Introduction

In recent years, rapid and sensitive detection of improvised explosive devices (IEDs) has become the most pressing issues with the increase of the globalization of terrorist acts during the last 2 decades, such as those in Oklahoma, U.S.A. (1995), Bali, Indonesia (2002, 2005), Pune, India (2010), Oslo, Norway (2011), Boston, U.S.A. (2013) and Abuja, Nigeria (2014)1,2,3. On-site screening is an effective way to reduce the risk of explosion, which demands the analytical methods with high requirements in terms of speed, sensitivity and robustness4,5,6,7,8,9,10.

The instrumental methods such as X-rays, neutron analysis, nuclear quadrupole resonance and colorimetric detection are always adopted for the bulk detection of organic and inorganic explosives, while the thermal desorption ion mobility spectrometry (TD-IMS) is proven to be a practical technique for trace detection (from ng to pg) of organic explosives such as 2,4,6-trinitrotoluene (TNT) and cyclo-1,3,5-trimethylene-2,4,6-trinitramine (RDX) et al11,12,13,14,15. IMS is a gas-phase ion separation and detection technique in a uniform electric field based on the mobility difference of gaseous ions16,17,18,19. More than 20,000 stand-alone IMS have been deployed at airports and subway stations worldwide for security applications, due to the advantages of fast speed, low cost, analytical flexibility, portability and commercial availability20,21,22,23. However, it still remains a great challenge for TD-IMS to detect trace inorganic explosives except for ammonium nitrate-fuel oil (ANFO) and sulphur (S) in black powder, owing to their ultra-low vapour pressure even at the typical maximum desorber temperature (≤280°C)13,20,24,25.

Inorganic explosives generally consist of an inorganic oxidizer such as potassium nitrate (KNO3), potassium chlorate (KClO3), or potassium perchlorate (KClO4) and a fuel such as carbon source, sulphur (S), sugar or powdered metals26,27. Inorganic explosives are extensively used in terrorist attacks owing to the readily available, low cost and legally purchased components. Ion chromatography (IC)28,29, capillary electrophoresis (CE)30,31, electrospray ionization mass spectrometry (ESI-MS)32,33,34 and electrospray ionization ion mobility spectrometry (ESI-IMS)35 have been performed to identify their characteristic ions, such as nitrate (NO3−), chlorate (ClO3−), sulphate (SO42−) and perchlorate (ClO4−) etc. contained within inorganic explosives36,37. However, the field-deployment of these techniques is hampered by the tedious and time-consuming procedures for handling the aqueous samples. Additionally, the relatively long analysis time (>5 minutes) of IC and CE and the relatively large bulk, high expensive cost of mass spectrometer make them not suitable for on-site screening as well.

Obviously, an analytical method which could rapidly and sensitively detect both inorganic and organic explosives is highly demanded. Herein, we firstly reported a simple method to detect inorganic explosives involving KNO3, KClO3 and KClO4 using TD-IMS via acid-enhanced evaporization for on-site detection. Samples on the swab were acidized by spraying H3PO4 and enhancements of signal for KNO3, KClO3 and KClO4 by a factor of over 3000 were achieved, realizing the detection of inorganic explosives at ng even pg level. Moreover, the common nitro-explosives and its mixture with the inorganic oxidizers and the frequently-used inorganic explosives such as black powder, firecracker and match head were sensitively detected. The influence of acidification process on the detection of organic explosives was also investigated.

Results and Discussion

Design of TD-IMS for analysis

A swiped-based TD-IMS apparatus operated in negative mode is schematically shown in Fig. 1. The key structure of IMS is similar to those reported in detail previously25,38 expect that an extra inlet named inlet 2 used for the sample inlet was punched in the reaction region at the vicinity of BN ion gate. Thus, only the purge gas flowed through the 63Ni source, so the pollution and acid corrosion of 63Ni source could be effectively avoided. In the test, samples were deposited on the swabs firstly (Fig. 1(a)) and then sprayed with minimal regime 3% H3PO4 (Fig. 1(b)). Afterwards, the swab was inserted into the thermal desorber with temperature of 180°C for IMS analysis (Fig. 1 (c)).

Figure 1
figure 1

Schematic of thermal desorption ion mobility spectrometer and the analysis process of inorganic explosives.

(a) Sampling: sampling swabs were swiped with samples; (b) Acidification: samples on the swab were sprayed with minimal regime 3% H3PO4; (c) Analysis: the swab was inserted into the thermal desorber and analysed by IMS.

