Chapter 1: The Application of Mass Spectrometry to Explosive Casework: Opportunities and Challenges

Intelligence, in a law enforcement context, is the collection and assessment of information for further use as a decision-making tool at, predominantly, three levels: strategic, operational and tactical. Each type of intelligence requires different information to be collected and can range from a broad to a targeted number of sources. The intelligence is then used to direct resources as part of a wider organisational policy or for a specific operation/investigation, including applying for warrants or engaging specific capabilities. It is within this context that forensic intelligence finds its place and significance.

Ribaux et al. define forensic intelligence as, “the accurate, timely and useful product of logical processing (analysis) of forensic case data (information) for investigation and/or intelligence purposes”.¹  Timeliness and accuracy of any type of intelligence are important with accurate intelligence provided too late being pointless, and inaccurate intelligence can be potentially disastrous. Disciplines such as DNA and fingerprints are easily recognised for providing forensic intelligence by identifying potential suspects or linking individual crimes based on profiles obtained from scenes. However, forensic intelligence is not only limited to these specific fields. Technical and physical attributes of devices have been exploited for forensic intelligence for many years in explosive device investigations. A recent highly publicised example is Malaysian Airlines flight, MH17, which was shot down over eastern Ukraine in 2014 with distinctive bowtie-shaped fragments recovered from the cockpit and captain’s body, providing the first forensic intelligence to support the involvement of military hardware in the incident.²  An additional critical piece of technical information to generate forensic intelligence from explosive devices is the identification and analysis of the devices’ explosive materials. This is typically produced based on the information gathered in two distinct phases: (1) rapid analyses in the field; and (2) in-depth analysis in the laboratory. Here the timeliness of the information plays a crucial role in determining whether what is being provided is intelligence or not. Any indication, or identification, of the explosive associated with the device that can be provided rapidly in the field is intelligence as it is timely and useful. At the same time, there is some latitude in how accurate this intelligence needs to be. While the presence or absence of an explosive must be as accurate as possible, the type and exact nature of it demand less accuracy. Once the forensic investigation moves into the laboratory, accuracy is paramount. Timeliness continues to be important but what would be considered useful intelligence changes. Efforts to confirm or refute field intelligence are critical at the commencement of laboratory examinations and need to be expedited, but, beyond that, given the time and the techniques available, for anything to be considered useful intelligence, it must be far more insightful (see Figure 1.1).

The development of portable detection technology for the real time analysis of samples at the scene has led to this type of intelligence becoming a crucial component in all bombing related incidents. The Bali bombings of 12 October 2002 provided a terrific example of the benefits of deploying portable detectors, as a “mobile laboratory”.³  Deploying portable equipment can assist in avoiding delays that may arise from the movement and transport of evidence from the scene to the laboratory, particularly from remote or distant locations, and it also provides an opportunity to triage evidence and prioritise, or exclude, certain items for further laboratory examination, saving the investigation valuable resources. Critically, although, rapid analysis in the field allows a forensic chemist to provide intelligence to the investigative leads at the scene. In Bali, the use of ion mobility spectrometry (IMS) in the mobile laboratory enabled detection of TNT in samples taken from two of the three scenes. This was later confirmed in the laboratory and corroborated by confessions. The success of the Bali bombing investigations, as well as numerous incidents since, was possible, despite portable mass spectrometers being unavailable. Recent advances in this area provide additional opportunities covered later in the chapter.

Sophisticated laboratory equipment, separately and holistically, plays crucial roles in providing information for forensic intelligence purposes. Laboratory equipment is generally considered to have greater sensitivity (i.e., able to detect smaller amounts) and more selectivity (i.e., able to detect more chemicals with greater confidence) than portable systems, offering the potential to produce confirmatory rather than presumptive results. These attributes are utilised to interrogate different characteristics of each explosive sample at a very low level to determine similarities and differences between other samples and/or to identify specific features that may indicate the sample’s background (i.e., production method).

Explosive sample attributes can be collated to create databases and while some databases exist, they tend to either be held within individual organisations or are more generic and publicly available over the internet. Databases held within specific law enforcement agencies are likely to be of high quality and enriched with, or linked to, broader case details including scenes and persons of interest; however, they can only be populated with samples analysed by that laboratory or from data shared directly with them. Thus, they are likely to be relatively small (depending on the jurisdiction of the agency and the number of seizures that are submitted). The need to maintain the operational capability and information security of the agencies necessitates heavily restricted access of such databases. Conversely, easily accessible databases contain characterised products procured from within the available market of the creator/s, an example being the online Smokeless Powder Database, which includes records of the manufacturer information, physical characteristics, explosives and additives present (as detected by GC–MS) of approximately 900 types of smokeless powder.⁴  Although these databases are likely to contain more entries, they are unlikely to be as well characterised as secure databases. As with all commercial product databases, ensuring that the database is up to date with the market, especially in situations where there are changes in formulations, is very difficult. As somewhat of a compromise, a small number of databases accessible via the internet, albeit with restricted access, are hosted by various agencies/organisations and available for interrogation once an application for access has been approved.⁵  Again, however, they are not exhaustive and tend to be limited to case samples analysed by the specific laboratory/group or those sampled from the marketplace. In each case, unless the specific case sample is being compared to a laboratory/user’s own database with additional case context, an external database can only provide limited information.

Field explosive detectors can be an important tool to provide forensic intelligence with critical evidence gathered from post-blast scenes, providing crucial leads as discussed previously, or by identifying linkages to other similar incidents or an organisation potentially responsible. Just as importantly, some very volatile explosives, such as triacetone triperoxide (TATP), and their decomposition products can be quickly lost. Early response to a post-blast scene with reliable portable trace detection capabilities increases the likelihood of identifying an explosive before it is lost, capturing key evidence. With time, the residual amount diminishes, posing challenges to even high sensitivity techniques within the laboratory.

A number of portable explosive detectors can be deployed to a scene to deliver this capability.⁶  Presently, the most commonly deployed explosive trace detectors (ETDs) are based on thermal desorption ion mobility spectrometry (TD-IMS). This section will focus specifically on portable mass spectrometers and compare their performance against the commonly used ETDs. Current ETDs are manufactured and sold as specific products using specific configurations, although some settings are user-editable, allowing for useful comparisons between different systems. Portable mass spectrometers are relatively new and strive to achieve lower false alarm rates with higher specificity, compared to the traditional ETDs. Other operational considerations also include (i) ease of use, (ii) device response time, and (iii) the sensitivity of the device and the degree of sample preparation required.

