Effect of multiwalled carbon nanotube functionalization on the gas sensing properties of carbon nanotube–titanium dioxide hybrid materials

Effect of multiwalled carbon nanotube functionalization on the gas sensing properties of carbon nanotube–titanium dioxide hybrid materials

Diamond & Related Materials 21 (2012) 1–6 Contents lists available at SciVerse ScienceDirect Diamond & Related Materials journal homepage: www.elsev...

1MB Sizes 0 Downloads 7 Views

Diamond & Related Materials 21 (2012) 1–6

Contents lists available at SciVerse ScienceDirect

Diamond & Related Materials journal homepage: www.elsevier.com/locate/diamond

Effect of multiwalled carbon nanotube functionalization on the gas sensing properties of carbon nanotube–titanium dioxide hybrid materials M. Sánchez ⁎, M.E. Rincón Centro de Investigación en Energía-Universidad Nacional Autónoma de México, Privada Xochicalco S/N. Col. Centro, Temixco, MOR, 62580, México

a r t i c l e

i n f o

Article history: Received 16 June 2011 Accepted 23 September 2011 Available online 5 October 2011 Keywords: Carbon nanotubes Titanium dioxide Gas sensor Ammonia Capacitor Titania

a b s t r a c t The effect of multiwalled carbon nanotube functionalization on the sensing properties of carbon nanotube– titanium dioxide hybrid materials during ammonia exposure was investigated. Strongly adherent films were evaluated by impedance spectroscopy, finding highly sensitive and completely reversible capacitance values when the materials were exposed to ammonia vapors at room temperature. The abundance of oxygenated functional groups caused by functionalization of carbon nanotubes in strong acid solutions correlates with the formation of carbon nanotube–titania hybrids with synergistic sensing properties, such as faster adsorption/desorption cycles, lower impedance values, and high sensitivity for ammonia. X-ray diffraction and atomic force microscopy studies showed that oxygenated functional groups on the carbon surface act as nucleation points for titania growth, resulting in thinner films with smaller crystallite size for the titania phase than those obtained with untreated carbon nanotubes. The better integration between both components produced films with unique sensing properties. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Carbon nanotubes (CNTs) and nanosized titanium dioxide (TiO2) have been extensively studied for gas sensing applications [1–12], and some road maps foresee that in the short term these materials will have an even more predominant role [13,14]. As an active sensing layer, nanosized TiO2 has been used as nanowires [2,15], nanotubes [4,5,15], and nanoparticles [1,3,6,15–24], showing fast adsorption/desorption cycles and enhanced sensor response [1– 3,6,15,16,19]. In these materials, research is focused on their high resistivity and cross sensitivity. In particular, TiO2 metal doping has been used to overcome these issues, in addition to prevent coarsening of nanosized grains, and delay the anatase to rutile phase transition at high temperatures [17,18,21–24]. Similarly, CNTs have shown interesting properties for use as sensors such as high conductivity, large aspect ratio, large surface area, and high sensitivity [7–12,25–27]. However the use of CNTs as ammonia or nitrogen dioxide sensors requires long recovery times due to the strong interaction of these gases with the CNT surface [7,8,25–27]. In order to overcome the drawbacks of metallic oxides and carbon nanotubes, and looking for novel properties, studies on sensors based on CNT/TiO2 composites and CNT–TiO2 hybrid materials have been reported [28–35]. Screen printed films based on multiwalled carbon nanotubes (MWCNTs) and TiO2 were used as room temperature

⁎ Corresponding author. Tel.: + 52 5556 229836; fax: + 52 5556 229742. E-mail address: [email protected] (M. Sánchez). 0925-9635/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.diamond.2011.09.010

