Opportunities to Improve Airport Passenger Screening with Mass Spectrometry (2004)

Mass spectrometry (MS) is an obvious candidate to consider for improving the IMS-technology-based trace detection systems currently deployed in airports across the United States. It has become the gold standard for resolving high-consequence analyses involving water, air, and ground pollution; pharmaceutical drug development and manufacture; treaty compliance verification relating to proliferation of nuclear materials; verification of employee drug abuse for prosecution and job termination; detection of performance-enhancing drugs in horses and athletes; and routine analysis in the chemical, drug, and fuel manufacturing industries.¹

While mass spectrometers have become one of the analytical mainstays of today’s chemistry and biotechnology laboratories, they have historically been large, complex systems that occupied the volume of several file cabinets, were operated by highly trained mass spectrometrists, and sold for around $250,000. The generalized use of the instruments by lab chemists and technicians has led to automated, self-calibrating, auto-tuning, benchtop units of reasonable size costing $50,000 to $100,000. These instruments are generally coupled with a gas chromatograph (GC) or a liquid chromatograph (LC) at the sample inlet to further enhance chemical selectivity. Some special-purpose instruments have been miniaturized for mobile applications,² though the performance and reliability of these miniaturized systems are still being assessed. In general, however, all of these instruments operate at high vacuum, and they need professional care and trained operators.

Mass spectrometry is not new to the TSA; as a result of a TSA program focusing on the use of MS technology, researchers concluded that “there are no major technical barriers to its use in the field for trace detection scenarios, provided effective sampling and introduction procedures for the specific application are employed.”³ Currently, the TSA is testing a personnel screener utilizing an MS-based system from Syagen Technology and a portal from Sandia National Laboratories.⁴ MS has also been

applied by others for the automatic analysis of samples acquired from boarding passes (see Box 2-1).⁵ In the committee’s view, such systems have the potential to add significantly to future trace detection capabilities for a variety of threat substances in the transportation context, as discussed below.

Mass spectrometers use four steps for analysis: vaporizing the sample; placing an electric charge on sample molecules to form ions; separating the ions based on their charge-to-mass ratio using an electric or magnetic field; and, finally, determining the number of separated ions having a particular mass-to-charge ratio (i.e., producing a “mass spectrum”). Some mass spectrometers that operate in this fashion are called quadrupole mass spectrometers (QMS), sector field mass spectrometers, and ion trap mass spectrometers. In a time-of-flight (TOF) mass spectrometer, ionized sample molecules in a vacuum are accelerated in a straight line so that they fly down an evacuated tube. By measuring how long the ions take to reach a detector at a fixed position and by taking into account the length of the flight, one can determine the mass-to-charge ratio and the number of ions detected at each sequential mass.⁶

Several varieties of hybrid and single analyzer mass spectrometers are used for analysis. One of the most common for utilizing multiple stages of mass spectrometry (MS/MS) is a combination of three quadrupole structures in sequence (triple quadrupole) and an ion-trap mass spectrometer. Increased chemical specificity is achieved in the triple quadrupole configuration because the first mass spectrometer can be used to select a single mass corresponding to the compounds of interest in the sample (targeted analysis), rejecting all the other compounds of differing molecular weight that could interfere with the detection and identification. Ions having the desired mass are then collided in the second quadrupole, and the resulting fragment ions are analyzed in the third quadrupole. This multiple mass analysis technique provides information about the structure of the original molecule and confirms the detection of the target compound. This same chemical specificity can be achieved in a single-ion-trap mass spectrometer by holding the ions of interest in a three-dimensional ion trap, colliding them with neutral gas molecules, collecting the fragments, and sweeping the fragments in order of ascending mass out of the trap for detection.

Additional chemical specificity can be attained by preseparation of target molecules from a complex mixture using chromatography in tandem with MS (C/MS or C/MS/MS). C/MS is generally accepted as the final word in a variety of applications that require high confidence in the identification of a substance. Laboratory-based mass spectrometers, especially when used in combination with other separation methods such as chromatography, have the ability to identify and quantify targeted chemicals based on molecular weight and structure in very complex mixtures at picomole (10^-12 mole) to attomole (10^-18 mole) levels.

