Nimmakayala V. V. Subbaraoa and Parameswar K. Iyer*ab
The fabrication of a two terminal sensor device based on a histidine substituted perylene diimide
(PDI-HIS) thin film for the sensitive detection and quantification of ammonia (NH3) vapors by monitoring
the changes in its current intensity is reported at room temperature under ambient conditions. The thin
film morphological variations of the drop cast PDI-HIS films before and after exposure to NH3 vapors
are characterized by FESEM and TEM confirming the diffusion/adsorption of the NH3 vapors. The
solution cast PDI-HIS thin film gas sensor device exhibited rapid, highly sensitive and selective vapor
phase response towards NH3 with a detection limit as low as 0.56 ppm which is much lower than the
maximum permissible limit set for NH3 (25 ppm) for prolonged exposure. Furthermore, control sensing
experiments performed using alkyl substituted PDI (PDI-n-octyl) demonstrated that the presence of
histidine groups at the imide position of PDI-HIS drastically affects the solid-state aggregation mode as
well as redox potential that ultimately enhances the sensing response of the device. The key performance
parameters of the device such as sensitivity, response/recovery time, selectivity, recyclability, stability and
detection limit demonstrated the protocol as simple, reliable, cost-effective and most efficient in
performing NH3 detection under very realistic conditions.
Introduction
Electrical sensors have received incredible attention owing totheir ease in the detection or analysis of information of variousanalytes through electrical signals.1 Moreover, the developmentof a semiconductor device is simple, since the signal does notnecessitate intricate detection tools. Among electrical sensors,three-terminal transistors and two-terminal resistors are themost promising candidates for the advancement of low-cost,portable, low power, ultrasensitive and selective applicationsin chemical, biological and physical monitoring.2 Such ultra-sensitive sensors can be utilized in detecting traces of varioustoxic gases and volatile organic compounds (VOCs) even atremote locations.3
Ammonia (NH3) detection has received immense attentionamong various volatile species owing to its severe effect on theenvironment as well as human health.1 NH3 is highly toxic andcorrosive that can be easily spread into the environment due to
its widespread applications in fertilizers, refrigeration systems,manufacturing of dyes, drugs, synthetic fibers, plastics, etc.4
Moreover, ammonium nitrate (NH4NO3) found in many explosivesgradually decomposes to release trace amounts of NH3 whichare essential to be monitored in order to prevent the lethalaccidents.5 Ammonia, in its flammability range of 15% to 28%by volume can easily explode or catch fire.6 Although, it findsapplications in various sectors, its exposure in high concentrationsis a major threat to the human health. The lower limit of humanNH3 perception by smell is around 50 ppm.7 However, even belowthis limit, it is irritating to the respiratory system, skin as well aseyes.8,9 The long term (8 h) permissible concentration of NH3 forworkers is 25 ppm.10 Therefore, it is highly desirable to designand fabricate a long-term-reliable, highly-sensitive, miniaturized,room-temperature-efficient and low power consuming NH3
gas sensor, which can detect and monitor NH3 concentrationsurrounding the environment in real time.
