IOP PUBLISHING
NANOTECHNOLOGY
Nanotechnology 18 (2007) 125502 (6pp)
doi:10.1088/0957-4484/18/12/125502
Columnar CeO2 nanostructures for sensor application Davide Barreca1,4 , Alberto Gasparotto2 , Chiara Maccato2 , Cinzia Maragno2 , Eugenio Tondello2 , Elisabetta Comini3 and Giorgio Sberveglieri3 1
ISTM-CNR and INSTM, Department of Chemistry, Padova University, Via Marzolo, 1-35131 Padova, Italy 2 Department of Chemistry, Padova University and INSTM, Via Marzolo, 1-35131 Padova, Italy 3 INFM-CNR, SENSOR, Brescia University, Via Valotti, 9-25133 Brescia, Italy E-mail:
[email protected]
Received 11 December 2006, in final form 19 January 2007 Published 21 February 2007 Online at stacks.iop.org/Nano/18/125502 Abstract CeO2 columnar nanostructures with tailored properties were synthesized by chemical vapour deposition on Si(100) and Al2 O3 substrates between 350 and 450 ◦ C and characterized in their structure, composition and morphology by means of a multi-technique approach. Their higher sensitivity in the detection of gaseous ethanol and nitrogen dioxide with respect to continuous CeO2 thin films opens interesting perspectives for the development of nanosized sensor devices.
to the development of quantum-confined structures [6]. Moreover, undesired coalescence events occurring at the high temperatures required by surface reactions can be conveniently avoided, leading thus to the obtaining of stable functional materials [6]. On this basis, the preparation of CeO2 anisotropic nanostructures has considerably developed over the last decade. In particular, several liquid-phase routes, such as sonochemical methods, self-assembly of cerium dioxide nanoparticles by the microemulsion technique, sol–gel processing into porous membranes serving as templates and hydrothermal methods have been adopted [7–10]. These approaches suffer from several drawbacks, including the difficult template removal and material recovery. In recent years, the chemical vapour deposition (CVD) method has gained a great deal of attention as an amenable route towards the production of supported anisotropic nanostructures, thanks to its versatility in tailoring the morphogenesis of the obtained systems [11, 12]. Nevertheless, previous investigations have focused on the use of metal nanoparticles to tailor the system architecture, resulting in undesired complications, as well as in the introduction of metal impurities during the preparation process. Recently, we have developed a template- and catalystfree CVD approach to ordered columnar CeO2 nanostructures on Si(100) substrates at moderate growth temperatures [13].
1. Introduction Rare-earth metal oxides have been extensively explored for several advanced applications, such as in electronics, optics and heterogeneous catalysis, thanks to their peculiar properties arising from the availability of the 4f shell. In this context, cerium dioxide (ceria CeO2 ) has been widely employed in different fields, such as UV blockers, three-way catalysts (TWCs), solid electrolytes for fuel cells and solid state gas sensing devices [1–3]. In the latter case, the sensing properties are based on surface reactions upon exposure to probe gases, inducing a variation in the system conductance due to electronic transfer with adsorbed species [4]. These applications offer interesting opportunities to investigate the influence of size, shape and composition on the system functional properties [5]. In particular, CeO2 anisotropic nanosystems have attracted increasing attention, thanks to their peculiar morphology and the higher surface-to-volume ratio with respect to 2D thin films and bulk materials. In addition, the effect of microstructure and particle size on the gas sensing performance is of significant importance, since reactions at the grain boundaries and significant depletion of carriers in anisotropic nanosystems can strongly modify the redox/transport properties, leading ultimately 4 Author to whom any correspondence should be addressed.
0957-4484/07/125502+06$30.00
1
© 2007 IOP Publishing Ltd Printed in the UK
Nanotechnology 18 (2007) 125502
D Barreca et al
The present paper focuses on a further insight into the system properties, devoted to the substrate influence on the sample microstructure and to the analysis of functional performances in gas sensors. Beside Si(100), polycrystalline Al2 O3 has also been considered in view of gas sensing tests. First, the system’s structural and morphological characteristics as a function of the preparative conditions were investigated by a multi-technique approach. Subsequently, the sensor activity has been characterized for selected probe gases, in particular CH3 CH2 OH (a reducing species), employed for breath analysers and food control utilization, and NO2 (an oxidizing gas), interesting for environmental applications [3, 6, 14–16].
