Forensic Sci Med Pathol (2017) 13:490–494

DOI 10.1007/s12024-017-9903-4

COMMENTARY

Nanoparticles in explosives detection – the state-of-the-art

and future directions
William J. Peveler 1 & Sultan Ben Jaber 1 & Ivan P. Parkin 1

Accepted: 30 June 2017 / Published online: 12 August 2017

# The Author(s) 2017. This article is an open access publication

Introduction One method to exploit the plasmons of AuNPs for sensing

is to use them in surface enhanced Raman spectroscopy
Nanoparticle-ligand systems can be targeted to specific analytes (SERS). A Raman spectrum is a powerful way to fingerprint
to effect a change in the properties of the nanoparticles. We will a molecule, using incident light to excite Raman active vibra-
examine two examples in which the electromagnetic properties tional modes (Fig. 1a), causing inelastic scattering of the pho-
of the nanoparticles (arising from their small size) are altered by tons, and giving rise to a unique spectrum that provides infor-
analyte binding, and can be applied as a transducer in a chemical mation on molecular shape and connectivity. The spectrum
sensing system for explosive analytes. The first property obtained from an unknown analyte can be compared to a
reviewed is the surface plasmon resonance (SPR) band of col- library of known spectra and used to identify a threat.
loidal gold nanoparticles (AuNPs), and the second, the fluores- However, Raman scattering is very weak, and so detection
cence of colloidal semiconductors (quantum dots – QDs). of low levels of analyte, enhancement is required.
Sensors based on these nanoparticle properties have the po- If a molecule is bound (chemi- or physisorbed) to a metal
tential to detect picomolar or lower concentrations of explosive surface, incident light (usually a monochromated laser pulse)
analytes, and can operate for both solution phase and gas phase excites the surface plasmons, inducing polarization in the
detection [1–3]. In addition, measurement of the signal produced bound molecules, increasing the amount of inelastic scattered
by the nanoparticle transducer uses standard scientific instrumen- light from the Raman vibrational modes (Fig. 1b). This leads
tation, making it easier to build complete detection systems from to a signal enhancement of up to E4, where E is the electric-
standard components - an important consideration [4]. field magnitude. The intensity of the SERS effect is largely
attributed to the monolayer of molecules absorbed to the nano-
particles, and is highly dependent on the form (and hence
plasmonic field) of the nanoparticles [5–8]. This adsorption
Example 1 - gold nanoparticles onto the substrate also creates new vibrational selection rules,
and surface-complex formation can lead to altered electronic
In AuNPs the free electrons of the metal surface interact properties of the absorbed molecule [9].
strongly with light causing large oscillations in the surface The best SERS enhancement is achieved by having strong
electromagnetic field. The particles therefore absorb light localized plasmons, that fall within the wavelength of the
strongly at the particular resonant frequencies of these elec- Raman laser excitation. For this reason gold and silver are often
trons, giving rise to SPR bands. chosen, as their SPR bands are typically within 400–800 nm,
which is easy to access with a visible laser [10]. In addition they
are chemically inert and thus stable against air, and strong ox-
* Ivan P. Parkin
idizing or reducing agents. Other materials have been success-
[Link]@[Link] fully applied as SERS substrates, such as other noble metals (Pt
and Pd) and even transition metals and semiconductors [11, 12].
1
Department of Chemistry, University College London, 20 Gordon St, Strong plasmon hotspots created in between individual par-
WC1H 0AJ, London, UK ticles can improve the SERS enhancement effect. This can be
Forensic Sci Med Pathol (2017) 13:490–494 491

