Title: Guide for the Selection of Commercial Explosives Detection Systems for Law Enforcement Applications Series: Law Enforcement and Corrections Standards and Testing Program Author: Charles L. Rhykerd, David W. Hannum, Dale W. Murray, and John E. Parmeter Published: National Institute of Justice, September 1999 Subject: Technology in law enforcement 111 pages 259,000 bytes ------------------------------- Figures, charts, forms, and tables are not included in this ASCII plain-text file. To view this document in its entirety, download the Adobe Acrobat graphic file available from this Web site or order a print copy from NCJRS at 800-851-3420 (877-712-9279 for TTY users). ------------------------------- U.S. Department of Justice Office of Justice Programs National Institute of Justice National Institute of Justice Law Enforcement and Corrections Standards and Testing Program Guide for the Selection of Commercial Explosives Detection Systems for Law Enforcement Applications NIJ Guide 100-99 ------------------------------- U.S. Department of Justice Office of Justice Programs 810 Seventh Street N.W. Washington, DC 20531 Janet Reno Attorney General Raymond C. Fisher Associate Attorney General Laurie Robinson Assistant Attorney General Noel Brennan Deputy Assistant Attorney General Jeremy Travis Director, National Institute of Justice Office of Justice Programs World Wide Web Site: http://www.ojp.usdoj.gov National Institute of Justice World Wide Web Site: http://www.ojp.usdoj.gov/nij ------------------------------- ABOUT THE LAW ENFORCEMENT AND CORRECTIONS STANDARDS AND TESTING PROGRAM The Law Enforcement and Corrections Standards and Testing Program is sponsored by the Office of Science and Technology of the National Institute of Justice (NIJ), U.S. Department of Justice. The program responds to the mandate of the Justice System Improvement Act of 1979, which created NIJ and directed it to encourage research and development to improve the criminal justice system and to disseminate the results to Federal, State, and local agencies. The Law Enforcement and Corrections Standards and Testing Program is an applied research effort that determines the technological needs of justice system agencies, sets minimum performance standards for specific devices, tests commercially available equipment against those standards, and disseminates the standards and the test results to criminal justice agencies nationally and internationally. The program operates through: The Law Enforcement and Corrections Technology Advisory Council (LECTAC) consisting of nationally recognized criminal justice practitioners from Federal, State, and local agencies, which assesses technological needs and sets priorities for research programs and items to be evaluated and tested. The Office of Law Enforcement Standards (OLES) at the National Institute of Standards and Technology, which develops voluntary national performance standards for compliance testing to ensure that individual items of equipment are suitable for use by criminal justice agencies. The standards are based upon laboratory testing and evaluation of representative samples of each item of equipment to determine the key attributes, develop test methods, and establish minimum performance requirements for each essential attribute. In addition to the highly technical standards, OLES also produces technical reports and user guidelines that explain in nontechnical terms the capabilities of available equipment. The National Law Enforcement and Corrections Technology Center (NLECTC), operated by a grantee, which supervises a national compliance testing program conducted by independent laboratories. The standards developed by OLES serve as performance benchmarks against which commercial equipment is measured. The facilities, personnel, and testing capabilities of the independent laboratories are evaluated by OLES prior to testing each item of equipment, and OLES helps the NLECTC staff review and analyze data. Test results are published in Equipment Performance Reports designed to help justice system procurement officials make informed purchasing decisions. Publications are available at no charge through the National Law Enforcement and Corrections Technology Center. Some documents are also available online through the Internet/World Wide Web. To request a document or additional information, call 800-248-2742 or 301-519-5060, or write: National Law Enforcement and Corrections Technology Center P.O. Box 1160 Rockville, MD 20849-1160 E-Mail: asknlectc@nlectc.org World Wide Web address: http://www.nlectc.org ------------------------------- The National Institute of Justice is a component of the Office of Justice Programs, which also includes the Bureau of Justice Assistance, Bureau of Justice Statistics, Office of Juvenile Justice and Delinquency Prevention, and the Office for Victims of Crime. ------------------------------- U.S. Department of Justice Office of Justice Programs National Institute of Justice Guide for the Selection of Commercial Explosives Detection Systems for Law Enforcement Applications NIJ Guide 100-99 Dr. Charles L. Rhykerd David W. Hannum Dale W. Murray Dr. John E. Parmeter Contraband Detection Technologies Department Sandia National Laboratories Albuquerque, NM 87185-0782 Coordination by Office of Law Enforcement Standards National Institute of Standards and Technology Gaithersburg, MD 20899-8102 Prepared for National Institute of Justice Office of Science and Technology Washington, DC 20531 September 1999 NCJ 178913 National Institute of Justice Jeremy Travis Director The technical effort to develop this report was conducted under Interagency Agreement No. 94-IJ-R-004, Project No. 97-028-CTT. This guide was prepared with the assistance of the Office of Law Enforcement Standards (OLES) of the National Institute of Standards and Technology (NIST) under the direction of Alim A. Fatah, Program Manager for Chemical Systems and Materials, and Kathleen M. Higgins, Director of OLES. This work was sponsored by the National Institute of Justice, David G. Boyd, Director, Office of Science and Technology. FOREWORD The Office of Law Enforcement Standards (OLES) of the National Institute of Standards and Technology (NIST) furnishes technical support to the National Institute of Justice (NIJ) program to strengthen law enforcement and criminal justice in the United States. OLES's function is to conduct research that will assist law enforcement and criminal justice agencies in the selection and procurement of quality equipment. OLES is: (1) Subjecting existing equipment to laboratory testing and evaluation, and (2) conducting research leading to the development of several series of documents, including national standards, user guides, and technical reports. This document covers research conducted by OLES under the sponsorship of the National Institute of Justice. Additional reports as well as other documents are being issued under the OLES program in the areas of protective clothing and equipment, communications systems, emergency equipment, investigative aids, security systems, vehicles, weapons, and analytical techniques and standard reference materials used by the forensic community. Technical comments and suggestions concerning this report are invited from all interested parties. They may be addressed to the Office of Law Enforcement Standards, National Institute of Standards and Technology, 100 Bureau Drive, Stop 8102, Gaithersburg, MD 20899-8102. David G. Boyd, Director Office of Science and Technology National Institute of Justice ACKNOWLEDGMENTS Many police officers and members of bomb squads helped evaluate our desired characteristics list for trace explosives detection equipment. They included: Arapahoe County Sheriff's Office Inv. Dan Davis, Sgt. Scott Linne, Sgt. R. R. Euchler, and Sgt. Joe Dempsey Charleston, S.C. Police Dept., Explosive Devices Team Cpl. Robert L. Kemp Eastern Kentucky University, College of Law Enforcement Prof. John T. Thurman Miami Dade Police Dept., Bomb Disposal Unit Lt. Eric Castaner National Law Enforcement & Corrections Technology Center - Rocky Mountain, University of Denver Research Institute Karen Duffala National Law Enforcement & Corrections Technology Center - Southeast Region Bill Nettles, Tommy Sexton New Mexico State Police Major Mike Francis, Major Roger Payne, Major Bill Relyea Prof. Gary Eiceman (New Mexico State University - Dept. of Chemistry) helped us with the protocol for evaluating explosives detection equipment. CONTENTS FOREWORD 1. INTRODUCTION 2. MARKET SURVEY OF COMMERCIALLY AVAILABLE EXPLOSIVES EQUIPMENT o 2.1 Portability o 2.2 Type of Item Being Screened o 2.3 Cost Range o 2.4 Throughput Rate o 2.5 Trace Explosives Detection o 2.6 X-ray Explosives Detection o 2.7 Canine Detection o 2.8 Novel Detection Techniques 3. SUGGESTED TECHNOLOGIES FOR PORTABLE, SEMI- PORTABLE, AND FIXED-SITE APPLICATIONS o 3.1 How to Choose Explosives Detection Technology for Specific Applications (Tables 7-9) o 3.2 Practical Notes Concerning Tables 7-9 4. DESIRABLE CHARACTERISTICS FOR EXPLOSIVES DETECTION EQUIPMENT FOR POLICE WORK o 4.1 Trace Systems o 4.2 X-ray Systems o 4.3 A Practical Example of How to Use Tables 7-11 5. CALIBRATION OF EXPLOSIVES DETECTION SYSTEMS 6. PROTOCOL FOR THE EVALUATION OF COMMERCIAL TRACE DETECTION SYSTEMS o 6.1 Introduction o 6.2 The Basics of Instrument Operation o 6.3 Protocol for Characterizing a Trace Explosives Detector 7. WARNING: DO NOT BUY BOGUS EXPLOSIVES DETECTION EQUIPMENT 8. CONCLUSION Appendix A-1 Suggested Readings Appendix A-2 Glossary of Terms Appendix A-3 Explosives: Nature, Use, Effects, and Applications TABLES Table 1. Cost Ranges for Trace and X-ray Explosive Detection Systems Table 2. Throughput Rates Table 3. Trace Detector Technologies and their Acronyms Table 4. Trace Explosives Detection Systems Table 5. X-ray Detector Technologies and Their Acronyms Table 6. X-ray Explosives Detection Systems Table 7. Matrices for Law Enforcement - Portable: Easily Carried by One Person Table 8. Matrices for Law Enforcement - Semi-Portable: Can fit in the Trunk of a Police Car Table 9. Matrices for Law Enforcement - Fixed Location, Dedicated Systems Table 10. Characterization of Commercial Trace Explosive Detection Systems Table 11. Characterization of Commercial X-ray Explosives Detection Systems Table 12. An Example of a Logbook Page FIGURES Figure 1. Vapor Concentrations of Neat High Explosives in Saturated Air at 25 degrees C Figure 2. Schematic of Ion Mobility Spectrometer (IMS) Operation Figure 3. IMS Spectrum, a Plot of Signal vs. Drift Time Figure 4. Commercial IMS Explosives Detection System, Barringer Ionscan 400 Figure 5. Commercial IMS Explosives Deetection System, Graseby Plastec Figure 6. Commercial IMS Explosives Detection System, ION Track Instruments Vapor Trace Figure 7. Commercial IMS Explosives Detection System, Intelligent Detection Systems ORION Figure 8. Schematic Diagram of Operation of an Electron Capture Detector (ECD) Figure 9. Commercial Electron Capture Detector (ECD), Ion Track Instruments Model 97 Figure 10. Photo of a Surface Acoustic Wave Detector, the EST Model 4100 Figure 11. Photo of a Thermo-Redox Detector, Sintrex/IDS EVD-3000 Figure 12. Photo of a Field Ion Spectrometer (FIS) by MSA Figure 13. Photo of the EXPRAY Field Test Kit for Explosives Figure 14. X-ray Explosives Detection System for Personnel, AS&E Body Search Figure 15. X-ray Explosives Detection System for Personnel, Rapiscan Secure 1000 Figure 16. Control Chart Using 5 ng of TNT from June 16, 1998 to July 13, 1998 COMMONLY USED SYMBOLS AND ABBREVIATIONS A ampere ac alternating current AM amplitude modulation cd candela cm centimeter CP chemically pure c/s cycle per second d day db decibel dc direct current C degree Celsius F degree Fahrenheit dia diameter emf electromotive force eq equation F farad fc footcandle fig. figure ft foot ft/s foot per second g acceleration g gram gr grain H henry h hour hf high frequency Hz hertz i.d. inside diameter in inch IR infrared J joule L lambert L liter lb pound lbf pound-force lgf-in pound-force inch lm lumen ln logarithm (base e) log logarithm (base 10) M molar m meter min minute mm millimeter mph miles per hour m/s meter per second N newton N-m newton meter nm nanometer No. number o.d. outside diameter ohm (ohm) p. page Pa pascal pe probable error pp. pages ppm parts per million qt quart rad radian rf radio frequency rh relative humidity s second SD standard deviation sec. section SWR standing wave ratio uhf ultrahigh frequency UV ultraviolet V volt vhf very high frequency W watt wavelength (wavelength) wt weight PREFIXES d deci (10[-1]) da deka (10) c centi (10[-2]) h hecto (10[2]) m milli (10[-3]) k kilo (10[3]) u micro (10[-6]) M mega (10[6]) n nano (10[-9]) G giga (10[9]) p pico (10[-12]) T tera (10[12]) COMMON CONVERSIONS (See ASTM E380) 0.30480 m =1ft 4.448222 N = lbf 2.54 cm = 1 in 1.355818 J =1 ft-lbf 0.4535924 kg = 1 lb 0.1129848 N m = lbf-in 0.06479891g = 1gr 14.59390 N/m =1 lbf/ft 0.9463529 L = 1 qt 6894.757 Pa = 1 lbf/in[2] 3600000 J = 1 kW-hr 1.609344 km/h = mph ------------------------------- NIJ GUIDE 100-99 GUIDE FOR THE SELECTION OF COMMERCIAL EXPLOSIVES DETECTION SYSTEMS FOR LAW ENFORCEMENT APPLICATIONS* This document includes a variety of information that is intended to be useful to the law enforcement community in the selection of explosives detection techniques and equipment for different applications. It includes a thorough market survey of all trace and x-ray based commercial detection systems known to the authors as of October 1998, including company contact information along with data on each system's cost, size, and uses. Information is also included on some additional novel detection technologies, and on such standard techniques as canine and physical search. Brief technical discussions are presented that consider the principles of operation of the various technologies. These may be ignored by readers who find them too technical, while those wanting additional technical information can obtain it from the extensive list of references that is included as an appendix. Other sections of the document present matrices listing the most highly recommended detection techniques for a variety of scenarios, a list of desirable characteristics for explosives detection equipment for law enforcement work with charts rating commercial systems against these criteria, and a standard test protocol for the evaluation of trace detection equipment. In addition to a reference list, the appendices include a section providing basic information about different types of explosives and explosions. Any law enforcement personnel having comments or questions are encouraged to contact the authors at Sandia National Laboratories. 