Analytical and Bioanalytical Chemistry

, Volume 395, Issue 2, pp 301–313

Current trends in the detection of peroxide-based explosives


  • Raychelle M. Burks
    • Department of ChemistryUniversity of Nebraska
    • Department of ChemistryUniversity of Nebraska

DOI: 10.1007/s00216-009-2968-5

Cite this article as:
Burks, R.M. & Hage, D.S. Anal Bioanal Chem (2009) 395: 301. doi:10.1007/s00216-009-2968-5


The increased use of peroxide-based explosives (PBEs) in criminal and terrorist activity has created a demand for continued innovation in the detection of these agents. This review provides an update to a previous 2006 review on the detection of PBEs, with a focus in this report on luminescence and fluorescence methods, infrared and Raman spectroscopy, mass spectrometry, and electrochemical techniques. Newer developments in gas chromatography and high performance liquid chromatography methods are also discussed. One recent trend that is discussed is an emphasis on field measurements through the use of portable instruments or portable assay formats. An increase in the use of infrared spectroscopy and mass spectrometry for PBE analysis is also noted. The analysis of triacetone triperoxide has been the focus in the development of many of these methods, although hexamethylene triperoxide diamine has received increased attention in PBE detection during the last few years.


Peroxide-based explosivesTriacetone triperoxideHexamethylene triperoxide diamineExplosives analysis


Triacetone triperoxide (TATP) and hexamethylene triperoxide diamine (HMTD) were first synthesized in the 1880s using simple recipes calling for just three ingredients: hydrogen peroxide (H2O2), an acid, and acetone (in the case of TATP) or hexamine (in the case of HMTD) [16]. In the intervening years, these peroxide-based explosives (PBEs) have seen little to no military or civilian use owing to their extreme sensitivity to mechanical stress, limited stability, high volatility, and lower explosive power compared with nitro-based explosives [3, 5, 714]. Nitro-based explosives such as trinitrotoluene (TNT) may be more powerful, but the intensity of PBE explosions is substantial and destructive [7, 8]. The power of PBEs, along with their simple synthesis from readily available materials, has led to their increased use in improvised explosive devices (IEDs) for criminal and terrorist activities.

Terrorist attacks using PBEs first occurred in Israel in 1980 [15]. However, PBE detection methods received little attention prior to a series of high-profile terrorist plots during the last decade. These plots included an attempt on American Airlines transatlantic flight 63 using a PBE IED, the Casablanca explosions in 2003, the 2005 London public transportation attacks, and a UK transatlantic flight bombing attempt in 2006. These events made the fast and reliable detection of PBEs and their precursors a research priority [8, 1620]. Designing a detection scheme for PBEs is no easy task given their sensitivity to mechanical stress and low stability, lack of UV absorbance or fluorescence, and limited solubility [2, 5, 7, 8, 12, 21, 22]. These challenges have recently been overcome. Today there are an array of techniques for the quick and reliable detection of PBEs, their precursors, and degradation products [8, 9, 18, 2326].

This journal presented a review of PBE detection in 2006 [8], which provided an excellent overview of established and new methods. This current article focuses on PBE detection trends that have appeared over the last 3 years or work that was not included in the previous review. This review is organized according to the detection mode that was employed and includes work that has focused on the two most commonly encountered PBEs (i.e., TATP and HMTD) along with their precursors and degradation products. The structures of TATP and HMTD, along with key properties of these explosives, are given in Table 1. In addition, select methods targeting the analysis of H2O2 in explosives have been included in this current review because H2O2 is a precursor for and degradation product of TATP and HMTD and is also used in IEDs [15]. Table 2 summarizes the PBE detection techniques that are highlighted in this review.
Table 1

Key physical and chemical properties of triacetone triperoxide (TATP) and hexamethylene triperoxide diamine (HMTD)

The melting points, densities, detonation velocities, and trinitrotoluene (TNT) equivalence data were taken from [7]. TNT equivalence compares blast over pressure or impulse of the explosive of interest with a similar amount of TNT. Values for the TATP and HMTD vapor pressures and enthalpies of sublimation were acquired from [76]. The TNT vapor pressure value is from [76].

Table 2

Recent detection and analysis methods reported for peroxide-based explosives (PBE)




Detection scheme





\( LG/PVA\xrightarrow{{{H_2}{O_2}}}LG\;degradation\;product/PVA \)

<1 wt% H2O2 (g)


\( 6M{o_2}{O_5}\left( {OH} \right)\xrightarrow{TATP}12Mo{O_3} + 3{H_2}O + {\left( {C{H_3}} \right)_2}CO \)




\( PolyF - 1\xrightarrow{{{H_2}{O_2}}}fluorescein \)

300 ppm (10 min)

H2O2(g), 30 ppm (30 s)

H2O2(l), 1 ppm (5 min)

H2O2 (l)



\( \surd_H^{+} \)

\( {H_2}\left( {Salen} \right)\xrightarrow{{{H_2}{O_2}Z{n^{II}}}}Zn\left( {Salen} \right) \)

10 nM TATP




\( pyrene\;sulfoxides\xrightarrow{{{H_2}{O_2}}}pyrene\;sulfones \)

100 nmol TATP


\( trimeric\;SID\xrightarrow{{{H_2}{O_2}}}3fluorophore\;reporters \)

1 μM H2O2, approximately microgram TATP



\( nano - CRET\xrightarrow{{{H_2}{O_2}}}nano - CRET + \lambda \)