In the past decade, there have been numerous reportson NH3 detection based on nanostructured metal oxides,11–16
plasmophores,17 conducting polymers,18–22 carbon nano-tubes23–25 and nanostructured graphenes.26–29 However, a veryfew small molecule based sensors30–35 with improved propertiessuch as lower operating temperature, high stability, fastresponse/recovery time, high selectivity and low detection limithave been reported.36 Among small molecules, perylene diimidederivatives (PDIs) are an important class of materials for the
a Centre for Nanotechnology, Indian Institute of Technology Guwahati, Guwahati,
781039, Indiab Department of Chemistry, Indian Institute of Technology Guwahati,
Guwahati-781039, India. E-mail: [email protected]; Fax: +91 3612582349;
Tel: +91 3612582314
† Electronic supplementary information (ESI) available: It includes syntheticschemes, characterization spectra, I–V curve and sensing study of PDI-n-octyl.See DOI: 10.1039/c5tc02521d
Received 13th August 2015,Accepted 15th September 2015
fabrication of devices due to their numerous features suchas high absorption coefficient, good chemical and thermalstability, better electronic properties and excellent photostabil-ity. Owing to their unique features, these materials have foundversatile applications in organic field-effect transistors (OTFTs),organic solar cells and sensor devices.37–39 Recently, varioustypes of PDIs composed of self-assembled aggregates, nanorods,nanobelts, nanofibres etc. have been developed for detectingvolatile species such as amines, explosives, etc.40,41 Molecularmodification has a vital role in improving the performance ofsuch PDI-based organic semiconductor devices. In this context,modifications of PDIs have been achieved by either substitutingelectron withdrawing groups at the core position42 (so calledbay position) or introducing solubilizing groups at the imideposition.43 Amino acids have achieved much attention as activegroups to tune the structure of PDIs at the imide position, sincethey can deliver a range of supramolecular structures withvariable degrees of arrangement due to the site-specific hydro-gen bonding.44 Though, most of the efficient sensor devicesreported for NH3 detection are based on the functionalizationof perylenes at the core position45,46 the imide position hasbeen modulated mainly for providing solubility to the system.Hence, meagre attention has been paid in designing perylenederivatives with suitable receptors at the imide position that canbe utilized as tools to fabricate devices with a specific aim ofsensing gases such as NH3 with high sensitivity and selectivity.
Herein, we report the fabrication of a low cost two terminalsensor device using a PDI-HIS (Scheme 1a) thin film as anactive layer and systematically investigated its sensing responsetowards NH3 vapors by measuring the variation of currentintensity. The sensor displayed notable features viz. (i) solutionprocessibility, (ii) fast signal response/recovery time (28 s/40 s),(iii) lower limit of detection at sub-ppm levels (0.56 ppm), (iv)high selectivity over other volatile species, and (v) operationalability at room temperature under ambient conditions. A controlsensing experiment was also performed using simple alkylsubstituted PDI i.e. PDI-n-octyl (Scheme 1b) to confirm theeffect of amino acid attached onto the imide position of PDIon sensing behavior. Finally, we have demonstrated that thevariation of substituent groups at the imide position of perylene
dyes remarkably affects the solid-state aggregation modes aswell as redox potential that cause the significant impact on thegas responses.
Experimental sectionMaterials and characterization
Perylene-3,4,9,10-tetracarboxylic dianhydride, octylamine, andzinc acetate were purchased from Sigma-Aldrich and were usedas received without any further purification. Histidine andimidazole were purchased from Himedia Leading BioSciencesCompany and Alfa Aesar, respectively. Spectroscopic gradesolvents were used for all the experiments. The 1H-NMR and13C-NMR spectra were recorded on a Bruker 600 MHz NMRspectrometer. The high resolution mass spectra (HRMS) wererecorded on a Micromass Q-TOF ESI-MS instrument (modelHAB273). Field emission scanning electron microscope(FESEM) images were recorded on a Sigma Carl ZEISS scanningelectron microscope. Transmission electron microscopic (TEM)studies were done using a Tecnai G2 F20 S-twin JEOL 2100transmission electron microscope. The UV-Vis absorption spectrawere recorded on a Perkin Elmer Lambda-35 spectrometer. FT-IRwas recorded on a Perkin Elmer spectrometer with samplesprepared using KBr pellets. Electrochemical measurements werecarried out using a CH instruments Model 700D series. Thethickness of the deposited films was optimized using a profilo-meter (Dektat-150). An electrochemical workstation consisted of athree electrode system viz. Ag/AgNO3 as the reference electrode, aplatinum wire as the counter electrode and glassy carbon as theworking electrode. A 0.1 M tetra n-butyl ammonium hexafluoro-phosphate (TBAPF6) in acetonitrile (CH3CN) and ferrocene wasused as a supporting electrolyte and an internal reference,respectively at a scan rate of 50 mV s�1 in an inert atmosphere.
