Ultrafast Gas Chromatography-Tandem Differential Mobility Spectrometry: Toward A New Generation of On-Site, Real-Time Trace-Explosives Detection

Abstract

In the defense and security sector, rapid detection of trace quantities of threat materials is paramount. Traditional instrumentation typically relies on standalone ion mobility techniques due to being inexpensive, portable, and highly sensitive. However, these techniques face limitations when handling complex samples, suffering from low resolving power (often less than 100) and ion-suppression effects, which can lead to false-positive and false-negative results. Here, we present a foundation to the solution through the hyphenation of the flow field thermal gradient gas chromatograph (FF-TG-GC) developed by HyperChrom with a tandem differential ion mobility spectrometer (DMS-DMS) developed in-house at New Mexico State University. The FF-TG-GC demonstrates the ability to separate a variety of nitroaromatic compounds of explosive significance in 20 s using a nitrogen carrier gas, highlighting the potential to offer selectivity advantages without substantially compromising high-throughput demands. These selectivity advantages are illustrated by the successful application of the FF-TG-GC-DMS-DMS to the detection and identification of single-nanogram loadings of 18 explosives and related substances in the presence of interfering materials, such as lactic acid, musk, and diesel. Furthermore, the system is capable of mitigating in-source ion-suppression effects by chromatographic separation of target analytes from background interference prior to ionization.

The defense and security sectors apply a variety of analytical techniques to the detection and identification of threat materials, such as explosives and chemical warfare agents. Ion mobility techniques are especially prominent in explosive trace detectors (ETDs) for situations where on-site, high-throughput, and sensitive screening for explosives is required¹ due to meeting a number of desirable characteristics of trace-explosive detection systems.^2,3 For example, ion mobility techniques can exploit selective ionization chemistry of explosive compounds,⁴⁻⁶ produce results within seconds of sample introduction, and are small enough to be made portable,^7,8 all while being robust, sensitive, and cost-efficient.⁹

However, ion mobility techniques can encounter obstacles when employed for direct determinations of explosives, particularly with the first step of detection: ionization. Adverse ionization chemistry occurs with complex samples, such as matrix-dependent ion–molecule reactions,¹⁰ which can lead to ineffective ionization of analytes, ultimately prohibiting detection. Even if ionization transpires effectively, explosives generate only a small number of ionized species.^4,10,11 The limited number of ionized species used for determinations of an explosive combined with the relatively low resolving power of ion mobility systems⁹ can result in a benign compound being mistaken for that of an explosive.^12,13 Furthermore, unfavorable changes in mobility properties of ions can cause false-negative results by moving target species outside of the mobility criteria used to identify an explosive compound. For example, changes in ion mobility properties can occur due to a high moisture content^14,15 or from formation of adducts in the ion source.¹⁶ Consequently, these results can lead to low confidence in identification and detection ability by the equipment operators.

There have been a variety of improvements to ion mobility technology to mitigate the risk of false-positive and false-negative results occurring. For example, tandem ion mobility¹⁷ with fragmentation capabilities¹⁸⁻²² have moved the technique toward improved signal-to-noise ratios and specificity through multiple dimensions of ion processing. Hyphenation of ion mobility with mass spectrometry¹² can also provide clear compositional information on the nature of an ion generated and separated by ion mobility. Additionally, employment of atmospheric modifiers^10,23 and the continued push toward higher resolution ion mobility spectrometers enable measurements to be taken with improved separation between ions leading to a reduced potential to misidentify compounds. All of these developments effectively provide solutions to recurring problems of analysis of explosives by ion mobility through increasing the selectivity and reducing the obfuscating effects of chemical noise, ultimately offering increased resilience against false-positive results compared to single-stage ion mobility techniques. However, these techniques can still suffer from in-source ion-suppression effects.²⁴

An additional method to improve system selectivity is the inclusion of gas chromatographic (GC) separation prior to the ion mobility analyzer^25,26 where target analytes are separated from the other constituents (i.e., the chemical background) of a sample prior to ionization. Such separation prevents all components of a sample entering the ion source simultaneously, instead enabling the components to be collected as fractions and ionized individually reducing competition within the source and mitigating the adverse effects of sample complexity on ionization effectiveness. Accordingly, GC is a powerful complementary technique for detection modes that require sample ionization in the gas phase. Despite the additional selectivity advantages generated by GC, the technique has not been widely adopted in mainstream, high-throughput detection systems for threat materials as analytical turnaround times of chromatographic-based methods are high. Even for methods that employ “fast GC”, the runtimes are still minutes long,^25,27 which makes the technique difficult to reconcile with any high-throughput demand from the defense and security sector, especially when standalone ion mobility system are capable of meeting this demand with subminute turnaround times.