IR and Raman spectroscopy



Gas-phase FTIR with PLS-DA




FTIR, GC-FTIR, and Raman microscopy




Hollow-fiber MIR QCL gas sensor

240 ng TATP



Hollow-fiber or open-path MIR QCL gas sensor

TATP, low nanogram (hollow fiber), 5 ppm per meter (open path)

 [17, 44]


MIR QCL device (walk-through portal)

15 ppb H2O2



Fiber-coupled MIR QCL device (handheld)





18 ppb TATP, 3 ppb acetone


Raman field-portable device (FirstDefender, Ahura Scientific)




Raman microscopy





1 pg HMTD

MS and IMS



IMS (ItemiserFX, General Electric)

1.9 μg TATP (E-mode), 0.8 μg TATP (N-mode)



Aspiration IMS (ChemPro100i, Environics)

Low milligrams per cubic meter TATP



Headspace GC-MS

<0.1 ng TATP




5 ng TATP



CH4 (g) and NH3(g) GC/PICI-MS or GC/NICI-MS, EI-MS

50 pg–2 ng TATP



Na+ adduct ESI-MS

62.5 ng TATP



Alkali metal DESI-MS

Low nanogram TATP or HMTD








Laser TOF MS





Low parts per billion TATP

Electrochemical methods

 [68, 70]



\( {H_2}{O_2}\xrightarrow{PB - electrode}2H{O^{-} } \)

250 nM TATP (UV),


300 nM HMTD (UV),

\( \surd_H^{+} \)

50 nM TATP (laser), 55 nM TATP (H+)

 [16, 71]


\( \surd_H^{+} \)

\( {H_2}{O_2}/ROOH\xrightarrow{{\begin{array}{*{20}{c}} {F{e^{II/III}}EDTA} \\ {GC - Electrode} \\ \end{array} }}H{O^{-} }/R{O^{-} } + H{O^\bullet } \)

890 nM TATP (H+), 30 μM HMTD (pH)




MPc chemiresistor

50 ppb–40.1 ppm H2O2

Other methods




1 mM TATP, 0.5 mM HMTD



Field-portable GC device (zNose, Electronic Sensor Technology)

Low parts per trillion TATP



Differential scanning microcalorimetry


LOD limit of detection, LG lissamine green, PVA poly(vinyl alcohol), PolyF-1 poly[3′,6′-bis(1,3,2-dioxaborinane)fluoran], H2Salen N,N′-ethylenebis(salicylaldimine), SID self-immolative dendrimer, CRET chemiluminescent reactor, FTIR Fourier transform IR, PLS-DA partial least squares regression with discriminant analysis, GC gas chromatography, MIR mid-IR, QCL quantum cascade laser, PAS photoacoustic spectroscopy, SERS surface-enhanced Raman spectroscopy, IMS ion mobility spectrometry, MS mass spectrometry, SPME solid-phase microextraction, PICI positive ion chemical ionization, NICI negative ion chemical ionization, EI electron impact, ESI electrospray ionization, DESI desorption electrospray ionization, DAPCI desorption atmospheric pressure chemical ionization API atmospheric pressure ionization, TOF time of flight, SPI single photoionization, PB Prussian blue, HPLC high-performance liquid chromatography, EDTA ethylenediaminetetraacetate, MPc metallated and metal-free phthalocyanines

aA check mark denotes the method directly monitors PBEs.

bA single check mark corresponds to simple H2O2 monitoring. A check mark with notation indicates the method indirectly monitors PBEs by first producing H2O2 by photodecomposition (√UV or √laser), PBE acid digest (\( \surd_H^{+} \)), or low pH (√pH).

Luminescence methods

Presumptive tests based on changes in color, fluorescence changes, or chemiluminescence can provide quick and reliable results for a variety of target analytes. Such luminescence-based methods were reviewed previously in this journal with regards to the detection of explosives [9]. The methods presented here were recently introduced and target PBEs and/or the precursor H2O2. In an effort to detect H2O2 through a simple color test, Mills et al. [27] encapsulated the triarylmethane dye lissamine green (LG) in poly(vinyl alcohol) (PVA) to monitor the bleaching of LG by H2O2. Experiments conducted in solution showed that H2O2 bleaching of LG through rapid oxidative degradation is slow at pH values significantly below the pKa of H2O2 (11.75). However, if LG is placed in a largely neutral polymeric environment, it is made particularly vulnerable to oxidative bleaching by H2O2 vapors. When blue-green LG/PVA films cast on glass discs were placed above a 50% (w/w) aqueous H2O2 solution, significant bleaching was observed in less than 5 min. Adjusting the film thickness allowed bleaching of LG/PVA by vapors above a 1% (w/w) H2O2 solution. Although the exact bleaching mechanism is unknown, it is known that the bleaching is due to degradation of LG and the mechanism is probably similar to the H2O2 -induced oxidative degradation of another triarylmethane dye, phenolphthalein. Specificity was a problem for the LG/PVA films, as researchers noted other volatile strong oxidizing agents such as ozone, chlorine, and nitrogen dioxide produced bleaching. The authors stated that although this trait is undesirable, LG/PVA films were found to be rapid sensors for strong oxidizing agents and had possible applications in the area of PBE detection [27].

