Report of the Fundamental Classification Policy Review Group

Appendix G to the
Report of the
Fundamental Classification
Policy Review Group


Report of the
Nuclear Materials Production
Working Group


January 15, 1997


Dr. Paul T. Cunningham, Chair
Los Alamos National Laboratory






The Nuclear Materials Production Working Group (NMPWG) comprised seven technical experts and four support staff (see Annex A). To begin its portion of the task, the NMPWG looked at the scope of activities within the nuclear materials arena and determined the boundaries of that scope (see Annex B). The boundaries set by our working group expanded from beyond nuclear materials to include the production and processing of all materials used in a nuclear weapons package but did not move into the manufacturing of those materials into weapons components (this was considered to be part of another working group's responsibilities). The boundaries were expanded beyond nuclear materials to include all weapons materials, even at the risk of overlap with other working groups, to ensure that areas relevant to the classification review were not overlooked.

The NMPWG then developed principle statements that were submitted to the Review Steering Committee and consequently considered to be the building blocks of our classification recommendations. The principles developed (both General and Nuclear Materials Production-Specific) were used as guidelines when determining future classification policy recommendations in Nuclear Materials Production areas. The General policy statements developed by the NMPWG are not meant to supersede or contradict those developed by the Steering Committee; they are specifically for the NMPWG.

A. General Principles

B. Nuclear Materials Production-Specific Principles

Once principles were identified, the next step was to develop a list of stakeholders. Stakeholders were invited to our working group meetings and, in addition to presentations by some, round-table discussions were initiated regarding current classification policy and what should be kept or changed in the DOE classification system (see Annex C). The positions of those stakeholders were considered, discussed at length, and working group members given writing assignments for white papers. The working group members then compiled and revised the papers, verifying technical content and eliminating conflicting policy statements. The resulting product of the NMPWG forms the body of this report.


A. Uranium Isotope Enrichment

1. Issue

Identify classified information related to gaseous diffusion, gas centrifuge, and laser isotope separation for fissile materials that could and/or should be declassified.

2. Background

Production of fissile materials is a necessary step toward construction of a nuclear weapon, but by itself is insufficient to achieve such a weapon. Both heavy (usually fissile) and light elements are of interest to weapons designers, depending on the sophistication of the design. Enrichment of naturally occurring isotopes is one method of obtaining such materials without use of reactors or other neutron sources. Also, reactors could be used for fissile isotope production techniques using suitable target materials with separation of the transmuted materials following irradiation.

Three national security risks are related to isotope separation. Small-scale, potentially clandestine, production of fissile isotopes could be of interest to nonnuclear states or terrorists. Production of hydrogen and lithium isotopes for thermonuclear weapons production could enhance the capabilities of established nuclear-capable states. Finally, isotope separation technologies could facilitate an increase in scale or reduction of cost in an existing weapons program.

Since the Manhattan Project, the U.S. has controlled selected information on the acquisition of fissile material by means of isotope separation. Isotope separation, particularly uranium enrichment, is one method of obtaining concentrated fissile material for nuclear weapons. The theory of isotope separation by numerous methods is widely documented in the open literature. Moreover, designs of uranium enrichment technology that specify construction materials and general dimensions are also available for the early developmental stages of several technologies. Even with this broad availability of basic information, dedicated national efforts have been required to achieve practical uranium enrichment capabilities. The protracted efforts of the nations that have succeeded in achieving enrichment capabilities have shown that uranium isotope separation is a demanding technology. Each developer has had to rediscover and master many of the details to yield practical technology.

Because uranium enrichment technologies are also used to produce low-enriched uranium (LEU) for nuclear electric generation, there are commercial incentives for nations, and more recently private companies, to develop and build uranium enrichment plants. Nonetheless, only eight countries (U.S., United Kingdom, Netherlands, Germany, France, Japan, Russia, and the Peoples Republic of China) have sizable uranium enrichment capabilities. In five of these cases, the commercial use of the technology evolved from defense production plants. A small number of other countries have limited enrichment efforts, presumably for military programs. The technology can be mastered by concerted multiyear efforts; however, the scale of these efforts makes concealment difficult. The recent experience with Iraq is one case where a country nearly had a useful enrichment capability before worldwide recognition.

Several technologies exist for isotope separation. Gaseous diffusion and gas centrifuge are currently well-established, large-scale techniques used to enrich uranium. Other techniques using calutrons, plasma separation, chemical exchange, vortex separation, and gas nozzles have to date not been used for large scale production, but would work on a small scale where economics are not a dominant factor, as exemplified by the South African experience.

The classification and export control regimes have been quite successful in restricting the flow of U.S. technology to other countries. Only many years after the U.S. declassified an isotope separation technology (electromagnetic separation) was information on this technology exploited by Iraq as a significant part of its clandestine attempt to develop nuclear weapons. The nations that form the Nuclear Suppliers Group have been supportive of effective controls based in part on the protection of sensitive technologies. Because of the large investments required by nations for their commercial nuclear fuel enrichment plants, there are also commercial incentives to maintain control of the technology.

3. Current Policy

Technical information that would enable other nations to attain or improve on a nuclear weapon's capability is of importance to the national defense and security of the U.S. The Atomic Energy Act (AEA) requires the classification of atomic energy information including production of fissile materials. DOE policy is to classify the key technical and operational information for methods of isotope separation that have a reasonable potential for the separation of "practical" quantities of SNM or other isotopes of military interest.

To be considered "practical," a technique must produce one or more critical masses in approximately one year. Within the U.S., this test has been applied to all methods described above. Details of performance, design, and certain concepts are presently considered Restricted Data (RD). Scientific bases and published data concerning nonfissile isotopes are considered unclassified information. The detailed application of nonfissile elements is also unclassified if no method of separating fissile isotopes is revealed. Data relating to the operations of enrichment facilities remain RD in many cases to protect historical production data and not because of the potential for revealing sensitive technology.

The current uranium enrichment classification policy is based on the protection of core technologies in the design, manufacturing, and operation of each process. The core technologies are the key features that enable practical or efficient use of the process. The classification guides are technology oriented and highly specific; most of the guides are classified documents themselves. There are seven DOE classification guides for fissile isotope separation that provide headquarters guidance for field application. These guides cover all uranium and plutonium isotope separation technologies. For new proposed isotope separation technologies for fissile material production, the DOE policy considers the concepts and small-scale experiments as unclassified until sufficient evidence accumulates to indicate the capability for making practical quantities of fissile materials. This policy promotes research on laboratory-scale isotope separation with major scientific benefits, avoids the DOE giving unwarranted credibility to proposers of new enrichment techniques, and is efficient in the use of government resources. The need to classify nongovernmental isotope separation research because of its usefulness to fissile material production has occurred only rarely.

4. Revised Policy

Based on the lessons learned over the past 50 years in limiting information flow to potential proliferants, the current policy for classification of uranium enrichment information should remain intact, with certain refinements, as an essential element of the U.S. policy to control the availability of weapons-usable materials.

The potential for nuclear proliferation persists even for so-called obsolete technologies. The world is changing, and advanced enrichment technologies are now being commercialized. The need to remediate existing sites and to allow construction of new sites increases the need to reduce the amount of RD to only the core data needed to actually proliferate the technology. Reduction of the amount and complexity of RD developed as a result of its derivative connection to historic production data will offer substantial relief to waste treatment and will allow increased public access. It should also enhance the actual protection of sensitive information by allowing clearer guidelines to be developed, thus reducing internal and site-to-site inconsistencies.

After the proposed privatization of the U.S. Enrichment Corporation (USEC), the active technical experts in production scale enrichment will be almost entirely in the private sector in the U.S. The restructuring of the U.S. enrichment enterprise, which started with the passage of the Energy Policy Act of 1992, has triggered a fundamental change in the way DOE will need to address future classification actions regarding uranium enrichment. The Nuclear Regulatory Commission (NRC) will assume responsibility for regulating USEC's handling of classified information. The continued DOE responsibility for decommissioning and decontamination of the three enrichment sites and the centrifuge development facilities does delay the classification implications of restructuring for a number of years.

5. Recommendations

a. General Recommendations

The working group recommends the following:

b. Specific Isotope Separation Technology Recommendations

Gaseous Diffusion


Identify gaseous diffusion information that may be suitable for declassification.


Gaseous diffusion is the current production leader for LEU and was the historical method used to produce the majority of fissile uranium for weapons in the U.S., Britain, and France. Russia also used this technology before switching to the gas centrifuge.

