ZnO decorated luminescent graphene as a potential gas sensor at room temperature

CARBON

x x x ( 2 0 1 1 ) x x x –x x x

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/carbon

ZnO decorated luminescent graphene as a potential gas sensor at room temperature Gaurav Singh a,b,1, Anshul Choudhary a,b,2, D. Haranath a, Amish G. Joshi a, Nahar Singh a, Sukhvir Singh a, Renu Pasricha a,* a b

National Physical Laboratory, Council of Scientific and Industrial Research, Dr. K.S. Krishnan Road, New Delhi 110 012, India Netaji Subhas Institute of Technology, Faculty of Technology, Delhi University, New Delhi 110 078, India

A R T I C L E I N F O

A B S T R A C T

Article history:

We present a simplistic single step synthesis and a detailed study of the remarkable room

Received 16 June 2011

temperature gas sensing and photoluminescence (PL) properties of zinc oxide (ZnO) deco-

Accepted 25 August 2011

rated graphene oxide sheets (GrO). Investigation of opto-electronic properties reveal near

Available online xxxx

UV to blue PL and semiconducting behavior of ZnO–GrO sheets. ZnO nano-crystallites serve the dual purpose of acting as a nano-spacer between dried graphene sheets as well as a primary sensing transducer for the gas sensing applications. PL has been used as a tool to study the defects associated with the surface of the nanocrystallite’s trap levels and/or acceptor–donor recombinations. Time-resolved PL was used to determine free carrier or exciton lifetimes, a vital parameter related to quality of composite and device performance. Results are presented for the detection of common industrial toxins like CO, NH3 and NO for concentrations as low as 1 ppm at room temperature. A large sensor response and quick recovery time was observed at room temperature with preferred selectivity towards electron donor gases like CO and NH3.  2011 Elsevier Ltd. All rights reserved.

1.

Introduction

Carbon based nanostructured materials, especially the onedimensional (1-D) carbon nanotubes [1] and of late even the two dimensional graphene [2] have been of tremendous scientific and technological interest due to their unique physical and chemical properties [3]. Graphene being a perfect 2-D material with high specific surface area (2600 m2/g) available for sensing and metallic conductivity [4,5] at room temperature, is an ideal choice for sensing applications. Ultra-high sensitivity of order of single molecule detection has been claimed for NO2 [6]. These properties also make the 2-D graphene an ideal platform to load metal or semiconductor nanoparticles, organic and biological molecules for photo-

catalytic, optoelectronic, cellular imaging and drug delivery applications [7–10]. There have been reports involving surface doping of graphene sheets with Pt, Pd, Ag or Au nanoparticles mainly focusing on preventing aggregation of graphene sheets upon drying thereby getting improvements is various properties [7,11]. Very recently Cao et al. [12] and Williams and Kamat [9] suggested that graphene-semiconductor nanostructures, such as Gr-CdS and Gr-ZnO have shown improved optical switching and catalytic applications respectively. In the vast area of gas-sensing using nanostructures mostly of the gas sensors are based on polycrystalline tin oxide (SnO2) and zinc oxide (ZnO) systems that are renowned for their high sensitivity and chemical stability. In addition their low production costs and miniature sizes – have made

* Corresponding author: Fax: +91 11 45609310. E-mail address: [email protected] (R. Pasricha). 1 Nanoscale Science Program, University of North Carolina at Charlotte, NC-28233, United States. 2 Department of Physics, Indian Institute of Science Education and Research, Sector 81, Mohali 140 306, India. 0008-6223/$ - see front matter  2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.carbon.2011.08.050

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

2

CARBON

x x x ( 2 0 1 1 ) x x x –x x x

them ubiquitous for many practical applications [13,14]. However, these devices have high power consumption due to their high working temperatures generally of order of 250–600 C [14] in addition to poor chemical specificity. Accordingly, the development of metal/semiconductor-oxide sensors that can operate at room temperatures with high sensitivity and low production cost has attracted much attention. Composites involving coating of thin layer of tin-oxides on single and multi-walled carbon nanotubes (CNT’s) have shown potential for room temperature sensing [15]. Taking advantage of the excellent properties of the unrolled CNT’s, we herein report synthesis of luminescent ZnO decorated graphene sheets for potential application in commercially viable gas sensors at room temperature. The ZnO decorated reduced graphene oxide sheet (GrO) sheets were the result of single step synthesis wherein simultaneous reduction of graphite oxide (GO) and ZnO occur in ethanolic solution. The presence of ZnO nanoparticles prevents aggregation of graphene sheets upon drying. These ZnO decorated reduced graphene oxide (ZnO–GrO) sheets display distinct optoelectronic characteristics as compared to independently made ZnO nanoparticles, emitting near-UV to blue luminescence when excited with far UV radiation. Blue luminescence from solution processed organic compounds and their derivative is also significant for development of low cost nano devices [16,17]. We consider that this study is the first step towards the goal of utilizing composites of chemically reduced graphene with metal-oxides for gas sensing as well as opto-electronic applications. It is to be noted that the primary sensing mechanism is not based on charge transfer between adsorbed gas molecules and graphene oxide sheets leading to change in its conductance [18,19], instead, we propose a mechanism where, luminescent ZnO nanoparticles act as active sensing and transducer element and graphene oxide acts like a high conducting mesh. It amplifies the transduction resulting in large change in conductance as compared to previous results reported for chemically derived graphene based sensors [20]. Preliminary gas sensing results are presented for CO, NH3 and NO gas molecules at room temperature up to 1 ppm concentrations.

2.

Experimental

2.1.

