M. Nakahara, Y. Sani, Y. Koma, M. Kamiya, A. Shibata, T. Koizumi and T. Koyama, "Separation of Actinide Elements by Solvent Extraction Using Centrifugal Contactors in the NEXT Process," J Nucl Sci Technol, vol. 44, no. 3, pp. 373-381, 2007.

J. D. Law, T. G. Garn, D. H. Meikrantz and J. Warburton, "Pilot-Scale TRUEX Flowsheet Testing For Separation of Actinides and Lanthanides from used Nuclear Fuel," Separation Science and Technology, vol. 45, pp. 1769-1775, 2010.

H. M. Mineo, H. Isogai, Y. Morita and G. Uchiyama, "An Investigation into Dissolution Rate of Spent Nuclear Fuel in Aqueous Reprocessing," J. Nucl. Sci. Technol., vol. 41, no. 2, pp. 126-134, 2004.

R. Jubin, "Spent Fuel Reprocessing," in Introduction to Nuclear Chemistry and Fuel Cycle Separations Course, Consortium for Risk Evaluation With Stakeholder Participation, Nashville, 2008.

T. Takata, Y. Koma, Sato, Koji, Kamiya, Masayoshi, A. Shibata, K. Nomura, H. Ogino, Koyama, Tomozo and S.-i. Aose, "Conceptual Design Study on Advanced Aqueous Reprocessing System for Fast Reactor Fuel Cycle," J Nucl Sci Technol, vol. 41, no. 3, pp. 307-314, 2004.

K. Ohyama and K. Nomura, "Development of Uranium Crystallization System in “NEXT” Reprocessing Process," in Proceedings of International Conference on Advanced Nuclear Fuel Cycles and Systems, Boise, 2007.

Appendix D: Crystallization of Uranium in Nitric Acid Data
The following graph depicts the solubility of uranium in nitric acid, as a function of uranium and nitric acid concentration, plus temperature. Moving from right to left, the uranium concentration decreases with temperature before reaching a minimum point, where water and nitric acid crystallize. To avoid this co-crystallization, a crystallizer must operate to the right of this point. The graph shows that as the concentration of nitric acid increases, the minimum point shifts from right to left and a higher UNH crystallization is possible. To maximize crystallization, the feed uranium and nitric acid concentration should be maximized as well [6].

Figure 3 Solubility Curves of Uranyl Nitrate
The above equilibrium data is supported by experimental data. Crystallization tests at high uranium concentration (300-600 g/l) and high nitric acid concentration (4-6M) to simulate spent fuel dissolver solution have yielded 70-80% uranium recovery at 10-20°C, and up to 95% at -10°C [7]. The results for a run with 500g U/l and 5M nitric acid are summarized below:
Figure 4 Uranium and nitric acid concentration in mother liquor (left) and recovery of uranium (right)

Appendix E-1: Equipment Diagrams
The following appendix features schematic diagrams of certain non-standard key equipment to aid in the visualization and understanding of their functioning, as described in Section 2.1 of the Process Description, and also as discussed in the non-key unit sizing section.

Figure 5 Fuel Bundle Shearer [8]

Figure 6 Continuous Rotary Dissolver [9]

Figure 7 Rotary Crystallizer

Crystallizer is scaled down from a tested device, having a feed rate of 1380L/hr and a residence time of 1 hour [6]. The dimensions of internal rotary cylinder with the blades of the tested crystallizer device are 11 centimeters of internal diameter and 50 centimeters by length [2]. Since the amount of uranium crystals to be dealt with is much less, the dimensions are rescaled by reducing the original volume of 0.00475m3 by half to 0.002375m3 in order to design for smaller blades.

Cost Table

Appendix T: Works Cited for Appendices

[1] Advamacs, "Concentration Calculator," [Online]. Available: http://www.trimen.pl/witek/calculators/stezenia.html. [Accessed November 2012].

[2] D. P. Jackson, "NWMO Background Papers: Technical Methods; Status of Nuclear Fuel Reprocessing, Partitioning, and Transmutation," NWMO, 2003.


[4] L. Johnson and J. Tait, "Source terms for 36Cl in the assessment of used fuel disposal," Atomic Energy of Canada Limited Technical Record, 1992.

[5] J. Tait, I. Gauld and G. Wilkin, "Derivation of Initial Radionuclide Inventories for the Saftey Assessment of the Disposal of Used CANDU Fuel," Atomic Energy of Canada Limited Report, 1989.

[6] G. Faure, Principles and Applications of Geochemistry, Upper Saddle River: Prentic Hall Inc., 1998.

[7] R. Hart and G. Morris, "Crystallization temperatures of uranyl ntirate-nitric acid solutions," Prog. Nucl. Energy III, p. 544, 1958.

[8] T. Chikazawa, T. Kikuchi, A. Shibata, T. Koyama and S. Homma, "Batch Crystallization of Uranyl Nitrate," J. Nucl. Sci. Technol., vol. 45, no. 6, pp. 582-587, 2008.

[9] T. Todd, "Spent Nuclear Fuel Reprocessing," in Nuclear Regulatory Commission Seminar, Rockville, 2008.

[10] R. Jubin, "Used Fuel Reprocessing," in CRESP Nuclear Fuel Cycle Course, 2011.

[11] Rousselet-Robatel, "ROUSSELET ROBATEL MODEL LX MULTISTAGE CENTRIFUGAL EXTRACTOR," [Online]. Available: http://www.rousselet-robatel.com/products/pdfs/multistage-centrifugal-extractor-operatingprinciple. pdf. [Accessed 2012].

[12] P. A. Haas, R. D. Arthur and S. W.B., "Development of Thermal Denitration to Prepare Uranium Oxide and Mixed Oxides for Nuclear Fuel Fabrication," Oak Ridge National Laboratory, Oak Ridge, 1981.

[13] S. Jeong, S. Park, S. Honge, S. C.S. and S. Park, "Electrolytic production of metallic uranium from U3O8 in a 20kg batch scale reactor," J. Radioanal. Nucl. Chem., vol. 268, no. 2, pp. 349-356, 2006.

[14] T. Takata, Y. Koma, Sato, Koji, Kamiya, Masayoshi, A. Shibata, K. Nomura, H. Ogino, Koyama, Tomozo and S.-i. Aose, "Conceptual Design Study on Advanced Aqueous Reprocessing System for Fast Reactor Fuel Cycle," J Nucl Sci Technol, vol. 41, no. 3, pp. 307-314, 2004.

[15] R. Herbst and M. Nilsson, "Standard and advanced separation: PUREX processes for nuclear fuel reprocessing," in Advanced Separation Techniques for Nuclear Fuel Reprocessing and Radioactive Waste Treatment, 2011, pp. 141-173.

This web page last updated June 21, 2021

Home Energy Nuclear Electricity Climate Change Lighting Control Contacts Links