Thin Solid Films 520 (2012) 2461–2466

Contents lists available at SciVerse ScienceDirect

Thin Solid Films

j ourna l homepage: www.e lsev ie r .com/ locate / ts f

Electron back-scattered diffraction of crystallized vanadium dioxide thin films onamorphous silicon dioxide

Felipe Rivera a,⁎, Laurel Burk b, Robert Davis a, Richard Vanfleet a

a Department of Physics and Astronomy, Brigham Young University, Provo, Utah 84602, USAb University of Nebraska at Lincoln, Lincoln, Nebraska 68588, USA

⁎ Corresponding author.E-mail address: rivera.felipe@yahoo.com (F. Rivera).

0040-6090/$ – see front matter © 2011 Elsevier B.V. Alldoi:10.1016/j.tsf.2011.10.014

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 April 2011Received in revised form 25 September 2011Accepted 10 October 2011Available online 15 October 2011

Keywords:Electron Backscattering DiffractionVanadium dioxideOrientation imaging microscopyTexture

Crystalline films and isolated particles of vanadium dioxide (VO2) were obtained through solid phase crystal-lization of amorphous vanadium oxide thin films sputtered on silicon dioxide. Electron back-scattered dif-fraction (EBSD) was used to study the crystals obtained in the thin films, to differentiate them fromdifferent vanadium oxide stoichiometries that may have formed during the annealing process, and to studytheir phase and orientation. EBSD showed that the crystallization process yielded crystalline vanadium diox-ide thin films, semi-continuous thin films, and films of isolated particles, and did not show evidence of othervanadium oxide stoichiometries present. Indexing of the crystals for the orientation study was performedusing EBSD patterns for the tetragonal phase of vanadium dioxide, since it was observed that EBSD patternsfor the monoclinic and tetragonal phases of vanadium dioxide are not distinguishable by computer automat-ed indexing. Using the EBSD patterns for the tetragonal phase of vanadium dioxide, orientation maps showedthat all VO2 crystals that were measurable (approximately the thickness of the film) had a preferred orienta-tion with the c-axis of the tetragonal phase parallel to the plane of the specimen.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Vanadium is a transition metal whose oxides undergo asemiconductor-to-metal transition at some characteristic tempera-ture [1–6]. Vanadium dioxide (VO2) has been extensively studied be-cause its transition temperature occurs near room temperature, 68 °C[1,3,6,7]. In single crystals, vanadium dioxide undergoes several dras-tic, abrupt, and reversible changes in its properties during its phasetransition. These changes to its properties include: 1) A structuralchange from a low temperature semi-conducting monoclinic phaseto a high-temperature tetragonal metallic phase; 2) a resistivitychange of several orders of magnitude; and 3) a sharp change in op-tical transmittance in the infrared region. These optical and electronicproperties that vanadium dioxide exhibits due to its phase transitionincrease the potential use of this material for various technical appli-cations [5,8–15]. Several studies have focused on deposition andcharacterization of VO2 thin films on various substrates, however,the nature of the phase transition and its mechanism are still underdebate [4,6,16–24]. Changes in transition temperature, hysteresis,and sharpness of the transition have been attributed to variations instoichiometry, particle size, stress, misorientations between grains,morphological faults, and other “imperfections.” However, whilespecific bulk properties of the reported films are well characterized,

rights reserved.

characterization of their microstructure is often limited to particlesize and morphology [8,9,11,14,25–28]. We add here a detailedstudy of the microstructure of VO2 thin films on SiO2 surfaces formedby solid phase crystallization.

During the transition from the high-temperature tetragonal phaseto the low-temperature monoclinic phase, the atomic arrangementsundergo both subtle changes in position as well as a break in symme-try. The reference structure to this discussion is the higher symmetrytetragonal phase and the phase transition described is then from highto low temperature (see Fig. 1). Along the c-axis of the tetragonalphase, there is a 0.88% expansion; the vanadium atoms move frombeing collinear with a spacing of 2.85 Å to being staggered with alter-nating spacings of 3.2 Å and 2.6 Å with a 166.7° angle. Along the a-axis, there is a 0.44% contraction; the vanadium atoms again movefrom their collinear state to a staggered one with alternating spacingsof 4.3 Å and 4.8 Å at an angle of 171.5°. Due to the staggering of theatoms along the a- and c-axes, the unit cell doubles along these direc-tions. There is also a 0.44% contraction along the b-axis (to greaterprecision, this is a slightly larger contraction than in the a-axis), how-ever the vanadium atoms remain lined up and there is no doubling ofthe unit cell in this direction.

