The percentage of uranium 235 is increased, if necessary, in an enrichment plant. Various methods of enrichment exist, but, at the industrial level, France has only used gaseous diffusion. We present this technology and three others that the CEA has studied or is still studying.

GASEOUS DIFFUSION

Uranium arrives at the plant in the form of solid UF6. It is vaporized and advantage is taken of the difference in the molar masses of the three isotopes to separate them selectively by passage of UF6 through a porous wall, a “barrier.” The lightest isotopes, uranium 235 and 234 pass more easily than uranium 238. Because enrichment by means of a single barrier is very slight, it is necessary to repeat the operation a great number of times.

The elementary unit in enrichment is the stage, which is composed of

--a diffuser containing barriers;

--a compressor which forces the UF6 to pass through the barriers;

--an exchanger which removes the heat generated by the compressor.

The stages are placed in a series. The part of the flux that passes through the barrier goes to the following stage; the part that does not pass is directed towards the lower stage. The stages are joined into a whole of ten to twenty units that constitute a group. Several groups constitute the cascade.

UF6 is introduced into the center of the cascade. The UF6 that has been enriched in uranium 235 is withdrawn at one end and the depleted UF6 at the other.

France studied the enrichment procedure in laboratories and pilot units at Saclay, at Grenoble, and at Pierrelatte/Tricastin. Two plants were put into service, a “military” plant at Pierrelatte and the Georges Besse plant at Tricastan belonging to Eurodif-Production. The Pierrelette plant ceased operation in 1996. Georges Besse is scheduled to cease operation in 2012-13 after a yet-to-be built centrifuge plant has been in operation for about five years.  Gaseous diffusion's heavy use of electricity is dooming the technology [NucF 22.xii.03].

This procedure uses the effect of centrifugal force on gaseous UF6 contained in a recipient turning very fast around an axis. The intensity of the centrifugal forces being proportional to the mass of the bodies, the atoms of uranium 238 are chased to the edge. The gas in the middle of the recipient is enriched in uranium 235 while the gas near the wall is depleted. The rate of enrichment is low. Thus, a very great number of centrifuge units are needed.

France studied the procedure at Saclay and at Pierrelatte, but the studies were interrupted at the end of 1983. However, the studies recommenced.  The four-year contract between the CEA and the French government signed January 2001 specified that the CEA would abandon work on Silva by the end of 2003 to concentrate on ultra-centrifugation, which Cogéma had come to regard as the replacement for gaseous diffusion enrichment. France intended to get a firm grasp on basic centrifuge technology in preparation for acquiring advanced technology from another source [NucF 5.ii.01]. 

In November 2003, Urenco and Cogéma announced that they had signed an agreement under which Areva, parent company of  Cogéma, would acquire 50% ownership of Urenco's Enrichment Technology Co. (ETC).  This company is to furnish Areva with technology and actual centrifuges for a centrifuge plant at Tricastan.  The plant, Georges Besse II, is scheduled to begin operation at reduced capacity in 2009.  At full capacity it will produce 7.5 million SWU/yr.  The agreement had to be approved by the European Union; and an intergovernmental agreement had to be signed by Germany, the Netherlands, the United Kingdom (the three countries that own Urenco), and France [NucF 22.xii.03, NucF 16.ii.04]. The final agreement was signed July 3, 2006 [NucF 17.vii.06].

Enrichment takes place during repeated exchanges between an aqueous phase with trivalent uranium and an organic phase with tetravalent uranium. Solutions were prepared using natural uranium and not UF6. Four elements constituted the mixture: a solvent, mercury, hydrochloric acid, and uranium. The procedure could not be used for the production of highly enriched uranium.

The program took place at Saclay, at Grenoble, and at Pierrelatte. The CEA shut it down in 1988. The most obvious reason was the more promising future of the Silva process.

