- Email: [email protected]
The analysis of nitrate explosive vapour samples using Lab-on-a-chip instrumentation
Journal of Chromatography A, 1602 (2019) 467–473
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
The analysis of nitrate explosive vapour samples using Lab-on-a-chip instrumentation Vitor Taranto a,∗ , Maiken Ueland a , Shari L. Forbes b , Lucas Blanes c a b c
Centre for Forensic Science, University of Technology Sydney, Australia Département de chimie, biochimie et physique, Université du Québec à Trois-Rivières, Canada Oswaldo Cruz Foundation (FIOCRUZ), Curitiba PR, Brazil
a r t i c l e
i n f o
Article history: Received 6 January 2019 Received in revised form 29 May 2019 Accepted 2 June 2019 Available online 3 June 2019 Keywords: Explosive detection Portable instrumentation Microchip electrophoresis
a b s t r a c t The detection and analysis of explosives and explosive-related compounds is a heightened priority in recent years for homeland security and counter-terrorism applications. This study aimed to evaluate the use of a commercial Lab-On-a-Chip (LOC) instrument for the analysis of explosive vapours, with the longterm goal of developing a portable instrument for passively detecting explosives in air samples. A simple method to collect explosive vapour residues was developed using a glass vial containing varying amounts of the target explosives (1 mg/mL). Standards were diluted to the desired concentration in 150 L of acetone to facilitate the evaporation. The top of the vial was covered with a circular 0.5 cm diameter filter paper and exposed to a range of temperatures from 22 ◦ C to 80 ◦ C for 15 min. Following evaporation, the filter paper chads were folded and inserted into the LOC wells containing the separation buffer for the analysis, avoiding any further extraction step. After successfully separating and detecting eight explosives via liquid analysis, three explosives were chosen as targets for the vapour analysis experiments. 1,3,5Trinitrobenzene (TNB), 2,4,6-Trinitrotoluene (TNT), and 2,4,6-Trinitrophenylmethylnitramine (Tetryl) were successfully separated, detected and identified following the vapour extraction of explosive standards onto filter paper chads. Limits of detection for the liquid analysis were demonstrated to be 2.32 ng for TNB, 2.35 ng for Tetryl, and 3.25 ng for TNT. The minimum detectable mass found for the vapour analysis was 6.03 for TNB, 9.99 ng for TNT, and 14.22 ng for Tetryl. The average recovery from the paper chads was 29% for Tetryl, 47% for TNB, and 75% for TNT (n = 4), comparable with findings from previous studies. Results show that a minimum temperature of 40 ◦ C is necessary to vaporize the compounds using acetone, while the best results were achieved when heating the vial to 80 ◦ C. The use of a filter paper to collect the explosives residues, avoiding any additional extraction step, and the ability to analyze these compounds using a LOC instrument, makes this approach a future alternative method for explosive residues detection in the headspace. © 2019 Elsevier B.V. All rights reserved.
1. Introduction An increasing focus on counter-terrorism and homeland security globally demands fast and reliable detection and analysis of explosives and related compounds [1]. The investigation of explosives detection in the headspace is important for the screening of large areas for commercial and improvised explosive devices, however explosives detection through air samples is notoriously difficult due to the low volatility of these compounds [2]. Nitroaro-
∗ Corresponding author at: Centre for Forensic Science, School of Mathematical and Physical Sciences, University of Technology Sydney, PO Box 123, Broadway NSW 2007, Australia. E-mail address: [email protected] (V. Taranto). https://doi.org/10.1016/j.chroma.2019.06.003 0021-9673/© 2019 Elsevier B.V. All rights reserved.
matic compounds are widely used in the chemical industry for the production of explosives [3], and their detection is of paramount importance for national security as modern military explosives are generally nitrogen-containing aromatic compounds [2,4]. Compounds such as 2,4,6-trinitrotoluene (TNT) and the more volatile and soluble 1,3-dinitrobenzene (DNB) are two of the most representative nitro-substituted aromatic explosive compounds [5]. The detection of explosive substances in the air is hampered by their low vapour pressures [6]. For example, Tetryl has a vapour pressure of 7.41 × 10−12 atm at room temperature (25 ◦ C), while nitroaromatics such as TNT and TNB have vapour pressures around 9.15 × 10-9 and 2.00 × 10-8 atm, respectively [7]. The diffusion of vapours from explosives is dependent on the temperature and airflow around the object; as the temperature increases, the vapour pressure also increases, so that some explo-
468
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
sives can yield a vapour pressure four times higher at 100 ◦ C than at 25 ◦ C [6]. The need for fast detection of explosives has generated a demand for rapid, sensitive and reliable methods, which could be made rugged and portable for field analysis [8–10]. Numerous analytical methods have been tested for detection of explosives and its breakdown products in air, water and soil [1,11]. Some instruments such as ion mobility spectrometry (IMS) and infrared (IR) detection based platforms have been widely researched for miniaturization and portability and are currently broadly used in airports, with proven parts per billion detection limits [1]. Other notable portable techniques developed for gas and vapour sampling are colorimetric and fluorescence quenching tests, chemiluminescence sensors, and mass spectrometry (MS) devices. All of these techniques have portable instrument representatives currently on the market for the detection of hazardous materials in gas and vapour samples. However it is important to note that IMS sampling requires contact with the surface of interest (i.e. swabbing) to achieve peak sensitivity [12], and when used for vapours and gas analysis needs more than one pre-concentration step to perform its analysis. IMS and MS detection systems require ionization of the analytes, which can be troublesome in a real time scenario due to competitive ionization pathways with field interferents [1,13]. Luminescence sensors and colorimetric tests may be described as utilising either direct or indirect detection methods, dependent on whether they use the fluorescence that the sample may emit or its effect on a fluorescent material, such as quenching [1]. Quenching techniques have already been considered in the literature as a subjective activity needing a confirmatory test with a better qualitative instrument and may also suffer from field interferents [14]. In general this tests are portable and easy to use but lack sensitivity, reliability and reproducibility [1]. Field deployable instruments must be small, cost efficient, rugged and be able to perform real time detection of hazardous materials among troublesome field activities [15]. Lab-on-a-chip (LOC) devices have promising features for the development of portable instruments such as the compact size and portability, quick analysis time, low cost, small sample and reagent consumption, and the possibility to integrate with other systems [16]. These devices result from the development of a technique know as microfluidics [17], which consists of pumps, valves, flow sensors, separation capillaries, and chemical detectors integrated on a single substrate or as compacted modules [18]. LOC devices are portable and capable of performing extremely fast, cost-effective separations [19], and represents an attractive alternative for the rapid analysis of forensic samples [20], including explosives [21,22]. The integration of capillary electrophoresis (CE) into LOC devices allows for the fast analysis of chemicals with low reagent consumption [15]. CE is an analytical technique used to separate electrically charged molecules based on their electrophoretic mobility. The separation occurs in a silica capillary tube connected between two buffer reservoirs by applying high voltage. The electric field produced by the high voltage separates the electrically charged analytes in the electrolyte while passing though the capillary. The separation is dependent on the difference in charge and friction caused by the size of the molecule inside the capillary [23,24]. Nitroaromatic compounds can be directly detected by ultraviolet–visible spectroscopy (UV–vis) [25]. This detection approach is widely used in conventional capillary electrophoretic analysis of nitroaromatic explosives and has been recently employed in microfluidic chips [21] creating the microchip-CE. The Agilent Bioanalyzer 2100 (Bioanalyzer) is a compact commercial LOC instrument fitted with both LED-Induced Fluorescence (LED-IF) and Laser Induced Fluorescence (LIF) detection systems [26]. The instrument can be made portable by the addition of an
external battery as it requires an electrical output to function. This instrument was developed for the analysis of DNA, RNA, proteins and fluorescent cell cytometry, but has been recently used beyond this scope for the analysis of amphetamines [26] and detection of explosives, both directly [27] and after extraction from soil samples [28]. This study aimed to develop a method to extract and analyse nitroaromatic explosives from vapour samples, using a simple and fast pre-concentration method with the aid of a small paper chad. After extraction the paper chad was inserted straight into the microchip and directly analysed by the LOC system, avoiding any other pre-concentration or extraction steps. The long-term goal of this method, which is currently in early stage development, is to detect commercial and improvised explosives by sampling the air around areas of interest, such as airports and cargo areas. This project was conceived following recent achievements in explosive extraction using paper devices found in literature [27–30]. The article describes a new method to extract explosive residues from vapour samples and directly identify them using a commercial LOC instrument.
2. Material and methods 2.1. Reagents Common nitroaromatic explosives standards were chosen to perform the experiments. TNT, DNB, 1,3,5-trinitrobenzene (TNB), methyl-2,4,6-trinitrophenylnitramine (Tetryl), 3-nitrotoluene (3NT), 2,4-dinitrotoluene (2,4-DNT), 2-amino-4,6-dinitrotoluene (2-A-4,6-DNT), 4-amino-2,6-dinitrotolune (4-A-2,6-DNT) were purchased from AccuStandard (New Haven, CT, USA) at a certified concentration of 1000 g/mL in acetonitrile. Analytical grade solvents (methanol, acetonitrile, and acetone) were purchased from ChemSupplies Pty Ltd (Gillman, SA. Australia). Sodium dodecyl sulphate (SDS) was obtained from Sigma-Aldrich (St. Louis, MO, USA) and sodium tetraborate from Fluka (Switzerland). Ultrapure grade water (18.2 M.cm−1 ) was obtained from a Sartorius 611 ® water purification system. DNA 1000 Dye Concentrate (blue) was © obtained from Agilent Technologies .
