Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2020 Jul 14;5(29):18073–18079. doi: 10.1021/acsomega.0c01615

Nanoporous Silica Preconcentrator for Vapor-Phase DMNB, a Detection Taggant for Explosives

Coco Day †, Nathan Rowe ‡, Tanya Hutter §,∥,*
PMCID: PMC7391368  PMID: 32743181

Abstract

graphic file with name ao0c01615_0008.jpg

The detection of trace amounts of explosives in the vapor phase is of great importance. Preconcentration of the analyte is a useful technique to lower the detection limit of existing sensors. A nanoporous silica (pSiO2) substrate was evaluated as a preconcentrator for gas-phase 2,3-dimethyl-2,3-dinitrobutane (DMNB), a volatile detection taggant added by law to plastic explosives. After collection in pSiO2, the DMNB vapor was thermally desorbed at 70 °C into a gas chromatography–mass spectrometry sorbent tube. This was analyzed for the total mass of DMNB collected in pSiO2. The loading time and loading temperature of pSiO2 were varied systematically between 15 and 60 min and 5–20 °C, respectively. The preconcentrator’s performance was compared to that of a nonporous substrate of the same material as a control. The collection efficiency of pSiO2 was calculated as approximately 20% of the total DMNB that passed over it in 30 min, at a concentration of 0.5 ppm in N2 carrier gas. It had enhancement factors compared to the nonporous substrate of 12 and 16 for 0.5 and 4.1 ppm DMNB, respectively, under the same conditions. No advantage was found with cooling pSiO2 below room temperature during the loading phase, which removes any need for a cooling system to aid preconcentration. The low desorption temperature of 70 °C is an advantage over other preconcentration systems, although a higher temperature could decrease the desorption time.

1. Introduction

With the rise of global terrorism, as well as of other security and military applications, the reliable detection of explosives has risen greatly in importance over the last several decades. The commonly used trace detection techniques include colorimetric sensors, which have the advantage of being portable and hand-held. However, they have significant disadvantages, such as the high incidence of false positives and negatives.1 Another widely used technique is ion mobility spectrometry (IMS), currently employed for the trace particle and vapor detection of explosive compounds.2 However, the swabbing and transfer of particles into the spectrometer are done manually,3 providing a potential for human error.

Vapor-phase detection is particularly useful for continuous, long-term monitoring and requires less human action than the currently common techniques detailed above. However, many explosives have very low vapor pressures, making their detection in the vapor phase difficult.4 A common vapor detection method is by the use of dogs, the detection limit of which is very low; however, they have high through-life costs in comparison with equipment.1

2,3-Dimethyl-2,3-dinitrobutane (DMNB; Figure 1) is the most widely used of three detection taggants, which are added, by law, into plastic explosives during their manufacture.5,6 Such taggants, although present in small quantities, have a much higher vapor pressure than the explosives themselves and therefore are ideally suited for vapor-phase detection.4 Currently, however, such detection of DMNB is relatively immature.

Figure 1.

Figure 1

Open in a new tab

Structure of the DMNB molecule.

Preconcentration techniques for the enhancement of signals are extremely useful in vapor-phase detection. This is particularly the case in large, open spaces, where the concentration of even relatively high-vapor-pressure compounds is low. In this work, nanoporous silica, which has been studied for many decades as a sensor material,7 is used as the preconcentrator. Produced by in-house etching of crystalline silicon wafers,8 this material consists of a network of nanometer-scale interconnected pores, with a very high surface area, which makes it adept at trapping analytes.

Silicon and silica are most often used in preconcentration applications as a matrix to hold an adsorbent. For example, devices including a micromachined silicon grid and silica nanoparticles have been used to hold carbon-based adsorbents for the preconcentration of benzene and other volatile organic compounds, respectively.9,10 The preconcentration of orthonitrotoluene has also been done using various adsorbents packed into a silicon micropillar array, achieving maximum preconcentration factors, compared to a system without adsorbents, of the order of 10.11

Porous silicon, also, has been used for a similar purpose: Camara et al. have published several papers on carbon (in the form of powder, nanotubes, or graphitized carbon) deposited inside porous silicon microchannels. This material is then used as a preconcentrator for gaseous benzene and nonexplosive aromatic nitro compounds, such as nitrobenzene and dinitrotoluene (DNT).12−14 A preconcentration factor (the ratio between the maximum measured concentration during desorption and the injected concentration) of about 60 was achieved for benzene; the preconcentration factors for other compounds were not reported.14 High desorption temperatures between 200 and 300 °C were required throughout.

