Mesoporous ceramic oxides as humidity sensors: A case study for gadolinium-doped ceria

Sensors and Actuators B 216 (2015) 41–48

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb

Mesoporous ceramic oxides as humidity sensors: A case study for gadolinium-doped ceria L. Almar a , A. Tarancón a , T. Andreu a , M. Torrell a , Y. Hu b , G. Dezanneau b , A. Morata a,∗ a Advanced Materials for Energy, Catalonia, Institute for Energy Research (IREC), Jardí de les dones de negre, 1, Planta 2, 08930 Sant Adrià del Besòs (Barcelona), Spain b Laboratoire Structures, propriétés et Modélisation des Solides, CentraleSupelec, CNRS, Grande voie des vignes, 92295 Chatenay-Malabry Cedex, France

a r t i c l e

i n f o

Article history: Received 11 February 2015 Received in revised form 1 April 2015 Accepted 8 April 2015 Available online 17 April 2015 Keywords: Mesoporous Humidity sensors Ceramic Ceria Thermal stability

a b s t r a c t Mesoporous materials have been studied as high performance sensing materials due to their singular microstructure and extremely high surface-to-volume ratio. However, the lack of stability of these nanostructures is assumed as one of the major drawbacks toward their application in real devices. In this work, this limitation is overcome by the synthesis of thermally stable mesoporous gadolinium doped ceria. Humidity sensors were fabricated and tested under different (i.e. humidity and temperature) conditions. The mesoporous layers were attached to the substrate at 900 ◦ C preserving mesoporous structure intact. This process at high temperature provides the layer with a mechanical strength and allows self-cleaning cycles at high temperatures if required. The humidity sensing mechanism is presented and discussed in detail by means of impedance spectroscopy. An ionic type of conduction mechanism is corroborated. Fast response and recovery, as well as very low hysteresis and no drift are observed. It was also shown that the response of the devices can be straightforwardly tuned by changing layer thickness or pore size, allowing to fulfill sensing needs of different applications. All the mentioned properties joined to the simplicity of the fabrication and the flexibility of the used fabrication route for synthesizing any other metal oxide make this kind of devices a potential group for developing high performance and fast gas sensors. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The humidity control is of utmost importance as it affects both living organisms and materials. Therefore, humidity sensors are currently used in many fields such as manufacturing processes, environmental and health monitoring or domestic applications [1]. Porous ceramics form a common group of materials to sense humidity [2,3]. The adsorption mechanism of the water molecules on the surface of the ceramic oxides is well known (Fig. 1). First, a chemisorbed layer is formed on the available sites of the oxide surface, causing dissociation of water molecules to surface hydroxyls. After the first adsorbed layer is formed, subsequent layers are physisorbed. The molecules in the first physisorbed layer are bounded through double hydrogen bonds and cannot move. At higher humidity levels, water molecules are singly bounded and form a network with a liquid-like behavior. Thus, the water molecules from the second layer can move or rotate freely (like

∗ Corresponding author. Tel.: +34 606522639. E-mail address: [email protected] (A. Morata). http://dx.doi.org/10.1016/j.snb.2015.04.018 0925-4005/© 2015 Elsevier B.V. All rights reserved.

in bulk liquid water) and the Grotthuss mechanism i.e. the proton transport becomes dominant [2]. From the surface adsorption mechanism, it is clear that large pore volumes are desirable in order to get highly sensitive humidity sensors. Moreover, the response and recovery times of the ceramic sensors can be controlled by the pore size. The detection limit can be set very low by decreasing the pore size [3]. In general homogeneity and controllability of porous structures are important properties in sensor applications. Mesoporous materials provide high area structures with a very precise shape, which make them perfect candidates for sensing applications [4–7]. In this respect, many studies have reported high sensitivity in silica based mesoporous humidity sensors [8–11]. However, during the fabrication of the sensor, the mesoporous powder is usually attached to the electrodes at very low temperatures in order to preserve the nanostructure. This limitation makes these devices mechanically unstable and affects reproducibility, limiting their viability for commercial applications. Moreover a heat treatment at temperatures over 400 ◦ C is normally used in commercial ceramic humidity sensors to recover the initial state [1,3]. However, if this heat treatment process is

42

L. Almar et al. / Sensors and Actuators B 216 (2015) 41–48

Fig. 1. Multi-layer adsorption of condensed water on the surface of the ceramic oxide CeO2 .

