Abstract
The birthplace of the Bomb doesn't have an accessible database on all the processes involved in nuclear weapons manufacturing. To bolster nonproliferation efforts, it should.
The Washington, D.C.-based Institute for Science and International Security knew in 2002 that the Iranian regime was planning to construct a heavy water reactor, years in advance of the actual construction activity. Using commercial satellite imagery, the institute observed the aboveg-round construction of a heavy water manufacturing plant in the industrial Iranian city of Arak. Coupled with the institute's knowledge of nuclear weapons infrastructure, the imagery led to the conclusion that the reactor was soon to follow.
If the heavy water plant had been built underground and operated secretly, in a fashion similar to Iran's Natanz uranium enrichment plant, would the institute's analysts have been able to detect the facility? Not likely. After all, the Natanz facility was discovered only after the Iranian opposition group, the National Council of Resistance of Iran, released information about the site in August 2002. Without this information (or “human intelligence”), the United States would likely have failed to discover the facility, as it was well concealed.
It should not be surprising then that Los Alamos National Laboratory staff and management have been interested in adding to national capabilities in detecting illicit nuclear activity. Lab management has outlined a set of “grand challenges” for science and engineering that are intended to enhance the intellectual vitality of the laboratory. One of the challenges, to “sense and anticipate nuclear and biological threats at a global scale in real time,” draws the imagination of many lab scientists and engineers who specialize in detection technologies. The lab envisions a solution to this challenge whereby many detector technologies would be interlinked with information systems that could provide alerts for clandestine nuclear facilities, loose nukes, loose nuclear materials, etc. (This article only discusses the nuclear threat.)
This grand challenge presents one dilemma, as far as the nuclear nonproliferation mission is concerned: How do scientists know what detectors to develop unless they have some a priori knowledge of the signal? In other words, lab scientists need to understand what it is they are looking for. Less than a year ago, Los Alamos's Threat Reduction Directorate held a town hall meeting where a hundred or so scientists and engineers shared ideas on methods for detecting the signatures of clandestine nuclear proliferation activities. The meeting, which was spread over two days, was lively, and scientists discussed a variety of chemical, acoustic, infrared, and nuclear sensors. Smart people engaged in happy talk, and then each went his or her separate way. I hope readers will forgive me for not mentioning more detail about the ideas discussed; the meeting was indeed classified.
Since two heads are better than one, the reader might assume that 100 smart heads would be rather brilliant. Instead, the conversation in the meeting's hallways told a different story. Only one or two people in attendance actually had knowledge of nuclear weapon manufacturing processes, and nearly everyone wanted to talk to them. In other words, virtually every speaker or participant wanted to know if his or her detector concept, which had been developed for another purpose, could be applied to the detection of clandestine nuclear facilities, loose nukes, or loose nuclear materials. The irony of the situation was apparent: At a meeting in Los Alamos, the mecca of U.S. nuclear weapons design and manufacturing, information about nuclear weapons manufacturing was in scarce supply.
A hypothetical example illustrates this point. Suppose that lab scientists developed a new and very sensitive instrument for detecting sulfur compounds that are transported in water. Such an instrument could be used for say monitoring effluents downstream from coal mines. With this promising detection technology in hand, lab scientists might ask, how can this instrument be useful in detecting and preventing nuclear proliferation? A Los Alamos employee or an employee from another Energy Department lab who is expert in the use of this new instrument might try to find if sulfurous compounds are emitted in any stage of the nuclear weapons manufacturing process. Even if this employee had access to “secret restricted data,” he would have trouble doing this. A set of telephone calls would be followed by meetings behind closed doors arranged one at a time. There is no guarantee that the right connection would ultimately be made, because others would have little to gain by taking time out of their activities to make the personal connections.
In principle, a centralized repository of information renders the problem of finding information trivial, even in a large, dispersed network like Los Alamos or the Energy Department. This central directory may be expensive to establish and maintain, and may be impractical for highly classified, compartmentalized information. But I believe that these obstacles can and should be overcome.
The over-compartmentalization of information has hindered the research capacity of los alamos scientists since the Manhattan Project, though management has taken steps to reduce it.
The over-compartmentalization of information has hindered the research capacity of Los Alamos scientists since the Manhattan Project, though management has taken steps to reduce it. Physicist Richard Feynman once complained that because the head of computing at the original Los Alamos site had not been allowed to know the lab's true mission, he used its advanced computing resources to solve integral calculus problems in an effort simply to entertain himself and pass the time. When Feynman pushed J. Robert Oppenheimer, the director of the operation, to allow a broader need-to-know authorization so that the staff would know the importance of their work, the productivity changed almost overnight, according to Feynman.