1. Introduction The primary purpose of this document is to provide law enforcement agencies with information that should aid them in the selection and utilization of explosives detection equipment. The document is thus more practical than technical, emphasizing advice about the capabilities of different technologies, and what technologies are likely to work best in various applications. A wide variety of factors are considered that may be important to purchasers of detection equipment, including cost, sensitivity, portability, ease of use, etc. Some technical information is included in sections describing how the various detection technologies work, but the level of detail is not great. Readers finding this material too technical can skip it while still making use of the rest of the document, and readers desiring more technical detail can obtain it from the suggested readings in Appendix A-1. The remainder of this document is divided into several sections as follows: Section 2 presents a market survey of currently available explosives detection equipment. In tables 4 and 6, specific information is listed on 26 trace detection systems and 90 x-ray based detection systems. To the knowledge of the authors, this information is complete as of October 1998. The information in these tables includes the type of detector or detection technology used in each system, cost, recommended uses, system size and weight, and vendor contact information. In the case of x-ray based systems, the cost information can only be approximate, because most vendors are hesitant to quote a specific price for these (usually expensive) systems. The table on trace detection systems also includes sensitivity information as provided by the manufacturer, but it must be remembered that this information comes from the manufacturer and is not based on independent tests by a third party. Some independent test data exist for a few systems, and in the experience of the authors, the claims made by the manufacturers are usually not out of line with the system's true performance. However, all independent test results known to the authors are either classified or unclassified controlled nuclear information, so these results are not included in this document, which is intended for public release. Parties interested in what independent test data exist should contact the authors. Also included in section 2 are definitions of commonly used terms such as throughput rate and portability, a discussion of explosive vapor pressures and the issue of vapor sampling versus swipe (particulate) sampling in trace detection, and information on how the different trace and x-ray systems operate and what their general capabilities are. Finally, other techniques are discussed, including canine detection (already familiar to most law enforcement agencies, for drugs if not for explosives) and a few novel (though usually expensive and not always fully developed) detection technologies that have appeared in recent years. Section 3 presents three matrices that give recommendations about what technologies and systems to use for a variety of applications. These matrices provide a quick reference point for anyone having a specific application in mind, and wanting to know what sort of detection system he or she should consider purchasing. Five factors are included in defining the applications in these matrices: system portability, presence or absence of an explosives background in the area where screening will be performed, throughput rate, the type of item to be screened (people, packages, vehicles, etc.), and system cost. As a general rule, these matrices do not point to a single system or technique that is considered "best" in each circumstance, but rather point to several options that may meet the user's needs. Needless to say, in some cases there may be several detection systems that can do the desired job, while under other circumstances there may be no system that does everything the potential purchaser would like it to do. The latter is unfortunately most often true when severe cost restraints are placed on the system to be purchased, as is usually the case in this era of rapidly advancing technology but extremely limited budgets for technology in law enforcement. It must be stressed that the matrices in section 3 are intended to point the reader in the right direction, and they are a starting point rather than a solution in choosing a detection system. They are not a substitute for detailed discussions with both the vendor(s) and a knowledgeable third party. Section 4 discusses various characteristics and performance parameters that could be used to judge both trace and x-ray based detection systems, and defines ideal and nominal capabilities or characteristics for these systems. Defining these parameters is to some degree arbitrary since the "ideal" will of course depend upon the specific application. The definitions used are based on the best judgment of the authors and some feedback received from several law enforcement agencies, but it is really up to each potential purchaser to determine what the requirements are for his or her application. Tables 10 and 11 rate various commercial detection systems as ideal, nominal, or subnominal for the different parameters considered, and this allows the reader to focus on those parameters that are most important to him/her and to make rapid comparisons. Once again, these tables serve as a starting point for obtaining information and should supplement but not replace detailed discussions with the vendor(s) and outside experts. Section 5 briefly discusses the issue of system calibration. Since calibration is very system specific, little can be said in general about this topic. The best advice is to discuss the calibration procedure thoroughly with the vendor of the equipment, and if possible learn it hands-on from the vendor at an onsite installation and training visit. Section 6 provides a protocol for the testing and evaluation of trace detection systems. Such testing is also rather system specific, but this protocol has been made as generic as possible. The protocol should be of interest to users wanting to determine performance parameters for the specific unit they have purchased, and to those wanting to monitor the performance of the system over an extended period of time. The protocol also includes some basic information about sampling, ensuring a detector is free of contamination, etc. For some users in the law enforcement community, the protocol may be of little interest, and these users can skip section 6 without losing any content that is crucial to understanding the rest of the document. Section 7 contains a brief warning about buying equipment that may not be based on sound scientific principles. Briefly stated, detectors that appear to make unprecedented claims about detection capabilities may be based on faulty science, and in extreme cases could prove to be fraudulent. When dealing with technologies that appear to be new and report exciting new capabilities, it is especially important to discuss the purchase with an outside expert before making a final decision. Section 8 provides a brief summary and conclusions section. Appendix A-1 provides a list of suggested readings relating to the topic of explosives detection. Appendix A-2 provides a glossary of terms used in explosives detection, many of which may not be familiar to the average reader. Finally, Appendix A-3 contains an introduction to different types of explosives, their uses, and their properties. Any law enforcement agencies desiring more information about this document or explosives detection in general are invited to contact the authors[1] of this document. The contact points are as follows: 2. Market Survey of Commercially Available Explosives Detection Equipment In this section four different divisions of explosive detection technology are discussed. Covered in the greatest detail are (1) trace detection technologies and (2) x-ray based detection systems, since these are by far the most widely developed technologies. Canine detection, which is really a form of trace detection already very familiar to most law enforcement agencies, is discussed more briefly. Finally, some novel detection technologies are briefly discussed. Most of these novel technologies are not fully commercially developed, but they are mentioned here for the sake of completeness. Readers should note that they might become more readily available and also cheaper in the near future. Before the specific technologies are discussed, four key characteristics that help describe explosive detection systems and applications are defined and addressed. These characteristics are portability, the type of item being screened, system cost, and throughput rate. Some of this material is covered further in section 4, but a brief introduction is necessary at this point. 2.1 Portability Portability simply refers to the ease with which a detection system can be moved from one location to another. Depending on the application, portability may or may not be important. If a system needs to be carried by a detective who is screening a room for explosives, portability is clearly important. On the other hand, if it is only to be used as a dedicated system for screening people at a single entrance to a courthouse, portability is not important. In general, systems are divided into three categories: portable, semiportable, and fixed site. A system is considered portable if it weighs less than 9.1 kg (20 lb) and can be easily carried by one person (or, in the case of a canine, led by one person). It is considered semiportable if it does not meet these definitions, but can be moved easily by two people, fit in no more than two boxes, and be easily stored in the trunk of a police car. Systems that are too large or too heavy to meet the definition of semiportable are considered fixed-site systems. These can be very large and heavy, and include personnel portals and many baggage screening systems. 2.2 Type of Item Being Screened Explosives detection can be used in a variety of applications. Two major categories are search applications and screening of individual items. Search applications involve situations where a bomb is suspected of being in a general area, but the exact location is unknown. This would include, for example, searching a building or property grounds for a bomb, once a bomb threat has been communicated. In most search applications, canine detection will be the detection method of choice because of the dog's rapid mobility and its ability to follow the scent to its source. For screening of individual items, a wide variety of technologies can be useful in different situations. In general, the type of item screened will fall into one of four categories: people, hand-carried items, mailed or shipped items, and vehicles. Personnel screening involves detecting bombs or explosive material that is usually hidden under clothing. It can occur in many circumstances, ranging from suspect apprehension to screening large numbers of people entering a courthouse or some other facility. Screening of hand-carried items will usually occur alongside personnel screening. Specific items in this category include briefcases, backpacks, purses, hand- carried bags and packages, etc. Mailed and shipped items are used increasingly to transport bombs; these items can include letters, small packages, and large shipping crates. Vehicle screening can involve both single vehicles (e.g., suspect apprehension) and large numbers of vehicles at checkpoints. The vehicles involved can range from the smallest cars to fully loaded tractor-trailers. 2.3 Cost Range Cost is one of the few characteristics associated with a detection system that can be fully quantified, and hence it is one of the easiest to get a handle on when comparing different systems. Obviously, a purchaser will want to know the exact cost of any system he or she is thinking about buying. Nevertheless, it is convenient when starting to look at different systems to divide them into low-, medium-, and high-cost ranges. The definitions of these ranges are given in table 1; note that they differ for trace systems and x-ray systems. Cost ranges are used when discussing x- ray systems. Often x-ray manufacturers are reluctant to quote an exact price until they have talked to the potential buyer. Note that most trace systems cost less than $75K (several less than $30K), while most x-ray systems cost more than $100K. The authors realize that all of these numbers may seem staggeringly high to most police departments, and that cost will usually be one of the most limiting factors in making a procurement decision. Nevertheless, the ranges listed are convenient reflections of current costs in state-of-the-art explosives detection equipment. 2.4 Throughput Rate When screening individual items, throughput rate refers to the number of items that can be screened per unit time. It can also be expressed in terms of the time required to screen a single item. For example, if personnel are being screened at a checkpoint using a personnel portal that processes five people every minute, the throughput rate can be expressed as 5 persons/min, 300 persons/h, etc. Alternatively, the screening time can be expressed as 12 s per person. Throughput rate is typically an issue only when large numbers of items need to be screened rapidly. Table 2 quantifies high, medium, and low throughput rates for the items discussed in section 2.2. 2.5 Trace Explosives Detection Explosive detection techniques can be divided generally into two categories: bulk detection and trace detection. In bulk detection, a macroscopic mass of explosive material is detected directly, usually by viewing images made by x-ray scanners or similar equipment. In trace detection, the explosive is detected by chemical identification of microscopic residues of the explosive compound. These residues can be in either or both of two forms: vapor and particulate. Vapor refers to the gas- phase molecules that are emitted from a solid or liquid explosive because of its finite vapor pressure. Particulate contamination refers to microscopic particles of solid material that adhere to surfaces that have, directly or indirectly, come into contact with an explosive material. Explosive vapor pressures and their implications for detection are discussed in the next subsection, while the following subsection considers particulate contamination. Thereafter, several additional subsections are included on the different trace detection technologies that are currently available for explosive detection. Note that one consequence of using trace detection is that a valid alarm may be recorded for the object (person, package, vehicle, etc.) being screened, even if the object does not contain a concealed bomb. This can happen if the object has been contaminated with trace explosive material for any number of reasons, legitimate or otherwise. For example, when screening people it is possible to record a positive alarm for nitroglycerin from a heart patient using nitroglycerin tablets for medication purposes. For this reason, alarm resolution is always a key issue when utilizing trace detection technologies. 