Synthesis of PDI-HIS and PDI-n-octyl
PDI-HIS and PDI-n-octyl were synthesized using a modifiedreported procedure.47,48 A mixture of perylene-3,4,9,10-tetracarboxylic dianhydride (300 mg, 0.76 mmol), histidine(260 mg, 1.67 mmol) and/or octyl amine (216 mg, 1.67 mmol),zinc acetate (catalytic amount) and 5.0 g of imidazole washeated at 140 1C for 8 h under continuous stirring. The reactionmixture was then allowed to cool and poured into water. To themixture, 2.0 M HCl was added under continuous stirring toobtain a precipitate. This was then centrifuged and washedseveral times with water followed by drying under vacuum toget the red colored solid product. PDI-n-octyl was furtherpurified by column chromatography using hexane–chloroformas an eluent.
The detailed synthetic scheme and the characterizationspectra are presented in Schemes S1, S2 and Fig. S1–S4 ofthe ESI.†
Device fabrication and characterization
For the fabrication of a two terminal sensor device, low costmicroscopic glass slides (1 cm � 2 cm) were used as substrates.The glass substrates were cleaned in piranha solution (3 : 1/H2SO4 : H2O2) for 1 h and washed several times with deionizedwater followed by sonication. The cleaned substrates were thendried under vacuum at 100 1C. A channel of 30 mm length (L)and 2000 mm width (W) was obtained by depositing 150 nmthick aluminum (Al) electrodes on to cleaned glass substratesby thermal evaporation under high vacuum o10�6 mbar. Fromthe stock solution of PDI-HIS (10�3 M) in DMSO, a volume of15 mL was drop-cast over the channel between the electrodes.The solvent was fully dried by heating on a hot plate for 30 minat 80 1C to form a thin film (thickness 60 nm) across thefabricated Al electrodes. All the electrical characterizations ofthe devices were carried out under ambient conditions using aKeithley 4200-SCS semiconductor parameter analyzer at roomtemperature.
Vapor phase detection
For the vapor sensing experiment, the fabricated device waskept in a chamber of approximately 0.06 L in volume. A certainvolume of the analyte was injected into the test chamberby using a micropipette. The concentration of vapors wascalculated using the following eqn (1);49
Cppm ¼VmLDg mL�1
Mg mol�1VmL� 2:24� 107 (1)
where, Cppm is the required vapor concentration, VmL is thevolume of the liquid analyte, Dg mL�1 is the density of the liquid,VmL is the volume of the test chamber and Mg mol�1 is themolecular weight of the liquid analyte. All the subscriptsare the corresponding units. Sensing studies using variouscommercially available common analytes were carried out byinjecting the sample in a vial placed inside the chamberadjacent to the device at room temperature and under ambientconditions.
Results and discussionSensing studies
Sensing experiments were carried out using both PDI-HIS andPDI-n-octyl fabricated two terminal sensor devices with simplearchitecture as shown in Fig. 1a. The devices were kept in achamber and connected to a Keithley 4200 SCS semiconductorparameter analyzer to perform electrical characterization. ForI–V measurements, the devices were tested in a vacuum firstand then under ambient conditions by sweeping the voltage
from �10 V to +10 V to check their stability in the realenvironment. The I–V curve obtained for a PDI-HIS fabricatedthin film demonstrated a good conducting behavior (Fig. 1b)compared to a PDI-n-octyl fabricated device (Fig. S5 in the ESI†).
To investigate the sensing response of a PDI-HIS film-baseddevice towards NH3, the concentrations of NH3 were system-atically varied from 100 ppm to 2 ppm and changes inthe current intensity were observed. As shown in Fig. 2a, asignificant increase in current was observed after exposing100 ppm NH3 vapors to the chamber that was seen enhancingcontinuously with very short response time. Upon turning offthe NH3 source, current intensity recovered quickly to its initiallevel. Thus, it can be concluded that the device exhibitsexcellent response and recovery time for the NH3 detection.To perform the quantification by this device, the chamber wasexposed to 50, 30, 20, 10, 5 and 2 ppm NH3 vapors to obtainsimilar responses with lower current intensity that variedaccording to the concentrations. The corresponding sensitivity(S) was calculated using the formula S = DI/I0, where DI is thechange in current intensity upon NH3 vapor exposure and I0 isthe initial current in absence of NH3 vapors. The sensitivityof the device was then plotted as a function of differentNH3 concentrations. The curve showed a linear response withincrease in the concentration of ammonia as shown in Fig. 2b.