Similarly, Tudisco et al. have functionalized a porous silica surface with an aromatic selective receptor called “EtQxBox” for the selective adsorption of nitrobenzene. Although this material seems promising, it has not yet been developed into a preconcentrator or integrated into a sensor system.15

However, the authors believe that the use of porous silica itself as a preconcentrator for the vapor-phase detection of DMNB has never been reported at the time of writing. The only group that appears to have used plain silicon for a similar purpose is that of Giordano et al. that fabricated a silicon nanowire array to be used as a preconcentrator for DNT and trinitrotoluene. They achieved detection limits of the order of 0.1 ppb using mass spectrometry (MS), although the preconcentration factor was not reported.16

Different preconcentration systems that do not include porous silicon or silica have also been investigated for the detection of DMNB and other explosives. The Almirall group investigated the addition of solid-phase microextraction (SPME) to an IMS system.3,17 Volatile compounds are extracted by absorption onto a polymeric coating, placed inside the injection port of the IMS system, and desorbed by heating to high temperatures. For DMNB, the limit of detection of the SPME–IMS system was calculated to be 0.31 ng, sampled from a quart-sized can, compared to 2.23 ng from a traditional filter paper IMS method. A temperature of 260 °C was found to be optimal for the desorption of DMNB.3

In a later paper by the same group, gas chromatography–MS (GC–MS) analysis was used to quantify the amount of analytes absorbed onto the polymer-coated fiber used for SPME.17 Sampling times of the order of minutes allowed the collection of 20 ng of DMNB from the headspace of a vial containing 1 g of commercial plastic explosives; this was approximately 10 times the limit of detection of the IMS sensor. However, the vapor-phase concentration of DMNB was not measured, and therefore the collection efficiency of the fiber was not calculated; it is not known how it would perform in a larger space, with a lower DMNB concentration. Preconcentration factors were not calculated in either paper.

The Zellers group used a graphitized carbon-based preconcentrator to detect DMNB in combination with a microfabricated GC column.18,19 Although this system had a low limit of detection, described as “sub-ppb concentrations”, the setup was complicated, with many separate microfabricated parts. In addition, thermal desorption was carried out at 250 °C, a relatively high temperature. A preconcentration factor of 4500 was reported for DMNB, although this is not directly comparable to the factors reported for other technologies above, as it was calculated differently. This preconcentration factor was defined as the ratio of the volume of the N2 carrier gas containing DMNB that passed through the microfocuser to the peak generated by the microfocuser.18

A different preconcentration technique was used by Gruznov et al., consisting of a woven metal wire grid, also in combination with a chromatography column.20 It was found that a relative humidity of 26% decreased the amplitude of the GC peak for DMNB by 7 times, and thereafter dry air was used for their measurements. A limit of detection of 0.1 pg/cm3 was achieved, compared to 2.7 pg/cm3 without the preconcentrator, although the preconcentration factor was not calculated and desorption conditions were not stated.

In this work, the capacity of a nanoporous silica (pSiO2) preconcentrator for the DMNB vapor was quantified. The preconcentrator can be simply and cheaply produced, and although in this work it was quantified using GC–MS, it would equally be compatible with integration into the existing IMS systems. Desorption was carried out at 70 °C, much lower than that in most of the existing literature.

DMNB vapor of known concentration was passed over the pSiO2 preconcentrator in a flow cell for a given length of time. During this time, pSiO2 was held at a specific temperature by a Peltier heating/cooling device. Then, DMNB was thermally desorbed by heating pSiO2 to 70 °C into a GC–MS sorbent tube, which was then analyzed to find the total amount of DMNB stored in it. The effects of varying the loading time and temperature of the preconcentrator were investigated, as well as that of varying the vapor concentration of DMNB to which the preconcentrator was exposed. A nonporous piece of surface-oxidized silicon was used as a control.