Current Policy

A significant number of gaseous diffusion documents remain classified to protect production rates for intermediate uranium enrichment prior to 1980. If these production rates were declassified, there would be a substantial decrease in the classified document holdings at the gaseous diffusion sites.

Gaseous diffusion is a relatively unattractive technology for a proliferator because of its high energy utilization and the physical size required for the thousands of stages of equipment required to produce highly enriched uranium (HEU). The current classification policy does not cause undue restrictions on the commercial use of this technology for production of low-enriched fuel.

Revised Policy

Because of the narrow area of classified information, there does not appear to be a significant benefit to declassifying the core technologies relative to the moderate proliferation risk that might be created by such an action. A proposal to declassify some features of the gaseous diffusion membrane (i.e., the barrier) to facilitate the use of this technology for producing inorganic membranes has been considered by DOE. This classified technology has been examined by cleared personnel from commercial firms and judged to have significant advantages for the manufacturing of nonuranium, enrichment-capable membranes for a large number of commercial applications. Some companies have stated that they would consider operating a classified manufacturing facility if the products would be sold as unclassified. A number of studies have evaluated the potential for discovery of key details of the classified manufacturing process by detailed examination and reverse engineering of the products.

The overall conclusion is that reverse engineering of the barrier manufacturing process is difficult, if possible at all; however, some continue to argue on theoretical grounds that manufacturing processes should remain completely protected. Recently, it has been proposed that controls could be imposed on manufactured products to ensure protection of classified manufacturing details. This proposal appears to offer a positive method of allowing productive use of the technology with very limited risk. A major benefit of this approach is that neither specific information about the uranium barrier nor barrier manufacturing information would be declassified.


The working group recommends the following:

Gas Centrifuge


Identify gas centrifuge information that may be suitable for declassification.


Gas centrifuge technology is used in uranium enrichment production facilities in Great Britain, Germany, Netherlands, Japan, and Russia. The U.S. gas centrifuge program was terminated by DOE in 1985 as part of a selection process for an advanced enrichment technology. Several other countries are known to have investigated, and in some cases deployed, centrifuge capability. The Nuclear Suppliers Group has worked to restrict the flow of design information, special materials, and key components to potential proliferants.

As part of the privatization of the U.S. enrichment business, a group of electric utilities has teamed with a European centrifuge company, the Uranium Enrichment Corporation (URENCO), to build a centrifuge enrichment plant in Louisiana. The company, Louisiana Energy Services, Inc., has applied for an NRC license to build an enrichment plant with centrifuges imported from Europe. This has raised the issue of the classification of an imported technology item. An agreement between the U.S. and the European countries involved (Britain, Germany, and the Netherlands) will ensure that all imported technology is treated in the U.S. as RD, exempt from the AEA provisions for exchange of RD with foreign nationals; however, under the agreement, any new RD generated in the U.S. including modifications to centrifuge designs, will be subject to all AEA provisions.

The general issue of importing technology that is classified in the U.S. is addressed more fully in a later section.

Current Policy

Because gas centrifuge technology is known to be highly proliferable, DOE has maintained strict classification on the technology following the program termination in 1985. Minor declassification actions have occurred to provide access to information for worker health studies and to facilitate use of related technology on other programs.

The technological feasibility of gas centrifuge enrichment of plutonium has been evaluated. The existing classification for gas centrifuge technology should be retained with alteration.

Revised Policy

In general, strict classification controls should be continued on this technology. Currently there are three separate classification guides on gas centrifuge technology and DOE is considering a consolidation into one guide. Some of the "fact of" associations that are now classified have complicated the cleanup and reuse of extensive centrifuge facilities for other missions. These facilities are now prime candidates for reuse by private firms as part of DOE's privatization initiatives.


The working group recommends the following for the gas centrifuge:

Declassifications should be evaluated for some of the "fact of" associations to facilitate the cleanup, waste management, and reuse of former centrifuge facilities.

Laser Isotope Separation


Recent laser techniques have the potential to compete with existing enrichment technologies for commercial and defense-related isotope production. Laser-based processes using molecular and atomic forms of uranium, plutonium, and nonfissile isotopes have been developed. The selective process can be in the form of disassociating a molecule, inducing a selective chemical reaction, or altering the charge state of an atom. Even the use of lasers to impart momentum to selected atoms in an atomic beam has been attempted. These techniques also have been studied for plutonium purification in producing weapons-grade material from nonweapons-grade materials. All have been successful to varying degrees. Molecular isotope separation (MLIS) development was approaching industrial scale for the production of both uranium and plutonium isotopes, but further development and deployment ceased in the mid-1980s. Presently the U.S. is moving toward commercialization of LEU production for light-water power reactor fuel using the atomic laser isotope separation (U-AVLIS). Such facilities are expected to be in operation by the turn of the century.

France and Japan have declared similar ambitions. The application of laser separation to other isotopes of interest to the industrial and medical communities is also proceeding. Although economically driven, some of these techniques can be used for much smaller scale covert separation installations that may be difficult to detect and that require little in the way of obvious nuclear technology support such as nuclear power generators or experimental reactors. In the future, other techniques undoubtedly will be developed.


The working group recommends the following:

B. Stable Light Isotopes

1. Issue

To what extent should details of the production or separation of stable light isotopes used in the manufacture of nuclear weapons be protected? (Tritium production is achieved by transmutation and is treated elsewhere in this document.)

2. Background

The technologies employed in separating light isotopes are generally unclassified and many are practiced commercially. These include thermal diffusion, distillation, chemical exchange, laser methods, and derivatives of mass spectroscopy. Deuterium is readily separated from single-mass hydrogen by electrolysis or by distillation of a hydrogenous compound. Other stable elements have been isotopically separated for weapons applications, but these uses are primarily for research purposes. Lithium-6 is used in thermonuclear weapons and has been obtained from natural lithium on an industrial scale by taking advantage of slight solubility differences of different isotopes.

3. Current Policy

Generally, the methods and operating parameters of light isotope separation are unclassified. Some features of the separation of lithium isotopes by DOE-sponsored facilities remain classified.

4. Revised Policy

Classification guidance for lithium enrichment was recently reviewed and revised. Remaining classified information relates to processing details; declassification of this information could assist in development of thermonuclear weapons. In further evaluating the trade-off of continued classification and release of information, a risk/benefit analysis(1)

was performed. This evaluation demonstrated the utility of a Multiattribute Decision Theory (MADT) approach and concluded that there was no clear driver for further declassification.

5. Recommendations

Enrichment technology related to any of these light isotopes should be unclassified, with the exception of certain details concerning the Lithium-6 enrichment process selected by the U.S. for weapon-materials production.


1. Issue

Most information about the technologies, facilities, and process steps used in reactor production of tritium, plutonium, and special isotopes has been declassified. Should technical details of the fabrication of fuel and targets used for nuclear material production (in reactors or accelerators) also be declassified?

2. Background

Savannah River Site (SRS) reactors used as the driver fuel an alloy composed of HEU and aluminum. This alloy was cast into billets, surrounded with an aluminum alloy, and then coextruded through dies to produce long, hollow cylindrical tubes that had a thin circular region of fuel surrounded, on the inner and outer diameters, by 0.03 inches of aluminum cladding. The empty inner region allowed coolant to flow through the tubes during irradiation to maintain low fuel temperatures.

Fuel used in the most common type of research reactor, the low-temperature Materials Test Reactor, is similar. This reactor design uses a uranium-aluminum alloy, or a dispersion of uranium oxide or silicide in an aluminum matrix, also clad with aluminum. The SRS design and process differs primarily in the geometry of the fuel, making it suitable for the specific design of large, high-throughput, low-temperature materials production reactors. The process used to coextrude billets into thin fuel regions while simultaneously extruding the cladding used at Savannah River is unique in the reactor fuel fabrication community.

The SRS operation used essentially the same process to fabricate targets for tritium production. Here, a lithium-aluminum alloy was cast into billets, surrounded with cladding material, and coextruded into target tubes that were nested, together with the fuel tubes, into assemblies for irradiation. The lithium-6 absorbed neutrons and was partly converted to tritium. At the low temperatures of SRS reactor operation, tritium was retained by the alloy matrix and the target cladding.

Extrusion of aluminum alloys into a variety of shapes, many more complex than the ones used for the SRS reactors, is common in industry. Coextrusion of two types of aluminum-based material is relatively rare, but has seen commercial application.

