Detection of explosives based on surface-enhanced ... - OSA Publishing

Detection of explosives based on surface-enhanced Raman spectroscopy Hainer Wackerbarth,* Christian Salb, Lars Gundrum, Matthias Niederkrüger, Konstantin Christou, Volker Beushausen, and Wolfgang Viöl Laser-Laboratorium Göttingen e.V., Photonic Sensor Technology, Hans-Adolf-Krebs-Weg 1, D-37077 Göttingen, Germany *Corresponding author: hainer.wackerbarth@llg‑ev.de Received 10 February 2010; revised 24 April 2010; accepted 30 April 2010; posted 3 May 2010 (Doc. ID 123925); published 4 August 2010

In this study we present a device based on surface-enhanced Raman scattering (SERS) for the detection of airborne explosives. The explosives are resublimated on a cooled nanostructured gold substrate. The explosives trinitrotoluene (TNT) and triacetone triperoxide (TATP) are used. The SERS spectrum of the explosives is analyzed. Thus, TNT is deposited from an acetonitrile solution on the gold substrate. In the case of TATP, first the bulk TATP Raman spectrum was recorded and compared with the SERS spectrum, generated by deposition out of the gas phase. The frequencies of the SERS spectrum are hardly shifted compared to the spectrum of bulk TATP. The influence of the nanostructured gold substrate temperature on the signals of TATP was studied. A decrease in temperature up to 200 K increased the intensities of the TATP bands in the SERS spectrum; below 200 K, the TATP fingerprint disappeared. © 2010 Optical Society of America OCIS codes: 240.6695, 300.6450.

1. Introduction

The increased use of explosives by terrorists over the past few years has led to the technological upgrading of existing detection systems and to the development of new technologies. The growing threats to the civilian population, as well as the security response of governmental and enforcement agencies compel researchers to continuously modify, update, and improve detection methods. Because of the wide range of explosives and the many differences in their physical properties, several detection devices detect only certain types of explosives and fail to detect others. For example, many detection devices readily detect conventional explosives made of organic nitro and nitrate compounds, but are not able to find explosives made of inorganic nitrates or nonnitrogenous compounds. In particular, many nitrogen-based detection devices fail to detect explosives such as triacetone triperoxide (TATP); in other 0003-6935/10/234362-05$15.00/0 © 2010 Optical Society of America 4362

APPLIED OPTICS / Vol. 49, No. 23 / 10 August 2010

words it is “transparent” outside the laboratory. Moreover, TATP is not only potent but also fairly easy to synthesize. It can be made using materials that are commercially available, and there is no need for sophisticated equipment [1]. These three characteristics, its transparency, potency, and simplicity to produce, have led to an increased usage of TATP in terrorist bombings. Raman spectroscopy is a vibrational technique that provides specific spectral information about molecules enabling the identification of analytes by their “fingerprint” spectrum. Neat explosives have been extensively studied by Raman spectroscopy [2–4]. However, its use is so far limited to bulk detection; the detection limits are too high to be practically useful in safety devices. The sensitivity of the inherent weak Raman process can be increased by the surface-enhanced Raman scattering (SERS) effect [5]. SERS is a sensitive spectroscopic tool with detection limits down to the picogram and femtogram level. This is particularly attractive because it combines high sensitivity with high information content

for establishing molecular identity. The potential of SERS for trace analysis has been explored actively in the past decade. It has been already applied as a sensor for nitro explosives in homeland defense and landmine detection [6,7]. A further advantage of SERS is the quenching of fluorescence, which is a known obstacle in Raman spectroscopy of explosives [3]. Here, we present a device based on resublimation on a cooled nanostructured surface for the detection of explosives such as TATP and trinitrotoluene (TNT). 2. Material and Experimental Setup