2.2. Instrumentation All experiments were performed on the Agilent 2100 ® Bioanalyzer . Separations were performed on the standard RNA ® 500 microchips . The microchip is manufactured in borate silica and allows the analysis of 12 samples in 1800 s. Microchannels with a depth of 10 m, width of 50 m, and a separation channel length of 15 mm, connects 16 wells. Out of the sixteen wells, twelve are used for sample analysis, while the other four are used as buffer reservoirs or waste collection wells. The chip layout can be seen in Fig. 1, as an adapted image from Lloyd, 2013 [31]. The separation channels lay between wells A4 and C4 and are filled with the background electrolyte. Wells B4 and D4 were used for waste, and all other wells were filled with samples. Each chip can be used up to three times after washing with milli-Q water and NaOH 1 M. The samples were injected using pinched mode. During injection, 100 V was applied to A4 and C4 to avoid sample diffusion, and a 1400 V potential difference was applied for 40 s. The separation was conducted using 1500 V for 100 s. Indirect detection was performed using laser emitting diode-induced fluorescence (ex = 635 nm, det = 680 nm). Data was collected with the Agilent 2100 Expert software (Agilent technologies, Waldbronn, Germany).
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
469
Fig. 1. RNA 500 microchip design (actual size 17 mm2 ). Figure adapted from Lloyd, 2013 [31]. The chip design and pattern belongs to Agilent technologies. All rights reserved.
2.3. Liquid analysis Liquid analyses were performed by the direct injection of liquid standards into the sample wells of the microchip in the desired concentration. Individual stock solutions of explosive standards (1000 ng/L) were diluted to the desired concentration into a sodium tetraborate (10 mM; pH 9.2) and SDS (50 mM) buffer. All solutions were stored in glass containers at 4 ◦ C in an explosion proof fridge. The background electrolyte used during the analysis was comprised of 10 mM sodium tetraborate (ph 9.2), 50 mM SDS, and Agilent Bioanalyzer DNA 1000 dye® 2% (v/v). The dye was used as a background for indirect detection by fluorescence quenching using LIF. The dye concentration was optimized to provide the best signal response. The background electrolyte was sonicated for 10 min and filtered using a 0.25 m syringe filter (Sartorius AG, Goettingen. Germany) prior to priming the chip. The chip was primed by pipetting 9 l of the background electrolyte into C4. Air pressure was applied for 45 s using a chip priming apparatus. Samples were then added to the remaining wells for analysis. Calibration curves were constructed by pipetting 9 l of the diluted explosive standard directly into the microchip sample wells at 1, 5, 10, 15 and 20 ng/l.
Fig. 2. A paper chad (0.5 mm diameter) is hole punched from a filter paper grade TM 1 (Whatman ) and inserted in a 2 ml vial cap. The vial, which contains explosive standards diluted in organic solvents, is subjected to heating in a dry block for 15 min. After extraction the paper chad is retrieved for analysis.
2.4. Vapours extraction procedure Following the liquid analysis results, three explosive standards were chosen and diluted to the desired concentrations into organic solvents to facilitate evaporation. The solution was placed into a 2 ml glass vial. A circular 0.5 mm diameter paper chad was hole punched from a WhatmanTM qualitative filter paper Grade 1 sheet, supplied by Sigma-Aldrich (St. Louis, MO, USA). The paper chad was then inserted in the cap of the 2 mL glass vial containing the explosive mixture. The vial was exposed to temperatures ranging from 22 ◦ C to 80 ◦ C for 15 min using a dry heat block (Fig. 2). Methanol, acetonitrile, acetone, and water were compared as extraction solvents. Different amounts of each solvent were tested with equimolar amounts of the target explosives to determine the optimal extraction solvent, solvent volume, and temperature.
Fig. 3. After extraction the paper chad was folded twice and inserted directly into the sample well of the microchip containing 5 L of buffer. After insertion, 4 L of buffer was added to the filter paper and the sample was analyzed. Figure adapted from Ueland et al., 2016 [28].
of sample or background electrolyte the microchip was inserted into the instrument for analysis (Fig. 3). Calibration curves for each explosive were constructed by vapour extraction of diluted explosive standard solutions at 5, 10, 20, 30 and 40 ng/l. The recovery rates from the paper chad were determined by correlating the response achieved during the liquid analysis of the explosive standards at 20 ng/l and the vapour extraction of the standards in the same concentration (n = 4). 3. Results and discussion
2.5. Vapour analysis 3.1. Liquid analysis After the extraction process, the paper chad was retrieved from the vial cap, folded twice in order to fit into the sample reservoir, and placed directly into the lab on a chip injection well containing 5 L of the background electrolyte. 4 L of electrolyte was added to obtain a final volume of 9 L. After all wells had been filled with 9 L
The samples were electrokinetic injected, electrophoreticaly separated, and detected by fluorescence. This involves the application of high voltage in the system, which creates an electric field for CE separations [26]. The addition of SDS to the background
470
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
Fig. 4. Electropherogram of the separation of 9 l of a 20 ng/L mixture of eight explosives diluted into buffer, pipetted directly into a sample well of the microchip. 1) TNB; 2) 1,3-DNB; 3) TNT; 4) Tetryl; 5) 2,4-DNT; 6) 3-NT; 7) 2,6-DNT, 8) 2-A-4,6DNT. Background electrolyte (10 mM Borate : 50 mM SDS; pH 9,2), injection 1500 V for 40 s, separation 1500 V for 100 s.
electrolyte allows separation of the explosives, which are neutral molecules. In the method known as micellar electrokinetic chromatography (MEKC), the addition of a surfactant to the electrolyte forms micelles, which act as a pseudo stationary phase to allow the separation of neutral molecules [25]. The indirect detection of the samples caused by fluorescence quenching is due to the addition of dye to the background electrolyte, which causes the explosives to appear as negative peaks in the electropherograms. The background electrolyte produced consistent baselines across all standards and sample runs. It was possible to separate, detect and identify eight explosives in a single run (Fig. 4) using liquid analysis. 9 L of a 20 ng/L solution containing TNT, DNB, TNB, Tetryl, 3-NT, 2,4-DNT, 2-A-4,6DNT, and 4-A-2,6-DNT diluted in buffer (10 mM sodium tetraborate 10 mM and 50 mM SDS; pH 9.2) were pipetted directly into the sample well of a previously primed microchip. All other wells were filled with 9 L of samples or buffer, and the microchip was inserted into the instrument for liquid analysis. The background electrolyte comprised of 10 mM sodium tetraborate, 50 mM SDS, at pH 9.2, and 2% v/v of Agilent DNA 1000 dye® . Liquid analysis calibration curves were constructed by pipetting 9 L of the explosives standards directly into the microchip sample wells at 1, 5, 10, 15 and 20 ng/L. The respective LOD’s were calculated as 3.3 / slope of the calibration curve, where is the slope deviation [32]. The minimum detectable masses were calculated as 9 times the LOD once 9 L of each standard was pipetted directly into the sample wells per run. The achieved values for liquid and vapour analysis as well as the recovery rates of the explosives after vapour extraction can be seen in Table 1. 3.2. Vapour extraction Following the liquid extraction results, TNB, Tetryl and TNT were chosen for the vapour analysis experiments. These three standards were chosen due to their retention times, sensitivity and availability. In order to facilitate the evaporation of the explosive standards, organic solvents were added to the mixture. After the targets for the extraction experiments were chosen, different organic solvents were compared. Methanol, distilled water, acetone and acetonitrile were added individually to the explosive solutions resulting in equimolar mixtures of 40 ng/L of each explosive standard per solvent. The mixtures were left to evaporate for 15 min at room temperature (between 22 and 25 degrees), 40, 60, and 80 ◦ C in a dry heat block. This experiment aimed to determine the best solvent and temperature to achieve the optimal extraction process. The results show that, for the three explosives, the optimal tem-
perature and solvent were 80 ◦ C and acetone, respectively (Fig. 5 a–c). Results show that, in vapour samples, the solvent with the lowest evaporation temperature was the most effective in the collection of explosive standards residues. In the comparison between acetone (56 ◦ C), methanol (64 ◦ C), acetonitrile (82 ◦ C), and water (100 ◦ C), acetone proved to be the most conducive for this experiment. This result is probably due to the higher concentration of analytes delivered per time provided by the solvent with more volatility. The higher volatility of acetone represents a higher flow of this solvent through the paper chad than the other solvents tested, facilitating the transportation of explosive residues. Although no comparison could be found in the literature for this type of extraction, previous literature results vary and methanol has been suggested for recovering explosive residues from soil [28], while acetone was used for recovery and post-blast analysis of organic explosives [33], and a review observed the use of both methanol and acetone for explosives detection in different media [34]. In the experiment described in this study, acetone was the best solvent for the recovery of explosives from vapour samples due to its high volatility. As previously described the higher volatility of this solvent may have aided in carrying a higher amount of explosives residues per time due to its higher flow when heated. Another theory would be that due to its higher volatility acetone was easily removed from the paper chad after transporting the vapours, while less volatile solvents, which present a higher evaporation temperature may have remained with the explosive residues in the filter paper acting as masking agents. 3.3. Vapour analysis The vapour analysis lead to clear negative peaks resulting from the indirect detection of the analytes after fluorescence quenching. The three target explosives were successfully separated, detected and identified using the developed method for vapour extraction and analysis. The explosives were diluted in acetone to 30 ng/L.150 L of the mixed solution was placed into a 2 mL glass vial. The vial was covered with a cap containing a paper chad and heated to 80 ◦ C for 15 min in a dry block, achieving full evaporation of the liquid. The paper chad was retrieved from the plastic cap, folded in two and directly placed in a sample well of a previously primed chip, which contained 4 L of buffer. After the paper chad insertion in the sample well, 5 L of buffer were added. All the other wells were filled with 9 L of samples or buffer before the run could take place. Recovery rates for the target explosives were found by comparing the peak area of the four replicate analyses of 20 ng/L TNB, TNT and Tetryl samples after liquid and vapour analysis and are shown in Table 1. Fig. 6 shows the comparison between the peak responses for both analyses. The electropherograms show the separation of three explosives in a 20 ng/L mixture. While the liquid analyses were performed by pipetting 9 L of the explosive mixture in buffer directly into the sample well of the microchip, the vapour extraction analyses were performed after 150 L of the explosive mixture in acetone was evaporated, with the aid of a dry heat block, onto a 0.5 cm filter paper. The filter paper was then folded twice and inserted in the sample well of the microchip, avoiding any other extraction steps. For both experiments, all other wells were filled with samples or buffer. The background electrolyte was composed of buffer plus dye at 2% (v/v). As can be seen in Fig. 6 the separation of the three compounds was clear and the baseline was consistent on both analysis methods. The vapour analysis showed a delayed separation when compared to liquid analysis, which can be observed by comparing the time elapsed between the electroosmotic flow peak and the explosives peaks on both runs. The same baseline noise can be observed on both electrophero-
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
471
Table 1 Values for the average retention time and minimum detectable masses for direct injection and vapour extraction in the paper chad, and average recovery rates for the paper chad when compared to the direct injection (n = 4). EXPLOSIVES
TNB TNT TETRYL
LIQUID ANALYSIS
VAPOUR ANALYSIS
Average Retention Time (s)
Minimum Detectable mass (ng)
Average Retention Time (s)
Minimum Detectable mass (ng)
Average Recovery (%)
4.72 8.00 10.50
2.32 ± 0.7 3.25 ± 0.4 2.35 ± 0.4
5.36 8.81 10.23
6.03 ± 4.5 9.99 ± 2.3 14.22 ± 1.2
47.00 75.00 29.00
Fig. 5. Response analysis of 40 ng/L target explosive samples for different organic solvents at increasing evaporation temperature and constant time (15 min) after extraction in the paper chad.