Recent advances in gas chromatography by Boeker and Leppert with the development of the flow field thermal gradient gas chromatograph (FF-TG-GC)²⁸ have replicated conventional GC analyses in a fraction of the time with little loss of chromatographic resolution. Applying the FF-TG-GC to laboratory-based explosives analysis, Leppert et al. demonstrated the potential to analyze comprehensive mixtures of explosive compounds and related substances in under 1 min.²⁹ The FF-TG-GC is able to achieve this feat by combining a 2 m column with a negative temperature gradient from inlet to outlet—the short column enables rapid elution of analytes introduced into the system while the negative temperature gradient reduces diffusive band broadening processes of analyte fractions,²⁸ enabling the instrument to achieve high-resolution separations. Given this cutting-edge development in true high-speed, high-resolution gas chromatography with the FF-TG-GC, the technique is highly suited to complementing advancements in ion mobility technology for screening of threat materials while maintaining high-throughput options.

Accordingly, we present here the coupling of the FF-TG-GC to a tandem differential ion mobility spectrometer (DMS-DMS) capable of field-induced fragmentation (FIF)¹⁹ to illustrate the impact these advances can have on the field through three avenues of investigation. First, high-throughput FF-TG-GC methods were evaluated for high-speed sample turnaround without long chromatographic separation periods. Further to this objective, the future prospects of designing a portable system are considered by mitigating the requirement for gas cylinders to support operation by using nitrogen as the GC carrier gas. Nitrogen is inert, well-established in GC operation and can be produced by a generator—desirable characteristics for a portable system. Second, the ability of the DMS-DMS to provide additional system selectivity was assessed using three aspects of the technique: ion separation, fragmentation, and tandem operation. Finally, the potential of this combination for the mitigation of false-positive and false-negative results was assessed by application of the FF-TG-GC-DMS-DMS to the detection and identification of a comprehensive mixture of explosives in the presence of representative interfering sample sets: diesel, Javon Musk, and lactic acid.

Experimental Section

FF-TG-GC-FID

Chromatographic method development involved a FF-TG-GC (HyperChrom SA) coupled to the built-in flame ionization detector (FID). The FF-TG-GC employed a 2.2 m × 0.1 mm × 0.1 μm Rxi-5Sil MS column (Restek Corp.). The split/splitless inlet was operated with a 20:1 split ratio and maintained at 180 °C. A Topaz 4 mm ID single-taper liner with deactivated glass wool (Restek Corp.) was used. The carrier gas was nitrogen with a flow rate of 0.25 mL/min. A 0.4 m × 0.1 mm methyl deactivated capillary (Trajan) was used as a transfer line between the analytical column and FID by employing the FF-TG-GC’s second purged connector (Figure S1) using a Siltek-coated (Restek Corp.) SilTite Mini Union connector (Trajan). The transfer line was maintained at 180 °C, while the second purged connector was operated at 205 °C with an additional flow of 0.2 mL/min. The FID was operated at 250 °C with hydrogen flow of 30 mL/min, nitrogen flow of 30 mL/min, and air makeup flow of 300 mL/min. The temperature program used is specified in Table 1. Chromatograms collected on the FF-TG-GC-FID where compared to chromatograms collected using a GC-time-of-flight mass spectrometer with the same stationary phase. Elution order between the two systems were compared to confirm the identity of each peak present in the FF-TG-GC-FID chromatogram.

Table 1. FF-TG-GC-FID Temperature Program.

time point (s)	temperature (°C)	ramp rate (°C/s)	differential (°C)
0	75	5	40
3	90	10	45
7	130	25	65
11	230	15	110
15	290	6	145
20	320	0	160
25	320		165

FF-TG-GC-DMS-DMS

Following the chromatographic method development, the FID was replaced with an in-house developed DMS-DMS¹⁹ (Figure S3) operated with a body temperature of 60 °C and a makeup flow of 1 L/min at 80 °C. The methyl deactivated transfer line was replaced with an aluminum-coated 0.19 m × 0.22 mm × 0.1 μm HT-5(AQ) column (SGE). The temperature program used is detailed in Table 2. All other FF-TG-GC parameters remained unchanged from the FF-TG-GC-FID setup.