Broadly, there are two classes of portable mass spectrometers (Table 1.1). One type utilises gas chromatography in-tandem with a quadrupole mass analyser or toroidal ion trap, while the other works based solely on ion trap technology. The former has the advantage of better resolution of different chemical species and easy reference to the NIST MS commercial database, but at the expense of weight and size. In contrast, the latter works in a higher-pressure range (i.e., with a less stringent requirement for a pump to achieve a lower vacuum level), allowing them to be manufactured in a way that makes them more mobile.⁷ 

There are also transportable or “vehicle-portable” mass spectrometers, which (along with their accessories) are too bulky to be carried by operators but can be, relatively easily, retrofitted in vehicles or mobile laboratories. One such example is the Waters Radian™ ASAP. This mass spectrometer has been developed for the detection of drugs, but its library can be customised to include some explosives. Nonetheless, it is not an explosive trace detector and is only capable of detecting explosives in solution at tens of parts-per-million (ppm) concentration.

When considering the design of systems to utilise for the purposes of providing forensic intelligence at the earliest stages of an investigation, it is important to understand the performance of each system being considered.

For example, the comparison of two ionisation techniques, namely, secondary atmospheric pressure chemical ionisation (SAPCI) and secondary electrospray ionisation (SESI), and their performance in terms of sensitivity, selectivity and repeatability. Burns et al. conducted a study using a modified Waters Acquity QDa™ single quadruple mass spectrometer, to detect five explosives, namely, TNT, RDX, PETN, Tetryl and HMTD, on swabs using each of the ionisation techniques. Through electrospraying solvents during SESI, gaseous solvent ions generated interacted with the analyte droplets to ionise and subsequently measure the molecules of interest. In contrast, SAPCI allowed the ionisation of analytes, in either solid, liquid or solution form, by rapidly heating the sample using hot nitrogen gas to 400–500 °C and subsequently subjecting the evolved molecules to high voltage (3–6 kV), negating the need for the presence of a solvent.

Both techniques were able to the detect the five explosives, with SAPCI showing higher sensitivity and better specificity, especially in the absence of dopants. SAPCI is highly suitable for molecules with moderate polarity and a mass below 1000 amu, a level most explosives fall below. SESI, on the other hand, was more repeatable and its calibration curve is more linear over a wider concentration range.⁸  It is performance factors such as these which must be considered before selecting a system for use in a mobile laboratory. The following question should be considered: What is the overall purpose of the forward deployment of capabilities? Accuracy in quantitative properties, or the ability to detect as little as possible with as much confidence as possible? From a forensic intelligence perspective, detection capability would always be the preference. The ability to achieve this at a high confidence is of even greater value, in line with the definition provided by Ribaux et al.¹ 

Weighing at approximately 27 kg, the 1st Detect Tracer 1000™ is another benchtop linear ion trap mass spectrometer, offering sensitivity down to picogram levels. The equipment has been developed to detect RDX/HMX, TNT, nitro-esters (PETN, ETN, NG, and EGDN), nitrates (e.g., ammonium and urea), TATP, HMTD, tetryl, picric acid, R-salt and smokeless powders. Although the detector generally operates from an A.C. supply, it can be customised to be integrated with an uninterrupted power supply (UPS), so that the set can be made transportable and deployed in fields, on trolleys or unmanned ground platforms. Table 1.2 compares the 1st Detect Tracer 1000™ with the IonScan™ 500DT, IonScan™ 600 and NucTech™ TR2000DC, as analogous benchtop ETDs utilising IMS technology (Table 1.2). It is noteworthy that the new generation of benchtop mass spectrometers is aiming for a lower false alarm rate and the detection of a larger number of explosive substances.

It should be noted here that all the available portable products focus predominately on the detection of organic explosives. Although some detection capabilities of nitrate-based explosives do exist, these are limited to ammonium nitrate and urea nitrate chemicals. Generally speaking, inorganic explosives (see inorganic explosives further in this chapter) are not reliably detected with either IMS or portable MS technologies. Some systems, including GreyScan’s ETD-100™, are entering the market to address this; however, these do not rely on mass spectrometry and hence will not be included in the chapter.

To understand the applications of mass spectrometry in the provision of forensic intelligence from explosives, both organic and inorganic, an understanding of their chemistry is required. The following sections provide an outline as well as the techniques currently being used to interrogate samples for intelligence purposes.

Organic explosives are mono-molecular explosives containing a hydrocarbon backbone upon which various functional groups can be introduced to provide a structure that emits significant energy when its chemical bonds are broken. These explosives can be categorized into family types (for the most encountered species): nitro-alkanes (R-NO₂), nitro-aromatics (Ar-NO₂), nitramines (R-N-NO₂), nitrate esters (R-O-NO₂), and peroxides (R-O-O-R) (Scheme 1.1). While these are the most common families of organic explosives, others do exist and occasionally are observed in forensic casework (e.g., R-salt; an N-nitrosamine, and R–N–NO). Due to their structural similarities, species within a family often have similar properties in terms of sensitivity, thermal lability, and detection ability based on the instrument type.

In addition to mono-molecular explosives, other organic chemicals can be utilised as fuels for blended explosives. Perhaps the most common examples of these would include fuel oil and sugar/sugar alcohols, such as erythritol. Some commercial products may also contain organic additives to provide stability or specific physical properties to an explosive product. Examples of these additives include acid scavengers utilised in smokeless powders (e.g., ethyl centralite and diphenylamine) and plasticizers utilised in plastic explosives and smokeless powders (e.g., bis(2-ethylhexyl) sebacate, and dibutyl phthalate).

The analysis of organic explosives, fuels, and additives can often be accomplished with spectroscopic techniques, such as Fourier transform infrared (FTIR) spectroscopy or Raman spectroscopy, when bulk material is available for interrogation. The utility of mass spectrometric techniques becomes more powerful when conducting trace residue analyses of these materials. Common exceptions to this could include a finer, detailed analysis for discriminating between different commercial explosive products. This is often done when inter-comparing between smokeless powders, plastic explosives, and other products containing organic additives, taggants, and explosives; Table 1.3 reports a list of commonly encountered explosives, fuels, and additives with their accurate masses).

The analysis of organic explosives and common organic additives to explosive products is not only significant from a law enforcement perspective but also for environmental and pollution monitoring, transportation screening, and for defensive/security applications. Given the wide range of compounds that are of interest, a number of different laboratory-based techniques have been developed which are outlined below.

Due to the labile nature of many organic explosives, the utility of gas chromatography is limited, even beyond the analysis of vapour phases. However, there are specific areas of explosives analysis where GC–MS is routine and extremely beneficial. One such application is in the analysis of smokeless powders and other nitrocellulose-based propellants.^9–11  Due to the existence of international standards for the analysis of smokeless powders, this type of chemical discrimination between samples has become commonplace in many laboratories. Combined with the physical properties (e.g., grain shape, grain size distribution, and marker grains) of a smokeless powder, the chemical constituents can provide an extremely useful comparison point. The identification of nitrocellulose can also be achieved via GC–MS through trimethylsilyl (TMS) derivatives.¹²  These types of analyses can be conducted from a residual perspective and used for comparisons with samples of known origin for explosives or investigations related to organic gunshot residues. Complete identification of a smokeless powder can be limited due to similar formulations and physical attributes.