2. Results and Discussion

The procedure followed for the collection of the sample is outlined below:

  • 1.

    The output of the preconcentrator-containing flow cell was connected directly to the exhaust; no sorbent tube was attached.

  • 2.

    The temperature of the silica substrate was set via the Peltier device; DMNB in N2 at a fixed concentration was passed through the cell for a defined amount of time (loading phase).

  • 3.

    The valve was switched, so that pure N2 passes through the cell; a sorbent tube and a pump were attached to the output from the cell. The pump was switched on for 10 min, while the silica substrate temperature was increased at a rate of 30 °C/min to 70 °C (release phase).

  • 4.

    The pump was switched off, and the sorbent tube was removed and immediately capped; the output of the silica flow cell was connected back to the exhaust.

  • 5.

    The silica substrate was kept at 70 °C for a further 50 min to be cleaned completely.

This procedure was repeated for each sorbent tube collected. A photograph of the setup can be seen in Figure 2 and another image taken closer to the cell and the sorbent tube is shown in Figure S1 in the Supporting Information.

Figure 2.

Figure 2

Open in a new tab

Photograph of the gas generation setup.

The experimental procedure outlined above was followed under systematically varied conditions, including DMNB concentrations, loading times, and loading temperatures. The concentration used was either 0.5 or 4.1 ppm; the loading time was varied from 15 to 60 min and the loading temperature from 5 to 20 °C. The values of the parameters tested are given in Table 1.

Table 1. Summary of the Experimental Parameters.

sample no. substrate type DMNB vapor conc./ppm loading time/min loading temp./°C
1 porous 0 30 5
2 porous 0.5 15 5
3 porous 0.5 30 5
4 porous 0.5 60 5
5 porous 0.5 30 10
6 porous 0.5 30 20
7 porous 4.1 30 5
8 porous 4.1 30 20
9 nonporous 0.5 30 5
10 nonporous 0.5 30 10
11 nonporous 0.5 30 20
12 nonporous 4.1 30 5
13 nonporous 4.1 30 20
Open in a new tab

The collected sorbent tubes were analyzed by GC–MS to find the total mass of DMNB contained inside them, which corresponds to the amount of DMNB desorbed from the pSiO2 preconcentrator in the 10 min release phase. The raw data are given in Table S1 in the Supporting Information.

Each sample collection was repeated twice, to check reproducibility, and the average was taken and plotted. The standard deviation is shown in the following graphs as error bars. However, there were two samples (2 and 9), where only one collection was made; in these cases, no error could be calculated.

2.1. Effect of Loading Time

Figure 3 shows the effect of loading time on the mass of DMNB collected. The mass increases with loading time, within the range of times tested. It is expected that after a certain loading time—when the system reaches equilibrium—a plateau value would be reached and the preconcentrator would cease to collect more DMNB.

Figure 3.

Figure 3

Open in a new tab

Effect of loading time on the DMNB mass collected in the first 10 min of desorption, after loading of 0.5 ppm DMNB at 5 °C (samples 2, 3, and 4).

The error in the mass of DMNB collected was fairly large, especially for the 60 min loading. This is likely to be due to human error in the attachment of the pump and sorbent tube; although it was done as quickly as possible, some of the collected DMNB may have been “lost” in the N2 flow before the sorbent tube was added.

2.2. Effect of DMNB Concentration

The concentration of DMNB in the N2 carrier gas that passed over the preconcentrator was also varied; the results are shown in Figure 4. In order to make a comparison, the pSiO2 preconcentrator was removed from the flow cell and replaced by a nonporous piece of surface-oxidized silicon wafer, of the same type and dimensions as that of the piece that was used to make the preconcentrator. This provided simply a flat surface in the place of the preconcentrator, with a much lower surface area and no porosity, but of the same material. The effect of concentration in this case is also shown.

Figure 4.

Figure 4

Open in a new tab

Effect of the DMNB vapor concentration on the mass of DMNB collected in the first 10 min of desorption, after loading at 5 °C for 30 min (samples 1, 3, 7, 9, and 12). The dashed line denotes the expected decrease in mass collected by the flat surface to 0 g at 0 ppm; this data point was not actually taken.

