Ds

Ma

b

a

ARR2AA

KGTHSM

1

tssAwrcb

eamtmmfismpc

0d

Sensors and Actuators B 162 (2012) 7– 13

Contents lists available at SciVerse ScienceDirect

Sensors and Actuators B: Chemical

journa l h o mepage: www.elsev ier .com/ locate /snb

etection of triacetone triperoxide (TATP) using a thermodynamic based gasensor

atin Amania,b, Yun Chua, Kellie L. Watermana, Caitlin M. Hurleya, Michael J. Platekb, Otto J. Gregorya,∗

Department of Chemical Engineering, University of Rhode Island, Kingston, RI 02881, USADepartment of Electrical Engineering, University of Rhode Island, Kingston, RI 02881, USA

r t i c l e i n f o

rticle history:eceived 31 July 2011eceived in revised form2 September 2011ccepted 6 November 2011vailable online 17 November 2011

a b s t r a c t

Triacetone triperoxide (TATP) is commonly used in improvised explosive devices (IEDs) due to its rel-atively simple preparation and readily available precursors. In this study, a small footprint gas sensorutilizing metal oxide catalysts was fabricated and tested, which is capable of detecting the heat of reac-tion during the catalytic decomposition of TATP. Due to its relatively high vapor pressure compared toother explosives, TATP is an ideal target molecule for this type of gas sensor. TATP and its decompo-

eywords:as sensorATPydrogen peroxide

sition products, hydrogen peroxide and acetone were successfully detected using this thermodynamicbased sensor platform. Each tested catalyst exhibited a specific response at a given temperature andthe detection limit of TATP was determined to be in the range of parts per million. X-ray photoelectronspectroscopy (XPS) was employed to study the oxidation state of selected catalysts prior to and afterexposure to the target gases. Morphology of the catalysts in the as-deposited and tested condition was

electr

putteringetal oxide

followed using scanning

. Introduction

Triacetone triperoxide (TATP) has been widely used in IEDs byerrorists, since its precursors (acetone, hydrogen peroxide and atrong acid) are readily available and its preparation is relativelyimple. For example, in 2001 a suicide bomber targeted Americanirlines with a TATP trigger [1] and it is also believed that TATPas used in the 2005 London subway bombings. Consequently, the

eliable detection of TATP using a portable, small footprint sensorontinues to draw interest, however, relatively few studies haveeen published to date.

Unlike many other explosives, TATP contains neither metalliclements nor nitro groups, and therefore, has no significantbsorption in the ultraviolet region and does not fluoresce [2]. Thisakes it very difficult to detect with conventional spectroscopy

echniques, especially when combined with its instability andolecular structure. Specifically, the presence of peroxide bondsakes the TATP molecule sensitive to extreme shock, heat and

riction, but at the same time this makes it degrade readily. TATPs typically a crystalline solid compound at low temperatures butublimes readily at room temperature, which makes it an ideal

aterial for vapor phase detection. Several studies have been

ublished on detecting TATP, including ion mobility spectra (IMS)ombined with Raman spectrometry, mass spectrometry (MS),

∗ Corresponding author. Tel.: +1 401 874 2085.E-mail address: gregory@egr.uri.edu (O.J. Gregory).

925-4005/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.snb.2011.11.019

on microscopy (SEM).© 2011 Elsevier B.V. All rights reserved.

nuclear magnetic resonance (NMR), fluorescence spectroscopy andabsorption spectroscopy [3–8]. Each of these detection methodshas limitations, which include slow responses and/or low sensi-tivity [2], and all these techniques require large scale analyticalinstrumentation which is very difficult to implement at venueswhere a portable or hand held device is required. TATP analysis byhigh performance liquid chromatography has been touted as animproved technique, coupled with UV irradiation and electrochem-ical detection [9], but it still requires considerable analysis timeand is not suitable for in situ monitoring under field conditions.Finally, all these techniques are based on the detection of solidTATP and often involve swabbing techniques, thus cannot be usedfor continuous in situ monitoring. More recently, sensor platformsbased on colorimetric sensor arrays have been developed whichhave detection limits as low 2 ppb [10]; however, this methodis based on a one-time use, swabbing technique, and cannot beused for continuous real time detection. Capua et al. have alsorecently demonstrated a multisensory array capable of detectingppb quantities of TATP vapor continuously, in real time [11].