Externally applied stress on crystal grains can have impact on theobserved properties of the crystal. Phase transitions are one clear casewhere, for instance, externally applied pressures can significantly im-pact the temperature of a transition. Recent work with single-crystalbeams of vanadium dioxide suggests structural phase changes due toexternal stresses [8,15,19,29–31]. In the case of thin films, the

Fig. 1. Diagram of the structural changes during the phase transition (not to scale)along the respective (A,B,C) crystal directions of the tetragonal phase. The solid circlesrepresent the positions of the vanadium atoms in the high-temperature tetragonalstructure, and the white circles represent the new positions in the low-temperaturemonoclinic structure. The axis labels correspond to the tetragonal phase.

2462 F. Rivera et al. / Thin Solid Films 520 (2012) 2461–2466

interactions between grains and with the substrate could either aid orhinder the changes expected from the phase transition and thus po-tentially shift the temperature of the local phase transition. In thecase of VO2 where each crystal axis has its own direction (expansionor contraction) of change, the relative alignment of the grains mayhave impact on the observed properties.

Knowing the microstructure of the thin films may aid in under-standing the phase transition in VO2 thin films. Vanadium dioxide isoften deposited and studied on amorphous SiO2-type substrates,such as silica or glass, but no detailed studies of crystal orientationhave been done [5,9,10,13,25,27]. Preferred orientations of the grainson these type of substrates may create the type of stresses that affectthe phase transition of individual vanadium dioxide grains within thefilm. In this work, solid phase crystallization conditions of VO2 thinfilms were explored and film morphology and grain orientationwere characterized by scanning electron microscopy and electronback-scattered diffraction.

2. VO2 film formation and morphology

The samples consisted of a silicon wafer with a thermally grownamorphous silicon dioxide layer approximately 380 nm in thickness.A layer of amorphous vanadium oxide (VOx), approximately 50 nmin thickness, was sputtered on top of the thermally grown oxide bymeans of reactive DC magnetron sputtering. Smaller pieces of thiswafer were then cleaved and annealed under various conditions. ALindberg-Blue 800 W tube furnace with a 1 in tube diameter wasused to perform the anneals. Individual samples were annealed attemperatures ranging from 200 °C up to 1000 °C under argon gasflow at approximately 600 sccm. The annealing times ranged from17 h down to 5 min. All the anneals were performed under atmo-spheric pressure and given 10 min before heating for the argon topurge the furnace.

A Philips XL30 S-FEG scanning electron microscope (SEM) wasemployed to determine the morphology of the films after each

anneal. Using micrographs obtained from the SEM the particleswere measured. Due to the irregular shape of most grains, the longestdimension was measured and used to compare grain sizes. Surfacescanning electron micrographs were obtained with accelerating volt-ages ranging from 10 kV up to 25 kV.

After annealing, granular crystalline thin films, crystalline semi-continuous thin films, and crystalline isolated particles were observed(Fig. 2). Enough surface mobility and VO2 crystal growth atop the sil-icon dioxide layer were observed at temperatures above 500 °C suchthat isolated grains taller than the original film were formed. Temper-atures ranging from 400 °C to 500 °C also showed nucleation andgrowth of crystals, but the films were more continuous. In this latterrange, higher temperatures again showed higher mobility on theSiO2 surface forming semi-continuous thin films as low as 440 °C,while temperatures below 440 °C produced continuous crystallinethin films.

A summary of the annealing results is shown in Table 1. Therewere no electron back-scattered diffraction patterns observed withtemperatures lower than 400 °C with the use of the SEM. However,work with a FEI Technai F20 (200 kV field-emission) transmissionelectron microscope (TEM) shows nano-crystallites beginning toform for anneals as low as 350 °C.

