Synthesis of nanocrystalline ZnO nanobelts via pyrolytic decomposition of zinc acetate nanobelts and their gas sensing behavior

Surface Science 606 (2012) 715–721

Contents lists available at SciVerse ScienceDirect

Surface Science journal homepage: www.elsevier.com/locate/susc

Synthesis of nanocrystalline ZnO nanobelts via pyrolytic decomposition of zinc acetate nanobelts and their gas sensing behavior A. Tarat a,⁎, R. Majithia b, R.A. Brown a, M.W. Penny a, K.E. Meissner c, T.G.G.Maffeis a a b c

Multidisciplinary Nanotechnology Centre, Swansea University, SA2 8PP, United Kingdom Material Science and Engineering Program, Texas A&M University, USA Department of Biomedical Engineering, Texas A&M University, 337 Zachry Engineering Center, TX 77843-312, USA

a r t i c l e

i n f o

Article history: Received 24 October 2011 Accepted 16 December 2011 Available online 30 December 2011 Keywords: Zinc oxide Zinc acetate Nanobelts Gas sensing CO AFM XPS XRD

a b s t r a c t The pyrolytic decomposition of layered basic zinc acetate (LBZA) nanobelts (NBs) into nanocrystalline ZnO NBs is investigated using scanning electron microscopy (SEM), atomic force microscopy (AFM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and photoluminescence (PL). We also report on the gas sensing response of the resulting ZnO nanomaterial to CO. The LBZA NBs are grown at 65 °C in an aqueous solution of zinc acetate dihydrate. AFM and SEM results show as-grown products possess the characteristic layered structure of the LBZA crystals. XRD and XPS results show that annealing as-grown products at 210 °C in air causes a transformation from zinc acetate to nanocrystalline ZnO NBs via thermal decomposition. The ZnO crystalline domain size increases with temperature from 9.2 nm at 200 °C to 94 nm at 1000 °C, as measured from XRD. SEM shows evidence of sintering at 600 °C. The thickness of the NBs, determined via AFM, ranges from 10 to 50 nm and remains approximately constant with annealing temperature. XPS confirmed the chemical transformation from zinc acetate to ZnO and showed a significant remaining zinc hydroxide component for the ZnO NBs consistent with published results. PL measurements at room temperature show a blue shift in peak emission as the nanobelts change from LBZA to ZnO at 200 °C. Above this transition temperature, the ZnO nanobelts possess strong band edge emission at 390 nm and little broad band emission in the visible region. The AFM and SEM images reveal that the crystallites within the nanobelts orientate in rows along the long axis during annealing. This structure provides a high surface area to volume ratio of aligned nanoparticles which is beneficial for gas sensing applications. Gas sensors fabricated from 400 °C annealed nanobelts showed a response of 1.62 when exposed to 200 ppm of CO in dry air at 400 °C, as defined by the ratio of resistance before and during exposure. This indicates that ZnO nanostructures obtained by thermal decomposition of LBZA NBs could provide a cost effective route to high sensitivity gas sensors. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Nanoscale semiconductor metal oxide structures have, for the past 10 years, generated great research interest due to their novel properties and potential applications [1]. Among these semiconductors ZnO has received much attention due to its attractive intrinsic properties and the availability of a wide range of nanoscale shapes and geometries [2]. ZnO is piezoelectric, has a wide band gap (3.36 eV) and a large exciton binding energy (60 meV), leading to a broad range of potential applications in microelectronics [3], optoelectronics [4], sensing [5,6], healthcare [7], energy harvesting [8] and photovoltaics [9]. ZnO nanostructures are particularly relevant to applications where the requirement for a high surface area to volume ratio is paramount, such as gas sensors and dye-sensitized solar cells. The