2.5.1 Vapor Pressures of Explosives To have a good understanding of the trace detection of explosives, it is important to understand the concept of vapor pressure. The vapor pressure of a solid or liquid substance at a given temperature is the gas phase pressure of the substance that exists at equilibrium above the surface of the solid or liquid. All solids and liquids emit a certain amount of vapor at all temperatures above absolute zero (-273 degrees C), and at a given temperature the amount of vapor emitted is characteristic of the particular substance. As an illustrative example, consider a piece of solid (2,4,6- trinitrotoluene (TNT)) placed in a jar with the lid closed. Before the TNT is placed in the jar, there is no TNT vapor present in the jar, but once the TNT is inside with the lid shut, the pressure of gas-phase TNT in the jar will increase as vapor molecules are emitted by the solid. Eventually, a state of dynamic equilibrium will be reached, where the number of vapor molecules emitted by the solid per unit time is the same as the number per unit time readsorbed by the solid and the walls of the jar. There will then be a constant pressure of TNT gas in the jar, and the quantitative value of this pressure is the vapor pressure of TNT at the prevailing temperature. Note that the vapor pressure of a chemical at a specific temperature is the maximum pressure of the gas that may exist above a solid or liquid. If the system has not yet reached equilibrium, the actual pressure of the vapor may be less than the vapor pressure, but never more. For convenience, vapor pressures are often expressed not in true pressure units but as relative concentrations in saturated air. Such concentrations are proportional to the true vapor pressure, and they often provide a clearer picture of the amounts of vapor that are involved. Figure 1 shows the vapor concentrations in saturated air of several high explosives at room temperature (25 degrees C or 77 degrees F). Note that the vertical axis of figure 1 has an increasing logarithmic scale, so that each higher mark corresponds to a factor-of-ten increase in vapor pressure. The horizontal axis displays the molecular weights of the various compounds, and is not important in the following discussion. It can be seen that the vapor pressures of the explosives shown vary widely. For example, the vapor pressure of ethylene glycol dinitrate (EGDN) is about 10(9) times (or one billion times) higher than the vapor pressure of HMX (homocyclonite, or octogen). In general, the explosives can be broken into three groups based on their vapor pressures. The high vapor pressure explosives include EGDN, nitroglycerin (NG), and 2,4-dinitrotoluene (DNT). These explosives have saturated vapor concentrations in air close to or greater than one part per million (1 ppm), which means that at equilibrium there will be roughly one molecule of explosive vapor per every million air molecules. The medium vapor pressure explosives have saturated vapor concentrations in air near one part per billion (1 ppb), a factor of about 1,000 lower than the high-vapor-pressure explosives. The medium-vapor-pressure group includes TNT and NH(4)NO(3) (ammonium nitrate). Finally, the low-vapor-pressure explosives have saturated vapor concentrations in air near or below the one part per trillion (1 ppt) level, approximately an additional factor of 1,000 lower than the medium-vapor-pressure explosives. The low-vapor-pressure group includes HMX, RDX (cyclotrimethylenetrinitramine or cyclonite), and pentaerythritol tetranitrate (PETN). These vapor pressures are for pure materials. Vapor pressures for mixtures containing these explosives may be even lower. The relative values of the vapor pressures mentioned above have important implications for the trace detection of explosives. The high-vapor-pressure explosives are relatively easy to detect from their vapor using detectors such as ion mobility spectrometers or electron capture detectors. Thus dynamites, which usually contain EGDN and/or NG as an explosive ingredient, can usually be detected from their vapor. Detecting these compounds by swiping surfaces for particle contamination (see next section) is also possible in some cases, but it may be less effective than with lower-vapor-pressure explosives, because the high-vapor pressures cause small particles to evaporate rapidly. In other words, in the case of particle sampling, the evidence may not be present long enough to make a detection. The medium-vapor-pressure explosives can sometimes be detected from their vapor, but in many cases this will test the limits of sensitivity for gas- phase detection, and particle detection based on surface swiping is usually preferred. Ammonium nitrate is a somewhat special case because it is almost always used in quantities of hundreds or even thousands of pounds in devices such as car bombs, and not in small bombs that could be carried on a person or shipped through the mail. Thus when ammonium nitrate is used, there is likely to be lots of contamination present to make a swipe-based detection, and various bulk detection techniques (e.g., x-ray) should also be effective. The low-vapor-pressure explosives do not produce enough vapor to be detected from their vapor in any but the most exceptional circumstances, and efforts to detect these compounds using trace technology must focus on swipe collection of particulate material. This makes swiping the preferred collection technique when dealing with plastic explosives such as C-4, semtex, and detasheet, which contain RDX and/or PETN as the explosive ingredient. The vapor pressure of a substance can be expressed as: where P(v) is the vapor pressure, P(0) is a constant with the same units as P(v), DG is the free energy of sublimation (for a solid) or vaporization (for a liquid) in units of J/mole, R is the gas constant in units of J/K x mole and T is the temperature in degrees K. An important point that can be gleaned from this equation is that P(v) depends upon the temperature as discussed above, and in fact the value of P(v) will increase exponentially with increasing temperature. Because of this exponential dependence, the effect of temperature on vapor pressure is quite dramatic. For example, for solid TNT near room temperature, the vapor pressure approximately doubles with every 5 degrees C (9 degrees F) increase in the temperature of the solid. Thus one cubic centimeter of air that is saturated with TNT vapor will contain about 0.096 ng of TNT at 25 degrees C, 0.19 ng of TNT at 30 degrees C, and 0.38 ng of TNT at 35 degrees C (1 ng = 10(-9) g = one billionth of 1 g). This means that one possible way to increase the chances of a successful vapor detection if a package or suitcase is suspected of containing a bomb is to heat the object. However, this is not always possible, and it can lead to interference problems if the object also contains another material that is more vaporous than the explosive. It should be pointed out that the numbers given for TNT vapor are very small compared to the amount of TNT contained in a typical particle in a fingerprint, which might contain several micrograms of TNT (1 ug = 1,000 ng). For a high-vapor-pressure explosive such as NG, the vapor concentration in air will be about 1,000 times higher than in the case of TNT. Therefore, the amount of NG vapor in a cubic centimeter (ccm) of saturated air will start to approach the amount present in a typical piece of particle contamination. A detailed report on the vapor pressures of several common high explosives has been published by Dionne et al. (Ref. 57 in App. 1). This study investigated the vapor pressures of TNT, RDX, PETN, NG, and NH(4)NO(3) over a wide range of temperatures. In each case, an empirical formula was derived for the vapor pressure over a certain temperature range. For example, in the case of TNT, it was found that the vapor pressure could be calculated from: where Log P(v)(ppt) is the base ten logarithm of the vapor pressure in units of parts per trillion (ppt), and T is the temperature in degrees K. This equation is valid for temperatures between approximately 21 degrees C and 144 degrees C. Similar equations for the other explosives, referred to collectively as the Dionne equations, provide a convenient means for estimating the vapor pressures of these explosives at different temperatures. A final important point about vapor pressures and vapor detection of explosives involves the low-vapor-pressure plastic explosives based on RDX and PETN. It has already been pointed out that RDX and PETN have extremely low vapor pressures, and the vapor pressures of the plastic explosives containing these compounds are even lower, due to the presence of oils and plasticizing agents that give the plastic explosive its form and consistency. When these explosives are manufactured, they are often spiked with a high-vapor-pressure nitro-compound called a taggant. Common taggants include ortho-mononitrotoluene (o-MNT), para- mononitrotoluene (p-MNT), and dimethyldinitrobutane (DMDNB). These taggants have vapor pressures similar to NG or EGDN, and their presence makes vapor detection of plastic explosives possible. However, relying on vapor detection with plastic explosives is still very risky, because old or homemade samples of plastic explosives will not contain the taggant. Nevertheless, detection of one of the taggants using gas-phase sampling with a trace detection system should be interpreted as possibly indicating the presence of a plastic explosive. 2.5.2 Particulate Contamination Particulate contamination, which can also be referred to simply as particle contamination, consists of microscopic solid particles, often with masses on the order of a few micrograms. Explosives in general tend to be rather sticky, and a person handling a macroscopic piece of the solid material will quickly transfer large amounts of such contamination to his or her hands. This material will then be transferred to any additional surfaces that are touched by the hands, which will likely include parts of the person's clothing as well as doorknobs, tabletops, and other objects he/she contacts. While it is hard to make generalizations, a typical fingerprint will contain many particles, often with a total mass on the order of 100 ug. For low- and medium- vapor-pressure explosives at room temperature, this amounts to more material than would be present in a liter of air saturated with vapor by a factor of 1,000 to 1,000,000. Thus, for these explosives, the ability to make detections based on particulate contamination is crucial, as was alluded to in the preceding section. Particulate contamination is usually sampled by wiping the surface to be screened with a swipe pad provided by the manufacturer of the trace detection system being used. Once this is done, the swipe pad can be inserted into a sampling port on the instrument, and in a matter of seconds it can be analyzed for the presence of explosives. This works best with briefcases and similar small packages. When screening people, this sort of surface swiping will necessitate physical contact with the test subject, and in some situations this may be considered excessively invasive. It should be noted that while careful handling of the explosive and the proper use of disposable gloves can greatly reduce the spread of particulate contamination, reducing it to zero is extremely difficult. Most bomb builders and carriers will not have the expertise required to do a clean job, so this method of sampling has very wide applications. 2.5.3 Trace Technologies The following subsections discuss specific trace detection technologies. A listing of different trace technologies and their acronyms is given in table 3. 2.5.3.1 Ion Mobility Spectrometry Ion mobility spectrometry (IMS) is one of the most widely used techniques for the trace detection of explosives and other contraband materials. The principle of operation of an IMS is shown in figure 2. The spectrometer consists of two main sections: the ionization region and the drift region. In a typical IMS, ambient air is drawn into an inlet port at the rate of a few hundred cubic centimeters per minute (ccm/min). The purpose of the instrument is to analyze this air for explosives or other compounds of interest, which may be present in the air in the form of vapor or airborne particulate matter. The air first enters the ionization region, where electrons interact with the incoming molecules to form positive or negative ions. In the case of explosives, it is negative ions that are formed. The source of the ionizing electrons is a small, sealed piece of metal that has been coated with a radioactive material, usually nickel-63 (63Ni{28}). Once ions are formed, they are periodically admitted into the drift region through an electronically shuttered gate. The ions are drawn through the gate by a static electric field, which pulls them towards a metal collection plate at the far end of the drift region. This "drift" of the ions from one end of the drift region to the other occurs at atmospheric pressure, with many collisions between the ions and the various molecules present. The time that it takes the ions to travel the length of the drift region is called the drift time, and is a complex function of the charge, mass, and size of the ion. Typical drift times are on the order of a few milliseconds (1 ms = 0.001 s). The current collected at the metal plate is measured as a function of time, and an IMS spectrum is a plot of ion current versus time, with different peaks representing different specific ions. An IMS spectrum of an air sample containing two types of explosives is shown in figure 3. Sometimes an additional gas called the dopant or carrier gas is admitted into the IMS to aid in the ionization process; very commonly methylene chloride or some other gas that easily forms chloride ions is used. Ions from this gas usually form the largest peak in the IMS spectrum, commonly known as the reactant ion peak or RIP, which serves as a reference peak. There are a number of features of IMS that make it attractive for the trace detection of explosives. This technique has probably been more widely developed for commercial applications of trace explosives detection than any other. A number of companies market IMS systems, including Barringer, Graseby, and Ion Track Instruments (see table 4). By the standards of most technology-based explosives detectors, IMS systems are moderately priced, with several systems in the $30K to $50K range. Upkeep costs vary from system to system, but are moderate in most cases. Most IMS systems are small and portable enough to be moved around in the trunk of a police cruiser, and can be operated by a person with only a few hours of training. These instruments have response times of only a few seconds, the proven ability to detect a number of key explosives, subparts per billion sensitivity in some cases, and audio and visual alarms that tell the operator when an explosive has been detected, and what type. The most effective means for collecting a sample for presentation to one of these systems is surface swiping, but vacuum collection of samples is also possible with many systems. Figures 4-7 show photos of some commercial IMS systems. Like all detection techniques, IMS also has certain weaknesses and drawbacks. As mentioned above, a radioactive material is used as the source of ionizing electrons in the ionization region. This source typically has a strength of about 10 mCi and does not pose any health risks if the system is operated properly, but simply having such a source may lead to some extra paperwork and regulatory oversight. Several attempts have been made to develop an IMS with a non-radioactive electron source such as a plasma discharge, but to date no such systems are commercially available. Most IMS systems do not run off batteries but rather require an electrical outlet, and this limits some field applications. There is a nontrivial warmup time, usually on the order of 10 min, associated with these systems. The drift time associated with a given ion is dependent on atmospheric pressure and can thus change during inclement weather or when the spectrometer is moved more than a few hundred feet in elevation. This requires little more than routine, periodic recalibration, but users need to be aware of this potential problem. Like other technology- based trace detection techniques, IMS systems cannot yet compete with canines in their ability to follow a scent to its source. Another drawback of IMS in some applications is that the peak resolution is not outstanding, and two different ions of similar size and mass may appear to give only a single peak rather than two distinct peaks in an IMS spectrum. One method of attacking this problem is to prefraction the molecules entering the IMS by first passing the incoming gas through a gas chromatograph (GC). A GC column is essentially a hollow tube, usually packed with beads that are coated with a special chemical substance, referred to as the stationary phase. This coating interacts more strongly with some molecules than with others, so if a gas flow containing different types of molecules is admitted into the GC, molecules that interact more strongly with the stationary phase will take longer to pass through the column. This means that an originally random mixture of different molecules can be sorted by type, with each species exiting the GC at a different time. The time it takes a certain molecule to travel through the length of the GC column is referred to as the retention time. If two molecules have identical drift times in an IMS, they will almost certainly have different retention times in the GC, and their