Gaseous Diffusion Applications

Much of the impetus for the awakened interest and utilization of inorganic membranes recently came hom a history of about forty or fifty years of some large scale successes of porous ceramic membranes for gaseous diffusion to enrich uranium in the military weapons and nuclear power reactor applications. In the gaseous diffusion literature, the porous membranes are referred to as the porous barriers. For nuclear power generation, uranium enrichment can account for approximately 10% of the operating costs (Charpin and Rigny, 1989]. 

Natural uranium contains close to 0.72% by weight of the only fissionable isotope and more than 99% non-fissionable U, and a trace quantity of U. For uranium to be useful as the fuel to the nuclear reactor, the level needs to reach 1 to 5% (more often 3 to 5%) while most of the nuclear weapons and submarines require a concentration of at least 90% The separation of those uranium isotopes, very similar in properties, can not be effected by chemical means. 

In contrast, the actual separation factor using alumina membranes is only 1.0030 [Isomura, et al., 1969]. In practice, up to thousands of membrane tubes are arranged in a counter-current cascade configuration to achieve the required degree of separation for example, over 1,200 stages required for 3% and over 4.000 stages for 97% even with gas recirculation. 

Concurrently in the U.S., membrane development efforts were sponsored and initiated by the Atomic Energy Commission for gaseous diffusion. Large scale plants were built at Oak Ridge (Tennessee) and subsequently at Paducah (Kentucky) and Portsmouth (Ohio) in the early 1950s with a total capacity of approximately 150 t/d processed. The confidential membranes are believed to be made of nickel and high-nickel alloys but other membrane materials possibly have also been considered. 

In addition to the U.S. and France, other countries such as the Soviet Union, China and England were also involved in using presumably inorganic membranes for its gaseous diffusion opierations although little has been documented. Ceramic membranes were also made by the anodic oxide process (to be discussed later in Chapter 3) in Sweden for military and nuclear applications. 

---

The next stage of development went to polymer chemists and development engineers, as the expertise of Roy Plunkett was really in fluorine chemistry. The first great application was in the separation of the isotope U-235 from U-238 by gaseous diffusion of UFe to make atomic bombs, as the gas uranium hexafluoride was exceedingly corrosive and destroyed conventional gaskets and seals. PTFE was just what was needed to form the diffusion membrane, as it was not attacked by fluorine. When peace returned, PTFE registered the trademark of Teflon in 1944. 

Fluorine is used in the separation of uranium, neptunium and plutonium isotopes by converting them into hexafluorides followed by gaseous diffusion then recovering these elements from nuclear reactors. It is used also as an oxidizer in rocket-fuel mixtures. Other applications are production of many fluo-ro compounds of commercial importance, such as sulfur hexafluoride, chlorine trifluoride and various fluorocarbons. 

The interest of using fine-pore thin-film ceramic or metal membranes for isotope separation (e.g. uranium) is still apparent even after years of production practice [Miszenti and Nannetti, 1975 Sumitomo Electric Industry, 1981]. Isotopes other than uranium, such as those of Ar or Ne [Isomura, et al., 1969 Fain and Brown, 1974], can also be separated by gaseous diffusion. The membrane materials having been successfully tested for these specific applications include alumina, glass and gold. 

Gas flow method. The above methods have a common characteristic. That is, they do not discriminate between the active pores and the "passive" pores. For separation applications, only the "active" pores arc most relevant. A nondestructive technique for inferring the pore size distribution of a membrane is the gas flow method. Having been developed and tested at Oak Ridge Gaseous Diffusion Plant [Fain. 1989], it departs greatly from the above methods in that the pore size distribution is based on gas flow... 

For process engineering calculations it is almost inevitable that experimental values of D or f), even if available in the literature, will not cover the entire range of temperature, pressure, and concentration that is of interest in any particular application. It is, therefore, important that we be able to predict these coefficients from fundamental physical and chemical data, such as molecular weights, critical properties, and so on. Estimation of gaseous diffusion coefficients at low pressures is the subject of Section 4.1.1, the correlation and prediction of binary diffusion coefficients in liquid mixtures is covered in Sections 4.1.3-4.1.5. We do not intend to provide a comprehensive review of prediction methods since such are available elsewhere (Reid et al., 1987 Ertl et al., 1974 Danner and Daubert, 1983) rather, it is our purpose to present a selection of methods that may be useful in engineering calculations. 

Unsteady-state Conditions. Arnold (2) has integrated the Maxwell-Stefan equation for gaseous diffusion in the case of the semi-infinite column, or diffusion from a plane at which the concentrations are kept constant into a space filled with gas extending to infinity, both for vaporization of a liquid into a gas and absorption of a gas by a liquid. It is possible that the resulting equations could be applied successfully to liquid diffusion for similar circumstances, provided that an assumption analogous to Dalton s law for gases can be made and that D is assumed to remain constant. The direct application to extraction operations of such equa-... 

H. F. Henry et al., "Criticality Data and Nuclear Safety Guide Applicable to the Oak Ridge Gaseous Diffusion Plant, K 1019, 5th Rev. (May 1959). 

The original solid angle method was developed at the Oak tUdge Gaseous Diffusion Plant and reported in 1956 (over 25 years ago). The area of application was initially for small planar arrays of.highly enriched uranium solutioa... 

H. F. HENRY, J. R. KNIGHT, and C E. NEWLAN, General Application of Theory of Neutron Interaction, K-1309, Oak Ridge Gaseous Diffusion Plant (Nov. IS, 1956). 

In natural uranium ores, the fraction of the atoms of the fissile isotope is about 0.72%. For many commercial applications, like production of fuel for light water reactors or several types of research reactors and other nuclear functions, its fraction must be increased, that is, isotope enrichment is carried ont. The main isotope separation methods, or isotope enrichment processes, ntilize the small differences in between the mass of U-235 and U-238. The two major commercial methods that have supplied most of the enriched uranium to date, gaseous diffusion and gas centrifuges, use the only gaseous compound of nraninm, nranium hexafluoride (UFg), as the feed material. Both methods utilize the difference between the mass of UFg (349 Da) and UFg (352 Da) where the mass ratio difference that is 0.86%. The product and tails of the enrichment process are also with the same chemical form, but the isotope composition of the material is altered in the enrichment process. Schematic diagrams of the principle of operation of these methods can be found on the web and in many textbooks, so will not be shown here. 

As it turned out, the properties of Teflon were ideal for an immediate and important application in the development of the first atomic bomb. Uranium hexafluoride (LIF5), which was used to separate fissionable by gaseous diffusion (see... 

---