⁎ Corresponding author. Tel.: + 44 7400696296. E-mail address: [email protected] (A. Tarat). 0039-6028/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.susc.2011.12.010

thermal decomposition of layered basic zinc acetate (LBZA) into ZnO provides a simple, high yield, low cost route to the mass production of nanostructured ZnO. The hydrothermal growth of LBZA and subsequent thermal decomposition into ZnO has been reported by a few groups over the last 10 years using zinc acetate and a variety of chemicals [10–14]. Most studies use water to dissolve zinc acetate although methanol has also been employed [11]. Ammonia is often used to control the pH, and surfactants have been reported too [14]. Temperatures between 30 and 60 °C and growth times of 12– 24 h have been reported. However, the published optimum growth parameters are sometimes contradictory. For example Cui et al. reported an optimum pH of 7.2 with no growth above a pH of 8 [10] while Zhang et al. obtained their highest yield at a pH value of 8.2 [14]. Here we report on a simplified method using only zinc acetate and deionized water to produce LBZA NBs, and on the subsequent thermal decomposition into ZnO nanocrystalline NBs. Samples are characterized using scanning electron microscopy (SEM), X-ray Diffraction Spectroscopy (XRD), Atomic Force Microscopy (AFM), X-ray Photoelectron Spectroscopy (XPS) and Photoluminescence (PL).

716

A. Tarat et al. / Surface Science 606 (2012) 715–721

We also present the first investigation of the gas sensing capabilities of polycrystalline ZnO NBs obtained via pyrolytic decomposition of LBZA NBs. 2. Experimental The LBZA NBs were made in a simple one-step process in which 3.29 g Zn(CH3COO)2 2H2O (Sigma Aldrich) was dissolved in 100 ml deionized water using a magnetic stirrer for 5 min at room temperature to make an homogenous solution with a pH of 6.8. The solution was then heated to 65 °C for 20 h in a dry oven. The as-grown product, consisting of a jelly-like phase of mature LBZA crystals (Fig. S1) was filtered using a vacuum filtration system. A white, thin membrane subsequently formed on the filter paper. The membrane was returned to the oven to be dried at the same temperature (65 °C) for 15 min. The LBZA NBs were then annealed in air using a furnace at temperatures from 110 °C to 1000 °C for 10 min. The structure and morphology of the products were characterized by AFM (NanoWizard® II NanoScience, JPK instruments), field emission SEM (Hitachi S4800), XRD (Bruker D8 diffractometer using CuKα radiation and fitted with LynxEYE detector), XPS (VG Escalab II) and PL using a pulsed Nitrogen laser with a wavelength of 337.1 nm. The PL was performed at room temperature and the XRD diffractogram acquired in θ–2θ mode. For the AFM and SEM measurements, a thin layer of the as-grown solution (jelly-like phase) was deposited onto clean silicon substrates or directly onto Aluminum SEM stub so that single NBs could be imaged. Gas sensors were fabricated on a 5 × 4 mm2 alumina substrate with a platinum heating track on the backside and platinum interdigitated electrodes on the top. The substrates were dipped in the as-grown solution and coated with a layer of LBZA materials. The substrates were then annealed at 400 °C for 10 min. The sensing measurements were conducted in a flow of 0.4 l min− 1 of dry synthetic air. The temperature of the sensor was calibrated using an infrared pyrometer and a Keithey 2000 multimeter was used to measure the resistance of the sensing film. 3. Results and discussion 3.1. As-grown LBZA NBs Fig. 1(a) shows a low magnification SEM image of a thick network of LBZA NBs grown at 65 °C, dispersed on a clean silicon substrate and dried at 65 °C. The NBs show typical ribbon-like morphology with widths ranging from 200 nm to several μm and lengths of up to 250 μm. The NBs width appears uniform along their length. The high magnification SEM image of Fig. 1(b) shows two types of typical