Fig. 6. LOC analysis of explosive standards using (I) direct injection of liquid standards and (II) vapour extraction of liquid standards in filter paper. Explosives used: 1.TNB; 2. TNT; 3. TETRYL at 20.ng/L. Background electrolyte (10 mM Borate : 50 mM SDS; pH 9,2), injection 1500 V for 40 s, separation 1500 V for 100 s.
grams, although it is more evident in the vapour analysis due to the peak size of the standards. The peak size on the vapour extraction analysis was considerably smaller compared to the liquid injection analysis. This lower yield is to be expected due to the affinity of the target molecules for the mobile phase (acetone) and stationary phase (filter paper). Additionally, in the vapour extraction analysis, consideration is required for sample degradation due to heating, as well as the presence of organic solvent in the filter paper, which can act as a masking agent. Moreover, in this type of experiment we were not able to vaporize one hundred percent of the explosive sample, as observed by the coloured residue present in the bottom of the vial after the liquid content had been fully evaporated. Limits of detection and the minimum detectable masses for the three target explosives were calculated. Calibration curves were constructed for each explosive after vapour extraction from solutions at 10, 20, 30, and 40 ng/L. After the vapour extraction process the paper chads were folded and placed directly into the microchip sample wells. The LOD’s were calculated as 3.3/slope of the calibration curve where = error in the slope [32]. The three explosives targeted for the vapour extraction and detection of residues in the headspace were readily detected and visualised using fluorescence
quenching. The minimum detectable amounts were 6.03, 9.99, and 14.22 ng for TNB, TNT and Tetryl, respectively. The recoveries from paper chads after vapour extraction, when compared to the direct injection were 47, 75, and 29% for TNB, TNT and Tetryl, respectively. The minimum detectable masses of the selected explosives and the recovery rates from the paper chads can be seen in Table 1. Intra and interday variations for the vapour extraction process were determined over a three day period by comparing the peak area of the four replicate analyses. The intraday variation across the three day period was 7.9% RSD for TNB; 3.22% RSD for TNT; and 4.5% RSD for Tetryl. The interday variation was RSD: 4.7% (TNB); 1.7% (TNT); and 1.2% (Tetryl). All three target explosives were successfully detected after the vapour extraction procedure, and the achieved LOD’s are comparable to the conventional methods used on bench CE instruments [35] proving the efficiency of microchip models. The proposed extraction process of vapours from explosive standards onto a filter paper chad proved to be suitable for the detection of nitroaromatic explosives. As can be seen in Table 1 the recovery rates for the three target explosives are similar to those described in the literature using different methods [28,36]. The proposed method was able to achieve recovery rates as high as 75% for TNT, which can be compared to studies conducted by Batlle et al. [36] that were able to achieve recovery rates of 86 and 90% of TNT in acetone using HPLC to determine explosives in vapour phases. Ueland et al. [28] described recovery rates of explosives from soil samples, using paper chads analysed by the Bioanalyzer, ranging from 53 to 82%. In addition, the insertion of the paper chad directly into the sample well did not block the channel or interfere with the sample separation, avoiding an extra step for the pre-concentration or extraction of the explosives. Directly inserting the paper chad into the chip sample wells represents a significantly simplified extraction process. Results show that a minimum temperature of 40 ◦ C is necessary to vaporize the compounds using this solvent. Nonetheless, the optimized extraction process for this preliminary study has shown positive results as demonstrated by the LODs and recovery rates achieved. The lower response observed from Tetryl during the vapour extraction experiment can be explained by different factors. The negative peaks observed in the electropherograms are due to the fluorescence quenching of the dye present on the
472
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
background electrolyte [27,37,38], hence explosives with higher fluorescence quenching power, such as TNB and TNT will have a higher impact in this experiment than nitramines, such as Tetryl and RDX, which proved to yield a lower quenching power [37]. Moreover, nitramines are known to be less volatile and sensitive than the nitroaromatics [2], which significantly impacts the amount of explosive residues present in the headspace and can explain why Tetryl was readily detected during liquid injection experiments but the same was not seen in the vapour extraction method. The current method was optimized for the analysis of three target explosives, which were successfully detected, separated and identified after a fast and simple process of extraction, using low cost equipment and an innovative platform. The ability of this instrument to analyse explosive mixtures, rapidly and at relatively low costs demonstrates the potential of this method for explosive residues detection in gas and vapour samples. Air samples analysis of explosives is of pivotal importance for military and anti-terrorist actions [11] and this study shows an innovative and promising technique for the detection of explosive residues in the headspace. However it is important to note that given the respective LODs and considering the saturated vapour pressure for each of the three target compounds the amount of air to be sampled in a real case scenario using the current method would be considerably large. For example TNT which had a minimum detectable mass from vapour analysis calculated to 9.9 ng and a saturated vapour pressure at 25 ◦ C of 9.15 ppbv [7] would need approximately 100 mL of air sampled at 100% efficiency. Considering that trace explosives vapours in the environment are generally calculated at concentrations of at least two orders of magnitude below the saturated vapour concentration, the volume required for the correct detection of TNT would be 10 L. Applying the same concept for the other two target analytes a total of 4 L would need to be sampled for the correct analysis of TNB (saturated vapour pressure at 25 ◦ C 20 ppbv [7]) and approximately 1400 L would have to be sampled for the correct analysis of Tetryl (saturated vapour pressure at 25 ◦ C 0.0074 ppbv [7]). Those values indicate the need for a pre-concentration procedure. By utilizing an easy to acquire paper chad, which can be directly inserted into the analysing instrument, we were able to simplify the pre-concentration and extraction methods. Moreover, with the use of a paper chad we were able to achieve a fast and reliable pre-concentration process without the need for ionization or other complex procedures. It is important to consider that this is a preliminary test, which has showed the potential and suitability of microchip electrophoresis analysis for the detection of explosives in the gas and vapour samples. The results point to the viability of the use of this method in future portable instruments for field work as the proposed platform with the use of a LOC device presents only one preconcentration step (a simple paper chad), does not need complex procedures for the analytes extraction and has decades of research behind it.