The sensitivity of the coextrusion technology centers on the ability to produce uniform cladding and fuel content over a 10-foot length of active fuel or target. The most important process parameters are the preheating temperatures for extrusion dies, die geometries, lubricants, and rates. Control of these parameters becomes increasingly difficult when the physical characteristics of the fuel or target and the cladding diverge. At Savannah River, the usual fuel enrichments ranged from 60% to 93% uranium-235; thus, only about 10% of the fuel volume was filled with uranium. Greater difficulty was encountered and process control was required in tests that used lower-enriched uranium fuel, which required more uranium to make up the total requirement for the fissile uranium-235. These problems were largely avoidable through the use of uranium oxide or silicide, dispersed in solid aluminum, which again behaved (in coextrusion) very similarly to aluminum for fuel enrichments down to about 35 weight percent.

Uranium-aluminum fuel technology, with aluminum cladding, is widely available for research reactors and is the subject of considerable unclassified research in support of U.S. nonproliferation policy. As part of this nation's nonproliferation policy, the Reduced Enrichment for the Research and Test Reactors (RERTR) program has endorsed and encouraged the use of LEU, or uranium enriched to less than 20 weight percent in uranium-235, in research and test reactors. Thus, a great deal of work has supported the preparation of fuel regions that contain a higher ratio of uranium to aluminum, through more difficult casting and fabrication techniques and the use of cermet fuels (dispersion of oxides or silicides in aluminum). Fabrication parameters have been published by independent investigators in both the U.S. and the international community in the open literature. Very little effort is required to obtain key fabrication parameters related to the melting, casting, degassing, and extrusion of these materials.

At the concentrations typically used at SRS, lithium-aluminum alloys for target elements behave very similarly to the aluminum cladding in coextrusion. Once conditions for full-scale production were established, the site was able routinely to maintain quality production of target materials with intact claddings.

In power reactors, which operate at temperatures above the melting point of aluminum, oxide-based or refractory-metal-based fuels and targets are required. For tritium production, ceramics such as lithium aluminate or lithium oxide are the most practical high-temperature target materials. The higher temperatures and the ceramic form of the target material place much greater demands on systems that will retain the tritium product because tritium gas is mobile and can diffuse through most barrier materials and claddings. The critical techniques for target design and fabrication focus on the composition, fabrication, and performance of these barriers.

Lithium alloys and lithium ceramics are widely fabricated and used in industry. However, there has been little industrial need to couple their fabrication with a requirement to retain gaseous materials (particularly hydrogen). Thus, the goals and parameters of target technology remain unique to tritium production.

3. Current Policy

Many technologies for the fabrication of aluminum-based fuels and targets are unclassified and widely disseminated. Information concerning Hanford fuel and target technology, and similar technology used in the early campaigns at SRS, is unclassified. Technology supporting the RERTR fuel programs has been aggressively disseminated by the U.S.

Currently, certain parameters of the SRS fabrication processes are classified CRD to protect the specific large-scale coextrusion process. These parameters include times, temperatures, or pressures used in melting, casting, outgassing, and extrusion for both fuels and targets. Components, geometries, and feed materials (including the aluminum alloys used for the fuels and targets and claddings) are unclassified.

Technology for tritium targets suited to high-temperature reactors is covered under separate guidance. Early N-Reactor technology from Hanford, using a lithium aluminate ceramic and specialized cladding techniques, has been declassified. Improved techniques, including those related to targets that would be compatible with light-water or pressurized-water power reactors, are classified CRD to prevent the spread of technology to potential proliferants.

4. Revised Policy

Both the aluminum-based fabrication techniques for low-temperature SRS reactors and the advanced techniques for high-temperature tritium targets are candidates for possible declassification. Because of their differences, they are evaluated separately.

a. SRS Fuel and Target Technologies

Declassification of these technologies is supported by the following points:

Risks from declassification of these technologies are very small. Coextrusion of long, thin fuel tubes and tritium-producing targets is unique to the SRS reactors. If this reactor design were to be chosen by a proliferant nation, declassification would make the engineering development of the industrial-scale process less difficult. However, other simpler designs, such as those used in materials test reactors, are accessible and much less expensive.

The "wholesale" duplication of the SRS reactor design and technologies would be easier for a weapons state, if it wished to build expanded capacity. Declassification would modestly reduce the amount of testing required to optimize the high-throughput fabrication processes, but would not significantly enhance a weapons state's ability to succeed.

b. High-Temperature Target Technology

Declassification of high-temperature target technology would provide limited benefits, as follows:

Conversely, declassification would pose a significant risk. Development of targets that could be irradiated clandestinely in a power reactor would be made easier for a proliferant nation or a weapons state that wishes to expand its tritium production capability. Light-water reactors are commonly available throughout the world. Technology for gas retention in materials at high temperatures is very specialized and extremely difficult to apply to production-scale targets.

5. Recommendations

The working group recommends that parameters related to preparation and coextrusion of aluminum-alloy and developmental aluminum-dispersion reactor fuel and targets at SRS be declassified. New technologies that may be applicable to higher-temperature-reactor designs, or to the coextrusion of alloys other than aluminum, should be evaluated for classification as they are developed.

Specialized technology related to gas retention in high-temperature targets should remain classified, including critical parameters for the fabrication of high-temperature lithium compounds and barrier materials, bonding to cladding, and any detailed processing techniques for irradiated targets that are related specifically to the barrier materials.


1. Issue

Should design details of the accelerator production of tritium (APT) facility be classified?

2. Background

The U.S. is considering a nonreactor approach to producing tritium. This approach utilizes a high-power, linear proton accelerator that creates neutrons by bombarding a heavy metal target such as tungsten or lead with a high-energy (~1000 MeV) proton beam. Neutrons are produced by spallation and by evaporation of neutrons from the resulting, excited heavy metal nucleus. These neutrons are then moderated to thermal energies and captured in a blanket that surrounds the neutron source. This blanket contains either helium-3 or lithium-6, both of which will capture the thermalized neutrons to form tritium. This tritium can then be extracted from the blanket material, purified, and packaged for shipment to the customer.

All of the technologies of this production process are well known and developed; that is, high-energy proton accelerators, heavy-metal spallation targets, and isotopic separation of tritium from other hydrogen isotopes. These technologies have been developed, tested, and improved through unclassified programs both in the U.S. and internationally. Some of the tritium processing technology has been developed at the SRS and that specific technology remains either classified or UCNI. This classification of Secret Restricted Data information is to protect production/processing quantities and inventories, not because the technology itself is classified.

The unclassified literature abounds with papers and descriptions of these technologies. It is impossible to consider classifying these technologies at this point. Within the APT project the only information that is classified is the annual production rate. There is a question whether even this information should be classified. It is impossible to relate the production rate of a tritium production facility to tritium inventories in any weapon system. Thus even knowing the number of warheads in the active stockpile, and the annual tritium production capability, one cannot determine quantities of tritium in any given weapon system.

The issue for the working group centers on the question of whether the details of the engineering design of the integrated APT facility should be classified. While there seems to be little doubt that it is impossible to classify the individual technologies, the question centers on whether the U.S. would achieve a net benefit if the details of an integrated APT production facility were classified. Would access to the detailed engineering design information, specifications, drawings, etc., allow a proliferant to gain significant schedule and cost benefit relative to tritium production?

3. Current Policy

Current policy states that tritium inventories and production rates are classified information. At this point, no other details of the APT are classified.

4. Revised Policy

The benefits that might accrue under the classification of the APT detailed design include the additional development time and money a potential proliferant would have to dedicate to developing this technology and the resulting possibility that the proliferant may then choose not to pursue an APT option. However, it is questionable whether a proliferant would choose this technology under any circumstances. The current baseline design of the U.S. APT includes a helium-3 blanket for tritium production. The only source of helium-3, in any significant quantity, is the decay of tritium. The U.S. has conserved much of the helium-3 produced by tritium decay over the past few decades. However, to use this option, one must have an initial tritium inventory such as the U.S. inventory produced in heavy-water reactors in the 1950–1988 time period. The lithium-6 blanket concept can be developed for an APT, but again the technology for irradiating lithium-6 in a nuclear reactor is a well-developed technology. Before proliferants would have great interest in producing tritium, they must first have access to fissile materials, presumably generated in a nuclear reactor. If they have such a reactor, it could be used for tritium production using a lithium-6 target.