As noted above, any mass spectrometer requires a method of forming ions from the molecules in the sample. The highest chemical specificity and (usually) sensitivity are achieved if the ions formed are those of the molecule itself (molecular weight) rather than fragments of the molecule. Two common methods of soft ionization achieve this: matrix-assisted laser desorption/ionization (MALDI) and

electrospray ionization (ESI). In MALDI, the sample is mixed with a larger quantity of an organic molecule (the matrix), which is selected for its ability to efficiently absorb radiation from a laser. When the dried sample-matrix mixture is exposed to a laser beam, the matrix absorbs the laser energy and transfers it to the sample, typically forming positive ions with a single charge. In ESI, a sample in a liquid is continuously aerosolized into a fine spray near a needle maintained at high electrical potential, and the droplets take on a charge from the electric field. As the charged droplets evaporate, the charge is transferred to organic molecules in the sample, forming molecular ions that can be separated in the MS.

The uniqueness of mass spectrometry lies in its chemical specificity. Because it directly measures a fundamental property of the target molecule—its molecular weight—it affords a highly specific means of identifying the molecule. By contrast, IMS systems measure a secondary and less specific property of the target molecule—the time it takes for the molecule to drift through a tube filled with a dense gas—and the identity of the molecule is inferred by reference to calibration standards. Since different molecules may have similar drift times, IMS inherently has less chemical specificity than MS.

Since there is no quantitative calibration of airport IMS systems⁷ and no systematic reporting of screened objects and alarms, there are no reliable data with which to properly asses the current IMS instrumentation in terms of probability of false alarms and probability of detection. Instead, another method of comparison is adopted here (described below) that might be used in the future for comparisons of technologies when operational data are unavailable.

The chemical specificity of an instrumental method can be quantitatively estimated on a consistent basis using a metric called “informing power.” In a discussion of the informing power of tandem mass spectrometry (MS/MS), Yost and Fetterolf⁸ use information theory to give figures of merit for chemical resolution of various analysis techniques. The committee has estimated the informing power of IMS using the same method (see Appendix A) and compares it with the previously calculated informing power of MS/MS systems in Table 2-1. The informing power of the tandem QMS/QMS configuration is on the order of 10,000 times greater than that of IMS.

The substitution of an MS/MS analyzer for the analyzer of an ion mobility spectrometer with the same ionization technique would yield an informing power (chemical specificity) increase of approximately 10,000. The higher chemical specificity of MS means that significantly smaller quantities of target molecules can be detected in the presence of relatively large quantities of background molecules. This enables the detection threshold to be lowered without increasing the false alarm rate.

In the case of trace detection in the airport scenario, the problem of identification is made easier in that there are only a few hundred threat compounds and/or organisms that might conveniently be used by terrorists. Thus, one does not have to determine the identity of an unknown substance from scratch but rather must determine whether it contains one of these ~200 compounds or organisms. If a mass spectrometer is used, the spectrum obtained would be compared electronically against a library of reference spectra, and a software algorithm would determine if a match occurred. These algorithms and spectral libraries already exist for electron ionization, though it is not a foregone conclusion that electron ionization will be the ionization method of choice. These libraries may be searched for over 300,000 reference spectra; typically, two measures of confidence in the spectral match are given as part of a search report. If GC or LC is also used, the chromatographic elution time can be used to further increase the confidence of the identification. If MS/MS is used, each molecule of interest would be identified by two or three masses: that of the parent (molecular weight) ion (or parent less some known ion loss), indicative of the molecular weight, and the masses of one or two ions that result from intentional fragmentation of the parent ion, for confirmation. These ions would be determined by running standards.

With highly energetic molecules such as explosives or with the complex macromolecules that are present in microorganisms, the process of preparing, vaporizing, and ionizing the sample often causes the target molecule (and background molecules as well) to break up into fragments. This tendency is minimized by the use of soft ionization (see above for two common methods), but the effect can also be utilized in the collision (fragmentation) process of MS/MS analysis to confirm the identity of the material. The pattern of fragments formed is characteristic of the structure of the parent molecule or organism, and under appropriate conditions this spectrum and the molecular weight can be used to identify the target molecule/organism by comparing them to the reference library. This detection strategy makes MS/MS systems much more flexibe than IMS systems. If additional threat agents become a concern, the MS/MS molecular species and reference library can be expanded to accommodate them.