Metal oxide based gas sensors have been extensively studiedand have been implemented in several commercial applications,such as O2 and CO detection in internal combustion engine exhaustsystems as well as the detection of other toxic, inflammable gasessuch as NH3, NO2, SO2 and H2 at concentrations as low as 1 ppm

[12,13]. These sensor platforms also offer several advantages, suchas compatibility with conventional CMOS processing, low cost andsmall dimensions, but have been criticized for their lack of longterm stability and poor selectivity. The majority of these platforms

8 M. Amani et al. / Sensors and Actuators B 162 (2012) 7– 13

Table 1Sputtering conditions for various metal oxide catalysts.

Catalyst

Cu2O Cu2O–CuO SnO2−x ZnO V2O5 WO3–TiO2

Target diameter (mm) 150 150 150 150 125 150Target material Cu CuO SnO2 ZnO V W–Ti (90/10)Power density (W/cm2) 1.77 1.98 1.98 1.98 2.45 1.77Voltage (V) 1100 900 1150 950 1600 1400Gas pressure (Pa) Ar: 1.1 PaO2: 0.3 Pa Ar: 1.4 Pa Ar: 1.4 Pa Ar: 1.4 Pa Ar: 1.1 PaO2: 0.3 Pa Ar: 1.1 PaO2: 0.3 PaDeposition rate (�m/h) 0.17 0.9 0.8 1.1 0.06 0.13

TATP

otraiattbc

tdbsairbsTs

ftpltoTdtfeT

impregnated with nanogram quantities of TATP crystals.1 The sen-sors were mounted using a standard DB-9 connector mounted in a13 mm diameter PVC tube.

Fig. 1. Schematic of apparatus used for the detection of

perate by monitoring the changes in the electrical properties ofhe catalyst due reactions with a target gas [14]. However, theseeactions are typically observed at relatively high temperatures,nd the electrical properties of the catalyst film tend to vary due tossues such as cracking or sintering of the film. While these issuesre still of concern in the thermodynamic sensors described here,he heat effect generated by the catalyst is relatively constant overime compared to the electrical response and provides a more sta-le response which should provide additional information whenoupled with a conductometric sensor.

In order to satisfy the need for a fast and accurate method of con-inuously screening TATP, extensive efforts have been devoted toeveloping an innovative and effective sensor platform that is capa-le of monitoring explosives, such as TATP, in real time [15]. TATPublimes readily due to its relatively high vapor pressure, and forms

complex vapor upon decomposition [16,17]. Preliminary stud-es suggest that using our gas sensor platform, TATP has a similaresponse to hydrogen peroxide, one of its decomposition products,ut also exhibits its own characteristic signal. Consequently, byelecting metal oxide catalysts sensitive to the peroxide bonds inATP, as well as its organic by-products, a reliable sensor with highensitivity and high selectivity can be produced.

In this study, a thermodynamic based gas sensor was developedor the detection of TATP and its precursors. This sensor measureshe heat generated or absorbed by a metal oxide catalyst in theresence of TATP and its decomposition products in air at very

ow concentrations. The technique employs a digital control sys-em which enables a thin film microheater, coated with a metalxide catalyst, to be scanned over a selected temperature range.he electrical power difference due to catalytic reaction (powerifference between the microheater in air and the microheater in

he presence of TATP) was measured as a function of temperatureor various metal oxide catalysts. Utilizing arrays of microheaters,ach with a different catalyst it was possible to uniquely identifyATP in real time, while simultaneously minimizing false positives.

, H2O2 and acetone using a thermodynamic gas sensor.

2. Experimental

2.1. Fabrication of thermodynamic gas sensors

Over a hundred thin film microheaters were fabricated on per-forated 96% pure compact alumina substrates using an MRC 822sputtering system to deposit 2 �m thick nickel films at room tem-perature. The wafer was then annealed at 900 ◦C in nitrogen for 5 hto densify and stabilize the electrical resistivity of the nickel films.Various metal oxide catalysts having a nominal thickness of 0.5 �mwere then deposited directly over the nickel microheaters usingeither metal or ceramic targets in a MRC 8667 sputtering system.The detailed sputtering conditions for the deposition of the metaloxide catalysts are listed in Table 1. The sensors were subsequentlyannealed in air at 650 ◦C to further improve their stability. Priorto and after testing, the oxidation state of the metal oxide cata-lysts was characterized using XPS and the morphology of the oxidecatalysts was followed using SEM.