sensors for acetone and ammonia [30], demonstrating the importance of poorly coordinated Ti sites for defining the direction and magnitude of charge transfer between the active sensing layer and the adsorbate [31]. More recently, thin compact films of MWCNT–TiO2, prepared by sol gel and dip coating techniques were tested as resistive ammonia sensors, showing a different sensing mechanism than the one observed in screen printed films [32]. Evidently, the weak or strong interaction between MWCNTs and TiO2 depends on the preparation method, and will have an impact on the sensing properties of the compound material explaining the conflicting data found in the literature, where responsive [32] and not responsive [33] materials were obtained. To study the source of interaction in MWCNT–TiO2 in more detail, the effect of MWCNT functionalization on the gas sensing properties of MWCNT–TiO2 films was studied by impedance spectroscopy (IS). Changes in capacitance and resistance indicated that functionalization of CNTs causes a better integration between both components resulting in materials with unique sensing properties. 2. Experimental 2.1. Film preparation techniques Commercial MWCNTs (Nanostructured & Amorphous Inc., 95 wt.% CNT, outer diameter b10 nm, length: 5–15 μm) were refluxed in aqueous solutions of sulfuric and nitric acids (JTBaker) at 100 °C for 6 h to remove amorphous carbon and to promote the formation of oxygenated functional groups grafted to

2

M. Sánchez, M.E. Rincón / Diamond & Related Materials 21 (2012) 1–6

a

b

Fig. 1. Sensor configuration: (a) top view; (b) side view.

A(200)

R(210)

A(004)

A(101)

the CNT surface. The functionalization was performed varying the concentration of nitric acid (2.5, 7.5 and 12.5 M), and using a fixed concentration of sulfuric acid (0.5 M). With exception of non-treated CNTs, all the others nanotubes were subsequently refluxed for 5 h in 3 M hydrochloric acid (JT-Baker) solution at 100 °C. Titania films were deposited from a sol gel bath containing 82 mL of 2-propanol (Sigma-Aldrich), 0.5 mL of concentrated hydrochloric acid, and 8 mL of titanium isopropoxide (SigmaAldrich). Films of MWCNT–TiO2 were prepared in a trilayer system on indium tin oxide substrates (ITO, Delta Technologies, Rs = 13–15 Ω, 1 in. × 1.5 in.). First, a titania layer of 10 immersions was deposited by dip-coating, at 30 mm/min dipping/withdrawing speed. Films were annealed in air at 400 °C for 5 min after each immersion and at 500 °C in air after the 10th immersion for 1 h. The second layer (CNT) was deposited by spray of aqueous CNT inks prepared with triton X-100 (CNT/triton ratio of 1/6 wt.). The two layer system was annealed in air for 20 min at 300 °C to dry and to burn out impurities. Finally, a second titania layer of 10 immersions was deposited by dipcoating and the whole system was annealed at 400 °C for 30 min. The hybrid materials containing functionalized carbon nanotubes were labeled as CNT-2.5, CNT-7.5, and CNT-12.5 according to the concentration of nitric acid used in the functionalization, while those

Intensity (a.u.)

4 3 2

2.2. Characterization An Alpha Step profilometer (Tencor Instruments) was used to measure film thickness. Fourier Transform Infrared (FTIR) spectra were obtained in a Bruker FTIR-5000 spectrophotometer in the range from 400 to 4000 cm −1. X-ray diffraction (XRD) studies were carried out in a Rigaku Dmax 2200 diffractometer with CuKα radiation (λ = 1.5405 Å), using the Bragg–Brentano configuration in the 2θ range 10–70°; the JADE (Materials Data, Inc.) software was employed for the analysis of chemical composition and the Debye– Scherrer equation [36] was used for crystallite size estimation. Surface topography was studied by atomic force microscopy (AFM) in a Nanosurf Easyscan unit (Nanosurf AG, Switzerland), using the software Gwyddion (Czech Metrology Institute) for image processing. 2.3. Sensing experiments The measurement system for the gas sensing experiments was described elsewhere [30,31]. Basically, the sensor configuration (see Fig. 1) consists of a two electrode cell with the counter electrode connected to silver lines printed on the MWCNT–TiO2 film surface, and the working electrode connected to the ITO substrate. Impedance spectroscopy was performed with an Autolab PGSTAT302N potentiostat/galvanostat unit (Eco Chemie), at room temperature (27 °C), in a closed cell kept in the dark. Measurements were done at open circuit potential, with a perturbation potential of 5 mV, in the range from 1 MHz to 10 Hz. Frequency sweeps were taken continuously so that the time needed to apply the same frequency between two consecutive sweeps was ~ 1 min. Dry air (5 L/min) was used to establish the baseline, followed by measurements in ammonia/nitrogen (N2)

Table 1 Crystallite size computed from XRD characterization using Debye–Scherrer equation.