The neutrons generated in the APT target can be used for things other than tritium production. They can be captured to produce fertile fuel materials. Again, these materials can be produced by neutron capture in a nuclear reactor. The argument can be made that an accelerator-based system might be an attractive option for a neutron source for a proliferant nation. Certainly, at least for a linear accelerator, the construction of such a facility (the accelerator would be 0.6-1.5 kilometers long) will be quite obvious and could, therefore, be readily identified by the U.S. or an international agency such as the International Atomic Energy Agency (IAEA). One can imagine the use of a cyclotron (a circular particle accelerator) to produce the high-energy protons. This system would have a much smaller "footprint" than the linear accelerator and thus might be more difficult to identify. However, one cannot operate cyclotrons with the high beam current planned for APT. Thus, to obtain comparable neutron fluxes several cyclotrons must be constructed, thus requiring construction of a facility with a large, obvious footprint.

The risks that would result from classification of the APT details are those of significant cost, schedule, and resource impact on the U.S. capability to produce tritium. Classification of this technology may adversely affect the goal of tritium production by the end of 2005. The cost to the government of classifying this technology will be considerable. It will limit the pool of available contractors for plant design and construction to those who can work within classified projects. This will tend to reduce competition within U.S. industry for this activity. The impact of classifying this technology will be similar to that experienced in the U.S. Strategic Defense Initiative Office when portions of the Ground Test Accelerator (GTA) program at Los Alamos were classified. The results included significant impact on cost and schedule of the GTA project. The cost and schedule impact within the U.S., if the APT technology is classified, could well be more than a proliferant country could possibly gain by having access to this technology.

5. Recommendations

The working group recommends that APT technology, including the design details of the integrated plant facility, not be classified. It is further recommended that the Fundamental Classification Policy Review Group consider whether there is any value gained in maintaining the classification of U.S. tritium inventory (total) and tritium production capacities. The current classification policy on APT should be retained.

Details of target technology that would enable significant SNM production utilizing an accelerator otherwise engaged in legitimate activities should be protected as RD.


A. Plutonium Processing

1. Issue

Are current classification and information control policies regarding plutonium processing appropriate?

2. Background

During the 1950s, as part of the Atoms for Peace program, the Atomic Energy Commission declassified virtually all information relating to the extraction of plutonium from irradiated reactor fuels.

In more recent times, as a result of increased sensitivity to the risks of proliferation, the DOE has been identifying incremental developments in plutonium extraction refining and waste minimization as UCNI. As discussed elsewhere in this report in greater detail, this use of UCNI has in part been due to historic interpretation, by the Office of General Counsel, that the Atomic Energy Act has no provision for reclassification of information.

3. Current Policy

Topic 4257 of the UCNI Topical Guideline for Nuclear Nonproliferation contains the following guidance:

Unclassified information concerning research and development to acquire applied scientific information including "bench-top" chemistry, that reveals current, new, or improved plutonium processing technology, equipment or methods for primary or secondary processes in sufficient detail to permit implementation of the technology to produce practical quantities of plutonium...UCNI.
Under this guidance, new techniques to extract plutonium from scrap and/or waste streams that allow the process residuals to be disposed of as low level waste must be protected since they also produce a plutonium-rich product. This determination was based on the belief that these techniques could be used to obtain practical quantities of plutonium from scrap and/or waste materials. Review of this determination with plutonium processing experts has led the NMPWG to conclude that this determination is flawed. Specifically,

Therefore, making these processes UCNI is not serving any significant nonproliferation objective but it is preventing the dissemination of useful cleanup technologies to the general nuclear community.

4. Revised Policy

As noted elsewhere in this report, the NMPWG does not believe that UCNI is an appropriate vehicle for protection of truly proliferation-sensitive information regarding plutonium processing. Truly sensitive information needs to be protected as RD. Examples of technologies that should be protected as RD include:

Such developments would permit the construction of small, easily hidden, clandestine SNM processing facilities, which would be a significant proliferation concern.

5. Recommendations

As noted elsewhere in this report, the NMPWG recommends that UCNI no longer be used as a vehicle for controlling access to technical information relating to SNM production. The Office of Declassification should perform a topic-by-topic review of UCNI guides covering production information and identify that information specifically warranting protection as RD. After consultation with the Office of General Counsel, this information, narrowly described, should be redefined as RD. Once all UCNI topics have been so addressed, the UCNI guidelines covering these topics should be withdrawn from use.

If it is found that some of the information currently protected as UCNI legitimately requires protection but cannot be redefined as RD due to prior declassifications, then a change to the Atomic Energy Act specifically authorizing reclassification should be sought. Such a provision could be based on the current provisions for reclassification found in Executive Order 12958, Section 1.8(d). Alternative approaches to allow reclassification as NSI should also be considered.

Further, the Office of Declassification should develop a formal policy regarding compilations including formally declassified information which recognizes the sensitivity of such compilations and allows for them to be protected, either as RD or as NSI, under the provisions of Executive Order 12958, Section 1.8(e).

B. Uranium Processing

1. Issue

Are current classification and information control policies regarding uranium processing appropriate?

2. Background

Because generic applications exist for enriched, normal, and depleted uranium in nuclear weapons, virtually all information relating to uranium processing has been declassified. Recently, the fact that intermediate assay uranium (20% < assay < 90%) has weapons applications was declassified. However, details concerning the specific applications of intermediate assay uranium remain classified.

Additionally, due to nonproliferation concerns, much unclassified information concerning uranium processing is currently being identified as UCNI.

3. Current Policy

As discussed above, much information about uranium and its uses in weapons has been declassified. Information which remains classified includes: quantity-related information, such as number, mass, and shape of produced components; total quantity of uranium worked; etc. Chemical processing steps used to prepare enriched uranium metal and to recover contaminated wastes are generally unclassified, as are many details of the design and construction of the equipment used to perform the chemical processes. Similarly, the mechanical steps involved in uranium metal fabrication are generally considered unclassified.

Current policy regarding the use of intermediate enriched uranium in weapons, while acknowledging that it is used, allows no further elaboration as to where in the weapon this material is found. This is causing some difficulty as it requires classification of otherwise nonsensitive information associated with weapons dismantlement.

Current UCNI guidance for uranium processing is found in the UCNI Topical Guide for Nuclear Nonproliferation, TG-NNP-1. Specific topical guidance mirrors that for plutonium, which is described in an earlier section.

4. Revised Policy

A policy which permits the association of intermediate assay uranium with a general location in unspecified weapons, without elaboration, would resolve the current difficulties. The NMPWG believes that this declassification would cause no harm to the national security.

Further, as noted elsewhere in this report, the NMPWG does not believe that UCNI is an appropriate vehicle for protection of truly proliferation-sensitive information regarding SNM processing. That information which is truly sensitive needs to be protected as RD. Examples of technologies that should be protected as RD include:

Such developments would permit the construction of small, easily hidden, clandestine SNM processing facilities, which would be a significant proliferation concern.

5. Recommendations

The NMPWG recommends that a policy which permits the association of intermediate assay uranium with identified types of weapons components and subassemblies, without further elaboration, be adopted. Other than this specific declassification, the NMPWG feels that existing classification guidance regarding uranium processing should be retained.

As noted elsewhere in this report, the NMPWG recommends that UCNI no longer be used as a vehicle for controlling access to technical information relating to SNM production. Recommendations relating to UCNI found elsewhere in this report apply equally to uranium processing.

C. Tritium Processing

1. Issue

Should information about tritium processing technologies associated with the weapons program remain classified, when the same technologies and similar applications in the U.S. and international fusion programs are unclassified?

2. Background

Tritium is available on the commercial market from Canada as well as France for such applications as exit signs and remote runway lights. General information on how to produce and process tritium is readily available and well known. Most of the commercial tritium available in the world is produced by the Canadians as a by-product from their heavy-water power reactors. The Canadians recover 1-2 kilograms per year from this source. Much of this tritium will be used to support the international development of fusion energy.

Virtually all tritium processing technologies associated with DOE's weapons program, other than reservoir loading operations, are currently being developed and applied by the international community in support of fusion energy programs. The fusion community is planning to eventually produce its own tritium supply by "breeding" tritium. The U.S. DOE has been an active participant in these programs and has been active in transferring tritium technology from defense applications to the fusion energy applications. The proposed International Thermonuclear Experimental Reactor (ITER)—a large-scale, fusion energy engineering demonstration project—will develop and install tritium processing and storage systems that will handle kilogram quantities of tritium. ITER's tritium systems will involve state-of-the-art technologies for tritium purification, storage, isotopic separation, and tritium measurement and monitoring.