2.2. Gas sensor testing apparatus and protocol

A schematic of the test bed used for all the gas sensor exper-iments is shown in Fig. 1, where multiple Alicat Scientific massflow controllers were used to mix the target gases at various con-centrations while maintaining a constant flow rate of 100 SCCM.H2O2 and acetone were introduced into the system by bubblingair through sealed flasks containing the liquid precursors, whileTATP was introduced into the system by passing air over filter paper

1 Prepared by Professor Jimmie Oxley and Professor James Smith at the Universityof Rhode Island.

nd Actuators B 162 (2012) 7– 13 9

ttpwrwewctwtccmtgsmrp

3

3

cFb(snff2ctam

Fig. 2. XPS spectra of Cu2O films prepared by sputtering in an Ar/O2 plasma froma metal target and CuO–Cu2O films prepared by sputtering in an Ar plasma from aCuO target.

M. Amani et al. / Sensors a

A digital control system employing a LabView program was usedo heat the microheaters to a series of predetermined set pointemperatures. These temperatures were selected using the tem-erature coefficient of resistance (TCR) of the nickel film, whichas independently calibrated. After the sensor reached equilib-

ium at each temperature set point, the total power dissipationas recorded for a period of 75 s. The average amount of heat gen-

rated or absorbed by the catalyst in the presence of the target gasas calculated as the power required to maintain the sensor at a

onstant temperature, as well as the percent response which washe fraction of total power generated by the catalyst. These testsere conducted using both a dynamic pulsed gas scheme, where

he sensor is initially run in air, followed by 2 min in various con-entrations of the target gas, and then run in air again, withouthanging temperature. Static testing was also performed by deter-ining the power dissipation in air at each temperature, cooling

he sensor to room temperature and running again in the targetas at the same temperatures. Using the static test protocol, theensitivity of electronics and drift in the resistance of the thin filmicroheaters allow for power differences greater than 8 mW to be

eliably detected, while the dynamic response technique allows forower differences as low as 1 mW to be reliably detected.

. Results and discussion

.1. Chemical characterization

The oxidation states of the copper oxide catalysts prepared fromeramic and metallic targets were determined by XPS (Fig. 2). Inig. 2, the Cu 2p spectra of those copper oxide films preparedy reactive sputtering in an Ar/O2 plasma from a metallic targetdashed line) were compared to copper oxide films prepared byputtering in an Ar plasma from a CuO target (solid line). The sig-ificant difference between the spectra is that the films prepared

rom ceramic targets contain satellite peaks which are typicallyound 7–10 eV above the main peaks associated with the copperp3/2 and 2p1/2 electrons, respectively for Cu2+ compounds, and

orrespond to shake-up electrons with 2p3d9 character [18]. Inhis film, the copper 2p3/2 and 2p1/2 peaks were found at 929.65nd 949.15 eV (Cu+), respectively; furthermore, curve fitting of theain 2p3/2 peak revealed a second peak with lower amplitude at

Fig. 3. XPS spectra of stoichiometric SnO2 films prepared by sputtering in an Ar/O2

plasma from a ceramic target and non-stoichiometric SnO2−x films prepared bysputtering in an Ar plasma from a SnO2 target.

Fig. 4. XPS spectra of (a) ZnO and (b) V2O5 films.

10 M. Amani et al. / Sensors and Actuators B 162 (2012) 7– 13

ore an

9bscOgpcp1rTfi1

Fig. 5. SEM micrographs of (a) CuO–Cu2O after testing, Cu2O (b) bef

31 eV (Cu2+). The broadening of the 2p3/2 peak towards higherinding energies combined with the presence of the satellite peaksuggests that these copper oxide films prepared by sputtering fromeramic targets in an Ar plasma have mixed oxidation states [19].n the other hand, by sputtering in Ar/O2 plasma from a metal tar-et, a narrow 2p3/2 peak at 929.65 eV was observed and no satelliteeaks were present, in the spectra indicating that the film is largelyomposed of Cu2O. Non-stoichiometric SnO2−x films were also pre-ared by sputtering from an oxide target in an argon plasma. A.12 eV shift in both the O 1s and the Sn 3d peaks was observedelative to those peaks for a SnO2 reference film, as shown in Fig. 3.

