Characterization of rhenium oxide films and their application to liquidcrystal cells

E. Cazzanelli,1 M. Castriota,1 S. Marino,1 N. Scaramuzza,1,a� J. Purans,2 A. Kuzmin,2

R. Kalendarev,2 G. Mariotto,3 and G. Das4

1Department of Physics, LICRYL-INFM-CNR and CEMIF.CAL, University of Calabria, Ponte P. Bucci,Cubo 31C, I-87036 Rende (Cosenza), Italy2Institute of Solid State Physics, University of Latvia, Riga LV-1063, Latvia3Department of Computer Science, University of Verona, Strada le Grazie 15, 37134-Verona, Italy4Bio-NanoTechnology and Engineering for Medicine, Magna Græcia University of Catanzaro, Viale Europa,88100-Germaneto (Catanzaro), Italy

�Received 28 March 2009; accepted 23 April 2009; published online 3 June 2009�

Rhenium trioxide exhibits high electronic conductivity, while its open cubic crystal structure allowsan appreciable hydrogen intercalation, generating disordered solid phases, with protonicconductivity. Rhenium oxide thin films have been obtained by thermal evaporation of ReO3 powderson different substrates, maintained at different temperatures, and also by reactive magnetronsputtering of a Re metallic target. A comparative investigation has been carried out on these films,by using micro-Raman spectroscopy and x-ray diffraction. Two basic types of solid phases appearto grow in the films: a red metallic HxReO3 compound, with distorted perovskite structures, like inthe bulk material, and ordered HReO4 crystals based on tetrahedral perrhenate ions. Because of itsconduction properties, the electrical and electro-optical behaviors of ReO3 films deposited onstandard indium tin oxide/glass substrate have been tested inside asymmetric nematic liquid crystalcells, showing an appreciable capability of rectification of their electro-optical response, in similarway to tungsten trioxide. © 2009 American Institute of Physics. �DOI: 10.1063/1.3138812�

I. INTRODUCTION

The cubic rhenium trioxide crystal, ReO3, which givesthe name to a specific perovskite-type crystal structure, be-longs to the space group Oh

1 �Pm3�m� with the lattice param-eter a0=3.7504 Å.1 The crystal lattice is composed of ReO6

octahedra joined by corners, and the Bravais unit cell con-tains 1 f.u. of ReO3. Rhenium trioxide appears red coloredand shows metallic conductivity below 500 K, so that it issometimes called “covalent metal.”2–5 In fact, the electronicconfiguration of rhenium, for the valence state correspondingto ReO3, can be written as

Re6+:�Xe��6s0��4f14��5d1�

with an unpaired electron in the band associated with d or-bital, responsible for a metal-like electronic conductivity.Tungsten, next neighbor in the periodic table, does not havethis unpaired electron in the conduction band of WO3, whileintercalated HxWO3 has an additional 5d1 polaronic electronresponsible for the electrochromic properties.6 So the inves-tigation of rhenium oxide in comparison with tungsten triox-ide looks quite interesting.

The surface reactivity of ReO3 catalyzes reactions withthe air moisture, leading to proton intercalation in the near-surface layers and to the formation of disordered phasesshowing an appreciable ionic conductivity. Therefore, thiscompound can be interesting for several applications whereelectronic and ionic conductivities are concerned, for in-stance, in the field of solid state batteries, as well as in elec-trochromic devices, possibly mixed with other oxides, and in

birefringent cells based on liquid crystals, where their inser-tion can modify the usual electro-optic response of the de-vices. For developing such possible applications is, however,necessary to deposit it in form of thin solid films, and thistask appears nontrivial, because Re ions can have differentvalence states, giving different oxides.7–9

The high stability of cubic ReO3 lattice at normal pres-sure was explained by the interaction between phonons andconduction electrons.10,11 However, lower symmetry phaseswere found in ReO3 at higher pressures.1,12–14

Moreover, experimental evidences of disordered solidphases, coexisting with the standard bulk crystal, are foundfor samples where surface effects are remarkable. First at all,near-surface layers of commercial ReO3 powders have a pe-culiar chemical composition and a defective structure, differ-ent from that of the bulk crystal. It has been reported15 thathydrogen concentration is high in the first surface layers atroom temperature.