In addition to propellant analysis, extracts of plastic explosives for the discrimination of plasticizing agents have been proven to be useful. The plasticizers confer different physical properties to the final explosive products, but the identification of these plasticizers and some chemical taggants can help narrow down the field of possible products when analysing an unknown plastic explosive material.¹³  Chemical taggants are an important part of plastic explosive history. In 1991, a convention on the marking of plastic explosives was held in Montreal, Canada, during which several countries agreed to add chemical taggants to manufactured plastic explosives to enhance their detectability. The agreement became effective in 1998. Currently, approximately 155 countries worldwide have signed the agreement with the most common chemical taggants being 2,3-dimethyl-2,3-dinitrobutane (DMDNB), 4-nitrotoluene (4-NT), and ethylene glycol dinitrate (EGDN).

Sugar and sugar alcohol characterization can also be accomplished through analysis on GC–MS after derivatization. Although this process is more complex and less straightforward compared to a simple extraction, it can provide useful information to the forensic investigator. A common derivatizing agent that has been utilised for the analysis of sugar residues is again, TMS.¹⁴  In fact, TMS derivatization has also been proven useful in the discrimination between Pyrodex® and Triple Seven® residues by derivatizing fuels (e.g., dicyandiamide, benzoic acid, and 3-nitrobenzoic acid) within each respective product.¹⁵ 

Finally, the analysis of fuel oil and other ignitable liquid products by GC–MS is routine in many forensic laboratories, of which most follow international standards for the isolation, analysis, and classification of these products.¹⁶  Some of these products have been utilised as fuels for commercial and improvised explosives. The classification and inter-comparison of such fuels can be extremely beneficial when determining the uniqueness of a liquid fuel among different samples. For instance, fuels such as diesel fuel and petrol (gasoline) exhibit considerable similarity among different refineries throughout the world but may show seasonal differences as well as differences arising from the crude material originally utilised or refinery processes (e.g., some petrol formulations contain n-alkanes to a greater extent than others). However, some specialty products can be unique in their composition (to include additives for scents or distinct physical properties) and provide greater clarity on the type of product utilised.

These techniques involve introducing the sample into the GC system by volatilising the material, or fractions of it, and either introducing this gas into the system or collecting it onto an amenable solid fibre which is introduced and heated again to release the captured compounds. While one would think that explosives would readily vaporise due to their chemical instability, this is often not the case. In fact, many explosives have a low vapour pressure (10⁻⁴ torr or lower at 27 °C (80 °F)). The common explosives that have a relatively higher vapour pressure include triacetone triperoxide (TATP), diacetone diperoxide (DADP), nitromethane (NM), nitroglycerine (NG) and ethylene glycol dinitrate (EGDN).¹⁷  The low vapour pressure limits the usefulness of techniques like headspace and solid-phase micro-extraction (SPME). However, some components of commercially produced explosives can be targeted in addition to these select few volatile organic explosives. Some such targets include smokeless powder and plastic explosive components that have more volatile compounds. Common detection limits for these types of components using SPME or headspace techniques range from the low picogram (pg) range to the single nanogram (ng) range.^18,19  Several studies have explored the use of SPME or headspace for the analysis of organic peroxide explosives with detection limits in the single nanogram range.^20,21  A particular study utilised ambient laser desorption in conjunction with SPME for the analysis of some nitro-aromatic explosives. In this study, non-aromatic explosives were also examined (HMX and RDX), but were not detectable using this technique, perhaps showing a specificity of the SPME fibre for specific classes of materials.²² 

Liquid chromatography offers a wider ability to characterise organic components related to explosives since high temperatures are not involved in the separation of the chemical species. Because of this, LC–MS related techniques are perhaps the most used mass spectrometry class for the analysis of organic explosives. Depending on the sensitivity of the detector, these techniques are routinely capable of detecting explosive-related compounds at nanogram levels with some techniques capable of reaching detection limits in the femtogram (fg) region. The ability to ionise, the ionisation method, stability in solution, and selectivity of the detector are common factors that can affect the sensitivity.

Commonly, two types of ionisation methods are utilised in LC–MS analysis, electrospray ionisation (ESI) and atmospheric pressure chemical ionisation (APCI). To improve sensitivity, it often becomes advantageous to enhance ionisation through the use of adduct ions. Common ionisation enhancement solutions include acetate, formate, nitrate, or chloride salts in negative mode ionisation. While negative ion modes are most commonly used for explosives, certain targets such as picric acid, DPA, and the centralites are well suited for positive ionisation detection. In addition, because of its amine structure, HMTD is also readily observed in the positive ionisation mode with APCI.²³ 

High-resolution mass spectrometry (HRMS) is becoming more commonplace with Orbitrap and time-of-flight (TOF) instruments increasing the selectivity of the mass detector. Some applications using these techniques achieve mass accuracy within a few part-per-million (ppm), narrowing the range of possible chemical molecules through mass spectrometry alone. As this technique becomes more routinely used, one may expect that the use of direct analysis through various ionisation techniques may increase in the forensic setting. Currently, most HRMS applications still involve the use of an LC component for separation. Case study #1 (Section 1.6.2) nicely illustrates an example of the application of HRMS.

A recent study utilised HRMS for the analysis of eleven sugar and sugar alcohol residues and validated the method with post-blast samples. This particular method involves a post-column addition of an ammonium nitrate solution to generate adduct ions, delivering detection levels from a high to low ng level when within a mixture.²⁴  Sugars and sugar alcohols can be utilised as fuels for explosive mixtures as well as precursors to produce nitrated sugar alcohols such as ETN and MHN.

Other than isotopic profiles, chemical impurity profiles obtained from LC–MS as a result of the synthesis of home-made explosives (HME) can also be used in forensic casework to differentiate or associate samples. This can become a challenge because impurity levels change due to acid catalysed degradation over time, affecting the certainty of the forensic intelligence obtained regarding the production and origin of the explosive material.²⁵  With the combination of LC–MS and IRMS, predictions of possible linkage between suspected precursor materials and HME samples (ETN, for example) in forensic explosive investigations has been shown to be feasible.²⁶  However, caution must still be taken when combining these techniques as the repeatability of the data obtained using different instruments will have an impact on the robustness of the dataset for forensic intelligence purposes.

Unfortunately, due to their structure, organic explosives are mostly not suitable for MS–MS applications. There are exceptions to this statement, but when these organic molecules fracture in the mass spectrometer, the negative charge typically stays with similar species revolving around the nitro-group, thus making non-characteristic ions. Notable exceptions to this are aromatic explosives that can easily stabilize a radical or negative charge as part of their structure (e.g., TNT, dinitrotoluenes, and dinitronapththalenes). Under these circumstances, MS–MS can help easily and quickly identify the nitro-aromatic explosive present. In addition, some related compounds such as the centralites and diphenylamine may prove to be useful analytes for this type of analysis.

