Chapter 3. Observations.
Equipment commercially available at the time of the 9/11 attacks was limited in its capability. PVT radiation detectors could detect radiation but could not identify isotopes, and shielding SNM might defeat detection. Radiographic equipment could reveal dense objects, but relied on operator skill to flag potential threats. It might be possible to hide a nuclear artillery shell in a cargo of dense objects, and it would be difficult to pick out a small piece of SNM. Resolving alarms required time-consuming methods, such as using hand-held radioisotope identification devices or unpacking a container.
Capabilities of existing systems can be improved incrementally, such as by using different detector material, computers, algorithms, or CONOPS (e.g., scan time).
Systems now under development have the potential to reduce false positives (speeding the flow of commerce) and false negatives (improving security). Fission that neutrons or x-rays induce in SNM generates unambiguous signals. Dual-energy radiography detects high-Z material automatically. EZ-3D reveals high-Z material hidden in medium-Z material, and might be able to differentiate SNM from other high-Z material. These approaches detect useful signatures, but have drawbacks as well, such as low signal strength, complexity, high cost, or large size. The task is to utilize these signatures and minimize drawbacks in a system that can be fielded. Other technologies, such as improved detector material and improved algorithms, also have the potential to improve detection capability.
It is difficult to predict the schedule of new detection technologies. In March 2008, the Royal Society, drawing on a workshop of experts, issued a report on nuclear detection that found, "In the medium term (5-10 years), there are promising opportunities to develop new technologies, such as muon detection systems. In the long term (10-20 years) detection could benefit from advances in nanotechnology and organic semiconductors." (166) In 2008, the company developing the muon tomography system thought the system could be commercially available in 2011. As of early 2010, that date had slipped to 2012 and the company had not passed its Test Readiness Review, a step to indicate whether a system is ready for its proof-of-concept demonstration. In 2008, some thought that nanocomposite scintillator technology could be available for transfer to industry by September 2009, but the project was canceled in January 2010.
It is difficult to evaluate prospects for R&D projects. Based on tracking the technologies presented in this report, it appears extremely difficult to evaluate how likely an R&D project is to succeed, even for the agencies that fund them, and one should not confuse a technical explanation and briefing slides with prospects for success. To succeed, a project must overcome many hurdles between concept and deployment. (1) The concept has to be scientifically sound. This is not always a given for projects that push the state of the art. (2) Even if scientifically sound, the underlying science must be transformed into a prototype through engineering. But materials may prove impossible to develop; laboratory-scale proof-of-concept equipment, where size and complexity are not a concern, may prove difficult to shrink in size; and algorithms may be unstable or may be confused by background radiation. (3) The prototype must be made into a system that is rugged enough to survive the bumps, vibrations, heat, cold, rain, humidity, dust, salt air, gasoline fumes, and whatever else people and nature may throw at it. (4) There must be a workable concept of operations: if it takes 1000 seconds to perform a scan, or if the false positive and false negative rates are too high, or if the operator cannot use the equipment easily, the equipment is useless. (5) The system must be affordable, however defined. It is hard to predict if a concept will make it past the next hurdle, let alone all five (and any others).
Here are several examples drawn from this report. Nanocomposite scintillators held the promise of being a gamma-ray detection material that would be sensitive, yet inexpensive and easy to produce on a large scale. Early research started in 2004, but DNDO and DTRA terminated the project in 2010. The AS&E CAARS project appeared promising, but encountered unspecified technical problems and DNDO terminated it; however, some of its technology is being applied to another project. Conversely, SAIC's CAARS depended on the development of an "interleaved" accelerator, one that could switch x-ray beams between two energy levels many times a second. An earlier attempt to develop such an accelerator failed, but SAIC's subcontractor, Accuray, was able to develop one that exceeded requirements by a substantial margin, contributing to the system's ability to differentiate among up to 15 bands of Z rather than simply indicating whether material in a cargo container was high-Z or not. This enhanced capability could help CBP agents search for contraband as well as SNM.
It is easier, less costly, and potentially more effective to accelerate a program in R&D than in production. DTRA believes that a significant increase in funding for proton beam technology, a standoff detection technology in early R&D, might shorten time to deployment by several years by enabling researchers to consider many technical alternatives simultaneously to determine the most promising approach faster. It is hard to attain large schedule gains by accelerating production; such gains may entail high cost, such as multiple shifts or more production lines; and a rush to production may cause a project to fail. While R&D projects may also fail, more risk is tolerable in R&D because the investment is much less.