performed, the sensing properties of the silica based mesoporous sensors become poor [12,13], due to either loss of the mesoporosity originated by the thermal treatment at high temperature or the poor attachment of the sensing layer to the substrate. Mesoporous metal oxides with higher thermal stabilities would be ideal candidates, if they could keep their microstructure after long operating times and periodic thermal regeneration cycles, preserving their sensing properties and increasing their lifetime. In a previous work, the authors developed a novel strategy to stabilize mesoporous ceramic oxides up to high temperatures [14]. Contrary to widely employed high temperature stabilization treatments inside the mesoporous template, resistance to thermal processes is achieved by getting the simultaneous presence of amorphous and crystalline phases before the template removal, taking profit of the so called self-limited grain growth regime [15–17]. Using this strategy, stabilization of mesoporous structures is achieved up to temperatures as high as 1000 ◦ C. Among the ionic type of ceramic sensors, nanostructured CeO2 has shown high sensitivity and fast response [18–20]. Moreover, it has been reported that metal doping changes the charge density surface of the ceramic oxide [20]. The strong electric field generated around the surface of the material is expected to increase the ionization of the water molecules [21]. In the present work, gadolinium-doped ceria (CGO) temperature-stable mesoporous humidity sensors were fabricated and tested under different (i.e. humidity and temperature) conditions. The sensitivity, response and recovery times and the sensing mechanisms based on the analysis of the impedance spectra are presented and discussed. 2. Materials and methods 2.1. Fabrication of the mesoporous humidity sensors Mesoporous gadolinium doped ceria (CGO) powder was synthesized as a replica of the KIT-6 silica template by a multi-step infiltration filling both interpenetrated channels of the replica, as described in detail elsewhere [14] and from now on named here CGO. In addition, mesoporous powder was fabricated by filling one of the two interpenetrated channels of the original silica template (from now on named CGO-D). Both mesoporous replicas were calcined at 600 ◦ C. A structural analysis and comparison between CGO and CGO-D is presented in the Supplementary Information and in Section 3.6. The main difference between the two powders lies in the pore size which is 3.9 nm for CGO and 11.3 nm for GCO-D. Both powders present very similar specific surface around 120 m2 /g. Sensors were fabricated from ethanol inks based on mesoporous powder (30 wt.%). Inks of CGO were airbrushed on a commercial alumina substrate with interdigitated gold electrodes (Planar IDE Au, Synkera) and sintered at 900 ◦ C for 1 h to ensure mechanical strength and thermal stability (Fig. 2). The so-prepared mesoporous

layer has dimensions of 6.35 mm length and two different thicknesses were tried 12–13 ␮m (named CGO) and 4 ␮m (Section 3.5, named CGO-F). 2.2. Methods of characterization The characterization of the specific surface area is determined by means of Brunauer–Emmet–Teller (BET) method. Nitrogen physisorption measurements at 77 K were performed in a Micromeritics Tristar 3000 surface analyzer. The total volume of pores was obtained from a single point adsorption method at P/P0 = 0.999 and the pore size distribution from the analysis of the desorption branch of the isotherms using the Barrett–Joyner–Halenda (BJH) method. Transmission electron microscopy (TEM) was used in order to confirm the mesoporous structure of the sintered replicas by a JEOL JEM 2010 (200 kV) electronic microscope. The microstructure and thickness of the humidity sensors were analyzed by scanning electron microscopy (SEM) using a ZEISS AURIGA instrument. X-ray diffraction (XRD) was performed in a Bruker D8 automated diffractometer with Ni filter and Lynx Eye detector, to study the crystallite and the periodic structure of the material. Wide angle XRD spectra were recorded in the 2 range from 20◦ to 100◦ and low angle XRD spectra from 0.5◦ to 3.5◦ . The sensors were electrically characterized by impedance spectroscopy using a frequency response analyzer (Novocontrol Alpha-A) in a range of frequencies from 1 MHz to 0.1 Hz and an AC signal of 300 mV. Samples were placed in a Linkam chamber and the humidity level was fixed with a commercial humidification system (Bronkhorst) composed of a liquid flow controller (LFC) and a controlled evaporator mixer (CEM) from 0 to 3 ml/min of H2 O in a temperature range from 30 to 90 ◦ C. 3. Results and discussion 3.1. Microstructural characterization Humidity sensors were fabricated with mesoporous gadolinium-doped-ceria (CGO) powder obtained as replica of the KIT-6 silica template. A complete micro-structural characterization of the CGO powder was recently presented [14] (see also Supplementary Information). After deposition and a sintering process at 900 ◦ C, the layers are homogeneous and firmly stuck to the alumina substrate. The SEM images of Fig. 3a and c show a continuous crack-free layer of the sensing material forming a network with a thickness of ∼12–13 ␮m, highly desirable to facilitate the conduction of the water molecules along the surface. A very fine and ordered microstructure with pores in the nanometric range can be observed in the high magnification SEM image (Fig. 3b). As expected from the previous studies, the thermally stabilized mesoporous CGO oxides maintain the microstructure after the