In the airport context, the time required for acquisition and analysis of a sample is a critical factor in determining where and in what circumstances the trace detection technique can be used. The current system of IMS trace detectors for explosive residues on passenger carry-on bags takes on the order of 1 minute from sample acquisition to final analysis. It is probable that an MS or MS/MS-based system would take about the same amount of time or slightly longer, because for explosives, the sample collection and preparation requirements for IMS and MS are essentially the same.

Once a sample has been acquired by wiping or collecting as with IMS, the time required to acquire a mass spectrum can be made short (on the order of 1 second), and the electronic analysis of the resulting spectrum takes about the same amount of time. However, in some cases (e.g., biological analysis), sample collection and preparation, as well as preseparation of complex mixtures, might add significantly to the analysis time. Gas or liquid chromatography used as a preseparator at the MS inlet might add several seconds to several minutes to the analysis time⁹; however, by using an MS/MS configuration, the chromatographic separation on the front end may be avoided, saving significant analysis time. Preparation of complex biological samples for MS analysis (e.g., lysing cells and digesting polypeptides) may also add considerable time to the analysis of potential biological agents. There is

considerable literature on the sampling and ionization of explosive material,¹⁰ but the optimum methods for vaporizing and ionizing the specific combinations of threat materials likely to be encountered remain to be determined.

The same dry wiping techniques used with current IMS systems can also be used with MS-based detection systems. However, as discussed in Chapter 1, automated sampling (either of hand luggage or in a portal arrangement) would be better because it would avoid the many uncertainties of manual sampling. Any automated sampling system would incorporate a condensation or concentration step to counteract the air flow dilution (such a step is now included in portal sampling systems).

As a trace detection technology, MS-based systems may have the same generic limitations as all trace detection technologies, discussed in Chapter 1. The threat preparation and delivery scenario must predict a high probability that threat residue will be deposited on items to be interrogated. The sampling method must effectively access the residues if present. This trace technique is unable to distinguish innocently acquired, but real, explosive compounds such as might be found on persons who take nitroglycerin heart medication or on, say, hunters, law enforcement personnel, and mining engineers, from explosive compounds on individuals who prepared bombs. However, as discussed below, MS-based systems can address several of the specific limitations of current IMS systems.

Given the unpredictable efficiency of sample acquisition, discussed above, and the possibility that a terrorist would try to minimize the presence of threat substance residues on his or her hands and luggage, it is desirable to reduce alarm thresholds below current levels to increase the probability of detection of trace residues or vapor. Whereas IMS systems have alarm levels in excess of 1 nanogram of explosive, MS-based systems should be capable of alarm levels on the order of 100 picograms. Given the 10,000-fold greater chemical specificity of MS compared with IMS (see above), this lower alarm level should be achievable without increasing the false alarm rate due to interfering compounds in the sample background.¹¹ While the lowered detection limit may increase the number of alarms caused by detection of innocently acquired explosive residues, the higher chemical specificity of MS-based systems might enable these systems to compensate for this problem. For example, nitroglycerin pharmaceuticals will contain other characteristic compounds that can be detected in the mass spectrum and used to distinguish medication residues from explosive residues, avoiding false alarms raised by heart patients.

As discussed in Chapter 1, currently deployed IMS trace detectors are designed to detect specific explosives only. The flexibility inherent in MS-based systems should make them capable of detecting a much broader range of threat substances, including other improvised explosives, chemical warfare agents, and biological agents.

The efficacy of mass spectrometry in detecting explosives is well established. Jehuda Yinon, an internationally known forensic scientist and author of many papers and several books¹² on detection of terrorist materials, concluded as follows:

McDonald et al.¹⁴ have reported a gas chromatography/mass spectrometry (GC/MS)-based laboratory method for confirming nine nitrogen-containing explosives (EGDN, DMDB, NG, PETN, RDX, HMX, NT, DNT, and TNT) by their molecular weight at subnanomole amounts using methylene chloride chemical ionization and detection of negative ions. (This is nearly the same ion source chemistry as is used in IMS.) The time for analysis was less than 10 minutes. Another approach that substitutes glow discharge ionization¹⁵ and another stage of MS for GC (in order to reduce the analysis time to seconds) is being tested by the TSA in a portal form.¹⁶

Picogram quantities of explosive residues on packages and luggage, as well as on passenger boarding passes, have been detected using the portable tandem MS detector described in Box 2-1 and shown in Figure 2-1.