In an interesting use of nanomaterials for PBE detection, Apblett et al. [28] used molybdenum hydrogen bronze [MoHB; Mo2O5(OH)] to detect and deactivate TATP. Owing to its high acidity and metallic properties, MoHB is capable of shuttling electrons and protons to peroxide- and nitro-based explosives, leading to their decomposition to nonexplosive compounds. Researchers added a suspension of MoHB in butanol, which was dark blue, to solid TATP, TATP in toluene, or water. The reaction between TATP and MoHB was found to lead to the disappearance of the suspension’s blue color. Excess TATP resulted in a yellow color due to the formation of peroxo-molybdenum complexes. This reaction and its accompanying color changes were dramatic enough for it to be run as a titration, with a persistent blue color of the sample solution marking the end point. Researchers also made test strips based on this reaction, noting that exposure to either TATP or H2O2 vapors rapidly bleached the blue color. This reaction was noted to be general in nature, occurring between H2O2 or ROOH and MoHB, as represented in Eqs. 1 and 2.
$$ 2M{o_2}{O_5}\left( {OH} \right) + ~{H_2}{O_2} \to 4Mo{O_3} + ~2{H_2}O $$
$$ 2M{o_2}{O_5}\left( {OH} \right) + ~ROOH \to 4Mo{O_3} + ~{H_2}O~ + ~ROH $$
Concerning the reactions between MoHB and TATP, analysis by gas chromatography (GC)-mass spectrometry (MS) of the headspace above the reaction mixture confirmed that the main ROH species formed was acetone. Given the common response of this method using either H2O2 or TATP, along with the indication that other organic peroxides can produce the same color change, there is a clear possibility of false positives in this approach. The limit of detection (LOD) for this method was not addressed in the initial report of this technique [28].
Another assay that can be easily integrated into a test strip format is the fluorimetric method introduced by Sanchez and Trogler [29]. This method targeted H2O2 vapor and related liquids because it was noted that residual H2O2 may be present in bulk TATP and HMTD, with H2O2 being both a PBE precursor and a degradation product. After the polymer poly[3′,6′-bis(1,3,2-dioxaborinane)fluoran] (PolyF-1) had been synthesized, thin films were fabricated by drop-casting the polymer onto sheets of filter paper (4 cm2). Exposure of PolyF-1 to H2O2 led to oxidation of the polymer and formation of fluorescein for use in detection, as shown in Fig. 1. This method had an LOD for H2O2 vapors of 300 ppm when a 10-min PolyF-1 exposure time was used (note the LOD dropped with increased exposure time). For liquid-based H2O2 samples, a 30-s PolyF-1 exposure time gave an LOD of 30 ppm, and 5 min of exposure gave an LOD of 1 ppm. It was stated that the specificity of boronic esters toward H2O2 oxidation makes PolyF-1 a highly sensitive and selective sensor for H2O2. Under ambient conditions and under UV light, PolyF-1 showed little response to radical oxygen species and other oxidants found in the atmosphere or generated by a UV lamp (λ = 302 nm). The authors suggested that the lower vapor pressure of organic peroxides relative to H2O2 precludes their possible interference in this method. Previous solution-phase studies also indicated that there was little to no response from possible liquid interferents [29].
Fig. 1

Polymer-based H2O2 sensor [29]. Fluorescence response of a 10 μg cm−2 film of poly[3′,6′-bis(1,3,2-dioxaborinane)fluoran] (PolyF-1) to 2.9 ppm H2O2 vapor. The solid line at 0 min represents the baseline fluorescence intensity of the PolyF-1 film. The dashed line represents the fluorescence emission of 10 μg cm−2. (Courtesy of W. Trogler)

A fluorescence detection method by Germain and Knapp [30] targeted H2O2 by using a chelator formed by reaction with H2O2. Taking advantage of the ability of H2O2 to convert C-B bonds to C-O bonds, these researchers designed a boronated prochelator that can be easily converted to the chelator N,N′-ethylenebis(salicylaldimine) by means of H2O2 deboration. This reaction could be easily monitored. The addition of H2O2 to a methanol solution of prochelator and Zn(acetate)2 resulted in a fluorescence signal with maximum emission at 440 nm. The LOD for H2O2 in this method was below 10 nM. Substituting benzoyl peroxide for H2O2 gave a similar fluorescence response. TATP solicited no such response, indicating TATP could not deboronate the prochelator. TATP was also subjected to acid digest using 1 M acetic acid to produce H2O2, giving an 80-fold increase in fluorescence signal relative to the standard prochelator/Zn2+ solution. The authors suggested that benzoyl peroxide was hydrolyzed by the low levels of water present in the reaction mixture but that, overall, organic peroxides would not produce fluorescence in this method [30].

Malashikhin and Finney [31] took advantage of florescence detection by investigating the use of various sulfur-containing pyrene derivatives in the presence of methyltrioxorhenium as visual sensors for TATP. These researchers settled on the oxidation of pyrene sulfoxides to sulfones, on the basis of their observation that these reactions gave the greatest fluorescence signal compared with other sulfur oxidation reactions. TATP did not react directly with the pyrene sulfoxide profluorophores that were tested; however, rapid oxidation was achieved using the H2O2 produced through UV irradiation of TATP. The resulting pyrene sulfones displayed a fivefold increase in fluorescence after 15 min of reaction relative to the profluorophores. A 90-min reaction gave a fluorescence signal visible to the naked eye for 100 nmol TATP that had been subjected to UV irradiation. It was noted that oxidants such as tert-butyl hydroperoxide, NaOCl, LiClO4, K2Cr2O7, and air did not appreciably react with their profluorophores. However, KMnO4 did undergo such a reaction. Although the profluorophores were stable in visible light, they were not stable with prolonged exposure to UV radiation.