DOE's support for both defense and fusion programs has often led to major inconsistencies and confusion regarding the classification/declassification of tritium processing information. Basic tritium technology information from a "defense" facility may be considered Secret, whereas the same information from a "fusion" support facility will be treated as unclassified and allowed to be freely disseminated in the open literature. These inconsistencies have often led to classification situations that require significant time and expense to unravel and resolve, consume valuable resources, damage credibility, and result in little or no protection of the information perceived to be sensitive.

3. Current Policy

The current classification guidance for tritium processing states that significant information concerning large-scale tritium processing is SRD. It also states that tritium extraction, separation, storage, and pumping technologies are well known; however, details that reveal operational parameters or sequences of parameters, that give plant tritium or deuterium quantities for weapons, or that reveal ratios of deuterium and tritium are SRD. Such data include times, temperature, pressure, and assays. This policy and its interpretive site guides have led to the classification of virtually all tritium processing systems and technology-related information at the production sites. This practice has severely limited the exchange of valuable tritium technology between the weapons and fusion programs.

To overcome some of these limitations, the national laboratories have developed special classification guides that have essentially declassified many areas of tritium processing technology for use in the fusion program. This has led to major inconsistencies within the DOE complex, whereby tritium technology information at one site may be treated as SRD, while at another site it is treated as unclassified and is published in the open literature. For instance, technology for hydrogen purification using palladium diffusers or membranes and for tritium storage on metal hydrides is readily available and used all over the world. However, many details about applications at weapons production sites are classified even though accurate weapons information would not be revealed.

Another example of classification inconsistency is in the area of isotope separation. Cryogenic distillation (CD), which is the most common method used today for separating tritium from deuterium, is largely unclassified provided details revealing accurate production rates of tritium for weapons are not revealed. However, while some details of other technologies for isotope separation, such as the thermal cycling absorption process and thermal diffusion (TD), are unclassified, yet others are still regarded as SRD. Protecting one method of tritium isotope separation while supporting the technology transfer of another makes very little sense, especially when one of the technologies (TD) has not been used for tritium production for well over a decade.

4. Revised Policy

A revised classification policy for tritium processing technologies is needed that achieves the overall goals of necessity, practicality, and consistency. Information on tritium processing technologies should not be classified only because the information is associated with a weapons production application. It is proposed that with the exception discussed below regarding extraction, all tritium processing information (including production applications) be unclassified unless the information reveals specific details about tritium weapons components. Furthermore, elimination of the UCNI designation for tritium processing would be desirable. The proliferation of information on tritium technologies and their many applications in fusion energy cannot be effectively controlled and restricted, especially when the U.S. continues to support an open exchange of technological information for the international fusion program.

One area of tritium processing that still requires selective classification is new extraction techniques for tritium targets associated with advanced production technologies. Details of tritium extraction systems that could reveal advanced tritium target designs should continue to be protected. This is especially true for targets designed for use in light-water power reactors in order to make it difficult for potential proliferant nations to produce tritium in their power reactors. Consideration should also be given to the protection of any novel or difficult extraction techniques associated with the accelerator production of tritium. On the other hand, much of the existing tritium extraction technology associated with lithium-aluminum targets used in SRS reactors is well known and is not difficult to apply. The international fusion energy community is exploring the use of lithium blanket systems, which would utilize similar extraction methods for the future production of tritium in fusion reactors. Thus, protecting details of traditional lithium-aluminum technology (including times, temperatures, and efficiencies of extraction) does not appear to be warranted, provided that accurate tritium production rates, inventories, or weapon information are not revealed.

There are numerous benefits to be gained by implementing the revised policy outlined above. The benefits of sharing information and exchanging tritium processing technology with the fusion community are manifold. Much of what has been learned in over 40 years of tritium processing experience in the DOE weapons program would be very useful in the design, operation, and maintenance of future fusion energy facilities. Some of the most valuable information is not necessarily associated with new technology, but rather with lessons learned from tritium operating experience. In this area, the U.S. can contribute significantly to the overall success of the fusion program. Large-scale tritium processing and material data, previously requested by ITER design personnel, have often been delayed or restricted due to current classification policy and guidance. Some of this information could play an essential role in the long-term reliability and safety of ITER tritium systems.

The DOE weapons program would also benefit from a less stringent classification policy on tritium technologies and processing experience. Collaborative efforts with the fusion community would undoubtedly lead to improved methods of purifying and storing tritium. Synergism between the two programs could result in significant R&D efficiencies and cost savings. There will be a significant effort by the ITER program to develop and install state-of-the-art tritium processing systems; some of the technology advancements will be useful in the weapons program. Much of what is developed for ITER systems could subsequently be adapted to existing DOE tritium facilities.

Additional benefits from such a revised policy could be realized through improved U.S. foreign relations resulting from increased international scientific cooperation. The U.S. is one of four major parties participating in the ITER project, each competing to be the ITER home site. Enhanced U.S. cooperation along with its willingness to openly share relevant tritium experience gained from defense programs could enhance the chances for siting ITER in the U.S. In addition to fusion energy, advancements in hydrogen-related energy systems could be assisted by application of hydrogen isotope storage and transfer technologies developed for the weapons program.

The risks of this revised classification approach are very small for two principal reasons. First, revealing information about tritium processing technologies from the weapons program would not significantly enhance an existing nuclear weapons state's capability to develop boosted nuclear devices. As part of fusion and hydrogen energy programs worldwide, all of the relevant technologies associated with tritium processing are well developed and openly known. All of the nuclear weapons states possess the resources and technical sophistication for acquiring and applying these technologies. For an emerging nuclear weapons state, the difficulty would be in the design details of a boosted weapon. Therefore, the most cost-effective way to impede nuclear states from acquiring the technology for boosted nuclear devices is to continue to protect critical weapons design information. Second, proliferant nations or terrorists lack the technical sophistication to develop boosted weapons, even though they could acquire "book knowledge" of tritium processing technologies from the open literature and possibly acquire small stocks of tritium from the open market. Proliferants would focus their resources on developing simple nuclear devices.

5. Recommendations

The NMPWG recommends that the classification policy regarding tritium processing technologies and applications be revised to provide consistent treatment of information associated with defense programs and the fusion program. Specific recommendations are as follows:


1. Issue

Should the simple association of specific nonnuclear materials with the production of nuclear weapons or SNM be classified?

2. Background

The production of nuclear weapons and SNM involves the use of some highly engineered, and often unique, materials. Some of these materials are incorporated into weapons and test devices. Others are used in the production of SNM or components. These materials often represent significant financial investment and are the end result of many years of development. The ways in which many of these materials are used in the production of nuclear weapons or SNM are considered sensitive and are often classified. In a few cases, simple association of specific materials with the program in general or with specific sites has been considered to reveal highly sensitive information, and such simple association has been classified.

Classification of simple associations of specified materials with specific sites can greatly complicate waste disposal and reporting activities. Current classification guidelines, when addressing these materials, provide no minimum concentration values, and therefore wastes containing trace quantities are considered classified. This practice increases the cost of disposal, as dedicated classified burial grounds must be developed and maintained, and makes it difficult to work openly with regulatory agencies and the public. The situation is complicated further by ongoing improvements in analytical techniques. Materials at concentrations that were not detectable using older techniques can now be identified, thus making literal compliance with current classification requirements more difficult.

3. Current Policy

Under current classification guidance, the simple association of certain materials with the production of nuclear weapons or SNM, in general, or with specific sites or activities, is classified. A primary source of classification guidance on this subject is Classification Guide TCG-WM-1, the Classification Guide to Weapons Materials. Related classification guidance is also found in program guides, such as the guides covering uranium enrichment activities.

4. Revised Policy

A blanket declassification of all materials associated with the weapons program was considered by the working group. Under such a blanket declassification, the simple association of any material with any site would be declassified. Consultation with weapon designers from Los Alamos and Lawrence Livermore National Laboratories revealed concern that simple association of certain materials with the weapons program could reveal key aspects of weapons design.

The working group, while respecting the concerns expressed during its discussions on this subject, believes that the amount of information determinable from the presence of a substance in a waste stream (consisting of a large number of different constituents) is small. Specifically, the working group believes that the analytical effort required to identify and characterize trace amounts of most sensitive materials likely to be found in waste streams far surpasses the value of the information. In the extreme case, someone would virtually have to know what to look for in order to determine that it was present. In such cases, analysis of waste streams would only confirm that a material was present and would provide limited new information regarding the use of the material. Therefore, it should be possible to declassify the presence of materials in sufficiently low concentration in waste streams. Such a change in classification policy would immediately resolve many classified waste issues.