2. Experimental details CVD preparation of columnar ceria nanostructures was performed by a custom-built cold-wall reactor under optimized conditions (1:1 N2 + O2 reaction atmosphere; total pressure = 10 mbar; substrate temperature = 350–450 ◦ C). At variance with previous reports [17], the introduction of water vapour into the reaction atmosphere was carefully avoided in order to limit the formation of nucleation centres in high density, resulting ultimately in a continuous film with well-interconnected grains. The precursor Ce(hfa)3 ·diglyme (Hhfa = 1,1,1,5,5,5-hexafluoro-2,4pentanedione; diglyme = bis(2-metoxyethyl)ether) was synthesized according to the literature [17] and vaporized at 80 ◦ C throughout each deposition. p-type Si(100) wafers and polycrystalline Al2 O3 were used as substrates and suitably cleaned prior to deposition in order to minimize the presence of surface contaminants. For comparison, continuous CeO2 films were prepared at the same growth temperatures by adding H2 O vapour to the reaction mixture, as reported elsewhere [17]. X-ray diffraction (XRD) measurements were performed by a Bruker D8 Advance diffractometer, equipped with a G¨obel mirror and a Cu Kα source (40 kV, 40 mA). Field emissionscanning electron microscopy (FE-SEM) and energy dispersive x-ray analysis (EDX) were performed by a Zeiss SUPRA 40VP instrument operated at acceleration voltages lower than 10 kV, equipped with an Oxford INCA x-sight x-ray detector. X-ray photoelectron spectroscopy (XPS) analyses were performed on a Perkin Elmer 5600ci spectrometer with an Al Kα source. A schematic diagram presenting the sensor structure is proposed in figure 1. Gas sensor fabrication involved the deposition of two Pt interdigitated electrodes by platinum sputtering with shadow masking on the specimen surface (spacing between electrodes = 190 μm). The sensor operating temperature was fixed keeping constant the resistance of a platinum heater, deposited on the backside of the Al2 O3 substrate, and calibrated using an IR camera in order to control its temperature coefficient. A good temperature uniformity, checked by the IR camera, could be achieved on the sensing area thanks to the meander structure (compare figure 1) and the thin (250 μm) alumina substrate. The flow-through technique was used to determine the gas-sensing properties of the ceria nanostructures. A constant synthetic air flow (0.3 l min−1 ), mixed with the desired amount of gaseous species, was passed through a stabilized sealed chamber maintained at 20 ◦ C, atmospheric pressure and
Figure 1. Schematic diagram presenting the sensor device structure.
constant humidity level (40%), as measured by a humidity sensor. For the production of water-saturated air, synthetic dry air was first flowed through a water reservoir kept at a constant temperature and, subsequently, through a condenser maintained at the temperature of the test chamber. The desired relative humidity was obtained by mixing with dry air. Electrical data were registered by a personal computer that controlled a picoammeter (Keithley 486) and a multiplexer (Keithley 7001 SWITCH SYSTEM) via GPIB (IEEE 488). All the connections were made by coaxial cables and BNC from the test chamber to the picoammeter. The sensor was bonded on a microelectronic standard case and inserted in a standard socket into the test chamber. The characterization was carried out by a volt– amperometric technique at a constant bias of 1 V, chosen in the ohmic range in order to yield a linear current variation. DC measurements were preferred to AC ones in view of an easier technological application of the obtained systems. The conductance has been calculated as the ratio between the measured current (sampling time = 60 s) and the applied voltage. Due to the low measured conductance values, a smoothing function which computes a moving average of the acquired data was used. For an n-type semiconductor such as ceria, conductance or resistance values increase upon exposure to reducing or oxidizing gases, respectively. In order to evaluate the sensor response, it is a well established practice to consider the relative variations of the increasing quantity, i.e. conductance or resistance. In the case of ethanol, a reducing gas, the response S is defined by equation (1):
S = (G f − G 0 )/G 0 = G/G
(1)
while in the case of NO2 , an oxidizing species, formula (2) is used: S = (R f − R0 )/R0 = R/R (2) where G 0 (R0 ) are the baseline conductance (resistance) values measured in air flow, while G f (R f ) are the corresponding steady state values upon gas exposure. The sensor response was investigated in the working temperature range 200–500 ◦ C. For the detection of ethanol 2
Nanotechnology 18 (2007) 125502
D Barreca et al
and nitrogen dioxide, the optimal working temperatures were 400 and 200 ◦ C, respectively. Nevertheless, the results reported in this work were obtained at 200 ◦ C also for ethanol in order to show the low-temperature performance of the synthesized materials and to compare the sensor response to the two probe gases at the same temperature. As a consequence, even the response of a homogeneous layer to the same analytes (figure 6) was obtained at the same working temperature. All samples were annealed ex situ in air at 600 ◦ C for 10 h in order to thermally stabilize the deposited material before sensing measurements.