Synthesis of GO and GrO

GO was synthesized using widely reported Hummers method [21]. The GO thus synthesized was thereafter re-suspended in ethanol and centrifuged at 2500 rpm for 30 min to remove large residual particles. The extracted solution is further centrifuged at 6000 rpm for 10 min to resulting in a uniformly dispersed suspension. Lithium hydroxide (LiOHÆH2O), was then added to the solution, and the reaction vessel was further sonicated at room temperature, and left undisturbed for 24 h for the reduction of GO to GrO.

2.2.

Synthesis of ZnO

Zinc oxide nanoparticles were synthesized as per method described in literature [22,23] with zinc acetate (Zn (Ac)2Æ2H2O) as precursor in an ethanol solution, followed by sonication.

LiOHÆH2O was added to the solution, and the reaction vessel was further sonicated at room temperature, allowing the particles to complete the growth process.

2.3.

Synthesis of ZnO–GrO

0.6 wt.% GO in ethanol was taken and mixed with (10 ml, 0.01 M) Zn (Ac)2Æ2H2O. The resulting solution was kept in sonication bath for 10 min at room temperature. After sonication, (10 ml, 0.002 M) LiOHÆH2O solution in ethanol was added to the above solution base. The resulting solution was again sonicated for 5 min and final solution was left undisturbed for 24 h to ensure complete reduction of GO and Zn (Ac)2Æ2H2O. Varying concentration of ZnO–GrO were synthesized by carefully adjusting the molar ratio of ZnO to GO. The GO/ZnO ratio was varied from 0 to 2 wt.% in steps of 0.15 wt.%, with resultant as ZnO–GrO-1 (GO-0.15 wt.%), ZnO–GrO-2 (GO-0.6 wt.%), ZnO– GrO-3 (GO-1.2 wt.%) and ZnO–GrO-4 (GO-2 wt.%). The composition with 0.6 wt.% GO termed as ZnO–GrO-2 was found to be most optimal for gas sensing applications with regard to sensitivity and selectivity towards gas analysts. Thin films of solution of ZnO decorated graphene oxide sheets of ZnO–GrO-2 were prepared by spin coating to ensure the uniformity of films. Hence on they will addressed as ZnO–GrO.

2.4.

Gas sensing

For all electrical measurements films deposited on ITO glass substrate were used with four-probe dc measurement technique to eliminate role of contact resistance. The measurements were performed in linear sweep mode from 10 to +10 V. Keithley 4200 (semiconductor characterization system) was used for all I–V measurements and data acquisition. Dry nitrogen was used for flushing. Gas sensing of the ZnO–GrO films was studied using an in house gas sensor assembly with Mass flow controllers to control the gas flow. The film was placed in gas chamber and vacuum was activated. After vacuum was maintained for sufficient time, I–V measurements were taken using the four point probe measurement.

2.5.

Characterization

X-ray diffraction (XRD) of the film deposited on glass slides was performed using a Rigaku X-ray diffractometer. The sample was prepared by depositing films onto a glass slide by repeated drop drying. Fourier transform infrared spectroscopy (FTIR) spectrum (400–4000 cm 1) was measured using a Nicolet 5700 FTIR spectrometer with pure KBr as the background. GO and GrO powder was filtered, repeatedly washed, and dried before mixing with KBr. The mixture was dried and compressed into a tablet for measurement. The UV–vis spectrum (190–1100 nm) was measured using a UV1800 Shimadzu UV spectrophotometer. Ethanol was used for background signal.

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

CARBON

xxx (2011) xxx–xxx

The photoluminescence (PL) spectra were obtained at an excitation wavelength of 340 nm on Perkin Elmer LS 55 setup with a 390 nm filter on emission side. High resolution transmission electron microscopy (HRTEM) studies of the GO, GrO, and ZnO–GrO nanosheets were performed using the Tecnai G2F30 S-Twin (FEI; Super Twin lens with C s 1.2 mm) instrument operating at an accelerating voltage at 300 kV, having a point resolution of 0.2 nm, and with a lattice resolution of 0.14 nm with attached energy-dispersive X-ray spectrometer (EDS/EDAX). Image processing was performed using Digital Micrograph software (Gatan). The HRTEM samples were prepared by drying a droplet of the GO and ZnO– GrO suspension on a carbon coated copper grid. X-ray photoelectron spectroscopy (XPS) studies have were carried out on pure GO and ZnO–GrO using a Perkin Elmer 1257 model. Experiments were performed at average base pressure of 4.8 · 10 8 torr at 300 K with a non-monochromatic AlKa line at 1486.6 eV, and a hemispherical sector analyzer capable of 25-meV resolution. The overall instrumental resolution was about 0.3 eV. Pass energy for general scan and core level spectra kept at 100 and 60 eV respectively. The work function of the analyzer was calibrated with Au.

3.

Results and discussion

GO was synthesized as per the procedure explained under experimental section. GO thus synthesized was dried and redispersed in ethanol and centrifuged at speeds varying from 6000 to 7000 rpm to remove large residual particles and obtain a uniformly dispersed solution which was stable for more than 6 weeks. The redispersed GO in ethanol was mixed with stoichiometric amount of zinc acetate dihydrate (Zn(Ac)2Æ2H2O). To this lithium hydroxide monohydrate (LiOHÆH2O) solution in ethanol was added and was left undisturbed for 24 h to ensure complete reduction of GO and ZnO precursor, leading to the formation of required ZnO decorated GrO sheets, hereby referred as ZnO–GrO.

3.1.