peaks can thus be temporally resolved because they will enter the IMS at different times. A combined system of this type is referred to as GC/IMS, and such instruments are marketed by Intelligent Detection Systems (see table 4) for approximately $75K. 2.5.3.2 Chemiluminescence Most explosive compounds, including all of those shown in the preceding chart of gas phase concentrations in saturated air, contain either nitro (NO(2)) or nitrate (NO(3)) groups. The compounds commonly used as taggants in plastic explosives also contain NO(2) groups. Detectors based on chemiluminescence take advantage of this common property of most explosives by detecting infrared light that is emitted from electronically excited NO(2) molecules, denoted as NO(2)*. In a chemiluminescence system, explosive molecules are first pyrolyzed to produce nitric oxide (NO). The NO molecules are then reacted with ozone (O(3)) in an evacuated reaction chamber maintained at a pressure of about 3 torr = 0.4 kPa. This reaction produces the excited state molecules, NO(2)*. A photomultiplier situated behind a red light filter is used to detect the infrared photons of a characteristic frequency that are emitted when the NO(2)* molecules decay to form unexcited NO(2). The signal output measured by the photomultiplier is directly proportional to the amount of NO present in the reaction chamber, and this signal is thus used to detect the presence of explosives in a chemiluminescence system. The overall sequence of reactions can be summarized as follows (where the chemical equations have not been balanced): Used alone, chemiluminescence is not capable of identifying what type of explosive molecule is present. Indeed, all that can be said is that a molecule must initially have been present that decomposed to yield NO molecules, and such molecules include not only explosives and taggants but also species found in fertilizers, some perfumes, and other potential interferents. For this reason, chemiluminescence detectors are not used alone but are fitted with a front-end gas chromatograph (GC), as described in the section on ion mobility spectrometry. The GC allows different molecules that are detected with the chemiluminescence detector to be specifically identified based on their GC retention times, and the resulting system is referred to as a GC/chemiluminescence (GC/CL) detector. Systems of this type are marketed by Thermedics (see table 4). The best-known GC/chemiluminescence system is the Thermedics Egis. It is capable of analyzing samples in 18 s, and because of its high sensitivity and excellent selectivity it is a popular system with laboratory researchers and forensic analysts. However, its cost of approximately $150K is 2 to 3 times the cost of a typical IMS system. One nice feature of chemiluminescence detectors is that they do not utilize a radioactive ionization source, and thus avoid some of the paperwork and regulatory oversight that may be associated with IMS detectors. A drawback of chemiluminescence systems is their inability to detect explosives that are not nitro-based. 2.5.3.3 Electron Capture Detectors An electron capture detector or ECD detects explosives and other types of molecules having high electron affinities. A schematic diagram of a typical ECD detector is shown in figure 8. In an ECD, a vapor sample is drawn into an inlet port, and this vapor mixes with a stream of inert carrier gas (usually helium or argon). The gas flow then travels through an ionization region to an exhaust line. In transit, the gas flow passes through a chamber with a radioactive material that acts as an electron source, as in an IMS. The source material is usually either nickel-63 (63Ni{28}) or tritium. The emitted electrons become thermalized through collisions with the gas in the chamber, and eventually are collected at an anode. Under equilibrium conditions, there is thus a constant standing current at the anode. The basic principle behind an ECD is that this standing current is characteristic of the gas mixture being drawn into the system. If the gas mixture originally consists, e.g., of helium and room air, the standing current will be reduced if the vapor of an explosive enters the chamber. This happens because the explosive molecules have a high electron affinity and thus a tendency to capture free electrons and form stable negative ions, leaving fewer electrons to reach the anode. Thus, a reduction of the measured standing current is evidence that an explosive or some similar species is present. As with a chemiluminescence detector, the ECD by itself cannot distinguish individual types of explosives from each other or certain interferents, so a gas chromatograph is placed on the front end to allow temporal identification of different explosives. Several companies market detectors of this type (see table 4), referred to as GC/ECD detectors. A photo of a commercially available GC/ECD device is displayed in figure 9. This type of detector has a rapid response and typical sensitivities of about 1 ppb for most electron-capturing compounds. This sensitivity is somewhat less than the sensitivity of a typical IMS or chemiluminescence system, but it is still adequate for some applications. However, GC/ECD detectors usually cost less than IMS or chemiluminescence systems, with several systems available for $20K or less. These systems also tend to be smaller, lighter, and more easily portable. As with an IMS, the fact that the instrument has a radioactive ionization source can lead to some extra regulatory oversight. Another problem with ECD systems is that they require an ultrapure carrier gas, usually contained in a small cylinder, and the availability of this carrier gas can put limits on field applications. 2.5.3.4 Surface Acoustic Wave Sensors Surface acoustic wave (SAW) sensors are usually coupled with a front end GC, as is the case with ECD and chemiluminescence detectors. The principal component of a SAW sensor is a piezoelectric crystal that resonates at a specific, measurable frequency. When molecules condense on the surface of this crystal, the resonant frequency shifts in proportion to the mass of material condensed. The frequency shift also depends upon the properties of the material being deposited, the surface temperature, and the chemical nature of the crystal surface. In a typical system, the exit gas from the GC is focussed onto the SAW resonator crystal using a carefully positioned and temperature controlled nozzle. A thermoelectric cooler maintains the SAW surface at sufficiently low temperatures to ensure efficient trapping of the molecules of interest. The crystal surface can also be heated in order to desorb vapors and thus clean the surface. The temperature of the surface allows control of sensor specificity, by preventing adsorption of species with vapor pressures above a certain level. This feature is useful in distinguishing between high and low vapor pressure explosives. During sampling, vapors are concentrated in a cryo-trap before being desorbed into the GC for temporal separation. SAW sensors are marketed by Electronic Sensor Technology, Inc. (see Table 4). Pictured in figure 10 is EST's Model 4100. Total analysis time, including sample concentration in the cryo-trap, is typically 10 s to15 s. The system is advertised to have parts per billion sensitivity to certain types of explosives. It is about the size of a large briefcase, and is operational within 10 min of startup. The cost is about $25K, similar to some GC/ECD systems. 2.5.3.5 Thermo-Redox Detectors Thermo-redox technology is based on the thermal decomposition of explosive molecules and the subsequent reduction of NO(2) groups. Air containing the explosive sample is drawn into a system inlet at a rate of approximately 1.5 L/min. The air is next passed through a concentrator tube, which selectively adsorbs explosive vapor using a proprietary coating on the tube's coils. The sample is then pyrolyzed to liberate NO(2) molecules, and these molecules are detected using proprietary technology. Those interested in additional information on this technology should contact Intelligent Detection Systems (see table 4), which now markets the thermo-redox based equipment formerly marketed by Scintrex. The thermo-redox system currently marketed, the EVD-3000, is a hand- held unit that costs approximately $23K, pictured in figure 11. It can analyze both vapor and particle samples, and contains no radioactive source. Since only the presence of NO(2) groups is detected, this technology cannot distinguish among different explosives and potential interferences that contain NO(2) groups. Thus, the system identifies the presence of an "explosive-like" material, without identifying a specific explosive. Furthermore, the technology cannot detect explosives that do not contain NO(2) groups. 2.5.3.6 Field Ion Spectrometry Field ion spectrometry (FIS) is a relatively new trace detection technology (1994) that is related to IMS. It incorporates a unique ion filter using dual transverse fields, which allows interferences to be eliminated electronically, without the use of GC columns, membranes, or other physical separation methods. FIS is similar to IMS in that it involves separating and quantifying ions while they are carried in a gas at atmospheric pressure. Furthermore, both systems utilize soft ionization methods that yield spectra where the species of interest produce the main features (i.e., under proper conditions there is little decomposition of the analyte). In FIS, ions enter an analytical volume defined by a pair of parallel conducting plates where they execute two motions. The first is a longitudinal drift between the plates due to the bulk motion of a clean, dry carrier stream of air. The second is an oscillating motion transverse to the bulk flow velocity that occurs as the ions respond to an asymmetric, time- varying electric field imposed between the two plates. In response to the asymmetric field, the ions tend to migrate towards one of the plates where they will be neutralized. A second DC field is simultaneously established across the plates and can be used to balance or compensate for the drift introduced by the primary field. The DC field intensity needed to compensate for the AC field induced drift depends on the mobility of the particular ion under investigation, so that only specific ions can pass completely through the analytical volume and into the collection area for detection. Therefore, the device can be tuned to selectively pass only the ions of interest. Scanning the DC field intensity produces a spectrum of ion current versus field intensity that is known as an ionogram. The sole manufacturer of FIS sensors is Mine Safety Applications (MSA) --see table 4. Their system, pictured in figure 12, can currently be purchased for about $30K. The sensor has no moving parts except for a small recirculation fan and no consumables except a replaceable calibrator and gas purification filters. The size of the system is 0.022.66 m3 (0.8 ft3), excluding a computer for control and display. The manufacturer has reported detection limits for explosives such as TNT, RDX, and PETN in the low picogram range. To our knowledge, there have not yet been any independent tests to verify this. A response time of 2 s for a single target molecule plus another 5 s for each additional target molecule has been reported. Like an IMS, an FIS uses a small radioactive source for ionization. Because of the newness of this technique, the current systems may be better adapted to laboratory research than to routine field applications, but this could change in the future. 2.5.3.7 Mass Spectrometry Mass spectrometry (MS) has long been one of the most powerful techniques available for laboratory chemical analysis. Although it is rarely used in routine field applications and may thus be of little interest to law enforcement agencies doing explosive detection, it is discussed briefly here for completeness and because of this widespread laboratory use. While there are different types of mass spectrometers, this is basically a magnetic filtering technique. Molecules are first ionized and then passed through a magnetic filter, which allows ions to be identified based on their charge-to-mass ratio. In some systems, the MS is connected to a front-end GC. Mass spectrometers have excellent specificity for identifying different ions, and some systems have sub-picogram sensitivity. Mass spectrometers tend to be expensive. Table 4 lists one mass spectrometer system that costs $70K, with an advertised sensitivity in the low parts per billion range for some analytes. 2.5.3.8 EXPRAY Field Test Kit EXPRAY is a unique, aerosol-based field test kit for the detection of what the manufacturer refers to as Group A explosives (TNT, DNT, picric acid, etc.), Group B explosives (Semtex H, RDX, PETN, NG, smokeless powder, etc.), and compounds that contain nitrates that are used in improvised explosives. Detection of explosive residue is made by observing a color change of the test paper. EXPRAY can be used in a variety of applications, and although in some aspects it does not perform as well as many of the other trace detectors discussed in this section, it costs only $250. This very low cost, coupled with simplicity and ease of use, may make it of interest to many law enforcement agencies (see the EXPRAY kit in fig. 13). The EXPRAY field kit[2] is comprised of the following items: o one can of EXPRAY-1 for Group A explosives, o one can of EXPRAY-2 for Group B explosives, o one can of EXPRAY-3 for nitrate-based explosives (ANFO, black powder, and commercial and improvised explosives based on inorganic nitrates), o special test papers which prevent cross contamination. Initially, a suspected surface (of a package, a person's clothing, etc.) is wiped with the special test paper. The paper is then sprayed with EXPRAY-1. The appearance of a dark violet-brown color indicates the presence of TNT, a blue-green color indicates the presence of DNT, and an orange color indicates the presence of other Group A explosives. If there is no reaction, the same piece of test paper is then sprayed with EXPRAY-2. The appearance of a pink color indicates the presence of Group B explosives, a group which includes most plastic explosives. If there is still no reaction, the same paper is then sprayed with EXPRAY-3. The appearance of pink then indicates the presence of nitrates, which could be part of an improvised explosive. If EXPRAY-2 is applied after a positive result with EXPRAY-1, a change to pink color indicates a double or triple base explosive. The chemistry associated with these color changes is proprietary. Note that the order of spraying is critical, and all three sprays should be used in order to perform a complete test. The EXPRAY system has certain limitations. Obviously, it is not always possible to identify the specific type of explosive present when a positive result occurs. Sampling is entirely by surface swiping; there is no method for obtaining a vapor sample. Furthermore, not all types of explosives can be detected. Some nitrate esters and the chlorate group give a negative result. These include mixtures of potassium chlorate, sodium chlorate, and potassium nitrate with sugar, sulfur, and/or carbon. In addition, only the specific colors mentioned above can be judged a positive detection. Other discoloration is possible, but should be judged negative. The manufacturer claims that EXPRAY can detect particles as small as 20 ng. However, tests performed at Sandia National Laboratories show limits of detection for TNT that are 10 to 20 times greater than this. Nevertheless, this kit is still sensitive enough to make detections in many real world situations. 