The Practicality of Pulsed Fast Neutron Transmission Spectroscopy for Aviation Security (1999)

The White House Commission on Aviation Safety and Security recommended the deployment of explosives-detection equipment, including x-ray computed tomography (CT) based explosives-detection systems (EDSs), the only detection method that has passed certification testing by the Federal Aviation Administration (FAA). Several other technologies are under development and have been tested in laboratory settings but have not passed certification testing.

One area of research being pursued by the FAA is accelerator-based nuclear detection technologies that detect explosives by measuring the elemental composition of the material under examination. These technologies exploit the high nitrogen and oxygen content present in most explosives. Pulsed fast neutron transmission spectroscopy (PFNTS), one of these element-specific detection technologies, also has the potential for generating low-resolution tomographic images (NRC, 1998; Overley, 1987). However, PFNTS also has a number of practical limitations, including large size and weight, the necessity of radiation shielding, and the regulatory and safety issues associated with using a nuclear-based technology (NRC, 1993, 1997).

In 1993, the FAA requested that the National Research Council (NRC) assist the agency in assessing its explosives-detection program. The NRC responded to this request by convening the Committee on Commercial Aviation Security (CCAS), which has produced two interim reports (NRC, 1996, 1997) containing recommendations for structuring the research portfolio for the FAA's explosives-detection program. The committee's recommendations addressed bulk explosives detection, trace explosives detection, combined technologies, and human factors. In the second interim report (NRC, 1997), the CCAS recommended that the FAA not pursue accelerator-based technologies for primary screening of checked baggage and not fund development projects for large accelerator-based hardware. The CCAS concluded that the detection performance of these methods should be better understood before the FAA addressed airport integration issues and recommended that existing laboratory equipment be used to determine the detection limits of PFNTS for Class A¹ explosives (NRC, 1997).

In 1994, the FAA awarded Tensor Technology a two-year grant to build a multidimensional neutron radiometer (MDNR) airline security system. This project included building an airline security system, transporting the system to a nuclear accelerator, and testing the MDNR to determine its sensitivity for detecting explosives concealed in suitcases (Tensor Technology, 1998). The detection performance of the MDNR showed that it could potentially meet the probability of detection required for FAA certification for all but one of the required explosives categories. Based on these test results and in light of the recommendations of the CCAS, the FAA awarded Tensor a six-month cooperative agreement grant to compare the performance of PFNTS with the performance of other, currently available technologies for primary screening of passenger baggage for explosives and for screening cargo in airports.

In 1998, the FAA requested that the NRC review and evaluate Tensor Technology's assessment of PFNTS in light of the CCAS's recommendations (see Box ES-1) and technical developments since the second interim report. In response to the FAA's request, the NRC convened the Panel on Assessment of the Practicality of Pulsed Fast Neutron Transmission Spectroscopy for Aviation Security under the auspices of the CCAS. The panel was charged with evaluating

the practicality of PFNTS for primary screening of passenger baggage or for screening cargo, compared to currently available x-ray CT-based systems.

The panel examined the principles of operation of PFNTS and the results of laboratory-based blind tests on explosives in cluttered passenger bags. Some PFNTS tests demonstrated detection levels consistent with the FAA's EDS certification standards, but two important deficiencies were revealed. First, PFNTS did not demonstrate an ability to detect Class A explosives, an important class of explosives that most alternative technologies also have problems detecting. Second, PFNTS, when used with a two-dimensional area neutron detector,² had a higher false alarm rate than the FAA's EDS certification criteria allow.

Tensor provided a conceptual design of a PFNTS-based explosives-detection device (the MDNR) for implementation in an airport rather than a laboratory setting. The panel found that the MDNR provided a reasonable baseline conceptual design for assessing a PFNTS-based explosives-detection device for airport implementation. One of the unique characteristics of the MDNR design is the use of a cyclotron rather than a linear accelerator, which was used in all previous PFNTS testing. The use of a cyclotron would reduce the size of PFNTS and, therefore, make it more practical for airport integration. However, the reduction in size entails a substantial increase in system weight attributable to the heavy magnets in the cyclotron.

One could argue that the accelerator weight of 20 tonnes (22 tons) is not a severe penalty compared to the 109-tonne (120-ton) weight of the cyclotron shield enclosure and the 528-tonne (581-ton) weight of the vault enclosure required for any neutron-producing accelerator system. The panel identified size, weight, radiation shielding requirements, and complicated baggage flow constraints as problems that could arise during airport integration. The panel identified the use of commercial parts for the cyclotron and the detector electronics as an advantage of the MDNR over many other laboratory-based nuclear detection technologies.

The panel compared the practicality of the MDNR conceptual design to x-ray CT-based EDSs for implementation in airports. The laboratory-demonstrated explosives-detection performance of the MDNR was inferior to the laboratory-based performance of certified CT-based systems, both in the detection of Class A explosives and in the false alarm rate. The MDNR testing protocol differed from the CT system certification testing protocol, however, in the distribution of the subclasses of Class A explosives and the way false alarm statistics were collected. A direct comparison of the detection performance of x-ray CT and MDNR would require a significant increase in the statistical database for MDNR and a common test protocol. However, the available test results suggest that, even in the area of Class B explosives, the MDNR did not show a significant performance advantage over x-ray CT-based EDSs.