morphology, a NB with a uniform flat top surface and a NB made up of two distinct thick layers. The different shapes could be a consequence of the layered growth mechanism. Arrows point towards fine steps at the edge of the flat NB, showing evidence of lamellar growth. The surface of the NBs appears smooth and free of defects. It had been previously reported that the pH had to be raised to 7.2 or 8.2 in order to produce a high yield of NBs [10,14]. This was not found to be the case in this study, as a pH of 6.8 yielded large quantities of NBs. At temperatures under 45 °C we found that no NBs were produced whereas increasing the growth temperature to higher than 80–85 °C resulted in precipitation of micron sized hexagonal ZnO crystals in a variety of sizes. The optimum NB growth temperature for our parameters was found to be 65 °C. With every other parameter kept constant, we investigated the role pH played in the formation and quantity of the final product; pH less than 6 resulted in no LBZA NBs and pH higher than 8 resulted in mostly ZnO nanoparticles, including micron sized hexagonal crystals (Fig. S2). This is in stark contrast to Zhang et al (optimum pH of 8.2) [14] but in good agreement with Cui et al [10] (optimum pH of 7.2) who also reported similar ZnO precipitates at higher pH. The XRD diffractogram of as-synthesized LBZA NBs is presented in Fig. 2 and shows that the NBs are crystalline with the characteristic main zinc acetate 001 peak at 6.67°, corresponding to an interplanar spacing within a single layer of 1.32 nm and confirming their composition as Zn5(OH)8(CH3COO)2.2H2O [10,12]. The peaks at 13.35° and 20.07° are assigned to the 002 and 003 reflections and correspond to interplanar distances of 0.66 nm and 0.44 nm, respectively. The peaks labelled 001 (b) and 002 (b) at 4.48° and 8.90° are assigned to the first and second order reflections corresponding to an interplanar distance of 1.97 nm for the first order reflection. This second group of reflections was also reported by Zhang et al. [14] and Cui et al. [10]. Our diffractogram also shows weaker peaks at 14.56° and 21.87°, which may be attributed to the second and third order of a third group of reflections. The first order peak for this reflection would then be at 7.28°, corresponding to an interplanar spacing within a single layer of 1.21 nm, but would be masked by the main 001 (a) peak at 6.67°. The two peaks at 14.56° and 21.87° could also be discerned on the diffractogram published by Zhang et al. [14], however they did not appear as intense and were not discussed. AFM imaging of the as-grown NBs is shown in Fig. 3. Fig. 3(a) shows several NBs of various widths and distinctive longitudinal terraces clearly apparent on the top of some NBs. Fig. 3(b) shows a close up of a double NB, similar in morphology to the SEM image of Fig. 1(b). The cross sectional line profile of Fig. 3(d) shows that the height of each of the 2 layers is about 20 nm. The flat NBs displayed in Fig. 3(a) have a height of 22 and 24 nm, as shown in Fig. 3(c). Overall the height of the LBZA NBs ranges from 10 to 50 nm. To the best of

Fig. 1. SEM images of LBZA NBs after drying at 65 °C. (a) Thick membrane showing the range of widths and lengths, and (b) isolated NBs showing the layered structure. The arrows in (b) point to steps at the edge of the NB, confirming the layered morphology of the NB.

0

5

003 (a)

002 (a)

002 (b)

001 (b)

Intensity (a. u.)

001(a)

A. Tarat et al. / Surface Science 606 (2012) 715–721

10

15

20

25

2 Theta (degrees) Fig. 2. low angle XRD diffractogram of the as grown LBZA nanobelts showing characteristic zinc acetate peaks corresponding to interplanar spacings of 1.32 nm and 1.97 nm for the 001(a) and 001(b) reflections, respectively. The arrows at 14.56° and 21.87° point to weaker peaks suggesting a third interplanar spacing at 1.21 nm.

our knowledge the height of LBZA NBs has not been systematically measured before, although Cui et al. reported an approximate thickness of 50 nm from side view SEM images [10]. 3.2. Thermal decomposition into ZnO NBs As previously reported, heating the LBZA NBs results in the formation of polycrystalline ZnO via thermal decomposition [10–14]. This is confirmed by the XRD data of Fig. 4, which shows the changes in the diffraction spectra with increasing annealing temperatures. Annealing at 110 °C did not change the crystal structure significantly but after the 210 °C anneal, the XRD scan showed distinct wurtzite ZnO peaks (JCPDS # 01-079-2205). This is in good agreement with Cui et al. who reported a transition to ZnO after annealing in air at