Silva is based on the difference in the frequencies at which atoms of the isotopes of a given element are ionized. A laser beam illuminates vapor of uranium metal or uranium metal alloy and selectively ionizes the atoms of uranium 235, removing an electron from each and leaving them with a positive charge. The uranium 235 is then collected on negatively charged plates. The uranium 238, still neutral, condenses on collectors on the roof of the separator. An enrichment line has two distinct parts:

A chain of lasers composed of two types of lasers: copper vapor lasers that emit photons of fixed energy and organic dye lasers that emit photons with a wavelength tuned to the resonant energy of uranium 235. The copper vapor lasers optically “pump” (increase the energy of) the dye lasers. Liquid dyes, diluted in another liquid such as ethyl alcohol, determine the “color” of the colored lasers.

One or more separators: In each separator, free electrons bombard the uranium metal and vaporize it. The uranium 235 is ionized, and the flux of uranium 235 and that of uranium 238, which contains most of the impurities, are condensed and collected. The source of the free electrons that vaporize the uranium after it enters the separator is an electron gun that is located under a crucible that holds the uranium. A magnetic field curves the stream of electrons released by the gun.

In 1996 and 1997 the CEA and Cogéma conducted a series of demonstrations of various aspects of the process: the integration of the process: the integration of the process (the big separator Aldébaran was to produce several kg of uranium enriched to commercial level); the lasers; the collection of enriched uranium; the uranium evaporator. [défi ii.97]. According to a February 1998 article, these tests and later research were to furnish Cogéma with all the elements necessary for making "an industrial decision" in 2001 [défi ii.98].

 A new prototype separator, Menelas, was to be inaugrated in 1999, but the budget of 434 million francs for Silva in 1999 was the lowest since 1995. In January 2001 the French government and the CEA signed a four-year contract that stipulated that the CEA will cease work on Silva by the end of 2003 to concentrate on enrichment by ultra-centrifugation. By the time that the Silva program is closed down "scientific and technical demonstration implying system integration" is to have been accomplished in order for the process to be ready to be taken up again at some future date [NucF5.ii.01]

The CEA completed construction of a pilot installation, Menphis, at Pierrelatte in early 2003.  In November the CEA conducted a demonstration production run in the pilot, which produced more than 200 kg of uraniuim enriched to 2%-3% uranium 235 and a metric ton of depleted uranium.  Menphis featured "a full-scale system integration of separator, atomic vapor production and laser."  According to Patrice Bernard of the CEA, the CEA was so encouraged by the demonstration that it has begun looking into the implementation of the technology on a commercial scale, perhaps in twenty years [NucF 19.i.04].  France's Court of Auditors (Cour des Comptes) has a less positive view of the Silva program.  It charged in its report on public sector spending in 2003 that the CEA wasted money by keeping research on Silva going after the technology's lack of feasibility had become apparent [NucF 16.ii.04].

Enrichment of UF6 Resulting from Reprocessing

The presence of uranium 234 and uranium 236 in the UF6 containing reprocessed uranium (Repu) necessitates overenrichment in uranium 235 of between 0.4% and 0.7% if [Castaing 81-82]. In other words, more uranium 235 is needed in Repu fuel than in standard fuel.

Use of Repu has a major impact on the choice of an enrichment process.

If fuel is to be fabricated using only reprocessed uranium, centrifugation is better suited than gaseous diffusion to enrichment of the uranium, particularly because of centrifugation’s modular installations with relatively small capacity. A given quantity of UF6 containing reprocessed uranium can be enriched quite rapidly without necessarily bringing it into contact with natural UF6. In addition, the modules are easier to cleanse of uranium 234 and 236 than are the units of a gaseous diffusion plant.

In 1994, Cogéma, Eurodif, and Urenco signed an agreement concerning a feasibility study for the construction of a unit for enriching Repu by centrifugation at Tricastan. According to Cogéma, the project was put off indefinitely because of the state of the market and existing centrifuge enrichment capacity. However, in November 2003, Cogéma through its parent company Areva, signed an agreement to acquire 50% of Urenco's Enrichment Technology Co . That company is to furnish Areva with technology and centrifuges for a new centrifuge enrichment plant, Georges Besse II..   Cogéma hopes to begin operation at Georges Besse II at 5% capacity in 2007.  One of Georges Besse's "double modules" of 500,000 SWU/yr, will be "devoted in principle to enrichment of reprocessed uranium" [NucF 22.xii.03].