his is consistent with the shift typically described in the literature,or SnO [20]. Fig. 4 shows the XPS spectra of the 3d and 2p electronsn (a) ZnO and (b) V2O5. The Zn 2p3/2 and 2p1/2 peaks were found at022.65 eV and 1045.5 eV, respectively. In the spectra for vanadium

d (c) after testing and (d) V2O5, (e) ZnO and (f) SnO2−x after testing.

oxide, the 2p3/2 and 2p1/2 core levels were found at 514.25 eV and520.65 eV, respectively, both of which are in agreement with thosevalues typically observed in ZnO and V2O5 films [21,22]. Moreover,no significant shifting of these peaks was observed after testing ata maximum temperature of 650 ◦C.

3.2. Structural characterization

Fig. 5 shows the microstructure of several oxide catalysts sput-tered from metallic and ceramic targets in different Ar/O2 partialpressures. CuO–Cu2O films with mixed oxidation states, were sput-

tered from a ceramic target in argon plasma, and show surfacesdevoid of any fine structure. They also exhibit relatively smoothrounded protuberances with minimal surface area, typical of anas-sputtered film (Fig. 5a). However, the Cu2O films reactively

M. Amani et al. / Sensors and Actuators B 162 (2012) 7– 13 11

Fig. 6. Percent response in 8 ppm TATP and 9 ppm H O as a function of temperature for (a) tin oxide and (b) tungsten oxide catalysts, measured using the static testinga

smacvtiTvtnfifimSbmImt

3

atopaap4iTTfwratrl

investigate if TATP is indeed decomposing into H2O2 in the vicinityof the catalyst, which is not the primary decomposition productbased on a number of related studies. For example, Oxley andSmith [16,17] have determined that acetone and not hydrogen

2 2

pproach.

puttered from a metallic target in Ar/O2 mixtures, exhibited aicrocrystalline surface morphology within the rounded protuber-

nces of the as-deposited films, unlike those films prepared from aeramic target (Fig. 5b). When this catalyst surface was exposed toarious target gases, including H2O2 and TATP at a high tempera-ure, there was no apparent change in the fine microstructure butnstead an increase in the cluster/grain size was observed (Fig. 5c).he reactively sputtered V2O5 catalysts prepared from a metallicanadium target in an Ar/O2 plasma (Fig. 5d), revealed the forma-ion of acicular needles as well as a very fine fibrous morphology orano-hair on the surface of the sensor. The surface area of these veryne microstructural features, associated with reactively sputteredlms produced from metallic targets, was estimated to be orders ofagnitude greater than that of typical sputtered film morphology.

nO2−x and ZnO films were also deposited from a ceramic targety sputtering in an argon plasma (Fig. 5e and f), and a morphologyore similar to that of sputtered CuO–Cu2O (Fig. 5a) was observed.

n the case of SnO2−x, some micro-cracking of the films due to ther-al stress due to high temperature testing became apparent over

ime.

.3. Gas sensing properties

The responses of a tin oxide catalyst and a tungsten oxide cat-lyst to TATP and hydrogen peroxide are shown in Fig. 6. Here,he percent response to 8 ppm TATP and 9 ppm H2O2 as a functionf temperature were measured using the static testing approach. Arecipitous drop in the response for tin oxide occurs until a temper-ture of 300 ◦C, at which point the response increases dramaticallys a function of temperature. We believe this is due to the decom-osition of TATP by the catalyst, for which the response peaks at00 ◦C. It should be noted that the response for hydrogen perox-

de using the same catalyst has a similar behavior and tracks theATP response closely. The response of a tungsten oxide catalyst toATP and hydrogen peroxide is shown in Fig. 6b and has very dif-erent characteristics than that of tin oxide. No significant responseas observed at temperatures below 380 ◦C, at which point the

esponse then increased dramatically until a maxima was observed

t 460 ◦C. As was the case for the tin oxide catalyst, the response ofhe sensor to H2O2 was very similar and closely tracks the TATPesponse. In both cases the response to hydrogen peroxide wasarger in magnitude as might be expected since the decomposition

of the TATP is not complete during the test cycle and the startingconcentration of TATP is 8 ppm whereas the starting concentrationof H2O2 is 9 ppm.