A modest amount of money spent in R&D can avoid looming problems. For example, GADRAS, a widely used algorithm for detecting SNM and other materials, runs on the standard Microsoft Windows operating system (OS) for personal computers. Microsoft introduces new generations of OSs from time to time. Typically, new OSs will support programs written for several generations of previous such systems. However, the Graphic User Interface (GUI) for GADRAS is written with the Visual Basic 6.0 compiler, which Microsoft no longer supports. At some point, Microsoft will likely introduce a new OS that will no longer support applications that are written with this compiler; GADRAS would then be unavailable to its users until it is updated. According to Dean Mitchell, who created GADRAS, updating that algorithm to run on current-generation OSs would avoid that problem, at a cost of perhaps $1 million a year for two years. (167)
R&D that leads to products that many systems can use may have a large impact on detection capability. Many detector systems have common elements--an accelerator, gamma-ray detector material, computers, algorithms--so improving any of these "building blocks" might improve the capability of many detector systems, including those in the field. Improved gamma-ray detector material can improve sensitivity, reduce cost, or both. An improved algorithm can boost performance. A more powerful computer permits the use of a more powerful algorithm, which may reduce false positives and false negatives.
On the other hand, it may not be possible to upgrade systems simply by swapping new components for old. Edward McKigney listed possible difficulties in the (hypothetical) case of upgrading systems by substituting higher- for lower-performance detector material:
(1) Detector modules that cannot detect light with high efficiency would need to be redesigned. This is particularly relevant for existing portal monitors that use plastic scintillator material, where the optical design is poor. (2) Electronics for converting signals from detectors into data for algorithms ("readout electronics") that are not suitable for high-resolution readout and analysis, or are mismatched for the technology (such as if the old electronics read electrical charge while the new ones read optical signals), would have to be replaced. (3) Data analysis algorithms that cannot process signals from the new detector module would have to be replaced. (4) The volume of data from the new detector module might be greater than the existing algorithms, data transmission system or computers could handle, requiring new computers, algorithms, data transmission system, or some combination. (5) Electrical power systems would have to be changed if the power requirements for the old and upgraded systems did not match. So, at the extremes, it might be possible to upgrade only the detector module, or the only features of the old system that would remain after an upgrade would be the wide spot in the road and the guard shack. I would recommend that the next generation of detector systems should be more modular so upgrades could be done while retaining as much of the value of deployed systems as possible. (168)
Synergisms may arise between technologies. Beams of neutrons or high-energy gamma rays used to induce fission in SNM may harm some products, expose stowaways to high doses of radiation, and require shielding to protect workers. Improved detector material and algorithms could lower the amount of radiation required for this technique, perhaps making it more usable for scanning containers.
Technical advances can place two systems in competition unexpectedly. Work is underway on several systems designed to induce fission as a way of detecting SNM. CAARS was not begun as a system of this sort. However, DNDO is investigating a technology add-on to give it that capability. If work proceeds on that path, CAARS and other such systems could be in competition.
Competition at the R&D level may be desirable. William Hagan, Assistant Director, Transformational and Applied Research, DNDO, states,
if we can squeeze additional functionality out of a system, we want to do that. This will cause various approaches to be in competition for achieving a capability at the R&D stage but that is what we want to do so we can drive towards the most effective. More generally, I think that having multiple organizations pursuing the same R&D goal is a good thing because it allows for different approaches or more capable organizations to compete for the objective. This is a very effective mechanism in R&D. A classic recent example is the race to decode the human genome. Another is the race for commercial space flight. This kind of competition goes on all the time in the basic research community and I think we should encourage it. There is, of course, some limit to this, but we are far from that limit right now for radiation detection. (169)
The competitive position of systems in R&D may change over time. Technology development is dynamic. This report presents several examples. The SAIC CAARS overcame a key technical hurdle, the development of an interleaved accelerator, resulting in better performance than expected. The AS&E CAARS encountered problems that led to its termination. The Rapiscan Eagle, with an added algorithm to detect high-Z material, became a competitor to CAARS through the JINII program. Decision Sciences Corporation addressed problems with the original concept for its muon tomography scanner, such as using boron-10 instead of helium-3 in drift tubes because of the latter material's scarcity and designing a top/bottom scanner rather than a top/bottom/both sides scanner to make the footprint more compatible with traffic lanes at ports.
"Concept of operations" (CONOPS) is crucial to the effectiveness of detection systems. CONOPS details how a detection system would be operated to gather data and how the data would be used. Without it, a detection system would be valueless. Since CONOPS and systems are mutually dependent, the design of each must take the capabilities and limitations of the other into account.
Current equipment to detect and identify SNM makes use of two main signatures of this material, opacity for radiography to detect SNM, and gamma-ray emissions for spectroscopy to detect and identify SNM. However, as discussed in Chapter 2 and the Appendix, SNM has many signatures in addition to opacity and gamma-ray emissions, and some systems under development attempt to make use of these other signatures. If systems utilizing these other signatures were to be deployed, methods that might be used in an attempt to hide or mask opacity and gamma-ray signatures would not necessarily defeat these systems under development--complicating any terrorist attempts to smuggle nuclear weapons or SNM into the United States. At the same time, these future systems tend to be more costly and complex than current systems; whether the added benefits are worth the added costs is a political decision.