Direct analysis in real time (DART) and desorption ionisation (DI) have found some utility in the analysis of explosives under ambient conditions. Desorption techniques can encompass thermal and laser applications, and this type of analysis could become more routine as ionisation techniques advance with technology. Perhaps the most significant advantage to these types of techniques is the minimal sample preparation they require. The distinct disadvantage is the ability to create gas-phase molecules from less volatile components. As discussed earlier, however, specific classes of explosives and compounds of interest are suitable for analyses via these techniques such as those found in smokeless powders.^11,27  As technology continues to advance for ambient ionisation techniques, one could suspect that these ionisation techniques paired with HRMS could offer powerful screening techniques in the near future.

Isotope ratio mass spectrometry (IRMS) has garnered increasing interest in the field of explosive analysis, and forensics generally, over the past several years. Its current use in casework is still limited, but significant research has started to explore the utility of IRMS and some if its limitations. IRMS has shown some usefulness in being able to differentiate certain explosives from each other, when other conventional analyses could not. Most IRMS research has focused on the isotopes of carbon, nitrogen, and oxygen. Although oxygen is not a commonly studied isotope, it can prove to be useful in some research studies since not all analytes of interest for explosives need external oxygen for the combustion and ionisation process for IRMS. At the heart of IRMS lies the need to understand the range of isotope ratios observed within a particular compound or a particular set of material. Even further, an understanding of statistical models is needed to aid in drawing conclusion from such measurements.

Studies have been performed with some organic explosives including PETN, RDX, and Tetryl, showing the ability to possibly correlate between some samples and a greater ability to differentiate between samples whose isotope ratios are different.^28–30  In addition, other work has focused on inorganic analytes such as ammonium nitrate, black powder, and potassium perchlorate.^31,32  Again, some correlation can be found among manufacturers and possibly among areas of origin. However, the most beneficial aspect of IRMS is the ability to show differences among samples.

Some studies have suggested that isotope patterns may be able to link (or differentiate) original material and post-blast explosive residues.³²  This could also mean that for some substrates, a chemical precursor could be linked (or differentiated) to a synthesised explosive and/or its residues.²⁶  However, further work is needed in this area to draw any practical conclusions, including studies of variances/similarities in chemical precursors from different products, different combinations of fuels/oxidisers and how those ratios affect post-blast products, matrix effects, etc.

While the desire to correlate samples together is common across various chemical disciplines, caution is required when it comes to such statistical models needing a large (and continuing) dataset to determine the true “background” of isotopic patterns for a particular compound.³²  With the globalisation of goods and materials, it can be easily assumed that commercial and amateur manufacturers of explosives would likely opt for the easiest and least expensive sources of chemical precursors. Thus, a defensible statistical model (likely continually updated) is required before the sourcing of an explosive compound could truly occur.

It is for these reasons that IRMS is unlikely to find application in a traditional forensic case in a courtroom for some time in terms of “source identification”. However, it should be expected that IRMS use will continue to grow for intelligence purposes and most definitely for differentiation purposes in a forensic setting. An example of the use of IRMS for forensic intelligence is outlined in case study #3 (Section 1.6.4).

Unlike organic explosives, inorganic explosives are not based on a hydrocarbon backbone. Inorganic explosives most commonly consist of a mixture of an oxidiser and a fuel. The oxidiser or both the oxidiser and fuel can be an inorganic compound. There are also mono-molecular inorganic explosives such as ammonium nitrate and ammonium perchlorate. Oxidisers, or oxidising agents, are usually oxygen-rich ionic solids that decompose at moderate to high temperatures, liberating radical oxygen to react with the fuel to sustain the combustion reaction that propagates through the mixture. The selection of the oxidiser used greatly influences the explosive properties of the mixture. The melting point, the heat of decomposition and the oxygen content of the inorganic oxidiser will affect the sensitivity to ignition as well as the explosive power of the mixture (Table 1.4 reports examples of solid oxidisers).

Inorganic explosives can be grouped into various “families” with similar explosive performance and forensic properties depending, predominantly, on the oxidiser utilised in the mixture.

The most commonly available and used inorganic oxidisers and inorganic salts are nitrate-based. Potassium nitrate, for example, is a very commonly used oxidiser in explosive mixtures, has a low melting point and is therefore readily ignitable; however it is highly endothermic and will require a highly exothermic fuel to help achieve rapid burning rates (Table 1.5 covers examples of various nitrate salts used in commercial explosives).

One of the most reactive and common oxidisers is potassium chlorate, owing to its low melting point (356 °C)³⁸  and exothermic heat of decomposition. Potassium perchlorate is less sensitive to initiation because of its higher melting point (610 °C).³⁸  Chlorate and perchlorate salts are most commonly observed in pyrotechnic formulations and black powder substitutes.

Azides are substances that contain the N₃-group and can exist as inorganic salts. The most commonly encountered and discussed azides in the explosives industry, are lead, silver and mercury azide.

Certain azides are commonly used in primer compositions within firearm ammunition cartridges.

Homemade explosive (HME) mixtures are explosive products that can be easily produced by mixing commonly available materials and require only two basic constituents: a fuel and an oxidiser. Inorganic explosives are typically a blended mixture of an inorganic salt and a powdered metal, non-metallic elements or organic compounds. The most common oxidisers used are the inorganic salts, such as nitrates and nitrate salts, chlorate and perchlorate salts. HME mixtures may have different performance characteristics compared to their commercial or military explosives equivalents. However, the accessibility of precursor chemicals, the general simplicity of blending oxidiser and fuel selections, as well as the vast range of relative ratios at which the mixture will still produce a viable explosive, make these types of homemade mixtures attractive for illicit use despite their potential lower performance characteristics (Table 1.6 reports examples of uses of inorganic chemicals in terrorist attacks).

FTIR and Raman spectroscopic techniques are typically used to analyse inorganic explosives and precursor chemicals. However, when only trace amounts of material are present, these spectroscopic techniques may no longer be adequate for the sample.

Unlike organic explosives, the analysis of inorganic explosives and their post-blast residues can be particularly challenging. When considering the very low sensitivity of chromatography and mass spectrometry detection equipment and the prevalence of the inorganic salts and compounds in the environment, interpretation of the analytical data can be a complex endeavour. The forensic reporting scientists must interpret the data obtained from the analysis of solvent extracts and provide their expert opinion based on what is deemed significant or insignificant, considering the environmental background and its contribution to the matrix. Inorganic anions and cations commonly associated with inorganic explosives are present in the environment, either naturally abundant or through contamination from common consumer products. When considering the implications of a potential misinterpretation or false positive result, any exhibit analyses performed in a forensic laboratory should be accompanied by associated blanks of the sampling media (e.g., swabs) and controls from the field where possible (e.g., control soil samples). Having pre-determined limits of quantitation (LOQs) will assist in determining which results are significant and warrant interpretation (0.2 mg L⁻¹ to 0.5 mg L⁻¹ are some of the values observed in the literature for IC–MS).^39–41 

While other techniques, such as capillary electrophoresis (CE), can be used to analyse inorganic explosives and their residues, ion chromatography (IC) is utilised in many but not all forensic laboratories. IC is advantageous due to it being a highly sensitive and selective analytical technique for water soluble inorganic species. Sample preparation is easy involving a water extraction of a surface or grain, followed by filtration and dilution (if necessary). The IC lacks specificity, in that the detector alone is insufficient for determining a solute based on retention time alone; however when coupled with a mass spectrometer, IC–MS can conclusively identify inorganic explosive compounds. IC–MS detects ions (positive and negative) in solution and relies on electrospray ionisation, detecting a range of ionic species from low-mass ions such as chloride to higher mass ions such as perchlorate. Typically, the level of detection for both anions and cations is in the low mg L⁻¹ range (approximately 0.3 mg L⁻¹ or less).