The gas sensor response to TATP using a number of differentcatalysts is shown in Fig. 7. Here, the response to 8 ppm TATPas a function of temperature using several catalysts includingWO3–TiO2, V2O5, SnO2−x, Nb2O3 and ZnO was measured usingthe static testing approach. Each of the five oxide catalysts showa unique response as a function of temperature, and show fea-tures ranging from almost no peak response for Nb2O3 to welldefined peaks for ZnO. Due to the catalytic decomposition ofTATP, it is expected that the target gas consists not only of amixture of air and TATP vapor, but also contains some H2O2vapor. To verify this, several experiments were undertaken to

Fig. 7. Response as a function of temperature using several catalysts to detect TATP(8 ppm) including WO3–TiO2, V2O5, SnO2−x and ZnO, measured using the statictesting approach.

12 M. Amani et al. / Sensors and Actuators B 162 (2012) 7– 13

d (b)

ppao

F4

Fig. 8. Response of (a) CuO–Cu2O an

eroxide is the predominant decomposition product of highurity TATP in air and they have measured the activation energynd determined the decomposition kinetics for the degradationf TATP.

ig. 9. Response of SnO2−x in several concentrations of (a) acetone and (b) H2O2 at15 ◦C.

Cu2O in acetone and H2O2 at 330 ◦C.

Fig. 8 shows the response of a CuO–Cu2O and Cu2O catalyst to8 ppm acetone and H2O2 at 330 ◦C after a 2-min exposure to thetarget gas. From this figure it can be seen that the relative responseto acetone is considerably greater than that of H2O2, which is con-sistent with other investigations using thermodynamic based gassensors for detection [23]. These studies support our findings thatsignificant exothermic heat effects are observed for hydrocarbons.Not only is the magnitude of the response to acetone for bothCuO–Cu2O and Cu2O significantly greater than that of hydrogenperoxide but it is also of opposite sign, which implies that thesensitivity of the thermodynamic sensor platform to acetone is sig-nificantly greater than that of the other gaseous products derivedfrom the decomposition of TATP. Furthermore, this figure indicatesthat CuO–Cu2O (Fig. 8a) is a more selective catalyst for hydrogenperoxide than that of Cu2O (Fig. 8b), since both a smaller heat ofreaction in the presence of acetone and a more negative heat ofreaction for H2O2 were observed. All of this suggests that theseresults can be further improved upon by tailoring the sputteringconditions and thus, the oxidation state of the catalyst films. Theeffect of acetone and H2O2 concentration on the magnitude of theresponse using tin oxide as the catalyst is shown in Fig. 9, wherethe temperature was held constant at 415 ◦C (maximum sensitiv-ity to H2O2 and TATP shown in Fig. 6a). Again the responses are ofopposite sign, and it was found that the TATP response tracks thatof hydrogen peroxide, which can be readily detected at concentra-tions as low as 3.5 ppm. Based on the results presented in Fig. 9, weshould also be able to reliably detect low levels of acetone in the gasstream generated by the decomposition of TATP. Since we do notsee any evidence of acetone in the output signal, we can concludethat the TATP is detected as hydrogen peroxide. However, it has alsobeen extensively cited in the literature [16,17] that under ambientconditions TATP primarily decomposes into acetone. Therefore, webelieve that we are sensing the catalytic decomposition of TATP,which is a much more specific reaction than the decomposition ofTATP to acetone in air ambient, and this specific mechanism shouldhelp mitigate false positives using this sensor platform.

4. Conclusion

A small footprint, thermodynamic based gas sensor capable

of detecting TATP at ppm levels in air under ambient conditionswas demonstrated. This robust sensor platform has the potentialof exhibiting the necessary sensitivity, selectivity and responsetime to be used for the continuous monitoring of TATP and its

nd Act

pvIobaSwpucaboaTt

A

rRTtuUt2o

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

M. Amani et al. / Sensors a

recursors. This would be particularly important in closed spacedenues where the public may be exposed to potential threats fromEDs employing TATP as the triggering device. A number of metalxide catalysts were able to detect TATP using the thermodynamicased gas sensor described within. We have shown that TATP vaport concentrations of 8 ppm in air can be detected readily usingnO2−x and ZnO catalysts, and its peroxide and acetone precursorsere able to be detected at even lower concentrations. Based on ourreliminary results, arrays of microheater sensors can be fabricatedsing multiple catalysts and combining the characteristic responseurves to uniquely identify the target molecules of interest. Inddition, the current thermodynamic detection scheme can easilye integrated into an existing micro-hotplate based MEMS metalxide gas sensing platforms by modifying the heater electronics anddding a mass flow controller to eliminate hydrodynamic effects.his can add a second independent variable to minimize false posi-ives and improve long term stability in existing gas sensor systems.