2.5.4 Table on Commercial Trace Detectors Table 4 gives information on various commercially available trace explosives detection systems. It must be emphasized that all information is subject to change, and the potential buyer should always check with the manufacturing company to obtain the most up-to-date information, including information on new systems. Companies can also provide advice on the suitability of their products for specific applications, though such advice will clearly not be disinterested. Bear in mind that all costs are approximate, and will depend to some degree on the exact options and accessories purchased. 2.6 X-ray Explosives Detection X-ray technologies are used to search for explosives in a variety of situations. X-ray systems are most often used for screening luggage, packages, mail, and other relatively small items. All of the systems involve irradiation of a target item with x-rays, followed by detection of an image created by x-rays that are either transmitted or backscattered by the item. The process of personnel screening is problematic because of privacy concerns and the perceived health problems associated with x- rays, but there are currently two different x-ray based personnel scanners on the market in the U.S. Vehicle screening systems are also available, though these are large, expensive, and require passengers to exit the vehicle while it is being screened. 2.6.1 X-ray Technologies The capabilities of x-ray systems range from those that produce a black and white picture to those that measure the effective atomic number (Z) of the screened items. The black and white images must be viewed and subjectively interpreted by an operator. Systems measuring effective Z can automatically alarm in the presence of materials that have an effective Z that is in the correct range for explosives. For most x-ray systems on the market, detection of a suspicious object is not automatic, and suspicious objects need to be identified by the person viewing the images. Two commercial x-ray systems for personnel screening (soft x-ray) are shown in figures 14-15. Many baggage-screening systems are also available and are similar to systems deployed at airport checkpoints. Such airport systems are undoubtedly familiar to the reader and are not pictured here. X-ray imaging systems can be divided into two broad types of categories: o systems that simply produce an image o systems that detect explosive-like materials and may generate automatic alarms. The simple black and white images from fluoroscopic imaging, image storage panel, and standard transmission systems cannot identify the actual explosive material, but do allow the operator to see wires, batteries, detonators and other bomb components. Fluoroscopic imaging refers to a transmission x-ray system, where the transmitted x-rays form an image of the objects investigated on a fluorescent screen. This is the simplest type of system, exposing the entire object with a cone of x-ray energy, often for extended periods. The fluorescent screen can be viewed in real time using a 45 mirror (to avoid standing in direct line with the x-rays), or can be used in conjunction with a video camera and a video image storage panel. Use of the camera and image storage panel minimizes the x-ray dose to the object being viewed, by allowing the image to be captured during a brief exposure. Even with a camera and storage device, this type of system is not film safe. The main reasons for the continued use of fluoroscopic imaging and image-storage panel systems are low cost and portability. A higher level of technology is the standard transmission x-ray. These systems use a fan or flying-spot of x-rays to scan the object as it is carried past the scanner by a conveyor belt. A black and white image is produced directly with a linear array of sensing diodes. The resulting image is stored in digital memory and displayed on a TV or computer screen for operator viewing. These systems are often found at airport checkpoints, where carry-on bags are screened. While standard transmission systems deliver a smaller x-ray dose to the object being screened (they are film safe), they produce black and white transmission images similar to the fluoroscopic systems and are subject to the same detection limitations. Specifically, standard transmission systems cannot identify the actual explosive material, but do allow the operator to see wires, batteries, detonators and other bomb components. In order to provide an operator with identification of explosives-like (low-Z) materials, other x-ray technologies like backscatter, dual energy, or computed tomography may be employed. Backscatter systems produce an image from x-rays that are scattered from the screened object. Because low-Z materials are more efficient at scattering x-rays, explosive-like materials are imaged as bright in the backscatter image, while they are barely visible in the transmitted image. The backscatter system produces two images and both the backscatter and transmission images are displayed. The backscatter image is usually most effective for the detection of low-Z materials such as explosives, narcotics, and plastic handguns while the transmission image is most useful for viewing metals. Backscatter systems cannot discriminate between various low-Z materials (e.g., between C-4 military plastic explosive and harmless plastic). In order to specifically distinguish explosive-like materials from common low-Z materials, other technologies are employed. Dual energy x-ray systems yield superior material discrimination through comparison of the attenuation of x-ray beams at two energies. Thus, identification of low-Z materials can be achieved by using incident x-ray beams of two distinct energies. Materials of specific Z numbers (the same effective Z as explosives) can be clearly highlighted for the operator by adding color to the image. A material that has a high Z number (metals) is often colored green, while low-Z materials are colored orange, and materials with the same Z as explosives are red. Some standard transmission systems may have color instead of black and white displays. Do not confuse dual-energy systems with colorized single energy transmission systems, which do not color the image on the basis of Z number, but rather color the image based on the level of x-ray energy that is transmitted. These simple single-energy colorized systems have never been proven more effective than the single-energy black and white systems, although their displays may look similar to a dual-energy display. Computed tomography (CT) is an even more sophisticated x-ray technique in which two-dimensional images ("slices") through an object are added together to produce a three dimensional image. Along with the three- dimensional image, the effective Z number is calculated and materials with the same Z number as explosives can be identified. This is essentially the same technique that may be familiar to the reader from medical CAT scans. Development of a dual-energy CT scanner is underway. This system may be capable of detection with lower false alarms than the single-energy CT scanners. Any system that can determine that low-Z materials are present can have an automated alarm function added. Systems that have been converted into automated detection systems include backscatter, dual energy, and the CT. In fact, CT systems are most useful as automated detection systems because the image they produce may be difficult for the operator to interpret (the image is often heavily distorted). 2.6.2 Table on Commercial X-ray Detectors Table 6 lists commercially available x-ray detection systems that can be used to detect explosives. Note that most of these systems are not easily portable, and are intended for screening at fixed checkpoints rather than for field use. Since costs vary widely depending on options and accessories selected, specific prices are not quoted, but rather the price category of each system is listed. Low cost is defined as less than $70K, medium cost as $70K to $300K, and high cost as greater than $300K. Again, it is very important to have thorough discussions with the vendors before making a purchase decision. 2.7 Canine Detection Trained canines provide a reliable and time-proven method for detecting concealed explosives. There is in principle no explosive compound that dogs cannot be trained to sniff out, and dogs used in the U.S. military are typically trained to detect nine different explosives. Compared to technology-based "sniffer" systems, dogs have the advantages of (1) superior mobility and (2) the ability to rapidly follow a scent directly to its source. Because of these advantages, canines are an excellent choice for explosives detection applications that involve a significant search component. Such applications include the search of vehicles, warehouses, luggage and cargo, aircraft, buildings, offices, and areas exterior to buildings such as parking lots, property perimeters, etc. Dogs are not usually used to search people because of the liability issues involved should a dog bite someone. Disadvantages of canines compared to trace technologies include: limited duty cycle (i.e., a dog works about 1 h before requiring a break), the need for regular retraining, and the inability to communicate to the handler the type of explosive that is detected. Law enforcement agencies can obtain canines from various civilian contractors. The best contact point is probably a law enforcement agency in the same area that already utilizes canines. The purchase cost is typically $5K to $10K per dog, and the initial training usually costs an additional $6K to $12K. This training typically involves two people and lasts up to three months. Once a dog has been procured, monthly training sessions are required. These are often performed in-house, so that the main cost is the time of the personnel involved. Feeding and caring for a dog (veterinary bills, etc.) typically costs approximately $1.6K per year. 2.8 Novel Detection Techniques This section lists several additional explosives detection techniques that have received at least some commercial development. Since most of these techniques are still in the process of development, it is especially important to contact the companies involved to obtain the latest information. (1) Thermal Neutron Activation (TNA): This technique is based on the interaction of neutrons with the nitrogen atoms contained in an explosive compound. A system of this type is produced by Science Applications International Corporation (SAIC), phone: (800)962-1632, fax: (619)646- 9718. The system is capable of detecting a number of explosives and has been used to screen check luggage in airports. However, it is very large 1363.6 kg (3,000 lb) and costs approximately $900K. Smaller and cheaper systems may be developed in the future, as the technology progresses. (2) Pulsed Fast Neutron Analysis (PFNA): This new and promising technique is also being developed by SAIC, which should be contacted if additional information is desired. It is also based on the interaction of neutrons with the explosive material, which results in the emission of gamma rays with characteristics that are indicative of the specific material. Preliminary testing has shown that many materials, including but not limited to explosives, yield unique signatures when this technique is applied. It has been used primarily for luggage screening. (3) Quadrupole Resonance: This is a promising new technology that has been extensively developed by Quantum Magnetics, Inc., phone: (619)566-9200. This technique uses pulses of radio-frequency energy to excite nitrogen nuclei within an explosive material, which then (4) emits photons of characteristic frequency when they relax. A number of systems are available for screening mail, packages, boxes, etc. Cost varies from $65K to $400K depending on the system selected. (5) Portable Isotopic Neutron Spectroscopy (PINS) Chemical Assay System: This system has been jointly developed by EG&G Ortec (phone: (423)482-4411) and the Idaho National Engineering and Environmental Laboratory. It is also based on irradiation of the explosive material with neutrons to produce gamma rays of a characteristic energy. It is designed primarily for field use in military applications, where it identifies explosives, nerve agents, and blister agents associated with firing ranges and munitions burial sites. 3. SUGGESTED TECHNOLOGIES FOR PORTABLE, SEMI- PORTABLE, AND FIXED-SITE APPLICATIONS This section presents three matrices (tables 7-9) that provide suggestions on which explosives detection technologies are likely to work best in a variety of different circumstances and applications. The emphasis is on what methods will work best in general, and not on what specific system should be purchased from a specific manufacturer. When two or more technologies will work equally well, all are listed. Note that high technology solutions are not always the best solutions. In a number of cases, canine detection or physical search (manual search) by police officers or security guards may prove more effective. 3.1 How to Choose Explosives Detection Technology for Specific Applications (Tables 7-9) The three matrices consider five factors that law enforcement personnel will want to weigh in making a procurement decision: o Degree of Portability: Table 7 is for portable applications, table 8 for semiportable applications, and table 9 for fixed-site applications. The definitions of portable, semi-portable, and fixed-site are as given in section 2.1. o Presence or Absence of an Explosives Background: For some applications that are primarily at a fixed site, the presence of background explosives contamination at the site could present problems. For example, if a detector is to be deployed near a bunker where explosives are stored, there will probably be a considerable amount of particulate contamination in the general area, transported by wind, people's feet, etc. In such cases, it may be undesirable to use trace detection and perhaps also canine detection in the area, because the contamination will lead to many nuisance alarms that need to be resolved. Bulk detection technologies or physical search are often preferable in such situations. o Item to be Screened: There are four categories of items - people, hand- carried items, mailed items, and vehicles, as discussed in section 2.2. o Throughput Rate: Alternatives are provided for low, medium, and high throughput rates, as defined in section 2.4. o System cost: Low-, medium- and high-cost range alternatives are considered, with the cost ranges defined as in section 2.3. Recall that the cost ranges are defined differently for trace and x-ray systems, in order to reflect the realities of the current market. 