717

150 °C. Annealing at higher temperatures generally increased the intensity of the wurtzite ZnO peaks and decreased their FWHM, indicating an increase in crystallite size with temperature. SEM images of products obtained after annealing the as-synthesized LBZA NBs at increasing temperatures are shown in Fig. 5, while AFM images taken after annealing at 110 °C and 400 °C are shown in Fig. 6. Additionally, AFM images for the 200 °C, 600 °C, 800 °C and 1000 °C anneals are shown in the supplementary information (Figs. S3–S6). After the 110 °C anneal the morphology and surface of the NBs appears to be altered, with the emergence of clear lines along the long axis of the NB and the introduction of structural damage. These are apparent on both SEM and AFM images and therefore are not a result of beam damage. They may be caused by the removal of water from the LBZA compound. Thermogravimetry measurements [15] have shown a 6% weight loss for LBZA at a temperature of 100 °C, attributed to the loss of intercalated water, which could explain the partial break-up of the structure observed in our experiments. After annealing at 210 °C and 400 °C, the surface morphology of the NBs clearly shows a polycrystalline nature, with interconnected nanoparticles arranged in a chain-like pattern along the long axis (Fig. 5(b–d) and Fig. 6(b)). This structure could facilitate electron transport along the ZnO NBs and help lower the resistance of devices fabricated from the NBs, while retaining the high surface to volume ratio of the nanoscale particles. Such performance would be particularly beneficial for gas sensing devices and dye sensitized solar cells. The crystallite size increases after each subsequent anneal, as suggested by the XRD data. Sintering of the crystallite also appears to have taken place after the 600 °C anneal. After annealing at 800 °C and 1000 °C the sintering process intensifies. Yet the overall ribbon shape of the structures remains even after the 1000 °C anneal, similar to results reported by Cui et al [10]. However, the unidirectional arrangement of the nanoparticle chains is lost after annealing at 800 °C, because of the sintering process. The thickness of the NBs was not affected by the annealing process and remained within the 10 to 50 nm range (Figs. S3–S6).

(a)

(b)

60

(c)

20

Height (nm)

Height (nm)

30

10 0

(d)

40 20 0

0

0.5

1

1.5

Lateral distance (um)

2

0

0.5

1

1.5

2

Lateral distance (um)

Fig. 3. (a) and (b) AFM images of as-grown LBZA NBs, showing the characteristic layered structure. (c) and (d) height profiles taken across the dashed arrows in (a) and (b).

(112)

(103)

(110)

(100)

(102)

1000°C 800°C 600°C 400°C 210°C

(003)

(002)

Normalised Intensity (a. u.)

(002)

A. Tarat et al. / Surface Science 606 (2012) 715–721

(101)

718

110°C 65°C

10

20

30

40

50

60

70

80

2 Theta (degrees) Fig. 4. XRD measurements of as-grown LBZA NBs (65 °C), and after annealing at increasing temperatures from 110 to 1000 °C.

The SEM, XRD and AFM analyses of the thermal decomposition of the LBZA NBs therefore show a range of ZnO NBs with temperaturedependent crystallite size, which could be used to tailor sensitivity in gas sensing devices. Table 1 summarises the grain sizes extracted from SEM and XRD. The sizes from both techniques are in very good agreement, considering that XRD averages over a large area and measures film properties such as coherence length perpendicular to the substrate surface, while SEM measures only a very small part of the sample providing two-dimensional information in the substrate plane. Thus, the shape of the particle will affect the measurements differently. This could explain the discrepancies at 400 °C and 600 °C. The large standard deviation observed at temperatures of 600 °C and above is likely to be due to the sintering process, resulting in the creation of large particles while retaining some of the initial small particles, as shown in the SEM images. Along with XRD, XPS was used to compare the composition of annealed ZnO NBs with LBZA NBs. Fig. 7 shows the O1s core level scans of the as-grown LBZA NBs and the ZnO NBs produced by annealing at 400 °C in air for 10 min. The ZnO NBs spectrum could