Although there are only three data points, this graph shows a nonlinear relationship, suggesting that at high concentrations of DMNB, the mass collected by the preconcentrator will plateau. The mass of DMNB collected on the flat surface appears to plateau at a lower concentration than that on the porous surface.

2.3. Effect of Preconcentrator Temperature

The temperature of the preconcentrator during the loading phase was varied, and the amount of DMNB adsorbed was compared to that of the nonporous surface of the same material. The results are shown in Figure 5. On the flat surface, approximately 5 times as much DMNB was deposited at 5 °C than at 10 and 20 °C, at both concentrations. This suggests that at 5 °C, the DMNB gas that passes close to the cold surface is above its saturation vapor pressure, giving a driving force for deposition. At higher temperatures, DMNB is below its saturation vapor pressure, and therefore only physical adsorption occurs.

Figure 5.

Figure 5

Open in a new tab

Effect of loading temperature on the DMNB mass collected in the first 10 min of desorption, for porous and nonporous substrates, at different DMNB concentrations, with a 30 min loading time (samples 3, 5, 6, 7, 8, 9, 10, 11, 12, and 13). No data were collected for 4.1 ppm DMNB at 10 °C.

Unfortunately, the saturation vapor pressure of DMNB is difficult to be determined accurately, and therefore, values from the literature vary: one equation suggests 12.5 ppm at 20 °C,21 whereas another source gives 2.7 ppm at 25 °C.22 However, it is clear that the concentrations used in this work are close to the saturation vapor pressure at room temperature.

In order to generate DMNB vapor, solid DMNB is heated in a vapor generator to 75 or 100 °C, but the gas is likely to cool significantly while passing through the tubing, before reaching pSiO2. In addition, it flows continuously over the surface, and consequently its temperature is not well-defined.

On the pSiO2 surface, significantly more DMNB was collected than on the flat surface at all temperatures, proving the effectiveness of the preconcentrator. However, no difference was observed in the adsorption of DMNB at the three temperatures. This can be explained using a kinetic argument: because of the much larger surface area to be covered, the loading phase would have to be significantly longer for multilayers to begin to build up, and therefore only physical adsorption of individual DMNB molecules occurs. The data imply that the mechanism for this physical adsorption is not strongly temperature-dependent.

The mechanisms of adsorption onto porous surfaces are not well understood. Even for ordered mesoporous materials, adsorption is heterogeneous and difficult to predict;23 this is even more true for disordered pores such as those in pSiO2 used in this work.

That said, the authors believe that the small diameter and curvature of the pores may result in a higher enthalpy of adsorption of a DMNB molecule in pSiO2 compared with a flat surface. If a molecule is released from one internal surface, it is likely to immediately collide with and adsorb onto another internal surface. The average pore size is 23 nm (see Figure S2 in the Supporting Information), which is only about 10 times the size of a DMNB molecule, and some pores are significantly smaller than this. Therefore, a majority of the DMNB molecules that pass close to the pSiO2 surface are captured and held in the pores, at all temperatures studied.

2.4. Calculations of Collection Efficiency and Capacity of pSiO2

2.4.1. Collection Efficiency of Preconcentrator

The authors believe that it would be instructive to make an approximate calculation of the collection efficiency of the preconcentrator. To this end, for sample 3, six sorbent tubes were filled, one for every 10 min interval during the release phase. This allowed for the analysis of the mass of DMNB released from the porous silica over time (while held at 70 °C). These data are given in Table S2 in the Supporting Information and are plotted in Figure 6.

Figure 6.

Figure 6

Open in a new tab

Fitted exponential decay curve and raw data (blue points) for sample 3. One measurement was taken at each release time.

To quantify this further, an exponential decay curve (eq 1) was fitted to the data and integrated to obtain the area underneath, corresponding to the total amount of DMNB collected inside the porous silica, as 2780 ng. This assumes a 100% collection efficiency of the sorbent tube.

2.4.1. 1

Furthermore, integration of this curve between 0 and 10 min was used to find the amount released in the first 10 min to be 1680 ng; this equates to approximately 60% of the total amount of DMNB within the porous silica. Approximately 84% was released within the first 20 min and approximately 94% in the first 30 min.