3.2 Practical Notes Concerning Tables 7-9 o If a low-cost option is listed, it can also be used in principle for the situation defined by the medium-cost and high-cost boxes just below it. However, it is sometimes the case that there is a trade-off between cost and performance, and sensitivity or some other performance characteristic may suffer if the cheapest possible system is chosen. o In some checkpoint screening applications, a high throughput rate may make it impossible to uniformly screen all items passing through the checkpoint. In these cases, a possible option is random screening. Random screening means that only a randomly chosen fraction of the passing items will be screened for explosives. For example, with a personnel portal, every fifth person passing the checkpoint might be screened, rather than all persons. The advantage of random screening is that it can reduce a high throughput situation to a medium or low throughput situation and thus provide more time to screen the individual items that are screened, while still retaining some deterrent effect. The disadvantage, obviously, is that it becomes possible for a bomb to pass through the checkpoint in an item that is not screened. In general, uniform screening is to be preferred to random screening whenever possible, but random screening is always an alternative. In tables 7-9, random screening is recommended only in cases where uniform screening would almost certainly be too slow. o Shaded boxes in the three matrices represent combinations of cost and throughput rate where it is difficult or impossible to perform uniform screening with current technologies. In many cases, these combinations are also unlikely to occur. For example, it will probably never be necessary to screen people with a high throughput rate using a portable system. A portable system for personnel screening is needed only for applications such as searching an apprehended suspect in the field, and in such an application a high throughput rate is not needed. The officers involved can, within reason, take as much time as they need to search the suspect. o Metal detection is listed in a few places in these matrices where options are rather limited, but it must be remembered that metal detection is of limited value in searching for explosives. It can find primers and metallic bomb components, but not the explosives themselves. o In the matrices, the range of recommended trace or x-ray systems in a particular situation is given by using the numbers assigned to detection systems in tables 4 and 6. Thus, for example, "Trace 1-7" indicates that any of the first seven systems listed in table 4 might be a satisfactory choice for the given situation. Readers of this document are encouraged to study tables 7-9, and focus on those circumstances they believe will be most important to their intended applications. Once again, it must be emphasized that these matrices serve as a starting point, and not an endpoint, in choosing an appropriate detection system. Obtaining product literature, speaking with the vendor(s), and holding discussions with an outside expert are still necessary to make the best-informed decision possible. 4. DESIRABLE CHARACTERISTICS FOR EXPLOSIVES DETECTION EQUIPMENT FOR POLICE WORK Sections 4.1 and 4.2 present tables and discussion dealing with desirable characteristics for trace and x-ray explosives detection equipment that might be used in law enforcement work. In the case of trace detection systems, 22 different characteristics are discussed in table 10. A much smaller number of defining characteristics for x-ray systems is in table 11. 4.1 Trace Systems In looking at desirable characteristics for detection systems, a useful approach is to define "ideal" and "nominal" characteristics, and then to classify each detector (when possible) as ideal, nominal, or subnominal with regards to that particular characteristic. While the definitions of ideal and nominal are necessarily somewhat arbitrary and application dependent, this categorization allows comparisons to be made easily and rapidly. However, it must be remembered that a law enforcement agency needs to procure a system based on the characteristics that are most important to it, and these will vary widely. In other words, there is no "one size fits all" method of choosing detection equipment for law enforcement work. For some applications, it may be crucial that a system meet the "ideal" standard for one or more characteristics, while for other characteristics a subnominal rating may be perfectly acceptable. Note that no system in table 10 meets all of our ideal characteristics, and it is very likely that no system will meet any other set of "ideal" specifications that might be defined. The enumeration of "ideal" and "nominal" characteristics in this section should not be construed as providing a list of requirements; it is intended only to provide information. As with the matrices in section 3, this table should serve only as a starting point in making a procurement decision. Furthermore, it must be remembered that some of the characteristics considered are not fully quantifiable, and in such cases the associated "ratings" (ideal, nominal, subnominal) are necessarily subjective and could be open to debate. With these caveats in mind, the following characteristics are considered: (1) System Size: Ideal capability: The system can be easily carried by one person. Nominal capability: The system fits easily in the trunk of a standard police car. Comments: At present, few commercial systems meet the ideal capability. The two that do are the Scintrex/IDS EVD-3000 and the ITI Vapor Tracer. Most commercial trace detection systems meet the nominal capability, though a few of the larger ones such as the Thermedics EGIS do not. Obviously, personnel portals also do not meet the nominal capability. (2) System Weight: Ideal capability: The system consists of one piece that weighs less than 9.1 kg (20 lb), so it can easily be hand-carried by a single operator. Nominal capability: The system can be contained within two crates or packages, each with a weight of 22.73 kg (50 lb) or less. Comments: Only a few commercial trace detection systems meet the ideal capability, but most meet the nominal capability. (3) Cost: Ideal capability: The system costs less than $30K. Nominal capability: The system costs less than $100K. Comments: Most of the commercial systems now available cost $75K or less. Few are less than $25K, although some like the Scintrex/IDS EVD-3000 and the ITI Vixen are. These two systems are based on thermo-redox and ECD technology. Those technologies are typically less sensitive compared to systems employing chemiluminescence or IMS technology. The typical cost for IMS systems is $40K to $75K. (4) Explosives Detected: Ideal capability: The system can detect TNT, RDX, PETN, NG, EGDN, DNT, ANFO (ammonium nitrate), double base smokeless and black powders, and common taggants such as DMDNB and mononitrotoluenes. Nominal capability: The system can detect TNT, RDX, PETN, NG, and black powder. Comments: These are the common high explosives that police officers are likely to encounter in everyday situations. Some less common explosives that are likely to be very rare in real-life situations are not considered here. Little emphasis is placed on chemical taggants because (a) there is lots of untagged plastic explosive material around (C-4, semtex), and (b) it is assumed that most sample collection will utilize surface swiping, in which case the explosives are easily detected without taggants. Most commercial systems do not detect all explosives equally well in practice. For example, IMS systems can detect nitrate-based explosives such as ANFO and black powder, but not as effectively as they detect TNT, RDX, etc. All of the commercial systems that the authors are familiar with can detect the nominal list; most can detect the ideal list. (5) Ruggedness: Ideal capability: The system is encased and protected to withstand wind, dust, and rain. It must also withstand the routine shocks and vibration associated with transporting it. Nominal capability: The system is generally rugged for field applications, but will not be operated during inclement weather (and thus exposure to dust, rain, and severe winds will be minimized). It must also withstand the routine shocks and vibration associated with transporting it. Comments: All current systems meet the nominal capability. Some testing would be needed to determine which systems meet the ideal capability. (6) Field Operation (Power): Ideal capability: The system can operate on batteries. Nominal capability: The system can operate from a standard 110 V AC power outlet. Comments: While battery operation is ideal, many commercial systems do not offer this option. All systems meet the nominal capability. (7) Maintenance: Ideal capability: Nominal maintenance required, perhaps every six months, and can be performed by a law enforcement officer. Nominal capability: Monthly replacement of consumable chemicals, filters, etc. Comments: Most systems will meet the nominal capability. (8) System Calibration: Ideal capability: A simple calibration procedure should be provided by the manufacturer and well documented in a manual. Under normal weather conditions, not more than one recalibration should be required per day, assuming continual operation of the equipment. Nominal capability: As for the ideal case, except that recalibration up to once per hour is permissible, assuming continual operation of the equipment. Comments: Most current systems meet the ideal capability under normal weather conditions. However, pressure changes during storms can rapidly change peak positions in IMS spectra, resulting in the need for frequent recalibration during inclement weather. This may be a problem with some other technologies as well. Police departments may wish to consider questions such as "How many days a year are there severe storms in this region?" before purchasing a system. Note also that moving an IMS system to a new elevation may also require recalibration, due to the pressure change involved. For most systems, this will not be an issue if the elevation change is less than plus/minus 91.44 m (plus/minus 300 ft). (9) Start-up Time: Ideal capability: Less than 10 min from a warm start, and less than 20 min from a cold start. Normal capability: Less than 1 h from a cold or warm start. Comments: For most systems, the start-up time will be considerably less if the system has been turned off briefly after running for a long time (warm start) than if it has been turned off for a long time (cold start). Thus, it may be desirable to have a system running continually at headquarters, and turn it off only when it needs to be transported to a nearby locale. This would certainly result in quicker start- ups in field applications. (10) Sample Collection Mode: Ideal capability: All target explosives can be detected using vapor (vacuum) collection. Nominal capability: All target explosives can be detected using swipe collection. Comments: It has been our general experience that IMS and chemiluminescence systems have the sensitivity to detect all the target explosives in a vapor collection mode (though the authors have relatively little experience with ANFO and black powder). The ECD and thermo-redox systems fail to detect at least some of the explosives in this collection mode. Data concerning vapor collection with GC/SAW systems and some less commonly used technologies was not reviewed. All systems should be adequate when collecting samples via swiping surfaces. For most applications, surface swiping should be permissible. However, there are a few applications where vacuum collection may be preferable. For example, obtaining samples from a suspect's clothing is perhaps best done by vacuuming, since physical contact with the suspect (swiping his clothing) might be considered excessively invasive (i.e., a violation of his right to privacy, or unreasonable search without probable cause, etc.). Surface swiping will also lead to greater concern about cross contamination of samples. (11) Sample Collection Time: Ideal capability: Ten seconds to 5 min (variable depending upon how the sample is obtained). Nominal capability: One minute to 5 min (variable depending upon how the sample is obtained). Comments: All systems known to the authors can meet the ideal capability. (12) Sample Analysis Time: Ideal capability: The system gives an "answer" less than 30 s after a sample is inserted into the instrument. Nominal capability: The system gives an "answer" less than 2 min after the sample is inserted into the instrument. Comments: Most if not all commercial trace detection systems meet the ideal capability. In general, this means that sample collection rather than sample analysis will usually be the rate limiting step. (13) Limits of Detection: Ideal capability: Can detect 100 ug of each target explosive in the vapor collection mode at least 95 percent of the time. Nominal capability: Can detect ug of each tarrget explosive in the swipe collection mode at least 95 percent of the time. Comments: The 100 ug figure is chosen because this is approximately the amount of particle residue contained in a typical fingerprint. Some current systems meet the ideal capability for all target explosives, while virtually all of them meet the nominal capability. For swipe collection, the true detection limit will depend to at least some degree on the surface the explosive is deposited on. Note that for ANFO the detection limit is likely to be less important than for the other explosives, because ANFO is almost always used in very large quantities. (14) Alarm and User Notification of Detection: Ideal capability: The system has both an audio and a visual alarm, and tells the user what type of explosive has been detected. A spectrum may be displayed, but the system must not require the user to interpret the spectrum in any way. Nominal capability: The system should have an audio alarm or a visual alarm. It will tell the operator that an explosive or explosive-like material has been detected, but will not identify the type (in this sense, it will be similar to an explosives sniffing canine). Comments: Most trace detection systems meet the ideal capability. A few, like the Scintrex/IDS EVD-3000, meet only the nominal capability. (15) False Positive Rate: Ideal capability: Less than 1 percent in laboratory tests. Nominal capability: Less than 5 percent in laboratory tests. Comments: These numbers should be based on laboratory tests because it ought to be possible to obtain clear-cut answers by doing the proper studies. Until such studies are performed, no definite answers can be given. However, all systems can meet the nominal capability, and many will probably meet the ideal capability. (16) Nuisance Alarm Rate: Ideal capability: Less than 1 percent when handled by police officers that are reasonably clean but may have had recent contact with firearms and ammunition. Nominal capability: Less than 5 percent when handled by police officers that are reasonably clean but may have had recent contact with firearms and ammunition. Comments: Nuisance alarms are alarms caused by true detections, but where the material detected is only innocuous contamination that does not result from the presence or handling of an illegal bomb or any other threat item. Some systems could give nuisance alarms when handled by law enforcement personnel that have been in contact with weapons. (17) Probability of Detection: Ideal capability: Greater than 99 percent for a swiped fingerprint. Nominal capability: Greater than 95 percent for a swiped fingerprint. Comments: Testing needs to be done to determine these values, but all systems should meet the nominal capability and several will meet the ideal capability. A swipe-friendly (smooth and hard) surface with deposited explosive is assumed. Note that this capability sets the false negative rate, since the false negative rate is simply one minus the probability of detection. (18) Data Storage: Ideal capability: The data obtained can be stored in the system and printed out later. Nominal capability: Data can be saved if the system is connected to a computer at the time of data collection. Comments: Most current trace detection systems meet the nominal capability. (19) Ease of Use: Ideal capability: No factory training required. Could be operated by an average police officer with one day of training. Nominal capability: No factory training required. Could be operated by an average police officer with no more than three days of training. Comments: This is a very important criterion, but it is also subjective and difficult to define. Most current trace detection systems probably meet the ideal capability as defined here. (20) Legal Issues: Ideal capability: The system should be certified to meet certain judicial standards, such as the Dow and Frye standards. There should be a proven history of data obtained with the system standing up in court. Nominal capability: There must be a reasonable expectation that the data obtained with the system can stand up in court, based on the record of similar instruments. Comments: Some detectors, such as the Barringer Ionscan, have undergone certain forms of legal certification. Probably all of the detectors considered could get such certification. Determination of the current legal status of each instrument should be based on up-to-date conversations with the vendor. (21) Drug Detection: Ideal capability: The system can also detect key illegal drugs such as cocaine, heroine, marijuana, and methamphetamine when operated in the proper mode. Nominal capability: No drug detection capability. Comments: It is up to each law enforcement agency that procures a detector to decide if they would like to use it as both an explosives detector and a drug detector. Most IMS-based systems can perform drug detection, but only when operated in a mode where explosives cannot be detected. Thus, these systems can detect drugs and explosives, but not at the same time. The chemiluminescence, ECD, and thermo-redox systems cannot detect drugs. (22) Radioactivity: Ideal capability: The system contains no radioactive source. Nominal capability: The system can contain a sealed radioactive ionization source with a strength of less than 50 mCi. Comments: Chemiluminescence detectors contain no radioactive source and hence meet the ideal capability. GC/SAW and thermo-redox detectors also contain no radioactive source. On the other hand, IMS and ECD detectors have a radioactive source. Several companies are doing research to develop nonradioactive IMS sources (Corona Discharge IMS), but none are commercially available yet and it is not certain how soon these will be available for reliable field use. The potential problem with these radioactive sources is one of paperwork and convenience, than one relating to human health, because the low energy electrons emitted are easily stopped by skin. Police departments need to be sure that if they purchase a detector with a radioactive source, they can transport it freely within their area of operation. Keeping a logbook of where the system goes each day is probably reasonable and even desirable, but any paperwork that is required beyond this would be a burden and could inhibit rapid field deployment. Most commercial detectors will be acceptable for police use with regards to this issue. 