4. Conclusion A simple, innovative, rapid and cost-effective method to collect explosive vapour residues was developed using a 2 mL glass vial topped with a 0.5 m filter paper chad. The paper chad was used as a pre-concentration device during the extraction of vapours from explosives. This technique presents itself as a viable alternative to eliminate complex procedures, keeping the selectivity, sensitivity and reliability of a well-known method such as microchip-CE. When compared to IMS and MS detection techniques the proposed platform presents itself as a better alternative as it does not requires ionization of the analytes. The analyte ionization in a complex real life scenario can be problematic due to competitive ionization path-
ways with interferents, which can lead to a reduction in selectivity and sensitivity. Also, the use of a LIF detector presents a clear advantage over other instruments, as the quenching mechanism is selective to nitro-containing analytes, improving the selectivity for target analytes over interferents. When compared to other fluorescence quenching technologies the proposed platform again holds an advantage as the separation allows for another dimension of selectivity leading to a potentially enhanced sensitivity. This method was able to detect and quantify vapour emitted from explosive standards after evaporation. The extraction process used minute amounts of standards and reagents, and low cost materials. The explosives were detected through fluorescence quenching and separation occurred in under 20 s, representing a fast and accurate approach for the detection of explosives in vapour samples. Moreover this method has proved to be reliable as the results were replicable and the intraday values were within the expected normal values, even when compared to other instruments such as HPLC [39]. The use of a filter paper to collect explosives residues in the headspace provides an inexpensive approach for the detection of explosives. The Bioanalyzer presents itself as a viable alternative for the analysis of explosive mixtures, rapidly, reliably and at relatively low cost using small amounts of standards and reagents. The LOC device does not require pumps or gas for its operation and can analyse up to 12 samples in under 1800 s. Its low weight, small size and robust nature demonstrates excellent potential for portability and field deployment. It is important to highlight that the explosive threat is not one from a diluted explosive standard on a laboratory bench but instead, from the concealed real explosives in civil areas. Explosives are known to have low volatility and because of that, are incredibly difficult to detect in the headspace [40]. Concealed explosives are even more problematic to detect, as the amount of vapours in the air for detection are limited [8]. Therefore, this experiment in no way reflects detection of concealed explosives in real scenarios as the paper chads are placed only centimetres away from the liquid standards, maximizing the amount of vapours extracted in the pre-concentration step. Heating the liquid standards also considerably raises their volatility facilitating the analysis and analytes detection. Nevertheless, this experiment demonstrates a new and promising platform that can be used for vapour analysis and should be explored more thoroughly. The next development steps will focus on improving the detection limits of the instrument and optimizing the extraction procedure and pre-concentration step. After the instrument detection capabilities have been properly unveiled, experiments will be conducted with real samples using the same pre-concentration step, also aiming to improve the robustness of the platform. Solid samples of real explosives will be stored into sealed containers and air samples will be extracted by vacuum pump onto the filter paper chad. Analysis will be conducted with a LOC instrument and LIF detection to establish whether it is possible to detect explosives from air samples using this proposed platform and extraction method. The ultimate goal is to create a platform that does not need the pre-concentration step and can constantly perform sampling and analysis. The vapour and gases would be fan forced drawn into the instrument which would constantly run the analysis of the air samples. To the best of our knowledge, this is the first report of a combination of filter paper chads and LOC instruments to extract and directly analyse explosive residues from vapour samples. This approach provides a new and promising method to identify explosives in the headspace.
Conflicts of interest There are no conflicts of interest to declare.
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
Acknowledgements The corresponding author is supported by a Science Without Borders Scholarship of the National Council for Scientific and Technological Development of the Government of Brazil (CNPq).201677/2014-8. The opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors, and the financial support by the Government ministry does not constitute an express or implied endorsement of the results or conclusions of the research. References [1] J.S. Caygill, F. Davis, S.P. Higson, Current trends in explosive detection techniques, Talanta 88 (2012) 14–29. [2] D.S. Moore, Instrumentation for trace detection of high explosives, Rev. Sci. Instrum. 75 (8) (2004) 2499–2512. [3] J. Feltes, et al., Gas chromatographic and mass spectrometric determination of nitroaromatics in water, J. Chromatogr. A 518 (1990) 21–40. [4] D. Badgujar, et al., Advances in science and technology of modern energetic materials: an overview, J. Hazard. Mater. 151 (2–3) (2008) 289–305. [5] Y.T. Yew, A. Ambrosi, M. Pumera, Nitroaromatic explosives detection using electrochemically exfoliated graphene, Sci. Rep. 6 (2016) 33276. [6] M. Krausa, A.A. Reznev, in: N.S. Series (Ed.), Vapour and Trace Detection of Explosives for Anti-Terrorism Purposes, vol. 167, Springer Science & Business Media, 2004. [7] R.G. Ewing, et al., The vapor pressures of explosives, Trac Trends Anal. Chem. 42 (2013) 35–48. [8] S. Singh, Sensors—an effective approach for the detection of explosives, J. Hazard. Mater. 144 (1–2) (2007) 15–28. [9] M.M. Houck, J.A. Siegel, Fundamentals of Forensic Science, Academic Press, 2009. [10] Y.H. Lee, et al., Dipyrenylcalix [4] arene—a fluorescence-based chemosensor for trinitroaromatic explosives, Chem. Eur. J. 16 (20) (2010) 5895–5901. [11] A. Hakonen, et al., Explosive and chemical threat detection by surface-enhanced Raman scattering: a review, Anal. Chim. Acta 893 (2015) 1–13. [12] R.G. Ewing, et al., A critical review of ion mobility spectrometry for the detection of explosives and explosive related compounds, Talanta 54 (3) (2001) 515–529. [13] R.M. Alberici, et al., Ambient mass spectrometry: bringing MS into the “real world”, Anal. Bioanal. Chem. 398 (1) (2010) 265–294. [14] S.J. Toal, et al., Visual detection of trace nitroaromatic explosive residue using photoluminescent metallole-containing polymers, J. Forensic Sci. 52 (1) (2007) 79–83. [15] E.T. da Costa, et al., Unmanned platform for long-range remote analysis of volatile compounds in air samples, Electrophoresis 33 (17) (2012) 2650–2659. [16] Q. Lu, et al., Sensitive capillary electrophoresis microchip determination of trinitroaromatic explosives in nonaqueous electrolyte following solid phase extraction, Anal. Chim. Acta 469 (2) (2002) 253–260. [17] J. Wang, et al., Single-channel microchip for fast screening and detailed identification of nitroaromatic explosives or organophosphate nerve agents, Anal. Chem. 74 (5) (2002) 1187–1191.
473
[18] P. Gravesen, J. Branebjerg, O.S. Jensen, Microfluidics-a review, J. Micromechanics Microengineering 3 (4) (1993) 168. [19] V. Dolnik, S. Liu, Applications of capillary electrophoresis on microchip, J. Sep. Sci. 28 (15) (2005) 1994–2009. [20] E. Verpoorte, Microfluidic chips for clinical and forensic analysis, Electrophoresis 23 (5) (2002) 677–712. [21] M. Pumera, Analysis of explosives via microchip electrophoresis and conventional capillary electrophoresis: a review, Electrophoresis 27 (1) (2006) 244–256. [22] J. Wang, Microchip devices for detecting terrorist weapons, Anal. Chim. Acta 507 (1) (2004) 3–10. [23] J.R. Petersen, A.A. Mohammad, Clinical and Forensic Applications of Capillary Electrophoresis, Springer Science & Business Media, 2001. [24] D.R. Baker, F. Foret, Capillary electrophoresis: techniques in analytical chemistry series, Anal. Biochem. 231 (1) (1995) 274. [25] S.A. Oehrle, Analysis of nitramine and nitroaromatic explosives by capillary electrophoresis, J. Chromatogr. A 745 (1–2) (1996) 233–237. [26] A. Lloyd, et al., A rapid method for the in-field analysis of amphetamines employing the Agilent Bioanalyzer, Anal. Methods 3 (7) (2011) 1535–1539. [27] A. Pesenti, et al., Coupling paper-based microfluidics and lab on a chip technologies for confirmatory analysis of trinitro aromatic explosives, Anal. Chem. 86 (10) (2014) 4707–4714. [28] M. Ueland, et al., Capillary-driven microfluidic paper-based analytical devices for lab on a chip screening of explosive residues in soil, J. Chromatogr. A 1436 (2016) 28–33. [29] A.W. Martinez, et al., Patterned paper as a platform for inexpensive, low-volume, portable bioassays, Angew. Chemie Int. Ed. 46 (8) (2007) 1318–1320. [30] A.K. Yetisen, M.S. Akram, C.R. Lowe, Based microfluidic point-of-care diagnostic devices, Lab Chip 13 (12) (2013) 2210–2251. [31] A. Lloyd, The Analysis of Amphetamine-type Stimulants Using Microchip Capillary Electrophoresis, 2013. [32] A. Shrivastava, V. Gupta, Methods for the determination of limit of detection and limit of quantitation of the analytical methods, Chron. Young Sci. 2 (1) (2011) 21. [33] S. Calderara, D. Gardebas, F. Martinez, Solid phase micro extraction coupled with on-column GC/ECD for the post-blast analysis of organic explosives, Forensic Sci. Int. 137 (1) (2003) 6–12. [34] J.I. Steinfeld, J. Wormhoudt, Explosives detection: a challenge for physical chemistry, Annu. Rev. Phys. Chem. 49 (1) (1998) 203–232. [35] J.P. Hutchinson, et al., Identification of inorganic improvised explosive devices by analysis of postblast residues using portable capillary electrophoresis instrumentation and indirect photometric detection with a light-emitting diode, Anal. Chem. 79 (18) (2007) 7005–7013. [36] R. Batlle, et al., On-line coupling of supercritical fluid extraction with high-performance liquid chromatography for the determination of explosives in vapour phases, J. Chromatogr. A 963 (1–2) (2002) 73–82. [37] R.V. Taudte, et al., A portable explosive detector based on fluorescence quenching of pyrene deposited on coloured wax-printed PADs, Lab Chip 13 (21) (2013) 4164–4172. [38] A. Ponnu, E.V. Anslyn, A fluorescence-based cyclodextrin sensor to detect nitroaromatic explosives, Supramol. Chem. 22 (1) (2010) 65–71. [39] R.V. Taudte, et al., Development of a UHPLC method for the detection of organic gunshot residues using artificial neural networks, Anal. Methods 7 (18) (2015) 7447–7454. [40] J. Yinon, Counterterrorist Detection Techniques of Explosives, Elsevier, 2011.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
The analysis of nitrate explosive vapour samples using Lab-on-a-chip instrumentation Vitor Taranto a,∗ , Maiken Ueland a , Shari L. Forbes b , Lucas Blanes c a b c
Centre for Forensic Science, University of Technology Sydney, Australia Département de chimie, biochimie et physique, Université du Québec à Trois-Rivières, Canada Oswaldo Cruz Foundation (FIOCRUZ), Curitiba PR, Brazil
a r t i c l e
i n f o
Article history: Received 6 January 2019 Received in revised form 29 May 2019 Accepted 2 June 2019 Available online 3 June 2019 Keywords: Explosive detection Portable instrumentation Microchip electrophoresis
a b s t r a c t The detection and analysis of explosives and explosive-related compounds is a heightened priority in recent years for homeland security and counter-terrorism applications. This study aimed to evaluate the use of a commercial Lab-On-a-Chip (LOC) instrument for the analysis of explosive vapours, with the longterm goal of developing a portable instrument for passively detecting explosives in air samples. A simple method to collect explosive vapour residues was developed using a glass vial containing varying amounts of the target explosives (1 mg/mL). Standards were diluted to the desired concentration in 150 L of acetone to facilitate the evaporation. The top of the vial was covered with a circular 0.5 cm diameter filter paper and exposed to a range of temperatures from 22 ◦ C to 80 ◦ C for 15 min. Following evaporation, the filter paper chads were folded and inserted into the LOC wells containing the separation buffer for the analysis, avoiding any further extraction step. After successfully separating and detecting eight explosives via liquid analysis, three explosives were chosen as targets for the vapour analysis experiments. 1,3,5Trinitrobenzene (TNB), 2,4,6-Trinitrotoluene (TNT), and 2,4,6-Trinitrophenylmethylnitramine (Tetryl) were successfully separated, detected and identified following the vapour extraction of explosive standards onto filter paper chads. Limits of detection for the liquid analysis were demonstrated to be 2.32 ng for TNB, 2.35 ng for Tetryl, and 3.25 ng for TNT. The minimum detectable mass found for the vapour analysis was 6.03 for TNB, 9.99 ng for TNT, and 14.22 ng for Tetryl. The average recovery from the paper chads was 29% for Tetryl, 47% for TNB, and 75% for TNT (n = 4), comparable with findings from previous studies. Results show that a minimum temperature of 40 ◦ C is necessary to vaporize the compounds using acetone, while the best results were achieved when heating the vial to 80 ◦ C. The use of a filter paper to collect the explosives residues, avoiding any additional extraction step, and the ability to analyze these compounds using a LOC instrument, makes this approach a future alternative method for explosive residues detection in the headspace. © 2019 Elsevier B.V. All rights reserved.