L. Almar et al. / Sensors and Actuators B 216 (2015) 41–48

43

Fig. 2. Scheme of the alumina substrate with interdigital electrodes (left) and the sensor after deposition of the ceramic oxide mesoporous layer (right).

Fig. 3. SEM images of the CGO mesoporous humidity sensor after the electrochemical measurements.

attachment to the alumina substrate at high temperatures (see Fig. S2 for a detailed structural characterization) [14]. 3.2. Humidity sensing capabilities of a mesoporous CGO film The response to humidity of the fabricated films is evaluated. Sinusoidal excitation signals of different frequencies are imposed and the impedance is measured. Fig. 4 shows the tendency of the impedance at different frequencies for several humidity levels at a controllable fixed room temperature (T = 30 ◦ C). In the low frequency region (f = 1–10 Hz) the device presents the highest sensitivity, 10 Hz is a good tradeoff between measurement velocity and selectivity, and will be the frequency used in the following. A detailed analysis of the impedance response of the sensors is presented in the next section. There, the choice of the low frequency range for on line monitoring of the sensor response is justified. The impedance does not change at higher frequencies (f = 104 –105 Hz) where the response of the sensor becomes independent of the humidity. This indicates that, in this higher range of frequencies, it is short-circuit by the parallel resistance. The relative resistance (R0 /RRH ) versus the relative humidity (RH) at room temperature (100% RH = 3% H2 O) is presented in Fig. 5 for the here-fabricated mesoporous CGO humidity sensor and compared to different cerium oxides based humidity sensors extracted from the literature [18–20]. In this figure, R0 indicates the resistance of the sensors at 25% RH and R(RH) the resistance values

at the corresponding humidity level. Dense bulk cerium oxides show almost no change in resistance on the whole range of relative humidities, while for nanostructured materials sensitivity starts to become significant. The here-proposed CGO mesoporous material sensor shows an exponential decrease of the logarithm of the resistance with the increase of humidity. This behavior has been

Fig. 4. Impedance versus % H2 O curves of the CGO humidity sensor at different frequencies (T = 30 ◦ C).

44

L. Almar et al. / Sensors and Actuators B 216 (2015) 41–48

Fig. 5. Relative resistance (R0 /RRH ) versus relative humidity for different ceria based humidity sensors at room temperature.

reported as arising from an ionic type humidity sensing mechanism [3,22,23]. Only complex nanostructured configurations like CeO2 nanowires present a sensitivity comparable to the here presented material. In order to be reliable, a sensor must present low drift baseline and low hysteresis. It reflects the capability of the system for measuring the same signal when exposed to equal conditions, independently of its previous sensing history. Fig. 6 presents the impedance of the layer when changing the humidity levels from dry to humid (3% H2 O) and then measuring back to the initial dry level. The adsorption and desorption curves practically overlap showing a hysteresis less than 0.1% H2 O, thus making remarkable the reliability and the highly reversible characteristics of this type of mesoporous ceramic sensors. The response and recovery of the CGO mesoporous humidity sensors at a frequency of 10 Hz (T = 30 ◦ C) is shown in Fig. 7a. No drift in the dynamic response was observed during a test performed for 50 min (13 cycles). Raising and recovery signals present an exponential behavior. Time constants, defined as the time that a sensor needs to reach 63.2% value of the steady state, are 1.3 and 19.1 s, respectively. Another sensor specification is its response time, which is defined as the time that a sensor needs to reach 90% of the steady state value after a step change. In our case, the device invests 2.9/44.0 s to react to an increase/reduction of humidity. This is a quick response–recovery characteristic to humidity changes, compared to the generally reported values. For instance, Pokhrel