TATP and TNT as solids (500 μg=carrier) were purchased from Explotech GmbH (Cologne, Germany). TNT in acetonitrile (1 mg=1 ml) was brought from Sigma-Aldrich Chemie GmbH. According to Dubnikova et al. TATP can safely be prepared by adding 5 drops of concentrated sulfuric acid to a mixture of acetone (5:6 g) and hydrogen peroxide (30%, 11:3 g) at 0 °C [8]. The mixture was kept at room temperature for 24 h. TATP precipitated as white crystals, which were filtered and air dried. For the Raman measurements, a standard system (Kaiser Optical Systems Inc., Ann Arbor, Michigan, USA) was used; the excitation wavelength of 785 nm (linewidth 0:06 nm) was provided by a GaAlAs diode laser (Invictus, Kaiser Optical Systems, Inc.). The excitation and scattered light were guided by a multimode optical fiber equipped with probe head onto the sample. The power of the laser emission was approximately 100 mW at the probe head. The scattered light was coupled by a confocal aperture in the optical fiber and transmitted to the spectrograph (Ph AT System, Kaiser Optical Systems Inc.), which uses holographic transmissive gratings to perform filtering and dispersion functions. The diffracted light was recorded with a CCD camera (iDus, Andor Technology plc.). The spectral resolution was 5 cm−1. All SERS spectra were recorded on a commercial nanostructured gold surface (Klarite, D3 Technologies, Ltd.) [9]. The SERS substrate consists of a gold-coated periodic square lattice of inverted pyramidal pits with a pitch of 2 μm and a depth of 0:8 μm (Fig. 1). It is produced using conventional optical lithography on a ð100Þ oriented silicon wafer followed by an anisotropig etching resulting in the described morphology. The modular device consists of an enrichment, analytical, and detection unit (Fig. 2). For the evaporation cell and analysis cell, vacuum components

Fig. 2. (Color online) (A) Schema of the modular device for the detection of explosives, (B) technical drawing of the analysis cell and evaporation cell.

are used. The detection unit encompasses the diode laser, the spectrometer, and the computer with software. The laser beam is guided by the excitation fiber to the probe. The beam is focused at the end of a Kovar tube on the nanostructured surface. The scattered light is collected and guided via the collection fiber to the entrance slit of the spectrometer. The explosives are evaporated from the carrier in the enrichment chamber. The chamber can be heated up to 200 °C. The explosive that has been released flows into the analysis unit. There, the gaseous analyte resublimates on the cooled nanostructured gold substrate. The temperature of the gold substrate can be set up to 120 K by the Stirling cooler. 3. Results and Discussion

Fig. 1. (Color online) Scanning electron microscope images of the (a) Klarite substrate and (b) pits of the Klarite substrate.

In situ detection and identification of explosives has a significant impact on the safety and security of the civilian population. To capture the explosives on the surface resublimiation out of a gas phase on a cooled metal surface was used. The explosives were detected by Raman spectroscopy. By using a 10 August 2010 / Vol. 49, No. 23 / APPLIED OPTICS

4363

nanostructured gold surface, the signals of the explosives adsorbed at the surface are enhanced due to the SERS effect by several orders of magnitude. Therefore, we built a modular device (Fig. 2) for the detections of explosives based on SERS spectroscopy. We have used the explosives TNT and TATP. TNT is an aromatic nitro compound [Fig. 3(a)], which is relatively insensitive to shock and cannot be exploded without a detonator. Therefore, it is the most favoured chemical explosive, extensively used in explosive ordnance and for demolitions. In contrast, the organic peroxide, TATP [Fig. 3(b)], is highly susceptible to heat, friction, and shock. Both explosives can be identified by their characteristic Raman bands. TNT solved in acetonitrile was dropped on the gold substrate. The acetonitrile was evaporated by heating the substrate to 30 °C for 10 min. The dominant acetonitrile bands in the SERS-spectrum are located at 381 cm−1 and 920 cm−1 [Fig. 4(a)]. These bands disappeared after the evaporation and the TNT bands appeared [Fig. 4(b)]. We have dropped the TNT solution on quartz glass and a smooth gold surface. Quartz glass did not show the TNT bands in the spectrum; instead the characteristic SiO2 band at 521 cm−1 can be observed. At the smooth gold surface the bands are visible with very low intensity, whereas the Klarite shows under same experimental conditions the enhancement of the TNT bands [Fig. 5]. This indicates the surface-enhanced Raman effect of the nanostructured gold surface. Seven bands of the TNT can be attributed to vibrational modes, namely 322 cm−1 (framework distortion mode), 792 cm−1 (C─H out-of-plane bend), 824 cm−1 (nitro-group scissoring mode), 1207 cm−1 (C6 H2 ─C vibration), 1355 cm−1 (NO2 symmetric stretching vibration), 1542 cm−1 (NO2 asymmetric stretching vibration), and 1616 cm−1 (C═C aromatic stretching vibration) [10]. Additionally, the recorded spectrum shows a band at 187 cm−1 . The SERS spectrum of TNT resublimated out of the gas phase (temperature nanostructured surface −20 °C) is shown in Fig. 4(c). The acquisition time was up to 10 s. Although this SERS spectrum shows a higher noise level compared to the one where the TNT was deposited by evaporation of acetonitrile most of the bands can be detected. The dominant feature is the band at 1351 cm−1 . In combination with the bands at 792 cm−1 and 825 cm−1 , the TNT can be identified clearly. The slight shift up to 4 cm−1 of the bands arises from the different adsorption process, i.e., TNT deposited from the acetonitrile solution or resu-