IC–MS systems typically utilise electrospray ionisation (ESI) as the interface between the liquid-based IC system and the gas-based MS system. In ESI, the analyte, dissolved in a liquid phase, emerges from a metal capillary to which a high voltage is applied. During the ionisation process, the resulting field at the needle tip charges the surface of the emerging liquid, dispersing it into a fine spray of charged droplets. The droplets are evaporated using a nitrogen gas stream, decreasing their size and increasing their charge density. At a point where the charge density on the surface becomes too great, the droplet breaks into finer droplets. This process repeats until ions (cations and anions) are desorbed from the droplets into the ambient gas entering the mass spectrometer. ESI is a soft ionisation technique, where little fragmentation of ions occurs.

Mass spectral identification is based on comparison of the relative ratios of the ion isotopes to those of a standard reference material. Since ion ratios are sensitive to concentration in solution, a calibration curve should be generated using serial dilutions of a standard reference material as points of comparison for the data. Most IC–MS systems in use are low-resolution providing reasonably accurate masses to approximately ±0.1 amu. There are scientific publications which discuss IC interfaced with a high resolution Orbitrap MS such as the Thermo Q-exactive (±0.0002 amu). Table 1.7 summarises ions associated with various explosive mixtures and the expected masses these ions, including their isotopes.

One limitation of IC–MS is the inability to detect inorganic species that are not water soluble, such as iron oxide, lead chromate, lead tetroxide (red lead), lead dioxide and, importantly, aluminium. Another challenge utilising this technique is when the matrix is complicated by environmental contamination or by multiple species present in the sample. Inorganic salts dissociate in aqueous solutions into their respective anions and cations. Where multiple anions and cations are identified, it is not possible to confirm the originating inorganic species.

IC–MS may be able to provide forensic intelligence where impurities or unexpected ions are observed in the analytical sample. Through comparison with commercial products, the data could provide information which would suggest that the sample is homemade. Significant amount of research is focussed on optimising explosive mixtures and compounds for their intended use while ensuring the safety of personnel working with them.⁴²  Commercial manufacturers have strict quality control processes governing the materials that may be used in energetic materials. HME, on the other hand, is far less likely to have been manufactured to the same standards raising the possibility that the final product may contain residues that differ from what is expected in similar commercial products. Of these, cations tend to be most useful given they are often present as spectator ions in explosive synthesis reactions. For example, the identification of K⁺ or Na⁺ ions in a nitro-aromatic or nitrate ester may indicate the employment of an improvised nitration process using potassium or sodium nitrate (KNO₃ or NaNO₃) rather than nitric acid.

Metals, and their oxides, are incorporated into many explosive formulations as fuels with their addition also providing a number of different benefits. Elements such as strontium and barium, as ionic salts, are commonly included in pyrotechnic mixtures as colourants, while sulfur (a non-metal) has been used in black powder formulations for hundreds of years.⁴³  Out of all the metals available, one of the most important and commonly used metals is aluminium. The addition of aluminium into a formulation increases the flame and temperature released by the explosive formulation as well as the duration of the pressure wave produced. Different grades of aluminium are used in a number of different explosive applications, ranging from high quality aluminium particles used in high performing military formulations, detonators and commercial blasting slurries⁴⁴  to improvised explosives utilised by terrorists incorporating lower quality particles sourced from readily available commercial or domestic products. An example of an improvised explosive incorporating a metal fuel, and widely recognised as a favorite among terrorists, is ammonium nitrate and aluminium (AN–Al). Metals are also a key component in the broader design of improvised explosive devices or munitions as circuitry, including switches and power sources, containers and significantly as fragmentation elements to increase the lethality, or as a liner in sophisticated shape-charge devices increasing their penetration. Of course, the analysis and exploitation of these components are important aspects of a forensic investigation and provision of intelligence to investigations. Although these components will not be included in the discussion below, given the focus of the chapter on explosives, their analyses can be conducted using the same techniques and this chapter can be used to inform their exploitation.

Forensic investigations of metal constituents in explosives commonly utilise a spectroscopic technique, such as X-ray fluorescence (XRF) spectroscopy or scanning electron microscopy coupled with energy dispersive X-ray spectroscopy (SEM–EDS), for analysis. Both of these are non-destructive techniques, always a favourable attribute in forensic applications, and capable of accommodating both large (millimetre) and small (sub-millimetre) sample sizes. SEM–EDS also has the additional advantage providing high magnification morphological information. Indeed, research has shown that morphological data alone, from optical as well as electron microscopy, can provide important forensic intelligence.^45,46  While these techniques have significant advantages, XRF as a technique is unable to detect elements lighter than sodium and is less sensitive generally to light elements, while SEM–EDS is poor at detecting nitrogen and carbon (both non-metallic elements of interest to explosive analyses) and more susceptible to variable surface topographies.⁴⁷  Both techniques are also relatively insensitive in terms of detection, on the order of 0.1% w/w, which, if using elemental composition solely as an attribute, may provide limited value in forensic intelligence.

With respect to high sensitivity across a wide range of elements, inductively coupled plasma–mass spectrometry (ICP–MS) is the gold standard technique. ICP–MS differs in the way that it produces ions for analysis in the mass spectrometer. As outlined in previous sections, most MS applications used in forensic science use chemistry to separate species before applying physical principles, in the form of electric charges or electron bombardment, in the MS source to produce ions for measuring. ICP, on the other hand, produces ions via an argon plasma (i.e., ionised gas) created by loading a copper coil which is wound around a ‘torch’ containing argon with a high-frequency alternating current at relatively low power.^48–50  This current, together with a high-voltage discharge, starts an ionisation process in which a certain number of argon ions generate additional argon ions, which propagate further, resulting in the generation of a large amount of heat. The result is a stream of argon plasma at an extremely high temperature (approx. 10 000 K). Samples are introduced into the plasma stream via a nebuliser where they are aerosolised, with droplets that are fine enough being carried into the torch by the argon gas. Once it reaches the torch, the fine droplets evaporate leaving behind the dried analytes which are subsequently decomposed and dissociated into their individual atoms and finally ionised. The amount of ionisation that occurs differs for each element. One of the favourable conditions of using argon as the plasma source is its high ionisation potential. The amount of energy required to produce a singly charged argon ion (i.e., its first ionisation potential, or IP) is higher than that of almost all other elements (F, Ne, and He are the exceptions), while lower than the ionisation potential for doubly charged ions (i.e., second IP) for all but a small number of elements.⁵⁰  This property creates a stream of predominantly singly charged ions from the sample emanating from the torch for analysis using a mass spectrometer. Given the different energies required for ionisation by each element, none are ionised equally. Ions closest to argon in the periodic table have a first IP most similar to argon and therefore are ionised the least. These include fluorine, neon and helium which are not ionised at all, and nitrogen, oxygen, chlorine and bromine which experience a low degree of ionisation. While some of these, namely, nitrogen and oxygen, and to a lesser extent chlorine, are important elements, when it comes to explosives these are all easily measured using other techniques. Fortunately, most other elements are far more easily ionised with elements consisting of electron structures where only one electron is present in the outer shell exhibiting the highest proportion of ionisation. Argon’s first IP is also lower than, or relatively similar to, the second IP of almost all other elements, meaning most elements will not form doubly charged ions and the few that do will only generate a very small proportion, simplifying the ultimate interpretation of the resulting mass spectra.