3. Results and discussion The system structure and its crystallinity were characterized by XRD. Irrespective of the growth surface, all the synthesized specimens had the cubic CeO2 structure [18]. In fact, typical diffraction patterns for Al2 O3 -supported systems (figure 2) displayed peaks located at 2ϑ = 28.6◦ , 33.1◦ and 47.5◦ , ascribed to (111), (200) and (220) ceria reflections, in addition to alumina signals. Notably, the relative peak intensity was directly affected both by the deposition temperature and the growth substrate. As concerns Al2 O3 -supported specimens, the I111 /I200 intensity ratio was higher than that of the powder spectrum and the value expected for cubic ceria was approached only at the highest adopted temperatures (450 ◦ C, see inset in figure 2) [18]. A different trend has been observed on Si(100) substrates. In this case, the I111 /I200 ratio was lower than that of the powder CeO2 spectrum up to 400 ◦ C, and underwent a significant increase at higher growth temperatures. The observed variation might be explained by taking into account the concurrence of two phenomena: (i) the presence of preferred orientations, depending both on the growth temperature and the used substrate; (ii) an anisotropic crystallite growth [19]. In particular, the latter hypothesis, also justified by the observed columnar morphology (see below), might explain the high I111 /I200 intensity ratio, especially for Al2 O3 -supported samples at T 400 ◦ C. In fact, at variance with Si(100), the polycrystalline alumina nature and its high surface corrugation suggests a predominance of an anisotropic growth mechanism over a substrate effect on the structural orientation. Figure 3 reports a representative FE-SEM cross-sectional micrograph for a Si(100)-supported specimen, displaying columnar ceria nanostructures with an average length of ≈330 nm and a density of 1010 columns cm−2 . A qualitatively similar morphology was observed on alumina substrates (figure 4), in spite of their different structural features with respect to Si(100). The most relevant difference was due to the higher lateral dimensions for Al2 O3 -supported systems (≈80 nm versus ≈40 nm) and to a more disordered growth on alumina owing to its remarkable roughness. Interestingly, irrespective of the adopted substrate, the observed nanostructures present spiral-like heads, associated with a spiral dislocation mechanism occurring along the crystallographic direction perpendicular to the substrate surface [13–20], that points out a direct connection between the system morphology and the anisotropic crystalline microstructure. The growth directions correspond to [100] and [111] for Si(100) and Al2 O3 -supported samples, respectively.
Figure 2. XRD spectra of CeO2 samples deposited on Al2 O3 as a function of the deposition temperature. Vertical bars mark the cubic CeO2 peak positions [18], while stars (*) indicate reflections attributed to the Al2 O3 substrate. Inset: I111 /I200 intensity ratio versus substrate temperature for Si(100) and Al2 O3 -supported specimens. For comparison, the ratio reported for ceria powders [18] is displayed as a dashed line.
In fact, the nature of the latter substrate can hardly induce an anisotropic growth or texturing along directions different from the principal [111] one. Due to the high surface roughness of the alumina substrates, no marked variations of the morphology of CeO2 /Al2 O3 columnar nanostructures were observed as a function of the deposition temperature. Notably, ex situ annealing in air at 600 ◦ C did not induce appreciable morphological variations, as confirmed by FE-SEM measurements. As concerns the system’s chemical composition, XPS surface analyses showed the clear predominance of Ce(IV) over Ce(III) species. Fluorine contamination (