Fig. 7. (a) Dynamic response of the mesoporous CGO humidity sensor at 10 Hz for 13 cycles (T = 30 ◦ C). The inset figure shows one cycle of recovery and response where the red line corresponds to the exponential fitting. (b) Response and recovery times of several humidity sensors from literature. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

et al. showed a Cr2 O3 –WO3 composite sensor with a response and recovery of 25 min and 7 min, respectively [24]. Rezlescu et al. made a Sn4+ and/or Mo6+ substituted Mg ferrites sensors with a response time between 3 and 5 min [25]. Zhang et al. presented a Ba-doped CeO2 NWs sensor with a response and recovery time of 3 min [19]. Hu et al. showed a sensor of Mn-doped CeO2 nanorods with a response of 2 min and a recovery of 3 min [20]. Hao et al. presented a CGO nano-sized based sensor with a response of 40 s and a recovery of 210 s [21]. The most comparable result was reported by Fu et al. who fabricated a fast humidity sensor based on CeO2 nanowires (NWs) with a response and recovery both about 3 s [18] (Fig. 7b). Apart from the material chosen, the origin of our quick response and recovery could be attributed to the ordered and interconnected channels characteristic of this kind of 3D mesoporous structures, helping the water molecules to diffuse easily through the pores. 3.3. Mechanism of humidity sensing

Fig. 6. Hysteresis of mesoporous CGO humidity sensor at 10 Hz (T = 30 ◦ C).

Different conduction mechanisms have been identified analyzing the impedance spectra of porous ceramic humidity sensors. The mechanism of conduction is widely reported to change from low water concentrations to high water concentrations, where the corresponding impedance spectra change from a line usually fitted as an equivalent circuit formed by a constant phase element,

L. Almar et al. / Sensors and Actuators B 216 (2015) 41–48

45

Fig. 8. (a) Equivalent electrical circuit used to fit the impedance spectra. Impedance spectra of the CGO mesoporous sensor at 30 ◦ C for (b) 0.08, (c) 1.2 and (d) 2.4% H2 O. The cross points indicate the experimental data and the line the corresponding fitted curve.

to a semicircle (fitted by a resistance and capacitance in parallel) and a Warburg line at lower frequencies. The impedance spectra in those works are usually collected in a frequency range from ∼50 Hz to 100 kHz and applying an AC voltage of 1 V. The humidity sensing mechanism has been extensively studied and even though the equivalent circuits used may differ from one author to another, the explained conduction mechanism is the same [6,8,11,26]. In the present work, impedance spectroscopy measurements were performed in a wide frequency range (1 MHz to 0.1 Hz) at an AC amplitude of 300 mV. Fig. 8 shows the experimental data (cross points) obtained from the impedance measurements for 0.08, 1.2 and 2.4% of H2 O (T = 30 ◦ C). In all the samples, the higher frequency range of the impedance spectra was fitted with an equivalent circuit formed by a resistance (Rfilm ) in parallel with a capacitance (C) and a constant phase element (CPE (Q, n)) in series (Fig. 8a). The curves arising from the fitted equivalent circuit are represented as a solid red line in Fig. 8. Rfilm represents the protonic conduction on the CGO surface while C is a capacitance intrinsic of the device. The decrease of the module of the impedance with the increase of humidity observed reflects the enhancement of proton conduction. At lower frequencies, diffusion processes appeared to be dominant. When humidity is introduced in the atmosphere, a Warburg line (fitted in the circuit as a CPE with a fixed exponent n = 0.5) appears representing the migration of electro-active species in the adsorbed layer toward the electrode (mass-transfer process). The inset of Fig. 8b shows the impedance spectra in the commonly collected frequency range from 10 Hz to 100 kHz, usually fitted by a constant phase element (infinite resistance) for dry atmospheres. Nevertheless, when extending the frequency range, what was considered a Warburg like element results in a typical RC arc.