1. Introduction An increasing focus on counter-terrorism and homeland security globally demands fast and reliable detection and analysis of explosives and related compounds [1]. The investigation of explosives detection in the headspace is important for the screening of large areas for commercial and improvised explosive devices, however explosives detection through air samples is notoriously difficult due to the low volatility of these compounds [2]. Nitroaro-
∗ Corresponding author at: Centre for Forensic Science, School of Mathematical and Physical Sciences, University of Technology Sydney, PO Box 123, Broadway NSW 2007, Australia. E-mail address: [email protected] (V. Taranto). https://doi.org/10.1016/j.chroma.2019.06.003 0021-9673/© 2019 Elsevier B.V. All rights reserved.
matic compounds are widely used in the chemical industry for the production of explosives [3], and their detection is of paramount importance for national security as modern military explosives are generally nitrogen-containing aromatic compounds [2,4]. Compounds such as 2,4,6-trinitrotoluene (TNT) and the more volatile and soluble 1,3-dinitrobenzene (DNB) are two of the most representative nitro-substituted aromatic explosive compounds [5]. The detection of explosive substances in the air is hampered by their low vapour pressures [6]. For example, Tetryl has a vapour pressure of 7.41 × 10−12 atm at room temperature (25 ◦ C), while nitroaromatics such as TNT and TNB have vapour pressures around 9.15 × 10-9 and 2.00 × 10-8 atm, respectively [7]. The diffusion of vapours from explosives is dependent on the temperature and airflow around the object; as the temperature increases, the vapour pressure also increases, so that some explo-
468
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
sives can yield a vapour pressure four times higher at 100 ◦ C than at 25 ◦ C [6]. The need for fast detection of explosives has generated a demand for rapid, sensitive and reliable methods, which could be made rugged and portable for field analysis [8–10]. Numerous analytical methods have been tested for detection of explosives and its breakdown products in air, water and soil [1,11]. Some instruments such as ion mobility spectrometry (IMS) and infrared (IR) detection based platforms have been widely researched for miniaturization and portability and are currently broadly used in airports, with proven parts per billion detection limits [1]. Other notable portable techniques developed for gas and vapour sampling are colorimetric and fluorescence quenching tests, chemiluminescence sensors, and mass spectrometry (MS) devices. All of these techniques have portable instrument representatives currently on the market for the detection of hazardous materials in gas and vapour samples. However it is important to note that IMS sampling requires contact with the surface of interest (i.e. swabbing) to achieve peak sensitivity [12], and when used for vapours and gas analysis needs more than one pre-concentration step to perform its analysis. IMS and MS detection systems require ionization of the analytes, which can be troublesome in a real time scenario due to competitive ionization pathways with field interferents [1,13]. Luminescence sensors and colorimetric tests may be described as utilising either direct or indirect detection methods, dependent on whether they use the fluorescence that the sample may emit or its effect on a fluorescent material, such as quenching [1]. Quenching techniques have already been considered in the literature as a subjective activity needing a confirmatory test with a better qualitative instrument and may also suffer from field interferents [14]. In general this tests are portable and easy to use but lack sensitivity, reliability and reproducibility [1]. Field deployable instruments must be small, cost efficient, rugged and be able to perform real time detection of hazardous materials among troublesome field activities [15]. Lab-on-a-chip (LOC) devices have promising features for the development of portable instruments such as the compact size and portability, quick analysis time, low cost, small sample and reagent consumption, and the possibility to integrate with other systems [16]. These devices result from the development of a technique know as microfluidics [17], which consists of pumps, valves, flow sensors, separation capillaries, and chemical detectors integrated on a single substrate or as compacted modules [18]. LOC devices are portable and capable of performing extremely fast, cost-effective separations [19], and represents an attractive alternative for the rapid analysis of forensic samples [20], including explosives [21,22]. The integration of capillary electrophoresis (CE) into LOC devices allows for the fast analysis of chemicals with low reagent consumption [15]. CE is an analytical technique used to separate electrically charged molecules based on their electrophoretic mobility. The separation occurs in a silica capillary tube connected between two buffer reservoirs by applying high voltage. The electric field produced by the high voltage separates the electrically charged analytes in the electrolyte while passing though the capillary. The separation is dependent on the difference in charge and friction caused by the size of the molecule inside the capillary [23,24]. Nitroaromatic compounds can be directly detected by ultraviolet–visible spectroscopy (UV–vis) [25]. This detection approach is widely used in conventional capillary electrophoretic analysis of nitroaromatic explosives and has been recently employed in microfluidic chips [21] creating the microchip-CE. The Agilent Bioanalyzer 2100 (Bioanalyzer) is a compact commercial LOC instrument fitted with both LED-Induced Fluorescence (LED-IF) and Laser Induced Fluorescence (LIF) detection systems [26]. The instrument can be made portable by the addition of an
external battery as it requires an electrical output to function. This instrument was developed for the analysis of DNA, RNA, proteins and fluorescent cell cytometry, but has been recently used beyond this scope for the analysis of amphetamines [26] and detection of explosives, both directly [27] and after extraction from soil samples [28]. This study aimed to develop a method to extract and analyse nitroaromatic explosives from vapour samples, using a simple and fast pre-concentration method with the aid of a small paper chad. After extraction the paper chad was inserted straight into the microchip and directly analysed by the LOC system, avoiding any other pre-concentration or extraction steps. The long-term goal of this method, which is currently in early stage development, is to detect commercial and improvised explosives by sampling the air around areas of interest, such as airports and cargo areas. This project was conceived following recent achievements in explosive extraction using paper devices found in literature [27–30]. The article describes a new method to extract explosive residues from vapour samples and directly identify them using a commercial LOC instrument.
2. Material and methods 2.1. Reagents Common nitroaromatic explosives standards were chosen to perform the experiments. TNT, DNB, 1,3,5-trinitrobenzene (TNB), methyl-2,4,6-trinitrophenylnitramine (Tetryl), 3-nitrotoluene (3NT), 2,4-dinitrotoluene (2,4-DNT), 2-amino-4,6-dinitrotoluene (2-A-4,6-DNT), 4-amino-2,6-dinitrotolune (4-A-2,6-DNT) were purchased from AccuStandard (New Haven, CT, USA) at a certified concentration of 1000 g/mL in acetonitrile. Analytical grade solvents (methanol, acetonitrile, and acetone) were purchased from ChemSupplies Pty Ltd (Gillman, SA. Australia). Sodium dodecyl sulphate (SDS) was obtained from Sigma-Aldrich (St. Louis, MO, USA) and sodium tetraborate from Fluka (Switzerland). Ultrapure grade water (18.2 M.cm−1 ) was obtained from a Sartorius 611 ® water purification system. DNA 1000 Dye Concentrate (blue) was © obtained from Agilent Technologies .