1 20

containing non-treated carbon nanotubes were labeled as CNT-P (from pristine).

25

30

35

40

45

50

2θ (o) Fig. 2. XRD diffractograms of MWCNT–TiO2 films, 1: CNT-P; 2: CNT-2.5; 3: CNT-7.5; 4: CNT-12.5. A: anatase, R: rutile.

Film

Crystallite size (nm)

CNT-P CNT-2.5 CNT-7.5

59 47 23

M. Sánchez, M.E. Rincón / Diamond & Related Materials 21 (2012) 1–6

3

a

Fig. 3. FTIR results of MWCNTs as a function of HNO3 concentration, 1: nontreated; 2: 2.5 M; 3: 7.5 M; 4: 12.5 M.

b

(0.01–0.4 vol.% ammonia in 150 mL/min N2). After 10–15 min, N2 flow was stopped and air was injected again to recover the baseline. Complex non-linear least squares fitting and pseudocapacitance estimation [37] were done using the Zsimpwin software (Princeton Applied Research). Sensor response (S) was calculated from Eq. (1):

Sð%Þ ¼

h  i Y 0;NH3 −Y 0;air =Y 0;air  100

ð1Þ

where Y0,air and Y0,NH3 are base admittances in air and ammonia, respectively. 3. Results and discussion

c

3.1. Structural characterization Thickness of MWCNT–TiO2 films shows a strong correlation with the degree of functionalization of CNTs, yielding CNT-P and CNT12.5 as the thickest (~ 1.5 μm) and thinnest (~0.5 μm) films, respectively, whereas the thickness of carbon-free TiO2 films was ~0.3 μm. XRD patterns of the films (Fig. 2) show anatase (JCPDS 21–1272) as the principal crystalline phase, and a sizable decrease in intensity and broadening of its main peaks [(101), (004,) and (200) crystalline planes] as the acid treatment becomes more intense. Apparently, the role of functional groups attached to carbon nanotubes is to facilitate the nucleation of titania crystallites, increasing its number and decreasing its size, as it is shown in Table 1. FTIR spectra of MWCNTs with and without acid treatment are presented in Fig. 3. Absorptions at 3744 and 3620 cm −1 correspond to the stretching of aromatic C\H and O\H bonds respectively, while 1741 cm −1 adsorption is related to the stretching of C_O bonded to a carboxylic group, 1620 and 677 cm −1 correspond to vibrations and combinations of free water, 1514 cm −1 to the stretching of single C\C coupled with a double bond (C_C), and 1395 cm −1 to the bending of carboxyl groups [38]. It is clear from this figure that as the intensity of the acid treatment increases so does the absorption signals related to oxygenated functional groups. The effect of CNT functionalization on the topography of the three layer films is evident from the sequence of AFM images shown in Fig. 4, where film height (vertical scale) decreases as the acid treatment intensifies. The average height and roughness of these films (Fig. 5) show similar tendencies than those observed for film thickness and crystal size, indicating that the effect of functionalization on the overall film microstructure is a combination of improved CNT dispersion and generation of more TiO2 nucleation sites.

d

Fig. 4. AFM images of the hybrid films: (a) CNT-P; (b) CNT-2.5; (c) CNT-7.5; (d) CNT-12.5.

4

M. Sánchez, M.E. Rincón / Diamond & Related Materials 21 (2012) 1–6

a 3

2

-Zi (kohms)

Average height (μm)

a

1

NH3 0.4% Time

1

Air

0

0 ITO

2

TiO

0

T-P T-2.5 T-7.5 -12.5 T CN CN CN

1

2

3

4

3

4

Zr (kohms)

CN

b

b 0.6

4

3 0.4

-Zi (kohms)

Roughness (μm)

2

0.2

2

1

0

0 ITO

2

TiO

T-P T-2.5 T-7.5 -12.5 T CN CN CN

0

1

Fig. 5. Statistical analysis of AFM images: (a) average height; (b) roughness.