A strong hydrogen emission from the solid ReO3 is ob-served for temperatures above 200 °C, indicating that inter-calated protons leave the host oxide. This temperature rangeis the same where the sublimation of the solid ReO3 occurs,and it can be exploited to obtain crystalline films by evapo-ration.

The proposed reaction of pure ReO3 with water is asfollows:

�1 + x�ReO3 + xH2O = HxReO3 + xHReO4

leading to a defective system, HxReO3, but also to the for-mation of HReO4.

Both hydrogenated compounds are experimentally ob-served as amorphous layers15 by electron microscopy mea-

a�Author to whom correspondence should be addressed. Electronic mail:scaramuzza@fis.unical.it.

JOURNAL OF APPLIED PHYSICS 105, 114904 �2009�

0021-8979/2009/105�11�/114904/7/$25.00 © 2009 American Institute of Physics105, 114904-1

surements. The insertion of protons induces in the host crys-tal structural modifications similar to those due to highpressure treatments, as indicated by x-ray and neutron scat-tering measurements.16 A good evidence of phase change,with a lowering of the starting cubic symmetry, is providedby repeated x-ray diffraction �XRD� measurements carriedout on ReO3 powders exposed to air moisture for longtimes.17

In the present work, the structural evolution toward de-fective solid phases is presented for a variety of systems,obtained from pure cubic ReO3 via different treatments.These systems are films deposited on different substrates bythermal evaporation of powders and films obtained by reac-tive magnetron sputtering.

As structural characterization techniques, both XRDmeasurements and extensive micro-Raman investigationshave been used.

An indirect evidence of the ionic conductivity has beenobtained for the thin films obtained by thermal evaporationof ReO3 powders: after the structural investigations, the filmsdeposited on glasses coated with indium tin oxide �ITO�have been tested in asymmetric nematic liquid crystal �NLC�cells.

A rectifying effect is expected for the electro-optic re-sponse of the NLC layer in such cells when an oxide layerwith appreciable ionic conductivity is deposited on one ofthe two ITO electrodes:18,19 it has been experimentally mea-sured for ReO3 films, giving results quite similar to thosepreviously observed in cells containing WO3 films.

II. EXPERIMENTAL METHODS

A. Sample preparation

Commercial polycrystalline powder of ReO3 �a nominalpurity of 99.9%�, from Metalli Preziosi SpA, constitutes thestarting material to produce some of the derived specimens,the thermally evaporated films. It has a red color, and itscrystalline character was checked by x-ray powder diffrac-tion.

Thermal evaporation has been performed on the quartzwindow of the optical oven Linkam TMS 600, in ambientatmosphere and also in a reducing gas mixture Ar-5% H2,but no significant difference of the outcome has been ob-served by Raman spectroscopy. In both cases the powder washeated between 200 and 250 °C, while the window wasnominally at room temperature.

Evaporation of ReO3 powder on glasses and on ITO-coated glasses, later used for electro-optical test in liquidcrystal cells, have been performed by keeping the sublimat-ing powders and the substrate within an oven, kept at con-stant temperature of 210 °C, for variable times, up to about24 h.

For another set of samples, rhenium oxide thin filmswere deposited on glass substrates by reactive magnetronsputtering in a plasma-focusing dc magnetic field at a dis-charge power of 100 W. Metallic rhenium �99.99%� plateswere used as sputtering targets. A gas mixture of argon andoxygen was used as sputter atmosphere.

The argon partial pressure was set at 0.040 Pa during thefull pumping step, before discharge, while the oxygen partialpressure was set at 0.0067 Pa, giving an O2 /Ar ratio of about17%. The working pressure in the chamber during the sput-tering process was increased up to about 4 Pa. The distancebetween the target and the substrate was 8 cm. The filmthickness was in the range 400–1000 nm.