(a) virtual states mixtures, deconvolution can be a challenge [25]. A second prob-
lem is the dependence of the Raman signal on the SERS sub-
strate. Particular vibrational modes in the Raman spectrum can
2 be enhanced or suppressed depending on the binding mode of
}
1 vibrational states the analyte, and if the substrate (e.g. AuNP concentration and
0 - ground state
(i) (ii) (iii) aggregation amount) is not identical in every instance then the
fingerprints may differ slightly, causing loss of specificity.
(b) Therefore, a key requirement is the development of cheap, iden-
Rayleigh scattering tical SERS substrates, that exhibit powerful enhancement, ensur-
Incident Raman scattering ing strong and repeatable spectral fingerprints can be obtained on
laser
each use.
Enhanced Raman Recently we reported a new method of designing such sub-
scattering strates based on a new enhanced Raman technique – photoin-
duced enhanced Raman spectroscopy – PIERS. It was shown
Plasmonic field that by combining semiconducting TiO2 substrates with plas-
enhancement monic AuNPs, and pre-irradiating the material, a PIERS sub-
AuNP
strate was created that displayed an order of magnitude enhance-
Fig. 1 a Basic inelastic scattering modes that give rise to Raman spectra.
ment over conventional SERS techniques [2]. We performed
(i) Rayleigh scattering - elastic scattering that is discarded from the sensing of DNT, TNT, RDX and PETN explosives in solution,
spectrum. (ii) Stokes Raman scattering, where the change in wavelength with high quality spectral finger prints obtained even at sub-
of the emitted photon gives the Raman signal. (iii) Anti-Stokes Raman nanomolar concentrations (Fig. 2). In particular nanomolar con-
scattering - much weaker than the Stokes equivalent. b Electromagnetic
enhancement of Raman scattering for a physisorbed analyte on an AuNP
centrations of DNT and TNT were detected both in solution and
surface. Both incident laser light and outgoing Raman light are enhanced in the vapor phase, demonstrating that this PIERS technique
by the particle plasmonic field. Adapted from Guerrini & Graham [9] might have interest for stand-off detection of explosives. We
also showed that via the pre-irradiation step the substrates could
achieved through aggregation of the particles, either in solu- be fully cleaned over several cycles, meaning the same substrate
tion by trapping at an interface or chemical aggregation with can be used multiple times, unlike many commercial SERS
linking molecules, or by solvent removal [13–15]. substrates on the market today.
In recent years SERS on nanoparticulate colloids has been We undertook a thorough investigation into the mechanism
heavily applied to the detection of illicit materials, such as of this enhancement and suggest that it arises from interaction
explosives [16, 17]. The technique has been shown to be ca- between the irradiated TiO2 and gold nanoparticles, causing
pable of detecting a range of high explosives, with good de- improved charge transfer and electromagnetic enhancement at
tection limits (into the nanomolar region or below), even with the surface of the substrate (Fig. 2). Further investigations into
raw colloidal solutions of Au and AgNPs, such as those ap- the underlying mechanism and optimization for field detection
plied by the group of Hernández-Rivera [18, 19]. By aggre- of explosives, and other threats, are ongoing.
gating AuNPs at an interface, Edel et al. created a regular
monolayer array, which showed enhanced sensitivity to a
range of compounds, including some explosives [20]. Example 2 - quantum dots
A more targeted approach has been taken by utilizing NPs
functionalized with cysteine to form Mesienheimer complexes Quantum dots (QDs) are semiconducting nanoparticles which
with nitroaromatics [21, 22]. Xu et al. demonstrated enhanced are small enough to confine a generated hole-electron pair
detection of DNT by using cyclodextrin coated triangular (exciton) within all 3 spatial dimensions, leading to quantiza-
nanoprisms of gold [23]. SERS also extends beyond the tion of the energy levels. This causes the electronic structure
nitroaromatics, for example it has been shown that RDX can of the material to sit between a classical semiconductor, and a
be detected at concentrations as low as 0.15 mg/L in ground molecular material (Fig. 3a).
water samples [24]. This illustrates a key benefit of SERS, that The result of this quantization of states is that the nano-
it is label free, requiring no special binding groups to target particles display sharp photon absorption and emission
particular analytes, but that it can become more targeted with bands, and the band gap is closely related to the size of
designed substrates. the nanoparticle. The fluorescence arises from photoexci-
SERS is a very useful technique for detection of explosives as tation of the nanoparticles, causing exciton formation.
it can can detect solution or vapor phase materials at very low Recombination of this exciton will then occur through
concentrations. The individual Raman fingerprint of each differ- radiative (fluorescence) or non-radiative (trap-states, oxi-
ent molecule makes specificity high, however in complex dation, energy transfer) pathways. The fluorescence from
492 Forensic Sci Med Pathol (2017) 13:490–494