2

Zr (kohms)

CN

c

2

IS curves of CNT-P and CNT-12.5 films are shown in Figs. 6 and 7 respectively. Figs. 6(a) and 7(a) correspond to anisotropic Cole–Cole graphs used to magnify the difference between the curve in air and those obtained during the transition from air to 0.4 vol.% of ammonia. For CNT-P, the impedance in air is described by one semicircle at higher frequencies, while the transition from air to ammonia requires two semicircles. During the transition, the high frequency semicircle decreases in size up to negligible values once a steady state is reached (i.e. when air is excluded from the system). Isotropic Cole–Cole [Fig. 6 (b)] and Bode [Fig. 6(c)] graphs depicting the response in air and 0.4 vol.% ammonia/N2, confirmed a substantial difference in the characteristic frequency in both environments, appearing in the kHz range in air, and in the Hz range in ammonia. The good reversibility of CNT-P can be appreciated from the similitude of the air-baseline before and after ammonia sensing. For CNT-12.5, Fig. 7(a) shows distorted semicircles with the maximum at almost the same frequency in both air and ammonia, and no transition regime (i.e. two semicircles evolving into one). The impedance values are ten times lower for CNT-12.5 than for CNT-P, and it cannot be accounted for by thickness differences. The high reversibility of this sensor can be appreciated in the Cole–Cole plot of Fig. 7(b), where the responses in air before and after ammonia adsorption are overlapped. Moreover, Fig. 7(c) confirms the similarity in the characteristic frequencies in air and in ammonia (~10 kHz). The sensible difference in characteristic frequencies of CNT-P and CNT-12.5 is related to the times required for

Zi (kohms)

3.2. Impedance spectroscopy

1

0 101

102

103

104

105

106

Frequency (Hz) Fig. 6. IS results of CNT-P films sensing 0.4 vol.% of ammonia: (a) transition from air to ammonia/N2, every curve was taken with a difference of 1 min; (b–c) response in air and at 0.4 vol.% of ammonia, (b) Cole–Cole graph, (c) Bode graph. White square: measurement in air before exposure to ammonia, triangle: measurement in 0.4 vol.% of ammonia, black square: measurement in air after exposure to ammonia.

adsorption/desorption cycles (desorption transition not shown), ~8 min for CNT-P and ~ 2 min for CNT-12.5 [Figs. 6(a) and 7(a)]. For CNT-12.5, the decrease in impedance when sensing an electron donor molecule such as ammonia indicates the n-type conductivity of the material in spite of the use of CNTs (p-type conductivity) and in contrast with the behavior of CNT-P.

M. Sánchez, M.E. Rincón / Diamond & Related Materials 21 (2012) 1–6

5

a 0.3

-Zi (kohms)

Air Time

0.2

Fig. 8. Electrical equivalent circuit used to fit the experimental impedance data.

0.1

NH3 0.4%

0 0

0.2

0.4

0.6

0.8

Zr (kohms)

τ ¼ RC

b 0.8

-Zi (kohms)

0.6

0.4

0.2

0 0

0.2

0.4

0.6

0.8

Zr (kohms)

c 0.2

Zi (kohms)

For CNT-P, the higher frequency sub-circuit fits the response in air, while the lower frequency sub-circuit fits the response in ammonia/ N2, and both sub-circuits are required during the transition from air to ammonia/N2. In contrast, CNT-12.5 requires only the high frequency sub-circuit to fit the experimental results in both air and ammonia/N2. The values of the circuit elements are presented in Table 2, with the relaxation times (τ) calculated from the Eq. (2).