The response and recovery times are considered as importantparameters for any gas-sensing devices. The response time is thetime needed for a sensing device to reach 90% of total currentchange after the supply of analyte vapors, whereas the recoverytime is the 90% of current change to return to its originalposition after the analyte vapor source is turned off. The response
Fig. 1 (a) Schematic diagram of the device structure and (b) I–Vcharacteristics of the PDI-HIS thin film.
Fig. 2 (a) Response of the PDI-HIS sensor device towards variousconcentrations of NH3 vapors ranging from 100 ppm to 2 ppm and (b)sensitivity of a device as a function of NH3 concentration.
time of the device to 100 ppm of NH3 was calculated to be verylow as 28 s, whereas the recovery time was found to be 40 s(Fig. 3a). To the best of our knowledge, such low cost highstability devices have not been reported yet with such a remark-able response/recovery time for NH3 detection and high sensitiv-ity making the current protocol highly reliable and useful for therapid detection of NH3 vapors. To further verify the viability ofthis system for practical applications, the recyclability of thedevice was checked by exposing certain concentration of NH3
vapors repeatedly after certain interval of time into the testchamber. Almost similar increment in current intensity (Fig. 3b)was observed after each exposure confirming the feasibility ofthe sensor device for realistic use.
The stability of the sensor device is another crucial and mostvital criterion for practical applications. After leaving the deviceunder ambient conditions for B4 weeks, the sensor device wasagain tested with different NH3 concentrations (Fig. 4a). It wasobserved that the change in current intensity in the devicesshowed excellent reversibility and stability under continuousoperation and storage conditions with negligible to no loss inactivity even after prolonged testing or storing under ambientconditions.
The limit of detection (LOD) is another vital parameter to bedetermined for a sensor device. To calculate the LOD, acalibration curve was constructed by plotting the maximumcurrent intensity against the concentration of NH3 (Fig. 4b).The curve demonstrates a good linear relationship with thecorrelation coefficient (R2) value of 0.9994. The limit of detection(LOD) was calculated using the formula LOD = 3.3s/S, where, ‘s’is the relative standard deviation of the current response of the
device in the absence of NH3 and ‘S’ is the slope of the calibrationcurve. The LOD was found to be 0.56 ppm which is compatiblewith the minimum permissible limit set for NH3 in the workingenvironment i.e. 25 ppm for long exposure upto 8 h.
Effect of humidity and film thickness on sensing
All the sensing experiments were carried out in the laboratoryatmosphere (relative humidity-RH Z 60%) that provides anenvironment closer to the ultimate condition of the sensordevices. To further study the effect of humidity on the NH3
response, sensing experiments were carried out by exposing50 ppm NH3 on three different devices at humidity levels RH0%, 60% and 90%, respectively (Fig. S6 of ESI†). In a vacuum(RH 0%), the sensor device gave highest response for NH3 anddecreases by B18% on increasing the RH upto 60%. However,on further increasing the humidity levels from 60% to 90%, theresponse of the device towards NH3 decreases by only B8%.Low sensitivity of the device at higher humidity is consistentwith the weak adsorption/diffusion of ammonia vapors on thesurface of the film. It is believed that under higher humidityconditions, there is a probability of a strong competition betweenwater and ammonia molecules to get adsorbed/diffused at thereceptor sites (histidine units) of the PDI-HIS film. Thus, it can beconcluded that the sensor device is functional over a wide rangeof humidity levels and can be used for sensing NH3 under realenvironmental conditions.