4.2 X-ray Systems Table 1, dealing with desirable characteristics for x-ray based detection systems, is somewhat different than table 10. Most significantly, there are far fewer characteristics that are considered. Two of the most important are the capability to detect low-Z materials and the presence of an automatic alarm. For both characteristics, each system has this capability as a standard feature (+), has it available as an option (=), or does not have it (-). Mobility is also considered, with each system classified as portable (P), mobile (M - mounted in a vehicle), or for use at a fixed site (F). Cost ranges are listed as defined earlier. Note that two cost ranges are listed for some systems (e.g., Low/Med); this means that the cost range may change depending upon what options are purchased. Once again, this table is intended only to provide information, and it can only be a starting point in making a procurement decision. 4.3 A Practical Example of How to Use Tables 7-11 Consider a law enforcement officer responsible for responding when suspicious packages are reported. Assume that any technology that the officer would use must be portable, easily carried by one person. Table 7 (Portable) is selected, because the additional technologies listed in table 8 (semiportable) and table 9 (fixed-site) are too cumbersome for this application. The Hand-Carried Item description fits the suspicious package, so the second set of three rows of table 7 is selected. In most cases, there will be No Explosives Background, so the right half of table 7 is used, for that second set of three rows. This area contains nine combinations: one for each combination of three costs and three throughputs. In those nine areas in table 7, all the High Throughput boxes are shaded. These are shaded because it is unlikely that an officer would have to screen items at a high throughput in a field situation. In this example of a suspicious package, there is only one suspicious package to screen, so a relatively Low Throughput rate can be accepted. The Low Throughput column has now been selected, where it intersects with the Hand-Carried Item rows in table 7. The only variable left to consider is the cost of the technology. In the Low cost ($0K to $30K trace; $1K to $70K x-ray), several Trace (1- 7) and x-ray (1-9) technologies are listed. These numbers refer to table 4 for Trace systems and to table 6 for x-ray systems. (Consider only the trace systems to simplify this example.) Looking up the name of Trace system number 6 in table 4, the Scintrex/IDS EVD-3000, pictured in figure 11 is found. The EVD-3000 costs $23K according to the table, and it weighs 3.18 kg (7 lb). If there is uncertainty about how it rates against Trace systems 1-4, table 10, Characterization of Commercial Trace Explosive Detection Systems, can be consulted. In table 10, look at the columns labeled Maintenance and Limits of Detection. The EVD-3000 has subnominal (black) maintenance characteristics as does Trace system number 3, while numbers 1, 2, and 4 have nominal (blue) maintenance characteristics. But when the Limits of Detection column is reviewed, the EVD-3000 is the only one of these five systems that is nominal, while the other systems, numbers 1-4, are rated sub-nominal. So, a judgment has to be made whether maintenance or detection is more important to consider. Assume that better limits of detection are desired. Then the EVD-3000 would be selected over the other four trace systems. (If the $23K price tag was too high, a department might choose one of the cheaper systems, 1-5, on that basis.) What if better limits of detection are needed? What if the EVD-3000 only rated nominal, while we wanted an ideal rating? Go back to table 7, and look at the available options. The higher cost selections have not yet been considered. Trace systems number 8 and 9 are listed in the Medium cost ($30K to $100K) box. Comparing systems 8 and 9 in table 10, system 9 has more pluses, i.e., ideal attributes than 8, and only costs $3K more. Look at that system for comparison with the EVD-3000. Trace system number 9 is found in table 4, where it is listed as the Ion Track Instruments ITMS Vapor Tracer, pictured in figure 6. The Vapor Tracer costs $38K and weighs 3.18 kg (7 lb) according to that table. Its characterization from table 10 is that the Vapor Tracer has a nominal maintenance rating and an ideal limits of detection rating. If the main consideration is detection limit, and if $38K is an acceptable price for the organization, then the Vapor Tracer would be selected over the EVD- 3000. 5. CALIBRATION OF EXPLOSIVES DETECTION SYSTEMS Commercially available explosives detection systems operate using different principles, and it is difficult to make generalizations about appropriate calibration procedures. One generalization that can be made is that all systems made by reputable companies should come with a detailed user's manual, and this manual should give a recommended procedure for system calibration. Depending on the system, the recommended calibration procedure may not be fully quantitative; the emphasis is usually on ascertaining that the system works at some minimal level. The manufacturer should normally be questioned about calibration and maintenance before a system is purchased. In most cases, a company representative can provide onsite training when the system is first purchased, giving one or more employees at the purchasing site an opportunity to receive hands-on training from an expert operator. The importance of such training can hardly be overstressed. Commercial trace detection systems use a variety of sample types for calibration, but all calibration procedures involve challenging the system with a minute amount of one or more of the target explosives. Several systems come with sampling pads that have already been spiked with known amounts of certain explosives. These can be inserted into a sampling port for analysis, and the lack of a detection from such a pad would indicate that the system is not working properly. If a detection is made, the intensity of the signal can provide a semi-quantitative measure of the system's sensitivity to the explosive involved, though it would be more accurate to use sampling pads spiked with explosives from freshly made standard solutions. At least one commercial system comes with a lipstick-like substance that contains trace quantities of several different explosives. This system can be tested by challenging it with a sampling pad onto which a small amount of the "lipstick" has been rubbed. Some systems that are used primarily for gas phase sampling come with small vials of material that emit vapors when opened, and can be tested by holding the open vial up to the system's sampling inlet. Such systems could perhaps be calibrated more accurately using one of the explosives vapor generators that have been described in the literature [1-4], but a calibration of this type would be expensive and time consuming, and is probably beyond the scope of most law enforcement work. Bulk detection systems such as x-ray scanners are often calibrated and tested by challenging them with a threat object hidden in a piece of luggage or some other type of hand-carried item. This is not always a straightforward process, because many systems do not alarm automatically but rather require a person to determine whether a threat item is present on a displayed screen image. Thus human factors are involved and the calibration process is not perfectly defined. When looking for explosives, it is desirable not to work with real explosive materials, because for bulk testing it would be necessary to work with a macroscopic sample that is capable of detonation. For this reason, a material called a simulant is often used in place of the actual explosive. A simulant is simply an innocuous material that has similar properties to the explosive when probed with the type of incident radiation the system uses. For example, if a transmission x-ray system is being used to study detection of detasheet, a good simulant would be any material that has an x-ray mass absorption coefficient similar to detasheet. In the past, Vivid Technologies Inc., has sold simulant materials suitable for the testing of dual-energy transmission x- ray detection systems. Other manufacturers may also have identified simulant materials that they would be willing to sell, or provide with the systems that they sell. Other calibration measurements for x-ray systems include those using a step wedge. The step wedge as defined by the American Society of Testing and Materials (ASTM) is a series of steps of aluminum with a series of thin wires of various gauges attached beneath. The purpose of this test device is to determine that the system can produce different distinct image intensity levels for each step while clearly displaying the smallest specified wire beneath each step. Typical x-ray systems can image a minimum of 34AWG solid copper wire through a minimum of 10 steps. Many modern systems can perform at higher levels with 38AWG and 20 steps are common. 6. PROTOCOL FOR THE EVALUATION OF COMMERCIAL TRACE DETECTION SYSTEMS This document describes standard procedures for testing various important performance benchmarks of commercial, trace explosives detection systems. These benchmarks include probability of detection, detection limit, false-negative rate, false-positive rate, nuisance-alarm rate, interference response, throughput rate, sampling time, analysis time, and total processing time. Note that other test protocols have been developed by various agencies and organizations, and interested readers can contact the authors for further information. 6.1 Introduction The tests described in this document are intended to be general enough for application to all or most commercial trace detectors. Consequently, comparisons between the performance of various detectors and documentation of existing detectors in service might be made uniform or convenient. However, these tests are not replacements for user's manuals, and readers should become thoroughly familiar with the instrument through reading the user's manual provided by the manufacturer. If any instructions provided in this protocol contradict instructions given in the user's manual, operators should follow the instructions given in the user's manual, or at the very least contact the company marketing the equipment to discuss the matter. A procedure not included here, because it is very instrument specific, is that of instrument calibration, including the setting of alarm levels for various explosives. All calibrations should follow the procedures given in the user's manual. When a system is purchased, most vendors will provide some onsite training so that the calibration procedure can be learned directly from an expert operator employed by the company. 6.2 The Basics of Instrument Operation (I) Blank Samples: Verifying Freedom From Contamination When presenting any chemical detection system with a sample (also known as "challenging" the system), it is imperative that the system is clean (free of the substance being detected prior to the challenge). If the system is contaminated, an apparent detection may result from the contamination that was already present in the system and not from the sample being analyzed. Clearly, this could lead to false-positive results. For example, in a portal system that screens people for explosives, it might be concluded that a certain person has had contact with explosives, when in fact he has not. For this reason, trace detectors must be certified as clean after a positive response or alarm shows the presence of explosive material. In addition, the system must be verified as clean before analyzing the first sample. This is done by challenging the system with blank samples that are known not to contain explosive material. For example, if swipe pads for sample collection are being used, pads that have come straight out of the package supplied by the vendor and have not been exposed to any explosive can be tested. Thus at the start of each period of use of the instrument and after the processing of any sample that results in an alarm, challenge the system with a blank to see whether or not an alarm is recorded. If no alarm is recorded, proceed to process additional samples. (When doing rigorous testing, it might be desirable to run as many as three blanks between samples.) If an alarm is recorded, it almost surely results from contamination of the system. In this case, keep challenging the system with blanks until no alarm is recorded. In extreme cases of exposure to very large explosive masses, it might be necessary to let the system sit for several hours with the detector at a high temperature in order to purge the system of all the explosive contamination. (II) Collecting the Explosive Material: Swipe Versus Vapor Collection With most commercial trace explosives detection systems, there are two common means of collecting samples: swipe collection and vapor collection. In swipe collection, a sampling pad (usually supplied by the manufacturer) is wiped across a surface suspected of having residue of explosive material. This surface could be a tabletop, the outside of a package, a piece of luggage, clothing, and so forth. The sampling pad is then inserted into a sampling port on the instrument for analysis. In contrast, vapor collection involves the use of a small hand-held vacuum to collect airborne vapors or particles. Typically, vacuuming is performed just above the surface to be investigated. A collection filter is located inside the inlet of the vacuum, and air is drawn through this filter. The explosive material will be trapped on the filter. The filter is then removed and analyzed by the system in a manner similar to the analysis of a swipe sample. Vapor sampling of this sort is generally less sensitive than swipe sampling, but it is advantageous for screening people because it is not necessary to touch the person being screened. Thus, taking samples with vapor collection is regarded as less invasive than collecting swipe samples. (III) Explosives Solutions and Proper Handling Successful work with trace explosives detectors usually requires detection of explosive residues, amounts of material that are so small that the actual sample cannot be seen by the human eye. Most trace detectors will show a positive response to amounts of explosive material on the order of a few nanograms or less (one nanogram equals one billionth of one gram). Even in the preparation of solutions of known composition and content (standardized solutions), only microscopic amounts of material are needed to perform instrument testing. One benefit of this is that the usual hazards of working with explosives are obviated, since such tiny amounts of explosive cannot produce a detonation. A common method of handling microscopic amounts of explosive material is to use explosives dissolved in volatile organic solvents. For example, a typical solution might be TNT dissolved in methanol at a concentration of 1 ng of TNT per microliter of methanol. (Such solutions can often be purchased directly from commercial chemical suppliers.