Now, the use of 10 Hz frequency in the dynamic tests is further justified at the light of the impedance arcs. As it is observed in the figures, the module of the impedance at this frequency is changing significantly. Fig. 9 shows the evolution with humidity of the parameters obtained from the fitting of the equivalent circuit to the impedance spectra. The resistance of the film decreases around four orders of magnitude as the humidity increases and the capacitance of the device is constant as it does not depend on the sensing mechanism. The results are consistent with the Grotthuss conduction mechanism [2]. The logarithm of the resistance of the film changes exponentially as water molecules are adsorbed gradually, from

Fig. 9. Resistance and capacitance of the CGO film obtained from the impedance spectra analysis of the sensor at different humidity levels (T = 30 ◦ C).

46

L. Almar et al. / Sensors and Actuators B 216 (2015) 41–48

Fig. 11. Response and recovery of the CGO and CGO-F humidity sensors at 10 Hz and T = 30 ◦ C.

temperature). This would also explain the slower response and faster recovery time as the mobility of the ions increases. This high mobility promotes a higher adsorption–desorption exchange and explains why introducing humidity into the system results in a slower response rate, while recovering the initial dry conditions is faster when the temperature is higher. 3.5. Effect of thickness of the sensing layer

Fig. 10. Response and recovery times of the mesoporous CGO humidity sensor at 10 Hz and at different temperatures (T = 60, 70, 80 and 90 ◦ C).

the very first monolayer to the physical adsorption of water molecules. Even at low humidity levels, it is possible to observe the hopping of protons in the chemisorbed monolayer of water. As the humidity increases, the continuous adsorption of water molecules (physisorbed) likely formed multilayers showing a liquid-like behavior (Grotthuss mechanism). The cerium ions lead to a strong electric field around the surface increasing the ionization of the water molecules and affecting the deeper physisorbed water [18]. On the other hand the oxygen vacancies, created due to the gadolinium-doping, have been recently studied as active sites for the dissociation of water molecules creating two hydroxyl groups per vacancy [27]. Regarding the humidity sensing mechanism, it is clear that the main benefit would be for large surface to volume ratio materials. The here synthesized mesoporous material may play a key role in the conduction, as the interconnected network with double gyroid structure could ease the water molecules pass easily through the regular pores. This can also be in the basis of the previously observed quick response and recovery. 3.4. Effect of temperature The dynamic response and recovery curves were measured at different temperatures (T = 60, 70, 80 and 90 ◦ C) once again fixing the frequency at 10 Hz. The results are shown in Fig. 10a and b. The amplitude of the response (R) decreases when operation temperature increases. This effect is consistent with the above explained Grotthuss mechanism, as the chemisorbed layer becomes thinner when water molecules hold higher kinetic energy (at higher

The influence of the thickness of the sensing layer was also investigated. With this purpose, a thinner CGO mesoporous layer (thickness of ∼4 ␮m) was deposited and measured, denoted as CGO-F (CGO-Fine layer). Fig. 11 shows a comparison of the response and the recovery time of the above presented CGO (layer thickness ∼12–13 ␮m) and CGO-F (layer thickness ∼4 ␮m) mesoporous humidity sensors. The thickness of the mesoporous layer has influence in both the sensitivity and the response/recovery. The explanation of the different sensitivity is also consistent according to the conduction mechanism and the number of adsorption sites in each sample. The higher impedance of CGO-F under dry conditions is explained by the different thicknesses of the films, the sensing layer of CGO-F is around three times thinner than the layer of the CGO sensor and therefore accordingly the measured impedance value (the resistance follows the equation R = ·l/A, (A = t·H) being  the resistivity of the material, l the length between contact electrodes, A the cross-sectional area, t the thickness and H the length of the of the sensing film). On the other hand, higher thickness indicates more available adsorption sites and therefore, the amplitude of the response (R) for the thicker layer sensor CGO is higher. Moreover, the measured impedance values are in agreement with the work presented by Shirpour et al. about the proton conductivity of pure and gadolinium doped nanocrystalline CeO2 [28]. The response and recovery time of CGO-F mesoporous sensor are calculated to be and 12.8 s, compared to the values of 2.9 and 44.0 s of the thicker layer sensor. The thicker film shows a higher response rate related with the larger amount of pores and higher number of adsorption sites. On the other hand, the thinner film presents a higher recovery rate, related with the shorter path through which the H3 O+ ions have to travel across the film. 3.6. Effect of pore size A big advantage of mesoporous materials is the possibility of tuning their pore size by modifying synthesis parameters. In particular, the pore size of the replicas of the KIT-6 silica template can