To monitor the effect of film thickness on sensitivity, threedevices with variable thicknesses (30, 60 and 100 nm) wereoptimized and investigated for their sensing response towards100 ppm NH3 (Fig. 5 and Fig. S7 of ESI†). It was found that thesensitivity of the devices rises by B3.3 times on increasing thethickness from 30 nm to 60 nm with slight increment (1 s/4 s)in the response/recovery time. However, on further increasingthe thickness from 60 nm to 100 nm, the sensitivity increasesby only B1.3 times with much larger increment (26 s/16 s) inresponse/recovery time. The high response time in a thickerfilm may be attributed to the diffusion of a large number ofNH3 molecules into the film which thereby takes a much longertime to get recovered. These studies helped in concluding thatthe thickness of a film affects the sensor response and a devicewith film thickness of B60 nm is generally appropriate for NH3
detection considering the response/recovery time and bettersensitivity observed here.
Fig. 3 (a) Response and recovery time of the NH3 sensor and (b) recycl-ability test with 100 ppm of NH3.
Fig. 4 (a) Sensing response of the PDI-HIS sensor device with differentconcentration of NH3 before and after 4 weeks of storage under ambientconditions and (b) calibration curve for calculating the detection limit.
Fig. 5 Effect of film thickness on (a) sensitivity of the device and (b)response/recovery time after exposing 100 ppm NH3.
To validate the selectivity of the device, similar sensing experi-ments were also performed with eight common volatile organicsolvents (chloroform, acetone, methanol, isopropanol, ethanol,ethyl acetate, THF and hexane) by exposing five times higherconcentration than NH3 (Fig. 6a). Interestingly, the deviceresponses to even 1000 ppm of common organic analytes werenegligible in comparison to 200 ppm of NH3. To date, suchoutstanding selectivity and reusability of NH3 sensors withremarkably low detection limits has marked this sensor deviceas a rare example available in literature. To further explore theselectivity of this system, the device was also exposed to variousother common organic amines such as cyclohexyl amine,triethyl amine, diisopropyl amine, pyridine and diethyl amine,since all these vapors have biological or environmental con-sequences. From Fig. 6b, it can be concluded safely that theresponse of a sensor device towards various other amines wasinsignificant compared to NH3 even at much higher concentrations.
Sensing mechanism
Generally, the mechanism of organic semiconductor basedconductometric gas sensors involves the adsorption of analytesonto the surface via dissociation and/or diffusion followed bythe formation of a possible charge transfer complex thateventually leads to the variation of majority charge carriersthan in the current.45,50 Depending upon the chemical natureof both the species, the binding can be weak or strong via somechemical interactions such as hydrogen bonding, dipole–dipoleinteractions etc. The molecular self-assembly due to p–p inter-actions in such semiconductors also provides efficient path-ways for charge migration or transport. Change in redoxpotential is another major factor associated with the mobilityof charge carriers.1,30,32,42,51,52
To understand the sensing mechanism, the surface morphologyof the PDI-HIS fabricated film was initially studied before andafter exposure to NH3 vapors via FESEM and TEM, respectively.The PDI-HIS film showed an amorphous self-assembled nano-structured network throughout the surface (Fig. 7a and c) whichindicates the huge surface area available for the adsorption and/or diffusion of gas molecules (Fig. 7b and d) that subsequentlyleads to the remarkable change in the current and response
time. On introducing NH3 vapors onto the PDI-HIS fabricatedfilm, the current intensity displayed significant increment byseveral orders of magnitude with an excellent response time.This can be explained by donor–acceptor like complexation53
between n-type organic semiconductor materials i.e. PDI-HIS(electron acceptor) and NH3 (electron donor) molecules. Since,each PDI-HIS molecule consists of two –COOH units at the imidepositions that have the tendency to bind with NH3 through anacid–base interaction, numerous ion pairs are expected to formin the film that can drastically increase the ionic conductivity andconsequently the current. In addition, the presence of histidinegroups will further boost up the sensitivity of the device due tothe favorable hydrogen bonding interactions with NH3. Theaggregation of PDI-HIS molecules due to p–p interactions isanother possible factor responsible for such a high sensitivity,since it can provide efficient charge transport into the material.