3. Sample characterization through electronback-scattered diffraction

The annealed samples were identified and differentiated usingelectron back-scatter diffraction. The EBSD system consisted of aTSL/OIM detector with a 1024 by 1024 pixel CCD camera. The pat-terns acquired by the CCD are automatically indexed with the OIMsoftware (TSL OIM Data Collection V. 4.6). The EBSD patterns corre-sponding to the crystal structures of different vanadium oxides (VO,V2O3, VO2 monoclinic, VO2 tetragonal, V2O5) were calculated by theOIM software from the material's known structural parameters. Dif-ferentiation among the different vanadium oxides was initially per-formed as a test of this technique. Commercial powder samples ofVO2 and V2O5, obtained from Alfa Aesar, were used as control sam-ples. EBSD patterns were acquired from these control samples andcompared to calculated patterns. The OIM software correctly indexedand identified the structures for the commercial powders. This wasperformed to verify that the patterns calculated from the known ma-terials' structures would match those patterns acquired from the ac-tual samples.

Using a scanning electron microscope, with an accelerating volt-age of 25 kV, and a 70° incident angle, EBSD patterns were obtainedfrom the individual particles and thin films produced from theannealed samples described above. A representative EBSD patternobtained from the films is shown in Fig. 3A. The diffuse bandsshown in the pattern indicate the crystallinity of the spot beingprobed and the symmetry of those bands reveal the structure and ori-entation of the crystal. The degraded aspect of the EBSD patternshown in Fig. 3A can be attributed to the low signal to noise ratiofrom probing a small grain (approximately 300 nm long, and 60 nmdeep), and with it, associated problems such as grain boundaries,twinning, sub-grain boundaries, and other various strains in the lat-tice [32–34]. It was observed that the monoclinic and tetragonalEBSD patterns are too symmetrically similar for the software to dis-tinguish between them. Fig. 3B–D shows the same diffraction patternoverlaid with the calculated patterns from the tetragonal and two(M1 and M3 structures) of the monoclinic phases of vanadium diox-ide. As observed from Fig. 3a–d, the bands shown in the acquiredEBSD pattern are present in all three phases (tetragonal, M1, andM3) of VO2.

The locations of the calculated bands, their widths, and the anglesbetween these three phases of VO2, are too close to the acquired EBSDpattern that the OIM software cannot accurately differentiate among

Fig. 2. Scanning electron micrographs of: A) Anneal performed at 1000 °C for 6 min. (Plan View) B) 500 °C for 1 h. (Image taken at a 70° stage tilt) C) 450 °C for 12 h. (Plan View) D)440 °C for 12 h. (Plan View).

2463F. Rivera et al. / Thin Solid Films 520 (2012) 2461–2466

them. Furthermore, the monoclinic phases of vanadium dioxide (M1and M3 shown in Fig. 3C and 3D respectively) “suffer” from a pseu-do-symmetry to which the OIM software invariably gives a confi-dence index of zero since it cannot determine the orientation of thecrystal of the monoclinic phases. This inability to discern betweenthe phases of vanadium dioxide occurred in the observations madefrom both the commercial powder samples and the crystalline thinfilm samples. Nonetheless, EBSD quite accurately differentiated be-tween the structures of different stoichiometries in the vanadium

Table 1Average particle size resulting from annealing process in argon. The following labelsare used: NDP – “No Diffraction Patterns” observed – there was no evidence of EBSDpatterns; IC “Inconclusive” – there was evidence of EBSD, but the patterns could notbe adequately indexed; IP – “Isolated Particles”; LIP – “Large Isolated Particles”; finally,a “dash” (−) is placed for those conditions not tested.

Tmp. °C ~10 min ~30 min ~1 h >10 h

b300 – – – NDP400 – – NDP 50 nma

440 NDP NDP IC 100 nm450 NDP IC 200 nm 250 nm500 300 nm 500 nm (IP) – –

680 500 nm (IP) 500 nm (IP) – –

1000 LIP>1 μm LIP>1 μm – –

a The 400 °C film yielded EBSD patterns, but the quality of the patterns obtained wasquite faint and not easily indexed, thus, it was not possible to map the surface of thefilm.

oxides as their crystal structures are remarkably different. It wasfound that the post anneal in an argon environment at temperaturesabove 400 °C consistently produced crystalline vanadium dioxide, asdetermined by EBSD patterns.