A shift in fluorescence, rather than the generation of fluorescence, was used in an H2O2 method employing self-immolative dendrimers (SID) [32]. SIDs are unique molecules that upon a single activation event will self-eliminate their end groups. This process leads to complete dissociation of the dendrimer into separate building blocks. A fluorescent trimeric SID was synthesized that contained an aryl borate ester, a functionality that reacts with H2O2 under mild alkaline conditions (NaHCO3 solution, pH 8.3). Such a reaction begins a series of self-elimination events that cause the trimeric SID to release three “reporter” units. The release of reporter units redshifts the fluorescence signal of the SID from its maximum emission at 450 nm to 510 nm. An LOD of 1 μM was reported for H2O2 when this approach was used. These SID probes were also reactive with TATP under alkaline conditions, with detection being possible in the microgram range. Reaction times ranged from 90 min (for H2O2) to 120 min (for TATP).

A second H2O2 assay employing nanomaterials was a chemiluminescent nanoreactor (nano-CRET) method introduced by Wingert et al. [33]. Hollow calcium phosphate (CaP) nanoshells were fabricated by coating a phospholipid liposome with a nanometers-thick layer of CaP. Encapsulated inside these nanoshells was a fluorescein-enhanced chemiluminescent luminol system with hematin. Incoming H2O2 reacted with luminol, generating excited intermediates. A portion of these intermediates produced chemiluminescence at 425 nm, whereas others engaged in Förster resonance energy transfer with fluorescein molecules and produced fluorescence that was observed at 525 nm. Compared with the efficiency of light production in the same chemiluminescent reaction in bulk solution, the efficiency was increased by using the nano-CRET method owing to the improved proximity of reactive species. Use of simple micelles and liposomes gave a similar improvement in light production efficiency; however, the researchers sought to limit interferences by organic molecules by restricting entry into their liposome through the use of a CaP shell. The assay time was not explicitly stated in this report and the authors stated that quantitative determination of LODs is currently under way [33].

IR and Raman spectroscopy

One of the first analytical methods used to characterize and detect TATP or HMTD was IR spectroscopy [4, 8]. IR spectroscopy and the related technique of Raman spectroscopy are classic tools for the identification of unknown chemicals (see the references cited herein for details on the operation of these methods), although these methods can suffer from some difficulties when one works with complex mixtures. Both IR spectroscopy and Raman spectroscopy have been used to identify and characterize PBEs, along with related compounds [2, 8, 22, 3436]. Gas-phase IR spectroscopy and Raman spectroscopy are especially well suited to the analysis of PBEs given the relatively high vapor pressure of PBEs (Table 1), which results in good LODs and often means no sample preparation is required for this type of analysis. Pacheco-Londono and co-workers [24, 35, 3740] used IR and Raman spectroscopy to study PBEs, their precursors, and by-products, as well as structurally similar compounds. Recently, an IR spectrum pattern recognition process was created based on partial least squares regression with discriminant analysis [37]. In-flow gas-phase IR spectroscopy was used to generate spectra for TATP and select nitro-based explosives in the near-IR (NIR) and mid-IR regions. Solid explosives, in amounts ranging from 100 to 300 μg cm−2, were deposited and examined in the presence of an air flow of 80–120 mL/min at various temperatures. The researchers in this study found that the near-IR region offered statistically significant differences for identifying explosives in air, but no explicit LODs were stated.

GC/Fourier transform IR (FTIR), FTIR, and Raman microscopy were used to characterize and differentiate a collection of cyclic organic peroxides [35]. TATP, diacetone diperoxide (DADP), and tetraacetone tetraperoxide (TRARP) were synthesized in-house and analyzed to determine differences in the IR and Raman spectra for such similar peroxides. Differences were found in the ν(O-O), ν(C-O), δ(CH3-C), and δ (C-O) spectral bands. Although all of the cycloperoxides examined had a Raman signature with a ν(O-O) vibration, the researchers found that this band could be used to determine if a dimer or a trimer of a peroxide (e.g., DADP vs. TATP) was present. LOD values were not reported in this paper [35].

Oxley et al. [34] sought to identify IR or Raman spectral lines of high intensity in regions clear of peaks that result from atmospheric species. This research identified clear “windows” at 909–1,333, 2,083–2,273, and 2,381–2,630 cm−1 and set out to explore the detection of PBEs using these spectral regions. It was found for these windows there were no unique spectral features allowing for PBE differentiation, with a broad spectral region being required to make reliable PBE identifications. This work was qualitative in nature and no LODs were explicitly stated [34].

A focus on the mid-IR region and the use of quantum cascade lasers (QCLs) has resulted in methods for the trace detection of PBE vapors. Explosives show strong and distinct absorption bands in the mid-IR region of 5–10 μm (or 2,000–1,000 cm−1). This feature makes quick and sensitive probing of PBEs possible through the use of QCLs [41, 42]. Lambrecht et al. [43] used hollow fibers as compact IR gas sensors and monitored QCL mid-IR light absorption by TATP. This analysis took only seconds to conduct and gave an LOD of approximately 10 g/L or 240 ng. Recently, this group extended its investigation of a hollow-fiber QCL for standoff and extractive TATP detection, in addition to an open-path QCL [42]. For hollow-fiber detection, the LOD was in the low nanogram range. An LOD of 5 ppm/m was achieved for open-path experiments in a laboratory setting, but it was noted that a smaller LOD would be required for realistic standoff measurements.