Furthermore, to validate the proposed sensing mechanism,we have monitored the changes in the IR spectrum of PDI-HISbefore and after exposing it to NH3 vapors (Fig. S8 of ESI†). Theband observed at 3418 cm�1 corresponds to the stretchingvibrations of carboxylic –OH and/or histidine –NH, whichshowed a significant shift (3347 cm�1) immediately after exposingit to NH3 vapors confirming the adsorption/diffusion of ammoniavapors onto the surface of the film via hydrogen bonding inter-actions/acid–base interactions that subsequently leads to theincrement in ionic conductivity and current.
In order to elucidate the effect of the amino acid anchoringgroups on PDI-HIS, sensing studies were also performed usinga fabricated device of a model compound PDI-n-octyl that doesnot consists of any active groups (e.g. COOH) at the imideposition except long alkyl chains. Interestingly, no significantenhancement in the current (0.20 times) was observed afterintroducing NH3 vapors to the PDI-n-octyl fabricated film evenat much higher concentrations (Fig. S9, ESI†). These resultsconfirm that the solubilizing/anchoring group attached onto theimide position of PDI-HIS played a vital role in the signallingprocess. The active –COOH groups on either side of PDI-HIS are
Fig. 6 Sensing response of the PDI-HIS sensor device with vapors ofvarious common (a) organic solvents and (b) amines. Concentrations ofNH3 and other analytes were 200 and 1000 ppm, respectively.
Fig. 7 FESEM and TEM images of the PDI-HIS thin film (a), (c) before and(b), (d) after exposure of NH3 vapors. (Inset: SAED pattern of PDI-HIS).
responsible for the ultra-sensitivity, since they may possiblyincrease the number of H+ ions which in turn enhance themobility of protons in the molecule and ultimately the current.The morphology of the PDI-n-octyl film was studied via FESEM
and TEM analysis (Fig. 8) which confirms that PDI-n-octylmolecules exhibit single-crystalline belt like microstructuresunlike amorphous aggregates as observed in the PDI-HISfilm. The single-crystalline micro/nanostructures provide highlyefficient channels for charge carrier transport due to the largesurface area for diffusion of gas molecules.30
However, the PDI-HIS molecule showed a marvellous responsetowards NH3 compared to the PDI-n-octyl molecule even though itis amorphous in nature. This indicates that there is a possibility ofsome additional factors that are also involved in the sensingmechanism. Thus, to gain further insight into the mechanismand differential response of the NH3 vapors towards PDI-HIS andPDI-n-octyl fabricated devices, electronic parameters of both thesemiconductors were studied via cyclic voltammetry (CV) measure-ments (Fig. 9).
The values of reduction potential, HOMO/LUMO energiesand the estimated band gap are summarized in Table 1. TheLUMO energy level obtained for PDI-HIS (�4.07 eV) was signifi-cantly lower compared to PDI-n-octyl (�3.60 eV) that makes itmore capable for the occurrence of charge transfer30 betweenthe PDI-HIS and NH3 molecules that ultimately leads to thedramatic increment in current (B33 times). Thus, it can beconcluded that the presence of active receptors (histidine units)in PDI-HIS is one of the key factors for increasing the sensitivityof the device via lowering the LUMO energies that enabled theefficient charge exchange between PDI-HIS and NH3 molecules,making the sensor device highly sensitive towards NH3.
Comparative study
A comparative study was also carried out to demonstrate theadvantage of this sensing platform over the previously reportedworks. The various parameters of the PDI-HIS sensor devicetowards NH3 vapors are compared with recent literature andpresented in a tabulated form (Table 2). It is noteworthy that thesensing performance of the PDI-HIS thin film is not only superiorto other materials such as polymers, metal oxides, graphenes,CNT etc. but also the best and the most economical platformamong various other small molecule based sensors towards NH3.