2.2. Instrumentation All experiments were performed on the Agilent 2100 ® Bioanalyzer . Separations were performed on the standard RNA ® 500 microchips . The microchip is manufactured in borate silica and allows the analysis of 12 samples in 1800 s. Microchannels with a depth of 10 m, width of 50 m, and a separation channel length of 15 mm, connects 16 wells. Out of the sixteen wells, twelve are used for sample analysis, while the other four are used as buffer reservoirs or waste collection wells. The chip layout can be seen in Fig. 1, as an adapted image from Lloyd, 2013 [31]. The separation channels lay between wells A4 and C4 and are filled with the background electrolyte. Wells B4 and D4 were used for waste, and all other wells were filled with samples. Each chip can be used up to three times after washing with milli-Q water and NaOH 1 M. The samples were injected using pinched mode. During injection, 100 V was applied to A4 and C4 to avoid sample diffusion, and a 1400 V potential difference was applied for 40 s. The separation was conducted using 1500 V for 100 s. Indirect detection was performed using laser emitting diode-induced fluorescence (ex = 635 nm, det = 680 nm). Data was collected with the Agilent 2100 Expert software (Agilent technologies, Waldbronn, Germany).
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
469
Fig. 1. RNA 500 microchip design (actual size 17 mm2 ). Figure adapted from Lloyd, 2013 [31]. The chip design and pattern belongs to Agilent technologies. All rights reserved.
2.3. Liquid analysis Liquid analyses were performed by the direct injection of liquid standards into the sample wells of the microchip in the desired concentration. Individual stock solutions of explosive standards (1000 ng/L) were diluted to the desired concentration into a sodium tetraborate (10 mM; pH 9.2) and SDS (50 mM) buffer. All solutions were stored in glass containers at 4 ◦ C in an explosion proof fridge. The background electrolyte used during the analysis was comprised of 10 mM sodium tetraborate (ph 9.2), 50 mM SDS, and Agilent Bioanalyzer DNA 1000 dye® 2% (v/v). The dye was used as a background for indirect detection by fluorescence quenching using LIF. The dye concentration was optimized to provide the best signal response. The background electrolyte was sonicated for 10 min and filtered using a 0.25 m syringe filter (Sartorius AG, Goettingen. Germany) prior to priming the chip. The chip was primed by pipetting 9 l of the background electrolyte into C4. Air pressure was applied for 45 s using a chip priming apparatus. Samples were then added to the remaining wells for analysis. Calibration curves were constructed by pipetting 9 l of the diluted explosive standard directly into the microchip sample wells at 1, 5, 10, 15 and 20 ng/l.
Fig. 2. A paper chad (0.5 mm diameter) is hole punched from a filter paper grade TM 1 (Whatman ) and inserted in a 2 ml vial cap. The vial, which contains explosive standards diluted in organic solvents, is subjected to heating in a dry block for 15 min. After extraction the paper chad is retrieved for analysis.
2.4. Vapours extraction procedure Following the liquid analysis results, three explosive standards were chosen and diluted to the desired concentrations into organic solvents to facilitate evaporation. The solution was placed into a 2 ml glass vial. A circular 0.5 mm diameter paper chad was hole punched from a WhatmanTM qualitative filter paper Grade 1 sheet, supplied by Sigma-Aldrich (St. Louis, MO, USA). The paper chad was then inserted in the cap of the 2 mL glass vial containing the explosive mixture. The vial was exposed to temperatures ranging from 22 ◦ C to 80 ◦ C for 15 min using a dry heat block (Fig. 2). Methanol, acetonitrile, acetone, and water were compared as extraction solvents. Different amounts of each solvent were tested with equimolar amounts of the target explosives to determine the optimal extraction solvent, solvent volume, and temperature.
Fig. 3. After extraction the paper chad was folded twice and inserted directly into the sample well of the microchip containing 5 L of buffer. After insertion, 4 L of buffer was added to the filter paper and the sample was analyzed. Figure adapted from Ueland et al., 2016 [28].
of sample or background electrolyte the microchip was inserted into the instrument for analysis (Fig. 3). Calibration curves for each explosive were constructed by vapour extraction of diluted explosive standard solutions at 5, 10, 20, 30 and 40 ng/l. The recovery rates from the paper chad were determined by correlating the response achieved during the liquid analysis of the explosive standards at 20 ng/l and the vapour extraction of the standards in the same concentration (n = 4). 3. Results and discussion
2.5. Vapour analysis 3.1. Liquid analysis After the extraction process, the paper chad was retrieved from the vial cap, folded twice in order to fit into the sample reservoir, and placed directly into the lab on a chip injection well containing 5 L of the background electrolyte. 4 L of electrolyte was added to obtain a final volume of 9 L. After all wells had been filled with 9 L
The samples were electrokinetic injected, electrophoreticaly separated, and detected by fluorescence. This involves the application of high voltage in the system, which creates an electric field for CE separations [26]. The addition of SDS to the background
470
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
Fig. 4. Electropherogram of the separation of 9 l of a 20 ng/L mixture of eight explosives diluted into buffer, pipetted directly into a sample well of the microchip. 1) TNB; 2) 1,3-DNB; 3) TNT; 4) Tetryl; 5) 2,4-DNT; 6) 3-NT; 7) 2,6-DNT, 8) 2-A-4,6DNT. Background electrolyte (10 mM Borate : 50 mM SDS; pH 9,2), injection 1500 V for 40 s, separation 1500 V for 100 s.
electrolyte allows separation of the explosives, which are neutral molecules. In the method known as micellar electrokinetic chromatography (MEKC), the addition of a surfactant to the electrolyte forms micelles, which act as a pseudo stationary phase to allow the separation of neutral molecules [25]. The indirect detection of the samples caused by fluorescence quenching is due to the addition of dye to the background electrolyte, which causes the explosives to appear as negative peaks in the electropherograms. The background electrolyte produced consistent baselines across all standards and sample runs. It was possible to separate, detect and identify eight explosives in a single run (Fig. 4) using liquid analysis. 9 L of a 20 ng/L solution containing TNT, DNB, TNB, Tetryl, 3-NT, 2,4-DNT, 2-A-4,6DNT, and 4-A-2,6-DNT diluted in buffer (10 mM sodium tetraborate 10 mM and 50 mM SDS; pH 9.2) were pipetted directly into the sample well of a previously primed microchip. All other wells were filled with 9 L of samples or buffer, and the microchip was inserted into the instrument for liquid analysis. The background electrolyte comprised of 10 mM sodium tetraborate, 50 mM SDS, at pH 9.2, and 2% v/v of Agilent DNA 1000 dye® . Liquid analysis calibration curves were constructed by pipetting 9 L of the explosives standards directly into the microchip sample wells at 1, 5, 10, 15 and 20 ng/L. The respective LOD’s were calculated as 3.3 / slope of the calibration curve, where is the slope deviation [32]. The minimum detectable masses were calculated as 9 times the LOD once 9 L of each standard was pipetted directly into the sample wells per run. The achieved values for liquid and vapour analysis as well as the recovery rates of the explosives after vapour extraction can be seen in Table 1. 3.2. Vapour extraction Following the liquid extraction results, TNB, Tetryl and TNT were chosen for the vapour analysis experiments. These three standards were chosen due to their retention times, sensitivity and availability. In order to facilitate the evaporation of the explosive standards, organic solvents were added to the mixture. After the targets for the extraction experiments were chosen, different organic solvents were compared. Methanol, distilled water, acetone and acetonitrile were added individually to the explosive solutions resulting in equimolar mixtures of 40 ng/L of each explosive standard per solvent. The mixtures were left to evaporate for 15 min at room temperature (between 22 and 25 degrees), 40, 60, and 80 ◦ C in a dry heat block. This experiment aimed to determine the best solvent and temperature to achieve the optimal extraction process. The results show that, for the three explosives, the optimal tem-
perature and solvent were 80 ◦ C and acetone, respectively (Fig. 5 a–c). Results show that, in vapour samples, the solvent with the lowest evaporation temperature was the most effective in the collection of explosive standards residues. In the comparison between acetone (56 ◦ C), methanol (64 ◦ C), acetonitrile (82 ◦ C), and water (100 ◦ C), acetone proved to be the most conducive for this experiment. This result is probably due to the higher concentration of analytes delivered per time provided by the solvent with more volatility. The higher volatility of acetone represents a higher flow of this solvent through the paper chad than the other solvents tested, facilitating the transportation of explosive residues. Although no comparison could be found in the literature for this type of extraction, previous literature results vary and methanol has been suggested for recovering explosive residues from soil [28], while acetone was used for recovery and post-blast analysis of organic explosives [33], and a review observed the use of both methanol and acetone for explosives detection in different media [34]. In the experiment described in this study, acetone was the best solvent for the recovery of explosives from vapour samples due to its high volatility. As previously described the higher volatility of this solvent may have aided in carrying a higher amount of explosives residues per time due to its higher flow when heated. Another theory would be that due to its higher volatility acetone was easily removed from the paper chad after transporting the vapours, while less volatile solvents, which present a higher evaporation temperature may have remained with the explosive residues in the filter paper acting as masking agents. 3.3. Vapour analysis The vapour analysis lead to clear negative peaks resulting from the indirect detection of the analytes after fluorescence quenching. The three target explosives were successfully separated, detected and identified using the developed method for vapour extraction and analysis. The explosives were diluted in acetone to 30 ng/L.150 L of the mixed solution was placed into a 2 mL glass vial. The vial was covered with a cap containing a paper chad and heated to 80 ◦ C for 15 min in a dry block, achieving full evaporation of the liquid. The paper chad was retrieved from the plastic cap, folded in two and directly placed in a sample well of a previously primed chip, which contained 4 L of buffer. After the paper chad insertion in the sample well, 5 L of buffer were added. All the other wells were filled with 9 L of samples or buffer before the run could take place. Recovery rates for the target explosives were found by comparing the peak area of the four replicate analyses of 20 ng/L TNB, TNT and Tetryl samples after liquid and vapour analysis and are shown in Table 1. Fig. 6 shows the comparison between the peak responses for both analyses. The electropherograms show the separation of three explosives in a 20 ng/L mixture. While the liquid analyses were performed by pipetting 9 L of the explosive mixture in buffer directly into the sample well of the microchip, the vapour extraction analyses were performed after 150 L of the explosive mixture in acetone was evaporated, with the aid of a dry heat block, onto a 0.5 cm filter paper. The filter paper was then folded twice and inserted in the sample well of the microchip, avoiding any other extraction steps. For both experiments, all other wells were filled with samples or buffer. The background electrolyte was composed of buffer plus dye at 2% (v/v). As can be seen in Fig. 6 the separation of the three compounds was clear and the baseline was consistent on both analysis methods. The vapour analysis showed a delayed separation when compared to liquid analysis, which can be observed by comparing the time elapsed between the electroosmotic flow peak and the explosives peaks on both runs. The same baseline noise can be observed on both electrophero-
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
471
Table 1 Values for the average retention time and minimum detectable masses for direct injection and vapour extraction in the paper chad, and average recovery rates for the paper chad when compared to the direct injection (n = 4). EXPLOSIVES
TNB TNT TETRYL
LIQUID ANALYSIS
VAPOUR ANALYSIS
Average Retention Time (s)
Minimum Detectable mass (ng)
Average Retention Time (s)
Minimum Detectable mass (ng)
Average Recovery (%)
4.72 8.00 10.50
2.32 ± 0.7 3.25 ± 0.4 2.35 ± 0.4
5.36 8.81 10.23
6.03 ± 4.5 9.99 ± 2.3 14.22 ± 1.2
47.00 75.00 29.00
Fig. 5. Response analysis of 40 ng/L target explosive samples for different organic solvents at increasing evaporation temperature and constant time (15 min) after extraction in the paper chad.