Equation 2 gives the total mass MT, in ng, of DMNB that has passed over the porous silica preconcentrator during a given loading time t, in min

2.4.1. 2

where F is the flow rate in mL/min; c is the concentration in ppm; Vm is the molar volume, in L/mol, of an ideal gas at room temperature and pressure; and Mm is the molar mass of DMNB in g/mol. This equation includes the assumption that the analyte gas behaves ideally, which, at the low concentrations present, is acceptable.

In the case of sample 3, in which 0.5 ppm DMNB was passed over pSiO2 for 30 min at 125 mL/min, MT = 14,380 ng. The collection efficiency of the preconcentrator can therefore be calculated using the following equation

2.4.1. 3

where MC is the total mass collected by the silica, measured by GC–MS, and MT is the total mass passed over the silica, calculated using eq 2. For sample 3, the collection efficiency is approximately 2780/14,380 = 20%.

2.4.2. Capacity of the Preconcentrator

In order to calculate the approximate capacity of the preconcentrator for DMNB, first, the total surface area of pSiO2 must be calculated. For this, the BET surface area per gram of 120.4 m2/g was used (see Section 4). The mass of an 8 μm thick layer of porous silica is calculated to be approximately 1.87 × 10–3 g, from simple geometry and using a porosity of 50%. Therefore, the total surface area of the porous layer is estimated to be 0.23 m2. Greater detail of this calculation is provided in the Supporting Information.

Using this surface area, a rough calculation indicates that the mass of DMNB needed for the monolayer coverage of the porous surface is approximately 260,000 ng. By comparison, a flat surface of the same area would require only about 200 ng to cover it. Greater detail of these calculations is also given in the Supporting Information.

For the case of sample 3 discussed above, a total of 2780 ng DMNB was adsorbed onto pSiO2; this equates to approximately 2780/260,000 = 1% surface coverage.

This calculation is approximate, and in practice, the entire surface area is unlikely to be readily accessible to the DMNB molecules. In some of the experiments with a higher DMNB concentration or a longer loading time, the mass of DMNB collected was a few times higher than that for sample 3, as can be seen in Figures 3–5. Nevertheless, the maximum pSiO2 surface coverage during these experiments was still well under a monolayer.

These calculations provide an interesting insight into some of the previous experiments. In the results in Sections 2.2 and 2.3, the mass of DMNB collected from the flat surface at 5 °C was about 1000 ng. This is greater than the 200 ng which should constitute a monolayer coverage, according to the calculations in this section. There must, therefore, be multiple layers of DMNB on the surface, providing more evidence that the DMNB molecules close to the cold surface are above their saturation vapor pressure at this temperature.

3. Conclusions

In conclusion, this work has shown that the pSiO2 preconcentrator provides a substantial advantage over a flat surface in its ability to collect DMNB. This is especially apparent at a 20 °C loading temperature, where the enhancement factor—the ratio between the mass collected on the porous surface and the mass collected on the flat surface, under the same conditions—was approximately 12 at 0.5 ppm and 16 at 4.1 ppm. Further works investigating the preconcentration ability of this material at lower concentrations, which are more likely to occur in real-world conditions, will be carried out as a next step.

The collection efficiency of the preconcentrator is approximately 20% at 0.5 ppm DMNB in N2 carrier gas, after 30 min loading. The preconcentrator has a large surface area, and under the same conditions, it is estimated that only approximately 1% of the surface was covered. At lower real-world concentrations, it could reasonably be expected to take days to become saturated with DMNB. This suggests that it might be most useful as a permanent location-based device, for example, in a room or venue, where there is a risk of an explosive being planted. Used in this manner, it would passively collect DMNB vapor until release by heating in an existing system, such as IMS. The instantaneous released concentration in this situation would be higher than the background concentration in the room, aiding detection.

Finally, decreasing the loading temperature does not appear to have an impact on the collection ability of the preconcentrator for the reasons discussed above, thus eliminating the need for a cooling system in the final device. Although the silica was only heated to 70 °C in this work because of the limitations of the Peltier heating/cooling system used, it can withstand much higher temperatures. A higher release temperature could be used in future works, if necessary, to decrease the time taken for the preconcentrator to release the stored DMNB, hence improving the measurement time.