QCL-based systems are making impressive gains in the area of PBE detection in high-traffic areas such as airports and train stations. Ongoing research at Cascade Technologies (Stirling, UK) has focused on a walk-through portal using a quasi-continuous wave intrapulse QCL regime for the fast and reliable detection of explosive precursors such as ammonia and H2O2 [17, 44]. This portal has fans to create air flow across the walkway and IR spectra are collected in milliseconds. Researchers reported an LOD of 15 ppb for H2O2 in this approach.

For the close-up monitoring of suspicious materials, a handheld sensor was designed employing a fiber coupled continuous wave distributed feedback QCL which was operated at 1,235.1 and 1,245.3 cm−1. This sensor, shown in Fig. 2, was utilized for the detection of TATP in ambient air [41]. This sensor was placed about 1 cm above a few milligrams of TATP under ambient air conditions and gave distinctive and reproducible spectra. The researchers identified unique spectral features when they compared TATP and its precursor acetone. TATP and its precursor acetone were also the target of a QCL photoacoustic spectroscopy technique designed by Dunayevskiy et al. [45]. These researchers used an array of wavelengths (dubbed a “smart grid”) to bypass the interference of water vapor under ambient conditions. Distinct spectra were collected for TATP, acetone, and TNT, with LODs of 18 ppb for TATP and 3 ppb for acetone. Integration of this technique with walk-through portal devices may be possible in the future.
Fig. 2

Quantum cascade laser (QCL) based handheld sensor device [41], including the a general sensor design and b image of the sensor head. (Courtesy of C. Bauer)

Raman-based systems for the field analysis of explosives and other compounds of forensic interest are commercially available and have been described previously [46]. A study of a portable Raman device known as FirstDefender, made by Ahura Scientific (Wilmington, DE, USA), has recently been presented [47]. This device was introduced in 2005 [48] and incorporates a dispersive Raman spectrograph that includes a 785-nm laser and a charge coupled device detector along with a database of over 4,000 compounds and mixtures for vapor monitoring. This device is designed to allow rapid identification of suspect material through transparent containers such as plastic or glass bottles. An evaluation of this device found that discrimination is possible between TATP, HMTD, and organic peroxides such as methylenthylketone peroxide; however, LOD values were not reported. An FTIR handheld device known as TruDefender, also made by Ahura Scientific and introduced in 2008, offers many of the same features as FirstDefender [49].

Given the urgent need for portable and standoff detection ready devices, it is not surprising such devices have become the focus of much research. For standoff screening of bottles for liquid explosives precursors (e.g., H2O2 for PBEs), Stokes et al. [50] used Raman microscopy. These researchers used a Raman microscope with a long working distance lens and found that closed plastic bottles could be reliably screened for 30% H2O2 with an analysis time of 100 ms. At the same time, the liquid explosive combination of H2O2/water/ethanol [50, 51] could also be detected with component differentiation.

An approach that offers promise for pushing the boundaries of Raman detection limits is surface-enhanced Raman scattering (SERS). The metal surfaces that are employed in this technique enhance the Raman signal via the large electromagnetic fields that are present on the small gaps between metal nanoparticles [52]. Taking advantage of the additional waveguide ability of a cylindrical SERS substrate, Ko et al. [52] designed a substrate of alumina nanopores containing gold nanoparticle clusters for the detection of explosives, including HMTD. Fabrication of the substrate that was utilized in this work is illustrated in Fig. 3a. After fabrication of a porous alumina membrane, the surface of these pores was modified with polyethyleneimine, with the amine groups providing a convenient way to attach gold nanoparticles that were capped with cetyltrimethylammonium bromide. Figure 3b shows the Raman spectra of HMTD that were obtained at several concentrations on the SERS substrate. This approach gave an LOD of approximately 1 pg for HMTD precipitated on the substrate. Although this type of research makes SERS a promising method for PBE detection, there are practical limitations and factors (e.g., the need for an appropriate SERS substrate and related sample preparation) to consider when using this technique.
Fig. 3

Surface enhanced Raman scattering (SERS) detection of hexamethylene triperoxide diamine (HMTD) [52]. a SERS substrate fabrication (see the text for details) and b SERS spectra of HMTD; characteristic signature peaks are marked. PEI polyethyleneimine, NP nanoparticles. (Courtesy of V. Tsukruk)

Mass spectrometry and ion mobility spectrometry

Along with IR and Raman spectroscopy, MS was one of the first techniques used to analyze PBEs and related compounds [4, 8]. TATP synthesis by-products [53, 54] and acid degradation products [55], along with the thermal decomposition of TATP [56] and HMTD [5], have all been studied using MS techniques. MS methods have seen wide application in the field monitoring of explosives and narcotics.

One technique that has been of particular interest in this work is ion mobility spectrometry (IMS). Because they offer suitable sensitivity, reliability, and easy operation, IMS instruments are often found in airports, in government buildings, and at border crossings [8, 57, 58]. Studies of two commercially available IMS field-friendly instruments have recently been published [57, 58]. Using General Electric’s ItemiserFX, Oxley et al. [57] developed a method to detect explosives in hair. This instrument is geared toward detecting narcotics and explosives and can operate in both positive and negative ion modes. As most narcotics have a positive ion affinity, the detection of positive ions by this device is called the “N-mode”, whereas the negative “E-mode” is used for nitro-based explosives, which have a negative ion affinity. Experiments were run in both modes to test their use for nitro-based explosives and TATP. The high vapor pressure of TATP proved to be an experimental challenge owing to its quick desorption from hair. Longer TATP exposure times and amounts, in addition to larger hair samples, were required to detect this analyte in hair. LODs of 0.8 μg (N-mode) and 1.9 μg (E-mode) were reported. Another field-ready IMS instrument was studied by Rasanen et al. [58] for use in TATP detection. An aspiration-type IMS has been integrated with semiconductor gas sensors in a handheld device called ChemPro100i from Environics (Toronto, ON, Canada). TATP vapor was measured under ambient conditions with this device and gave an LOD in the low milligrams per cubic meter region, as verified by GC-MS. However, the detection of TATP in complex matrices when this device was used was not reported.

The high vapor pressure of TATP has been a boon for the application of headspace GC-MS in detecting this explosive. Stambouli et al. [19] designed a headspace GC-MS technique targeted at detecting trace TATP in postexplosion debris. For this study, debris was collected from the 2003 Casablanca explosion by the forensic laboratory of Moroccan Gendarmerie Royale. Both TATP and its by-product DADP were easily detected though extensive decomposition and/or fragmentation results from thermal degradation and MS ionization. Characteristic ion peaks were present for both TATP and DADP. The method developed was then used to examine postblast debris that had been collected in glass containers and hermetically sealed. The final procedure included heating the glass sample container for 30 min followed by sampling 1 mL of the headspace vapor for analysis. An LOD of 0.1 ng was reported for TATP, but the LOD of DADP was not provided.

GC-MS was combined with solid-phase microextraction by Kende et al. [59] for the trace analysis of TATP. These researchers used polydimethylsiloxane fibers to trap TATP vapors in the headspace of sample containers, followed by transfer of fibers to the injector of a GC system kept at 160 °C. A maximum signal was achieved when the exposure time of TATP to the fibers was 20 min. Electron impact MS was used with a trio of indicative ion peaks for compound identification, including the parent ion. Researchers examined a variety of model preblast and postblast samples, such as TATP-contaminated soil, with favorable results. An LOD of 5 ng for TATP was reported.

Sigman et al. [53, 54, 60] used a variety of MS modes to detect and characterize the fragmentation of TATP and its synthesis by-products when these chemicals were subjected to collision-induced dissociation (CID). Low-nanogram LODs were achieved for the GC-MS analysis of TATP using ammonia or methane positive ion chemical ionization and negative ion chemical ionization along with electron impact [60]. It was found that ammonia positive ion chemical ionization was the best overall method, as a diagnostic adduct [TATP + NH4]+ was consistently detected and gave LOD values of 0.5 ng (ion trap) or 0.1 ng (quadrupole). Sigman et al. [53, 54] next used electrospray ionization (ESI) MS to monitor both ammonia and sodium adducts of TATP and its oligoperoxide by-products. Sodium adducts for TATP, previously seen using desorption ESI (DESI) MS [61], were observed along with a new series of ions corresponding to [oligoperoxides + Na]+ [53]. An LOD of 62.5 ng was reported for TATP. This sodium adduct technique was used to analyze TATP synthesis products in postblast samples, with trace amounts of TATP and oligoperoxides being detected after detonation. TATP synthesis reaction mixtures, which include a variety of oligoperoxide by-products, received more attention in recent work [54] in which detailed CID mechanisms of sodiated and ammonium adducts were determined using deuterium isotopic labeling tandem MS experiments. The CID mechanisms differed for the sodiated and ammonium adducts; smaller oligoperoxide ammonium adducts formed cyclic peroxides, whereas sodium adducts did not. Both adduct forms were noted to undergo extensive fragmentation, as shown for ammonium adduct CID in Fig. 4. It can be seen in this example that the 314 m/z peak corresponding to TRARP [TRARP + NH4]+ is quite abundant, as is the 240 m/z peak for [TATP + NH4]+. Studies of various synthetic TATP batches revealed a variation in oligoperoxide distribution between batches, a feature that could prove to be useful in forensic analysis. Distribution of oligoperoxides was found to shift in preblast and postblast samples, an effect that was likely due to thermal decomposition.
Fig. 4

Analysis of oligomeric peroxides in synthetic triacetone triperoxide (TATP) samples by electrospray ionization mass spectrometry (MS) [54]. Product ion spectrum obtained from collision-induced dissociation of m/z 348 [H(OOC(CH3)2)4OOH + NH4]+. Major m/z peaks are identified. (Courtesy of M. Sigman)

Detection of sodiated and ammonium adducts of TATP was introduced by Cotte-Rodriquez et al. [61, 62]. Their DESI MS technique was discussed in detail in a previous review [8]. These same researchers recently presented a DESI MS method for the rapid detection of trace amounts of TATP, TRARP, and HMTD directly from ambient surfaces with no sample preparation [63]. In addition to the use of sodium and ammonium, this group investigated the use of potassium and lithium for complex formation. Positive ion DESI spectra for TATP and HMTD are shown in Fig. 5. Rapid detection (less than 5 s) was achieved for the target PBEs in complex matrices (e.g., diesel fuel) using single or multiple cation additives. The LODs were in the low-nanogram range. The use of desorption atmospheric pressure chemical ionization (DAPCI) MS was also explored for the detection of TATP and HMTD. Trace amounts of HMTD were easily detected by DAPCI when methanol vapor in nitrogen was used, but gave insufficient ionization for TATP. This effect was attributed to the lower proton affinity of TATP relative to HMTD. The higher proton affinity of HMTD is due to its two basic amine groups. For TATP detection, ammonium acetate was added to the DAPCI gas so that ammonium adducts could be monitored. Favorable results led to modification of the HMTD DAPCI regime to also include ammonium acetate. LODs for all experiments were in the low-nanogram range.
Fig. 5