Thus, the revised policy would be as follows:

Simple association of a material with the nuclear weapons program or with specific sites shall not be classified.

Note that under the proposed new policy, further elaboration or discussion of certain materials would remain classified.

5. Recommendations

The NMPWG recommends that the above policy be implemented.


1. Issues

Classification issues related to nuclear materials disposition will arise from actions associated with material security, stabilization, storage (including international monitoring of stored material), and ultimate disposal of excess material.

2. Background

Most classification concerns arise from the declaration that "all excess fissile material will be under international inspection." It is important to realize that the most likely inspection regimen will be that of the IAEA, where safeguards are intended to provide timely detection of host nation diversion of materials.

Under IAEA inspection, the host nation must expect the following:

The inventory of excess weapons plutonium that must be dealt with by the Fissile Materials Disposition Program (MD) will come from weapon pits and processing residues. Pit disassembly will probably use pyroprocessing, in which plutonium is removed by hydriding followed by dehydriding to metal or oxidation to plutonium oxide. This will supply feedstock material of sufficient purity for any of the disposition options. Residues in the weapons complex exist in a variety of forms having a wide range of chemical stability and concentrations of impurities. These residues will be processed chemically and/or repackaged as part of the material stabilization action to be taken in response to DNFSB Recommendation 94-1. Stabilization plans are presently in formative stages, with the result that forms and purities of residues entering MD may change from those anticipated today. However, it is safe to assume that those residues deemed sufficiently rich in plutonium to enter the disposition program will be in forms with adequate chemical stability to allow storage for a minimum of several decades.

Materials must be transportable. This requires compatibility with the shipping envelope as defined by:

10 CFR 71, Packaging and Transportation of Radioactive Material
49 CFR 173, General Requirements for Shipments and Packagings
49 CFR 178, Specifications for Packagings.
Materials also must be sufficiently stable for storage for a minimum of 15-20 years.

As described above, material feed expected for the disposition program will meet both of these conditions. The requirement for transportability is obvious; material must be moved from sites in the complex to the location of disposition facilities. With regard to storage, MOX and immobilization options (described below) will require research and development programs as well as time for construction and licensing of facilities, and, for the MOX/LWR option, licensing of LWRs to burn MOX fuel. Furthermore, the disposition campaign will take some years to accomplish, requiring extended storage of some materials. For the deep borehole option, drilling and emplacement is relatively simple and straightforward, but given the difficulty with qualification and licensing of the proposed repository at Yucca Mountain, qualification and licensing for the deep borehole could be a lengthy process.

3. Current Policy

Much of the current classification policy relevant to nuclear materials disposition has been driven in the past by concerns related to the release of specialized weapon design information. Today the greatest concern is in the arena of nonproliferation, specifically in regards to potential proliferant states and threshold states.

Several broadly worded classification/proliferation concerns relevant to nuclear material disposition activities are as follows:

These classification issues, which relate to weapon design and fabrication, are primarily important for items and material which are potentially considered as "excess" fissile material.

Many of the above classification concerns directly affect disposition activities. In fact, a significant amount of material may be placed under IAEA safeguards in the near future. The classification attributes of much of this material will limit the overall disposition program and may affect international negotiations.

For residual and waste material, the current classification policies associated with disposition are essentially the same as those for material processing, which were addressed earlier in this report. That information is largely unclassified but may be controlled as UCNI.

4. Revised Policy

The issues associated with the characterization of components must be addressed by the Design and Weaponization Working Groups.

The classification policies associated with the stabilization of materials and its treatment or processing for ultimate disposal ought to be consistent with, if not identical to, those dealing with the production and processing of materials. These policies are addressed earlier in this report.

The quantities of material that are stored or disposed of should be treated in the same fashion as material inventories. These issues are addressed elsewhere in this report.

5. The Disposition Options


Uranium will enter the disposition program as HEU. Its treatment is straightforward: it will be blended with natural or depleted uranium to reduce the content of the Uranium-235 isotope to less than 20%, which is no longer a weapon material. This process raises no issues regarding classification.


Long-Term Storage
This is the no-action alternative for disposition. The requirements are simply the long-term storage (LTS) criteria: plutonium metal or oxide containing greater than 50% plutonium by weight; and loss on ignition (LOI) less than 0.5%. Some oxides and metals containing less than 50% plutonium may be acceptably stable, but may also increase significantly the size of the storage facility or facilities.
Deep Borehole
This option would emplace material in a borehole several kilometers deep in a region of ancient and stable rock with little mobile groundwater. Any water present would be ancient and stagnant, probably high in ionic strength, and at elevated temperature. Isolation relies primarily upon the lack of water migration pathways and driving forces in the natural system to prevent long-term transport to the accessible environment rather than on engineered barriers. Fissile material may be emplaced directly in shipping product cans, incorporated into a grout, or immobilized into an engineered form prior to emplacement. For emplacement in shipping product cans or grout, there are no material feed requirements other than necessary stability and LOI to allow for transport and storage, that is, the initial pathway requirements. It may be desirable to fill shipping product cans with a packing material to remove free void space.

Emplacement in the borehole in an immobilized form will have the constraints described below for immobilization.

MOX fuels for thermal spectrum reactors require relatively high purity. Therefore, purification steps are assumed to be necessary as part of head-end processing for the MOX fabrication facility. This processing is being designed under the assumption that the feed material already meets long-term storage requirements, as stated above. For MOX fabrication, the general feed preferences include uniform quality with low impurity levels, low-fired oxide, and little or no chloride. (Once more, it is understood that additional conversion and stabilization may be required by MD to meet the LTS criteria.)

In this option, plutonium-bearing materials are immobilized, with or without high-level waste (HLW) or Cesium-137, by vitrifying in a borosilicate or other glass or incorporating in a zeolite, ceramic or metal matrix. These processes can accept greater impurity concentrations than can MOX fuel; however, the initial pathway requirements for stability and LOI must be met. For glass and ceramic options, the feed preferences are for low-fired oxide with low halides—5% to 10% would be acceptable, assuming the ability to blend Category III (impure) material with Category I (weapons grade) material. Plutonium metal is the feed option for the metal matrix, and metal or chloride is the choice for zeolite. A new option called Glass Material Oxidation and Dissolution System (G-MODS) can accept any material low in halides.

6. Recommendations

The NMPWG defers to the Design and Weaponization Working Groups with regard to those classification policies dealing with component characteristics. We recommend that the policies concerning material processing and inventories presented earlier in this report be applied to the disposition of material.


A. Historical Production of SNM and Tritium

1. Issue

Although the historical quantities of SNM (plutonium and HEU) produced have recently been declassified, the historical quantities of tritium produced remain classified.

The quantitative inventories of SNM located at most weapons complex sites as of December 1993 have been released to the public. The quantitative inventory of tritium continues to be classified.

Should historical production rates and corresponding inventories of all SNM and tritium produced by DOE be declassified and routinely released for public dissemination?

2. Background

SNM and tritium have been produced in large quantities to support U.S. nuclear weapons programs. The SNM is used in the production of primaries and secondaries. Tritium is used to boost the yield of fission weapons, and been used in the design of many weapons systems. SNM and tritium will continue to be used to support current and projected Department of Defense (DoD) stockpile requirements.

In addition, small amounts of tritium have been produced for nondefense uses, which include medical and industrial applications. The supply of material for these applications is drawn from the production and recycling process streams that support weapons requirements.

a. SNM

New SNM was produced by the DOE and its predecessor agencies to meet the demands of the DoD stockpile and naval reactor requirements. HEU was produced at Portsmouth through 1993 and at Oak Ridge through 1964. New plutonium was produced/extracted at Hanford and SRS through 1989.

The DOE weapons complex has satisfied the DoD's stockpile requirements for SNM through a combination of material reused from stockpile retirements and new material production. Material from retirements was recycled and blended with new production materials to support nuclear component build requirements.

