Lessons that can be learnt from an abandoned project

30 October 2000

The AVLIS programme was cancelled in 1999, but the technology development process has produced several successful spin offs.

One of the hardest scientific challenges has been to effectively and inexpensively separate out a desired isotope of a chemical element. Traditionally, this was carried out by gaseous diffusion or gas centrifuge.

In the early 1970s the US's Lawrence Livermore National Laboratory began experiments with lasers to enrich uranium. Over the years, funding of about $2 billion was invested to develop the technology at Lawrence Livermore By the end of the decade, important progress had been made (see panel). In 1985, the Department of Energy (DoE) selected AVLIS as having the best potential to provide a low-cost, environmentally sound method to enrich uranium. The DoE's goal was to replace, in an orderly way, the ageing and energy-inefficient gaseous diffusion plants in Ohio and Kentucky.

Development of plant scale laser and separator hardware began in 1986, while researchers were continually improving their performance and reliability. The early 1990s marked the first time that tests were carried out using full-sized components in integrated systems producing enriched uranium over many tens of hours. The Energy Policy Act of 1992 transferred the US government's uranium enrichment activities to USEC, which at the time was a government corporation thatsupplied the nuclear fuel industry with enrichment services through gaseous diffusion technology.

In July 1994, after a two-year period during which Livermore's AVLIS activities were on standby, USEC gave the go-ahead for advanced development. The technology was then transferred from DoE to USEC for commercialisation, representing the largest technology transfer in the Laboratory’s history.

Proving the technology

A pilot enrichment plant completed at Livermore in 1997 was the focal point for the AVLIS team of experts, who came from Livermore itself, along with Bechtel, Duke Engineering, Babcock and Wilcox, Lockheed Martin, AlliedSignal, and USEC.

Improvements included making key separation upgrades for better performance and reliability, introducing deformable optics for laser beam uniformity, and substituting compact diode-pumped solid-state lasers for copper lasers to energise the dye lasers. This latter change significantly reduced costs and space requirements.

The pilot plant operated for over 1.5 years, processing several thousand kg of uranium in a series of tests aimed at verifying component performance, operational lifetime, and economics. “Running the pilot facility was not an experiment. This was a 24-hours-a-day demonstration of industrial capability using full-scale hardware,” said Steve Hargrove, the senior scientist supporting the project from the Laboratory.

In tandem with the long-term tests, scientists and engineers began the first stages of designing and licensing a commercial plant.

By December 1997, the pilot plant had achieved its target for separator lifetime, which was 400 hours of continuous operation before needing refurbishment. However, the next three tests during 1998 fell well short of 400 hours due to unexpected component corrosion. The problem was a minor impurity in the uranium.

By March 1999, the pilot demonstrations had verified the projections that the technology could achieve enrichment at costs comparable to those of current US gaseous diffusion, even without the planned pilot improvements. Project managers were preparing final engineering improvements and additional 400-hour enrichment tests in order to address two remaining economic factors: separator lifetime and enrichment efficiency. The goal was to conclusively demonstrate within a few months the basis for projecting that the LIS plant could enrich uranium at costs equal to or below those of all existing technologies.

Development efforts to achieve high enrichment efficiency centred on improving laser beam uniformity and uranium vapour conditions. In several tests the plant achieved its enrichment efficiency goal of 80%. Engineering upgrades were in place to address half the remaining shortfall in forthcoming tests and to meet the 100% target during final design of the commercial plant.

Halting development

However, the tests were interrupted by USEC's decision in June 1999 to halt the development of the technology because a combination of near-term factors limited its funds. These factors included market-driven price declines for enriched uranium, significant cost increases to operate US gaseous diffusion plants, and the need to continue shareholder dividends.

USEC's decision was no reflection on the advantages and capability of AVLIS. For example, AVLIS only uses 5% of the electricity consumed by existing gaseous diffusion plants, and AVLIS facilities would cost less to build than those for other enrichment technologies such as centrifuge technology.