The asymmetric cells of NLC were realized by using astandard sandwich configuration, locked by metallic clamps;ITO-coated glasses were used in NLC cells as counterelec-trode with respect to the electrode covered by a rheniumoxide thin film deposited by thermal evaporation, playing therole of working electrode. After a careful cleaning in chromicmixtures and repeated cleansing with acetone, the counter-electrodes were covered with polyimmide and underwent arubbing process, to ensure a better planar alignment of theNLC molecules. For the working electrodes, on the contrary,no surface treatment has been performed because the rectifi-cation effect is supposed to be related to the ionic chargedistribution and motion at the oxide-liquid crystals interface.

Thus, the insertion of an alignment layer could stronglymodify the wanted phenomena. Moreover, the rhenium oxidelayer induced a homogeneously planar alignment of the liq-uid crystal molecules in all the prepared cells. The thicknessof the cells was ensured by stripes of Mylar �8 mm�, and thefinal value was deduced by analyzing the interference pat-terns in the transmittance spectrum of the empty cell, givenby a spectrophotometer.

The introduction of the liquid crystal in the space en-closed between the asymmetric glass plates was done veryslowly, to prevent any orientational alignment induced by theflow. The cell was filled with a NLC called BL001 by Merck�former E7�.

B. Characterization techniques

Structural phase analysis of the films was performed byXRD technique using PANalytical X’Pert PRO diffracto-meter, working in the Bragg–Brentano “�-�” configuration.Conventional x-ray tube with Cu anode, operated at 45 kVand 40 mA, was used as an x-ray source.

The vibrational properties of evaporated films were char-acterized by micro-Raman spectroscopy, taking into accountthe visual map and the Raman spectral map of the depositedfilms. A microprobe Horiba-Jobin-Yvon Labram was used,equipped with a charge coupled device detector, thermoelec-trically cooled. The low frequency detection limit, due to thenotch filter, was at about 200 cm−1. In all the experiments a50� Mplan Olympus objective with numerical aperture of0.70 was used. The power of the He–Ne laser �632.8 nmemission� at the exit of the objective was about 5 mW andthe laser spot size was about 2–3 �m. To avoid unwantedlaser-induced transformations, neutral filters of different op-tical densities �ODs� were used, usually OD=2 and OD=1.Electro-optical response of the NLC cells has been measuredas the transmitted light intensity through a crossed polarizersmicroscope, equipped with a photodiode for light intensitymeasurement. Transmittance of the cells has been studied forboth broad spectrum white light and He–Ne red laser line.

114904-2 Cazzanelli et al. J. Appl. Phys. 105, 114904 �2009�

Finally, measurements of the weak currents passing inthe asymmetric NLC cells have been carried out, by analyz-ing the voltage drop on a load resistance inserted in the cir-cuit. These current data, combined with the correspondingapplied external voltage, provide some cyclic voltammetryplot.

One connector acts as working electrode, while thecounterelectrode has been short circuited with the referenceone. In all electrical and electro-optical measurements on theliquid crystal cells, the polyimmide-coated electrode wasgrounded, so that the phase of the electric voltages applied tothe cell refers always to the ReO3-coated electrode.

III. RESULTS AND DISCUSSION

A. Films deposited by thermal evaporation

When the temperature increased above 200 °C, strongsublimation occurs for the ReO3 powders in air, inside anoptical oven for micro-Raman measurements, producing adeposition of thin films on the quartz window, which ismaintained in contact with the ambient atmosphere. Theabove said temperature threshold corresponds to thatreported15 for the hydrogen emission of the intercalatedHxReO3. The deposition of rhenium oxide film on a quartzsubstrate, in contact with outer atmosphere, has been alsoperformed by heating the powders in reducing atmosphere of5% H2 in argon. Thin films have been also deposited byevaporation on substrates of pure glass and ITO-coatedglasses, maintained at high temperature �210 °C� as the sub-limating powders, in air, within a great oven. The arrange-ment of the substrate, covering the crucible containing thepowders, allows to reach a high vapor pressure of rheniumoxide on the substrate, enough to obtain a stable depositionof a film, while in previous works9 such possibility was ex-cluded.

Typical optical microscopy images of the deposited filmsbelonging to the kinds listed above are shown in Figs.1�a�–1�d�. Raman spectra collected from the evaporated filmare shown in Figs. 2�a�–2�d�, corresponding to the images ofFigs. 1�a�–1�d�, respectively.