Structural and morphological studies

XRD peak profiles corresponding to bare ZnO (curve 1, red coloured) and ZnO–GrO (curve 2, black coloured) samples are shown in Fig. 1(a). All the major peaks could be indexed to crystalline hexagonal wurtzite structure of ZnO. ZnO grains show a preferred orientation towards (0 0 1) crystallographic plane represented by peak marked ‘d’ in Fig. 1a. Some minor peaks corresponding to minor phases Zn(OH)2 could also be identified. XRD profile of ZnO–GrO in curve 2 show an additional peak at 26.46 (marked as ‘j’) similar to that of carbon black [24] and one at 23.97, corresponding to the interlayer distance (d-spacing) of 0.38 nm, which is slightly more than the d-spacing of perfect graphene. The peak broadening in ZnO–GrO indicates small but finite degradation in crystalline structure of ZnO. The small bump near 22 and also 10 (inset of Fig. 1(a)) indicates that the GO is not completely reduced and hence the name graphene oxide. We call it reduced graphene oxide and not graphene as the chemical reduction alone always results in the presence of few defects and residual oxygen moieties.

3

FTIR spectra further confirmed the reduction of GO to GrO by LiOHÆH2O. Fig. 1(b) shows the curves 1 (black) and 2 (red) corresponding to GO and GrO respectively. GO mainly has carboxyl and hydroxyl groups. In curve 1, peak around 3458 cm 1 can be assigned to O–H stretching vibrations in hydroxyl groups and peak at 2934 cm 1 represents asymmetric stretching vibrations in the CH2. The C@O stretching vibrations in the carboxyl group are also visible around 1731 cm 1 and at around 1647 cm 1. The peak at 1381 cm 1 due to the hydroxyl group attached to the aromatic carbon backbone represents the O–H deformation. Peak at 1121 cm 1 can be assigned to C–O stretching vibrations. On comparing with the FTIR spectra of GrO (curve 2), we can observe the presence of O–H stretching vibrations around 3432 cm 1, asymmetric stretching and symmetric CH2 vibrations peaks at (2930 and 2859 cm 1), some minor (low intensity) C@O stretching vibrations in the carboxyl group are also visible around 1641 cm 1. A high intensity major peak at around 1490 cm 1 represents C-partial double bond-C ring stretch vibrations. All these indicate the formation of GrO. The UV–vis absorption spectra of ZnO and ZnO–GrO, with different GO concentrations are shown in Fig. 1(c). It is evident that graphene absorbs strongly in the UV region and may interfere with the excitation at 340 nm of ZnO nanoparticles in emission studies [9]. The peak shown as inset at 260 nm may be due to p–p* transitions of C@C [8]. XPS survey scan spectra are shown in Fig. 2(a) for GO (curve 1) and ZnO–GrO (curves 2 and 3) for the scans acquired before and after Ar ion sputtering. The presence of ZnO is clearly manifest from its various orbital levels. The Auger peak for O and Zn are also observed which are marked with dashed arrows. During photoemission studies, small specimen charging was observed which was later calibrated by assigning the C 1s signal at 285 eV. XPS high resolution C 1s core level spectra acquired in the range of 275–295 eV are shown in Fig. 2(a and b). Systematic XPS investigations were carried out on pristine GO and ZnO decorated GrO sheets. The presence of ZnO significantly affects the carbon functional groups in the system. To identify the changes in this functional group, core level spectra of C 1s were analyzed using peak fit programme and peak deconvolution was performed on all the three samples. However, deconvoluting the peaks into Gaussian components using appropriate positions and FWHM, reveal four functional groups: (i) the non-oxygenated ring C (C–C, 284.71 eV), (ii) the C in C–O bonds, (286.39 eV), (iii) the carbonyl C (C@O, 287.79 eV), and (iv) carboxylate carbon (–O–C@O, 289.26 eV). Furthermore the peaks observed at 280.7 eV (C–O–Zn), could be assigned to the formation of ZnO–GrO sheets. To resolve 2D structure of ZnO decorated Graphene sheet, Ar ion sputtering was performed using energy of 4 keV for 5 min at 1 · 10 6 torr, with 20 mA emission current. After sputtering survey scan spectra shows a sharp increase in the Zn (2p) level. The percentage areas for different functional groups obtained from XPS deconvolution analysis are listed in Table 1. It is evident from Fig. 2(a and b) and also the results summarized as Table 1, that significant reduction was observed for C in C–O, carbonyl carbon and carboxylate carbon while non-oxygenated carbon increased from 4.03% to 55.9% which

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

4

CARBON

x x x ( 2 0 1 1 ) x x x –x x x

Fig. 1 – XRD and spectroscopic studies (a) X-ray diffraction plots for (curve 1) pure ZnO and (curve 2) ZnO–GrO composites. The peaks marked as ‘d’ and ‘m’ corresponding to 001 and 002 crystallographic planes of hexagonal wurzite structure of ZnO respectively. (b) FTIR spectra of (curve 1) GO and (curve2) GrO (c) UV spectra of (a) pristine ZnO (b) ZnO–GrO-1 (c) ZnO–GrO-2 (d) ZnO–GrO-3 (e) ZnO–GrO-4 at room temperature. The inset at bottom shows the magnified spectra for ZnO–GrO-2.

Fig. 2 – X-ray photoelectron spectroscopy XPS survey scan spectra acquired at hm = 1486.6 eV for (a) GO (curve 1) and ZnO–GrO (curve 2). Various Zn and O (1s) state are marked with arrows and the dotted arrows showing their Auger transition. XPS peak fit deconvolution of C (1s) spectra of (b) GO (c) ZnO–GrO before sputtering and (d) ZnO–GrO after sputtering, depicting the change in the C functional groups.

shows the formation of GrO. The effect of argon ion sputtering causes further reduction in C–C species and increase in C–O–Zn which provide convincing evidence for the decorated layered structure of ZnO between the graphene oxide sheets. In other words, the sputtering of layered structure of ZnO decorated graphene by Ar ions could have led to increased exposure of ZnO nanaoparticles, which is evident from the enhancement of C–O–Zn species.