Fig. 3. Chemical structure of (a) trinitrotoluene (TNT) and (b) triacetone triperoxide (TATP). 4364


Fig. 4. (Color online) Raman spectra of (a) acetonitrile on a nanostructured gold substrate, (b) TNT deposited from acetonitrile on a nanostructured gold surface, and (c) TNT resublimated from the gas phase all at a nanostructured gold surface. Excitation wavelength 785 nm.

blimated from the gas phase, obviously resulting in different adsorption modes. Moreover, the temperature difference of the gold substrate, which was 20 °C for the TNT deposited from the acetonitrile solution and −20 °C for the resublimated TNT can contribute to the shift. Nevertheless, the bands can clearly be recognized. The spectral features of the SERS spectra should be sufficient to differentiate even between nitrogen based explosives. Lewis et al. have shown the difference of, e.g., 1,3,5 trinitro-1,3,5triazacyclohexane, (RDX) the major part of C4, nitroglycerine, and pentaerythrito tetranitrate by Raman spectroscopy [2]. Raman spectroscopy has been identified as a very useful tool for peroxide identification [11–13]. The

Fig. 5. Raman spectra of TNT (a) on a nanostructured gold substrate, (b) on a smooth gold surface, and (c) on a quartz glass substrate.

Raman spectrum of TATP crystals is shown in Fig. 6(a) and agrees well with the frequencies from Brauer et al. (Table 1) [14]. The SERS spectrum was recorded with an acquisition time of 7 s after the resublimiation at the nanostructured gold surface at 220 K. Due to the high vapor pressure of TATP (4:33 Pa), the concentration in the chamber was 70 parts in 106(ppm). For both spectra in Fig. 6 background subtraction was performed. The dominant peak is the sharp intense peak at 866 cm−1 , which is attributed to the peroxide stretching mode. The modes are assigned according to Brauer et al. [14] and Oxley et al. [15]. For the first time a SERS spectrum of TATP is analyzed. Generally, most of the characteristic TATP Raman bands are also present in the SERS spectrum (Fig. 6). The frequencies of the SERS spectrum do not shift more than 2 cm−1 compared to the Raman spectrum, indicating that the molecular environment has only little influence on the frequencies. This is plausible considering the high volatility and therefore weak intermolecular interactions in solid TATP. The band at 556 cm−1 dominates the Raman as well as the SERS spectrum. The low frequency (200 to 600 cm−1 ) bands remain more or less unchanged in intensity, besides the broad band at 242 cm−1 , which almost disappeared in the SERS spectrum. Oxley et al. found two shear modes at 236.4 and 237:3 cm−1 in their calculation [15]. A further splitting of theses modes can explain the broad peak at 242 cm−1 in the SERS spectrum. Generally, contaminations are a problem in SERS spectroscopy: a few molecules located at hot spots will contribute to the spectrum. This is obviously negligible here, inasmuch as the most intense bands in SERS spectrum can be unambiguously attributed to TATP modes. Changes in the intensity of the bands in the SERS spectrum compared to the Raman spectrum might be due the orientation on the surface. Vibrations perpendicular or horizontal to the surface are often enhanced to a different extent in SERS

Fig. 6. (Color online) Raman spectra of (a) TATP, (b) TATP resublimated from the gas phase at a nanostructured gold surface. Excitation wavelength 785 nm.