Full size image

Enhancement of acidification on the sensitivity

Ion mobility spectra for typical inorganic oxidizers of KNO3, KClO3 and KClO4 obtained without and with 3% H3PO4 acidfication were depicted in Fig. 2, respectively. Clearly, in lack of H3PO4 acidfication on the swab, only two weak product ion peaks with K0 of 2.18 and 2.05 cm2V−1s−1 were observed for 20 μg KNO3 and the signal intensity of characteristic peak (2.05 cm2V−1s−1) was only about 83 mV. Whereas, its response for 0.02 μg KNO3 instantly increased to 280 mV in the existence of H3PO4. It was over 3 times higher than that of 20 μg KNO3 obtained without acidification and the sensitivity were at least enhanced by a factor of 3000. Similarly, no responses for KClO3 and KClO4 even with quantity of 50 μg and 600 μg, respectively were obtained in the absence of H3PO4, while two strong ion peaks of 0.05 μg KClO3 were appeared at 2.16 and 1.88 cm2V−1s−1 and peaks with K0 of 2.15 and 1.77 cm2V−1s−1 were observed for 0.60 μg KClO4 in the assistance of H3PO4. And the signal intensities for ion peaks of KClO3 (1.88 cm2V−1s−1), KClO4 (1.77 cm2V−1s−1) were 184 and 179 mV, respectively. Furthermore, we also adopted other weak acids such as acetic acid (CH3COOH) and boric acid (H3BO3) to perform the acidification process. The results demonstrated that H3PO4 was the most suitable for the analysis due to the perfect sensitivity and its simple background (see Supplementary Fig. S1–2). It is noteworthy that the time taken for sample-to-sample analysis is less than 5 s and the analysis speed is fast enough for rapid on-site screening.

Figure 2
figure 2

Ion mobility spectra of 20 μg KNO3, 50 μg KClO3 and 600 μg KClO4 obtained without 3% H3PO4 and ion mobility spectra of 0.02 μg KNO3, 0.05 μg KClO3 and 0.6 μg KClO4 obtained with 3% H3PO4.

Full size image

Mechanism for the acid-enhancement

The product ions of inorganic oxidizers in 63Ni source were assigned by an atmospheric pressure negative ionization time-of-flight mass spectrometer (TOF-MS) (see Supplementary Fig. S3). The typical mass spectra were demonstrated in Fig. 3; the reduced ion mobilities (K0), mass-to-charge ratios (m/z) and corresponding product ions were summarized in Table 1. Accordingly, two product ions of KNO3 at m/z 62 and m/z 125 were assigned to NO3− and HNO3·NO3− and two ions of KClO3 at m/z 83, 85 and m/z 167, 169, 171 were attributed to ClO3− and HClO3·ClO3−, while the ions assigned to ClO4− (m/z = 99, 101) and HClO4·ClO4− (m/z = 199, 201, 203) were observed for KClO4. The product ions observed for KNO3 were same as those in previous experiments of ammonium nitrate (AN, the main component of ANFO) where NO3− and complex HNO3·NO3− were the dominant ions, likely formed by HNO3 (the decomposition product of AN)3. According to the observed product ions, gaseous inorganic acids, such as HNO3, HClO3 and HClO4 were believed to be released in the thermal desorber after acidification through the metathesis reaction beween the solid oxidizers and H3PO4 at high temperature of 180°C and then ionized by the reactant ions O2− in the ion source. Hence, inorganic oxidizers could be indirectly detected by identifying their gaseous inorganic acids in the existence of H3PO4. It was further verified by ion mobility spectra of trace HNO3 and HClO4 (see Supplementary Fig. S4) where the ion peak positions of HNO3 and HCIO4 were same as those of KNO3 and KClO4 with acidification in Fig. 2. The mechanism of acid-enhanced detection of inorganic salt oxidizers are summarized as equation (1) to (3) and X presents N or Cl atom.

Table 1 Reduced ion mobilities (K0), mass-to-charge ratios (m/z), product ions and quantitative analysis results obtained with 3% H3PO4
Full size table
Figure 3
figure 3

TOF-mass spectra of 20 ng KNO3, 50 ng KClO3 and 600 ng KClO4 obtained with 3% H3PO4.

Full size image

Quantitative analysis

To further investigate the sensitivities of inorganic explosives, we used ion peaks at K0 of 2.05, 1.88 and 1.77 cm2V−1s−1 for quantitative analysis of KNO3, KClO3 and KClO4. The results were summarized in Table 1; their detectable linear range were 2–60 ng, 10–120 ng and 50–600 ng, respectively and the RSD (n = 9) were 17.10%, 16.27% and 9.99%, respectively. The limits of detection (LOD, S/N = 3) for KNO3, KClO3 and KClO4 were 0.10 ng, 0.84 ng and 6.30 ng, respectively. These results exhibit the good quantitative capability of the current method for detection of inorganic explosive.