L. Almar et al. / Sensors and Actuators B 216 (2015) 41–48

47

thinner film sensor studied presents a lower sensitivity and response rate, although a higher recovery rate. Pore sizes of ∼11 nm showed a faster response and recovery, while pore sizes of ∼4 nm seem to limit the diffusion of the water molecules. The CGO mesoporous sensors fulfilled all the required characteristics of high performance humidity sensors. Moreover, the synthesis route can be extrapolated to fabricate any other metal oxide.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.snb.2015.04.018

References Fig. 12. Response and recovery of the CGO and CGO-D mesoporous humidity sensors at f = 10 Hz and T = 30 ◦ C.

be modified by filling one or two channels of its interpenetrated network. To study the influence of the pore size in the humidity sensing properties, mesoporous CGO was synthesized by filling only one of the two sets of interpenetrated channels and the fabricated sensor is denoted as CGO-D. The micro-structural characterization of the CGO-D synthesized powder is presented in the Supplementary Information. Fig. 12 shows a comparison of the response and recovery time between the CGO and CGO-D mesoporous humidity sensors. The response and recovery time of the CGO-D mesoporous sensor were calculated to be 2.3 and 35.8 s. Both response and recovery times are faster compared with the CGO sensor. The bigger pore size (over ∼10 nm) of the CGO-D seems to have a positive influence in the dynamic response of the humidity sensors. The smaller pores of the sensor CGO (∼4 nm) can limit the diffusion of the water molecules, reducing the rates of the adsorption–desorption processes. Differences shown in the impedance under dry conditions could be explained by the non-normalized thicknesses of both sensing layers. However the amplitude of the response (R) is very similar for both sensors attributed to the similar surface area (BET) (see Supplementary Information). 4. Conclusions Mechanically robust mesoporous gadolinium-doped ceria (CGO) humidity sensors were fabricated by the attachment of the sensing layer at high temperatures (900 ◦ C) keeping the microstructure intact. The layers are stable to temperature cycles, the self-cleaning required in commercial ceramic sensors at high temperature could be performed, if required, without detriment of the sensing properties. The resistance of the sensing film changes around four orders of magnitude within the measured humidity levels (0–2.4% H2 O) at the selected optimum frequency (10 Hz) and the response and recovery times were 2.9 and 44.0 s, respectively. No drift was measurable during the 13 cycles of the dynamic response experiment performed. Maximum hysteresis was only ∼0.1% indicating very high reliability. Impedance spectra were analyzed and the parameters obtained from the fitting of the equivalent circuits verified an ionic type of conduction mechanism. Moreover, the influence of temperature, thickness of the sensing layer and pore size were studied in order to check the consistency of the sensing mechanism and explore the possibility of adjusting sensing parameters to different applications. Higher temperatures imply higher mobility of ions and therefore, lower sensitivity and response and higher recovery. The