Table 1.

a

Comparision of TATP Raman and SERS Bands; Excitation 785 nma

Raman Bands Δν=cm−1

SERS Bands Δν=cm−1

Assignment

242 308 330 401 558 619 845 867 888 947 1450

242 301 326 402 556 619 845 866 890 949 1451

COOC sh ring deformation ring deformation ring bending COOC t OCO b, COOC t OCOO sh OO str CCO sym str CO str, OCO b, CCC b CH2 b

sh, shear; t, torsion; b, bending; sym, symmetric; str, stretching.

[16]. Finally, decomposition of the thermodynamically unstable TATP caused by the laser energy (785 nm excitation wavelength) was not observed. However, the identification of TATP can be unambiguously done by SERS. The ring torsion and bending bands at 308 and 401 cm−1 , the prominent C─O─O─C torsion mode at 558 cm−1 , and the four bands between 840 and 950 cm−1 are specific for the peroxide ring structure of TATP. The temperature dependence of the resublimiation of TATP on the gold substrate was studied. The TATP was evaporated in air at room temperature in the enrichment cell. Opening the valve allowed the air–TATP mixture to flow into the analysis cell. The final pressure in the analysis cell was about 2:3 × 104 Pa. Prior to the opening of the valve, the temperature of the gold substrate was set between 270 and 120 K. The SERS spectra at different gold substrate temperatures are shown in Fig. 7. The acquisition time for each spectrum was 10 s. The spectrum at 270 K clearly shows the characteristic bands of TATP (see above). As expected the lower the temperature of the substrate the better the signal of the molecules. This is clearly reflected in the spectra from 270 K to 200 K; the intensity of the bands increases, e.g., the intensity of the band at 556 cm−1 is nearly twice as much at 200 K compared to 270 K. This could be explained by the adsorption of more TATP molecules and by the temperaturedependence of the SERS effect published by Pang et al. [17]. In this publication a reversible change of the SERS signal with the change of substrate temperature over the range 15–300 K could be observed, with the intensity at 15 K being larger than that at 300 K by a factor of approximately 3. The enhancement of the Raman intensity, moreover, has been calculated over the experimental temperature range based on the electromagnetic model for SERS proposed by Gersten and Nitzan [18]. Chiang et al. showed in their calculations based on the electromagnetic model that the SERS effect survives at elevated temperatures and suggested that the variation of the plasmon frequencies with temperature could have a significant effect on the temperature 10 August 2010 / Vol. 49, No. 23 / APPLIED OPTICS

4365

2.

3.

4.

5.

6.

Fig. 7. Raman spectra of TATP resublimated from the gas phase at different temperatures of the nanostructured gold substrate. Excitation wavelength 785 nm.

dependence of the SERS effect [19,20]. At 160 K the characteristic TATP bands are hard to recognize, whereas at 120 K the TATP fingerprint has completely disappeared. This result is in contrast to the observation of Pang et al. [17]. As we have not performed our experiments under ultrahigh-vacuum conditions, contaminants and water from air could displace the TATP from the surface of the SERS substrate. 4. Conclusion

We have developed a device for the detection of airborne explosives based on SERS. TNT and TATP were studied by this device. The spectral features of the TATP SERS spectrum were analyzed by a comparison with neat TATP data and calculation from literature, and so the TATP SERS modes were identified. The device is equipped with a Stirling cooler to decrease the gold substrate temperature. Cooling up to 200 K of the gold substrate increases the signals of the adsorbed TATP. Cooling increases the resublimation of TATP and thus the intensity of the signals. Moreover, superimposed on this could be an increase of the intensity caused by the temperature dependence of the SERS effect. The results will pave the way for an implementation of this method as an explosive detector. However, the technique is not restricted to explosives; generally, airborne substances, e.g., drugs could also be detected. The authors gratefully acknowledge the support of the Federal Ministry of Ecomomics and Technology (BMWi), Germany. The study was supported by BMWi via VDI/VDE-IT: NanoSens—Nanostrukturierte photonische Gassensoren, InnoNet 16 IN0463. References 1. G. A. Buttigieg, A. K. Knight, S. Denson, C. Pommier, and M. B. Denton, “Characerization of the explosive triacetone

4366


7.

8.

9.

10.

11.

12.

13. 14.

15.

16. 17.

18.

19.