0.1

ð2Þ

Values of capacitance in polycrystalline materials are orders of magnitude lower in the grain bulk (10 −12 F), than in grain boundaries (10 −11 to 10 −8 F), or at the film/electrode interface (10 −7 to 10 −5 F) [39,40]. Therefore, the pseudocapacitance values presented in Table 2 suggest that the high frequency sub-circuit of CNT-P and CNT-12.5 could be assigned to events taking place at the grain boundaries of the films. Functionalization of carbon nanotubes promotes the formation of more integrated hybrid materials due to the formation of ester-like linkages between TiO2 and CNTs [41], and the complete coverage of CNT by a thin titania film. This intimate contact causes a unique response (semicircle) in both, air and ammonia, in contrast to the response of more segregated systems like CNT-P. Additionally, the decrease of R1 (the high frequency sub-circuit resistor) from ~10 3 Ω for CNT-P to ~10 2 Ω for CNT-12.5 confirms the superior dispersion and doping of CNTs caused by the acid treatment. For CNT-P sensors, CNTs are the active surface in the segregated system and undergo dedoping when exposed to ammonia [32], given that R2 (the low frequency sub-circuit resistor) increases from a no measurable value up to ~10 kΩ. As the aggressiveness of the acid treatment increases the individual characteristics of TiO2 and CNTs are gradually lost. The intimate contact between CNTs and titania nanoparticles reduces the size and contribution of the low frequency sub-circuit (i.e. the straight adsorption on CNTs). The relaxation times presented in Table 2 indicate that the introduction of functional groups removes the slow processes in both air and ammonia sensing. The dynamical behavior of the best sensor (CNT-12.5) during ammonia adsorption/desorption cycles is presented in Fig. 9. Here, the base admittance (Y0,1) of the constant phase element Q1 is plotted versus time. Fast adsorption and complete desorption are observed for this sensor at room temperature, without any additional aid to

0 101

102

103

104

105

106

Frequency (Hz)

Table 2 Values of the EQC elements for the best fit of the experimental data. Parameter

Fig. 7. IS results of CNT-12.5 films sensing 0.4 vol.% of ammonia: (a) transition from air to ammonia/N2, every curve was taken with a difference of 1 min; (b–c) response in air and at 0.4 vol.% of ammonia, (b) Cole–Cole graph, (c) Bode graph. White square: measurement in air before exposure to ammonia, triangle: measurement in 0.4 vol.% of ammonia, black square: measurement in air after exposure to ammonia.

Complex non-linear least squares fitting of the experimental results was accomplished by using the equivalent circuit (EQC) presented in Fig. 8, a resistance R0 in series with two sub-circuits RQ (i.e. a resistor R in parallel with a constant phase element Q).

R0 (Ω) R1 (Ω) Y0,1 (Ssn1) n1 C1 (F) τ1 (μs) R2 (Ω) Y0,2 (Ssn2) n2 C2 (F) τ2 (s)

CNT-P

CNT-12.5

Air

0.4 vol.% NH3

Air

0.4 vol.% NH3

96 3150 1.0 × 10−7 0.82 1.7 × 10−8 54

128

75 582 6.6 × 10−8 0.89 1.9 × 10−8 11

67 617 1.5 × 10−6 0.70 7.5 × 10−8 46

9610 2.9 × 10−5 0.62 1.3 × 10−5 0.10

6

M. Sánchez, M.E. Rincón / Diamond & Related Materials 21 (2012) 1–6

Acknowledgments

0.4

0.2

NH3 (%)

Y0 (Ssn)

10-6

10-7

References

0 0

10

20

30

40

50

60

70

Time (min) Fig. 9. Dynamical behavior of the base admittance of CNT-12.5 during ammonia sensing at room temperature. Symbols correspond to the sensor response and the dashed line to variations in NH3 concentration.

promote ammonia desorption (i.e., heating, high flow of inert gas, ultraviolet light). The sensitivity of CNT-12.5, defined as the variation of S with respect to ammonia concentration, was evaluated in the concentration range from 0.01 to 0.3 vol.% and it was found of ~0.6 (Fig. 10). Additionally, the response (~10 3%) of this material is superior to the values of ~10 1% reported for sensors based on CNTs [7–12].

