J Probabilistic Aspects of Taggant Recovery

An important problem in considering the forensic utility of taggants is contamination of the identification taggant in a sample that is being examined by a forensic scientist. Contamination can come from a variety of sources during the manufacture, transportation, and handling of an explosive before a blast and could be compounded further by several possible sources of environmental contamination at the site of the blast.

The likelihood of finding taggants that would unequivocally identify the source of the explosive would depend on the relative quantity of "correct" taggants compared to the quantity of contaminating taggants from different batches introduced from each source. If the exact relative amounts of taggant from each of these sources were known, the forensic reliability of a given taggant sample analysis could be stated with scientific precision by application of the laws of probability.

This appendix demonstrates how such a calculation would be made if all of the information on relative concentrations—which, in principle, might be determinable—were available as input data. The following exercise not only demonstrates how the data might be used to make a reliable estimate of probability, but also suggests an approach to obtaining reliable data for each source of contamination and therefore making an unequivocal identification. The discussion focuses on hypothetical tagging of ammonium nitrate (AN), for which contamination could arise in a variety of ways. Nevertheless, the general arguments presented here define the probabilistic logic that could be extended to other examples of tagging explosives.

Several possible points of contamination exist along the path from the manufacture



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 347
Containing the Threat from Illegal Bombings: An Integrated National Strategy for Marking, Tagging, Rendering Inert, and Licensing Explosives and Their Precursors J Probabilistic Aspects of Taggant Recovery An important problem in considering the forensic utility of taggants is contamination of the identification taggant in a sample that is being examined by a forensic scientist. Contamination can come from a variety of sources during the manufacture, transportation, and handling of an explosive before a blast and could be compounded further by several possible sources of environmental contamination at the site of the blast. The likelihood of finding taggants that would unequivocally identify the source of the explosive would depend on the relative quantity of "correct" taggants compared to the quantity of contaminating taggants from different batches introduced from each source. If the exact relative amounts of taggant from each of these sources were known, the forensic reliability of a given taggant sample analysis could be stated with scientific precision by application of the laws of probability. This appendix demonstrates how such a calculation would be made if all of the information on relative concentrations—which, in principle, might be determinable—were available as input data. The following exercise not only demonstrates how the data might be used to make a reliable estimate of probability, but also suggests an approach to obtaining reliable data for each source of contamination and therefore making an unequivocal identification. The discussion focuses on hypothetical tagging of ammonium nitrate (AN), for which contamination could arise in a variety of ways. Nevertheless, the general arguments presented here define the probabilistic logic that could be extended to other examples of tagging explosives. Several possible points of contamination exist along the path from the manufacture

OCR for page 347
Containing the Threat from Illegal Bombings: An Integrated National Strategy for Marking, Tagging, Rendering Inert, and Licensing Explosives and Their Precursors of a tagged explosive to the ultimate recovery of taggants by investigators at the site of a criminal bombing. The following analysis attempts qualitatively to cover the types and ranges of contamination that might be encountered and to suggest implications for law enforcement. It must be kept in mind that different types of explosives—e.g., cap-sensitive packaged explosives such as dynamite, as opposed to bulk fertilizer-grade AN—may be subject to distinct types or sources of contamination. It is presumed here that taggants have been supplied to the explosives manufacturers in uncontaminated form, with accurate records of codes employed and of schedules of taggant incorporation. SOURCES OF CONTAMINATION During the manufacturing process, improperly cleaned equipment may contain the residue of a previous batch of explosive that had its own, differently coded taggant. An explosive produced with contaminated equipment would thus contain some fraction, x, of differently coded taggant particles, while the remaining fraction 1 – x would be correct. The fraction x would depend strongly on the details of the particular manufacturing equipment and processes (and therefore on the specific product involved), on the frequency with which the taggant code was changed, and on the number of batches produced previously under the same code. As a result of these variations, x could reasonably be estimated to range anywhere from virtually zero to approximately 0.02. Stringent cleaning protocols for the processing equipment could hold this number close to an absolute minimum, but at some cost burden to the manufacturer and legitimate consumer. In packaged explosives, taggant purity is unlikely to be compromised to any significant extent during transportation and distribution. By contrast, these activities might be substantial sources of contamination for unpackaged bulk products. For example, creating and moving large heaps of AN as well as loading and unloading it from trucks and river barges would surely reduce taggant purity. Also, the comingling of products from different manufacturers, an occasional practice in the handling of bulk AN, might also reduce taggant specificity. If y denotes the fraction of contamination introduced during the transportation and distribution of an explosive, the reduction of the purity of the taggant as it existed at the factory output stage can be signified by 1 – y. Clearly, this factor will depend on the type of explosive involved and on the presence or absence of anticontamination measures taken during transportation and distribution, especially for bulk products such as AN. The factor 1 – y could be virtually 1 in favorable circumstances that might realistically apply to packaged explosives, but the committee estimates that a combination of contaminating effects especially relevant to bulk AN might at least temporarily reduce it to 0.8 (y = 0.2). At a bomb site, environmental contamination potentially introduces yet another source of error in identifying the "correct" taggant. For example, taggant particles driven outward from an explosion could become embedded in walls of a