Table 2. FF-TG-GC-DMS-DMS Temperature Program.

time point (s)	temperature (°C)	ramp rate (°C/s)	differential (°C)
0	60	0	25
10	60	1.75	25
130	270	0	150
140	270		150

Chemicals and Reagents

Certified reference standards of 2,4,6-trinitrotoluene (TNT), 1,3,5-trinitroperhydro-1,3,5-triazine (RDX), and the 8330B explosives mixture³⁰ were purchased in acetonitrile at 1000 ng/μL (Restek Corp.). Nitroglycerine (NG), ethylene glycol dinitrate (EGDN), and pentaerythritol tetranitrate (PETN) were purchased in methanol at 1000 ng/μL (Restek Corp.). Triacetone triperoxide (TATP) was purchased at 100 ng/μL in acetonitrile (AccuStandard). Diesel was purchased at 5000 ng/μL in dichloromethane (DCM, Restek Corp.). Javon Musk was purchased locally and 85% lactic acid in water was retrieved from internal stocks. All preparations and dilutions were performed in analytical grade methanol using Eppendorf mechanical pipettes, with samples manually injected into the FF-TG-GC at 0.5 μL using a 1 μL syringe (26 g/70 mm/cone tip, Agilent Technologies).

The 8330B explosives mixture contains nitrobenzene (NB), 2-nitrotoluene (NT), 3-NT, 4-NT, NG, 1,3-dinitrobenzene (DNB), 2,6-dinitrotoluene (DNT), 2,4-DNT, 1,3,5-tinitrobenzene (TNB), TNT, PETN, RDX, 4-amino-2,6-dinitrotoluene (4-Am-2,6,-DNT), 3,5-dinitroaniline (DNA), 2-Am-4,6-DNT, tetryl, and octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX).

DMS Atmospheric Modification

Dichloromethane was introduced into the makeup gas using a syringe pump and a 500 mL Hamilton syringe pumping the liquid into a heated chamber, vaporizing the DCM into the DMS gas flow. The flow rates used were 0.01, 0.10, 0.20, 0.30, and 0.50 mL/h, corresponding to 170, 1720, 3440, 5150, and 8590 ppm mass in air, respectively. Each flow rate change was left to equilibrate for 30 min before measurements were taken.

Single-Stage DMS Operation

The first DMS analyzer (DMS1) and the fragmentation unit were both inactive, while the second DMS analyzer (DMS2) was operated in a CV-scanning mode with SV varied between 63 and 157 Td depending on the experiment (Table S1).

Field-Induced Fragmentation

Fragmentation was achieved using a custom built modular integrated power supply unit, supplied by GAA Custom Electronics, operating between 157 and 314 Td. DMS1 remained inactive while DMS2 was operated in CV-scanning mode with SV varied between 63 and 157 Td depending on the experiment (Table S1).

Tandem DMS with Fragmentation

DMS1 was operated in ion-isolation mode at constant SV of 94 Td and CV of 0 Td. DMS2 was operated in CV-scanning mode with SV varied between 63 and 157 Td depending on the experiment (Table S1).

Performance Assessment

The interfering materials used for this study were diesel, Javon Musk, and 85% lactic acid in water. 1.25 ng mass loadings of target analytes were simultaneously analyzed with interfering materials at mass loadings of 10, 25, and 50 ng for lactic acid; 50, 125, and 250 ng for diesel and 50, 100, 200, and 500 ng for Javon Musk.

Results and Discussion

FF-TG-GC Coupling to DMS-DMS

Optimization of the temperature gradient (Table 1) enabled the FF-TG-GC-FID to achieve separation among a variety of nitroaromatic compounds in 20 s (Figure 1). Although there is no apparent resolution between DNB and 2,6-DNT at 16.6 s and TNB and TNT at 17.7 s, further separation of these compounds can be achieved using the DMS-DMS. These compounds have different ion mobility properties so that they are measured at different CVs over the ion mobility axis (see Atmospheric Modification and Separation Voltages). Accordingly, the FF-TG-GC initially presents itself as a favorable candidate for supporting detection of a number of explosives by ion mobility techniques; two compounds are unlikely to share both the same vaporization and ion mobility properties, leading to improved selectivity without severely impeding analysis times.

Figure 1.