Enriching uranium from its natural level (0.7% U-235) to 3-5% U-235, is achieved in a few passes with AVLIS, a great improvement over the hundreds to thousands of passes required by other processes. This translates into a much smaller plant and production costs substantially lower than those of either gaseous diffusion or gas centrifuge technology.

The system is very compact. A vacuum chamber holding one separator unit produces output equivalent to that of several thousand of the best commercial centrifuges. A commercial AVLIS plant would use 84 enrichment units, compared to more than 150,000 centrifuge machines.

AVLIS would also offer strong environmental, safety and health advantages. Instead of using uranium hexafluoride, the starting material required by other processes, AVLIS uses uranium metal, which is less hazardous. Compared to centrifuge or gaseous diffusion, the laser process requires about 30% less natural uranium ore to produce a comparable amount of enriched product, which also minimises the amount of uranium tailings by about 30%.


Although the AVLIS programme has been cancelled, the work has produced a number of interesting spinoffs. In a study for the National Academy of Sciences and the National Academy of Engineering, Livermore identified over 60 potential spinoffs in various industrial areas.

A promising application is enriching gadolinium for use in BWRs. Fuel rods now contain about 8% natural gadolinium, which is a mixture of seven isotopes. The odd-numbered isotopes, Gd-155 and Gd-157, are interesting because they absorb neutrons to control excess reactivity of burning fuel rods, which translates to more efficient use of the fuel and reduced waste.

AVLIS technology can be used to double the concentration of odd-numbered gadolinium isotopes, so reactor fuel would need only half its current amount of gadolinium, further contributing to fuel burnup efficiency. Enriched gadolinium could be an important constituent in future nuclear fuel designs because it could permit significant increases in fuel burnup time. Enriched gadolinium could also play an important role in MOX fuels that burn surplus plutonium from nuclear weapons. An AVLIS plant for gadolinium would be similar to a uranium enrichment system, but it would only require about one-tenth of the hardware to economically meet market demand projections.

Another product with similar benefits is enriched erbium-167, currently used in its natural form in fuel for PWRs. Test runs at Lawrence Livermore have verfied the enrichment process for both gadolinium and erbium, and have produced kilogrammes of each material.


Lawrence Livermore scientists have proposed using AVLIS technology for another DoE mission, one closely related to the enrichment of natural uranium. The idea is to use AVLIS to recover the energy value remaining in the tailings from several decades of uranium enrichment activities at US gaseous diffusion plants – activities that were undertaken for both commercial and defence purposes. AVLIS could claim up to 60% of DoE's tailings inventory with full cost recovery and generate a profit. It would also provide important environmental and energy conservation benefits.

In all, over 700 million kg of depleted UF6 containing 475 million kg of uranium has been generated by the US government. The material is stored in nearly 60,000 steel cylinders at Oak Ridge National Laboratory and at the Ohio and Kentucky gaseous diffusion plants. Although depleted UF6 does not present a serious radiological threat, it is a potential chemical hazard and is under safe management by DoE's cylinder and maintenance programme. The DoE is considering long-term plans to convert depleted UF6 into a more environmentally acceptable and nonhazardous form (either an oxide or uranium metal) before final disposition of the tailings.

Lawrence Livermore scientists note that an AVLIS plant could convert the tailings to metal (required for the AVLIS enrichment process), enrich the uranium to recover its energy value, and finally convert the leftover tailings to oxide for burial. Livermore projections show that AVLIS can profitably re-enrich about 60% of the tailings inventory (the proportion containing more than 0.24% of U-235). By recovering energy from existing tailings, the AVLIS plant would dispose of over 11 million kg of uranium tailings every year and provide enough fuel to generate some 40GW.

Revenue from enriched uranium produced by the plant, even at prices well below current market levels, would more than pay for both the plant's construction costs and all operating costs. The market is currently at about $80-85 per SWU. Even at $60 per SWU, the DoE could recover all of its costs as well as generate a profit of $2.4 billion over the 25-year life of the AVLIS plant. This profit could then be used to more than offset costs to clean up the remaining 40% of tailings not economical for AVLIS enrichment. Hargrove says: “There’s abviously a lot of value in tailings that is waiting to be exploited.”

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.