Table 1 – Percentage areas for different functional groups obtained from XPS deconvolution analysis.

C–C C–O C@O O–C@O C–O–Zn

GO

ZnO–GrO before sputtering

ZnO–GrO after sputtering

4.03% 57.9% 25.8% 12.3%

55.9%

39.6%

16.5%

24.9%

27.6%

35.4%

The heterostructure of these composites can be verified by the morphological analyses presented by the HRTEM images. Fig. 3a shows the almost transparent crumpled carbon sheets decorated with ZnO nanoparticles. The wavy highly crumpled structures of the carbon sheets indicate the presence of monolayer GrO sheets. Fig. 3b shows the clear micrograph of the high density of ZnO nanoparticles on the monolayer sheet. It should be noted that no nanoparticles were found scattered outside GrO sheets, even after sonication, which indicates that ZnO nanoparticles are strongly bound on graphene surface. The graphene oxide obtained by chemical reduction of oxy-functional groups would have many carbon vacancies and defects, which may enhance the interaction between ZnO nanoparticles and graphene [25]. Closer examination, Fig. 3c of the sheets elucidates that most of the ZnO particles are deposited in the contorted regions of the crumpled sheet. The average size of ZnO nanoparticles is 3–5 nm and they are fairly monodisperse. It is to be noted that the ZnO nanoparticles are synthesized without

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

CARBON

xxx (2011) xxx–xxx

5

Fig. 3 – High resolution transmission electron micrographs. (a) TEM images of the crumpled ZnO–GrO sheet. (b) The GrO sheets with the decorated semiconducting nanoparticles. (c) The bright HRTEM images of ZnO–GrO sheet showing the presence of nanoparticles in the contours of the crumpled sheets. (d) HRTEM image ZnO nanoparticles. (e) EDAX of the composite showing presence of Zn, O and carbon. (f and g) The reconstructed lattice of the marked areas in the HRTEM micrograph shown in (d), clearly showing the lattice for graphene (0.34 nm) and ZnO nanoparticles (0.28 nm) 100 plane. Inset of (g) shows front view of the wurtzite crystal structure for ZnO. One of the hexagons is marked with black outline to emphasize the atomic arrangement.

any size controlling agent by simple chemical reduction and the presence of fairly uniform sized nanoparticles could be due to the stabilizing effect provided by the graphene oxide. HRTEM image Fig. 3d clearly shows that ZnO nanoparticles are in crystalline state with lattice space of 0.28 nm corresponding to (1 0 0) plane. The EDAX spectra, Fig. 3e from the oxide nanoparticle decorated sheets clearly shows the presence of Zn. Fig. 3f and g shows the reconstructed images of the marked regions of Fig. 3d reconstructed using Inverse Fast Fourier transformation (IFFT). The images Fig. 3f and g evidently show lattice spacing of 0.34 nm corresponding to graphene and the crystal structure of wurtzite ZnO showing lattice spacing of 0.28 nm corresponding to (1 0 0) plane respectively. The reconstructed images noticeably show the hexagonal structure (outlined). The corresponding front view of the ZnO wurtzite structure is shown as inset with yellow showing oxygen’s atoms and brown showing zinc atoms. The highly dispersed metal nanoparticles on supports with larger surface areas have properties with advantages for catalytic activity and sensor sensitivity [26]. Furthermore, these monolayer sheets possess large surface areas, and particles can deposit on both sides of these sheets [27,28]. Thus, such integration of two-dimensional supports with large surface areas and the highly dispersed nanoparticles can be an exciting material for use in future nanotechnology.

3.2.

Photoluminescence and time-resolved PL

The PL spectrum of GO (curve 1) and GrO (curve 2) obtained at an excitation wavelength of 340 nm are shown in Fig. 4(a). The PL of as prepared GO shows a emission peak in the visible spectrum at 443 nm similar to that reported by Eda et al. [17] for liquid suspension of GO. Reduction of GO blue-shifts the PL spectra by 0.15 eV with resulting emission at 424 nm. The PL in such carbon systems is a consequence of geminate recombination of localized e–h pairs in sp2 clusters which essentially behave as the luminescent centres. Since the band gap depends on the size, shape and the fraction of sp2 domains, tunable PL could be achieved by controlling the sp2 sites. In other words, the PL intensity increases with an increase in sp2 content in the disordered carbon systems [29]. This undoubtedly shows the role played by lithium base in reduction of GO. Luminescence of GO have also been reported in red and near-infrared region [30,31] which can result from presence of multilayered and aggregated flakes [17]. Centrifugation is therefore done at early stage as illustrated in experimental section to remove these particles. Blue photoluminescence is also considered to be an indicative of the quality of graphene suspensions. The PL spectrum of bare ZnO (curve a) of Fig. 4(b) clearly shows a broad green emission peak at 550 nm, which is

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

6

CARBON

x x x ( 2 0 1 1 ) x x x –x x x

Fig. 4 – Photoluminescence studies (a) PL spectra GO (curve 1) and GrO (curve 2). (b) PL spectra of ZnO and of ZnO–GrO sheets at different GO concentrations (a) ZnO (b) ZnO–GrO-1 (c) ZnO–GrO-2 (d) ZnO–GrO-3 (e) ZnO–GrO-4. All the spectra were taken at room temperature at excitation wavelength of 340 nm.