Chromatogram collected during the analysis of (1) NB, (2) 2-NT, (3) 3-NT, (4) 4-NT, (5) DNB/2,6-DNT, (6) 2,4-DNT, (7) TNB/TNT, (8) Am-2,6-DNT, (9) DNA, (10) Am-2,4-DNT, and (11) tetryl by FF-TG-GC-FID using the temperature program in Table 1. The FID response is normalized against the largest peak.

Following assessment of the chromatographic speeds using the FID as a detector, the FF-TG-GC was coupled to the DMS-DMS. However, the use of polyimide-coated columns with ion mobility techniques is suboptimal when sustained elevated temperatures are involved. Heating of polyimide coatings results in off-gassing of highly electronegative imide-based compounds. These compounds can interfere with the ionization chemistry of target compounds and add an increased chemical background to the detector. Both events can inhibit the ability of ion mobility spectrometers to detect subnanogram quantities of target compounds. Consequently, if transfer lines are maintained at 100 °C or greater, then coupling of equipment to the DMS-DMS requires the use of a metal-coated column as off-gassing of imide-based compounds cannot occur.

The column available for this work was an aluminum-coated HT-5(AQ) column with an internal diameter of 0.22 mm and a film thickness of 0.1 μm. However, this column induces suboptimal chromatographic conditions due to the presence of a stationary phase and the GC-DMS transfer line being held at 180 °C. The combination of a stationary phase and isothermal conditions causes additional peak broadening in the analytes that have eluted from a temperature-programmed section of a chromatographic system, especially in low-volatility analytes like RDX and tetryl. At a length of 19 cm, the FF-TG-GC-DMS-DMS transfer line composes approximately 9% of the total chromatographic separating region, causing the peak broadening to be detrimental to the quality of the peak resolution.

Additionally, ions generated in the ion source can adsorb to the walls of the reaction region leading to further peak tailing as ions exit the source into the drift tube. The requirement to optimize sample gas flows and source geometry has been highlighted³¹ to not only improve sensitivity but to also quickly remove ions from the ionization region. This process enables reagent ions to quickly replenish, resulting in the signals generated from target compounds not overlapping when eluting close to one another. The combination of suboptimal transfer line conditions and a nonoptimized ion-transfer process for ultrafast GC results in an unsuitable interface between the FF-TG-GC and DMS-DMS, the consequence of which is displayed in Figure S4—low-volatility analytes begin to become indistinguishable.

Consequently, the 20 s runtime achieved with the FF-TG-GC-FID could not be replicated with the FF-TG-GC-DMS-DMS to the same degree of peak resolution during the period of this work. Instead, a 2 min method (Table 2) was adopted so that the necessary chromatographic separation of analytes could be achieved with the current interface and enable selectivity advantages of the system to be assessed.

However, despite the hardware limitations initially encountered, Figure 1 illustrates the potential for the FF-TG-GC to be used for rapid, additional separation for analysis of explosives. Future investigation should attempt to identify, develop, and implement hardware to facilitate effective coupling of the FF-TG-GC with an ion mobility spectrometer to accommodate rapid chromatographic elution.

Atmospheric Modification and Separation Voltages

Figure 1 highlights the potential to chromatographically separate a range of nitroaromatic compounds in 20 s with the exception of TNB with TNT and DNB with 2,6-DNT. As a separation technique, DMS can provide additional resolution between chromatographically unresolved signals, albeit with a modification to the DMS atmosphere; Eiceman et al. observed that in purified air the differential ion mobility spectra for the ions of nitroaromatic compounds are restricted to limited CV ranges.²³ However, the addition of DCM into the drift gas at 1000 ppm provides additional separation over the differential ion mobility drift region for all analytes investigated, including baseline-resolution between DNT and DNB.²³

Following the addition of dichloromethane at a constant SV of 110 Td, the same separation trends as seen for DNT and DNB by Eiceman et al.²³ could be observed for TNT and TNB (Figure 2). Distinction between the TNT and TNB peaks can be achieved at an addition of 0.1 mL/h of DCM into the atmosphere, but further resolution between the two compounds can be achieved with greater quantities of DCM. At 0.5 mL/h of DCM, TNT remains relatively unchanged at a CV of approximately 0 Td, whereas TNB moves from 0 to −0.6 Td, resulting in two resolved peaks attributed to each of the compounds. Accordingly, a DMS configuration that uses a DCM-modified atmosphere is an effective means of improving discrimination between compounds, if they remain unresolved after application of high SVs.