As a result of the treaties calling for a reduction of the size of the DoD stockpile, the inventories of SNM located at DOE sites now and in the future will be used to meet the currently projected stockpile requirements.

b. Tritium

Tritium is radioactive and must be replaced in weapons. Limited-life components (LLCs) have been deployed in the weapons stockpile, returned for recycling of tritium and removal of the decay product helium, and refilled and redeployed to the stockpile. An extensive infrastructure supports the extraction from the reactor targets that were used to produce the tritium, the "loading" of tritium into reservoirs, the processing and purification of material returned in LLCs or retirements from the DoD, and the storage of filled and returned reservoirs in the pipeline at SRS.

The DOE assigned to the SRS the responsibility for tritium production and recycling. In addition, a small quantity of tritium had previously been produced for defense purposes in Hanford reactors. In supporting all material demands, reactor production was scheduled to first meet "current" weapons tritium demands with the balance of available reactor production devoted to plutonium-239 and by-product isotopes. Since tritium production quantities were closely linked to specific DOE-supported missions, a correlation might be made that would allow the approximate amount of tritium in a specific weapon system to be inferred and, therefore, design information might be compromised.

In general, the production reactors at SRS and Hanford performed multiple missions, producing plutonium-239, tritium, and by-product isotopes. Although quantitative information on the historical production of plutonium-239 and by-product isotopes has been released to the public, the quantitative details on tritium production have remained classified to avoid releasing a complete "information link," which might reveal DoD stockpile content.

A detailed knowledge of the amounts of tritium that were produced or located in different portions of the production, processing, and recycling complex might indicate the size of the past and current weapons stockpile and plans for the future stockpile. General design features of weapons systems may then be calculated, including the average tritium content of past, current, or future weapons in the stockpile.

3. Current Policy

Historical quantities of plutonium and tritium produced at Hanford have been declassified, as have historical quantities of plutonium produced at SRS. However, the quantities of SNM and tritium allocated to the weapons program remain classified. The DOE has identified the quantitative inventories of SNM at most DOE sites, excepting Pantex, in a one-time release of information. Additionally, the DOE has declassified the fact that intermediate assay enriched uranium (20% to 93% Uranium-235) has weapons applications and has included the amounts of these materials produced in the total HEU production data declassified. Tritium produced at SRS remains SRD, since those data might be used to calculate the historical tritium production and stockpile sizes.

Information from which tritium production may be derived is classified only if it would reveal "accurate" production estimates. (The criterion for accurate production is classified CRD.) Lesser-quality information that could provide "estimated" production quantities is unclassified. However, the criteria for "accurate" and "estimated" production do not apply to the portion of the tritium supply that is, or was, allocated for military requirements. A knowledge of how much was produced might allow calculation of the amount allocated for military purposes.

4. Revised Policy

Historical production of HEU by the gaseous diffusion plants, by year and assay, should be declassified unless such a declassification would permit accurate information concerning assays and quantities used in specific weapons to be determined. Other than for the potential for this information, combined with historical weapons stockpile/production data to reveal weapons-specific design information, DOE has no reason for keeping this information classified; there may be military preparedness or other arguments for not implementing this change.

Within the DOE, a policy change involving the declassification of the historical quantities of SNM and tritium produced and/or the periodic release of information on existing inventories of SNM and tritium should be effected. The DOE will not release information that would compromise the integrity of the safeguards and security programs.

Additional sensitivities that bear consideration by the DOE include the desires of the DoD to protect its stockpile interests and the continuing development of agreements with other countries by the State Department concerning nuclear weapons and related inventories of SNM and tritium. The DOE has no reason for protecting the historical production rates; however, DoD may have reasons for not effecting a change in the DOE's proposed classification policy.

Benefits of declassification include the following:

Disadvantages of declassification include the following:

5. Recommendations

The working group recommends that:


A. Non-U.S.-Generated Restricted Data

1. Issue

How should classification be applied to RD generated outside the U.S. and voluntarily brought to the U.S. by its legal owners?

2. Background

When the Atomic Energy Act was drafted, an unwritten but pervasive assumption was that the U.S. had, and would continue to have, a virtual monopoly on the initial generation of RD. Further, portions of the Act presume that RD will only be generated by nations, although the Act covers "all data" and makes no specific reference to "national" vs. "private" RD. Therefore, while the Act provides for sharing RD on a limited basis with nations that sign Agreements for Cooperation with the U.S., the Act makes no explicit provision for protection of RD imported to the U.S. or from subnational (i.e., commercial) entities.

The worldwide development of nuclear technology has reached a point where significant information falling within the definition of RD is available from non-U.S. sources. For example, the European consortium URENCO wishes to sell or license its gas centrifuge uranium enrichment technology to Louisiana Energy Services. There are many other areas where research into peaceful uses of atomic energy are ongoing on an international basis. These include inertial confinement fusion and advanced isotope separation techniques.

The design of the URENCO gas centrifuge clearly falls within the definition of RD. Key features of the DOE-developed gas centrifuge design are classified. Without a specific declassification or a formal Agreement for Cooperation, there is no formal basis in current law for sharing these classified features with foreign nationals. This problem creates a situation where a uranium enrichment centrifuge, or its design, could be voluntarily brought to the U.S. by URENCO, but discussion of the device or its design with its foreign designers would be a violation of the Atomic Energy Act.

To handle this situation, DOE has adopted an interpretation of the Atomic Energy Act under which centrifuge information voluntarily brought to the U.S. by URENCO is "old" RD, and not subject to restrictions on being transmitted to foreign nationals. (Note that to preserve URENCO's intellectual properties rights, and consistent with nonproliferation requirements, the U.S. and Louisiana Energy Services will protect URENCO's data from disclosure.) If, however, URENCO makes improvements to their process or equipment while they are in the U.S., such improvements would be deemed "new" RD and would be subject to all Atomic Energy Act provisions regarding exchanges of information with foreign nationals.

This compromise, while appearing to adequately address the current situation, has yet to be tested. How it will actually be implemented remains to be seen. Precisely what would constitute "new" RD is not known to be fully defined, nor are conflict resolution procedures known to be in place. It is also not certain if the compromise reached in this case should be considered a precedent applicable to all similar situations arising in the future.

Analogous questions arose when Russia sent TOPAZ nuclear reactors to the U.S. Rather than continue to treat such situations on an ad hoc basis, DOE needs to develop a policy for handling non-U.S.-generated RD.

3. Current Policy

It is unclear if a broad, generally applicable policy for classification of non-U.S.-generated RD currently exists.

4. Revised Policy

A policy that assures that the public has the benefits of the best available technology for the peaceful exploitation of nuclear energy, regardless of the technology's nation of origin, is clearly required. Such a policy would avoid the potential disincentive to importation of technologies to the U.S. that the threat of classification of future developments represents. It is conceivable, under current classification policy, that a supplier would decline to offer for use in the U.S. new technologies because of the fear that evolutionary changes to those technologies would become classified in the U.S. and not available to the original technology holder.

The new policy must address such factors as the following:

5. Recommendations

The DOE should prepare a broad policy designed to facilitate the use of non-U.S.-generated RD within the U.S. Such a policy should, as a minimum, respect the intellectual property rights of the original technology developer while maintaining consistency with U.S. nonproliferation objectives and commitments. Once the agreement on a proposed policy was obtained, changes to the Atomic Energy Act would be proposed, if needed, to formalize implementation of the policy.

B. Unclassified Controlled Nuclear Information (UCNI)

1. Issue

Should UCNI continue to be used for the protection of sensitive but unclassified materials production information?

2. Background

Provisions for control of the dissemination of UCNI were added to the Atomic Energy Act (Section 148) in 1981. This section of the Act provides for control of:

...dissemination of unclassified information pertaining to

(A) the design of production facilities or utilization facilities;

(B) security measures (including security plans, procedures, and equipment) for the physical protection of (i) production or utilization facilities, (ii) nuclear material contained in such facilities, or (iii) nuclear material in transit; or

(C) the design, manufacture, or utilization of any atomic weapon or component if the design, manufacture, or utilization of such weapon or component was contained in any information declassified or removed from the RD category by the Secretary (or head of the predecessor agency of the Department) pursuant to section 142.

Such control may only be exercised if:
...the Secretary determines that the unauthorized dissemination of such information could be reasonably expected to have an adverse effect on the health and safety of the public or the common defense and security by significantly increasing the likelihood of (A) illegal production of nuclear weapons, or (B) theft, diversion, or sabotage of nuclear materials, equipment or facilities.
These provisions closely parallel the provisions of Section 147, which grants to the NRC authority to control the dissemination of "Safeguards Information." In addition to using these provisions for protection of safeguards information, the DOE has used UCNI to control the dissemination of information of potential proliferation significance in declassified areas. This use of UCNI has been extremely controversial and has been criticized by many stakeholders and was questioned by the National Academy of Sciences in their report, "A Review of the Department of Energy Classification Policy and Practice."