Conclusion
In conclusion, a new and efficient platform for the vapor phasedetection of NH3 using a two terminal sensor device based on
Fig. 8 (a) FESEM and (b) TEM image of PDI-n-octyl. Inset: SAED pattern ofPDI-n-octyl.
Fig. 9 Cyclic voltammograms of PDI-HIS and PDI-n-octyl films on aglassy carbon electrode in CH3CN.
Table 1 Frontier molecular orbital energies as estimated from cyclicvoltammetry and optical absorption
PDI-HIS nanostructures fabricated on a simple glass substrateby the drop-casting method has been developed. The sensingparameters viz. sensitivity, selectivity, recyclability, response/recovery time and stability were studied that revealed theexcellent performance of the device with a very low detectionlimit of 0.56 ppm for NH3. We have also demonstrated thatthe molecular assemblies of PDI-HIS nanostructures, redoxpotential and ionic groups at the imide position are key aspectsfor the remarkable response of the device towards NH3. Thus,the ability of the as-fabricated sensor device to detect NH3
vapors at room temperature and under ambient conditionswith high sensitivity, selectivity and low fabrication cost makesthe protocol economical and feasible for potential applicationsunder chemical, biomedical and competitive environmentalconditions.
Acknowledgements
Financial assistance from Department of Science and Technology(DST), New Delhi, (No. SB/S1/PC-020/2014/000034), DST-MaxPlanck Society, Germany (IGSTC/MPG/PG(PKI)/2011A/48), andDepartment of Electronics & Information Technology, (DeitY)No. 5(9)/2012-NANO (Vol. II), is gratefully acknowledged. Theauthors acknowledge valuable discussions with M. Adil Afroz.The Central Instruments Facility, IIT Guwahati is acknowledgedfor instrument facilities.
Notes and references
1 Q. Meng, F. Zhang, Y. Zang, D. Huang, Y. Zou, J. Liu,G. Zhao, Z. Wang, D. Ji, C. Di, W. Hu and D. Zhu,J. Mater. Chem. C, 2014, 2, 1264–1269.
2 L. Li, P. Gao, M. Baumgarten, K. Mullen, N. Lu, H. Fuchsand L. Chi, Adv. Mater., 2013, 25, 3419–3425.
3 A. K. Nayak, R. Ghosh, S. Santra, P. K. Guha and D. Pradhan,Nanoscale, 2015, 7, 12460–12473.
4 M. Appl, in Ammonia: Principles and Industrial Practice,Wiley-VCH, New York, USA, 1999.
5 S. Pandey, G. K. Goswami and K. K. Nanda, Sci. Rep., 2013,3, 2082.
6 C. Cannilla, G. Bonura, F. Frusteri, D. Spadaro, S. Trocinocand G. Neri, J. Mater. Chem. C, 2014, 2, 5778–5786.
7 S. Budarvari, in The Merck Index, An Encyclopedia of Chemi-cals, Drugs and Biologicals, ed. J. M. O’Neil, Royal Society ofChemistry, Cambridge, 12th edn, 1996.
8 J. Wang, P. Yang and X. Wei, ACS Appl. Mater. Interfaces,2015, 7, 3816–3824.
9 B. Timmer, W. Olthuis and A. Berg, Sens. Actuators, B, 2005,107, 666–677.
10 T. Grady, T. Butler, B. D. MacCraith, D. Diamond andM. A. McKervey, Analyst, 1997, 122, 803–806.
11 N. Kumar, A. K. Srivastava, R. Nath, B. K. Gupta andG. D. Varma, Dalton Trans., 2014, 43, 5713–5720.
12 P. K. Kannan and R. Saraswathi, J. Mater. Chem. A, 2014, 2,394–401.
13 J. Wang, F. Yang, X. Wei, Y. Zhang, L. Wei, J. Zhang, Q. Tang,B. Guo and L. Xua, Phys. Chem. Chem. Phys., 2014, 16,16711–16718.
14 A. Pistone, A. Piperno, D. Iannazzo, N. Donato, M. Latino,D. Spadaro and G. Neri, Sens. Actuators, B, 2013, 186,333–342.