Similar approaches taking advantage of the versatility of mass spectrometry have been used to detect chemical warfare agents, and commercial spectrometers are offered for both military and industrial use.^17,18,19,20 Presentations have been given documenting the detection in drinking water of 42 of the 48

FIGURE 2-1 The as-deployed Scentinel mass spectrometer trace explosives analyzer. SOURCE: Mass Spec Analytical, Ltd.

compounds listed on Chemical Weapons Convention schedules.²¹ These studies used a low-pressure photoionization/ion trap/TOF mass spectrometer.²² IMS spectrometers have not been considered for these diverse analysis needs.

Biological agents pose a greater challenge to MS-based detection systems because the target bioagents must be detected against a complex background of naturally occurring microorganisms and other aerosols found in the environment. Sample preparation is likely to be more complex with biological samples, and it is not yet clear which types of instruments will be best suited to the analysis. If the analysis focuses on whole proteins, TOF spectrometers with large mass ranges may be required; if detection of characteristic polypeptides or unique amino acid sequences is the goal, MS/MS and small spectrometers might be used. Single-particle aerosol mass spectrometers are also being developed that can analyze the composition of individual aerosol particles in real time.^23,24 Only further research on these issues will determine which, if any, MS technology will be appropriate. The encouraging thing is that

mass spectrometry has already shown significant capabilities for biological identification.^25,26

A paper by Randolph Long²⁷ of Edgewood Arsenal—an Army laboratory that has been conducting and funding research on chem/bio detection for more than 30 years—compares the three current contenders for the detection of chemical and biological agents and bioactive peptides and toxins: immunoassays; assays based on nucleic acid sequence; and mass spectrometry. In contrast to immunoassays, which are specific to a single agent and take 15-20 minutes, and assays based on nucleic acid sequence, which use polymerase chain reaction (PCR) amplification and detection with DNA arrays, mass spectrometry was characterized as follows:

Biological agent analysis will require considerable chemical specificity, and MS/MS configurations will be incorporated. MS/MS systems with a single-analyzer, ion trap mass spectrometer have been implemented by Oak Ridge National Laboratory in the chemical and biological mass spectrometer, Block II (CBMS).²⁸ The CBMS was developed with funds from the U.S. Army and the Department of Energy for use in chemical and biological warfare. It consists of an ion-trap MS/MS mass spectrometer with chemical ionization source and three specific inlets to handle various sample types to be analyzed.²⁹ Biodetection for bacterial, viral, and toxin targets was accomplished by concentrating particles in a respirable range (2-10 μm) into a quartz pyrolysis/thermolysis tube, where, after the addition of a derivatization reagent, the sample is thermolyzed. The mass spectra of the liberated chemical biomarkers are then processed to determine if targeted biological agents are present in the sample. Viruses and toxins are typically distinguished on the basis of protein fragments, while bacteria are distinguished on the basis of their fatty acid methyl ester profiles.³⁰

Another extensive program developing mass spectrometry for chemical and especially biological detection is being conducted at the Johns Hopkins University Applied Physics Laboratory (APL). With funding from DARPA as part of a systems approach to the detection of threats in the environment, APL has built and is testing an integrated chemical and biological detection system based on a miniaturized

FIGURE 2-2 Miniaturized mass spectrometer for bio-chem defense. The Johns Hopkins University Applied Physics Laboratory (APL) has developed a miniaturized time-of-flight mass spectrometer built into a suitcase-sized container for analyzing solids, liquids, and gases in the field. SOURCE: JHU/APL.

TOF mass spectrometer utilizing electron impact ionization for vapors and laser desorption-ionization for higher molecular weight threats (bioregulators, toxins, and microbes; see Figure 2-2).^31,32

Recent reports from Purdue University suggest that the identification of whole proteins from complex mixtures can be accomplished with lysated cells (E. coli) and mass spectrometry alone, using no chromatography or other preparation step.³³ For targeted biomolecules, this raises the possibility of significant reductions in sample preparation time.

MS-based systems face a number of challenges before they can be deployed in airports as trace detectors, discussed further below.

As noted above, the U.S. Army and DARPA have conducted proof-of-principle research and development, testing, and evaluation for both chemical and biological threat analysis using fieldable, rugged, specialized mass spectrometer systems. It is likely that much of the work these and other agencies have done to develop equipment concepts could be used directly or modified for TSA threat scenarios, but TSA needs to focus on its unique needs for a rugged, backbone mass spectrometer that would be useful for many threat detection scenarios.