20.

triperoxide and detection by ion mobility spectroscopy,” Forensic Sci. Int. 135, 53–59 (2003). I. R. Lewis, N. W. Daniel, Jr., N. C. Chaffin, P. R. Griffiths, and M. W. Tungol, “Raman spectroscopic studies of expolsive materials: towards a fieldable explosives detector,” Spectrochim. Acta Part A 51, 1985–2000 (1995). M. L. Lewis, I. R. Lewis, and P. R. Griffiths, “Raman spectrometry of explosives with a no-moving-parts fiber coupled spectrometer: a comparison of excitation wavelength,” Vibr. Spectrosc. 38, 17–28 (2005). F. T. Docherty, P. B. Monaghan, C. J. McHuge, D. Graham, W. E. Smith, and J. M. Cooper, “Simultaneous multianalyte identification of molecular species involved in terrorism using raman spectroscopy,” IEEE Sens. J. 5, 632–639 (2005). M. Fleischmann, P. J. Hendra, and A. J. McQuillan, “Ramanspectra of pyridine adsorbed at a silver electrode,” Chem. Phys. Lett. 26, 163–166 (1974). K. M. Spencer, J. M. Sylvia, J. A. Janni, and J. D. Klein, “Advances in landmine detection using surface-enhanced raman spectroscopy,” Proc. SPIE 3710, 373–399 (1999). K. M. Spencer, J. M. Sylvia, P. J. Marren, J. F. Bertone, and S. D. Christesen, “Surface-enhanced Raman spectroscopy for homeland defense,” Proc. SPIE 5269, 1–8 (2004). F. Dubnikova, R. Kosloff, J. Almog, Y. Zeiri, R. Boese, H. Itzhaky, A. Alt, and E. Keinan, “Decomposition of triacetone triperoxide is an entropic explosion,” J. Am. Chem. Soc. 127, 1146–1159 (2005). N. M. B. Perney, J. J. Baumberg, M. E. Zoorob, M. D. B. Charlton, S. Mahnkopf, and C. M. Netti, “Tuning localized plasmons in nanostructured substrates for surfaceenhanced Raman scattering,” Opt. Express 14, 847–857 (2006). X. Wang, S. Chang, J. Yang, J. Tan, H. Jia, H. Yin, X. Li, and G. Peng, “Detection of TNT in acetone using Raman spectroscopic signature,” Proc. SPIE 6622, 662219 (2008). V. Vacque, B. Sombret, J. P. Huvenne, P. Legrand, and S. Suc, “Characterisation of the O─O peroxide bond by vibrational spectroscopy,” Spectrochim. Acta, Part A 53, 55–66 (1997). P. Jacob, B. Wehling, W. Hill, and D. Klockow, “Feasibility study of Raman spectroscopy as a tool to investigate the liquid-phase chemistry of aliphatic organic peroxides,” Appl. Spectrosc. 51, 74–80 (1997). G. Socrates, Infrared and Raman Characteristic Group Frequencies, Table and Charts, 3rd ed. (Wiley, 2001). B. Brauer, F. Dubnikova, Y. Zeiri, R. Kosloff, and R. B. Gerber, “Vibrational spectroscopy of triacetone triperoxide (TATP): anharmonic fundamentals, overtones and combination bands,” Spectrochim. Acta A, Part A 71, 1438–1445 (2008). J. Oxley, J. Smith, J. Brady, F. Dunikova, R. Kosloff, L. Zeiri, and Y. Zeiri, “Raman and infrared fingerprint spectroscopy of peroxide-based explosives,” Appl. Spectrosc. 62, 906–915 (2008). M. Moskovits, “Surface enhanced spectroscopy,” Rev. Mod. Phys. 57, 783–826 (1985). Y. S. Pang, H. J. Hwang, and M. S. Kim, “Reversible temperature dependence in surface-enhanced Raman scattering of 1-propanethiol adsorbed on a silver island film,” J. Phys. Chem 102, 7203–7209 (1998). J. Gersten and A. Nitzan, “Electromagnetic theory of enhanced Raman scattering by molecules adsorbed on rough surfaces,” J. Chem. Phys. 73, 3023 (1980). H. P. Chiang, P. T. Leung, and W. S. Tse, “The surface plasmon enhancement effect on adsorbed molecules at elevated temperatures,” J. Chem. Phys. 108, 2659–2660 (1998). H. P. Chiang, P. T. Leung, and W. S. Tse, “Remarks on the substrate—temperature dependence of surface-enhanced Raman scattering,” J. Phys. Chem. B 104, 2348–2350 (2000).