Fig. 6. LOC analysis of explosive standards using (I) direct injection of liquid standards and (II) vapour extraction of liquid standards in filter paper. Explosives used: 1.TNB; 2. TNT; 3. TETRYL at 20.ng/L. Background electrolyte (10 mM Borate : 50 mM SDS; pH 9,2), injection 1500 V for 40 s, separation 1500 V for 100 s.
grams, although it is more evident in the vapour analysis due to the peak size of the standards. The peak size on the vapour extraction analysis was considerably smaller compared to the liquid injection analysis. This lower yield is to be expected due to the affinity of the target molecules for the mobile phase (acetone) and stationary phase (filter paper). Additionally, in the vapour extraction analysis, consideration is required for sample degradation due to heating, as well as the presence of organic solvent in the filter paper, which can act as a masking agent. Moreover, in this type of experiment we were not able to vaporize one hundred percent of the explosive sample, as observed by the coloured residue present in the bottom of the vial after the liquid content had been fully evaporated. Limits of detection and the minimum detectable masses for the three target explosives were calculated. Calibration curves were constructed for each explosive after vapour extraction from solutions at 10, 20, 30, and 40 ng/L. After the vapour extraction process the paper chads were folded and placed directly into the microchip sample wells. The LOD’s were calculated as 3.3/slope of the calibration curve where = error in the slope [32]. The three explosives targeted for the vapour extraction and detection of residues in the headspace were readily detected and visualised using fluorescence
quenching. The minimum detectable amounts were 6.03, 9.99, and 14.22 ng for TNB, TNT and Tetryl, respectively. The recoveries from paper chads after vapour extraction, when compared to the direct injection were 47, 75, and 29% for TNB, TNT and Tetryl, respectively. The minimum detectable masses of the selected explosives and the recovery rates from the paper chads can be seen in Table 1. Intra and interday variations for the vapour extraction process were determined over a three day period by comparing the peak area of the four replicate analyses. The intraday variation across the three day period was 7.9% RSD for TNB; 3.22% RSD for TNT; and 4.5% RSD for Tetryl. The interday variation was RSD: 4.7% (TNB); 1.7% (TNT); and 1.2% (Tetryl). All three target explosives were successfully detected after the vapour extraction procedure, and the achieved LOD’s are comparable to the conventional methods used on bench CE instruments [35] proving the efficiency of microchip models. The proposed extraction process of vapours from explosive standards onto a filter paper chad proved to be suitable for the detection of nitroaromatic explosives. As can be seen in Table 1 the recovery rates for the three target explosives are similar to those described in the literature using different methods [28,36]. The proposed method was able to achieve recovery rates as high as 75% for TNT, which can be compared to studies conducted by Batlle et al. [36] that were able to achieve recovery rates of 86 and 90% of TNT in acetone using HPLC to determine explosives in vapour phases. Ueland et al. [28] described recovery rates of explosives from soil samples, using paper chads analysed by the Bioanalyzer, ranging from 53 to 82%. In addition, the insertion of the paper chad directly into the sample well did not block the channel or interfere with the sample separation, avoiding an extra step for the pre-concentration or extraction of the explosives. Directly inserting the paper chad into the chip sample wells represents a significantly simplified extraction process. Results show that a minimum temperature of 40 ◦ C is necessary to vaporize the compounds using this solvent. Nonetheless, the optimized extraction process for this preliminary study has shown positive results as demonstrated by the LODs and recovery rates achieved. The lower response observed from Tetryl during the vapour extraction experiment can be explained by different factors. The negative peaks observed in the electropherograms are due to the fluorescence quenching of the dye present on the
472
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
background electrolyte [27,37,38], hence explosives with higher fluorescence quenching power, such as TNB and TNT will have a higher impact in this experiment than nitramines, such as Tetryl and RDX, which proved to yield a lower quenching power [37]. Moreover, nitramines are known to be less volatile and sensitive than the nitroaromatics [2], which significantly impacts the amount of explosive residues present in the headspace and can explain why Tetryl was readily detected during liquid injection experiments but the same was not seen in the vapour extraction method. The current method was optimized for the analysis of three target explosives, which were successfully detected, separated and identified after a fast and simple process of extraction, using low cost equipment and an innovative platform. The ability of this instrument to analyse explosive mixtures, rapidly and at relatively low costs demonstrates the potential of this method for explosive residues detection in gas and vapour samples. Air samples analysis of explosives is of pivotal importance for military and anti-terrorist actions [11] and this study shows an innovative and promising technique for the detection of explosive residues in the headspace. However it is important to note that given the respective LODs and considering the saturated vapour pressure for each of the three target compounds the amount of air to be sampled in a real case scenario using the current method would be considerably large. For example TNT which had a minimum detectable mass from vapour analysis calculated to 9.9 ng and a saturated vapour pressure at 25 ◦ C of 9.15 ppbv [7] would need approximately 100 mL of air sampled at 100% efficiency. Considering that trace explosives vapours in the environment are generally calculated at concentrations of at least two orders of magnitude below the saturated vapour concentration, the volume required for the correct detection of TNT would be 10 L. Applying the same concept for the other two target analytes a total of 4 L would need to be sampled for the correct analysis of TNB (saturated vapour pressure at 25 ◦ C 20 ppbv [7]) and approximately 1400 L would have to be sampled for the correct analysis of Tetryl (saturated vapour pressure at 25 ◦ C 0.0074 ppbv [7]). Those values indicate the need for a pre-concentration procedure. By utilizing an easy to acquire paper chad, which can be directly inserted into the analysing instrument, we were able to simplify the pre-concentration and extraction methods. Moreover, with the use of a paper chad we were able to achieve a fast and reliable pre-concentration process without the need for ionization or other complex procedures. It is important to consider that this is a preliminary test, which has showed the potential and suitability of microchip electrophoresis analysis for the detection of explosives in the gas and vapour samples. The results point to the viability of the use of this method in future portable instruments for field work as the proposed platform with the use of a LOC device presents only one preconcentration step (a simple paper chad), does not need complex procedures for the analytes extraction and has decades of research behind it.