It was further observed that the samples annealed at temperaturesnear 400 °C and below did not produce discernible EBSD patterns.However, TEM studies do show crystallites forming as low as350 °C. Crystallites in the order of 20 nm form in this range of temper-atures (350 °C to 400 °C) when annealed for times longer than 12 h(since the deposited film is ~50 nm thick, this put several grainsacross the thickness of the film). At 400 °C EBSD patterns were ob-served but were too weak to index. However, TEM studies with elec-tron diffraction and imaging confirmed the VO2 crystallinity of thefilms annealed at 400 °C and showed that the films had crystal grainsapproximately 50 nm in size covering the thickness of the film. There-fore, the inability to acquire EBSD patterns from these finer-grainedfilms was attributed to a combination of surface roughness, smallgrain size, and specimen drift rather than lack of crystallinity. In thisstudy, crystals greater than 100 nm in length and width with 50 nmthickness produced indexable EBSD patterns. This last observation in-dicates that the use of the EBSD technique for determining crystallin-ity of VO2 thin films is limited to films with particles larger than50 nm in diameter and covering the thickness of the film.

To further confirm the VO2 identification done using EBSD, wemeasured the resistance of the resulting thin films as a function oftemperature. The temperature–resistance measurements were

Fig. 3. A) Acquired EBSD pattern (with the background subtracted) obtained from a sample annealed at 1000 ° C. This pattern is representative of all the annealed samples. Thereflections in the EBSD pattern have been enhanced for clarity on the right side of the figure. The grainy appearance of the EBSD pattern is attributed to the (~200 nm) size ofthe grain as well as the digital background subtraction during acquisition. This same pattern could be indexed to all three possible phases: B) VO2 tetragonal phase; C) VO2 mono-clinic (M1) phase; and D) VO2 monoclinic (M3) phase.

2464 F. Rivera et al. / Thin Solid Films 520 (2012) 2461–2466

carried out by heating the samples in a resistively-heated stage whilesimultaneously monitoring and recording the temperature and resis-tance. The temperature measurements were obtained by using a ther-mocouple placed on the film, while the resistance was obtained witha 4-point-probe measurement. A resistance change of ~2.5 orders ofmagnitude was observed near the accepted transition temperatureof 68 °C. As shown in Fig. 2, the granular nature of the grains seemsto indicate that the electrical switching was dominated by percolationin the transition of individual particles [3,35]. This change in resis-tance was in good agreement with results, in comparable films, previ-ously obtained by different groups [1,9,35].

4. Orientation imaging microscopy

During EBSD characterization of the films and particles, it was ob-served that the particular patterns shown in Fig. 3A were repeatedthroughout the films and particles. These patterns remained consis-tent “in shape” differing mainly through a rotation or a slight transla-tion. The repeatability of these patterns indicated a preferredorientation for the thermally grown vanadium dioxide crystals,

despite the amorphous nature of the substrate. To quantitativelycharacterize grain orientation, orientation imaging microscopy(OIM) was performed. OIM was done by scanning the surface of thesample, collecting EBSD patterns throughout, and indexing the orien-tation of each pattern. In this manner, a map of the surface is createdcontaining information regarding crystal structure, grain size, and ori-entation of individual crystalline grains. From the data displayed inFig. 3A and from our tests with the commercial powders, it was deter-mined that EBSD yields information that allows the OIM software todistinguish between the different structures of the varied vanadiumoxides, but not between the monoclinic and tetragonal phases ofVO2. Although the structure of VO2 at room temperature is monoclin-ic, because the two phases cannot be distinguished and because theanneal temperatures were well above the monoclinic to rutile transi-tion temperature, we used the tetragonal (rutile) structure for index-ing and labeling the OIM results. In this case we are labeling accordingto the structure present while the nucleation and growth is takingplace. OIM data was acquired using the TSL OIM Data Collection V.4.6 software package and analyzed using the TSL OIM analysis V. 5.3software.

Fig. 5. Orientation density plot from ~300 grains of a sample annealed at 450 °C for17 h. The crystal directions plotted are A) [001], B) b100> and C) b110> for the tetrag-onal structure of VO2. (A) shows a texture with a preferred orientation for the growncrystals of VO2 with the 001 axis parallel to the substrate. D) Inverse pole figure show-ing a density plot of the crystal orientations normal to the surface of the sample.