Detection performances of actual samples

The acidification process did not affect the detection of organic explosives and typical nitro-explosives such as TNT and RDX could be sensitively detected by current method and their ion mobility spectra were displayed in Fig. S5 in the Supplementary Information. Clearly, product ion with K0 of 1.54 cm2V−1s−1 was observed for TNT and the product ion peaks at K0 of 1.65, 1.53 and 1.44 cm2V−1s−1 were observed for RDX. Additionally, the mixtures of TNT and inorganic oxidizers could also be sensitively measured as well and Fig. 4 shows the ion mobility spectra of 5 ng TNT mixed with 20 ng KNO3 and 50 ng KClO3, respectively. From the spectra, the characteristic ion peaks of TNT, KNO3 and KClO3 are clearly observed simultaneously for the mixtures.

Figure 4
figure 4

Ion mobility spectra of the mixture of 5 ng TNT and 20 ng KNO3 and mixture of 5 ng TNT and 50 ng KClO3 obtained with 3% H3PO4.

Full size image

Moreover, the oxidizers in typical inorganic explosives were also detected. Figure 5 shows the ion mobility spectra of black powder, firecracker and match head obtained in the existence of H3PO4. From the spectra, the characteristic ion peaks of KNO3, KClO4 and KClO3 were distinctly observed for black powder, firecracker and match head, respectively. It indicates that inorganic oxidizers in black powder, firecracker and match head are KNO3, KClO4 and KClO3, respectively and other substances in the explosives such as powdered metals and charcoals had little impact on the identification. In particular, only S was detected in the firecracker and black powder without H3PO4 acidification (see Supplementary Fig. S6), so the detection of both S and inorganic oxidizers all could help identification the materials used in IEDs. Finally, as shown in Supplementary Fig. S7, KNO3, KClO3, KClO4 in sugar (20 ng KNO3/sugar (4:1), 50 ng KClO3/sugar (4:1) and 600 ng KClO4/sugar (3:2)) could also be sensitively measured, which means that the common fuel in inorganic explosives did not affect the detection.

Figure 5
figure 5

Ion mobility spectra of 200 ng match head, 600 ng firecracker and 20 ng black powder obtained with 3% H3PO4.

Full size image

Conclusions

We demonstrated a new rapid and sensitive method based on IMS via acid-enhanced detection of oxidizers including KNO3, KClO3 and KClO4 to detect inorganic explosives. It owns the advantages of fast speed (about 5 s), low cost, analytical flexibility, portability, commercial availability and suitable for detecting inorganic explosives and organic ones at the level of ng even pg. Moreover, it is envisioned to be easily deployed to the rapid and sensitive on-site screening of both inorganic explosives and organic ones at airports, subway stations, sports venues and even blast site, being an effective way to screen the substances in the detonation scene and reduce the risk of explosion.

Methods

Experiment

A 500 ng μL−1 KNO3, KClO3 and 600 ng μL−1 KClO4 standard stock solutions and 3% H3PO4 solution were prepared by weighing sample and dissolving in distilled water. The standard solutions and test samples were obtained by gradually diluting the relevant stock solutions with distilled water or other volatile solvents. 20 ng KNO3/sugar (4:1), 50 ng KClO3/sugar (4:1), 600 ng KClO4/sugar (3:2) were made up by mixing KNO3, KClO3, KClO4 and sugar in corresponding proportion while match head, firecrackers, black powders, 2,4,6-trinitrotoluene (TNT) and cyclo-1,3,5-trimethylene-2,4,6-trinitramine (RDX) were commercial grade.

A known quantity sample solutions were added on or solid samples were rubbed on the swabs firstly and the solvent was evaporated to dryness on a hot plate of 50°C and then the swabs with samples were sprayed minimal regime 3% H3PO4 (several microlitres). After that the swab was inserted into the thermal desorber for analysis by IMS operated in negative mode. It is noted that enough acid content is essential for the sensitivity.

Operating process and conditions

The sample on the swab (2 cm × 8 cm) made of Teflon was heated to 180°C in the thermal desorber and vaporized into a gas flow that passed into the drift tube for analysis. In the 63Ni ionization source, the sample molecules were ionized and then the product ions were injected into the drift region through the BN gate with an injection pulse width of 300 μs. In the drift region, ions were separated according to their mobility with an electric field of 391 V/cm and then detected by the Faraday plate and the ion current was amplified with a gain of 109 V/A. The drift tube temperature was kept at a constant of 90°C. Air dried by silica gel, activated carbon and 13× molecular sieve traps was divided into drift, carrier and purge gases via flow controllers which were optimized at 800, 400 and 400 mL/min, respectively. The operating parameters were summarized in Table 2.

Table 2 The IMS operating parameters used in this work
Full size table

Calculation

The reduced ion mobility, K0 for unknown peaks were calculated with a standard via the following equation,

where K0a is the reduced mobility of the analyte, tds and tda is the drift time of the standard and the analyte, respectively. (TNT-H)− was used as the standard substance, whose K0 was 1.54 cm2V−1s−1 and its drift time was 11.28 ms under the same experimental conditions.