Figure 2.

DMS spectra showing how TNT and TNB peaks separate over the ion mobility axis with increasing DCM content at (A) 0.01, (B) 0.10, and (C) 0.50 mL/h at a constant SV of 110 Td. DMS response is normalized against the largest peak in each plot.

Further to providing improved separation between individual compounds, and also consistent with the observations of Eiceman et al.,²³ the addition of DCM also provides improved separation of the ions attributed to single compounds. For example, Figure S5 highlights that the chloride-doped molecular ion, nitrated ion, and dimers of specific compounds can be separated in the ion mobility dimension with DCM unlike when the compounds are analyzed in purified air. Consequently, this separation of ions enables identification of specific compounds to not rely on the detection of just a single peak. Instead, identification can be achieved using the detection of multiple peaks enabling a more robust process that is less easily mistaken with other single ions that may be detected at similar SV and CV values.

With an atmosphere modified with DCM and application of differing SV values, the DMS can provide additional selectivity advantages to chromatographic analyses in which some compounds are unresolved. Additionally, the ion mobility spectrometer can result in multiple separated ions attributable to a single compound increasing confidence in identification compared to identification from a single point.

Field-Induced Fragmentation

In the absence of separation between compounds analyzed by GC-DMS, methods to generate additional selectivity are necessary to resolve the identity of the compounds. This approach can be achieved through fragmentation of ions into smaller charged-species. There are a number of properties that can contribute to the process of fragmentation: thermal energy, field energy and ion-neutral collisions,^32,33 with the DMS-DMS employed in this study primarily focusing on field-induced fragmentation.

By activating the fragmentation unit, RDX begins to decompose into NO₂^– ions which appears at a CV of −1 Td with a SV of 110 Td. Additionally, NG produces the NO₃^– ion at approximately −0.7 Td with a SV of 110 Td, with the same process occurring for other nitrate esters like EGDN and PETN. NG also produces another fragment, which is tentatively assigned as the NO₂^– ion due to appearing at the same CV/SV value as the NO₂^– ion from RDX. FIF also provides fragments for ions in the positive polarity channel with DNB fragmenting at 314 Td in a purified air atmosphere. However, the identity of the fragment was not established during this study (Figure 3).

Figure 3.

DMS spectra showing fragmentation of (A) NG at 220 Td resulting in NO₃^– (fragment 1) and NO₂^– (fragment 2), (B) RDX at 251 Td resulting in NO₂^– and (C) DNB at 314 Td in the positive channel when using a purified air atmosphere. DMS response has been normalized against the largest peak in each plot.

Not all ions fragment at 314 Td at 60 °C, though. For example, the in-source-generated ion [TNT-H]⁻ has been recorded to require temperatures as high as 700 °C to facilitate thermally induced fragmentation at atmospheric pressure.³⁴ However, the instrumentation used for this investigation is not able to accommodate such active fragmentation conditions with experimentation at the upper limit of the fragmentation unit (314 Td), resulting in no observable fragments attributable to TNT. Despite the apparent lack of TNT-selectivity afforded by the DMS and fragmentor, Jurado-Campos et al.²⁴ demonstrated an approach utilizing FIF to decompose the ion of an interfering material without compromising the signal generated by the [TNT-H]⁻ ion, ultimately enabling TNT to be identified in a chemically contested environment.

FIF may be a powerful addition to GC-DMS analysis when analytes and/or interfering materials are not resolved over the chromatographic or ion mobility separation modes. By combining both approaches—fragmentation of the target species or fragmentation of the interfering materials—selectivity of the system and confidence in a measurement can be improved from additional chemical information.³⁵

Tandem DMS Operation

During fragmentation of target analytes, common ions produced were NO₃^– and NO₂^–. However, the presence of these fragments could become obscured by the reactive ion peak (RIP). In a purified air atmosphere, the NO₃^– ion convolves with the O₂^– RIP (Figure 4A) and with a chloride-dopant present in the atmosphere, the NO₂^– ion convolves with the Cl^– RIP (Figure S6). In either instance, important diagnostic chemical information would be obscured by the respective RIPs, resulting in reduced identification potential.

Figure 4.

Topographic plots of FF-TG-GC-DMS-DMS spectra highlighting that (A) O₂^– RIP has similar mobility to NO₃^– fragments. When (B) DMS1 is operated in ion-isolation mode the O₂^– RIP is removed allowing NO₃^– fragments to be clearly resolved.