4. Experimental Section

4.1. pSiO2 Fabrication

A boron-doped p-type silicon wafer of dimensions 25 × 25 × 0.525 mm, with a resistivity of 0.01–0.02 Ω cm and orientation (100), was etched electrochemically in a 1:1 mixture of 48% hydrofluoric acid and ethanol, following the procedures detailed in the study by Sailor, Section 2.8.8 A current density of 120 mA/cm2, applied for 115 s, resulted in a cylindrical porous layer. This layer had a diameter of 1.5 cm, thickness of approximately 8 μm [measured by scanning electron microscopy (SEM), see Figure 7], and porosity of approximately 50% (calculated by the spectroscopic liquid infiltration method8).

Figure 7.

Figure 7

Open in a new tab

Cross-sectional SEM image of a cleaved porous silica wafer.

The freshly etched porous silicon was then thermally oxidized in oxygen at 800 °C for 16 h, resulting in a porous silica layer on top of a crystalline silicon substrate. This process improves stability because an uncontrolled native oxide layer otherwise forms on the surface of unoxidized silicon.

Nitrogen isotherm analysis (3Flex Analyzer, Micromeritics) was carried out on an oxidized sample, and the surface area per gram was found by the Brunauer–Emmett–Teller adsorption isotherm analysis to be 120.4324 m2/g. The Barrett–Joyner–Halenda adsorption pore size distribution, which is shown in Figure S2 in the Supporting Information, gave the average pore diameter as 23 nm. The material, therefore, is classified as mesoporous by traditional IUPAC definitions (as its pores are between 2 and 50 nm diameter).

4.2. Vapor Generation

A diffusion tube was filled with solid DMNB (Sigma-Aldrich) and placed inside an Owlstone Vapor Generator, which heated it to 75 or 100 °C under a 125 mL/min flow of N2 gas. A gravimetric calibration was carried out to determine the concentrations of DMNB at these two temperatures to be 0.5 and 4.1 ppm, respectively. This calibration was done according to the procedure detailed in the Owlstone guide, and details are given in Figure S3 and Table S3 in the Supporting Information.24

The gas then flowed via the polytetrafluoroethylene tubing into a four-port valve. A separate line of pure N2 was also connected into the valve, which was used to switch between DMNB-containing N2 and pure N2. The outputs of this valve were an exhaust line and a sealed polyether ether ketone cell containing the pSiO2 preconcentrator, attached by a thermally conducting rubber sheet to the Peltier heating/cooling device. A rubber O-ring of 1.5 cm diameter sealed the cell and ensured that only the porous part of the surface was exposed to the gas flow. A schematic of the cell is given in Figure S4 in the Supporting Information. The overall dimensions of the device—including the heat sink, the Peltier device, and the pSiO2 preconcentrator contained in the flow cell—are approximately 60 × 60 × 45 mm. Figure S5 in the Supporting Information shows a schematic of the whole gas generation and transportation system.

4.3. GC–MS Analysis

During the loading phase, Tenax TA sorbent tubes (Markes International) were attached to the outlet of the flow cell, as illustrated in Figure S5 in the Supporting Information. A Sidekick pump (SKC Ltd.) was used to draw gas at a rate of 1000 mL/min through the sorbent tube; the excess was made up with ambient air.

After collection, the sorbent tubes were stored in a refrigerator at approximately 4 °C. The analysis was done by thermal desorption GC–MS; all the parameters used are given in Tables S4 and S5 in the Supporting Information.

The sorbent tubes were thermally desorbed, and the released gas was collected on a cryo-focusing trap. This trap was then rapidly heated to desorb the analyte gas directly onto the GC column. A split flow may also be applied to reduce the amount of analyte, to enable the analysis of high-concentration samples.

The dominant ion peak in the DMNB mass spectrum was integrated in a window around the peak retention time. Information from known calibration standards was then used to quantify the mass of DMNB in the sorbent tubes. An example chromatogram is given in Figure S6 in the Supporting Information.