O 2N O 2N NO 2
N N
(a) Rayleigh scattering
Incident N
laser Raman scattering (c) 10-9 M NO 2 (d)10-5 M NO 2
5000
NO 2 3000 PIERS PIERS
Enhanced Raman 4000
scattering
NO 2 2000
1500

Pre-irradiation 1000
Au/AgNP 1000
with UV 500
0 0
SERS SERS

Intensity ( a.u.)
-4.3 400
TiO2 (R) 600
-5.1
300
-7.5 Au 400
Glass 200
(b) 200
100

0 0
1500 Raman 700 Raman

300 200

200 150
100
100
50
0 0
700 800 900 600 700 800 900 1000
500 nm
-1
Wavenumber (cm ) Wavenumber (cm-1 )
Fig. 2 Illustration of (a) the proposed PIERS mechanism, (b) SEM of a PIERS substrate and (c,d) exemplar sensing results for DNT and RDX. Adapted
from Nature Communications [2] Copyright 2016 Nature Publishing Group

QDs is easily tuned to the visible or near-IR region of the Introducing molecules around the surface of the QD will
spectrum, by choice of semiconductor material and parti- affect the rate of recombination of the exciton, and may also
cle size, making QDs useful fluorophores (Fig. 3b). disrupt its recombination. In particular, the conduction band elec-
tron may be lost to a local species in a process termed photoin-
duced electron transfer (PET). In PET the loss of the excited
(a) electron to the acceptor results in the prevention of recombina-
tion, and thus the loss of fluorescence. The more efficient this
process, the larger the fluorescence quenching of the system.
Energy

Eg Eg
HOMO-LUMO This electron transfer mechanism has been used extensive-
gap
ly to transduce the presence of analytes in proximity to a QD,
and thus form the sensor element in a chemical sensor
[26–28]. The surface of the QD can be targeted to certain
metal bulk QD single
analytes by the placement of receptors on the surface that
(b) semiconductor molecule preferentially bind the analyte and bring it into close proximity
with the QD, facilitating the PET mechanism, and quenching
Energy

the observed fluorescence [29].

Eg
QDs, therefore, have many properties of interest to chem-
ical sensing - they exhibit high fluorescence quantum yields,
are resistant to photobleaching, and have broad absorption
Particle Size giving rise to narrow emission bands. This means they lend
Fig. 3 a Schematic of band structures of metals, semiconductors, themselves well to a multiplexed or multichannel fluorophore
quantum dots (QD) and single molecules, showing the increase in band system, with a single excitation wavelength causing emission
gap (Eg) as confinement is increased. At the single molecule scale the from many different species of varying color. The surface of
energy bands become completely discrete, and the QD sits between this the particles is easily modified with targeting ligands to allow
system and the band model of semiconductors. b Graphic illustrating the
change in QD band gap and photoluminescence emission wavelength specific and sensitive fluorescence enhancement or quenching
(colour) with increasing particle size on association with an analyte via a PET mechanism.
Forensic Sci Med Pathol (2017) 13:490–494 493

(a) Functional surface (b) 300000

Tetryl RDX
250000

Intensity (AU)
200000

150000

Fluorescent Quantum Dot 100000

Titrate
with analyte 50000
Analyte binding causes
electron transfer (ET) 0
450 500 550 600 650 450 500 550 600 650
Wavelength (nm) Wavelength (nm)

(c)
6

Canonical2 - 2.94%
4

TNT
PETN
2
Fl. Intensity

RDX

DNT

Tetryl
Fluorescent fingerprint 0
for
-2
Good binding Poor binding Good binding -5 0 5 10 15
and good ET no ET or good ET Canonical1 - 96.10%
Fig. 4 Array based sensing of explosives. (a) Schematic of a QD array for PET based sensing of explosives. (b) Fluorescent quenching by two explosives. (c)
Statistical discrimination of the fluorescent fingerprints for different explosives. Adapted from ACS Nano [1] Copyright 2016 American Chemical Society