[1] B.M. Kulwicki, Humidity sensors, J. Am. Ceram. Soc. 74 (1991) 697–708. [2] G. Korotcenkov, Handbook of Gas Sensor Materials 1, 2013, pp. 389–404 (Chapter 18). [3] E. Traversa, Ceramic sensors for humidity detection: the state-of-the-art and future developments, Sens. Actuators B 23 (1995) 135–156. [4] B.J. Melde, B.J. Johnson, P.T. Charles, Mesoporous silicate materials in sensing, Sensors 8 (2008) 5202–5228. [5] A. Bearzotti, J.M. Bertolo, P. Innocenzi, P. Falcaro, E. Traversa, Humidity sensors based on mesoporous silica thin films synthesised by block copolymers, J. Eur. Ceram. Soc. 24 (2004) 1969–1972. [6] X. Liu, R. Wang, T. Zhang, Y. He, J. Tu, X. Li, Synthesis and characterization of mesoporous indium oxide for humidity-sensing applications, Sens. Actuators B 150 (2010) 442–448. [7] J. Zhao, Y. Liu, X. Li, G. Lu, L. You, X. Liang, F. Liu, T. Zhang, Y. Du, Highly sensitive humidity sensor based on high surface area mesoporous LaFeO3 prepared by a nanocasting route, Sens. Actuators B 181 (2013) 802–809. [8] W. Zhang, R. Wang, Q. Zhang, J. Li, Humidity sensitive properties of K-doped mesoporous silica SBA-15, J. Phys. Chem. Solids 73 (2012) 517–522. [9] W. Geng, R. Wang, X. Li, Y. Zou, T. Zhang, J. Tu, Y. He, N. Li, Humidity sensitive property of Li-doped mesoporous silica SBA-15, Sens. Actuators B 127 (2007) 323–329. [10] W.C. Geng, L.B. Duan, Q.Y. Zhang, Humidity sensing property of Al-doped mesoporous silica, Appl. Mech. Mater. 249–250 (2013) 992–997. [11] W. Geng, Q. Yuan, X. Jiang, J. Tu, L. Duan, J. Gu, Q. Zhang, Humidity sensing mechanism of mesoporous MgO/KCl–SiO2 composites analyzed by complex impedance spectra and bode diagrams, Sens. Actuators B 174 (2012) 513–520. [12] C.T. Wang, C.L. Wu, Electrical sensing properties of silica aerogel thin films to humidity, Thin Solid Films 496 (2006) 658–664. [13] Q. Yuan, W. Geng, N. Li, J. Tu, R. Wang, T. Zhang, X. Li, Study on humidity sensitive property of K2 CO3 -SBA-15 composites, Appl. Surf. Sci. 256 (2009) 280–283. [14] L. Almar, T. Andreu, A. Morata, M. Torrell, L. Yedra, S. Estradé, F. Peiró, A. Tarancón, High-surface-area ordered mesoporous oxides for continuous operation in high temperature energy applications, J. Mater. Chem. A. 2 (2014) 3134–3141. [15] J.L.M. Rupp, A. Infortuna, L.J. Gauckler, Microstrain and self-limited grain growth in nanocrystalline ceria ceramics, Acta Mater. 54 (2006) 1721–1730. [16] J.L.M. Rupp, B. Scherrer, A.S. Harvey, L.J. Gauckler, Crystallization and grain growth kinetics for precipitation-based ceramics: a case study on amorphous ceria thin films from spray pyrolysis, Adv. Funct. Mater. 19 (2009) 2790–2799. [17] J.L.M. Rupp, B. Scherrer, L.J. Gauckler, Engineering disorder in precipitationbased nano-scaled metal oxide thin films, Phys. Chem. Chem. Phys. 12 (2010) 11114–11124. [18] X.Q. Fu, C. Wang, H.C. Yu, Y.G. Wang, T.H. Wang, Fast humidity sensors based on CeO2 nanowires, Nanotechnology 18 (2007) 145503, 4 pp. [19] Z. Zhang, C. Hu, Y. Xiong, R. Yang, Z.L. Wang, Synthesis of Ba-doped CeO2 nanowires and their application as humidity sensors, Nanotechnology 18 (2007) 465504, 5 pp. [20] C.H. Hu, C.H. Xia, F. Wang, M. Zhou, P.F. Yin, X.Y. Han, Synthesis of Mn-doped CeO2 nanorods and their application as humidity sensors, Bull. Mater. Sci. 34 (2011) 1033–1037. [21] Y.W. Hao, J. Wang, Z.Y. Chen, Investigation on preparation of nano-size Gd0.2 Ce0.8 O2 − ␦ material and its humidity sensing properties, J. Mater. Sci. 45 (2009) 1361–1365. [22] T. Nenov, S. Yordanov, Ceramic sensor device materials, Sens. Actuators B 8 (1992) 117–122. [23] B.M. Kulwicki, Ceramic sensors and transducers, J. Phys. Chem. Solids 45 (1984) 1015–1031. [24] S. Pokhrel, K.S. Nagaraja, Electrical and humidity sensing properties of chromium(III) oxide–tungsten(VI) oxide composites, Sens. Actuators B 92 (2003) 144–150. [25] N. Rezlescu, C. Doroftei, E. Rezlescu, P.D. Popa, Structure and humidity sensitive electrical properties of the Sn4+ and/or Mo6+ substituted Mg ferrite, Sens. Actuators B 115 (2006) 589–595.