OCR for page 347
Containing the Threat from Illegal Bombings: An Integrated National Strategy for Marking, Tagging, Rendering Inert, and Licensing Explosives and Their Precursors nearby building that themselves are contaminated with similar but differently coded taggants in the explosives used for mining raw materials for the construction of buildings. Alternatively, a sufficiently powerful blast could pulverize building materials that then would fall on the blast site and inadvertently be recovered at the scene by forensic investigators. Another potential source of environmental contamination could be fertilizer (perhaps with AN as an ingredient) that itself was tagged and used repeatedly on lawns or gardens surrounding a bomb site. Although it would not be relevant in an urban setting, such a source of contamination could eventually become a significant issue where water runoff and/or animal ingestion and excretion tended to concentrate taggants locally. Environmental contamination might well be relatively unimportant in the early stages of a tagging program, but with the passage of time, slow environmental buildup could become an increasingly significant issue. To characterize these effects, let z represent the environmental contamination fraction, so that 1 -z is the taggant purity diminution factor. While substantial variation in the contamination parameter has to be recognized, in especially unfavorable circumstances z is estimated to be as high as 0.2. DISCUSSION The overall probability, p, that any given recovered taggant particle is "correct" is given by the product of factors: which would equal approximately 0.63 for the estimated, "worst case" upper limits mentioned above. Suppose that the recovery effort at a bomb scene turns up some small number of taggant particles, say 10. The chance that exactly j of these 10 possess the "correct" code for the explosives batch used is given by the binomial expansion expression: The specific outcome with respect to the numbers of correct and incorrect taggants found in any one bombing incident is a "roll of the dice." By way of illustration, suppose that p were equal to 0.65, indicative of heavy contamination, as conceivably could apply in the case of AN. The following outcome probabilities can be calculated from the expression shown above. j P(j,10) 0 0.0000 1 0.0005 2 0.0043

OCR for page 347
Containing the Threat from Illegal Bombings: An Integrated National Strategy for Marking, Tagging, Rendering Inert, and Licensing Explosives and Their Precursors 3 0.0212 4 0.0689 5 0.1536 6 0.2377 7 0.2522 8 0.1757 9 0.0725 10 0.0135 Although 7 correct taggants, with 3 incorrect, is the most probable outcome, it will occur only about one-quarter of the time, and other combinations also have nonnegligible probabilities. In more favorable circumstances, as, for example, with a packaged explosive subject only to light environmental contamination, and with p = 0.98, the results might be as follows. j P(j,10) 0 0.0000 1 0.0000 2 0.0000 3 0.0000 4 0.0000 5 0.0000 6 0.0000 7 0.0008 8 0.0153 9 0.1668 10 0.8171 Now the most probable outcome is 10 correct taggants. COMMENTS In the first example above, when p is equal to 0.65, the chance for the majority of the 10 recovered particles all to have an incorrect code cannot be ignored. The net occurrence probability for 7 or more of the 10 to possess an erroneous code is 0.0000+0.0005+0.0043+0.0212=0.0260 or 2.6 percent. This seems to be of sufficiently high magnitude to allow a legal challenge in court to the veracity of taggant evidence on the basis of the "beyond a reasonable doubt" criterion; still, such evidence might have forensic value in

OCR for page 347
Containing the Threat from Illegal Bombings: An Integrated National Strategy for Marking, Tagging, Rendering Inert, and Licensing Explosives and Their Precursors helping to identify a suspect. Such uncertainty is aggravated by the fact that p (let alone separate x, y, and z values) in any real-world case would likely be unknown. The estimates above assume that all of the incorrectly coded particles bear the same signature; if they were to differ, the opportunity for misinterpretation would be diminished, and if necessary, more elaborate combinatorial formulas could be invoked to provide quantitative estimates. Likewise, recovery of substantially more than 10 taggant particles would obviously increase the "signal-to-noise" ratio and would improve the ability to use this evidence in criminal proceedings. The contamination attributable to x (manufacture) and y (transportation and distribution), and to a lesser extent z (environment), would depend on the frequency with which the tagging code was changed. If the intervals were long (e.g., 6 months, as in the Swiss protocol), contamination would be insignificant except during manufacturing lot changes. Of course, such infrequent changing of the code implies reduced information content for the taggants. As a rough average, 1 pound of an explosive is required to dislodge 1,000 pounds of rock, mineral, or ore in mining operations. However, circumstances (surface mining versus tunnel mining, and the materials involved) can cause this ratio to vary somewhat. Each pound of explosive might be tagged at a level of approximately 10,000 particles per pound. Perhaps 1 in 1,000 of these can be expected to survive the mining blast in readable form, to be transported away from the mine. Depending on the subsequent processing steps used to produce usable construction material (such as wallboard, concrete, etc.), the concentration of taggants might be decreased or increased. In assessing the resulting environmental contamination that such building materials might produce when directly involved at a criminal bombing event, it is estimated that about 1 taggant particle per 100 pounds of material could be anticipated. Clearly, though, a substantial uncertainty applies to this estimate. The least equivocal message that emerges from these considerations is that more real-world data need to be accumulated to tighten each of the contamination estimates and their quantitative implications.