2465F. Rivera et al. / Thin Solid Films 520 (2012) 2461–2466

Fig. 4 shows an orientation map for a sample annealed at 450 °Cfor 17 h. Orientation maps color-code the crystal direction that is nor-mal to the surface of the sample, as well as misorientations betweengrains. The distribution of colors is indicative of a particular orienta-tion. The map shown in Fig. 4 is shaded according to the image qualityof the EBSD pattern obtained, thus a brightly colored area representsa clear EBSD pattern. For all the samples mapped with OIM, themajority of the indexed crystals did not have the b100> directionnormal to the surface. In the raw data some spots indexed with the[001] direction normal to the surface (shown as red in Fig. 4), howev-er the confidence of the indexed patterns was quite low (often zeroconfidence). It was observed that in these maps there were no crys-tals oriented with the [001] direction normal to the surface with aconfidence greater than 0.1. This map is representative of the dataobtained for anneals ranging from 440 °C up to 1000 °C where graingrowth was observed. This map indicates that vanadium dioxide crys-tals prefer to form with the c-axis of the tetragonal phase parallel tothe amorphous silicon dioxide surface. The orientation of the grainswas further studied in the samples and an orientation density plotis plotted in a pole figure (Fig. 5).

In a pole figure, three directions are represented: The labels (RD),(TD), and (ND) refer to directions relative to the sample's experimen-tal placement. The Normal Direction (ND) is perpendicular to thesample and is represented at the center of the plot; the Reference Di-rection (RD) refers to the direction along the surface of the samplethat faces the detector; finally the Transverse Direction (TD) is the di-rection normal to RD and ND. A particular crystal direction is thenplotted in the pole figure with respect to the sample directions.

Fig. 5 is an orientation density plot from a ~20 μm²(2.9 μm×7.1 μm) scan covering approximately 300 independentgrains from a sample annealed at 450 °C for 17 h. The orientationdensity was plotted in a pole figure. This particular plot was obtainedfrom a sample similar to the one shown in Fig. 2C. The color scale rep-resents the distribution of crystal orientations. Fig. 5A shows the lackof grains with the [001] crystal direction normal to the surface of thesample and shows them to be found evenly along directions parallelto the plane of the substrate. This is shown by low concentration of

Fig. 4. A) Orientation map from a sample annealed at 450 °C for 17 h. B) The color-codeof the grains represents the crystal direction that is normal to the surface of the sample.The brightness of each data point corresponds to the image quality of the EBSD patternindexed. C) Grain misorientation boundaries are color-coded. Note: There are very fewspots labeled with the C axis normal to the surface of the sample (red color), howeverthe confidence of the match for these spots is zero.

grains in the center of Fig. 5A (white area indicating no grainsfound). Fig. 5A shows that the [001] direction of the tetragonal VO2

grains is not found in any appreciable quantity within 60° from thenormal (grid set every 15°). The darker areas in Fig. 5A indicate the[001] direction concentrated within 15° from the plane. Scans fromthe other samples and areas also show this lack of grains with the[001] tetragonal direction in the center and concentrated in plane.The spotted patterns seen in the dark areas of Fig. 5A are attributedto the low number of grains in the scanned area. Similar randomizedspotted patterns appear in the other scans. Fig. 5B and c shows thatthe b100> and the b110> directions generally have an isotropic dis-tribution with a weak preference to be found normal to the surface.Fig. 5D shows the same orientation density in an inverse pole figure.This figure plots the crystal orientation within the grains that pointnormal to the surface. As mentioned before, the lack of grains withthe c-axis within 60° from normal to the surface is also evident by

Fig. 6. Fraction of data points where the c axis is closer to the normal than the givenangle. Notice that more than 90% of b001> crystal directions are within 20° of beingparallel to the surface (i.e. less than 10% is found lower than 70° from the normal).

2466 F. Rivera et al. / Thin Solid Films 520 (2012) 2461–2466

the white area of this figure; further, there is no distinct preferencefor the b100> and the b110> directions.

Fig. 6 is an attempt to make the results more quantitative andplots the fraction of the data points with the labeled direction closerto the normal direction than the given angle. Thus, for the [001] direc-tion of the tetragonal phase 90% of the observed points lie within 20°of being in-plane.