DOE use of UCNI to control the dissemination of declassified information has, in part, been based on an interpretation by the Office of General Counsel that the Atomic Energy Act has no provision for reclassification of information. This is significantly different from the Executive Order on National Security Information (NSI), which makes explicit provision for reclassification, provided the information meets the requirements for classification and "...the information has not previously been disclosed to the public under proper authority." While there have been exceptions, the general position of the Office of General Counsel has been that unless a declassification action was specifically limited in scope, new information in an area previously declassified cannot be (re-)classified as RD. Under this interpretation, new information that warrants protection from uncontrolled dissemination, in any area previously declassified, can only be protected as UCNI.

The working group has a fundamental disagreement with this interpretation. It does not agree that a declassification proposed and approved years before the development of new and potentially dangerous (i.e., high proliferation risk) technology can be automatically applied to such new developments. If this interpretation were reversed and new developments assessed for classification on their individual merits, much of the justification for, and controversy over, UCNI would disappear.

The NMPWG also finds that UCNI is being applied in an overly broad manner. Topic 4257 of the UCNI Topical Guideline for Nuclear Nonproliferation, TG-NNP-1, contains the following guidance:

Unclassified information concerning research and development to acquire applied scientific information including "bench-top" chemistry, that reveals current, new, or improved plutonium processing technology, equipment or methods for primary or secondary processes in sufficient detail to permit implementation of the technology to produce practical quantities of plutonium...UCNI.
Under this guidance, new techniques to extract plutonium from scrap and/or waste streams that allow the process residuals to be disposed of as low level waste must be protected because they also produce a plutonium-rich product. This determination was based on the belief that these techniques could be used to obtain practical quantities of plutonium from scrap and/or waste materials. Review of this determination with plutonium processing experts from Lawrence Livermore National Laboratory has led the working group to conclude that this determination is flawed. Specifically,

Therefore, making these processes UCNI does not serve any significant nonproliferation objective, but does prevent the dissemination of a useful cleanup technology to the general nuclear community.

On the other hand, the NMPWG agrees that, were a truly revolutionary process developed that would permit, for example, the extraction of plutonium from spent fuel in a small, easily hidden facility, such a development would pose a major proliferation risk. Such a development would warrant protection as RD.

The working group also notes that newly prepared compilations of information, such as reviews of the declassified electromagnetic separation or calutron process used during the Manhattan Project for uranium enrichment, are being classified as UCNI. While agreeing that compilations of unclassified information can easily divulge sensitive information, the working group does not agree that designating such compilations UCNI provides adequate protection. The working group notes that Section 1.8(e) of Executive Order 12958 specifically provides for classification, as NSI, of compilations of unclassified information. Such an approach appears to be less arbitrary than marking such information UCNI and would provide a higher level of protection to sensitive information.

3. Current Policy

DOE policy on UCNI is found in DOE Order 471.1, Identification and Protection of Unclassified Controlled Nuclear Information. A number of UCNI topical guidelines have also been promulgated.

4. Revised Policy

The NMPWG believes that UCNI information is too broadly defined and too weakly controlled to be useful for the protection of truly sensitive information relating to materials production. While it is conceivable that narrower guidelines and stronger controls could be applied to UCNI, a policy of information being either classified and unavailable to the general public or unclassified and available to the public is recommended.

UCNI, which fundamentally provides minimal protection to broad areas of information, is inconsistent with the overall policy the Fundamental Classification Policy Review is striving for, of building high fences around small areas of the most sensitive information.

5. Recommendations

The Office of Declassification should perform a topic-by-topic review of all UCNI topical guidelines covering materials production information and identify that information specifically warranting protection as RD. After consultation with the Office of General Counsel, this information, narrowly described, should be redefined as RD. Once all UCNI information has been so addressed, the UCNI guidelines covering these topics should be withdrawn from use.

If it is found that some of the information currently protected as UCNI legitimately requires protection but cannot be redefined as RD because of prior declassifications, then a change to the Atomic Energy Act, specifically authorizing reclassification, should be sought. Such a provision could be based on the current provisions for reclassification found in Executive Order 12958, Section 1.8(d). Alternative approaches to allow reclassification as NSI should also be considered.

Further, the Office of Declassification should develop a formal policy regarding compilations, including formally declassified information, that recognizes the sensitivity of such compilations and allows for them to be protected either as RD or as NSI under the provisions of Executive Order 12958, Section 1.8(e).


Nuclear Materials Production Working Group
Fundamental Review of Classification
Working Group Roster

(Paul T. Cunningham, Chairman)

Person Organization  Area of Expertise 
Paul Cunningham Los Alamos materials production chemistry
Earl Ault LLNL AVLIS (classification of)
Tom Gould WSRL tritium, Pu production
Mr. Stanley Keel OSD (AE) (Pentagon) material production
Dick Krajcik Los Alamos weapons design
W. Glenn Northcutt MMES metallurgy
Jim Rushton ORNL uranium enrichment
molten salt reactor
Beverly Bender NMRT plutonium
George Biggs FSS-16, LANL materials classification
Jeff Zarkin DOE/Office of Declassification classification policy
Dave Klein DynMeridian nuclear materials


List of Acronyms

AEA Atomic Energy Agency
AEC Atomic Energy Commission
APT Accelerator Production of Tritium
AVLIS Atomic Vapor Laser Isotope Separation
CD Cryogenic Distillation
CRD Confidential Restricted Data
DNFSB Defense Nuclear Facilities Safety Board
DoD Department of Defense
DOE Department of Energy
G-MODS Glass Material Oxidation and Dissolution System
GTA Ground Test Accelerator
HEU Highly Enriched Uranium
HLW High-Level Waste
HWR Heavy-Water Reactor
IAEA International Atomic Energy Agency
ITER International Thermonuclear Experimental Reactor
LANL Los Alamos National Laboratory
LEU Low-Enriched Uranium
LLNL Lawrence Livermore National Laboratory
LOI Loss On Ignition
LTS Long-Term Storage
LWR Light-Water Reactor
MD Fissile Materials Disposition Program
MLIS Molecular Laser Isotope Separation
MMES Martin Marietta Energy Systems
MOX Metal Oxide
NDA Nondestructive Analysis
NMPWG Nuclear Materials Production Working Group
NMRT Nuclear Materials Reconstruction Technology
NRC Nuclear Regulatory Commission
ORNL Oak Ridge National Laboratory
OSD Office of the Secretary of Defense
RD Restricted Data
R&D Research and Development
RERTR Reduced Enrichment for Research and Test Reactors
RTR Research Test Reactor
SNM Special Nuclear Materials
SRD Secret Restricted Data
SRS Savannah River Site
TD Thermal Diffusion
UCNI Unclassified Controlled Nuclear Information
URENCO Uranium Enrichment Corporation
USEC United States Enrichment Corporation
WSRL Westinghouse Savannah River Laboratory


Nuclear Materials Production Working Group

Ted Motyka (SRS/SRTC), Tritium Processing

Tracy Rudisill (SRS/SRTC), Actinide Processing

Dana Christensen, Joel D. Williams, Joe Martz (LANL/NMT), Plutonium Processing, Residues & Waste

Earl Ault (LLNL), UAVLIS

Mike Cappiello (LANL/LER-APT), APT

Douglas E. Fain (Lockheed Martin Energy Systems), Fabricating Inorganic Membranes @ ORNL

Richard A. Krajcik, Rod Schultz, Stephen B. Kemic, Merri M. Wood, Gary Wall, D. Reid Worlton (LANL), Use vs Application

Jim Anderson (LANL/APT), Accelerator Production of Tritium

George Eccleston (LANL/NIS-SG), IAEA / Nonproliferation Issues

Ernie Gladney (LANL/ESH-17), EPA

Tom Gould (SRS), Reactor Fuel and Target Technology

Dana Rowley, Jim Tyler, Hasel Slone, Debbie Wojtowkz, Leonard Haselman (LLNL), Classification Discussion Regarding Non-Nuclear Materials (Part 2)

1 Hardie, Wayne R. et al., Classification Review Assessment Using Risk/Benefit Decision Analysis (Applied to Lithium-Enrichment Technology). LA-UR-96-254, Nov. 1995.

Report of the Fundamental Classification Policy Review Group