15 G. K. Mani and J. B. B. Rayappan, Sens. Actuators, B, 2013,183, 459–466.
16 X. Liu, N. Chen, B. Han, X. Xiao, G. Chen, I. Djerdjc andY. Wang, Nanoscale, 2015, 7, 14872–14880.
17 N. Stromberg and A. Hakonen, Anal. Chim. Acta, 2011, 704,139–145.
18 N. Chartuprayoon, C. M. Hangarter, Y. Rheem, H. Jung andN. V. Myung, J. Phys. Chem. C, 2010, 114, 11103–11108.
19 E. Danesh, F. M. Lopez, M. Camara, A. Bontempi, A. V.Quintero, D. Teyssieux, L. Thiery, D. Briand, N. F. Rooij andK. C. Persaud, Anal. Chem., 2014, 86, 8951–8958.
20 P. Stamenov, R. Madathil and J. M. D. Coey, Sens. Actuators,B, 2012, 161, 989–999.
21 S. Chen and G. Sun, ACS Appl. Mater. Interfaces, 2013, 5,6473–6477.
22 S. Kang, J. Chun, N. Park, S. M. Lee, H. J. Kim and S. U. Son,Chem. Commun., 2015, 51, 11814–11817.
23 J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng,K. Cho and H. Dai, Science, 2000, 287, 622–625.
24 S. Cui, H. Pu, G. Lu, Z. Wen, E. C. Mattson, C. Hirschmugl,M. G. Josifovska, M. Weinert and J. Chen, ACS Appl. Mater.Interfaces, 2012, 4, 4898–4904.
25 X. Feng, S. Irle, H. Witek, K. Morokuma, R. Vidic andE. Borguet, J. Am. Chem. Soc., 2005, 127, 10533–10538.
26 R. Ghosh, A. Singh, S. Santra, S. K. Ray, A. Chandra andP. K. Guha, Sens. Actuators, B, 2014, 205, 67–73.
27 M. Seredych, J. Teresa and T. J. Bandosz, Langmuir, 2010, 26,5491–5498.
28 R. Ghosh, A. Midya, S. Santra, S. K. Ray and P. K. Guha, ACSAppl. Mater. Interfaces, 2013, 5, 7599–7603.
29 F. Yavari, Z. Chen, A. V. Thomas, W. Ren, H. M. Cheng andN. Koratkar, Sci. Rep., 2011, 1, 166.
30 Y. Huang, L. Fu, W. Zou, F. Zhang and Z. Wei, J. Phys. Chem.C, 2011, 115, 10399–10404.
31 X. Yu, N. Zhou, S. Han, H. Lin, D. B. Buchholz, J. Yu,R. P. H. Chang, T. J. Marks and A. Facchetti, J. Mater. Chem.C, 2013, 1, 6532–6535.
32 W. Huang, K. Besar, R. LeCover, A. M. Rule, P. N. Breysseand H. E. Katz, J. Am. Chem. Soc., 2012, 134, 14650–14653.
33 M. Mirza, J. Wang, D. Li, S. A. Arabi and C. Jiang, ACS Appl.Mater. Interfaces, 2014, 6, 5679–5684.
34 F. Zhang, C. Di, N. Berdunov, Y. Hu, X. Gao, Q. Meng,H. Sirringhaus and D. Zhu, Adv. Mater., 2013, 25, 1401–1407.
35 H. W. Zan, M. Z. Dai, T. Y. Hsu, H. C. Lin, H. F. Meng andY. S. Yang, IEEE Sens. J., 2012, 12, 594–601.
36 J. Zhang, S. Wang, M. Xu, Y. Wang, H. Xia, S. Zhang, X. Guoand S. Wu, J. Phys. Chem. C, 2009, 113, 1662–1665.
37 B. A. Jones, A. Facchetti, M. R. Wasielewski and T. J. Marks,Adv. Funct. Mater., 2008, 18, 1329–1339.
38 E. Kozma and M. Catellani, Dyes Pigm., 2013, 98, 160–179.