5. Discussion

Variations in stoichiometry, particle size, stress, misorientationsbetween grains, morphological faults, and other “imperfections” areoften cited [8,9,11,14,25–28] as explanations for variations in thetransition temperature and size of the hysteresis in thin films, butoften, the characterization of the film's microstructure is limited toreporting particle size dependence [11,26–28]. In some instances,when X-ray diffraction is used, it is often only employed to verifythe crystal structure matches that of vanadium dioxide, rather thanthe overall orientation of the deposited VO2 grains [8,9,14,25,29].We find that the use of amorphous SiO2 as a substrate for the solidphase crystallization of VO2 allows for the growth of stoichiometricVO2 crystals with a preferred orientation or texture: the c-axis ofthe tetragonal phase is parallel to the surface of the sample and ran-domly oriented within the surface of the sample.

Having the c-axis of the tetragonal phase parallel to the substrateposes several questions regarding the structural transition of VO2. Thec-axis (of the tetragonal phase) expands a small 0.88% on coolingthrough the phase transition. With all the c-axes in the film beingin-plane, there is a larger area expansion upon cooling than wouldhave been seen if the crystal grains had been three dimensionally ran-domly oriented. Perpendicular to that c-axis there is a smaller and op-posite 0.44% contraction that would partially compensate the changesin the c-axis. The resulting stresses on the film and variationsbetween individual grains may impact the local phase transition tem-perature. However, these expansions and contractions are quite small(less than 0.5%) and while they have potential impact on observedphase transitions, we have not seen the data that would let us decideif stresses of this small magnitude have a significant effect. Nonethe-less, recent reports on single-crystal micro-beams show evidencethat strains do have an effect on the structural transition of vanadiumdioxide and create local domains with differing structures, and as aconsequence, different electronic properties [15,31]. Since the use ofamorphous SiO2 substrates is common for the deposition of VO2

thin films, this in-plane orientation of the c-axis (in conjunctionwith particle size, grain boundaries, etc.) may help explain reportedsmall variations in transition temperature and hysteresis for similarlycomparable thin films of vanadium dioxide from different groups[1,9,35].

6. Conclusions

Electron back-scattered diffraction was shown as an effectivetechnique to characterize vanadium dioxide thin films. This techniqueis able to accurately differentiate crystalline vanadium dioxide crys-tals from other stoichiometries of vanadium oxide. StoichiometricVO2 crystalline thin films can be achieved through solid phase crystal-lization of amorphous vanadium dioxide deposited on a silicon diox-ide substrate. Annealing time and temperature give some controlover the particle size and film continuity.

This anneal process yields crystalline thin films (or isolated parti-cles) of stoichiometric VO2 with a preferred orientation or texturehaving the c-axis of the tetragonal phase parallel to the plane of thespecimen and randomly oriented within the plane of the sample.The c-axis of the tetragonal phase has a ~0.88% expansion as the

VO2 grains transition into the monoclinic phase. While the tworemaining a- and b-axes exhibit a ~0.44% contraction, these are notenough to relieve the film from stresses that may locally affect thetransition temperature.

OIM can yield information about the microstructure of the VO2 de-positions that might prove beneficial in understanding the differencesbetween the electrical and optical properties of VO2 films prepared bydifferent methods and on different substrates. However, due to therequirements for EBSD imaging, this OIM type of study was limitedto particles greater than 100 nm with a 50 nm thickness.

Acknowledgments

We would like to thank and acknowledge Kevin Coffey from theUniversity of Central Florida for providing the amorphous vanadiumoxide sputtered samples for this study and Jeffrey K. Farrer for hisguidance and insightful discussions regarding SEM and EBSD. Thisstudy was supported by Brigham Young University, along with theNational Science Foundation's Research Experience for Undergradu-ates (REU) project NSF#PHY-0552795 at BYU.

References

[1] F. Morin, Phys. Rev. Lett. 3 (1959) 34.[2] S. Kachi, T. Takada, K. Kosuge, J. Phys. Soc. Jpn. 18 (1963) 1839.[3] H. Verleur, A. Barker, C. Berglund, Phys. Rev. 172 (1968) 788.[4] D.H. Hensler, J. Appl. Phys. 39 (1968) 2354.[5] K.A. Khan, M.S. Rahman Khan, Pramana 38 (1992) 389.[6] A. Cavalleri, T. Dekorsy, H. Chong, J.C. Kieffer, R.W. Schoenlein, Phys. Rev. B 70

(2004) 161102.[7] K. Kosuge, J. Phys. Soc. Jpn. 22 (1967) 551.[8] D.H. Kim, H.S. Kwok, Appl. Phys. Lett. 65 (1994) 3188.[9] Y. Dachuan, X. Niankan, Z. Jingyu, Z. Xiulin, J. Phys. D: Appl. Phys. 29 (1996) 1051.