generally believed to arise from transition in defect states, particularly oxygen vacancies [32]. Interestingly PL from ZnO–GrO (curves b–e) was not only blue shifted as compared to that of bare ZnO but also is significantly quenched. This suggests that an additional pathways for the diminishing the charge carriers dominates because of the interactions between the excited ZnO and the GrO sheets. In other words the quenching behavior could be explained by interfacial charge transfer from ZnO to graphene [9]. Fig. 4(b) also shows the effect of increasing concentration of GO to the initial solution. As the concentration of GO is increased, the extent of PL quenching increases; an observation similar to that was reported by Williams and Kamat [9]. However, they didn’t observe any shift in PL measurements since their preparation involves anchoring of as-prepared ZnO particles over GO. In their case, the electrons from the n-type ZnO were primarily used up in reduction of GO to GrO upon irradiation of UV light. However, we propose another mechanism for the observed large blue-shift of 0.15 eV in PL. In our case due to simultaneous in-pot reduction of both ZnO and GO, the path of photo-catalytic reduction of GO by e -transfer from ZnO is eliminated. The interaction between GrO and ZnO, in single step reduction process, could be modeled on charge transfer dynamics as in SnO2–CNT or ZnO–SWCNT composites [15]. Since it is well known that ZnO intrinsically behaves as n-type semiconductor due to oxygen vacancies and zinc interstitials [33] and graphene behaves like a p-type semiconductor [20,34] in ambient atmosphere, hence n–p heterostructure or a depletion layer is formed at the GrO–ZnO interface. A blue shift of 0.15 eV can be envisaged as a significant step towards band gap tuning of graphene based heterostructures for future application involving opto-electronics and imaging. If the interaction between the excited ZnO and GrO is indeed route cause for quenching of PL emission, then it should be possible to time resolve the process from the PL emission decay. Hence, time-resolved PL (TRPL) was performed on the ethanolic solutions of intrinsic GO and ZnO–GrO composites to obtain their exciton decay lifetimes. Fig. 5 shows that the

exciton lifetime of ZnO–GrO dramatically decreases by many times as compared to intrinsic GO. The excitation and emission wavelengths of 340 and 424 nm were used for the experiment to record the lifetimes. The obtained decay patterns are multiexponential in nature and can be fit to a biexponential expression in order to derive decay time constants, which could be further used to calculate the average lifetimes of the photoluminescence. The inset shows the decrease in average lifetime as a function of various GO/ZnO concentrations. As prepared GO is insulating in nature because sp2 clusters are being interrupted by oxygen functional groups, which in turn decrease the creation of e–h pairs upon excitation. In the absence of ZnO, the PL emission decay has a average lifetime of 29.35 ns whereas on the formation of ZnO nanoparticles decorated GrO sheets, the average lifetime decreases drastically to 12.32 ns. The analysis of the average lifetimes conclude that the highly influential static interaction prevails between ZnO and GrO sheets, leading to increased number of localized states allowing radiative recombination to occur. Further, the exciton lifetime was found to be weakly dependent on the duration of stirring or the room temperature. In any case, the reduction process is also expected to produce defects such as vacancies in the ZnO–GrO composites. Increased number of defects and larger delocalization lengths in the reduced GrO result in higher probability of nonradiative recombination and, therefore faster decay of PL. The non-radiative recombination can also occur by tunneling or hopping mechanisms of e–h pairs into defect sites due to probable presence of the deep traps.

3.3.

Electrical properties

Fig. 6(a) shows the I–V conductivity plots of GO and GrO at room temperature. As expected the conductivity increases markedly upon reduction due to partial restoration of graphitic structure. However, it is to be noted that since the complete planar structure cannot be restored by chemical

Fig. 5 – Time-resolved photoluminescence decay curves of GO and ZnO–GrO composites. The inset shows the dependence of lifetime on GO concentration. The excitation and emission wavelengths of 340 and 424 nm were used for the experiment to record the lifetimes.

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

CARBON

xxx (2011) xxx–xxx

reduction, hence conductivity values are much lower than reported for single or bi-layer graphene sheets [35]. Also to be noticed is that ZnO–GrO have a very sharp increase in the conductivity value, almost >200· time than that of simple reduced graphene oxide as shown in Fig. 6(a), curves b and c. The increase might be due to the role played by the presence of ZnO nanoparticles in hopping mechanism of charge transfer in the defect riddled chemically reduced graphene oxide, by compensating for the degradation of electrical network and providing an alternative path for charge transfer. Fig. 6(a) also shows change in conductivity with increase in concentration of graphene oxide in ZnO–GrO composite. As expected we get higher values of electrical conductivity with increase in concentration of GO in starting solution. The composite system studied in this report exhibits not only monolayers but also bilayers of GrO, as deduced by 4 nm sheet heights (AFM not shown). The presence of bilayer and few-layer graphene is not necessarily indicative of a bad sample. There are studies to prove that bilayer graphene decorated with nanoparticles can exhibit conductivities significantly greater than those for single-layer graphene. Fig. 6(b) shows resistance (log) measurements with variation in temperature. We can observe that conductivity increases with rise in temperature implementing the semiconducting nature of resulting ZnO–GrO composites [36]. We also observe that up to 300 K ( 37 C) there is not much change in resistance of ZnO–GrO composites which can be correlated to previous results on temperature dependent conductivity measurements of graphene [37]. However, dramatic increase in conductivity beyond 300 K can be fitted reasonably well with the Arrhenius model, suggesting that thermally excited carriers begin to dominate electrical conduction at relatively higher temperatures [38].

3.4.

Gas sensing

3.4.1.