The resolution of false alarms is not part of the current EDS certification requirements, but it is an important consideration in the selection and fielding of equipment in airports. Laboratory test results indicate that the false alarm rate for PFNTS is between 13 and 25 percent. Evaluation of these results suggests that even an optimized PFNTS system would have a false alarm rate of at least 4 percent. In the absence of an acceptable alarm resolution protocol, this is an unacceptably high rate for airport implementation. Current x-ray CT systems in airports rely on operators to interpret high-resolution images to resolve automated alarms. Al-

though operator intervention could potentially lower the probability of detection for x-ray CT-based systems, it has been demonstrated to reduce the false alarm rate. The image produced by transmitted neutrons from a PFNTS-based device would not be sufficient for an operator to resolve automated false alarms. Unless a highly reliable alarm resolution method can be found, PFNTS would have to be combined with a high-resolution x-ray-based imaging technology for the purpose of alarm resolution.

The potential of PFNTS for screening small loose cargo packages has not been sufficiently explored using existing research accelerators. In addition, the characterization of air cargo (type, size, weight, delivery constraints, method of delivery to the airport) is not sufficient to develop a valid testing protocol, although the FAA provided the panel with a working threat definition (explosive type and quantity) to use in evaluating cargo inspection technologies. Based on existing cargo characterization data (FAA, 1996) and analytic estimates of neutron attenuation, the panel believes that PFNTS does not have a realistic potential for screening the full spectrum of cargo containers or pallets to this threat level. Cargo containers filled with cargo with high hydrogen content would attenuate the neutron transmission to a level comparable to the room neutron scattering background, thus rendering neutron transmission-based detection approaches ineffective. A thorough analysis of the potential of PFNTS for cargo scanning, however, will require a statistically significant set of explosives-detection test data with various types of cargo.

The greatest performance shortfall of PFNTS is its failure to detect Class A explosives. Unless this issue is resolved, PFNTS has no future as an explosives-detection technology in commercial aviation security. Even if this issue were resolved and the performance were equal to available FAA-certified EDSs, PFNTS in general—and the MDNR design specifically—has other disadvantages related to its size and weight that would preclude its selection for airport baggage scanning. Based on the FAA's current certification testing requirements for the detection of Class B explosives (as opposed to Class A explosives), PFNTS-based technologies would not be selected for primary screening of carry-on baggage, checked baggage, or cargo because of the difficulty of integrating these technologies into existing airport terminals and because of the safety issues associated with the operation of radiation-producing accelerators.

Tests have indicated that PFNTS has the potential for very low false alarm rates. If this potential were realized, then the PFNTS might play a role in aviation security but not as a primary EDS under current certification requirements. The panel concluded that only if the low false alarm rate is validated should the PFNTS-based system be taken through prototype development and demonstration in an airport environment. Even after airport testing, however, the system design would probably not be widely deployed in airports but would be placed on the shelf as a validated and characterized device that could be reevaluated as explosives threats changed or as regulatory requirements were refined.

At the first meeting of the Panel on Assessment of the Practicality of PFNTS for Aviation Security, the FAA asked that the panel address four questions during the course of the study. The panel and the NRC staff determined that these questions fall within the panel's Statement of Task. The questions are addressed below.

Question 1. Given the choice, would airlines select the PFNTS instead of equipment based on currently available x-ray CT for checked baggage inspection?

Answer. No, the airlines would not choose a PFNTS-based explosives-detection device for three reasons. First, PFNTS has not demonstrated an ability to meet the FAA's certification requirements for detecting Class A explosives. Second, the area detector configuration has not demonstrated an ability to meet the FAA's false alarm requirements. Third, because of the difficulties in deployment and integration of a PFNTS-based device, including size, weight, and safety issues, the airlines would choose currently available x-ray CT-based EDSs.

Question 2. If so, is their preference for this technology strong enough to justify the remaining costs to develop this technology (estimated to be $20 million to $30 million)?

Answer. Not applicable.

Question 3. Does PFNTS have any realistic potential for application to full cargo container inspections?

Answer. No. Based on Tensor's analysis of PFNTS for cargo screening of LD-3 containers containing single items (a reasonable projection given the sparseness of existing test data), PFNTS could only interrogate 57 percent of air cargo shipped in LD-3 containers. The other 43 percent contains large amounts of hydrogenous and, therefore, highly neutron-attenuating material, rendering PFNTS ineffective for screening. Further complicating the use of PFNTS for cargo screening is that containerized cargo is not always uniform in composition. The likelihood of false alarms from nonhomogeneous containerized cargo assembled by cargo consolidators has not been determined. Finally, the capability of PFNTS to detect explosives concealed in thick containers (e.g., the LD-3) has not been experimentally verified. Current estimates on neutron attenuation (based on crude exponential algorithms) are sufficiently accurate to raise con-

cerns about highly attenuating hydrogenous cargo. These estimates do not treat beam divergence from neutron scattering in the cargo and are, therefore, not sufficient to validate the PFNTS detection for low-attenuating scenarios.