[10] M. Lee, Thin Solid Films 365 (2000) 5.[11] R. Lopez, L.A. Boatner, T.E. Haynes, R.F. Haglund, L.C. Feldman, Appl. Phys. Lett. 79

(2001) 3161.[12] P.P. Boriskov, A.A. Velichko, A.L. Pergament, G.B. Stefanovich, D.G. Stefanovich,

Tech. Phys. Lett. 28 (2002) 406.[13] T.D. Manning, I.P. Parkin, M.E. Pemble, D. Sheel, D. Vernardou, Chem. Mater. 16

(2004) 744.[14] W. Yin, K.G. West, J.W. Lu, Y. Pei, S.A. Wolf, P. Reinke, Y. Sun, J. Appl. Phys. 105

(2009) 114322.[15] A. Tselev, J.D. Budai, E. Strelcov, J.Z. Tischler, A. Kolmakov, S.V. Kalinin, Nano Lett.

11 (2011) 3065.[16] A. Zylbersztejn, Phys. Rev. B. 11 (1975) 4383.[17] S. Shin, S. Suga, M. Taniguchi, M. Fujisawa, H. Kanzaki, A. Fujimori, H. Daimon, Y.

Ueda, K. Kosuge, S. Kachi, Phys. Rev. B. 41 (1990) 4993.[18] R.M. Wentzcovitch, Phys. Rev. Lett. 72 (1994) 3389.[19] J.M. Gregg, R.M. Bowman, Appl. Phys. Lett. 71 (1997) 3649.[20] V. Eyert, Annalen Der Physik 11 (2002) 650.[21] C. Hébert, M. Willinger, D.S. Su, P. Pongratz, P. Schattschneider, R. Schlögl, Eur.

Phys. J. B 28 (2002) 407.[22] S. Biermann, A. Poteryaev, A.I. Lichtenstein, A. Georges, Phys. Rev. Lett. 94 (2004)

5.[23] A. Liebsch, H. Ishida, G. Bihlmayer, Phys. Rev. B. 71 (2005).[24] M. Qazilbash, K. Burch, D. Whisler, D. Shrekenhamer, B.-G. Chae, H. Kim, D. Basov,

Phys. Rev. B. 74 (2006) 5.[25] T. Maruyama, Y. Ikuta, J. Mater. Sci. 28 (1993) 5073.[26] R. Lopez, T. Haynes, L. Boatner, L. Feldman, R. Haglund, Phys. Rev. B. 65 (2002) 1.[27] R. Lopez, L.A. Boatner, T.E. Haynes, L.C. Feldman, R.F. Haglund, J. Appl. Phys. 92

(2002) 4031.[28] R.A. Aliev, V.N. Andreev, V.M. Kapralova, V.A. Klimov, A.I. Sobolev, E.B. Shadrin,

Phys. Solid State 48 (2006) 929.[29] K. Nagashima, T. Yanagida, H. Tanaka, T. Kawai, J. Appl. Phys. 100 (2006) 063714.[30] J. Cao, Y. Gu, W. Fan, L.Q. Chen, D.F. Ogletree, K. Chen, N. Tamura, M. Kunz, C. Bar-

rett, J. Seidel, J. Wu, Nano Lett. 10 (2010) 2667.[31] A. Tselev, I.A. Luk'yanchuk, I.N. Ivanov, J.D. Budai, J.Z. Tischler, E. Strelcov, A. Kol-

makov, S.V. Kalinin, Nano Lett. 10 (2010) 4409.[32] R. Keller, Microelectron. Eng. 75 (2004) 96.[33] J.K. Farrer, C.B. Carter, N. Ravishankar, J. Mater. Sci. 41 (2006) 661.[34] S.I. Wright, M.M. Nowell, D.P. Field, Microsc. Microanal. The Official Journal of Mi-

croscopy Society of America, Microbeam Analysis Society, Microscopical Societyof Canada 17 (2011) 316.

[35] B.G. Chae, H.T. Kim, Phys. B (Amsterdam, Neth.) 405 (2010) 663.