In the security domain, QDs have been used for approx- discriminated using the array of just three QDs. It was also
imately 10 years as a fledgling sensor for explosives. Due shown that the QDs in the array could operate in the format of
to explosives’ (particularly nitro/conventional explosives’) a paper-based test, in addition to the solution-based assay.
electron deficient nature, they make excellent PET Limits of detection down to ppb levels were obtained, and
quenchers [30]. Initially antibody targeting of explosives most importantly the array could read out information on what
was used, but this is complicated by the procedures to type of explosive was present, rather than just if there was a
obtain, purify and conjugate the antibodies [31, 32]. A particular single explosive or not (Fig. 4).
cheaper and simpler targeting approach has been the for-
mation of Meisenheimer complexes, using surface amines
to bind nitroaromatics, such as TNT or picric acid [33–38]. The future?
This system can be highly specific towards nitroaromatic
materials, but is very difficult to apply more widely to Nanomaterials have shown great promise in the explo-
other explosives (even DNT). A novel donor/acceptor sys- sives detection field and it is likely that future commercial
tem for the sensing of explosives beyond TNT has been developments in this area will make some use of these
pioneered by Willner et al. They have utilized both a redox types of matter. Two key areas of interest are the Raman
couple based on NAD + /NADH and a donor/acceptor enhancing properties of plasmonic gold nanoparticles, and
scheme, based on surface bound dopamine derivatives, to the fluorescent nanomaterials such as quantum dots that
sense a range of explosives, including RDX [39, 40]. can be used in complex photonic systems for sensing dif-
To further this work we have recently exploited the attractive ferent types of explosive at low levels. Each of these has
optical properties of QDs to build an explosive sensing array of shown their worth in the laboratory and efforts must now
quantum dots. By combining multiplexed fluorophores with var- focus on more rigorous device design and field testing to
iable response do different explosives, with array statistics, it was move towards end-user applications.
possible to ‘fingerprint’ 5 different explosives and identify them
at low concentration [1]. This has potential applications in waste- Acknowledgements WJP is supported by an EPSRC Doctoral Prize
water management and testing, as well as drinking water evalu- Fellowship (EP/M506448/1). SBJ acknowledges the support of the gov-
ernment of Saudi Arabia, Ministry of Interior, King Fahd Security
ation in areas where explosives contamination is a health issue,
College (KFSC).
such as ordinance ranges and manufacturing sites, as well as for
safety management. Open Access This article is distributed under the terms of the Creative
In this system, the surfaces of the QDs were modified with Commons Attribution 4.0 International License (http://
[Link]/licenses/by/4.0/), which permits unrestricted use,
a range of supramolecular functionalities to control their se-
distribution, and reproduction in any medium, provided you give appro-
lective interactions with different explosives, and DNT, TNT, priate credit to the original author(s) and the source, provide a link to the
RDX, PETN and tetryl were successfully detected and Creative Commons license, and indicate if changes were made.
494 Forensic Sci Med Pathol (2017) 13:490–494