be fitted with O–Zn, OH and H2O components in good agreement with published data for ZnO [16]. The control scan of commercial ZnO powder is virtually identical to that of the ZnO NBs. The Zn2p to O1s ratio measured from the ZnO NBs is almost double that of the LBZA material, which is consistent with the stochiometry of the compounds. The LBZA O1s core level was fitted with a Zn–OH component and a Zn–O component, using the same energy offset between the two as for the ZnO NBs (1.4 eV). The FWHM of the components used to fit the LBZA NBs are much larger than those of the ZnO NBs, possibly because of charging effects broadening the peak. This could explain why no separate H2O component was observed as the energy separation between OH and H2O might be too small to be resolved. The charging of the insulating LBZA sample resulted in a 5.9 eV difference for the binding energy of the C 1 s core level between LBZA and ZnO. Therefore, in order to compensate for charging, the LBZA NB spectrum for O1s was normalized to the C 1 s energy and shifted 5.9 eV to lower binding energy. The Zn 2p core level peak shape was similar for both samples, possibly because the energy offset between Zn–O and Zn–OH is below the resolution of the instrument. Fig. 7 clearly shows the chemical transition from LBZA to ZnO, with the LBZA scan dominated by O–H bonding, which decreases after annealing, while the O–Zn bond increases. The PL spectra of the as-synthesized LBZA belts and the 400 °C annealed sample are presented on Fig. 8. The PL results show the transition from LBZA to ZnO after annealing at 400 °C which is in good agreement with the XRD and XPS data. The as-grown LBZA NBs have a broad emission centered around 415 nm. After annealing at 200 °C the characteristic ZnO near band edge emission at 385 mm is observed and the spectra taken from the higher temperature samples do not show any significant changes. Relative to the band edge emission, no visible emission is observed from the defect band in any of the annealed samples. The luminescence lifetime of the as-grown LBZA NBs is 3.0 ns at 380 nm and 4.4 ns at 410 nm. Following annealing, the lifetime of the ZnO band edge emission at 380 nm is shorter than 500 ps, the detection limit of our system. This matches well with short (b100 ps) luminescence lifetimes observed for ZnO nanoparticles [17] and again highlights the transition from LBZA to ZnO upon annealing.

Fig. 5. SEM images of ZnO polycrystalline NBs annealed at (a) 110 °C, (b) 210 °C, (c) 400 °C, (d) 600 °C, (e) 800 °C and (f) 1000 °C at 60 k magnification.

A. Tarat et al. / Surface Science 606 (2012) 715–721

719

(a)

(b)

30

(c)

40

Height (nm)

Height (nm)

50 30 20 10 0

(d)

20 10 0

0

1

2

0

3

0.2

Lateral distance (um)

0.4

0.6

0.8

1

Lateral distance (um)

Fig. 6. AFM images of ZnO NBs produced at (a) 110 °C and (b) 400 °C and corresponding height profiles (c) and (d), taken across the dashed arrows in the AFM images. The chainlike arrangement of the nanoparticles within the NBs is apparent in the 400 °C image.

3.3. Gas sensing behavior of ZnO NBs The gas sensing potential of the polycrystalline ZnO NBs was tested by measuring the change in resistance of a network of NBs annealed at 400 °C) as a function of CO concentration and temperature. The structure of the sensing network is shown in Fig. S7. The gas sensing apparatus, developed in-house, consisted of an array of four mass flow controllers (MFCs), a vacuum chamber, sensor holder, and a Keithley 2000 multimeter connected to a PC for data acquisition. Two of the MFCs were used to produce dry air from zero grade O2 and N2, while another MFC regulated a pre-mixed flow of 1000 ppm

400C

Intensity (a. u.)

H2O

O-Zn O-H

Table 1 Average size of ZnO grains/particles, extracted from XRD/SEM measurements, as a function of temperature.