4. Conclusions It has been shown that strong functionalization of MWCNTs is a key step in the formation of highly adherent, reversible, and sensitive MWCNT–TiO2 ammonia sensors working at room temperature. The abundance of oxygen-containing functional groups on the carbon nanotube surface promotes the nucleation of titania particles and the interaction at the molecular level. Films containing functionalized CNTs show a substantial reduction on the response and recovery times, most likely due to the promotion of a uniform coverage of CNTs by the titania phase, as well as the elimination of amorphous carbon and other impurities. Those films based on untreated CNTs show the presence of slow processes, which indicate the straight adsorption of ammonia on CNTs.

2.0

S(%)=0.57[NH 3] R2=0.97

1.5

S(%) x1000

Financial support from Dirección General de Asuntos del Personal Académico-Universidad Nacional Autónoma de México (DGAPAUNAM) IN104309-3, Proyecto Universitario de Nanotecnología Ambiental (PUNTA-UNAM), and Consejo Nacional de Ciencia y Tecnología (CONACYT-México) (49100), is gratefully acknowledged, as well as the fellowship (M. Sánchez) provided by CONACYT-Mexico. We thank R. Moran Elvira, M.L. Ramon Garcia, and G. Alvarado Tenorio for technical assistance.

1.0

0.5

0.0 0

0.1

0.2

0.3

NH3 (%) Fig. 10. Sensor response of CNT-12.5 as a function of ammonia concentration.