Detection systems have their limits. Systems to detect SNM at close range, such as at ports and land border crossings, are generally not applicable to detection of terrorists smuggling a weapon across a remote stretch of border. But that is not a flaw of the detection system. Detectors can work at "points," i.e., places where people or cargo may enter the United States legally. There, detectors attempt to find SNM or weapons that may be hidden in cargo. In contrast, at "lines," the vast distances between "points" along coasts or borders, any entry is illegal, so interdiction is a matter of law enforcement. Effective intrusion detection systems (TV cameras, seismic monitors) coupled with a CONOPS that provides rapid response may suffice, though they have a long way to go to become effective. At the same time, standoff radiation detection systems that have yet to be developed, mounted along borders at natural choke points or smuggling routes, might be of value for this mission.
Observations on Technical Progress and Congress
Congress has supported a broad portfolio of detection R&D projects that has created a pipeline with technologies expected to become available for operational systems from near-term to long-term. These systems exploit many signatures in addition to those of currently deployed systems, offering Congress the prospect of improved detection capability and a broader menu of choices. Several technical factors may influence the choice among technologies to support. For example: (1) Projects will advance at different rates. (2) Projects may benefit differently from an advance in a related technology. (3) As a project moves from research to development to deployment, cost and capability may vary from early projections.
Congress may wish to reevaluate current deployment decisions if it concludes that significantly more capable systems will be available in a year or two. Of course, any such decision would depend on comparing such factors as cost, footprint, ease of use, production rate, and the like for competing systems, and caution is necessary in assessing contending claims.
On the other hand, it is difficult for Congress to choose among contending technologies. Each requires evaluation in such terms as cost, scan time, ease of use, reliability, schedule, footprint, radiation exposure, spatial resolution, and ability to thwart shielding. Yet these data are difficult to obtain. Some are proprietary. Some are unknown: schedules may slip and costs rise, or technical advances may cause the opposite to occur. Developers of a technology tend to be its advocates, and see the strengths of their technology and a path to overcome its weaknesses. Even if these data can be obtained, it is necessary to weight data elements to support a choice among contending technologies. With many variables to be traded off against each other, how are weights to be assigned, and who decides? And can this weighting system function despite weaknesses in the data?
Congress has focused much attention on preventing terrorists from smuggling nuclear weapons or SNM into the United States in cargo containers. For example, P.L. 110-53, Implementing Recommendations of the 9/11 Commission Act of 2007, Section 1701, states, "A container that was loaded on a vessel in a foreign port shall not enter the United States (either directly or via a foreign port) unless the container was scanned by nonintrusive imaging equipment and radiation detection equipment at a foreign port before it was loaded on a vessel." While terrorists might attempt to smuggle in a nuclear weapon by other means, developing technology to scan containers at seaports is a reasonable place to start. Container-scanning technology can be modified for use in other situations, such as monitoring air cargo containers or passenger cars, which are easier to scan because they can contain much less shielding. Developing and deploying detection equipment for use at seaports ensures ruggedness and ease of use adequate for real-world applications, and forces governments at all levels to plan CONOPS.
More generally, some could argue that it is impossible to prevent terrorists from smuggling nuclear weapons into the United States, so there is no point in spending large sums in a futile effort. Congress has rejected that approach, and has appropriated, in total, billions of dollars to deploy available systems and to support R&D on advanced technologies. Supporters of the R&D and deployment approach assert that it offers several advantages.
* It has provided some capability quickly, increasing the odds of detecting weapons or SNM. An important example of this is the rapid deployment of passive radiation detectors to scan maritime cargo containers.
* This limited detection capability would help deter terrorists and would complicate plans to smuggle in weapons or SNM.
* Initial deployments provide data of use to subsequent deployments. They help refine what throughput, robustness, etc., front-line inspectors require of a system. They help refine CONOPS. They help define desirable features of an architecture. These results can make future technologies, systems, and architectures more effective.
* It has created an R&D pipeline that is intended to generate a steady stream of new technologies and systems.
* The resulting improvements in individual technologies, operations, and architectures can improve overall system effectiveness.
* As technologies become more capable, they can plug gaps in the current architecture. For example, remote detection might offer a way to monitor choke points in the United States or overseas that terrorists might pass through in transporting SNM or weapons.