References 21. Xu Z, Hao J, Braida W, Strickland D, Li F, Meng X. Surface-en-

hanced Raman scattering spectroscopy of explosive 2,4-
dinitroanisole using modified silver nanoparticles. Langmuir.
1. Peveler WJ, Roldan A, Hollingsworth N, Porter MJ, Parkin IP. 2011;27:13773–9.
Multichannel detection and differentiation of explosives with a 22. Dasary SSR, Singh AK, Senapati D, Yu H, Ray PC. Gold nanopar-
quantum dot array. ACS Nano. 2016;10:1139–46. ticle based label-free SERS probe for ultrasensitive and selective
2. Ben-Jaber S, Peveler WJ, Quesada-Cabrera R, Cortés E, Sotelo- detection of trinitrotoluene. J Am Chem Soc. 2009;131:13806–12.
Vazquez C, Abdul-Karim N, et al. Photo-induced enhanced 23. Xu JY, Wang J, Kong LT, Zheng GC, Guo Z, Liu JH. SERS detec-
Raman spectroscopy for universal ultra-trace detection of explo- tion of explosive agent by macrocyclic compound functionalized
sives, pollutants and biomolecules. Nat Commun. 2016;7:12189. triangular gold nanoprisms. J Raman Spectrosc. 2011;42:1728–35.
3. Smith RG, D’Souza N, Nicklin S. A review of biosensors and 24. Hatab NA, Eres G, Hatzinger PB, Gu B. Detection and analysis of
biologically-inspired systems for explosives detection. Analyst. cyclotrimethylenetrinitramine (RDX) in environmental samples by
2008;133:571–84. surface-enhanced Raman spectroscopy. J Raman Spectrosc.
4. Wolfbeis OS. Probes, sensors, and labels: why is real progress 2010;41:1131–6.
slow? Angew Chem Int Ed. 2013;52:9864–5. 25. Kasera S, Herrmann LO, Barrio JD, Baumberg JJ, Scherman OA.
5. Liu Z, Zhang F, Yang Z, You H, Tian C, Li Z, et al. Gold Quantitative multiplexing with nano-self-assemblies in SERS. Sci
mesoparticles with precisely controlled surface topographies for Rep. 2014;4:6785.
single-particle surface-enhanced Raman spectroscopy. J Mater 26. Freeman R, Willner I. Optical molecular sensing with semiconduc-
Chem C. 2013;1:5567–76. tor quantum dots (QDs). Chem Soc Rev. 2012;41:4067–85.
6. Tiwari VS, Oleg T, Darbha GK, Hardy W, Singh JP, Ray PC. Non- 27. Algar WR, Stewart MH, Scott AM, Moon WJ, Medintz IL.
resonance SERS effects of silver colloids with different shapes. Quantum dots as platforms for charge transfer-based biosensing:
Chem Phys Lett. 2007;446:77–82. challenges and opportunities. J Mater Chem B. 2014;2:7816–27.
7. Kedia A, Kumar PS. Controlled reshaping and plasmon tuning 28. Sun X, Wang Y, Lei Y. Fluorescence based explosive detection: from
mechanism of gold nanostars. J Mater Chem C. 2013;1:4540. mechanisms to sensory materials. Chem Soc Rev. 2015;44:8019–61.
8. Peveler WJ, Parkin IP. Rapid synthesis of gold nanostructures with 29. Hezinger AFE, Teßmar J, Göpferich A. Polymer coating of quan-
cyclic and linear ketones. RSC Adv. 2013;3:21919–27. tum dots – a powerful tool toward diagnostics and sensorics. Eur J
9. Guerrini L, Graham D. Molecularly-mediated assemblies of plas- Pharm Biopharm. 2008;68:138–52.
monic nanoparticles for surface-enhanced Raman spectroscopy ap- 30. Shi GH, Shang ZB, Wang Y, Jin WJ, Zhang TC. Fluorescence
plications. Chem Soc Rev. 2012;41:7085–107. quenching of CdSe quantum dots by nitroaromatic explosives and
10. Ghosh SK, Pal T. Interparticle coupling effect on the surface plas- their relative compounds. Spectrochim Acta Mol Biomol
mon resonance of gold nanoparticles: from theory to applications. Spectrosc. 2008;70:247–52.
Chem Rev. 2007;107:4797–862. 31. Goldman ER, Medintz IL, Whitley JL, Hayhurst A, Clapp AR,
11. Abdelsalam ME, Mahajan S, Bartlett PN, Baumberg JJ, Russell Uyeda HT, et al. A hybrid quantum dot-antibody fragment fluores-