cknowledgements

The authors wish to thank the Department of Homeland Secu-ity Center of Excellence for Explosives Detection, Mitigation andesponse and Sensor Tech, Savannah, GA for support of this work.he authors also thank Prof. Jimmy Oxley and Prof. James Smith athe University of Rhode Island for preparation of the TATP samplessed throughout the study. The authors would like to acknowledgeS Patent #7,329,389 B2, “Sensor Device and Method for Qualita-

ive and Quantitative Analysis of Gas Phase Substances”, Feb. 12,008, Michael L. Horovitz and Karl L. Anderson upon which somef the preliminary work was based.

eferences

[1] F. Dubnikova, R. Kosloff, Y. Zeiri, Z. Karpas, Novel approach to the detection oftriacetone triperoxide (TATP), its structure and its complexes with ions, J. Phys.Chem. A 106 (19) (2002) 4951–4956.

[2] I. Cotte-Rodriguez, H. Chen, R.G. Cooks, Rapid trace detection of triacetonetriperoxide (TATP) of complexation detections of desorption electrospray ion-ization, Chem. Commun. 5 (2006) 953–955.

[3] G.A. Buttigieg, A.K. Knight, S. Denson, C. Pommier, M.B. Denton, Characteri-zation of the explosive triacetone triperoxide and detection by ion mobilityspectrometry, Forensic Sci. Int. 135 (2003) 53–59.

[4] R. Rasanen, M. Nousiainen, K. Parakorpi, M. Sillanpaa, L. Polari, O. Anttalainen,M. Utriainen, Determination of gas phase triacetone triperoxide with aspirationion mobility spectrometry and gas chromatography–mass spectrometry, Anal.Chim. Acta 623 (2008) 59–65.

[5] J. Wilkinson, C.T. Konek, J.S. Moran, E.M. Witko, T.M. Korter, Terahertz absorp-tion spectrum of triacetone triperoxide (TATP), Chem. Phys. Lett. 478 (2009)172–174.

[6] A. Stambouli, A. El Bouri, T. Bouayoun, M.A. Bellimam, Headspace GC/MS detec-tion of TATP traces in post-explosion debris, Forensic Sci. Int. 146S (2004)S191–S194.

[7] C. Mullen, A. Irwin, B.V. Pond, D.L. Huestis, M.J. Coggiola, H. Oser, Detection ofexplosives and explosives-related compounds by single photon laser ionizationtime-of-flight mass spectrometry, Anal. Chem. 78 (2006) 3807.

[8] C. Shen, J. Li, H. Han, H. Wang, H. Jiang, Y. Chu, Triacetone triperoxide using lowreduced-field proton transfer reaction mass spectrometer, Int. J. Mass Spec-

trom. 285 (2009) 100–103.

[9] R.S. Ladbeck, U. Karst, Determination of triacetone triperoxide in ambient air,Anal. Chem. 482 (2003) 183–188.

10] H. Lin, K.S. Suslick, A colorimetric sensor array for detection of triacetone triper-oxide vapor, J. Am. Chem. Soc. 132 (2010) 15519–15521.

uators B 162 (2012) 7– 13 13

11] E. Capua, R. Cao, C.N. Sukenik, R. Naaman, Detection of triacetone triperox-ide (TATP) with an array of sensors based on non-specific interactions, Sens.Acuators B 140 (2009) 122–127.

12] N. Barsan, D. Koziej, U. Weimar, Metal oxide-based gas sensor research: howto? Sens. Actuators B 121 (2007) 18–35.

13] N.L. Hung, H. Kim, S.K. Hong, D. Kim, Enhancement of CO gas sensing propertiesof ZnO thin films deposited on self-assembled Au nanodots, Sens. Actuators B151 (2010) 127–132.