[3] Once obtained, the solutions can be used as received, or further diluted to provide lesser concentrations.) A known amount of explosive can be obtained from such a solution by withdrawing a known volume of solution (typically measured in microliters by a syringe) and then depositing this volume onto a sampling pad. A typical calculation is: Other terms for mass and volume may be used so long as the units are consistent throughout the formula. The volatile solvent will evaporate rapidly at room temperature and the measured amount of explosive material will remain on the sampling pad. While a thorough discussion of the proper handling of chemical solutions is beyond the scope of this document, always make sure that the work area is free of explosive residues, the syringes and glassware that are used are clean, and in particular that a syringe used to measure out a dilute solution is not contaminated with a more concentrated solution. (IV) Logbook and Record Keeping An easily overlooked aspect of making trace chemical measurements is record keeping on a sample and on the instrument used for analysis. In both laboratory and field use, a record of who, when, where, what, and why should be kept. This record or logbook is essential for legal purposes and other uses such as instrument maintenance. Notice that consecutive analysis numbers are given for each use of the instrument and that the sample's identity is recorded elsewhere in the Description column. Moreover, the table above is an example of minimum content for a logbook. Other details might include where the data are stored, a reference to the origin of the standards, etc. In addition, the results from the calibrations might be maintained elsewhere (see Sec. VII) in a type of maintenance record to anticipate or schedule routine repair and upkeep of instruments. Experimental parameters such as temperature and pressure should be included in the logbook as shown in the fourth column; these will show only minor changes in laboratory use, but may show wide variations in field use. (V) Data Storage and Backup Trace detectors may be connected to small computers where results of analyses are processed and findings are displayed. The level of technology on these detectors may range from hand-held units with on-board computers without data storage capabilities to laboratory models with modern computers and large storage capacities. In instances where analysis results are stored, efforts should be made to copy the results to another storage medium and the results archived for future use. The organization and records for such archives should be well maintained and secure, particularly for legal purposes. The possible corruption of records through technology failures (copying errors) or other failures (misplace records or damaged disks) must be avoided. In addition, hardcopy records might be maintained as the ultimate backup. (VI) Control Charts and Long-Term Instrument Performance A routine procedure with trace detectors should be the use of blanks and standards to guarantee freedom of contamination and proper responses to known amounts of chemicals. This determines the quantitative response of the detector and if it changes over time. Often loss in performance for a detector is seen in the quantitative response and can suggest need of maintenance. All trace detectors require routine maintenance, so the frequency and cost of such service should be anticipated. One helpful tool is the control chart. A control chart is a graph of calibration results of an instrument on a daily, weekly, or monthly basis. Some control charts may span years. In the control chart shown in figure 16, the response of an IMS based detector to 5 ng of TNT was monitored (given a certain amount of normal variability) from the middle of June until the middle of July. Such control charts are often created and kept by specialized personnel whose responsibility is to inventory and maintain field instruments or by regular laboratory personnel operating the detectors daily. Field or enforcement personnel should not ordinarily be responsible for this somewhat tedious aspect of instrumentation use. (VII) Experimental Parameters Experimental parameters are those quantities and conditions that define or describe the conditions under which an analysis or measurement occurred. When testing a commercial trace detection system, the experimental parameters will include both system operating parameters (e.g., detector temperature, alarm level, etc.) and external parameters (ambient temperature, ambient pressure, etc.). When determining detector performance, the values of all experimental parameters should be recorded. Changing any experimental parameter could change the final value of Probability of Detection [P(d)] or any other experimentally determined quantity, so the value of the experimentally determined quantity is only meaningful if all experimental parameters are defined. Some experimental parameters will have a large impact on the experimentally measured quantity, while others will have little or no impact. Minor changes in external parameters such as ambient temperature usually will have only a small impact. During any testing to characterize a detector's performance, the experimental parameters should be carefully monitored. 6.3 Protocol for Characterizing a Trace Explosives Detector This protocol includes procedures for determining the following benchmarks, each of which will be described in the sections below: o Probability of Detection o Detection Limit o False-Negative Rate o False-Positive Rate o Nuisance-Alarm Rate o Interference Tests o Throughput Rate o Sampling Time o Analysis Time o Total Processing Time 6.3.1 Probability of Detection - P(d) P(d) should be determined or defined for each type of sample collection. The procedure outlined below concerns the most common method for explosives screening where a known amount of explosive from a solution is deposited directly onto a sampling pad. However, a very similar procedure could be followed to determine a P(d) based on other types of sampling. 1. See that the system is turned on, calibrated, and that the alarm level is properly set. 2. Using a syringe, place a known amount of explosive in a solution of a volatile solvent onto the center of a sampling pad (e.g., 1 ng of TNT in 1 uL of acetonitrile). Use the type of sampling pad recommended by the manufacturer for that instrument. 3. Wait for the solvent to evaporate. 4. Present the sampling pad to the instrument as appropriate for that system. For many systems, this will mean inserting the pad into a heated sample port on the instrument. For others, this may mean rubbing the sampling pad over an inlet where material is sucked into the system. 5. Observe the system response, and record whether or not an alarm occurs. You will probably want to save the spectrum produced for future reference. 6. Present the system with clean pads (a blank), and observe whether or not an alarm is recorded for the explosive in question. If three consecutive clean pads produce no alarm, the system can be assumed to be clean and can again be challenged with a pad containing explosive material. 7. Repeat this procedure for a total of 20 measurements. P(d) is then = [number of alarms recorded/20]. P(d) can also be measured for the following sampling methods: (i) known explosive mass deposited onto a designated surface, then swiped; (ii) known explosive mass deposited onto a surface, then vacuumed; (iii) known vapor dose directed into sampling pad (perhaps contained in a vacuum device); and (iv) known vapor dose directed into an instrument using a vapor generator. Not all of these modes are possible with every instrument. 6.3.2 Detection Limit (DL) The DL is defined here as the lowest mass of explosive material with a P(d) of 0.9 or higher, i.e., the lowest amount that will cause the system to alarm on > 90 percent of the challenges. Note that this is dependent upon where the alarm level is set for a particular system. In principle, it is desirable to test the detection system with a variety of different masses for a given explosive, and the lowest mass that gives a P(d) of 90 percent or greater will be the DL. general procedure is outlined below. Once again the example of known explosive masses deposited directly onto a sampling pad of the appropriate type is used. 1. See that the system is turned on and properly calibrated, and that the alarm level is set at the appropriate level. 2. Challenge the system with a blank sample pad to verify that it is clean. If no alarm is recorded, proceed to the next step. If an alarm is recorded, repeat this step until no alarm is recorded. To save pads, this step may be repeated with the same pad initially if desired, but always challenge the system with at least one new clean pad and obtain a result of "no alarm" before moving on to the next step. 3. Challenge the system with a sample mass that is suspected of being well above the DL. This might take some guesswork, but the system manual and a little experience can be of assistance. If an alarm is recorded, move to the next step. If no alarm is recorded, the mass chosen was too small. After running a blank sample to verify system cleanliness, double the mass, and see if an alarm is recorded. Repeat this procedure until an alarm is recorded. 4. Perform a P(d) test for the mass where the alarm is recorded, as described above in the section on Probability of Detection. If P(d) for this mass is 90 percent or greater, this can serve as the starting point for the test performed in step (5). If P(d) is less than 90 percent, double the mass and again perform a P(d) test. Repeat this procedure until a mass is found that is above the DL, i.e., that has a P(d) of at least 90 percent. 5. Once such a point above the DL has been found, perform a series of tests where the system is challenged with ten different masses, ranging from no explosive to the mass determined in step (4) in equal increments. For example, if it was determined in step (4) that a mass of 2 ng was above the DL, the masses tested should be (0, 0.2, 0.4, 0.6, 0.8, 1.0, 1.2, 1.4, 1.6, 1.8, and 2.0 ng). Start with the lowest mass (no explosive) and work upwards, performing two tests for each mass and recording whether or not alarms are recorded. Make sure that a clean blank sample is run and that no alarm is recorded following each challenge with explosive material. 6. Find from the test in step (5) the lowest mass for which two alarms were recorded, and run a P(d) test with twenty challenges at this mass. This mass should be very close to the DL. If P(d) for this mass is found to be less than 90 percent, perform a P(d) test at the next highest mass, and keep moving up in mass until a mass is tested where P(d) is greater than 90 percent. If, on the other hand, P(d) is greater than or equal to 90 percent for the mass chosen initially, move down to the next lowest mass and perform a P(d) test, and continue to move down until a mass is reached with P(d) less than 90 percent. Once this procedure is complete, the lowest mass tested with P(d) > 90 percent can be taken as the DL. The DL can be determined more precisely by testing at additional explosive masses between the DL determined above and the mass immediately below it where P(d) < 90 percent. However, in practice, this more precise determination will rarely be worth the effort. 6.3.3 False-Negative Rate For a given set of experimental conditions, the false-negative rate is simply one minus the probability of detection. Thus, if P(d) = 0.8, the false negative rate is 0.2, or 20 percent. All experimental parameters must be identical for this to hold true. If any parameter is changed, determine a new P(d) before calculating the false negative rate. 6.3.4 False-Positive Test The following test procedure can be followed for a laboratory test of the false-positive rate. The example of clean sampling pads, though vapor tests with clean air could be used in some cases, is used. 1. Make sure the system is turned on, calibrated, and that the alarm level is set properly. 2. Challenge the system with a clean sample pad. If no alarm is recorded, the system can be assumed to be clean and the testing can begin. If an alarm is recorded, continue to challenge the system with clean pads until no alarm is recorded three consecutive times. 3. Challenge the system with clean sample pads 20 times, and record each time whether an alarm is recorded and for what explosive. If an alarm is recorded, interrupt the test by challenging the system with blank pads that are not counted towards the total of twenty. When three consecutive pads produce no alarm, the system can be taken to be clean again. At this point, continue where you left off in the test sequence of twenty pads. The false-positive rate is then [number of alarms in the test sequence of twenty pads/20]. A similar and somewhat more useful test could also be performed to determine the false-positive rate under realistic field conditions. However, in this case the situation becomes more complicated, because "false positives" obtained in the field may actually be due to real explosives contamination. If this is so, they should be classified as "nuisance alarms" rather than "false alarms." Therefore, some form of alarm resolution (questioning of people screened or further chemical analysis of samples) needs to be conducted in order to correctly classify the alarms. Alternatively, all alarms can simply be classified as "false positives," and the false-positive rate thus determined will represent an upper limit on the actual false alarm rate. 6.3.5 Nuisance-Alarm Rate Nuisance-alarms are alarms that result from the actual detection of an explosive, but where the explosive material present originates from an innocuous source rather than from a threat item. For example, when screening people for explosives at a checkpoint, a nitroglycerin alarm could result because the test subject is a heart patient with nitroglycerin tablets, rather than a terrorist attempting to smuggle a bomb containing nitroglycerin. Similarly, an alarm for TNT could result from a construction worker employed at a blasting site, an alarm for black powder or some related substance could result from anyone having recently handled firearms for a legitimate reason, etc. In law enforcement applications, nuisa