The proposed sensing mechanism

Carbon monoxide (CO) (22 ppm) gas in dry nitrogen was introduced at room temperature through the inlet valve of gas chamber. Fig. 7(a) shows the change in sensor (here ZnO– GrO) current with respect to time. It is clear from the graph that the response is fast resulting in the increase in the value of current by several magnitudes, from pico to nano ampere range. This may be on account of very low percolation threshold [39] of 0.1 vol.%, for room temperature conductivity of

7

graphene, the lowest reported value for any carbon-based composite except for those involving carbon nanotubes. At only 1 vol.%, the ZnO decorated graphene oxide nanosheets (ZnO–GrO) has a conductivity of, 0.1 S m, sufficient for many electrical applications [40]. Also the ZnO–GrO film demonstrates a cyclic response and good repeatability; a criteria essential for practical sensor. We propose that the following mechanism based on the different regions of the response curve. The region marked as (a) for curve 1, Fig. 7(b) shows that the initial volume of CO gas is consumed in establishing the percolation. It takes around 4–5 min for complete percolation during which current increases by 5000 times from pico levels to micro regime. After percolation is complete, the gas molecules start adsorbing on ZnO nanoparticles as shown by a slow increase in sensing current depicted by region (b) in Fig. 7(b). The full scale response assuming initial percolation conductivity as baseline is 24% and it takes 5 min to reach the saturation value. On purging (using dry nitrogen at mild pressure of 2 atm for 30 s), the sensor recovers 75% of its initial response rapidly within 2 min depicted by region (c) in Fig. 7(b) thus showing excellent recovery. Such rapid recovery may be due to breaking of percolation network as gas molecules are swept by purging gas. However, complete recovery further takes two more minutes as shown by region (d) for de-adsorption of gas molecules from ZnO nanoparticles surface. The figure also shows the sensor response to CO (1 ppm) gas, curve 2, Fig. 7(b) measured as described above. The total sensing duration is still around 10 min, however, response magnitude is significantly smaller and recovery time is also reduced. In addition for the purpose of comparison, a sensing experiment for CO (22 ppm) was conducted on bare ZnO nanoparticles and the response is shown in curve 3, Fig. 7(b). Bare ZnO nanoparticles shows no or negligible sensing at room temperature to CO (22 ppm). Similar experiments were conducted for NH3 sensing also and it was found that ZnO shows negligible response. Fig. 7(c) shows ZnO–GrO response to (1 ppm) of ammonia (NH3). The data is plotted as normalized current I/Imax, where, Imax is full scale response. The nature of response is similar to CO however there is considerable increase in percolation time to 8 min and then it takes 7 min for a 24.3% increase in sensing current after reaching percolation threshold denoting adsorption of gas molecules by zinc oxide nanoparticles. The total sensing time is approximately 15 min however, the response magnitude is quite large as compared

Fig. 6 – I–V studies (a) I–V characteristics of (curve a) GO (curve b) GrO (curve c) ZnO–GrO-1 (curve d) ZnO–GrO-2 (curve e) ZnO– GrO-3. (b) Temperature dependent I–V characteristics of ZnO–GrO-2 palette. Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

8

CARBON

x x x ( 2 0 1 1 ) x x x –x x x

Fig. 7 – Gas sensing (a) response of ZnO–GrO nanosheets to 22 ppm CO gas. (b) (curve I) Response of ZnO–GrO nanosheets to 22 ppm CO at room temp. (curve 2) Response curve of ZnO–GrO nanosheets to 1 ppm CO at room temp. (curve 3) Sensing response of bare ZnO film to 22 ppm CO gas. (c) Response curve of ZnO–GrO nanosheets 1 ppm NH3 at room temperature. (d) Response curve of ZnO–GrO nanosheets to 5 ppm NO at room temperature.

to CO sensing, even for 1 ppm, denoting strong affinity of ammonia to graphene as also reported earlier [4]. Fig. 7(d) shows response of ZnO–GrO composite sensor to 5 ppm nitric oxide (NO) gas in dry nitrogen at room temperature. Region (I) denotes establishing of percolation network taking average time of 6 min as in earlier cases. Region (II) denotes decrease in sensor current as it is an oxidizing gas. Sensor current decreases by a mere 3.5% over a period of 25 min before attaining saturation thus denoting poor sensor response to NO. Further the sensor response could not be recovered by nitrogen purging suggesting permanent chemisorption. The selectivity for the reducing gases can we further explained as the sensing mechanism of ZnO based chemical sensors is well established and is based on the formation of Schottky barrier due to chemisorbed oxygen on metal-oxide surface [13]. Reducing gases like CO and NH3 leads to decrease in partial pressure of oxygen which results in release of surface trapped electrons back to conductance band of ZnO hence the observed decrease in sensor resistance or increased current. In the same manner oxidizing gases (NO) increase the resistance by increasing the coverage of ion-sorbed oxygen [41]. At a low operation temperature, the low response can be expected as the gas molecules do not have enough thermal energy to react with the surface adsorbed oxygen species hence the low gas sensitivity shown in Fig. 6(b) by ZnO film. In ZnO–GrO films, two different depletion layers

co-exist. One is on the surface of ZnO particles and the other is in the interface between graphene oxide and ZnO. The adsorbed gas molecules will modulate both depletion layers; the one on the ZnO surface and the other at ZnO–GrO composite interface. As the electrical response indicates ZnO–GrO acts as a system which provides the net gas sensing mechanism and not the individual components hence a room temperature response by the sensor is possible. Moreover, it is expected that the larger specific surface area of ZnO nanoparticles on graphene oxide islands would further enhance the performance of ZnO as a gas sensor. Also the ZnO–GrO composite sensor shows poor response to oxidizing gases (here NO) and thus is selective. One of the substantial advantage of our reduced graphene oxide–zinc oxide composite sensor is the large electrical gain provided by graphene due to its near metallic conductivity resulting in an inherent-amplified sensing configuration as well as eliminating the need of thermal activation of ZnO nanoparticles thus enabling room temperature sensing. Table 2 presents a comparative analysis of gas sensing performance for graphene and ZnO based sensors. It is evident that ZnO based sensor has high sensitivity and moderate selectivity which can be further improved by doping and other methods but they require high working temperature. On the other hand, most of Graphene and/or CNT based, gas sensors work at room temperature but have poor selectivity (as evident from reduced graphene oxide sensing data in

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

CARBON

xxx (2011) xxx–xxx

2–3 2–5

Current work [43]

[42]

d

c

b

[20] [43] [20] [41] [44]

Gas sensitivity is expressed as % change in resistance of gas sensor. Gas sensitivity is expressed as % change in conductance of gas sensor. Excludes time required for establishment of percolation network. Undoped ZnO normally is not very sensitive to CO so typically ZnO doped with SnO2, Ni, Cu, etc. is used.