In the case of films deposited by evaporation in con-trolled Ar–H2 mixture �reducing environment�, Raman spec-tra are collected through the quartz window �Fig. 2�b��; theirspectral pattern is about the same as for the films evaporatedin air �Fig. 2�a��, with the remarkable difference that no sig-nal from carbon is observed �D band and G band, at 1370and 1600 cm−1, respectively�. In both the cases �Figs. 2�a�and 2�b��, the Raman bands of rhenium oxide are quite simi-lar to those of finely ground powders,20 showing remarkablesharp peaks at 240, 350, 470, and 990 cm−1, probably due totetrahedral surface species.

Rhenium oxide films deposited on glass substrate at hightemperature �210 °C� exhibits on the contrary a Ramanspectrum �Fig. 2�c�� more similar to that of the as-receivedcommercial powders, which has been attributed to the stan-dard bulk structure of ReO3 with some surface disorder.20

In the case of film deposition on ITO-coated glass, only

the strongest band of ReO3 are observable, as weak struc-tures superimposed to the substrate spectrum; this fact indi-cates a smaller thickness of the film spectrum, due to reduceddeposition times. It is interesting that such very thin films,when tested in NLC cells, induce in any case the expectedrectifying effect on the electro-optic response of the liquidcrystals.

(a)

(b)

(c)

(d)FIG. 1. �Color online� Optical microscopy images of �a� films grown byevaporation in air, on quartz; �b� films grown by evaporation in reducingmixture Ar:5%H2, on quartz, observed through the quartz window; �c� filmsgrown by evaporation in air, on glass, being substrate and powders at thesame temperature of 210 °C �the diagonal stripe of different color is ascratch in the ReO3 film�; and �d� films grown on ITO-coated glass, with thesame conditions as in �c�.

114904-3 Cazzanelli et al. J. Appl. Phys. 105, 114904 �2009�

A good additional evidence of the structural modifica-tions of the evaporated films with respect to the bulk ReO3 isgiven by an XRD pattern quite similar to that reported17 forH0.57ReO3, as shown in Fig. 3.

However, in many optical microscopy images �see, forexample, Fig. 4� small microcrystals are well observable,having different optic properties with respect to the back-ground. Thus, two basic kinds of solid phases are obtainedby the evaporation of rhenium oxide powders: for most ofthe deposited surface a defective, hydrogen-containing phasebased on corner-sharing octahedral ReO6 is found, while theother solid phases consist of rectangular and trapezoidalcrystals, giving Raman spectra with sharp peaks, quite simi-lar to that of perrhenates or ReO4

− ion in solution.21,22

A typical Raman spectrum collected from these crystalsis shown in Fig. 5. The strongest peak at 961 cm−1 corre-sponds to nondegenerate v1 mode �symmetric stretching� ofthe tetrahedral ReO4

− ion, while the other sharp peaks can beassigned to the other characteristic mode of a tetrahedralgroup: v2 at 337 cm−1, v4 at 375 cm−1, and two separatedcomponents of v3 at 891 and 928 cm−1. Some significantvariation of these frequency values can be measured forother crystals observed in the deposited films, depending onthe effect of different crystal fields on the frequency and thesplitting of the degenerate modes. The existence of differentcrystal phases based on perrhenate ions has to be consideredtoo. Obviously, for most of the films surface, the observedRaman spectrum results a sum from the red background andthe white microcrystals contributions. Some more Raman

(a)

400 600 800 1000 1200 1400 1600

Intensity(arb.u.)

Raman shift (cm-1)

243

345

470

740

890

990

1570

1300

(b)

400 600 800 1000 1200 1400 1600

Intensity(arb.u.)

Raman shift (cm-1)

350

470

240 750

885

992

(c)

400 600 800 1000

Intensity(arb.u.)

Raman shift (cm-1)

240

340

470

720

990

(d)

400 600 800 1000

Intensity(arb.u.)

Raman shift (cm-1)

348 470

9451000

FIG. 2. Raman spectra coming fromthe different films shown in Fig. 1, inthe same order: �a� films grown byevaporation in air, on quartz; �b� filmsgrown by evaporation in reducingmixture Ar:5%H2, on quartz, col-lected through the window, inside thechamber with controlled atmosphere;�c� films grown by evaporation in air,on glass, with substrate at 210 °C; and�d� films grown by evaporation in air,on ITO-coated glass, for shorter times,with the peaks assigned to rheniumoxide pointed by arrows. Peak fre-quency values in cm−1 are written forthe main bands.

10 20 30 40 50 600

100

200

300

400

500ReO3 evap. in air (present work )H0.57ReO3 powders (from ref. 17 )

Intensity(arb.u.)

2θ (degrees)

FIG. 3. Comparison of XRD data for our evaporated films �dotted line� andfor powders undergoing slow hydrogen intercalation and consequent struc-tural change �from Ref. 17�.

FIG. 4. �Color online� Optical microscopy image of rhenium oxide film,deposited by evaporation, including white trapezoidal monocrystals.

114904-4 Cazzanelli et al. J. Appl. Phys. 105, 114904 �2009�

and XRD studies performed with equipment of high spatialresolution would be necessary for further investigations ofsuch phases.

B. Films deposited by sputtering

Microscopic imaging and micro-Raman spectra havebeen performed also on films grown by sputtering of metallictargets. A typical microimage is shown in Fig. 6, for a filmdeposited on a pure glass substrate, without any heating.

As for films deposited by thermal evaporation, somechange from point to point is observed, but the two basicclasses of solid phases are observed. Representative Ramanspectra of a sputtered film on glass, coming from the back-ground between crystals �Fig. 7�a�� and from the crystals�Fig. 7�b�� can be compared.

These spectra have been collected by opening the spec-trometer slits and using high laser power, so that the intensityis much higher than in other spectra here reported, while thelinewidths of the peaks appear also much broader; however,the basic information they provide is about the same. Thebackground covering most of the film �Fig. 7�a�� shows a

broad band spectrum �main bands at 335 and 765 cm−1�,typical of a disordered, defective solid phase based oncorner-sharing ReO6 octahedra, maybe with some surfacespecies �peaks at 972 and 470 cm−1�. Moreover, many dis-tinct crystals can be observed, having a structure based onReO4 tetrahedral groups, which generates the typical Ramanspectrum of narrow strong peaks. The intensity ratio and thespecific frequency values here observed are somewhat differ-ent from those observed in the crystals grown by evapora-tion. In particular, in the case of spectrum from crystals �Fig.7�b��, a strong peak is observed at 340 cm−1, with a shoulderat 315 cm−1, for the low frequency bending mode, a doubletat 900 and 922 cm−1 is assignable to �3 antisymmetricstretching, and, finally, a peak at 990 cm−1 is assignable tothe symmetric stretching mode. These different values sug-gest the occurrence of a different crystal phase, based in anycase on tetrahedral ReO4 units.

C. Application to liquid crystal cells

After the discovery of the rectifying effect of a tungstentrioxide layer inserted into an asymmetric NLC cell,18 sev-

400 600 800 1000 1200 1400 1600

Intensity(arb.u.)

Raman shift (cm-1)

337

375

961

891

928

FIG. 5. Raman spectrum collected selectively from the crystal in the centerof Fig. 4. Peak frequency values in cm−1 are written for the main bands.

FIG. 6. �Color online� Optical microscopy image of ReO3 film deposited onpure glass, by reactive magnetron sputtering: single microcrystals are scat-tered on a homogeneous background.

(a)200 400 600 800 1000 1200

Intensity(arb.u.)

Raman shift (cm-1)

265335

470765

920

972

(b)200 400 600 800 1000 1200

Intensity(arb.u.)

Raman shift (cm-1)

990

922900

340

315

FIG. 7. Raman spectra from the sputtered film on pure glass: �a� spectrumcollected from zones between the crystals, showing broad bands and �b�spectrum from the crystals, with sharp peaks. Peak frequency values in cm−1

are written for the main bands.

114904-5 Cazzanelli et al. J. Appl. Phys. 105, 114904 �2009�

eral investigations have been carried out on various metaloxides, containing some amount of mobile protons as a con-sequence of the deposition process. Thus it is interesting totest the response of electrode coated by rhenium oxide,knowing the presence of mobile protons into the solid phasesof the film. As in other previously studied oxides,18,19 theapplication of a low frequency electric field �square waveshaped�, perpendicular to the electrode-liquid crystal inter-face, induces an asymmetric optical switching, so that theelectro-optic response of the NLC layer has the same fre-quency of the applied voltage and about the same shape, as itcan be seen in the plot shown in Fig. 8�a�. The usual electro-optic response of a symmetric NLC cell, shown for compari-son in Fig. 8�b�, exhibits, on the contrary, a modulation atdoubled frequency with respect to applied voltage and aquite different shape, short pulses instead of square waves. Inthese experiments, the light transmitted by the cell in acrossed-polarizer configuration was measured by a photodi-ode detector. The asymmetric response does not depend onthe thickness of the LC layers, while it depends on the am-plitude and frequency of the perturbing electric field.

In trying to understand this behavior, current versus ap-plied voltage measurements have been made, to investigatethe possible electrochemical phenomena associate to the

presence of the mixed conductor ReO3. In Fig. 9�a� the timedependence of the very weak current measures across thecell, compared with the triangular wave of the applied volt-age, are shown. By combining current versus voltage data, acyclic voltammetry plot is obtained, shown in Fig. 9�b�. It isclear from the shape of this voltammogram that the behaviorof ReO3 film is non-Ohmic. The asymmetric shape is similarto that provided by the WO3 films inserted in the cell. Thislast evidence seems to support an explanation of the unusualswitching response of the cell as due to a reverse internalelectric field. It could be associated with a deintercalation ofsmall ions, coming from the rhenium trioxide layer and mi-grating toward the ITO electrode during the anodic phase.18

IV. CONCLUSIONS

Deposition of rhenium oxide thin films has been per-formed by thermal evaporation of commercial powders ofReO3 and by reactive magnetron sputtering of metallic Retarget.

Micro-Raman spectroscopy indicates that such filmscontain different solid phases, some of them having disor-dered structures, with an appreciable amount of intercalatedhydrogen ions. More careful structural investigations areneeded to know the crystallographic details of such solidphases; however, two basic classes have been identified.

�i� HReO4 based crystals, transparent, insulating, and notvery reactive with atmospheric gases, showing Raman

(a)0 2 4

-6

-4

-2

0

2

4

6

t(sec)

Appliedvoltage(V)

0.00

0.02

0.04

0.06

0.08

0.10

0.12

Electroopticresponse(arb.u. )

(b)-1 0 1 2

-10

-5

0

5

10

t (s)

AppliedVoltage(V

)

0.04

0.06

0.08

0.10

0.12

ElectroopticalResponse(arb.u.)

FIG. 8. Electro-optic response vs time of NLC cells: �a� asymmetric cell,with a film of ReO3 deposited on one of the electrodes and �b� usual sym-metric cell, with both the ITO electrodes coated by a surfactant. In bothcases the amount of transmitted light through a crossed polaroids micro-scope, measured by a photodiode �triangles, right scale�, is plotted as func-tion of the time, together with the applied external voltage, square waveshaped �dashed line, left scale�.

(a) 0 20 40 60 80 100

-10

-5

0

5

10

t (s)

Appliedvoltage(V

)

-8

-6

-4

-2

0

Current(μA)

(b)-10 -5 0 5 10

-8

-7

-6

-5

-4

-3

-2

-1

0

1

Current(μA)

Applied voltage (V)

FIG. 9. �a� Current �solid squares, right scale� and applied voltage �triangu-lar shaped, dashed line, left scale� vs time, for the asymmetric NLC cell,with a film of ReO3 deposited on one of the electrodes; �b� cyclic voltam-metry curve derived from the combination of current and voltage datashown in the plot �a�.

114904-6 Cazzanelli et al. J. Appl. Phys. 105, 114904 �2009�

spectra characterized by narrow peaks, assignable tothe internal modes of ReO4 tetrahedra.

�ii� HxReO3 compounds based on corner-sharing ReO6

octahedra, more or less distorted by the hydrogen in-tercalation. For deposition on high temperature sub-strate �about 200 °C�, the films show a deep red colorand metal-like electronic conductivity, in agreementwith the resistivity data reported for ReO3 films an-nealed at 200 °C;9 moreover, a simple Raman spec-trum characterized by a broad band at about700 cm−1 can be found. On the contrary, for deposi-tions on substrates at lower temperatures, variablecolor films are observed, from deep blue to transpar-ent, and the Raman spectra exhibits both broad bandsand narrow peaks, corresponding to vibrations of oc-tahedral and tetrahedral units.

The actual films contain these two components in differ-ent amounts and with different textures. However, these re-sulting mixed films have a combination of electronic andionic conductivities able to give good results as rectifyinglayers in NLC cell, comparably to films of tungsten trioxide,well known also for its electrochromic applications.

AKNOWLEDGEMENTS

Special thanks are deserved to Tiziana Barone andGiuseppe De Santo for their help in the evaporation of ReO3

films and the Raman measurements. A.K. would like tothank the University of Trento and the CeFSA Laboratory ofITC-CNR �Trento� for hospitality and financial support. Thisresearch was partly supported by the Latvian GovernmentResearch Grant Nos. 05.1714 and 05.1717.

1J. E. Jorgensen, J. D. Jorgensen, B. Batlogg, J. P. Remeika, and J. D. Axe,Phys. Rev. B 33, 4793 �1986�.

2P. B. Allen and W. W. Schulz, Phys. Rev. B 47, 14434 �1993�.3T. P. Pearsall and C. A. Lee, Phys. Rev. B 10, 2190 �1974�.4C. N. King, H. C. Kirsch, and T. H. Geballe, Solid State Commun. 9, 907�1971�.

5T. Tanaka, T. Akahane, E. Bannai, S. Kawai, N. Tsuda, and Y. Ishizawa, J.Phys. C 9, 1235 �1976�.

6C. G. Granqvist, Handbook of Electrochromics �Elsevier, New York,1995�.

7M. Ishii, T. Tanaka, T. Akahana, and N. Tsuda, J. Phys. Soc. Jpn. 41, 908�1976�.

8M. Ghanashyam Krishna and A. K. Bhattacharya, Solid State Commun.116, 637 �2000�.

9M. Ohkubo, K. Fukai, M. Kohji, N. Iwata, and H. Yamamoto, Supercond.Sci. Technol. 15, 1778 �2002�.

10J. B. Goodenough, J. Appl. Phys. 37, 1415 �1966�; Prog. Solid StateChem. 5, 145 �1971�.

11A. Fujimori and N. Tsuda, Solid State Commun. 34, 433 �1980�.12J. D. Axe, Y. Fujii, B. Batlogg, M. Greenblatt, and S. Di Gregorio, Phys.

Rev. B 31, 663 �1985�.13E. Suzuki, Y. Kobayashi, S. Endo, and T. Kikegawa, J. Phys.: Condens.

Matter 14, 10589 �2002�.14J. E. Jorgensen, W. G. Marshall, R. I. Smith, J. Staun Olsen, and L.

Gerward, J. Appl. Crystallogr. 37, 857 �2004�.15S. Horiuchi, N. Kimizuca, and A. Yamamoto, Nature �London� 279, 226

�1979�.16P. G. Dickens and M. T. Weller, J. Solid State Chem. 48, 407 �1983�.17C. A. Majid and M. A. Hussain, J. Phys. Chem. Solids 56, 255 �1995�.18G. Strangi, D. E. Lucchetta, E. Cazzanelli, N. Scaramuzza, C. Versace, and

R. Bartolino, Appl. Phys. Lett. 74, 534 �1999�.19V. Bruno, E. Cazzanelli, N. Scaramuzza, G. Strangi, R. Ceccato, and G.

Carturan, J. Appl. Phys. 92, 5340 �2002�.20J. Purans, A. Kuzmin, E. Cazzanelli, and G. Mariotto, J. Phys.: Condens.

Matter 19, 226206 �2007�.21I. R. Beattie and G. A. Ozin, J. Chem. Soc. A 1969, 2615.22F. D. Hardcastle, I. E. Wachs, J. A. Horsley, and G. H. Via, J. Mol. Catal.

46, 15 �1988�.

114904-7 Cazzanelli et al. J. Appl. Phys. 105, 114904 �2009