Once a sample is transformed into ionised atoms, they then travel through an interface into the vacuum of the mass spectrometer and are focussed through electrically modulated lenses into the inner components of the mass spectrometer. The interface acting as the bridge between the atmospheric pressure/high temperature plasma and the low vacuum/temperature spectrometer is composed of metal cones each with small holes at their apex, allowing ions from the plasma core to pass through while deflecting extraneous plasma flow. The holes must be large enough to limit blocking while being small enough to facilitate high analysis performance by maintaining optimal vacuum conditions. Lower vacuum pressures reduce the number of gaseous molecules present, limiting the number of interactions between ions and molecules, thereby increasing the transmission efficiency of ions into the mass spectrometer, which, in turn, minimises peak broadening and background artefacts, increasing resolution and sensitivity. Once past the vacuum interface, ions and neutral masses not yet deflected, enter the lens component. Here electric charges are used to repel and attract the ions in order to focus them in order to maximise their transmission efficiency and the instrument’s performance. At the last stage of the lens assembly, the ions are electronically separated from any neutral masses and light photons by deflecting them from the axis along which they have travelled with the neutral species and photons to this point.

The next stage of an ion’s path through an ICP–MS depends on what type of configuration the system has, most commonly categorized as a ‘single-quad’ or ‘triple-quad’ system. For forensic intelligence applications of explosives, high sensitivity is required with triple quad systems preferred over single quad systems to achieve this. Because of this, only triple-quad systems are considered in this chapter. The superior ion sensitivity from this configuration is produced by utilising an isolating-removing-measuring process across the three different parts of the system. In the first quadrupole (Q1), masses of interest (including on-mass interferences) are isolated from the stream of ions from the source. The next stage involves a collision/reaction cell where a reaction gas is added which inhibits interference ions and reacts with other ions, usually in the form of polyatomic ions changing their mass, and removing them from the ion stream. Finally, the now “purified” ion stream consisting only of the target analyte then moves to the last quadrupole (Q2), where it is measured.⁵⁰  Using this triple-quad configuration routinely provides sensitivity in the range of parts per trillion (ppt), with some approaching detection limits in the parts per quadrillion (ppq) range. ICP–MS also allows simultaneous measurement of multiple elements (typically 20–30) in a single analysis and offers a large linear range spanning eight orders of magnitude with fine enough resolution to conduct isotopic measurements.

Sample introduction is a significant issue with the technique requiring species to be dissolved in liquid. When focussing on the analysis of metals, this generally requires digestion of the sample, using either acid or alkali solutions often at increased temperatures, increasing the risk of contamination due to the increased handling and use of chemical reagents.⁴⁷  An alternative to this is the use of laser ablation (LA), which offers the ability to directly sample solid material by imparting enough energy onto the material so that it transforms into a gas and this technique is used across various other forensic evidence types including glass and tapes.^52,53  As part of an LA–ICP–MS arrangement, the gas is then transferred into the ICP–MS system and analysed. The combination of LA with ICP–MS provides a number of advantages, including being semi-destructive, only requiring very small areas (approximately 100 µm) of the material to be consumed as a sample; little to no sampling preparation; and allowing the sampling to be conducted across multiple axes (i.e., 2D or 3D sampling). Results obtained from laser-ablated samples however tend to have greater variation in them given the chaotic nature of the ablative effect. Given this variation, it can be more advantageous to use more traditional techniques (i.e., XRF) in a forensic analysis of intact explosives and LA–ICP–MS for post-blast debris.⁵⁴ 

The volume of effort required to exploit trace elemental analysis of explosives for forensic intelligence is unquestionably greater than that listed in research publications. Given the relationship between explosives and national security, most case details and innovations relating to their analyses are tightly held on secure systems.

The published research efforts have predominantly involved inorganic mixtures (including sparkler compositions and ammonium nitrate based fertilisers) with C4 explosive formulations and homemade urea nitrate also investigated.^54,55  Sparklers are readily available products often encountered in relatively small pipe-bomb devices, while AN/CAN/urea fertilisers are important industrial/agricultural products manufactured and stored in large quantities potentially posing a national security risk and with a well-known history of use in terrorist devices (e.g., Oklahoma City bombing, Afghanistan/Iraq). Each study has attempted to create a database of explosive entities from different manufacturers/types/batches to provide a means to ‘match’ samples/sources; however, given the number of elements analysed, the resulting data are often complex.^54,56–59  To overcome this, the most common approach is to use statistical analyses with the goal of linking an unknown sample to one (or more) possible entities to establish a source. The statistical analyses utilised vary but generally include principle component analysis (PCA), linear discriminant analysis (LDA) or partial least squares-discriminant analysis (PLS-DA). Each of these computations is complex and automated with the resulting outcomes often needing to be accepted without the ability to verify the result by other means. This becomes challenging when being incorporated into intelligence where, ideally, the characteristics being used to establish a link are known and understood well enough to nominate the significance of the linkage. For this purpose, it seems that LDA, followed by a likelihood ratio/Bayesian analysis approach, is the most promising process moving forward as it is the most robust and utilises a process already widely accepted and used by the forensic community.

Data collected using laser ablation sampling become especially difficult to interpret. Coupled with the heterogeneous nature of compositions when sampled at a small scale, the variation in the plasma produced during the ablation process creates instances where trace elemental profiles of a sample as a whole cannot be easily established.⁵⁴  Taking this technique a step further and sampling post-blast debris via laser ablation is even more difficult. The chaotic nature of explosions can create a non-uniform spread of reaction products each with their own makeup. This makes forensic intelligence assessments almost impossible for such samples.

Below are some specific examples of forensic intelligence research and applications from explosive samples. These provide an appreciation of the outcomes that are delivered in the in real world, which at times, can be limited.

In an unpublished study, methyl ethyl ketone peroxide (MEKP) samples were synthesised under different reaction conditions, and purified and analysed using liquid chromatography mass spectrometry (LC–MS), an adaptation of published methodology.⁶⁰  The analyses were run on a Shimadzu LC–MS-8030, coupled with LC-30AD pumps, SIL-30AC autosampler and CTO-20AC column oven (see Table 1.8 for running conditions), delivering a limit of detection of 1 pg Reserpine.

MEKP is an organic peroxide formed by reaction of methyl ethyl ketone (MEK) with acidified hydrogen peroxide (Scheme 1.2 for possible reaction mechanisms).⁶¹ 

Possible peroxides formed (Scheme 1.3) with compounds 1–7, which have been identified in the literature,⁶²  and compound 8, which was deduced from the LC–MS analysis.

Species with m/z as high as 580 and 668 were detected (Figure 1.2).

This demonstrates the ability of laboratory analyses to detect and deduce precursors used in the explosive synthesis assisting investigations by opening up avenues of enquiry into suppliers/sources of chemicals and/or recipes and the level of knowledge of those who manufactured it.

As part of an investigation into an attack on a building, swabs of the damaged building surfaces were collected for the analysis of explosive residues. These swabs were heavily sooted and contained potential chemicals used by first responders as part of the matrix to be deconvoluted. To this point, no forensic intelligence had been provided as the initial analysis did not conclusively indicate the presence of any energetic material due to peaks being obscured by potential interferences. Further analysis was conducted on acetone extracts of the swabs utilising a UPLC-HRMS (Orbitrap) technique, relying on the Orbitrap’s higher selectivity. This method utilised nitrate adducts to enhance ionisation in the mass spectrometer (Figure 1.3) and required the nitrate adduct ion plus two isotopic ions for identification.

Upon analysis using this method, the laboratory was able to identify between approximately 1–50 pg of HMX and PETN residues on the swab of the building materials. RDX could not be identified, at the approximate 100–500 fg level, as the threshold required to detect the second isotopic ion was not met (Figure 1.4). However, the selectivity of the HRMS showed utility in being able to detect accurate masses of single mass unit ions with an interfering matrix.

A post-blast investigation of a vehicle was conducted in a residential area. The vehicle suffered minimal damage and it was clear that the device functioned only partially. Remnants from the explosive main charge were spread throughout the vehicle’s interior and later determined to be a mixture of ammonium nitrate and a medium petroleum distillate. Visual examination of the material provided forensic intelligence that the ammonium nitrate had been harvested from cold packs. Precursor chemicals recovered from the scene provided additional forensic intelligence, suggesting the possible manufacture and involvement of HMTD, most likely as an improvised detonator and booster. The vehicle was left at the scene for two days before being moved to an indoor facility for exploitation. During the first day of the vehicle exploitation, the interior of the vehicle was examined utilising an IMS system to determine if HMTD residues were present post blast, with one swab from the roof liner providing a positive HMTD response. The roof liner was subsequently cut out from the vehicle and packaged; however, the exhibits collected during this examination were not submitted to the laboratory for another three weeks following their collection.

Activated charcoal strips were used to absorb vapours from within the exhibit bags over an 18 hour period and desorbed according to a standard procedure. Portions of the roof liner, as well as other items recovered from within the vehicle, were also directly extracted with acetone and then filtered, concentrated and analysed using GC–MS (CI). None of the exhibits analysed in the laboratory confirmed the previous forensic intelligence of the presence of HMTD.

This example demonstrates that, although multiple opportunities for forensic intelligence can exist in a single case, often they can also be fleeting. Volatile explosives, such as HMTD, require that laboratory analyses on samples suspected to contain them be conducted as soon as possible to prevent the loss of the explosive material. Not only does this provide the greatest opportunity to formulate good intelligence, it is also important for court proceedings as presumptive on-site testing results from techniques such as IMS are not considered confirmatory results for court, and are not included in forensic reports.

A device was seized as part of an investigation with the main charge identified as pentaerythritol tetranitrate (PETN) using spectroscopic techniques. Unlike the previous case studies involving post-blast residues, where the identity of the explosives was a significant forensic intelligence goal, much was already known about this sample leading to more involved questions being asked. Where could this material have been obtained? Has it been stolen from a mining site or from the military? Has it been homemade? Crucially, each of these questions help to frame the overarching question, “How many more devices like this are out there, and what ongoing risk is posed to the community?”

PETN is typically used in commercial detonators, where it is present as a crystalline powder, as well as in detonating cords and boosters as part of a melt-cast formulation. Microscopic analysis was initially conducted to confirm the absence of fragments of plastic sheathing used to wrap PETN formulations into detonating cords. A combination of GC–MS and LC–MS, along with IC, was subsequently employed to explore different aspects of the sample. GC–MS was unable to confirm the presence of plasticisers. Analysis via LC–MS allowed the identification of pentaerythritol trinitrate, an intermediate product in the synthesis of PETN. While not utilising mass spectrometry, IC analysis also indicated the presence of ionic impurities, which may have been confirmed had an IC–MS been available in the laboratory. Holistically, the analyses provided the forensic intelligence that the PETN was homemade.

As the investigation progressed, a second PETN seizure from a known source was identified and it was hypothesized that the two shared the same source. A sample of the second PETN seizure was obtained and the carbon isotope ratios (δ13C) of each sample were measured. Published research exists on IRMS analysis of PETN; however, the value of the data is constrained by sample size limitations with most PETN production tightly secured as part of government apparatus and regulations. Additionally, PETN is produced with a very high (approx. 100%) purity through the nitration of pentaerythritol (PE) in a 1 : 1 PE : PETN ratio.^63–65  Carbon atoms within the PE precursor remain unaffected during this process; consequently, the corresponding isotope ratios of PETN directly relate to the ratios within the precursor. Although not available for this analysis, greater discrimination is provided by adding δ15N values to the dataset. The δ13C values for both seized samples with an approximately 1‰ difference were observed, three times greater than the measurement uncertainty, indicating different sources (see Figure 1.5) (the error bars represent expanded measurement uncertainty at 95% confidence intervals (k = 2)).

The analytical results, which allowed the differentiation between two samples, and the expert interpretation of the results in the context of the reaction process and wider research, offered two forensic intelligence outcomes:

1. The samples come from the one production facility however pentaerythritol supplies had changed;

2. The samples are unrelated and came from different production facilities, each with

   • their own supplier; or,

   • multiple suppliers.

The ambiguity of the intelligence provided is inherent to all real-world applications, given the reaction chemistry, and is compounded by the limitations of obtaining contemporary samples for analysis. Obtaining and sharing this data is one of the major challenges of providing forensic intelligence in casework with calls for addressing this highlighted in research.^66,67 

The exploitation of explosive samples using mass spectrometry to gather information for assessment and the provision of forensic intelligence is a field full of both numerous opportunities and challenges.

Generally, it is not feasible to build a single reference database for all explosives. Most commercially available explosives are controlled or prohibited items, which limits their availability. Large variabilities in the production of homemade explosives such as the type of precursors used, synthesis conditions, the stability of the explosives and its impurities, makes it extremely challenging to build a comprehensive database to provide forensic intelligence. Taking into consideration the challenges presented, the establishment of an explosive database to be used or shared internationally is a work in progress.