>5

Sensor response Response time (min) Recovery time (min) Reference

Summary

Based on the preliminary studies we strongly believe that very precise gas sensing to few ppb levels is quite possible taking into account the large sensing response shown for 1 ppm concentrations of gas analytes. We also investigated and discussed the optical and electrical properties of our material. Blue PL of ZnO–GrO Composites makes them attractive material for low cost blue optoelectronics. However, ZnO–GrO sheets synthesized by our single step simultaneous reduction offer tremendous potential for stable, high precision gas sensing applications. We need to further emphasize that this is an excellent process to stabilize the semiconductor nanoparticles without functionality on surface during formation. A lot of work needs to be undertaken in improvising the selectivity of the composites on the lines discussed above by either tailoring the graphene sheet edges with gas-specific binding molecules or using different ZnO nanostructure such as nano-rods, nanoplates, etc. and this can be taken as future studies.

a

30 10 –

– –

1% NH3 in air +11.1% DG 40 NH3 5 ppm in dry N2 +10% DR >10 NH3 – Sensor characteristics Sensing gas Analyte conc.

350 C Working temperature

NO2 1 ppm in air +80% DRa 3–6

CO 50 ppm in air Not sensitived

Room temperature

b

10

2 ppm in air +12% DG 40 NO2 5 ppm in dry N2  30% DR >10

4.

30

CO 22 ppm in dry N2 +24.3% DG 5c

NH3 1 ppm in dry N2 +24% DG 6c

NO 5 ppm in dry N2  3.5% DG 25

Table 1). Our ZnO–GrO composite sensor tries to bridge this gap by providing sufficient sensitivity with very good selectivity to reducing gases at room temperature. Also, it can be noted from Table 1, that, ZnO–GrO composite sensor shows enhanced response and quick recovery times as compared to room temperature graphene based sensors.

H2 4% vol in air +16% DR 9

Room temperature Room temperature

ZnO–GrO nanocomposite Pt-Graphene Reduced graphene oxide Zinc oxide Material

Table 2 – Comparison of gas sensing characteristics of different nano-structured materials.

9

Acknowledgments Gaurav Singh and Anshul Choudhary thank Prof. L.S. Tanwar and Dr. Rajesh Purohit, NSIT for their encouragement and support. The authors’ gratefully acknowledge Prof. R.C. Budhani, Director NPL for his kind support and valuable suggestions. We would also like to acknowledge Dr. Rajeev Singh and Mr. Amit Kumar for assistance with I–V measurements and Ms. Shweta Gupta for support during experiments.

R E F E R E N C E S

[1] Baughman RH, Zakhidov AA, de Heer WA. Carbon nanotubes – the route toward applications. Science 2002;297:787–92. [2] Novoselov KS, Geim AK, Morozov SV, Jiang D, Zhang Y, Dubonos SV, et al. Electric field effect in atomically thin carbon films. Science 2004;306:666–9. [3] Geim AK. Graphene: status and prospects. Science 2009;324:1530–4. [4] Stankovich S, Dikin DA, Dommett GHB, Kohlhaas KM, Zimney EJ, Stach EA, et al. Graphene-based composite materials. Nature 2006;442:282–6. [5] Geim AK, Novoselov KS. The rise of graphene. Nat Mater 2007;6:183–91. [6] Schedin F, Geim AK, Morozov SV, Hill EW, Blake P, Katsnelson MI, et al. Detection of individual gas molecules adsorbed on graphene. Nat Mater 2007;6:652–5. [7] Xu C, Wang X, Zhu J. Graphene–metal particle nanocomposites. J Phys Chem C 2008;112:19841–5. [8] Eda G, Fanchini G, Chhowalla M. Large-area ultrathin films of reduced graphene oxide as a transparent and flexible electronic material. Nat Nanotechnol 2008;3:270–4.