Acknowledgments

Thanks to Andrew Marr and Stephen Nicklin at DSTL and Dan Wood at the Metropolitan Police Service for their advice and expertise in the field of explosives. Thanks to William Winter for his technical help with the gas setup. This project was partially funded under the Innovative Research Call in Explosives and Weapons Detection 2016. This is a cross-government programme sponsored by a number of departments and agencies under the UK government’s CONTEST strategy, in partnership with the US Department of Homeland Security, Science and Technology Directorate. C.D. is supported by a Vice-Chancellor’s Award from the Cambridge Trust. The views expressed in this publication are those of the authors and not necessarily those of the funding contributors.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c01615.

  • Close-up photograph of the flow cell, sorbent tube, and pump; raw data for 10 min release experiments; pSiO2 pore size distribution; raw data for release time experiment; details of the calculation of the porous surface area; details of the gravimetric calibration of DMNB; results of gravimetric calibration; cross-sectional schematic of the flow cell; schematic of the gas-flow setup; parameters for thermal desorption and GC–MS; example chromatogram from the GC–MS analysis (PDF)

The authors declare the following competing financial interest(s): Dr Tanya Hutter is a shareholder at SensorHut Ltd.

Supplementary Material

ao0c01615_si_001.pdf (2.8MB, pdf)