Desorption electrospray ionization (DESI) MS detection of TATP and HMTD [63]. Positive ion DESI mass spectra of 10 ng TATP (a) and HMTD (b) deposited on paper in an area of 1 cm2. Methanol/water (70:30) doped with 10 mM NaCl was used as a spray solvent. Insets: Product ion MS/MS spectrum of (peroxide-base explosive + Na)+ complex. (Courtesy of R. Cooks)

Ammonia hydroxide treated TATP, HMTD, and tetramethylene diperoxide dicarbamide (TMDD) were analyzed by Pena-Quevedo et al. [64] using an AccuTOF DART instrument (JEOL USA, Peabody, MA) [64]. DART (direct analysis in real time) is the sampling component which is coupled to an atmospheric pressure ionization (API) time of flight (TOF) mass spectrometer. Compounds were synthesized in-house and characterized by Raman and IR spectroscopy. Reaction mixtures were subjected to minimal purification prior to MS analysis. Positive ammonia adducts were detected for all PBEs. For HMTD, the parent ion peak [HMTD + H]+ was more abundant than the ammonia adduct, whereas for TATP the ammonia adduct [TATP + NH4]+ was the only peak present. To explain the lack of a fragmentation pattern for TATP, the authors suggested that TATP is stabilized by ammonia. For TMDD, the [TMDD + NH4]+ peak was more abundant than the peak for the parent ion, [TMDD + H]+. It was reported that trace PBE analysis could be performed by this approach, although LOD values were not provided.

Mullen et al. [65, 66] used laser photoionization for MS studies of TATP. When comparing femtosecond and nanosecond laser pulses for the analysis of TATP vapor by TOF MS [65], they noted that a parent ion peak was only present in the femtosecond laser pulse spectra. This shorter pulse provided “softer” ionization and yielded more abundant acetone ion peaks compared with previously published GC-MS analysis data for TATP. Single photoionization (SPI) TOF MS was next used to detect a variety of explosives and related compounds in the gas phase [66]. In SPI MS, the parent molecule was directly ionized using a vacuum ultraviolet (VUV) photon. A single VUV photon was absorbed by a molecule and, if the photon energy was higher than the molecule’s ionization potential, an electron was removed. The limited ionization energy of VUV photons can efficiently ionize organic compounds but not gases such as nitrogen, oxygen, and water, because these gases have relatively high ionization potentials [66, 67]. A diagram of this SPI TOF MS system is shown in Fig. 6a. As can be seen in Fig. 6b, TATP underwent extensive fragmentation, with the acetyl ion (43 m/z) being the most abundant fragment ion, although the parent ion was also visible at 222 m/z. For TATP, an LOD in the low parts per billion range was achieved. To increase the application of SPI MS in the detection of explosives and narcotics, Schramm et al. [67] determined the ionization potentials of several such compounds using monochromatized synchrotron radiation from BESSY (i.e., Berliner Elektronenspeicherring-Gesellschaft für Synchrotronstrahlung). This latter work was qualitative, aimed at providing ionization potentials for an array of forensically important compounds.
Fig. 6

TATP detection by single photoionization (SPI) time of flight (TOF) MS [66]. a The SPI TOF MS instrument and b the SPI mass spectrum of TATP; parent molecular ion (222 amu) and a number of photodissociative products, including acetyl ion (43 amu), acetone ion (58 amu), C3H7O+ (59 amu), C3H7O2+ (75 amu), C3H6O4+ (106 amu), and diacetone diperoxide, C3H6O5+ (122 amu). (Courtesy of H. Oser)

Electrochemical methods

Explosives detection by electrochemical means was recently reviewed by Wang [25]. This previous review focused on sensors for commercial and homemade explosives, with portability and disposability being emphasized during the development of electrochemical techniques in this area. Although nitro-based explosives are the most popular target analyte in such work, PBEs are also receiving increased attention. For example, significant contributions to the field have been made in the use of Prussian blue (PB) modified glassy carbon disk electrodes. These electrodes were used to detect H2O2 that was generated from UV or laser treatment of TATP and HMTD [68]. The preferential electrocatalytic activity of PB toward H2O2 has led to PB being called an “artificial enzyme peroxidase” [25, 68, 69]. For TATP that was treated with a short-burst laser, an LOD of 50 nM was observed by this electrochemical method. When UV irradiation was used, the LOD for TATP was 250 nM and the LOD was 300 nM for HMTD. Researchers next monitored H2O2 produced from acid treatment of TATP with and without neutralization steps [70]. An LOD of 55 nM for TATP was observed when an acidic TATP solution was neutralized prior to amperometric measurements of stirred solutions. In the same report, a simplified experimental design was presented based on the fabrication of single-use PB-modified screen-printed electrodes and elimination of the neutralization step. An LOD of 18 mM was achieved using this one-step method. This higher LOD is likely the result of both the elimination of the neutralization step and the direct chronoamperometric monitoring of a nonstirred reaction solution. This approach with screen-printed electrodes required low reaction volumes (approximately 20 μL) and a 1-min assay time. HMTD was tested as well, but results for this analyte were not provided.