48

L. Almar et al. / Sensors and Actuators B 216 (2015) 41–48

[26] Z. Wang, L. Shi, F. Wu, S. Yuan, Y. Zhao, M. Zhang, The sol–gel template synthesis of porous TiO2 for a high performance humidity sensor, Nanotechnology 22 (2011) 275502, 9 pp. [27] R. Schaub, P. Thostrup, N. Lopez, E. Laegsgaard, I. Stensgaar, J.K. Nørskov, F. Besenbacher, Oxygen vacancies as active sites for water dissociation on rutile TiO2 (1 1 0), Phys. Rev. Lett. 87 (2013) 2661041–2661044. [28] M. Shirpour, G. Gregori, R. Merkle, J. Maier, On the proton conductivity in pure and gadolinium doped nanocrystalline cerium oxide, Phys. Chem. Chem. Phys. 13 (2011) 937–940.

Biographies Dr. Laura Almar obtained her degree in chemical engineering in 2009 in the Universitat Politècnica de Valencia (UPV). She received her master degree in nanoscience and nanotechnology at the University of Barcelona (UB), where she developed her PhD in the fabrication of mesoporous materials and their implementation in solid oxide fuel cells and gas sensors. Dr. Albert Tarancón is head of the nanoionics and fuel cells group of the advanced materials for energy department at the Catalonia Institute for Energy Research (IREC). His research interest is primarily concerned with materials for alternative energy technologies and their applicability in powering portable devices. He is deeply involved in the electrical characterization of materials for intermediate temperature solid oxide fuel cells, solid oxide electrolyzers, thermoelectric generators and Li-ion batteries. A particular field of interest is the integration of power generators in silicon technology.

Dr. Teresa Andreu received her degree in chemistry in 1999 and the PhD in 2004 in material science at the University of Barcelona. From 2004 to 2006, she worked in the R&D Department of MacDermid Inc. involved in plating on plastics and electroless deposition of nickel. In 2007, she joined the electronics department of the University of Barcelona, mainly focused on the synthesis of metal oxides using nanotemplates. Since 2009, she is a researcher of the Advanced Materials Area of IREC. Her current interests include synthesis and characterization of semiconductors and its application to chemical sensors, photocatalysis and electrocatalysis (http://www. researcherid.com/rid/F-1594-2011).

Dr. Marc Torrell received his chemistry degree in 2004 and the PhD in 2008 in material science at the University of Barcelona. In 2011 received his degree as material engineer at the Polytechnic University of Barcelona. He worked as a researcher in thermal spray at the Thermal Spray Centre of Barcelona in different fields related with materials science for energy. He had been working in Portugal in nanostructured thin films in the physics school of the University of Minho. Since 2013 is working in materials for energy at IREC (Barcelona) focused in solid oxide cells. Nowadays his main field of research is the mesoporous electrodes for solid oxide electrolyzers and fuel cells, among the solid oxide cells degradation. Dr. Yang Hu received his bachelor in materials science and engineering in Xi’an Jiaotong University (XJTU), China, where he also obtained his master in materials science. He attained his PhD in Ecole Centrale Paris (ECP) in the field of novel cathode materials for IT-SOFC applications. He is currently working as researcher in Centre for Materials Science and Nanotechnology Chemistry (SMNKJEMI), at the University of Oslo, where he works on the characterization of thin film materials for all-solidstate Li-ion batteries. Dr. Guilhem Dezanneau (France, 1973) is permanent CNRS researcher at CNRSSPMS. He obtained his PhD in 2001 from the Grenoble Institute of Technology. After a post-doc position at the Univ. of Barcelona, he obtained in 2005 his actual CNRS position. He is now head of the laboratory and manages the “Materials for hydrogen technologies” group. He (co-)supervised 7 PhD theses and obtained, as PI or coordinator for the lab., 4 multi-partner research projects representing more than 2 MD . His main activity concerns the theoretical and experimental study nanostructured functional oxides. He has a wide experience in the development of classical molecular dynamics approach for the description of ion-conduction compounds and is also expert in electrical measurements applied to ion-conducting or ferroelectric compounds, in their bulk or nanostructured form. Dr. Alex Morata is a researcher at the nanoionics and fuel cells group at the Catalonia Institute for Energy Research (IREC). He received his M.Sc. in Physics at the University of Barcelona (U.B.) in 2002. He carried out his PhD in the field of materials and film deposition for electrochemical devices. After working in the technology company Telstar as application engineer (2008–2010), he joined IREC. His current research topics are in the field of material fabrication with application in power devices and sensors (http:// www.researcherid.com/rid/C-6950-2011).