References

  1. Kangas M. J.; Burks R. M.; Atwater J.; Lukowicz R. M.; Williams P.; Holmes A. E. Colorimetric Sensor Arrays for the Detection and Identification of Chemical Weapons and Explosives. Crit. Rev. Anal. Chem. 2017, 47, 138–153. 10.1080/10408347.2016.1233805. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Thiesan L.; Hannum D.; Murray D. W.; Parmeter J. E.. Survey of Commercially Available Explosives Detection Technologies and Equipment 2004, Document No.: 208861, 2005.
  3. Perr J. M.; Furton K. G.; Almirall J. R. Solid Phase Microextraction Ion Mobility Spectrometer Interface for Explosive and Taggant Detection. J. Sep. Sci. 2005, 28, 177–183. 10.1002/jssc.200401893. [DOI] [PubMed] [Google Scholar]
  4. Ewing R. G.; Waltman M. J.; Atkinson D. A.; Grate J. W.; Hotchkiss P. J. The Vapor Pressures of Explosives. TrAC, Trends Anal. Chem. 2013, 42, 35–48. 10.1016/j.trac.2012.09.010. [DOI] [Google Scholar]
  5. Convention on the Marking of Plastic Explosives for the Purpose of Detection, 1991.
  6. Administrative Package for Ratification of or Accession to the Convention on the Marking of Plastic Explosives for the Purpose of Detection, 2005.
  7. Canham L. T. Silicon Quantum Wire Array Fabrication by Electrochemical and Chemical Dissolution of Wafers. Appl. Phys. Lett. 1990, 57, 1046–1048. 10.1063/1.103561. [DOI] [Google Scholar]
  8. Sailor M. J.Porous Silicon in Practice: Preparation, Characterization and Applications, 1st ed.; Wiley-VCH, 2011. [Google Scholar]
  9. Ivanov P.; Blanco F.; Gràcia I.; Sabaté N.; Ruiz A.; Vilanova X.; Correig X.; Fonseca L.; Figueras E.; Santander J.; Cané C. Improvement of the Gas Sensor Response via Silicon μ-Preconcentrator. Sens. Actuators, B 2007, 127, 288–294. 10.1016/j.snb.2007.07.032. [DOI] [Google Scholar]
  10. Akbar M.; Wang D.; Goodman R.; Hoover A.; Rice G.; Heflin J. R.; Agah M. Improved Performance of Micro-Fabricated Preconcentrators Using Silica Nanoparticles as a Surface Template. J. Chromatogr. A 2013, 1322, 1–7. 10.1016/j.chroma.2013.10.083. [DOI] [PubMed] [Google Scholar]
  11. Mohsen Y.; Sanchez J.-B.; Berger F.; Lahlou H.; Bezverkhyy I.; Fierro V.; Weber G.; Celzard A.; Bellat J.-P. Selection and Characterization of Adsorbents for the Analysis of an Explosive-Related Molecule Traces in the Air. Sens. Actuators, B 2013, 176, 124–131. 10.1016/j.snb.2012.10.042. [DOI] [Google Scholar]
  12. Camara M.; James F.; Breuil P.; Pijolat C.; Briand D.; de Rooij N. F. MEMS-Based Porous Silicon Preconcentrators Filled with Carbopack-B for Explosives Detection. Procedia Eng. 2014, 87, 84–87. 10.1016/j.proeng.2014.11.271. [DOI] [Google Scholar]
  13. Camara E. H. M.; Pijolat C.; Courbat J.; Breuil P.; Briand D.; de Rooij N. F.. Microfluidic Channels in Porous Silicon Filled with a Carbon Absorbent for Gas Preconcentration. International Solid-State Sensors, Actuators and Microsystems Conference. 2007; pp 249–252.
  14. Camara E. H. M.; Breuil P.; Briand D.; de Rooij N. F.; Pijolat C. A Micro Gas Preconcentrator with Improved Performance for Pollution Monitoring and Explosives Detection. Anal. Chim. Acta 2011, 688, 175–182. 10.1016/j.aca.2010.12.039. [DOI] [PubMed] [Google Scholar]
  15. Tudisco C.; Motta A.; Barboza T.; Massera C.; Giuffrida A. E.; Pinalli R.; Dalcanale E.; Condorelli G. G. Cavitand-Decorated Silicon Columnar Nanostructures for the Surface Recognition of Volatile Nitroaromatic Compounds. ACS Omega 2018, 3, 9172–9181. 10.1021/acsomega.8b01018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Giordano B. C.; Ratchford D. C.; Johnson K. J.; Pehrsson P. E. Silicon Nanowire Arrays for the Preconcentration and Separation of Trace Explosives Vapors. J. Chromatogr. A 2019, 1597, 54–62. 10.1016/j.chroma.2019.03.045. [DOI] [PubMed] [Google Scholar]
  17. Lai H.; Leung A.; Magee M.; Almirall J. R. Identification of Volatile Chemical Signatures from Plastic Explosives by SPME-GC/MS and Detection by Ion Mobility Spectrometry. Anal. Bioanal. Chem. 2010, 396, 2997–3007. 10.1007/s00216-010-3501-6. [DOI] [PubMed] [Google Scholar]
  18. Serrano G.; Sukaew T.; Zellers E. T. Hybrid Preconcentrator/Focuser Module for Determinations of Explosive Marker Compounds with a Micro-Scale Gas Chromatograph. J. Chromatogr. A 2013, 1279, 76–85. 10.1016/j.chroma.2013.01.009. [DOI] [PubMed] [Google Scholar]
  19. Collin W. R.; Serrano G.; Wright L. K.; Chang H.; Nuñovero N.; Zellers E. T. Microfabricated Gas Chromatograph for Rapid, Trace-Level Determinations of Gas-Phase Explosive Marker Compounds. Anal. Chem. 2014, 86, 655–663. 10.1021/ac402961t. [DOI] [PubMed] [Google Scholar]
  20. Gruznov V. M.; Baldin M. N.; Efimenko A. P.; Maksimov E. M.; Naumenko I. I.; Pronin V. G. Rapid Gas-Chromatographic Determination of Marking Agents Added to the Industrial Plastic Explosives in Air. J. Anal. Chem. 2015, 70, 207–212. 10.1134/s1061934814120053. [DOI] [Google Scholar]
  21. Jones D. E. G.; Lightfoot P. D.; Fouchard R. C.; Kwok Q. S. M. Thermal Properties of DMNB, a Detection Agent for Explosives. Thermochim. Acta 2002, 388, 159–173. 10.1016/s0040-6031(02)00033-3. [DOI] [Google Scholar]
  22. Containing the Threat from Illegal Bombings: An Integrated National Strategy for Marking, Tagging, Rendering Inert, and Licensing Explosives and Their Precursors; National Academies Press: Washington DC, 1998. [Google Scholar]
  23. Llewellyn P.Adsorption by Powders and Porous Solids; Elsevier, 2014; pp 529–564. [Google Scholar]
  24. Boyle B.The Complete Guide to Testing Chemical Sensors; Owlstone Nanotech Inc, 2012. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ao0c01615_si_001.pdf (2.8MB, pdf)

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES