Electrochromic thin films of sodium intercalated vanadium(V) oxide xerogels: Chemical bath deposition and characterization

Surface & Coatings Technology 277 (2015) 308–317

Contents lists available at ScienceDirect

Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat

Electrochromic thin films of sodium intercalated vanadium(V) oxide xerogels: Chemical bath deposition and characterization Metodija Najdoski a,b,⁎, Violeta Koleva c, Sasho Stojkovikj a,b, Toni Todorovski a,1 a b c

Institute of Chemistry, Faculty of Natural Sciences and Mathematics, Ss. Cyril and Methodius University, POB 162, Arhimedova 3, 1000 Skopje, Republic of Macedonia Research Center for Environment and Materials, Macedonian Academy of Sciences and Arts, Krste Misirkov 2, 1000 Skopje, Republic of Macedonia Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, G. Bonchev Str. Bldg. 11, 1113 Sofia, Bulgaria

a r t i c l e

i n f o

Article history: Received 26 February 2015 Revised 20 July 2015 Accepted in revised form 21 July 2015 Available online 29 July 2015 Keywords: Thin films Vanadium(V) oxide xerogels Chemical synthesis Electrochromism Optical properties Electrochemical properties

a b s t r a c t An optimized chemical bath method is applied to obtain well-structured thin films with composition Na0.33V2O5·nH2O (n = 1 and 1.3). The method is based on a controlled precipitation reaction that takes place in the system of sodium metavanadate and diethyl sulfate at 85 °C. The film structure, morphology and the changes occurring during prolonged aging are examined by XRD, IR spectroscopy, TG-DTA, SEM and AFM. The electrochemical and electrochromic properties are studied by cyclic voltammetry and UV–vis spectroscopy. The as-deposited thin films are characterized with high optical transmittance varying between 40 and 70% at the 500 nm visible region in dependence on film thickness. The Na0.33V2O5·nH2O thin films exhibit stable electrochemical cycling combined with relatively high electrochromic activity. The reproducibility of the transmittance variance of 55% after 500 cycles in the electrochromic cell is a promising result for the potential application of Na0.33V2O5·nH2O thin films in electrochromic devices. © 2015 Published by Elsevier B.V.

1. Introduction Vanadium(V) oxide and derived compounds have been extensively studied due to valuable chemical and physical properties which determine a wide range of applications in catalysis, high-energy lithium batteries and a variety of electric and optical devices. The synthetic procedures at ambient conditions usually produce hydrated vanadium(V) oxides, V2O5·nH2O, known as xerogels which adopt layered structures with V2O5 layers and interstitial water molecules [1–3]. Vanadium(V) oxide xerogels like crystalline V2O5 are typical intercalation compounds with multiple valence state of vanadium which enables redox-dependent properties [4–6]. Due to the high intercalation capacity (for instance, a lithium intercalation capacity about 1.4 times larger than that of crystalline V2O5) [7] they have a great potential for applications like reversible cathodes for lithium batteries [8,9], micro-batteries [6], supercapacitors [10], electrodes [11] and humidity sensors [12]. The reversible cation intercalation/deintercalation within the xerogel framework is concomitant with reversible reduction/oxidation of V(V) to V(IV) or to a lower ⁎ Corresponding author at: Institute of Chemistry, Faculty of Natural Sciences and Mathematics, Ss. Cyril and Methodius University, POB 162, Arhimedova 3, 1000 Skopje, Republic of Macedonia. E-mail addresses: [email protected] (M. Najdoski), [email protected] (V. Koleva), [email protected] (S. Stojkovikj), [email protected] (T. Todorovski). 1 Present address: Institute for Research in Biomedicine, Parc Científic de Barcelona, 08028 Barcelona, Spain.

http://dx.doi.org/10.1016/j.surfcoat.2015.07.041 0257-8972/© 2015 Published by Elsevier B.V.

valence vanadium state giving rise to easy color changes: yellow (V(V)), blue (V(IV)), green (V(III) or a mixture of V(V) and V(IV)) and violet (V(II)) [2,6]. The multi-colored electrochromism demonstrated by V2O5·nH2O xerogels makes them very attractive since it provides the opportunity to extend the range of functions of the electrochromic materials. Vanadium(V) oxide xerogels under the form of thin films on electroconductive glass substrates have been used in electrochromic devices [13,14], electrochromic mirrors [14], “smart windows” designed for architectural purposes to control light transmittance [15–17] and controlled reflectance mirrors for vehicles [18]. The thin film properties, including V2O5·nH2O xerogels, are well known to depend essentially on its microscopic characteristics [19,20] such as structure, crystallinity and morphology, which can be governed by the deposition method and the deposition parameters (kind and concentration of the precursors, rate of deposition, temperature, pressure, etc.). Therefore, the choice of the suitable synthetic procedure is a powerful tool for control and optimization of the material properties. In this regard, we have recently developed a simple chemical bath deposition method to obtain well-defined ammonium intercalated vanadium(V) oxide xerogels with the compositions (NH4)xV2O5· 1.3H2O (x = 0.15 and 0.30) [21,22]. The method is based on the direct acidification of NH4VO3 solutions by acetic acid at different temperatures between 50 and 85 °C. Through rational selection of the deposition parameters like vanadium concentration, temperature and deposition time, (NH4)xV2O5·1.3H2O thin films exhibiting high values of the transmittance variance (ΔT) of 55% at 400 and 900 nm were designed.

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

However, the use of the same simple synthetic procedure in the case of initial NaVO3, i.e. acidification of NaVO3 solution with acetic acid at 75 °C (pH = 3), leads to the formation of unstructured amorphous thin films as we have previously established [23]. In that case a further thermal treatment at 400 °C was needed in order to obtain crystalline films which represent a two-phase mixture of sodium vanadium oxides such as NaV6O15 and Na1.1V3O7.9. Thus prepared thin films exhibited insufficiently high ΔT values of about 20% in the voltage range of ± 2 V. The present work is focused on the examination of well-structured hydrated sodium vanadium oxide thin films with electrochromic properties. Such thin films are prepared by an optimized chemical bath method that ensures one-step deposition of the thin films at low temperature. For the purpose we have applied a different synthetic approach: instead of a direct acidification, we have used here an indirect acidification through the hydrolysis of diethyl sulfate present in the chemical bath. This controlled precipitation reaction gives rise to the deposition of thin films of vanadium(V) oxide xerogel with composition of Na0.33V2O5·H2O having well-organized layered structure in the nanoscale region. These film characteristics are advantageous to achieving electrochemical stability and high ΔT value of 55% which is reproducible for 500 cycles. The changes during the film aging are studied in respect to the film structure, V(V) reduction and electrochromic effect. It is worth emphasizing that the as-prepared well-structured Na0.33V2O5·H2O composition (solid or film) is highly beneficial for further obtaining a variety of chemical compositions. Thus, once obtained it can be used to produce either lower hydrated xerogels Na0.33V2O5·nH2O (0.3 b n b 1) or a single phase of NaV6O15 by thermal treatment at an appropriate temperature. All these compositions having a layered or tunnel structure can serve as host matrices for intercalation processes which open opportunities for different applications. From this point of view the difference with the previously studied (NH4)xV2O5· 1.3H2O compositions is obvious: their thermal treatment produces the well studied V2O5.

2. Material and methods Thin film deposition is performed onto commercially available glass substrates. They are coated with a conductive, transparent thin layer of SnO2:F (FTO) with 80% optical transparency in the visible spectrum and electrical resistance of 10–20 Ω/cm2. Before deposition, the substrates were cut into pieces with the dimensions 40 mm × 25 mm × 2 mm and cleaned in the following order: with detergent, alkaline solution, 1:1 diluted hydrochloric acid, hexane, acetone and rinsed with deionized water and dried at room temperature. Commercial diethyl sulfate (Sigma-Aldrich), propylene carbonate (Sigma-Aldrich), lithium perchlorate (Sigma-Aldrich), sodium metavanadate min 98 wt.% (Carlo Erba), ethanol 96% (Alkaloid), sodium hydroxide (Merck), hydrochloric acid (Merck), hexane (Merck) and acetone (Merck) are used without further purification. 2.1. Preparation of the thin films The films are deposited from a chemical bath with optimized composition and process conditions. The stock solution for preparation of the chemical bath is obtained by heating at 65 °C of 0.50 g sodium metavanadate and 250 ml deionized water. The chemical bath solution is prepared in a 120 ml beaker by mixing 100 ml sodium metavanadate solution (0.016 M) and 0.5 ml of diethyl sulfate. The addition of diethyl sulfate results in a change of the solution color from yellow to dark orange. The cleaned substrates are vertically supported to the wall of the beaker with the non-conductive side facing the wall. The deposition system is then heated up to 85 °C (deposition temperature) with continuous stirring and the achieved temperature and stirring are maintained during the deposition time. pH of the chemical bath during the deposition is about 5.5. The beginning of the deposition reaction is

309

observed as the appearance of turbidity of the liquid phase which is due to the formation of a solid substance (precipitate). Just at this moment the substrate is removed from the chemical bath and quickly, but carefully, is wiped with cotton soaked with ethanol (96%). This procedure is needed in order to remove weakly sticking grains while the well adherent grains remain on the surface. The substrate is then returned into the chemical bath at the same position and the deposition time starts to be measured. In such a manner we have prepared thin films for 5, 10 and 15 min deposition times. To prepare a thicker film, the film already obtained for the 15 min deposition time is reinserted into a fresh chemical bath for a further 15 min (2 × 15 min) which is done 5 min after the beginning of the deposition reaction in the second chemical bath. Thus the obtained film will be further designated as that prepared for the 30 min deposition time. By the deposition procedure used the films are deposited on the both sides of the substrates. To remove the thin film from the non-conductive side of the substrate, this site is carefully wiped with cotton wetted with 2 M aqueous solution of sodium hydroxide. Finally, the as-deposited thin films are wiped with cotton soaked with ethanol, rinsed with ethanol and left vertically to dry at room temperature. The thin films prepared for the 5 to 15 min deposition time have yellow color, while the thicker film obtained for the 30 min deposition time has yellow-brown color. The precipitate from the chemical bath is separated by vacuum filtration, washed with ethanol and dried in air at room temperature for 3–4 h. The fresh precipitate has brown color.

3. Characterization of the thin films The composition and structure of both thin films and precipitate from the chemical bath were examined using a Rigaku Ultima IV X-ray diffractometer with CuKα radiation. The thermal studies (TG and DTA) were carried by a LABSYS™ Evo apparatus (SETARAM) in a temperature interval of up to 500 °C in an airflow at a heating rate of 10 °C/min. Infrared spectra were recorded with a Perkin-Elmer System 2000 infrared interferometer using KBr disks. The morphology of the thin films was observed by scanning electron microscopy (JEOL JSM-5510). The topography of the film surfaces was examined by AFM in taping mode at room temperature using NanoScopeV system (Veeco Instruments Inc.). In-situ optical spectra of the thin films were recorded by a Varian Cary 50 Scan spectrophotometer ranging from 350 to 900 nm at voltages in the range of ±2.5 V. The electrochromic cell was a home-made cell of 3 mm thick window glass. The cell is actually a Vis cuvette on a squared glass base with holes that allow the cuvette to be attached to the spectrophotometer. The electrochromic cell was used as a twoelectrode system: one electrode was a blank FTO substrate and the other electrode was FTO substrate with a thin film. The distance between the electrodes was about 1 cm and 1 M LiClO4 in propylene carbonate (PC) was used as an electrolyte (30 ml). The surface of each electrode was about 8 cm2. The prolonged cycling up to 500 cycles was performed in the same two-electrode electrochromic cell with alternative square pulse voltage of +2.5 V and switching time of 60 s. The electrochemical behavior of Na0.33V2O5∙H2O thin films were examined by cyclic voltammetry in 1 M LiClO4 (PC) in a conventional three-electrode cell using a micro AUTOLAB II equipment (Eco-Chemie) in the potential range initially between −2.5 and +2.5 V, and then reduced to −1 and +1 V. The prepared thin film is the working electrode, the reference electrode is Ag/AgCl (3 M KCl) and the auxiliary electrode is a platinum wire. The CV curves are recorded at 10 and 50 mV/s scanning rates. The film thickness was measured by a Alpha Step D-100 profilometer (measuring parameters: stylus force 5 mg, length 8 mm, range 10 μm and speed 0.07 mm/s). All as-deposited Na0.33V2O5·nH2O xerogels thin films with different thickness have passed an adhesion tape test.

310

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

4. Results and discussion 4.1. Chemical consideration for film formation The chemistry of the deposition process is based on a controlled precipitation reaction resulting from the acidification of the NaVO3 solution that occurs in the presence of diethyl sulfate. The main idea for the precipitation process was taken from a previous study of one of the authors [24]. Above 65 °C the hydrolysis of diethyl sulfate takes place according to the reaction: ðCH3 CH2 OÞ2 SO2 ðaqÞ þ 4H2 OðlÞ → 2CH3 CH2 OHðaqÞ þ 2H3 Oþ ðaqÞ þ SO4 2− ðaqÞ:

The concentration of H3O+ gradually increases (decreasing pH to about 5.5) and the conditions for precipitation of vanadium(V) oxide xerogels are fulfilled [6]. 4.2. Composition and structure of the thin films Fig. 1 shows the X-ray powder diffraction (XRD) patterns of the asdeposited film (Fig. 1a) and the precipitate from the chemical bath, so called the brown sample (Fig. 1b). Besides the peaks due to the FTO substrate (PDF 46–1088) the two patterns are very similar, except the difference in the intensities of some peaks, for instance at 25.49, 29.85, 47.31 and 50.54° (2θ). In general, such a difference is reasonable and it is mainly due to the occurrence of texture in thin film growth [25]. The similarity between the XRD patterns evidences the same phase composition of the film and precipitate. This is of importance since we are able to undertake some studies that require more amount of the sample (for example TG-DTA analyses) using the precipitate and the obtained data are then referred to the film. Both patterns display comparatively small number of diffraction peaks, about ten peaks, all broad. The first peak centered at 7.89° (d = 11.20 Å) has much higher intensity than the other peaks located above 23° (2θ scale). The comparison with the patterns from PDF database showed that our diffractograms do not correspond exactly to any known X-ray powder patterns of vanadium compounds. However,

Fig. 1. XRD patterns of (a) as-deposited film, (b) brown precipitate and (c) greenish precipitate.

there is a great similarity with the patterns of V2O5·nH2O xerogel (PDF 40–1296) with the difference that our patterns exhibit additional peaks. From the literature survey on vanadium oxides [6,26–29] it is well known that the X-ray patterns of V2O5·nH2O xerogel are characterized by a small number of 3–5 broad peaks from the series of 00l reflections (missing 002 peak). Moreover, the intensity of the first order diffraction 001 located generally around d = 11.5 Å is much larger than that of the other terms. The structure of xerogel was resolved by Petkov et al. [29]. The xerogel is found to be an assembly of double V2O5 sheet forming slabs that are stacked along the c-axis of a monoclinic unit cell. The slabs are separated by water molecules. The basal distance between the layers depends on the amount of water and increases by step of about 2.8 Å for each water layer: 11.55 Å for n ≈ 1.5–1.6 and 8.75 Å for n ≈ 0.5 [30]. Due to the layered structure, vanadium(V) oxide gels are able to intercalate a wide variety of inorganic and organic guest species without change in the onedimensional stacking of the layers [6,31–35]. It was reported that the intercalation of cations like Na+ and TMA+ (tetramethyl ammonium) into the xerogel leads to the appearance of extra diffraction peaks in the diffraction patterns of M0.3V2O5·1.5H2O [31]. Durupthy et al. [31] have suggested that the observation of hkl set of reflections (instead of 00l only) is related to the loss of ordered stacking of the double V2O5 layers. It is important that our diffractograms resemble to a great extent the patterns given in the above paper. Based on all above we consider that our synthesis product in the form of precipitate or film is Na+ intercalated vanadium(V) oxide xerogel, NaxV2O5·nH2O, so that the oxidation number of vanadium in our samples becomes less than 5. The incorporation of cations between the layers of the gels is a result of ion-exchange reactions with the acid protons of the gels, so the amount of the intercalated ions is around 0.3–0.4 per mole of V2O5 [31,33–36]. For the sodium ions the equilibrium amount is found to be 0.33 [32]. Concerning our compositions (precipitate and film) we have one more argument in favor of the above ratio: during the annealing at 400 °C they completely transform into a single phase of monoclinic NaV6O15 (Na0.33V2O5) (PDF 86–120) without any additional diffraction peaks (XRD patters are not shown) which confirms a Na:V ratio of 1:6 in the as-prepared compositions. The formation of vanadium(V) oxide xerogel is further supported by IR spectroscopy (Fig. 2). It is clearly seen that the spectral characteristics of the synthesis product are different from those of the initial reagent NaVO3 (Fig. 2a). The IR spectra of the as-deposited scraped film (Fig. 2b) and precipitate (Fig. 2d) are practically identical (within the limit of the experimental resolution), thus confirming the same phase composition of the two samples. The IR spectra (Fig. 2b, d) are dominated by strong absorptions in the 1020–400 cm−1 region associated with the vibrations of the vanadium– oxygen framework. The band at 1012 cm−1 is attributed to the stretching vibration of terminal V_O groups (shorter V\\O bond), the band at 763 cm−1 is due to asymmetric stretching vibrations of the bridged V\\O\\V units (longer V\\O bonds), and the band at 514 cm−1 is assigned to the V\\O\\V symmetric stretch mixed with bending vanadium oxygen vibrations [36–38]. The absorption at 917 cm−1 is likely to be associated with a V_O stretch strongly perturbed by the water molecules [38]. In addition, a weak band near 1230 cm−1 appears but its origin is unclear (Fig. 2). The presence of water molecules in the V2O5·nH2O xerogel is clearly manifested by the bands at 3590 cm−1 (shoulder) and 3400 cm−1 (OH stretching vibrations) and at 1614 cm−1 (HOH bending vibration) (Fig. 2b, d). The band near 3600 cm−1 has been assigned to water molecules nearly free of hydrogen bonding, which are presumably directly bonded to vanadium through their oxygen atom [37–39]. The band below 3580 cm−1, in our samples at 3400 cm−1 has been attributed to the water molecules hydrogen bonded with the oxygen either of V2O5 [37, 40] or of other H2O molecules [37–39]. There is a third kind of water molecules, that also give rise to a νOH band around 3600 cm−1 since they are not involved in hydrogen bonds [37,38]. These water molecules

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

Fig. 2. IR spectra of (a) NaVO3, (b) as-deposited scraped film obtained for 30 min, (c) film aged 20 months, (d) fresh brown precipitate and (e) brown-greenish precipitate aged more than 3 years.

(n around 0.1–0.5) are supposed to be trapped into the cavities of the lattice, where they are strongly retained and thus leave the structure above 250 °C just before the crystallization of the sample [30,37,38]. It is noticeable that the band positions in the prepared Na0.33V2O5· nH2O are very close to those for pristine V2O5·nH2O xerogels: 1015, 760 and 515 cm−1 [37–39]. This finding evidences that the V2O5 framework in Na0.33V2O5·nH2O is not essentially affected by the intercalated sodium ions. All the same, the small downshift with 3 cm−1 of the band at 1012 cm−1 in the as-prepared xerogel (Fig. 2) implies more V(IV) sites than that in the pristine V2O5·nH2O xerogels which can be related to the presence of Na+ ions in the interlayer space. The water content in the as-prepared fresh Na0.33V2O5·nH2O precipitate (brown sample) was determined by the TG-DTA technique (Fig. 3). As seen the release of the water molecules is stepwise in accord with earlier reports [2,30,37]. The first step is developed in the temperature range of 40–150 °C (endothermic peak at 116 °C) with a mass loss of 4.40% equivalent to 0.5 mol H2O. These water molecules are the most weakly bonded ones. Between 150 and 270 °C the TG curve shows a gradual decrease with a small mass loss of 1.43% (0.16 mol H2O) without a distinct endothermic effect. A sharp endothermic effect appears at 293 °C which is accompanied by a sharp TG step with mass loss of 2.90% (0.33 mol H2O). This is the last amount of strongly retained water molecules. Their release is immediately followed by an exothermic effect at 373 °C due to the crystallization of monoclinic NaV6O15. According to the TG curve the total mass loss is 8.73% which corresponds to one mole H2O per Na0.33V2O5. From the XRD, IR spectroscopy and TG-DTA data we can conclude that the synthesis product in the form of film and precipitate is a vanadium oxide xerogel with the composition Na0.33V2O5·H2O. It is worth mentioning that over time (for instance, in about 7 days) both the precipitate and films turn greenish. The color change occurs regardless the storage conditions (at ambient conditions or closed vessel) and it is an indication for a reduction process at the V(V) sites. This

311

Fig. 3. TG-DTA curves for Na0.33V2O5∙H2O xerogel precipitates: (a) fresh and (b) aged 20 days.

process, however, does not appear to modify the XRD pattern of greenish precipitate (Fig. 1c) which is very similar to that of the fresh one and only differences in the relative intensity of some peaks such as at 25.44 and 32.74° (2θ) are observed (Fig. 1b). Since the IR spectroscopy can give valuable information on the change in the valence state of vanadium [41,42] we have further examined the IR spectra of the precipitate and film, both aged at different times, 20 days, 20 months and 3 years, and some of them are included in Fig. 2. The inspection of the IR spectra showed that the spectra of the precipitates aged 3 years and film aged 20 months match closely those of the corresponding samples aged 20 days, as well as the spectra of the films that are practically the same with the spectra of the precipitates. So that, as an illustration in Fig. 2 the IR spectra of a film aged 20 months (Fig. 2c) and a precipitate aged 3 years (Fig. 2e) are compared with the spectra of the corresponding fresh samples (Fig. 2b,d). By IR spectroscopy it was established that the electrochemical reduction of V(V) from V2O5 to V(IV) causes a considerable modification of the spectra as both the terminal V_O and bridging V\\O\\V stretching vibrations are mainly affected [41,42]. Therefore, in the presence of more V(IV) species we should expect a red shift of the band at 1012 cm−1 as well as a decrease in the intensity and frequency of the band at 761 cm−1 [41,42]. The close examination of the IR spectra in Fig. 2 reveals that even prolonged aging does not result in any significant difference in the band positions and intensities in comparison with the fresh samples. Only a very small frequency variation between 2 and 4 cm−1 is observed in some spectra, but these variations are not consistent and, particularly the former one is within the limit of the experimental resolution (Fig. 2). These spectral features provide evidence for the low degree of reduction of V(V) to V(IV) that takes place during the aging, so the amount of the newly obtained V(IV) is not high enough to affect the internal V\\O framework of the initial xerogel. We could also suppose that the reduction process occurs mainly on the surface of the particles, and thus the vibrational characteristics of the bulk material appear to be unchanged.

312

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

The sensitivity of the vanadium(V) oxide xerogels toward reduction under storage which is accompanied by progressive color change in green after some weeks or months was commented on by Livage [6], but the explanation of this phenomenon is still not clear. Unfortunately, we are not able to specify what species are responsible for the electron transfer process leading to the V(V) reduction. However, considering the easy hydrolysis of the weakly bound water as found in [43] we could suppose a possible role of the water molecules and protons in the reduction process. On the other hand, our IR absorption studies give valuable information that the degree of the reduction of V(V) in Na0.33V2O5·H2O xerogel does not increase during the prolonged aging at ambient conditions. Whereas the vibrations of the V\\O units are not affected, some changes are observed in the OH stretching mode region (Fig. 2). In the greenish samples the intensity of the band around 3600 cm−1 is increased and the band near 3440 cm−1 is upshifted (Fig. 2c, e) compared to the respective bands in the yellow/brown samples (Fig. 2b, d). Obviously, some changes related to the amount and state of the water molecules in the gel occur during storage and to elucidate this phenomenon we examined the thermal behavior of the greenish sample aged 20 days (Fig. 3). As seen the TG curves for the fresh and aged samples differ considerably from each other. Firstly, the total mass loss for the greenish sample is increased to 10.67% (vs. 8.73% in brown sample) which leads to a higher water content of 1.3 mol. This water amount is distributed in the following proportions: 1.1 mol (mass loss of 9.13%) are released between 40 and 205 °C (first strong endothermic effect at 175 °C), about 0.1 mol (mass loss of 1.32%) between 205 and 270 °C and the last amount of 0.1 mol between 270 and 320 °C. The comparison in the proportions between the two samples shows that the greenish sample contains more amount of water molecules that are removed below 200 °C than the brown sample, i.e. more amount of weakly bonded water molecules. Therefore, the “extra” water molecules accommodated during the storage are weakly bonded. This finding can explain the observed increase in the intensity of the ν(OH) band at 3595 cm−1 in the IR spectrum of the aged samples.

From the above data it follows that the vanadium reduction during aging of our Na0.33V2O5·H2O is accompanied by an increase in the water content to Na0.33V2O5·1.3H2O. Such phenomenon has been previously reported for reduced xerogels [44]. Babonneau et al. [44] have found that the reduced gels having V(IV)/V(V) of 16% contain 2.5 H2O per V2O5 instead of 1.6–1.8 H2O for gels with usual V(IV)/V(V) about 1 to 4%. 4.3. Morphology of Na0.33V2O5·H2O thin films As described in the literature [6,31] vanadium(V) oxide xerogels normally exist in the form of long ribbons. SEM images of two thin films prepared for the 15 and 30 min deposition times are depicted in Fig. 4. Among the studied films these are the films exhibiting the best optical properties (see Section 4.5) and their thickness determined by the profilometer is about ~150 and ~300 nm (for the 15 and 30 min deposition times, respectively). The cross-sectional SEM image of the film for the 30 min deposition time (Fig. 4d) gives a thickness of about 250 nm. The chemical bath deposition method produces thin films which surface is completely covered with the deposited material (Fig. 4a, b and c). Both films are dense and without any porosity. The SEM image of the thinner film recorded at a higher magnification (Fig. 4b) clearly shows the granular structure of the film. It is composed of nanograins, spherical and elongated, with sizes between 50 and 200 nm. Besides well separated nanograins, randomly oriented ribbon-like units (about 200 nm wide and 1–1.5 μm long) are also visible (Fig. 4a). The observations by SEM are supported by AFM data for the same Na0.33V2O5∙H2O thin films (Fig. 5). The 2D surface topography of the thinner film (Fig. 5a) shows nanograins with sizes in the range of 50–150 nm. In addition, there are elongated ribbon-like units with the following dimensions: width from 150 to 300 nm and length from 0.7 to 2 μm. The longer deposition time ensures the growth of the grains and, expectedly, the thicker film (Fig. 5b) exhibits larger grains with sizes from 250 to 700 nm as well as larger ribbons reaching to 1 μm

Fig. 4. SEM images of Na0.33V2O5∙H2O thin films with different thickness: (a) and (b) ~150 nm, (c) ~300 nm and (d) cross-section view of the thin film with ~300 nm thickness.

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

313

Fig. 5. AFM images of Na0.33V2O5∙H2O thin films: (a) and (b) are 2D images with profiles of films with 150 and 300 nm thickness, respectively; (c) and (d) are 3D images of films with 150 and 300 nm thickness, respectively.

wide and 3 μm long. Moreover, AFM data reveal that the films exhibit comparatively high surface roughness with amplitude of the grain height at around 600 nm for the thinner film and 250–500 nm for the thicker one, i.e. the thickness is not uniform throughout the whole film area. The 3D images, and especially those for the thinner film (Fig. 5c), clearly illustrate the way for the formation of the ribbons. We can see that the ribbons arise as a result of the coalescence of nanograins in a preferred direction. We suppose that only the grains having the most suitable orientation and that are situated closely to each other are able to coalesce and thus to form long ribbon-like units. Moreover, we can also speculate, that a possible growth mechanism of Na0.33V2O5∙H2O thin films is the “island” mechanism (Fig. 5d) in which the coalescence of the islands generates ribbon-like units and the latter, at a certain point, will form the compact film layer.

4.4. Electrochemical properties Firstly, we performed CV measurements in large potential range of ±2.5 V with a blank substrate. As expected, we observed only oxygen reduction that begins around −1.5 V and high polarization of the working electrode in the range of ±1 V with ~0 mA/cm2 current density. This

allows us to perform electrochemical analysis in the relevant voltage range of ±1 V. Fig. 6 shows cyclic voltammograms with five scans (10 mV/s) of two as-deposited films with different thickness. Both cyclic voltammograms exhibit three pairs of oxidation/reduction peaks. It is essential that the shape and position of the respective redox peaks are very close for the two films (they differ within ±0.04 V) which reflect the same electrochemical processes. The anodic peaks (A1, A2, A3) appear at −0.15, 0.16 and 0.74 V and the cathodic peaks (C1, C2, C3) are at −0.50, −0.16, and 0.38 V (Fig. 6a). The A2/C2 pair is the most intense one. The color changes observed in the three-electrode cell are the same as those in the two-electrode electrochromic cell. The survey from the literature shows that the cyclic voltammograms of vanadium(V) oxide xerogels display one [45], two [46] and three redox pairs [2]. The presence of one redox pair has been attributed to the amorphous nature of the films [45]. According to Benmoussa et al. [46] the appearance of two redox pairs reflects the crystalline nature of the films. The observation of three redox pairs in our films is consistent with data of Costa et al. obtained at the same scanning speed for vanadium oxide gel prepared on flexible polyethylene terephthalate/ indium tin oxide electrodes [2]. It should be mentioned that the general view of the CV curves presented by Costa et al. [2] also having an

314

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

Fig. 6. Five cyclic voltammograms of Na0.33V2O5∙H2O thin films with thickness: (a) 150 nm and (b) 300 nm at a scanning rate of 10 mV/s.

intensive A2/C2 redox pair as well as the voltage ranges for the anodic (from −0.23 to 0.64 V) and cathodic (from −0.8 to 0.3 V) peaks are very similar to our data. The calculation of the exchanged charge showed that the values for the extracted charge for the thin film with 150 nm thickness is 65.5 mC/cm2 and for the film with 300 nm thickness is 75.5 mC/cm2. The inserted charge values are much lower, 19.9 mC/cm2 and 22.3 mC/cm2 respectively. The higher values for the exchanged charge for thicker films are expected due to their higher capacity. The electrochemical reversibility for 200 cycles was monitored at scanning rate of 50 mV/s for the film with 300 nm thickness (Fig. 7). In conformity with the higher scanning rate all CV curves in Fig. 7 display one broad pair of oxidation/reduction peak instead of the three pairs well-separated at the lower scanning rate (Fig. 6b). The oxidation/reduction pair remains stable during the cycling with a slight shift of the peak maximum to lower potentials: from 0.48 V/−0.44 V for the 5th scan to 0.43 V/− 0.41 V for the 200th scan. Moreover, the peak maxima move to higher potentials than those of the most intensive A2/C2 pair at the lower rate. It is also observed that the peak area (exchanged charges) decreases during the cycling and this process is

Fig. 7. Cyclic voltammograms to 200 scans of thin film with 300 nm thickness at a scanning rate of 50 mV/s.

more pronounced up to the 100th cycle. Then, the electrochemical reversibility appears to be stabilized. As mentioned previously, over time, the as-deposited films changed its original yellow/yellow-brown color into greenish color. Fig. 8 compares the CV curves (up to 20 scans) of a fresh film (yellow-brown) and an aged 20 days film (brown-greenish) with the same thickness. Opposite to our expectations, the comparison shows that electrochemical behavior of the fresh and aged film during the initial scans does not differ considerably. In both cases there are three cathodic and three anodic peaks, however, the first anodic peak A1 is very well distinguished in the fresh film, while in the aged one it appears as a shoulder to the second A2 peak. The corresponding peak potentials show differences within ±0.06 V. With increasing scan number a clear difference between the fresh and aged film is observed (Fig. 8). During the longer cycling some changes occur and the fresh film is more essentially concerned. It is clearly seen (Fig. 8a) that after 15 scans some peaks disappear in the CV curves of the fresh film. The A3/C3 pair surely disappears, the A1/C1 pair strongly diminishes and the A2 peak is shifted to more positive values of the potential. Thus, the cyclic voltammograms of the fresh film transform from a system with three redox peaks to a system with only one redox pair, A2/C2, which is characteristic of amorphous films [47]. At the same time the aged film exhibits more stable cycling behavior and only the redox A1/C1 pair decreases in intensity. Regarding the exchanged charges, going from the 1st to the 20th scans for the fresh film the extracted charge slightly decreases from 41 to 37 mC/cm2, while the decrease in the inserted charge is more pronounced (from 15 to 8 mC/cm2). The film aging causes a reduction mostly of the extracted charges (from 34 to 30 mC/cm2 for 5th and 20th scans, respectively), while the inserted charges are similar to those for the fresh film (from 14 to 8 mC/cm2 for 5th and 20th scans), respectively. The peak disappearing in the CV curves with the increasing scan number is observed by other scientists in the cases of vanadium oxide xerogels [2,48] but the origin of this phenomenon is still not clear. Several possible reasons are commented in the literature. The deactivation of some redox sites of the film could be related either to amorphization of the film [45] or to crystalline phase transition

Fig. 8. Cyclic voltammograms of thin films with 300 nm thickness: (a) fresh film, 1st to 4th and 16th and 20th scans, (b) aged film, 5th, 7th, 10th, 15th and 20th scans (10 mV/s).

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

315

triggered by the oxidation/reduction cycles [2] or to changes in morphology and microstructure of the film leading to the formation of a more “homogeneous” film [48]. The observed redox peaks in the potential range ±1 V for Na0.33V2O5∙H2O thin films are related to step-wise reduction of V(V) to V(IV) concomitant with formation of different crystalline states LixV2O5 and, accordingly reversible V(IV) oxidation and Li-deintercalation [46,49,50]. 4.5. Optical properties Optical transmittance spectra in the 350–900 nm spectral range of the as-deposited thin films are given in Fig. 9. It is important that the four films exhibit very similar dependence of the optical transmittance on the wavelength, but the overall transmittance decreases with the film thickness. This is expected since more states in the thicker film are available for the photons to be absorbed. The thin films have relatively high transmittance (low absorbance) at wavelengths longer than 550 nm as the thinner films (obtained for 5, 10 and 15 min deposition time) exhibit transmittance between 70 and 90%, while the thick film (30 min deposition time) exhibits a lower transmittance between 55 and 70%. The marked reduction in the transmittance at wavelengths shorter than 550 nm is associated with the fundamental absorption. It is essential that the absorption edge for the four films appears in a narrow interval of wavelengths (365–385 nm). This means that the films with different thicknesses should have very close values of the optical band gap which implies the same or very close stoichiometry regarding the different vanadium sites. In this regard we have also compared the IR spectra of the scraped films obtained for the 5, 15 and 30 min deposition and we have not found any variation in the positions and intensity of the vibrational bands in comparison with the IR spectrum presented in Fig. 2b. Considering these data we believe that the change of the film color with the time of deposition from yellow to yellow-brown shade is mainly related to the film thickness and grain growth evidenced by AFM analysis (Fig. 5a, b) rather than the same changes in the vanadium oxidation state during the deposition. The optical transmittance of Na0.33V2O5∙H2O thin films in the 350–900 nm range is recorded in a two-electrode electrochromic cell using LiClO4/PC electrolyte. Our previous studies on (NH4)xV2O5· 1.3H2O thin films in the voltage range between ±1 and ±2.5 V have shown that the increased voltage results in increasing transmittance variance (ΔT) and shorter bleaching and coloration response times [30]. It is important to notice that the current densities achieved in the used two-electrode electrochromic cell are around 0.27 mA/cm2 at +2.5 V (at the highest voltage), but these values are still much lower than those achieved in the three-electrode cell (CV measurements): around 1.5 mA/cm2 at a significantly lower voltage of 0.3 V. This fact explains the longer response time (Fig. 10) observed in the two-electrode electrochromic cell.

Fig. 9. Optical transmittance spectra of as-deposited thin films with time of deposition.

Fig. 10. Response time of Na0.33V2O5∙H2O thin film with 300 nm thickness at 900 nm at ± 2.5 V.

The response time (τ) is calculated as the time required for the thin film to change its color from yellow/brown to obtain 90% [51] of the blue color (τc,90% — coloration response time) or vice versa, to obtain 90% of the yellow/brown color (τb,90% — bleaching response time). It is seen in Fig. 10 that the bleaching response time τb,90% is 10 min, while the coloration response time τc,90% is 17.5 min i.e. the reduction process followed with Li+-intercalation is almost two times slower than the oxidation process followed with Li+-deintercalation. The optical transmittance spectra recorded at ±2.5 V with a switching time of 20 min of four Na0.33V2O5∙H2O thin films prepared for different deposition times are shown in Fig. 11. It is seen that the dependence of the transmittance of the reduced forms on wavelength passes through a maximum at 530 nm (Fig. 11a) and 560 nm (Fig. 11b, c, d). The oxidized forms exhibit a steep increase in the transmittance up to 600 nm with further gradual increases up to 900 nm. It is important that the fundamental absorption edge for both reduced and oxidized films exhibits a red shift (to lower energy) with the increase in the film thickness. The red shift of the absorption edge reflects a decrease in the optical band gap and this effect has been attributed to the increase in the grain size and effective decrease in the imperfections at the grain-boundary regions [52]. Transmittance variance (ΔT) defined as: ΔT = T(bleached) − T(colored) is used to quantify the electrochromic effect of the prepared thin films. For all thin films the maximum ΔT value is obtained at 900 nm (Fig. 11). The three thinner films exhibit close ΔT values of 32–37%, while for the thicker film having 300 nm thickness ΔT reaches 55%. For the latter film the optical response during a prolonged cycling up to 500 cycles was also examined and the corresponding optical transmittance spectra are included in Fig. 11d. As seen the oxidized state is completely recovered during the cycling and the transmittance curve after 500 cycles well matches the initial curve. Some changes occur in the reduced states between the 5th and 100th cycles resulting in an increase in the transmittance between 400 and 500 nm and shift of the absorption edge to lower wavelengths (high energy) as well as a slight decrease in the transmittance between 500 and 700 nm. The observed blue shift of the absorption edge reflects a blue shift of the optical band gap and this shift could be related to the change of the film microstructure and increase in the imperfections during the lithium intercalation process [53]. Further changes between 100th and 500th cycles are not observed, so that the reduced state is stabilized which is evident from the overlap of the corresponding transmittance curves. Therefore, the data for the optical stability of the thin film after the 100th cycle in the electrochromic cell are in line with the data for its electrochemical stability. Between 700 and 900 nm the transmittance of the reduced forms in all cycles shows the same trend as the initial scan. It is very important that the ΔT value after 500 cycles is 53% at 900 nm i.e. it retains practically unchanged which is a demonstration of the good optical stability of as-deposited films. The comparison with the literature data for electrochromic vanadium(V) oxide xerogels

316

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

Fig. 11. In-situ optical transmittance spectra of Na0.33V2O5∙H2O thin films prepared for different deposition times: (a) 5 min, (b) 10 min, (c) 15 min and (d) 30 min; after 5 cycles in (a, b, c) and up to 500 cycles in (d).

[47,51,54,55] shows that the obtained results are promising and can serve as a base for further design of electrochromic devices. As was mentioned before, the as-deposited yellow/yellow-brown films turn greenish tint over time. Fig. 12 shows the change in the optical transmittance over time (up to 20 months) of the best film with 300 nm thickness. The optical transmittance spectra evidence that aging does not affect the transmittance in the 350–550 nm spectral region. As expected from the greenish shade of the aged film, however, a gradual decrease of the film transmittance at wavelengths longer than 550 nm (i.e. an increased absorption of the red light) is observed as the transmittance reduction in 20 days aging is about 10% at 900 nm. It is also seen that the transmittance reduction is most significant up to 10th day while the film structure still stabilizes (about 8%) and after that it is only about 2%. Further decrease of the transmittance is not recorded even for a prolonged aging of 20 months (Fig. 12). It is very important, however, that when such a film aged 20 days is placed in the electrochromic cell (Fig. 12b) it is completely recovered in its oxidized state and thus ΔT retains its high value (56%). 5. Conclusions Well-structured nano-sized thin films with compositions Na0.33V2O5· nH2O (n = 1 and 1.3) are successfully deposited by an optimized chemical bath method where the acidification of NaVO3 solution occurs through the hydrolysis of diethyl sulfate. The thin films demonstrate appreciable optical transmittance varying between 40 and 70% at 500 nm in dependence on the film thickness. The film structure and morphology, electrochemical behavior and electrochromic activity of fresh and aged thin films are studied. It is established that the long-term aging of the films does not affect their electrochromic properties. The thin films exhibit maximum transmittance variance (ΔT) at 900 nm which varies between 32 and 55% depending on the film thickness. The electrochemical

Fig. 12. Optical transmittance spectra of as-deposited thin film with 300 nm thickness: (a) over time and (b) brown-greenish film aged 20 days placed in the electrochromic cell.

M. Najdoski et al. / Surface & Coatings Technology 277 (2015) 308–317

reversibility and electrochromic reproducibility with ΔT retention up to 500 cycles have been demonstrated. These characteristics can be attributed to the well-defined structure and nanosized morphology of the thin films provided by the suitable preparation method. The results obtained are good prerequisite for further application of Na0.33V2O5·nH2O thin films in different electrochromic devices. Acknowledgment The authors thank the Bulgarian Academy of Sciences and the Macedonian Academy of Sciences and Arts for the financial support and the Alexander von Humboldt Foundation for providing the electrochemical equipment. References [1] W.D. Murphy, A.P. Christian, J.F. Disalvo, V.J. Waszczak, Lithium incorporation by vanadium pentoxides, Inorg. Chem. 18 (1979) 2800–2803. [2] C. Costa, C. Pinheiro, I. Henriques, C.A.T. Laia, Inkjet printing of sol–gel synthesized hydrated tungsten oxide nanoparticles for flexible electrochromic devices, Appl. Mater. Interfaces 4 (2012) 5266–5275. [3] A.L. Pergament, E.L. Kazakova, G.B. Stefanovich, Optical and electrical properties of vanadium pentoxide xerogel films: modification in electric field and the role of ion transport, J. Phys. D. Appl. Phys. 35 (2002) 2187–2197. [4] F. Cheng, J. Chen, Transition metal vanadium oxides and vanadate materials for lithium batteries, J. Mater. Chem. 21 (2011) 9841–9848. [5] D. Vernardou, State-of-the-art of chemically grown vanadium pentoxide nanostructures with enhanced electrochemical properties, Adv. Mater. Lett. 4 (2013) 798–810. [6] J. Livage, Vanadium pentoxide gels, Chem. Mater. 3 (1991) 578–593. [7] K. Takahashi, S.J. Limmer, Y. Wang, G.Z. Cao, Synthesis and electrochemical properties of single-crystal V2O5 nanorod arrays by template-based electrodeposition, J. Phys. Chem. B 108 (2004) 9795–9800. [8] J. Scarminio, A. Talledo, A.A. Andersson, S. Passerini, F. Deckers, Stress and electrochromism induced by Li insertion in crystalline and amorphous V2O5 thin film electrodes, Electrochim. Acta 38 (1993) 1637–1642. [9] V. Vivier, S. Belair, C.C. Vivier, J.Y. Nedelec, L.T. Yu, A rapid evaluation of vanadium oxide and manganese oxide as battery materials with a micro-electrochemistry technique, J. Power Sources 103 (2001) 61–66. [10] E.A. Olivetti, J.H. Kim, D.R. Sadoway, A. Asatekin, A.M. Mayes, Sol–gel synthesis of vanadium oxide within a block copolymer matrix, Chem. Mater. 18 (2006) 2828–2833. [11] V.S.R. Channu, R. Holze, E.H. Walker Jr., S.A. Wicker Sr., R.R. Kalluru, Q.L. Williams, W. Walters, Synthesis and characterization of lithium vanadates for electrochemical applications, Int. J. Electrochem. Sci. 5 (2010) 1355–1366. [12] H. Yin, K. Yu, H. Peng, Z. Zhang, R. Huang, J. Travas-Sejdic, Z. Zhua, Porous V2O5 micro/nano-tubes: synthesis via a CVD route, single-tube-based humidity sensor and improved Li-ion storage properties, J. Mater. Chem. 22 (2012) 5013–5019. [13] J. Legendre, J. Livage, Vanadium pentoxide gels: I. Structural study by electron diffraction, J. Colloid Interface Sci. 94 (1983) 75–83. [14] R. Ceccato, G. Carturan, Sol–gel synthesis of vanadate-based thin films as counter electrodes in electrochromic devices, J. Sol-Gel Sci. Technol. 26 (2003) 1071–1074. [15] C.G. Granqvist, Electrochromics for smart windows: oxide-based thin films and devices, Thin Solid Films 564 (2014) 1–38. [16] S. Papaefthimiou, E. Syrrakou, P. Yianoulis, An alternative approach for the energy and environmental rating of advanced glazing: an electrochromic window case study, Energy Build. 41 (2009) 17–26. [17] R. Baetens, B.P. Jelle, A. Gustavsen, Properties, requirements and possibilities of smart windows for dynamic daylight and solar energy control in buildings: a state of the art review, Sol. Energy Mater. Sol. Cells 94 (2010) 87–105. [18] N.I. Jaksic, C. Salahifar, A feasibility study of electrochromic windows in vehicles, Sol. Energy Mater. Sol. Cells 79 (2003) 409–423. [19] H.K. Koduru, H.M. Obili, G. Cecilia, Spectroscopic and electrochromic properties of activated reactive evaporated nano-crystalline V2O5 thin films grown on flexible substrates, Int. Nano Lett. 3 (2013) 24–31. [20] S. Beke, A review of the growth of V2O5 films from 1885 to 2010, Thin Solid Films 519 (2011) 1761–1771. [21] M. Najdoski, V. Koleva, A. Samet, Effect of deposition conditions on the electrochromic properties of nanostructured thin films of ammonium intercalated vanadium pentoxide xerogel, J. Phys. Chem. C 118 (2014) 9636–9646. [22] M. Najdoski, V. Koleva, A. Samet, Influence of vanadium concentration and temperature on the preparation of electrochromic thin films of ammonium intercalated vanadium(V) oxide xerogel nanoribbons, Dalton Trans. 43 (2014) 12536–12545. [23] M. Najdoski, V. Koleva, S. Demiri, Chemical bath deposition and characterization of electrochromic thin films of sodium vanadium bronzes, Mater. Res. Bull. 47 (3) (2011) 737–743.

317

[24] M. Najdoski, T. Todorovski, A simple method for chemical bath deposition of electrochromic tungsten oxide films, Mater. Chem. Phys. 104 (2007) 483–487. [25] M. Birkholz, Thin Film Analysis by X-ray Scattering, John Wiley & Sons, New Jersey, 2006 230. [26] R. Ceccato, S. Dirè, T. Barone, G. De Santo, E. Cazzanelli, Growth of nanotubes in sol–gel-derived V2O5 powders and films prepared under acidic conditions, J. Mater. Res. 24 (2009) 475–481. [27] T. Yao, Y. Oka, N. Yamamoto, Layered structures of vanadium pentoxide gels, Mater. Res. Bull. 27 (1992) 669–675. [28] J. Livage, F. Beteille, C. Roux, M. Chatry, P. Davidson, Sol–gel synthesis of oxide materials, Acta Mater. 46 (1998) 743–750. [29] V. Petkov, P.N. Trikalitis, E.S. Bozin, S.J.L. Billinge, T. Vogt, M.G. Kanatzidis, Structure of V2O5·nH2O xerogel solved by the atomic pair distribution function technique, J. Am. Chem. Soc. 124 (2002) 10157–10162. [30] P. Aldebert, N. Baffier, N. Gharbi, J. Livage, Layered structure of vanadium pentoxide gels, Mater. Res. Bull. 16 (1981) 669–676. [31] O. Durupthy, N. Steunou, T. Coradin, J. Maquet, C. Bonhomme, J. Livage, Influence of pH and ionic strength on vanadium(V) oxides formation. From V2O5·nH2O gels to crystalline NaV3O8·1.5H2O, J. Mater. Chem. 15 (2005) 1090–1098. [32] L. Znaidi, N. Baffier, D. Lemordant, Kinetics of the H+/M+ ion exchange in V2O5 xerogel, Solid State Ionics 28–30 (1988) 1750–1755. [33] Y.J. Liu, J.A. Cowen, T.A. Kaplan, D.C. De Groot, J. Schindler, C.R. Kannewurf, M.G. Kanatzidis, Investigation of the alkali-metal vanadium oxide xerogel bronzes: AxV2O5·nH2O (A = K and Cs), Chem. Mater. 7 (1995) 1616–1624. [34] Y.J. Liu, J.L. Schindler, D.C. De Groot, C.R. Kannewurt, W. Hirpo, M.G. Kanadzidis, Synthesis, structure, and reactions of poly(ethylene oxide)/V2O5 intercalative nanocomposites, Chem. Mater. 8 (1996) 525–534. [35] V.L. Volkov, G.S. Zakharova, Nanotubes in the V2O5·nH2O xerogel-hydroquinone H2O system, Russ. J. Inorg. Chem. 54 (2009) 36–39. [36] M. Malta, G. Louarn, N. Errien, R.M. Torresi, Redox behavior of nanohybrid material with defined morphology: vanadium oxide nanotubes intercalated with polyaniline, J. Power Sources 156 (2006) 533–540. [37] L. Abello, E. Husson, Y. Repelin, G. Lucazeau, Structural study of gels of V2O5: vibrational spectra of xerogels, J. Solid State Chem. 56 (1985) 379–389. [38] Y. Repelin, E. Husson, L. Abello, G. Lucazeau, Structural study of gels of V2O5: normal coordinate analysis, Spectrochim. Acta A 41 (1985) 993–1003. [39] C. Sanchez, J. Livage, G. Lucazeau, Infrared and Raman study of amorphous V2O5, J. Raman Spectrosc. 12 (1982) 68–72. [40] M.T. Vandenborre, R. Prost, E. Huard, L. Livage, Etude par spectroscopie infra-rouge de l'eau adsorbee sur un xerogel d'oxyde de vanadium, Mater. Res. Bull. 18 (1983) 1133–1142. [41] A. Šurca, B. Orel, G. Dražič, B. Pihlar, Ex situ and in situ infrared spectroelectrochemical investigations of V2O5 crystalline films, J. Electrochem. Soc. 146 (1999) 232–242. [42] M. Hajzeri, A. Šurca Vuk, L.S. Perše, M. Čolović, B. Herbig, U. Posset, M. Kržmanc, B. Orel, Sol–gel vanadium oxide thin films for a flexible electronically conductive polymeric substrate, Sol. Energy Mater. Sol. Cells 99 (2012) 62–72. [43] D.S. Yakovleva, V.P. Malinenko, A.L. Pergament, G.B. Stefanovich, Electrical and optical properties of thin films of hydrated vanadium pentoxide featuring electrochromic effect, Tech. Phys. Lett. 33 (2007) 1022–1024. [44] F. Babonneau, P. Barboux, F.A. Josein, J. Livage, Réduction des gels de V2O5, J. Chim. Phys. 82 (1985) 761–766. [45] J. Livage, Optical and electrical properties of vanadium oxides synthesized from alkoxides, Coord. Chem. Rev. 190–192 (1999) 391–403. [46] M. Benmoussa, A. Outzourhit, R. Jourdani, A. Bennouna, E.L. Ameziane, Structural, optical and electrochromic properties of sol–gel V2O5 thin films, Act. Passive Electron. Compon. 26 (2003) 245–256. [47] N. Ozer, S. Sabuncu, J. Cronin, Electrochromic properties of sol–gel deposited Ti-doped vanadium oxide film, Thin Solid Films 338 (1999) 201–206. [48] F.J. Anaissi, G.J.F. Demets, H.E. Toma, Electrochemical conditioning of vanadium(V) pentoxide xerogel films, Electrochem. Commun. 1 (1999) 332–335. [49] N.T.B. Bay, P.M. Tien, S. Badilescu, Y. Djaoued, G. Bader, F.E. Girouard, V.V. Truong, L.Q. Nguyen, Optical and electrochemical properties of vanadium pentoxide gel thin films, J. Appl. Phys. 80 (1996) 7041–7045. [50] S.F. Cogan, N.M. Nguyen, S.J. Perrotti, R.D. Rauh, Optical properties of electrochromic vanadium pentoxide, J. Appl. Phys. 66 (1989) 1333–1337. [51] W. Zhongchun, C. Jiefeng, H. Xingfang, Electrochromic properties of aqueous sol–gel derived vanadium oxide films with different thickness, Thin Solid Films 35 (2000) 238–241. [52] D. Sun, C.W. Kwon, G. Baure, E. Richman, J. MacLean, B. Dunn, S.H. Tolbert, The relationship between nanoscale structure and electrochemical properties of vanadium oxide nanorolls, Adv. Funct. Mater. 14 (2004) 1197–1204. [53] C.V. Ramana, R.J. Smith, O.M. Hussai, Grain size effects on the optical characteristics of pulsed-laser deposited vanadium oxide thin films, Phys. Status Solidi 1 (2003) R4–R6. [54] N. Ozer, Electrochemical properties of sol–gel deposited vanadium pentoxide films, Thin Solid Films 305 (1997) 80–87. [55] M.R.J. Scherer, P.M.S. Cunha, O.A. Scherman, U. Steiner, Enhanced electrochromism in gyroid-structured vanadium pentoxide, Adv. Mater. 24 (2012) 1217–1221.