Although extensively used for a variety of laboratory applications, commercially available chemical analysis systems (C/MS/MS) are not designed for an environment as harsh as an airport or other transportation arenas, nor are they designed for use by TSA security operators.³⁴ Given the range of airport deployment sites (e.g., baggage rooms, curbside check-in kiosks, passenger checkpoints), an ETD must be able to operate effectively under a variety of adverse conditions, including extremes of temperature, changes in barometric pressure, high humidity, and high levels of dust or other airborne particles. IMS systems have been known to fail under these adverse conditions, and substantial investment may be required to adapt MS-based systems for reliable use in these environments. MS-based systems can be robust, however, as demonstrated by contract laboratories in which millions of environmental analyses are run around the clock, with several mass spectrometers overseen in their automatic operations by technicians.

The costs of laboratory instruments depend on their complexity and the volume produced. GC/MS instruments that sell in the thousands of units per year are offered at about $50,000. LC-ion trap machines and TOF machines cost from $100,000 to $250,000 and are sold in quantities of hundreds per year. The annual maintenance contracts typically cost about 10 percent of the purchase price per year in a laboratory setting.

Once an analyzer configuration is selected, the operational, cost/benefit, and functional requirements can be specified, test protocols issued, and bids solicited from companies with MS product history to supply a rugged and cost-effective product or design. Cost goals must be given, but there are limits of design and pricing that depend on volume. To keep costs as low as possible, it would be desirable to involve multiple vendors in the design and production of an MS-based ETD system.

The configuration of the sample inlet (chromatograph, if needed) and the ionization, mass analysis, and ion detection methods will depend on the problem to be solved. It is not reasonable to expect that one inlet and ionization method will serve all threat materials or all threat scenarios. It is moreover possible to have more than one inlet or ion source in or attached to the MS. Ion trap analyzers, in particular, can perform several different types of mass spectrometer scans (MS/MS/MS...) and change the polarity of detected ions, all based on information from the previous scan. As can be seen from the examples so far, several different configurations have been chosen for solutions to chem/bio problems: systems utilizing gas or liquid chromatography, chemical vs. electrospray ionization, ion trap or TOF, MS/MS or MS alone, and positive/negative ion detection, among others. Choosing the most appropriate configurations will take time, expertise, in-depth knowledge of the manufacture of mass spectrometers, and perhaps additional research. A significant attempt should be made to select configurations that are extendable to as many threat scenarios and substances as possible.

Several groups and companies are working on related problems, and these efforts must be considered. For example, R.G. Cooks at Purdue University is leading an initiative on multiplexing and miniaturizing ion trap mass spectrometers.³⁵ These devices (Figure 2-3)³⁶ are ion traps and hence have MS/MS capability. They are miniaturized and have a single vacuum system; they are, however, the only devices having separate mass spectrometers and separate ion inlets, allowing different ionization and sample handling techniques.

FIGURE 2-3 Miniaturized mass spectrometer with inlet configuration enabling a single sample to be analyzed by several different MS techniques. SOURCE: Courtesy of R.G. Cooks, Department of Chemistry, Purdue University.

The Achilles heel in trace detection is improper manual sampling. To fit in better with current in-line baggage handling systems, ETDs with automated sampling systems will be needed. Two approaches are being developed. One uses a laser pulse to desorb particles from the bag; these are subsequently collected. The other uses pulsed air jets to dislodge particles. These approaches have the potential to automatically and systematically sample a large percentage of carry-on and checked baggage without relying on an operator. If successful for carry-on materials, a major flaw and bottleneck in passenger screening could be eliminated. The actual capabilities and understanding of these approaches need further funding and optimization, with deployment once they have been shown to be justified.

Passenger screening has been the primary reason for the development of trace detection. However, in currently deployed systems, neither the passenger’s body nor his or her clothing is sampled for residues of threat materials—only personal items and carry-on bags that are likely to have been touched by the passenger are sampled. Other than metal detectors, there is no currently deployed technology for screening passengers per se. Clearly, one scenario for bringing down an airplane is a suicide attack in which the terrorist straps an explosive or other threat material to his or her body. Such a device should be detectable by a portal sampler.

Passenger portals based on IMS technology for sampling and detection have been tested and are commercially available. As noted earlier, TSA has evaluated an MS-based portal system. Passenger portal systems will, however, require lower alarm thresholds in the future to compensate for the inefficiency of sample collection.