As grown

525

526

527

of CO in N2. The concentration of the CO in the test chamber was controlled by setting this flow rate as well as the synthetic air flow rate to the desired ratio. The last MFC was used for a compensatory flow of N2 while the sensor was not being exposed to the CO/N2 mixture. This ensures the total flow remained constant throughout the experiments at 400 sccm and, more importantly, that the oxygen concentration within the test chamber remained unchanged; thus, any change in the sensor's resistance was solely due to the presence/absence of CO. Care was taken when designing the apparatus to avoid pressure build up in the gas lines, which could greatly affect both the speed and magnitude of the sensor's response to CO. The response is defined as the ratio of the resistance in air and the resistance in the presence of CO. Fig. 9(a) shows the changes in resistance to decreasing concentrations of CO at 400 °C, from 200 ppm to 12.5 ppm. The data acquisition was stopped for a few minutes (less than 4) between each exposure in order to adjust the different flows for the next exposure. This explains the small discontinuities (less than 1 kΩ) in the resistance plot between exposures. The response was calculated using the stable region of the plot during CO exposure, after the initial dip, and the resistance immediately before exposure. The response time, defined as the interval between 10% and 90% of the resistance decrease was 15 s for the 200 ppm exposure in our set-up.

528

529

530

531

532

533

534

535

Binding Energy (eV) Fig. 7. O1s core level XPS spectra of as-grown LBZA NBs and after annealing at 400 °C. The raw data points are displayed as open circles while the fitted components are shown as solid lines.

Temperature XRD grain (°C) size (nm)

XRD standard deviation (nm)

SEM particle size (nm)

SEM standard deviation (nm)

210 400 600 800 1000

0.058 0.09 0.18 0.29 0.56

9.25 14.3 30.7 55.2 89.6

2.8 5.9 20.4 32.5 41

9.18 18.6 42.4 53.9 95.2

A. Tarat et al. / Surface Science 606 (2012) 715–721

Normalised intensity (a.u)

720

65C 400C

Wavelength (nm) Fig. 8. PL measurement of the as-grown LBZA structures and after annealing at 400 °C.

It is well known that CO sensing with metal oxide semiconductors is based on electron exchanges between the CO molecules and the oxygen species ionosorbed on the surface (O2−, O −, O 2 −) [18]. For example, CO reacts with O 2− according to the reaction 2CO + O 2− => 2CO2 + 2e − , therefore increasing the amount of conduction electrons and the overall conductivity for n-type semiconductors. In the case of polycrystalline sensing films, these reactions also control the height of the potential barrier between adjacent particles and therefore electron transport, leading to higher sensitivity. In the absence of reducing gas, the oxygen ions create a depletion region and lead to a high potential barrier, which is lowered when CO molecules react with the oxygen [19,20]. The response curves of Fig. 9(a) show a dip in the resistance within 30 s after exposure to CO, followed by a slower increase in resistance until equilibrium was reached after 5 min. The initial dip has been observed before [21] when sensing CO with SnO2 thin films, but its origin is not clear. It could be caused by competing reactions as CO

can also react directly with the sensing surface and lead to an increase in resistance. It is also possible that the initial decrease is caused by CO reacting with ionosorbed oxygen species, which is followed by a slower reaction between CO and the ZnO NBs, leading to an increase in resistance. Fig. 9(a) also shows that the resistance increases rapidly when the CO is turned off, but does not go back to its pre-exposure value. Instead, this rapid recovery is followed by a slower increase. This behavior is probably due to residual CO being gradually flushed from the test chamber and is inherent to the design and relatively small total flow rate (400 sccm). The background resistance was also affected by small periodical variations of the air temperature in the laboratory (caused by the air conditioning system) which induced an oscillation with an amplitude of 500 Ω and a period of 13 min. These factors account for the slight variations of the pre-exposure resistance for each concentration. Fig. 9(a) shows a stable, reproducible behavior for the sensor as well as a relatively low resistance of 83 kΩ at 400 °C, compared to published data for nanocrystalline ZnO, where the resistance of the sensors is in the MΩ range [22–24]. This is possibly a consequence of the chain-like structure of the polycrystalline ZnO NBs providing enhanced conduction along the long axis of the NBs (Fig. S7). The response as a function of CO concentration is shown on Fig. 9(b) and increases linearly from 1.04 at 12.5 ppm to 1.62 at 200 ppm. These values are comparable to published work [23] and could be considerably improved by optimizing the deposition process and thickness of the sensing film. Comini et al. showed that the sensitivity of devices fabricated with single crystal SnO2 NBs depends greatly on film thickness, which can also affect the optimum temperature [25]. The linear behavior of the response shape has been reported before [26] and is consistent with low gas concentrations. Fig. 9(c) shows the variation of the response to 200 ppm of CO at different temperatures, peaking at 400 °C. This is in good agreement