Partly as a result of the early lessons learned, a majority of the cleared U.S. citizens working as scientific and engineering staff at Los Alamos have broad (as opposed to compartmentalized) access authority to nuclear weapons information. Clearance levels are designated “sigma 1” through “sigma 10.” The 1954 Atomic Energy Act created a broad category of information known as “secret restricted data” that can be discussed in meetings among staff having the right sigmas. This includes most nuclear weapons manufacturing infrastructure information of a generic nature.
This article cannot go into a nuanced discussion of types of classified information. Rather, let me simply say that a user-friendly database of nuclear weapon manufacturing information classified at the secret-restricted-data level could dramatically increase the productivity of the sigma-cleared nuclear nonproliferation community. Again, it is a fortuitous yet ironic coincidence that this database is needed at the site where many of the world's nuclear weapons experts conduct their work.
Construction of such a database will certainly require great effort, but the end-product would be useful not only to Los Alamos, but to other Energy sites and government agencies. Where will all the needed worker-hours come from? The answer is tied to the budget priorities of the new administration and Congress. The expert workforce that is needed to populate the database is already working at the lab, though some job descriptions will have to change.
Take the hypothetical examples of the sulfurous-compound detector and the underground heavy water manufacturing plant. The database would contain at least seven different processes that may be involved in heavy water production. For each of these processes, a diagram that shows the inputs and outputs of a particular manufacturing process, as well as the steps that are taken along the way, would be provided. The three main feeds to one of these processes (the hydrogen sulfide process) are water, air, and aluminum metal. The outputs are heavy water, alumina, wastewater, and exhaust air. A drill-down into each of the outputs would reveal expected contaminants in each stream. For instance, the wastewater would be depleted in deuterium and contaminated with one or more sulfur compounds such as hydrogen sulfide. The alumina and the heavy water product would also likely be contaminated with sulfur compounds. Thus, the connection is made: the new sulfurous compound detector would indeed be relevant to the nuclear nonproliferation mission.
Because the full scope of the nuclear weapons manufacturing process essentially relies on all Los Alamos engineering and science disciplines, the development of this proposed database will also use all of these disciplines, as illustrated by the above example of heavy water production. Most clearly, the main disciplines needed for the database are chemistry, chemical engineering, and environmental engineering. Yet the Los Alamos's expertise in electrical, mechanical, and nuclear engineering could also help determine if any electromagnetic or nuclear radiation or sonic emissions are expected from a given process. Other diagrams in the relational database would require a general engineering background to construct and would show the input materials needed for construction; the major components, materials, and supplies needed for maintenance; and the personnel and transportation requirements. These information sets may be of use to other researchers developing pattern recognition technologies that combine data from a variety of sources.
Los alamos has begun to develop a compendium and analysis of the possible ways that nations or terrorist organizations could process and recover nuclear materials. But effluents and other signatures are not yet explicitly addressed in this program.
Indeed, much of the data needed for this database already exists. For example, a program in the early 1970's (which is still classified), cataloged potential effluents and signatures from nuclear facilities. This multi-million dollar, multi-year program was funded by the Defense Department, and the work was carried out at NNSA laboratories, including Los Alamos. The data exists in a set of reports and is not in the form of a database. More recently, Los Alamos has begun to develop a compendium and analysis of the possible ways that nations or terrorist organizations could process and recover nuclear materials. But effluents and other signatures are not yet explicitly addressed in this program. Lastly, personnel at Energy headquarters have access to updated assessments of the nuclear fuel cycle development capability for select countries using a graphical tool I will refer to as “WIZ” (not its actual name). The current emphasis of these assessments is Nuclear Suppliers Group equip-ment–especially trigger-list and dual-use equipment with both civilian or military applications. The database, however, is classified at a level beyond secret restricted data, and it does not contain flow-sheet or process effluent information, which are essential for determining the applicability of a particular environmental monitor.
This Energy Department database would be a good starting point for developing the more complete Los Alamos database. If the database was sanitized to the secret restricted data level, its controlling software could be modified to include flow-sheets, inputs, outputs, and effluents. This process will first require changes to the data structure itself, and the database's graphic user interface will need to be changed similarly. The computer science aspects of this problem, however, are of secondary concern. The foremost concern and the one that is likely to lead to the most useful outcome is the need to change how some divisions at Los Alamos operate, as was illustrated in the example.
If developed properly, the final product would link disparate scientific and engineering technologies from a broad range of disciplines to the nonproliferation mission. The database's construction may take years, but the end product will be useful to Los Alamos and other Energy sites and federal agencies. (The potential user community includes cleared scientific and engineering personnel in all Los Alamos divisions who are involved in detection and measurement technologies of any kind. Users outside Los Alamos would include other NNSA sites, Energy headquarters, and Defense personnel with “critical nuclear weapon design and information” clearances.)
The internal need for this project is driven by the stovepiping of knowledge and expertise between divisions and directorates within NNSA. Yet the overwhelming motivation for this project is the external need for the information that would be included in the database.