As with all science, forensic intelligence, evidence, or otherwise, the greater the number of similarities that are observed, the stronger the correlation between the samples being observed. In this respect, datasets from multiple MS analyses, including the combination of LA–ICP–MS and IRMS data, have also been combined in some studies, providing better results than single techniques only.⁵⁶  The establishment of databases where explosive samples are treated as entities within it, and the results of each technique listed as an attribute, could only lead to improvement. However, this will not overcome the challenges of open vs secure databases as discussed at the beginning of the chapter.

In an ideal world, law enforcement agencies across the globe, or subsets of them as formal trusted international and domestic partners, would share access to a secure database containing detailed chemical results. This would allow the comparison and potential linking of seized explosives, potentially identifying previously unknown terrorist or criminal networks and/or facilitating the monitoring of explosives trends over time and across regions. Of course, we do not live in an ideal world and chemical analysis capabilities vary broadly across countries, not to mention geopolitical barriers. This should not mean that such solutions cannot be endeavoured with like-minded and capable partners.

No studies yet identified, with respect to LA–ICP–MS analysis, have first separated samples into their individual explosive components prior to elemental analysis. Such an approach is often conducted in forensic laboratories for analyses using the technology currently used (as discussed previously). The inclusion of this step in forensic intelligence practices would improve the quality of the resulting assessments by providing a greater understanding of the origins of the correlations between samples. For example, from what explosive ingredient, say aluminium and ammonium nitrate, is a trace element profile observed? Analysing this formulation as a whole, or even as particles, could result in incorrect linkages if the presence of the trace element is due to the aluminium in one sample and the ammonium nitrate in the other.

Despite the challenges arising from variabilities in LA samples and trace, post-blast collections, there exists published literature on LA–ICP–MS analysis of organic gunshot residue (oGSR) particles.⁶⁸  This research included the collection of oGSR using traditional SEM stubs with a modified adhesive and subsequent analysis using separate techniques on each of the stub. One half of the sample underwent Raman spectroscopy analysis (necessitating the modified adhesive) while the other half was sampled and analysed using LA–ICP–MS (using optimised LA parameters to limit ablation of the aluminium stub). While demonstrating its success, the elements detected in the study were related to the ammunition itself (cartridge and/or projectile) or their primer compositions. While propellants can be encountered as charges in IEDs, the components detectable in this study are not. However, the study does highlight the opportunities to conduct LA–ICP–MS examinations on trace, post-blast samples. Previous studies have demonstrated that it is possible to capture explosive reaction products on SEM stubs metres away from detonations of multiple compositions including ammonium nitrate/aluminium and PE4 (RDX within a plasticiser matrix).⁶⁹  These were analysed using SEM–EDS (discussed previously), providing an ability to automatically image and analyse the sample. Particles of interest were highlighted based on their elemental composition. The superior sensitivity capability of ICP–MS in conjunction with LA sampling provides an avenue for trace elemental analyses for forensic intelligence studies on samples collected in a similar manner and it may be an avenue worth exploring.

Limitations too are also imposed by the explosive chemistry itself. Instances can occur at both ends of the spectrum, where either too little or much information is present for a meaningful determination. As demonstrated in case study #3, reaction mechanisms constrain the atoms are involved in explosive synthesis and can result in production of insufficient information. Transferring that to the final product will limit the determinations that can be made. Conversely, too much information can be gleaned from the analyses of some samples, for instance, from trace impurities from HMEs. Depending on the specifics of a reaction, the quality of its precursor chemicals, the steps taken to purify the final product, cleaning processes between batches, and so forth, a number of chemicals could be included in the final product at a range of concentrations. Analysis using highly sophisticated systems may become overwhelmed with data, particularly if the sample is not prepared properly. It can also become very difficult to link different batches of HME manufactured in the same clandestine laboratory if the trace chemical, or impurity, “profile” changes significantly. Often these types of laboratories will use whichever chemicals they can obtain to make their explosive of choice, which could even be dictated by the chemicals available.

Wastewater is already analysed in sewerage systems around the world for drugs⁷⁰  and is discussed extensively in Chapter 10. With the ongoing and ever evolving threat posed by terrorist utilisation of IEDs across the globe, the analysis of wastewater could provide valuable forensic intelligence for the early disruption of potential security threats. PhD research at the University College, London, has demonstrated the it is possible to detect peroxide and organic explosives using LC–MS analysis from wastewater collections in London, UK.⁷¹  The proposal within this study, suggesting the collaboration between forensic intelligence efforts and surveillance operations, which could include both physical and/or digital, to conduct targeted analyses of specific areas or addresses in real-world applications is a very valid one, especially when the specific case circumstances, for example, legislative restrictions or the operational security measures of targets, prevent covert activities to collect more direct samples from taking place. Achieving this is not without its challenges, including the complexity of the matrix, the dilution factor and the varying solubilities across the entire range of explosives and their precursors. Notwithstanding, opportunities exist to develop this methodology further and to deploy it into national security efforts.

The production of forensic intelligence from explosives samples has a history of over two decades, predominantly with the incorporation of portable systems into, or near, the crime scene to provide rapid presumptive results to feed into the investigation. Only in recent years however, has this been recognised as ‘forensic intelligence’, prompting more determined efforts to increase the scope of intelligence that forensic science can deliver. The improvement of portable systems, especially those using mass spectrometry techniques, will continue to be led by the private industry in the chemical detection market, in collaboration with law enforcement and military partners. The further development of these types of systems presents a significant opportunity to improve forensic intelligence produced during the earliest stages of an investigation from in-field analyses. Improvements to miniaturisation, sensitivity and specificity of current technologies are expected to continue as technology evolves, improving the accuracy and timeliness of forensic intelligence produced from the scene. Although technically challenging, this could be considered relatively easy given that each system can be improved upon in isolation. The most challenging aspect will be to incorporate multiple, highly sensitive, highly specific and complex analyses, including anticipated mass spectrometry, addressing current gaps such as inorganic explosives, from multiple samples into a coherent form allowing comparisons and assessments to be made to formulate meaningful intelligence. This cannot be done without sharing case data, most likely outside one’s own agency to partners requiring security solutions. Additionally, whether this can be achieved without complex multivariate data treatment is unknown. Is it enough to rely on PCA results, or minor differences in first derivative data, and call it intelligence or does the person producing it need to have some understanding of what the data treatment has simply highlighted and be able to explain it? To some extent, this will depend on the outcome with significant or critical intelligence perhaps necessitating human understanding and explanation. Remembering that a key parameter of forensic intelligence is its utility, the longer this process takes, the more insightful it should be, increasing the likelihood of it providing significant or critical intelligence and enhancing human understanding and reasoning. To this end, should complex multivariate data be used at all? Will artificial intelligence be accepted instead?

The opportunities are great, but so are the challenges.