85

250

80

Resistance (kOhm)

200 75 70

150

65 100

60 55

50

CO concentration (ppm)

(a)

50 45

0 0

10

20

30

40

50

60

70

80

Time (min)

CO response

CO response

(b)

1.6 1.4 1.2 1 0

100

CO concentration (ppm)

200

1.6

(c)

1.4 1.2 1 300

350

400

450

sensor temperature (C)

Fig. 9. Response to CO of a mat of ZnO NBs produced by annealing at 400 °C. (a) Variation of the resistance in the presence of various CO concentrations at 400 °C. (b) Sensor response as a function of CO concentration at 400 °C. The linear fit has a correlation coefficient of 0.99 and a p value of less than 0.01. (c) Response to 200 ppm CO as a function of sensor temperature. The response was extracted from the resistance plot in (a) using the resistance before CO exposure and the resistance at equilibrium during the CO exposure. The discontinuities between the end of an exposure and the start of the next are caused by time gaps of no more than 4 min.

A. Tarat et al. / Surface Science 606 (2012) 715–721

with previously published results for ZnO sensing. Ryu et al. reported an optimum operating temperature of 350 °C for CO [24] while Bai et al. also reported an optimum temperature of 400 °C for NO2 detection with their ZnO nanorod sensors [27]. Xu et al., who investigated the response of ZnO to various organic compounds [28], found that the highest sensitivity ranged from 370 °C to 420 °C, depending on the fabrication method. In order to check the effect of operating the sensor for several days at elevated temperature, including 8 h at 400 °C and 2 h at 425 °C, the sensor was inspected with SEM before and after the sensing experiments. The SEM shows an increase in particle size from 16.3 nm before sensing to 22.9 nm after (Fig. S8). This is consistent with Ostwald ripening of the nanoparticles [29] and indicates that sensing devices fabricated from LBZA NBs should be annealed at least to 500 °C in order to remain stable at their peak operating temperature. 4. Conclusion We have presented an investigation of the thermal decomposition of LBZA NBs into nanocrystalline ZnO NBs and their gas sensing behavior. The LBZA NBs were produced by a high yield, low cost, low temperature hydrothermal technique. The NBs have widths ranging from 200 nm to several μm and lengths in the 100 s of μms range. The height of the LBZA NBS ranges from 10 to 50 nm with morphologies typical of lamellar growth. Thermal decomposition into ZnO occurs after annealing in air at 210 °C and produces ZnO NBs consisting of nanoparticles arranged in a chainlike pattern along the long axis of the NBs. Annealing at higher temperatures results in an increase of the nanoparticle size and sintering was observed after 600 °C. The unidirectional pattern was lost after 800 °C. The response to 200 ppm CO of a network of ZnO NBs produced by annealing asgrown LBZA NBs reached a maximum of 1.62 at 400 °C. The response to CO was linear in the range 12.5 ppm to 200 ppm. The relatively low operating resistance of 83 kΩ is possibly a consequence of the chain-like arrangement of the nanocrystals within the ribbons. This is an attractive property for the integration of the sensors into existing circuitry. The sensitivity and repeatability of the sensor, coupled with the simplicity, high yield and low cost of the synthesis, demonstrate the viability of the material as a novel route for the mass production of high sensitivity sensors. Acknowledgements The authors would like to thank Dr Javier Jo and Alejandra Sancho for their help with the PL measurements and Dr Nattamai Bhuvanesh for his assistance with the XRD data.