[1] B. Karunagaran, P. Uthirakumar, S.J. Chung, S. Velumani, E.K. Suh, Mater. Charact. 58 (2007) 680–684. [2] L. Francioso, A.M. Taurino, A. Forleo, P. Siciliano, Sens. Actuators B 130 (2008) 70–76. [3] A.M. Ruiz, G. Sakai, A. Cornet, K. Shimanoe, J.R. Morante, N. Yamazoe, Sens. Actuators B 103 (2004) 312–317. [4] O.K. Varghese, D. Gong, M. Paulose, K.G. Ong, C.A. Grimes, Sens. Actuators B 93 (2003) 338–344. [5] G.K. Mor, O.K. Varghese, M. Paulose, K.G. Ong, C.A. Grimes, Thin Solid Films 496 (2006) 42–48. [6] B. Huber, A. Brodyanski, M. Scheib, A. Orendorz, C. Ziegler, H. Gnaser, Thin Solid Films 472 (2005) 114–124. [7] L. Valentini, I. Armentano, J.M. Kenny, C. Cantalini, S. Santucci, Appl. Phys. Lett. 82 (2003) 961–963. [8] T. Ueda, M.M.H. Bhuiyan, H. Norimatsu, S. Katsuki, T. Ikegami, F. Mitsugi, Physica E 40 (2008) 2272–2277. [9] W.S. Cho, S.I. Moon, Y.D. Lee, Y.H. Lee, J.H. Park, B.K. Ju, IEEE Electron Device Lett. 26 (2005) 498–500. [10] N.D. Hoa, N.V. Quy, Y. Cho, D. Kim, Sens. Actuators B 135 (2009) 656–663. [11] Y.T. Jang, S.I. Moon, J.H. Ahn, Y.H. Lee, B.K. Ju, Sens. Actuators B 99 (2004) 118–122. [12] D.R. Kauffman, A. Star, Angew. Chem. Int. Ed. 47 (2008) 6550–6570. [13] J. Rodel, A.B.N. Kounga, M. Weissenberger-Eibl, D. Koch, A. Bierwisch, W. Rossner, M.J. Hoffman, R. Danzer, G. Schneider, J. Eur. Ceram. Soc. 29 (2009) 1549–1560. [14] J. Hodgkinson, J. Saffel, J. Luff, J. Shaw, J. Ramsden, C. Huggins, R. Bogue, R. Carline, MNT Gas Sensor Roadmap, MNT Gas Sensor Forum, London UK, 2006. [15] M.H. Seo, M. Yuasa, T. Kida, J.S. Huh, K. Shimanoe, N. Yamazoe, Sens. Actuators B 137 (2009) 513–520. [16] M. Sánchez, M.E. Rincón, R.A. Guirado-López, J. Phys. Chem. C 113 (2009) 21635–21641. [17] N. Bonini, M.C. Carotta, A. Chiorino, V. Guidi, C. Malagu, G. Martinelli, L. Paglialonga, M. Sacerdoti, Sens. Actuators B 68 (2000) 274–280. [18] M.C. Carotta, V. Guidi, C. Malagu, B. Vendemiati, A. Zanni, G. Martinelli, M. Sacerdoti, S. Licoccia, M.L.D. Vona, E. Traversa, Sens. Actuators B 108 (2005) 89–96. [19] Z. Seeley, Y.J. Choi, S. Bose, Sens. Actuators B 140 (2009) 98–103. [20] Y. Shimizu, T. Okamoto, Y. Takao, M. Egashira, J. Mol. Catal. A Chem. 155 (2000) 183–191. [21] K. Shimizu, I. Chinsei, H. Nishiyama, S. Kakimoto, S. Sugaya, W. Matsutani, A. Satsuma, Sens. Actuators B 141 (2009) 410–416. [22] A.M. Ruiz, G. Sakai, A. Cornet, K. Shimanoe, J.R. Morante, N. Yamazoe, Sens. Actuators B 93 (2003) 509–518. [23] M.C. Carotta, M. Ferroni, D. Gnani, V. Guidi, M. Merli, G. Martinelli, M.C. Casale, M. Notaro, Sens. Actuators B 58 (1999) 310–317. [24] Y. Li, W. Wlodarski, K. Galatsis, S.H. Moslih, J. Cole, S. Russo, N. Rockelmann, Sens. Actuators B 83 (2002) 160–163. [25] J. Kong, N.R. Franklin, C. Zhou, M.G. Chapline, S. Peng, K. Cho, H. Dai, Science 287 (2000) 622–625. [26] L. Valentini, F. Mercuri, I. Armentano, C. Cantalini, S. Picozzi, L. Lozzi, S. Santucci, A. Sgamelloti, J.M. Kenny, Chem. Phys. Lett. 387 (2004) 356–361. [27] H.Q. Nguyen, J.S. Hu, Sens. Actuators B 117 (2006) 426–430. [28] S. Carrara, V. Bavastrello, D. Ricci, E. Stura, C. Nicolini, Sens. Actuators B 109 (2005) 221–226. [29] N.V. Duy, N.V. Hieu, P.T. Huy, N.D. Chien, M. Thamilselvan, J. Yi, Physica E 41 (2008) 258–263. [30] M. Sánchez, R. Guirado, M.E. Rincón, J. Mater. Sci. 18 (2007) 1131–1136. [31] R.A. Guirado-López, M. Sánchez, M.E. Rincón, J. Phys. Chem. C 111 (2007) 57–65. [32] M. Sánchez, M.E. Rincón, Sens. Actuators B 140 (2009) 17–23. [33] E.H. Espinoza, R. Ionescu, B. Chambon, G. Bedis, E. Sotter, C. Bittencourt, A. Felten, J.J. Pireaux, X. Correig, E. Llobet, Sens. Actuators B 127 (2007) 137–142. [34] N.V. Hieu, N.V. Duy, P.T. Huy, N.D. Chien, Physica E 40 (2008) 2950–2958. [35] T. Ueda, K. Takahashi, F. Mitsugi, T. Ikegami, Diamond Relat. Mater. 18 (2009) 493–496. [36] B.D. Cullity, Elements of X-ray Diffraction, Addison-Wesley, Massachusetts, 1978. [37] M.E. Orazem, B. Tribollet, Electrochemical Impedance Spectroscopy, John Wiley & Sons Inc., New Jersey, 2008. [38] B.H. Stuart, Infrared Spectroscopy: Fundamentals and Applications, John Wiley & Sons Inc, 2005. [39] J.E. Bauerle, J. Phys. Chem. Solids 30 (1969) 2657–2670. [40] J.C. Ruiz-Morales, D. Marrero-López, J.T.S. Irvine, P. Núñez, Mater. Res. Bull 39 (2004) 1299–1318. [41] X. Li, J. Niu, J. Zhang, H. Li, Z. Liu, J. Phys. Chem. B 107 (2003) 2453–2458.