Congress may wish to address gaps and synergisms in this portfolio. For example:
Gaps: Several systems may use helium-3 tubes for neutron detection, yet the supply is limited. Alternatives are available, but the longer developers take to switch to these alternatives, the longer it would take to deploy their systems because of the need to incorporate different detectors, modify algorithms, and test the revised system. Other gaps include sensors that can detect SNM at long range (e.g., over 100 m), sensors that can operate autonomously in remote areas, and large but inexpensive detectors that can distinguish SNM from other radioactive material.
Synergisms: A component, algorithm, or material developed for one system may be applicable to another. Projects are under way to develop more sensitive materials to detect gamma rays and neutrons. These materials can be used in systems that induce fission in SNM. Their improved sensitivity permits a smaller source (e.g., an accelerator) to generate the interrogation beam, reducing cost, complexity, and radiation dose. Similarly, if detector material offers only fair resolution of gamma ray spectra, then peaks in a spectrum may blur, requiring a complicated algorithm to deal with the uncertainties. Sharper resolution from improved materials would reduce these uncertainties, permitting simpler algorithms to be used. More powerful computers could support faster, more powerful algorithms, reducing scan time, false positives, and false negatives.
Minimizing gaps and maximizing synergisms have the potential to lead to more capable systems faster and at lower cost. Companies that considered using helium-3 for neutron detection might expedite deployment and reduce costs by sharing effort to develop an alternative neutron detector for their common use. Information on progress in developing more sensitive detector material would permit companies to incorporate such materials into their systems sooner, also speeding deployment and lowering costs. Is there a way that development could be shared or licensed so that companies, especially those working on government-funded projects, could avoid duplicating effort? And could this be done while retaining the benefits of competition?
In considering the Advanced Spectroscopic Portal, Congress and the Government Accountability Office examined in detail whether DNDO had followed proper procedures for testing competing systems. An alternative means by which Congress could address testing is to direct the executive agency in charge of a system to conduct specified tests. These tests would need to be designed, and perhaps overseen, by experts not affiliated with the relevant agency, company, or laboratory. Congress has ample access to the technical expertise required. The relevant congressional committees could consult with individual experts or with groups that have a long history of providing independent technical advice to the government, such as the American Association for the Advancement of Science, the JASON defense advisory group, the National Academy of Sciences, the National Council on Radiation Protection and Measurements, and the National Institute of Standards and Technology. In this way, Congress could seek a fair comparison between systems on variables of interest, such as scan times or the ability to detect specified targets in containers with specified cargoes, enhancing confidence in the test results and decisions based on them. Other alternatives exist. Congress could require DHS to establish an independent test and evaluation unit; obtain an outside review of DHS test and evaluation procedures; require DNDO to provide detailed reporting of each step in the acquisition process as it occurs; or provide for an external review of a program.
Observations on Technical Progress and Terrorism
Ongoing improvement in U.S. detection capabilities produces uncertainties for terrorists that seem likely to increase over time, adding another layer of deterrence beyond that of the capabilities themselves. Capability of fielded equipment may be upgraded. Terrorists may not know the capability or availability of future detectors. More advanced technologies should improve detection capability. It should be harder for terrorists to evade new systems than current ones. Detection may affect terrorists in another way. A nuclear weapon would be of immense value to them. Therefore, increasing the risk of detection would have a much greater deterrent effect for them than for drug smugglers, where detection and confiscation of drugs are part of the cost of doing business. The multiplication of technical obstacles to a successful terrorist attack may thus help deter an attack or an attempt to undertake a project to launch one.
At the same time, it is important to avoid the "fallacy of the last move." Herbert York, a former Director of Defense Research and Engineering, coined this term to argue that in the Cold War nuclear arms race, one side's actions typically led to countervailing actions by the other side. (170) The same principle applies to nuclear detection. This report suggests that some U.S. detection systems nearing readiness for deployment are more capable than current detectors. Yet if terrorists were to attempt to bring a nuclear weapon or SNM into the United States, they could use various techniques to evade detection by such systems, though these techniques might complicate the plot and increase the risk of detection by non-technical means. Further, the threat might increase in various ways, such as if new terrorist groups emerged or if more nations built nuclear power plants or nuclear weapons. For such reasons, Congress has funded, and executive agencies are pursuing, R&D with short- and longer-term time horizons. Also for such reasons, the global nuclear detection architecture may need to be updated from time to time. Thus, while the United States has an immense technical advantage in a competition of detection vs. evasion, and a pipeline of increasingly more capable technologies, it is important to recognize not only the dynamic aspects of advances in detection capabilities but also the dynamic aspects of the competition.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Detection of Nuclear Weapons and Materials: Science, Technologies, Observations|
|Publication:||Congressional Research Service (CRS) Reports and Issue Briefs|
|Date:||Jun 1, 2010|
|Previous Article:||Chapter 2. Advanced technologies: detecting SNM at a distance.|
|Next Article:||Appendix. The physics of nuclear detection.|