However, the concealment of the fragments could be mitigated by employing the tandem mode of the DMS. For example, Figure 4 highlights how the NO₃^– ions generated by nitrate esters could be observed in a purified air atmosphere by operating the DMS1 in an ion-isolation mode at an SV of 94 Td and CV of 0 Td. This operation prevents the O₂^– RIP from passing through to DMS2 while enabling mobility-selected ions to continue. Consequently, any resulting fragments cannot be obscured by the RIP allowing additional characteristic chemical information detected in the DMS2 to be used for identification purposes.

Performance Assessment

The 8330B explosive mixture, EGDN, and TATP were analyzed alongside samples of diesel, Javon Musk, and 85% lactic acid in water at various mass loadings to initially probe the effectiveness of the system’s selectivity-suite against the presence of interfering compounds. This allowed both the complexity of data interpretation and the magnitude of any suppression effects to be investigated.

Lactic Acid

Lactic acid has previously been established to cause interference in ion mobility systems for the detection and identification of explosive compounds.¹⁶ These outcomes occur through in-source reactions resulting in impaired detection limits for a number of explosives and affecting the qualitative compositions. Accordingly, lactic acid (85% in water) presents itself as a suitable initial benchmark to examine the robustness of the FF-TG-GC-DMS-DMS.

Due to achieving suitable chromatographic separation, Figure 5A illustrates how the presence of lactic acid and water did not inhibit the identification of any compounds analyzed during this study. No modifications had to be made to the DMS to provide further selectivity beyond what the FF-TG-GC alone contributes, with the system operating in a single-stage mode with a purified air atmosphere at a constant SV of 110 Td. Additionally, lactic acid and water only appear in the negative channel when analyzed by the DMS system; these compounds cannot cause false-positive results for analytes observed in the positive channel.

Figure 5.

Topographic plots obtained via FF-TG-GC-DMS-DMS of (A) the 8330B mixture and EGDN (1.25 ng each) analyzed with 85% lactic acid in water (50 ng), (B) NB, 2-NT, 3-NT, and 4-NT (1.25 ng each) analyzed in diesel (250 ng), and (C) TATP (1.25 ng) analyzed in diesel (125 ng).

As a result of separating the interferents away from the target analytes prior to ionization, in-source ion-suppression effects from a known interferent can been mitigated. Figure S7 (top) displays the results of 1.25 ng loadings of explosives analyzed with 0, 10, 25, and 50 ng of 85% lactic acid in water. Within the mass range tested, the interferent does not cause the peak intensity of target analytes to diminish at all; if ion-suppression effects were present during analysis, there would be a decrease in peak intensity for some or all analytes as the mass of lactic acid increased.

Diesel

Highly complex mixtures, like diesel, often make data interpretation and specific compound identification an arduous and involved task with any analytical technique. Accordingly, diesel is an example of a substance that could increase the possibility of false-positive and false-negative results by producing uninterpretable plots, so is a suitable benchmark for identifying compounds among complex mixtures.

Components from diesel appear only in the positive-ion channel as a continuous set of peaks over the chromatographic time scale. Accordingly, analytes in the negative channel can be observed without hindrance. However, TATP, NB, 2-NT, 3-NT, and 4-NT are recorded in the positive channel leading to difficulties in interpreting the chromatogram when analyzed in diesel. When combined with the DMS, though, the system is able to separate NB, 2-NT, 3-NT, 4-NT (Figure 5B) and TATP (Figure 5C) away from the diesel signals. In both instances, the DMS-DMS was operated in tandem mode without fragmentation enabled in a purified air atmosphere; DMS1 was operated at SV 94 Td and constant CV of 0 Td with DMS2 operated at an SV of 157 Td for the nitroaromatic compounds and an SV of 126 Td for TATP.

Given diesel’s complexity, it is unlikely to be entirely removed from the ionization source at the same time one of the target analytes is being ionized. To identify the impact this may have on ion-suppression effects, diesel was injected at 0, 50, 125, and 250 ng/μL simultaneously to the 8330B mixture and TATP at 1.25 ng mass loadings. Figure S7 (bottom) shows that most analytes were generally resistant to ion-suppression effects, likely as a result of being negative-channel compounds. However, TATP shows an exception to the trend with a slight decrease in peak intensity from 0.7 to 0.6 au with increasing diesel mass loadings. Despite this decrease, TATP can still be observed at 1.25 ng compared to the greatest mass (250 ng) of diesel injected simultaneously, which is likely due to the chromatography—the majority of the components from diesel have been separated from the elution region of TATP, leading to competition of only a small number of remaining components left in the source.