Question 4. What experiments, if any, should be pursued in the near future to further define this potential?

Answer. Because PFNTS does not appear to be either practical or currently desirable for airport deployment, the panel does not recommend that experiments addressing the airport integration of PFNTS be pursued at this time. However, experimental verification of Tensor's simulation of PFNTS performance for cargo screening might be useful.

The inability of PFNTS-based explosives-detection technologies to detect Class A explosives at the probability of detection level required for EDS certification is a critical limitation. PFNTS-based techniques also demonstrated unacceptable false alarm rates when using area detectors and did not demonstrate a viable approach for resolving alarms. Unless and until these limitations are overcome, there is no reason for the FAA to pursue other technical or operational issues associated with integrating the technology into an airport setting.

Recommendation. The FAA should not fund the development of a prototype multidimensional neutron radiometer-based explosives-detection device.

Recommendation. At current levels of explosive threat and with the current state of the art, the FAA should not deploy pulsed fast neutron transmission spectroscopy-based explosives-detection technologies or devices for primary screening of carry-on baggage, checked baggage, or cargo.

Recommendation. At this stage, the FAA should not fund the development of an airport test facility.

Laboratory testing has not demonstrated the PFNTS technology to be technically desirable. However, because the threat to aviation security is dynamic, the requirements for explosives-detection systems certification may change over time. At some point, if existing deployed technologies do not provide adequate protection, a need for new explosives-detection approaches could arise. Among the technologies currently in development, PFNTS shows the most promise of meeting more stringent certification testing requirements because it is an element-specific detection technology. Therefore, even though the deployment of a PFNTS prototype designed for integration into an airport (e.g., MDNR) is not desirable at this time, valuable research could be conducted on the application of PFNTS technologies to explosives detection.

The research recommendations in this report are directed toward improving the detection performance of PFNTS-based explosives-detection technologies. However, even if these recommendations are followed and the detection performance is improved, practical limitations to the deployment of PFNTS-based devices remain (e.g., size, mass). These practical limitations should be taken into account when evaluating the potential of PFNTS for explosives detection in airports.

If funding becomes available for the development of PFNTS technology for explosives detection, the greatest benefit would be derived by addressing the current shortfalls of the technology rather than by developing and assembling a prototype MDNR unit for integration into an airport. Because the panel has not studied, and is not acquainted with, the whole spectrum of research requests for explosives-detection technologies submitted to the FAA, the panel neither supports nor opposes the allocation of research funds for further research on PFNTS.

Recommendation. The research priorities for pulsed fast neutron transmission spectroscopy should be directly related to the current shortfalls of the technology. Three major problem areas that should be addressed are listed below in order of importance:

Consistent with the recommendations in the second interim report of the CCAS, this panel believes that if PFNTS continues to be pursued it should be researched and developed with current neutron sources, detector technology, and radiation modeling capabilities. However, it appears that no existing research facility can meet the requirements for demonstrating the potential of PFNTS for explosives detection. If efforts are made to acquire a more compatible research facility, the FAA should acquire an existing commercially available accelerator (rather than funding the development of a new accelerator).

If funding is allocated for the construction of a facility to develop PFNTS for explosives detection, the FAA would derive the greatest benefit from a facility that promotes broad

cooperation and long-term multidisciplinary research. It should be noted that the panel does not support or oppose the allocation of funds for such a facility.

Recommendation. If the FAA acquires an accelerator to meet the PFNTS testing requirements, it should be configured to support a broad range of research activities.

FAA (Federal Aviation Administration). 1996. The Composition and Handling of Passenger Aircraft Cargo. DOT/FAA/AR-96/83. Washington, D.C.: Federal Aviation Administration.

NRC (National Research Council). 1998. Configuration Management and Performance Verification of Explosives-Detection Systems. National Materials Advisory Board. Washington, D.C.: National Academy Press. Available on the Internet: http://www.nap.edu/readingroom/enter2.cgi?0309061962.html

NRC. 1997. Second Interim Report of the Committee on Commercial Aviation Security. National Materials Advisory Board. Washington, D.C.: National Academy Press. Available on the Internet: http:// www2.nas.edu/nmab/262e.html

NRC. 1996. First Interim Report of the Committee on Commercial Aviation Security. National Materials Advisory Board. Washington, D.C.: National Academy Press.

NRC. 1993. Detection of Explosives for Commercial Aviation Security. National Materials Advisory Board. Washington, D.C.: National Academy Press.

Overley, J.C. 1987. Element-sensitive computed tomography with fast neutrons. Nuclear Instruments and Methods in Physics Research B24/25: 1058-1062.

Tensor Technology, Inc. 1998. Advanced Studies on the Multi-Dimensional Neutron Radiometer (MDNR) Airline Security System. Washington, D.C.: Federal Aviation Administration.