AE. SERS at structured palladium and platinum surfaces. J Am cence resonance energy transfer-based TNT sensor. J Am Chem
Chem Soc. 2007;129:7399–406. Soc. 2005;127:6744–51.
32. Wilson R, Spiller DG, Prior IA, Bhatt R, Hutchinson A. Magnetic
12. Wang X, Shi W, She G, Mu L. Surface-enhanced Raman scattering
microspheres encoded with photoluminescent quantum dots for
(SERS) on transition metal and semiconductor nanostructures. Phys
multiplexed detection. J Mater Chem. 2007;17:4400–6.
Chem Chem Phys. 2012;14:5891–901.
33. Bai M, Huang S, Xu S, Hu G, Wang L. Fluorescent nanosensors via
13. Kasera S, Biedermann F, Baumberg JJ, Scherman OA, Mahajan S.
photoinduced polymerization of hydrophobic inorganic quantum
Quantitative SERS using the sequestration of small molecules inside
dots for the sensitive and selective detection of nitroaromatics.
precise plasmonic nanoconstructs. Nano Lett. 2012;12:5924–8.
Anal Chem. 2015;87:2383–8.
14. Taylor RW, Lee T-C, Scherman OA, Esteban R, Aizpurua J, Huang 34. Carrillo-Carrión C, Simonet BM, Valcárcel M. Determination of TNT
FM, et al. Precise subnanometer plasmonic junctions for SERS explosive based on its selectively interaction with creatinine-capped
within Gold nanoparticle assemblies using cucurbit[n]uril “Glue.”. CdSe/ZnS quantum dots. Anal Chim Acta. 2013;792:93–100.
ACS Nano. 2011;5:3878–87. 35. Pazhanivel T, Nataraj D, Devarajan VP, Mageshwari V, Senthil K,
15. Kundu S. A new route for the formation of au nanowires and ap- Soundararajan D. Improved sensing performance from methionine
plication of shape-selective au nanoparticles in SERS studies. J capped CdTe and CdTe/ZnS quantum dots for the detection of trace
Mater Chem C. 2013;1:831–42. amounts of explosive chemicals in liquid media. Anal Methods.
16. Dasary SSR, Senapati D, Singh AK, Anjaneyulu Y, Yu H, Ray PC. 2013;5:910–6.
Highly sensitive and selective dynamic light-scattering assay for 36. Singh K, Chaudhary GR, Singh S, Mehta SK. Synthesis of highly
TNT detection using p-ATP attached gold nanoparticle. ACS luminescent water stable ZnO quantum dots as photoluminescent
Appl Mater Interfaces. 2010;2:3455–60. sensor for picric acid. J Luminescence. 2014;154:148–54.
17. Dasary SSR, Rai US, Yu H, Anjaneyulu Y, Dubey M, Ray PC. Gold 37. Tian X, Chen L, Qing X, Yu K, Wang X, Wang X. Hybrid cadmium
nanoparticle based surface enhanced fluorescence for detection of tellurium quantum dots for rapid visualization of trace-level
organophosphorus agents. Chem Phys Lett. 2008;460:187–90. nitroaromatic explosives. Anal Lett. 2014;47:2035–47.
18. Jerez-Rozo JI, Primera-Pedrozo OM, Barreto-Caban MA, 38. Wang YQ, Zou WS. 3-Aminopropyltriethoxysilane-functionalized
Hernandez-Rivera SP. Enhanced Raman scattering of 2,4,6-TNT manganese doped ZnS quantum dots for room-temperature phos-
using metallic colloids. IEEE Sensors J. 2008;8:974–82. phorescence sensing ultratrace 2,4,6-trinitrotoluene in aqueous so-
19. Primera-Pedrozo OM, Jerez-Rozo JI, La Cruz-Montoya DE, Luna- lution. Talanta. 2011;85:469–75.
Pineda T, Pacheco-Londono LC, Hernandez-Rivera SP. 39. Freeman R, Willner I. NAD +/NADH-sensitive quantum dots: ap-
Nanotechnology-based detection of explosives and biological plications to probe NAD +−dependent enzymes and to sense the
agents simulants. IEEE Sensors J. 2008;8:963–73. RDX explosive. Nano Lett. 2009;9:322–6.
20. Cecchini MP, Turek VA, Paget J, Kornyshev AA, Edel JB. Self- 40. Freeman R, Finder T, Bahshi L, Gill R, Willner I. Functionalized
assembled nanoparticle arrays for multiphase trace analyte detec- CdSe/ZnS QDs for the detection of nitroaromatic or RDX explo-
tion. Nat Mater. 2013;12:165–71. sives. Adv Mater. 2012;24:6416–21.