Javon Musk

Musk is an interfering material that can be commonly encountered by defense and security personnel when screening for threat materials on people. Consequently, the substance acts as another useful test to initially probe the selectivity and robustness of the system.

The majority of the components in Javon Musk only appear in the negative channel of the DMS and do not impede identification of any positive-channel analytes. Furthermore, in the negative channel, Figure 6A highlights that all compounds can be resolved over the chromatographic separation period without significant DMS-DMS method development. However, EGDN and some components of Javon Musk do elute very close to one another, leading to the possibility of false outcomes. For example, in the absence of EGDN, Javon Musk could result in a false alarm for EGDN through minor variations in chromatographic retention time, raising a requirement to develop the methodology further to reduce this likelihood.

Figure 6.

Topographic plots obtained via FF-TG-GC-DMS-DMS (A) of the 8330B mixture and EGDN (1.25 ng each) analyzed with Javon Musk (500 ng) without DMS optimization, (B) with FIF enabled to generate NO₃^– fragments, which are mobility-separated from Javon Musk and (C) a magnified view of the NO₃^– fragments.

One mechanism to reduce the potential for this misidentification can be found in FIF. Fragmentation of EGDN (and other nitrate esters) results in the formation of the NO₃^– ion, which is baseline resolved from the parent ion and ions generated by Javon Musk. This is demonstrated in Figure 6B, in which the DMS-DMS fragmentation cell is operated at 210 Td while DMS1 is operating in an ion-isolation mode (SV = 94 Td, CV = 0 Td). Under this setup, the O₂^– RIP is removed via DMS1 and the fragmentation of EGDN generates NO₃^– ions, which are observed in DMS2 operating in a CV-scanning mode with an SV of 94 Td. Together, the chromatography, ion separation, fragmentation, and tandem DMS operation can produce a selective methodology to enable detection and identification of explosive compounds in Javon Musk and increase confidence in identification of analytes among abstruse data sets.

Conclusions

The work reported here illustrates the impact that recent advances in gas chromatography and ion mobility spectrometry can have on the field of explosives detection; together, the FF-TG-GC and the DMS-DMS can provide rapid separation and high selectivity through the use of gas chromatography, modified drift gas atmospheres, field-induced fragmentation, and multiple stages of ion mobility. The system successfully demonstrates the increased analytical performance by detecting key explosive compounds in samples of diesel, musk, and lactic acid. Furthermore, the FF-TG-GC creates potential to deliver additional selectivity gains though chromatographic separation without unsuitably long analysis times that a GC would typically introduce.

Further, to the initial probe into increased analytical performance, this investigation highlights a number of additional opportunities for research to move toward a deployable and operational system. First, equipment interfacing must be explored; coupling of the FF-TG-GC with the DMS-DMS highlighted the difficulties in maintaining fast chromatographic methodology with optimized interfacing-hardware required to facilitate high chromatographic resolution. Second, sample introduction during this work was achieved through liquid injection, which is an impractical method for field deployment. A more suitable method for sample introduction is swab desorption due to it being straightforward and effective. So, development, implementation, and optimization of a swab desorption unit for the FF-TG-GC should be perused. Third, the FF-TG-GC is currently designed as a laboratory-based benchtop instrument with limited options for portability. Effort to reduce the size of the system and dependence on laboratory infrastructure should be explored. Finally, an all-in-one system simultaneously employing all forms of available selectivity should be investigated; currently, the FF-TG-GC-DMS-DMS is optimized between experiments to suit specific requirements where, in practice, the specific chemical background cannot be anticipated nor should operators be expected to optimize methodology themselves. Once these steps have been taken, and given the well-established analytical techniques employed in this study, there is scope to apply this system to a wide range of trace detection activities, including screening for controlled drugs and chemical warfare agents.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.analchem.4c01464.

• Seven additional figures (Figures S1–S7) and one additional table (Table S1) providing clarity on method details, further experimental observations, and technical challenges (PDF)

Author Contributions

J.A.T. performed experimental design, writing, laboratory measurements, and data analysis. P.E.F. performed laboratory measurements, writing, and data analysis. G.A.E. performed experimental design and writing.

The authors declare no competing financial interest.