The TSA has an opportunity to directly compare MS technology with existing IMS technology in passenger screening portals. Deployment of both IMS and MS types with lower alarm levels would be a good test of mass spectrometry’s ability to detect at lower levels with fewer false positives than current technology. The probability of detection should be compared with the probability of false alarms at lower alarm thresholds for each technology in an airport scenario, and each technology should be characterized for operational and security adequacy. Concurrent studies could determine (1) the expanded list of threat materials that MS could address and (2) the extent of alarm level reduction.

In the committee’s view, MS-based portals represent the lowest-cost pathway to demonstrating the utility of MS-based detection. Portals with IMS-based detectors and portals with MS-based detectors should be compared side-by-side to determine the relative probability of detection and the probability of false alarms as alarm levels are lowered.

To identify a target molecule using a mass spectrometer, the spectrum obtained would be compared against a library of reference spectra, and a software algorithm would determine if there is a match. Such algorithms, formats, and spectra libraries already exist and will form the basis for those used in this application. Once the methods of analysis are chosen, corresponding libraries will need to be agumented. All commercial MS data systems allow libraries to be created based on standard samples of interest. Since the chemical specificity of the analysis technique allows for eliminating most if not all background signals, it is not necessary to run standards in the presence of all known backgrounds, as is always necessary with IMS.

There are several commercial algorithms available for searching spectra libraries of both chemical and biological compounds. These algorithms need to be customized and tested with the libraries using spectra obtained in the actual airport environments to better understand their robustness and sensitivity. As has been seen in other technology implementations for civil aviation security, such real-world tests are critical to the assessment and success of the technology.

Based on the discussion above, the committee offers two sets of findings and recommendations:

Finding 1: The trace detection systems currently deployed in airports have limited utility for the following reasons:

• The relatively low chemical specificity of IMS means that the instrument alarm threshold must be set high to avoid excessive false alarms; yet, lower alarm levels are desirable to account for inefficient manual and portal sampling techniques and, possibly, “cleaner” perpetrators.

• Detection is dependent on the use of blind sampling methods that cover only a small portion of the bag surface for acquisition of adequate residues for analysis.

• Current sampling protocols do not allow for the sampling of explosive residues or vapors that may be associated with a passenger’s skin or clothing.

• Currently deployed IMS systems are designed to detect only a specific list of explosives and cannot easily be reconfigured to detect an expanded list of explosive, chemical, and biological threat substances.

Recommendation 1: To address these deficiencies in the performance of explosive trace detectors, TSA should do the following:

• Place a high priority on the development and deployment of automated trace sampling hardware.

• Decrease the threat alarm threshold for ETDs systematically over time to improve the probability of detection of residues while keeping false alarms at current levels.

• Deploy passenger screening portals to enable the detection of explosive traces on passengers’ skin and clothing, and assess the acceptability and efficacy of the portals.

• Explore new technologies with higher chemical specificity that are capable of detecting a wider range of explosive, chemical, and biological threat materials.

Finding 2: Owing to their lower limit of detection, higher chemical specificity, and chemical flexibility, MS-based trace detection systems have the capability to address many of the limitations of IMS-based systems. Several development efforts are under way at universities and in the private sector to commercialize miniaturized, low-cost MS systems that can detect a range of threat materials.

Recommendation 2: TSA should establish mass spectrometry as a core technology for identifying an expanded list of explosives, as well as chemical and biological agents. Specifically, TSA should

• Create a prioritized list of threat materials that are likely to fit a residue scenario and a second list of materials that are not likely to fit the scenario.

• Determine appropriate MS sampling procedures, inlet configurations, ionization methods, and analysis strategies for relevant materials on this list.

TSA should execute an orderly program for development, testing, and deployment of mass spectrometry-based systems for residue threat scenarios that involve the expanded threat list. Rugged, miniaturized MS-based trace detection systems should be developed for use in an airport environment in order to achieve lower alarm thresholds without increasing false alarm rates and to provide versatility for threat substances not now detected.

A good way to bootstrap this entire process would be to purchase the best field-deployed instrument to gain experience and to test system applications. One option would be for TSA to purchase an instrument with the capabilities of the Scentinel instrument described in Box 2-1 and evaluate its effectiveness in an airport setting.

In Chapter 3, the committee suggests one possible example of an orderly program for development, testing, and deployment of mass-spectrometry-based systems.