4. Conclusion A simple, innovative, rapid and cost-effective method to collect explosive vapour residues was developed using a 2 mL glass vial topped with a 0.5 m filter paper chad. The paper chad was used as a pre-concentration device during the extraction of vapours from explosives. This technique presents itself as a viable alternative to eliminate complex procedures, keeping the selectivity, sensitivity and reliability of a well-known method such as microchip-CE. When compared to IMS and MS detection techniques the proposed platform presents itself as a better alternative as it does not requires ionization of the analytes. The analyte ionization in a complex real life scenario can be problematic due to competitive ionization path-
ways with interferents, which can lead to a reduction in selectivity and sensitivity. Also, the use of a LIF detector presents a clear advantage over other instruments, as the quenching mechanism is selective to nitro-containing analytes, improving the selectivity for target analytes over interferents. When compared to other fluorescence quenching technologies the proposed platform again holds an advantage as the separation allows for another dimension of selectivity leading to a potentially enhanced sensitivity. This method was able to detect and quantify vapour emitted from explosive standards after evaporation. The extraction process used minute amounts of standards and reagents, and low cost materials. The explosives were detected through fluorescence quenching and separation occurred in under 20 s, representing a fast and accurate approach for the detection of explosives in vapour samples. Moreover this method has proved to be reliable as the results were replicable and the intraday values were within the expected normal values, even when compared to other instruments such as HPLC [39]. The use of a filter paper to collect explosives residues in the headspace provides an inexpensive approach for the detection of explosives. The Bioanalyzer presents itself as a viable alternative for the analysis of explosive mixtures, rapidly, reliably and at relatively low cost using small amounts of standards and reagents. The LOC device does not require pumps or gas for its operation and can analyse up to 12 samples in under 1800 s. Its low weight, small size and robust nature demonstrates excellent potential for portability and field deployment. It is important to highlight that the explosive threat is not one from a diluted explosive standard on a laboratory bench but instead, from the concealed real explosives in civil areas. Explosives are known to have low volatility and because of that, are incredibly difficult to detect in the headspace [40]. Concealed explosives are even more problematic to detect, as the amount of vapours in the air for detection are limited [8]. Therefore, this experiment in no way reflects detection of concealed explosives in real scenarios as the paper chads are placed only centimetres away from the liquid standards, maximizing the amount of vapours extracted in the pre-concentration step. Heating the liquid standards also considerably raises their volatility facilitating the analysis and analytes detection. Nevertheless, this experiment demonstrates a new and promising platform that can be used for vapour analysis and should be explored more thoroughly. The next development steps will focus on improving the detection limits of the instrument and optimizing the extraction procedure and pre-concentration step. After the instrument detection capabilities have been properly unveiled, experiments will be conducted with real samples using the same pre-concentration step, also aiming to improve the robustness of the platform. Solid samples of real explosives will be stored into sealed containers and air samples will be extracted by vacuum pump onto the filter paper chad. Analysis will be conducted with a LOC instrument and LIF detection to establish whether it is possible to detect explosives from air samples using this proposed platform and extraction method. The ultimate goal is to create a platform that does not need the pre-concentration step and can constantly perform sampling and analysis. The vapour and gases would be fan forced drawn into the instrument which would constantly run the analysis of the air samples. To the best of our knowledge, this is the first report of a combination of filter paper chads and LOC instruments to extract and directly analyse explosive residues from vapour samples. This approach provides a new and promising method to identify explosives in the headspace.
Conflicts of interest There are no conflicts of interest to declare.
V. Taranto et al. / J. Chromatogr. A 1602 (2019) 467–473
Acknowledgements The corresponding author is supported by a Science Without Borders Scholarship of the National Council for Scientific and Technological Development of the Government of Brazil (CNPq).201677/2014-8. The opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors, and the financial support by the Government ministry does not constitute an express or implied endorsement of the results or conclusions of the research. References [1] J.S. Caygill, F. Davis, S.P. Higson, Current trends in explosive detection techniques, Talanta 88 (2012) 14–29. [2] D.S. Moore, Instrumentation for trace detection of high explosives, Rev. Sci. Instrum. 75 (8) (2004) 2499–2512. [3] J. Feltes, et al., Gas chromatographic and mass spectrometric determination of nitroaromatics in water, J. Chromatogr. A 518 (1990) 21–40. [4] D. Badgujar, et al., Advances in science and technology of modern energetic materials: an overview, J. Hazard. Mater. 151 (2–3) (2008) 289–305. [5] Y.T. Yew, A. Ambrosi, M. Pumera, Nitroaromatic explosives detection using electrochemically exfoliated graphene, Sci. Rep. 6 (2016) 33276. [6] M. Krausa, A.A. Reznev, in: N.S. Series (Ed.), Vapour and Trace Detection of Explosives for Anti-Terrorism Purposes, vol. 167, Springer Science & Business Media, 2004. [7] R.G. Ewing, et al., The vapor pressures of explosives, Trac Trends Anal. Chem. 42 (2013) 35–48. [8] S. Singh, Sensors—an effective approach for the detection of explosives, J. Hazard. Mater. 144 (1–2) (2007) 15–28. [9] M.M. Houck, J.A. Siegel, Fundamentals of Forensic Science, Academic Press, 2009. [10] Y.H. Lee, et al., Dipyrenylcalix [4] arene—a fluorescence-based chemosensor for trinitroaromatic explosives, Chem. Eur. J. 16 (20) (2010) 5895–5901. [11] A. Hakonen, et al., Explosive and chemical threat detection by surface-enhanced Raman scattering: a review, Anal. Chim. Acta 893 (2015) 1–13. [12] R.G. Ewing, et al., A critical review of ion mobility spectrometry for the detection of explosives and explosive related compounds, Talanta 54 (3) (2001) 515–529. [13] R.M. Alberici, et al., Ambient mass spectrometry: bringing MS into the “real world”, Anal. Bioanal. Chem. 398 (1) (2010) 265–294. [14] S.J. Toal, et al., Visual detection of trace nitroaromatic explosive residue using photoluminescent metallole-containing polymers, J. Forensic Sci. 52 (1) (2007) 79–83. [15] E.T. da Costa, et al., Unmanned platform for long-range remote analysis of volatile compounds in air samples, Electrophoresis 33 (17) (2012) 2650–2659. [16] Q. Lu, et al., Sensitive capillary electrophoresis microchip determination of trinitroaromatic explosives in nonaqueous electrolyte following solid phase extraction, Anal. Chim. Acta 469 (2) (2002) 253–260. [17] J. Wang, et al., Single-channel microchip for fast screening and detailed identification of nitroaromatic explosives or organophosphate nerve agents, Anal. Chem. 74 (5) (2002) 1187–1191.
473
[18] P. Gravesen, J. Branebjerg, O.S. Jensen, Microfluidics-a review, J. Micromechanics Microengineering 3 (4) (1993) 168. [19] V. Dolnik, S. Liu, Applications of capillary electrophoresis on microchip, J. Sep. Sci. 28 (15) (2005) 1994–2009. [20] E. Verpoorte, Microfluidic chips for clinical and forensic analysis, Electrophoresis 23 (5) (2002) 677–712. [21] M. Pumera, Analysis of explosives via microchip electrophoresis and conventional capillary electrophoresis: a review, Electrophoresis 27 (1) (2006) 244–256. [22] J. Wang, Microchip devices for detecting terrorist weapons, Anal. Chim. Acta 507 (1) (2004) 3–10. [23] J.R. Petersen, A.A. Mohammad, Clinical and Forensic Applications of Capillary Electrophoresis, Springer Science & Business Media, 2001. [24] D.R. Baker, F. Foret, Capillary electrophoresis: techniques in analytical chemistry series, Anal. Biochem. 231 (1) (1995) 274. [25] S.A. Oehrle, Analysis of nitramine and nitroaromatic explosives by capillary electrophoresis, J. Chromatogr. A 745 (1–2) (1996) 233–237. [26] A. Lloyd, et al., A rapid method for the in-field analysis of amphetamines employing the Agilent Bioanalyzer, Anal. Methods 3 (7) (2011) 1535–1539. [27] A. Pesenti, et al., Coupling paper-based microfluidics and lab on a chip technologies for confirmatory analysis of trinitro aromatic explosives, Anal. Chem. 86 (10) (2014) 4707–4714. [28] M. Ueland, et al., Capillary-driven microfluidic paper-based analytical devices for lab on a chip screening of explosive residues in soil, J. Chromatogr. A 1436 (2016) 28–33. [29] A.W. Martinez, et al., Patterned paper as a platform for inexpensive, low-volume, portable bioassays, Angew. Chemie Int. Ed. 46 (8) (2007) 1318–1320. [30] A.K. Yetisen, M.S. Akram, C.R. Lowe, Based microfluidic point-of-care diagnostic devices, Lab Chip 13 (12) (2013) 2210–2251. [31] A. Lloyd, The Analysis of Amphetamine-type Stimulants Using Microchip Capillary Electrophoresis, 2013. [32] A. Shrivastava, V. Gupta, Methods for the determination of limit of detection and limit of quantitation of the analytical methods, Chron. Young Sci. 2 (1) (2011) 21. [33] S. Calderara, D. Gardebas, F. Martinez, Solid phase micro extraction coupled with on-column GC/ECD for the post-blast analysis of organic explosives, Forensic Sci. Int. 137 (1) (2003) 6–12. [34] J.I. Steinfeld, J. Wormhoudt, Explosives detection: a challenge for physical chemistry, Annu. Rev. Phys. Chem. 49 (1) (1998) 203–232. [35] J.P. Hutchinson, et al., Identification of inorganic improvised explosive devices by analysis of postblast residues using portable capillary electrophoresis instrumentation and indirect photometric detection with a light-emitting diode, Anal. Chem. 79 (18) (2007) 7005–7013. [36] R. Batlle, et al., On-line coupling of supercritical fluid extraction with high-performance liquid chromatography for the determination of explosives in vapour phases, J. Chromatogr. A 963 (1–2) (2002) 73–82. [37] R.V. Taudte, et al., A portable explosive detector based on fluorescence quenching of pyrene deposited on coloured wax-printed PADs, Lab Chip 13 (21) (2013) 4164–4172. [38] A. Ponnu, E.V. Anslyn, A fluorescence-based cyclodextrin sensor to detect nitroaromatic explosives, Supramol. Chem. 22 (1) (2010) 65–71. [39] R.V. Taudte, et al., Development of a UHPLC method for the detection of organic gunshot residues using artificial neural networks, Anal. Methods 7 (18) (2015) 7447–7454. [40] J. Yinon, Counterterrorist Detection Techniques of Explosives, Elsevier, 2011.