721

Appendix A. Supplementary data Supplementary data to this article can be found online at doi:10. 1016/j.susc.2011.12.010.

References [1] Z.W. Pan, Z.R. Dai, Z.L. Wang, Science (New York, N.Y.) 291 (5510) (Mar. 2001) 1947. [2] Z.L. Wang, et al., Adv. Funct. Mater. 14 (10) (Oct. 2004) 943. [3] P.-C. Chang, Z. Fan, C.-J. Chien, D. Stichtenoth, C. Ronning, J.G. Lu, Appl. Phys. Lett. 89 (13) (2006) 133113. [4] M. Willander, et al., Nanotechnology 20 (33) (Aug. 2009) 332001. [5] Z. Fan, D. Wang, P.-C. Chang, W.-Y. Tseng, J.G. Lu, Appl. Phys. Lett. 85 (24) (2004) 5923. [6] Q. Wan, et al., Appl. Phys. Lett. 84 (18) (2004) 3654. [7] R. Yakimova, et al., Biosens. Bioelectron. 22 (12) (Jun. 2007) 2780. [8] Y. Qin, X. Wang, Z.L. Wang, Nature 451 (7180) (Feb. 2008) 809. [9] H.E. Unalan, et al., Appl. Phys. Lett. 93 (13) (2008) 133116. [10] Q.Y. Cui, K. Yu, N. Zhang, Z.Q. Zhu, Appl. Surf. Sci. 254 (11) (2008) 3517. [11] E. Hosono, S. Fujihara, T. Kimura, H. Imai, J. Colloid Interface Sci. 272 (2) (Apr. 2004) 391. [12] H. Morioka, H. Tagaya, J. Kadokawa, K. Chiba, Mater. Sci. 8 (1999) 995. [13] R.-Q. Song, a-W. Xu, B. Deng, Q. Li, G.-Y. Chen, Adv. Funct. Mater. 17 (2) (Jan. 2007) 296. [14] Y. Zhang, F. Zhu, J. Zhang, L. Xia, Nanoscale Res. Lett. 3 (6) (Jun. 2008) 201. [15] E. Hosono, S. Fujihara, T. Kimura, H. Imai, J. Colloid Interface Sci. 272 (2) (Apr. 2004) 391. [16] D. Byrne, E. McGlynn, M.O. Henry, K. Kumar, G. Hughes, Thin Solid Films 518 (16) (Jun. 2010) 4489. [17] K.H. Tam, et al., J. Phys. Chem. B 110 (42) (Oct. 2006) 20865. [18] D. Kohl, Sens. Actuators, B 1 (1–6) (1990) 152. [19] P.T. Moseley, Meas. Sci. Technol. 8 (3) (Jan. 1997) 223. [20] D. Yoon, Sens. Actuators, B 45 (3) (Dec. 1997) 251. [21] R. Sharma, Sens. Actuators, B 72 (2) (Jan. 2001) 160. [22] X. Jiaqiang, C. Yuping, C. Daoyong, S. Jianian, Sens. Actuators, B 113 (1) (Jan. 2006) 526. [23] J.X. Wang, et al., Nanotechnology 17 (19) (Oct. 2006) 4995. [24] H. Ryu, Sens. Actuators, B 96 (3) (Dec. 2003) 717. [25] E. Comini, G. Faglia, G. Sberveglieri, D. Calestani, L. Zanotti, M. Zha, Sens. Actuators, B 111–112 (Nov. 2005) 2. [26] S.K. Gupta, A. Joshi, M. Kaur, J. Chem. Sci. 122 (1) (Feb. 2010) 57. [27] S. Bai, X. Liu, D. Li, S. Chen, R. Luo, A. Chen, Sens. Actuators, B 153 (1) (Mar. 2011) 110. [28] H. Xu, X. Liu, D. Cui, M. Li, M. Jiang, Sens. Actuators, B 114 (1) (Mar. 2006) 301. [29] C. Pacholski, A. Kornowski, H. Weller, Angew. Chem. Int. Ed. Engl. 41 (7) (Apr. 2002) 1188.