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050

10

CARBON

x x x ( 2 0 1 1 ) x x x –x x x

[9] Williams G, Kamat PV. Graphene-semiconductor nanocomposites: excited-state interactions between ZnO nanoparticles and graphene oxide. Langmuir 2009;25:13869–73. [10] Sun X, Liu Z, Welsher K, Robinson J, Goodwin A, Zaric S, et al. Nano-graphene oxide for cellular imaging and drug delivery. Nano Res 2008;1:203–12. [11] Si Y, Samulski ET. Exfoliated graphene separated by platinum nanoparticles. Chem Mater 2008;20:6792–7. [12] Cao A et al. A facile one-step method to produce graphene– CdS quantum dot nanocomposites as promising optoelectronic materials. Adv Mater 2010;22:103–6. [13] Franke ME, Koplin TJ, Simon U. Metal and metal oxide nanoparticles in chemiresistors: does the nanoscale matter? Small 2006;2:36–50. [14] Comini E. Metal oxide nano-crystals for gas sensing. Anal Chim Acta 2006;568:28–40. [15] Wei B, Hsu M, Su P, Lin H, Wu R, Lai H. A novel SnO2 gas sensor doped with carbon nanotubes operating at room temperature. Sensor Actuat B Chem 2004;101:81–9. [16] Sheats JR, Antoniadis H, Hueschen M, Leonard W, Miller J, Moon R, et al. Organic electroluminescent devices. Science 1996;273:884–8. [17] Eda G, Lin Y, Mattevi C, Yamaguchi H, Chen H, Chen I, et al. Blue photoluminescence from chemically derived graphene oxide. Adv Mater 2010;22:505–9. [18] Kaniyoor A, Imran Jafri R, Arockiadoss T, Ramaprabhu S. Nanostructured Pt decorated graphene and multi walled carbon nanotube based room temperature hydrogen gas sensor. Nanoscale 2009;1:382–6. [19] Robinson JT, Perkins FK, Snow ES, Wei Z, Sheehan PE. Reduced graphene oxide molecular sensors. Nano Lett 2008;8:3137–40. [20] Fowler JD, Allen MJ, Tung VC, Yang Y, Kaner RB, Weiller BH. Practical chemical sensors from chemically derived graphene. ACS Nano 2009;3:301–6. [21] Hummers WS, Offeman RE. Preparation of graphitic oxide. J Am Chem Soc 1958;80:1339. [22] Spanhel L, Anderson MA. Semiconductor clusters in the sol– gel process: quantized aggregation, gelation, and crystal growth in concentrated zinc oxide colloids. J Am Chem Soc 1991;113:2826–33. [23] Kamat PV, Patrick B. Photochemistry and photophysics of ZnO colloids. In: Symp. Electron. Ionic Prop. Silver Halides, St. Paul, MN. Springfield, Va: The Society for Imaging Science and Technology; 1991. p. 293–9 [May 12–17]. [24] Cuong TV, Pham VH, Tran QT, Hahn SH, Chung JS, Shin EW, et al. Photoluminescence and Raman studies of graphene thin films prepared by reduction of graphene oxide. Mater Lett 2010;64:399–401. [25] Okamoto Y. Density-functional calculations of icosahedral M13 (M = Pt and Au) clusters on graphene sheets and flakes. Chem Phys Lett 2006;420:382–6. [26] Xing Y. Synthesis and electrochemical characterization of uniformly-dispersed high loading Pt nanoparticles on sonochemically-treated carbon nanotubes. J Phys Chem B 2004;108:19255–9.

[27] Li D, Kaner RB. Materials science – Graphene-based materials. Science 2008;320:1170–1. [28] Novoselov KS, Geim AK, Morozov SV, Jiang D, Katsnelson MI, Grigorieva IV, et al. Two-dimensional gas of massless Dirac fermions in graphene. Nature 2005;438:197–200. [29] Robertson J. Recombination and photoluminescence mechanism in hydrogenated amorphous carbon. Phys Rev B 1996;53:16302–5. [30] Liu Z, Robinson JT, Sun X, Dai H. PEGylated nanographene oxide for delivery of water-insoluble cancer drugs. J Am Chem Soc 2008;130:10876–7. [31] Luo Z, Vora PM, Mele EJ, Johnson ATC, Kikkawa JM. Photoluminescence and band gap modulation in graphene oxide. Appl Phys Lett 2009;94:111909-1–3. [32] Hu JQ, Li Q, Meng XM, Lee CS, Lee ST. Thermal reduction route to the fabrication of coaxial Zn/ZnO nanocables and ZnO nanotubes. Chem Mater 2003;15:305–8. [33] Look DC, Hemsky JW, Sizelove JR. Residual native shallow donor in ZnO. Phys Rev Lett 1999;82:2552–5. [34] Qazi M, Vogt T, Koley G. Trace gas detection using nanostructured graphite layers. Appl Phys Lett 2007;91: 233101-1–3. [35] Mattevi C et al. Evolution of electrical, chemical, and structural properties of transparent and conducting chemically derived graphene thin films. Adv Funct Mater 2009;19:2577–83. [36] Gilje S, Han S, Wang M, Wang KL, Kaner RB. A chemical route to graphene for device applications. Nano Lett 2007;7:3394–8. [37] Li X, Zhang G, Bai X, Sun X, Wang X, Wang E, et al. Highly conducting graphene sheets and Langmuir–Blodgett films. Nat Nanotechnol 2008;3:538–42. [38] Eda G, Mattevi C, Yamaguchi H, Kim H, Chhowalla M. Insulator to semimetal transition in graphene oxide. J Phys Chem C 2009;113:15768–71. [39] Su F, Zeng J, Bao X, Yu Y, Lee JY, Zhao XS. Preparation and characterization of highly ordered graphitic mesoporous carbon as a Pt catalyst support for direct methanol fuel cells. Chem Mater 2005;17:3960–7. [40] Stauffer D, Aharony A. Introduction to percolation theory. 2nd ed. Taylor & Francis; 1994. [41] Lupan O, Shishiyanu S, Chow L, Shishiyanu T. Nanostructured zinc oxide gas sensors by successive ionic layer adsorption and reaction method and rapid photothermal processing. Thin Solid Films 2008;516:3338–45. [42] Kaniyoor A, Imran Jafri R, Arockiadoss T, Ramaprabhu S. Nanostructured Pt decorated graphene and multi walled carbon nanotube based room temperature hydrogen gas sensor. Nanoscale 2009;1(3):382–6. [43] Lu G, Ocola LE, Chen J. Reduced graphene oxide for room-temperature gas sensors. Nanotechnology 2009;20(44): 445502-1–9. [44] Cho P-S, Kim K-W, Lee J-H. NO2 sensing characteristics of ZnO nanorods prepared by hydrothermal method. J Electroceram 2006;17(2–4):975–8.

Please cite this article in press as: Singh G et al. ZnO decorated luminescent graphene as a potential gas sensor at room temperature. Carbon (2011), doi:10.1016/j.carbon.2011.08.050