Acid treatment of TATP was also used in an electrochemical method introduced by Laine et al. [16]. This approach was based on the reactions in Eqs. 3 and 4. Iron(II) ethylenediaminetetraacetate (EDTA) was produced at a glassy carbon electrode by the reduction of iron(III) EDTA, as given in Eq. 3. Next, iron(II) EDTA reduced H2O2 and/or hydroperoxides that had been released by acid treatment of TATP, as shown in Eq. 4.
$$ F{e^{III}}ED TA\; + \;{e^{-} } \to F{e^{II}}EDTA $$
$$ F{e^{II}}EDTA~ + ~{H_2}{O_2}/ROOH \to F{e^{III}}EDTA~ + ~R{O^{-} }/H{O^{-} } + ~HO \bullet $$
An LOD of 890 nM for TATP was achieved in this technique. Later work with HMTD [71] indicated a separate acid digest step was not required. HMTD added to a pH 2.1 iron(II)/iron(III) EDTA solution spontaneously hydrolyzed to form simpler peroxides, including H2O2, and provided a similar sensor response (see Fig. 7). A slightly higher LOD of 30 μM was seen for HMTD in this modified method.
Fig. 7

Electrochemical detection of TATP and HMTD using an iron(II)/iron(III) ethylenediaminetetraacetate (EDTA) reaction [16, 71]. Chronoamperograms of a acid-treated TATP in 1 mM iron(III) EDTA and b increasing concentrations of HMTD added to a pH 2.1 iron(II)/iron(III) EDTA solution. Chronoamperograms were obtained by stepping to −400 mV (vs. Ag/AgCl). (Courtesy of F. Cheng)

The ability of H2O2 to induce current changes in phthalocyanine (Pc) p-type semiconductors was employed in a method designed by Bohrer et al. [72]. In this report, 50-nm-thick films were made of Pcs, both metallated and metal-free, forming chemiresistors for use as H2O2 vapor sensors. A host of MPcs (where M is Co, Ni, Cu, or H2) were tested, with H2O2 causing current losses in CoPc and current gains in NiPc, CuPc, and H2Pc. Other strong oxidants all caused current gains in all MPcs; only H2O2 showed a differential response. This was the first example of contrasting analyte redox behavior dependent on the metal center in the chemiresistor. Using all or just a combination of MPcs with opposite responses (e.g., CoPc and CuPc) gave a catalytic redox sensor array for the selective detection of H2O2. It was suggested that an MPc sensor array could be used to detect PBEs after conversion to H2O2 using UV irradiation. The maximum response time was 10 min for all MPcs and the current response was constant even when changes in humidity occurred. The LOD depended on the MPc that was tested. CoPc, the most potent catalyst for H2O2 redox, had an LOD of 50 ppb. For NiPc, CuPc, and H2Pc, the LODs were 40.1, 12.2, and 11.7 ppm, respectively.

Other methods

A variety of techniques making use of high-performance liquid chromatography (HPLC) have been developed for the detection of PBEs. Both HPLC-MS and HPLC with electrochemical detection have used to monitor TATP, DADP, and HMTD [8]. For the detection of TATP and HMTD, Schulte-Ladbeck et al. [73] developed a reversed-phase HPLC method with online IR detection using a CaF2 flow cell. TATP and HMTD were well resolved in this approach and gave LODs of 1 mM for TATP and 0.5 mM for HMTD. Spiked soil samples gave similar results.

GC is another method that is often used in PBE detection and characterization. A commercially available handheld GC device called zNose (Electronic Sensor Technology, Newbury Park, CA, USA) has been studied for the detection of vapors from explosives [74]. This “electronic nose” contains a solid-state sensor that provides LODs in the low parts per trillion range for a variety of explosives, including TATP. For possible future integration into a portable explosives detector, Zuck et al. [75] recently fabricated a microcalorimetry device. This differential scanning device was used to analyze 30–100-μm-size explosives particles in addition to nonexplosive material such as sugar and sea sand. The thermograms obtained were sufficiently unique to allow for differential detection. An LOD for TATP was not provided in this report, but the authors stated that work on the creation of a portable unit is ongoing.


The continued and increased use of PBEs in terrorist activities has made the development of detection methods for these explosives a research priority. In a previous review, method requirements for such work were outlined and it was noted that a variety of techniques would be needed to meet the desired goals of unambiguous identification, portability, easy operation, minimal analysis times, and low LODs in a variety of sample matrices [8]. As can be seen in this current review of recent developments in PBE detection, progress has been made to meet these goals, through the use of a variety of new assays and variations of more established methods. These methods have included techniques based on luminescence and fluorescence measurements, IR or Raman spectroscopy, MS, electrochemical methods, and separation techniques such as HPLC and GC.

Several trends have emerged since the previous review. First, there has been an emphasis on field measurements. This work has included methods involving commercially available portable instruments and the design of devices or assays that have the promise of being portable, such as a QCL handheld sensor [41], a QCL-based walk-through portal [17, 44], and luminescence techniques based on PVA/LG films [27] or MoHB nanoparticles [28] for use in test strips or badges. Clearly, given the high-security and high-traffic areas in which PBE detection is used, continued advancements in portability are still needed. Another trend has been an increase in the use of IR spectroscopy and MS for PBE analysis. For instance, a number of researchers have identified IR regions that can be used to identify PBEs or have elucidated fragmentation pathways for PBE detection by GC-MS [60], ESI MS [53, 54], or DESI MS [6163]. This type of work should allow analysts at forensic facilities to more easily integrate new methods, as they can employ their qualified instruments. TATP still appears to be the focus in the development of many of these methods, although HMTD has received increased attention in the last few years.


This work was supported by DARPA grant N66001-08-1-2.

Copyright information

© Springer-Verlag 2009