14] G. Eranna, B. C/Joshi, D.P. Runthala, R.P. Gupta, Oxide materials for developmentof integrated gas sensors: a comprehensive review, Crit. Rev. Solid State Mater.Sci. 29 (2004) 111–188.

15] W.H. Zhang, W.D. Zhang, L.Y. Chen, Highly sensitive detection of explosivetriacetone triperoxide by an In2O3 sensor, Nanotechnology 21 (2010) 315502.

16] J.C. Oxley, J.L. Smith, K. Shinde, J. Moran, Determination of the vapor den-sity of triacetone triperoxide (TATP) using a gas chromatography headspacetechnique, Propellants Explos. Pyrotech. 30 (2) (2005) 127–130.

17] J.C. Oxley, J.L. Smith, H. Chen, Decomposition of a multi-peroxidic compound:triacetone triperoxide (TATP), Propellants Explos. Pyrotech. 27 (4) (2002)209–216.

18] W.Y. Yang, W.G. Kim, S.W. Rhee, Radio frequency sputter deposition of singlephase cuprous oxide using Cu2O as a target material and its resistive switchingproperties, Thin Solid Films 517 (2008) 967–971.

19] C.I. Pearce, R.A.D. Pattrick, D.J. Vaughan, C.M.B. Henderson, G. van der Laan,Copper oxidation state in chalcopyrite: mixed Cu d9 and d10 characteristics,Geochimica 70 (2006) 4635–4642.

20] S. Hwang, Y.Y. Kim, J.H. Lee, D.K. Seo, J.Y. Lee, H.K. Cho, Irregular electrical con-duction types in tin oxide thin films induced by nanoscale phase separation,Journal of the American Ceramic Society 95 (1) (2012) 324–332.

21] S. Uthanna, T.K. Subramanyam, B.S. Naidu, G.M. Rao, Structure–composition–property dependence of reactive magnetron sputtered ZnO thin films, Opt.Mater. 19 (4) (2002) 461–469.

22] G. Silversmit, D. Depla, H. Poelman, G.B. Marin, R. Gryse, Determinationof the V2P XPS binding energies for different vanadium oxide oxidationstates (V5+ to V0+), J. Electron Spectrosc. Relat. Phenom. 135 (2–3) (2004)167–175.

23] E. Vereshchaginam, R.A.M. Wolters, J.G.E. Gardeniers, Measurement of reactionheats using a polysilicon based microcalorimetric sensor, Sens. Actuators A 169(2) (2011) 308–316.

Biographies

Matin Amani received his B.S. degree in electrical engineering at URI in 2011, andis currently working on his B.S. degree in chemical engineering and M.S. degree inelectrical engineering. His research interests include oxide thermoelectrics, nano-materials, sensors for harsh environments and gas sensors.

Yun Chu received his M.S. degree in chemistry from Beihang University in 2008. He iscurrently working on his Ph.D. in chemical engineering at URI. His research interestsinclude gas sensors and materials discovery using combinatorial chemistry.

Kellie L. Waterman received her B.S. degree in chemical engineering in 2011, andis in processes of obtaining her M.S. degree at WPI.

Caitlin M. Hurley is a senior undergraduate student studying chemical engineeringat URI.

Michael J. Platek received his B.S degree in physics from Indiana University ofPA (1984) and M.S. degree in physics from Wesleyan University (Middletown, CT)(1986). He is the laboratory manager for the URI Sensors and Sensors TechnologyPartnership. His interests include fabrication as well as characterization of nano-materials, development of gas sensors, bio-sensors and MEMS.

Otto J. Gregory received his B.S. degrees in chemical engineering and ocean engi-neering and his M.S. degree in chemical engineering from URI, in 1975 and 1977,respectively and his Ph.D. in engineering from Brown University in 1984. He is cur-rently Distinguished Engineering Professor at URI and Director of the URI Sensors

and Surface Technology Partnership. Dr. Gregory has published over 80 peer-reviewed journal articles and filed 25 US patents. His work has largely focused onsensors for harsh environments, wide bandgap semiconductors and chemical sen-sors, and has been funded by the gas turbine engine industry, NASA, DOE, Air Force,NSF and DHS.