Reactive evaporation of anomalous blue V0 2

Francine C. Case

Thin films of vanadium dioxide exhibit a thermally induced semiconductor-to-metal phase transition near670 C. In most deposition conditions, the transition is accompanied by large changes in optical properties atinfrared wavelengths, but with only slight visible contrast. Recently, reactive evaporation at high tempera-tures and in stringent process conditions of so-called blue V0 2 has been reported [G. A. Nyberg and R. A.Buhrman, J. Vac. Sci. Technol. A 2, 301 (1984)]. These films exhibit a striking blue-to-red transition intransmitted light. In this paper, a new technique for thin-film growth of anomalous vanadium dioxide will bepresented. The primary steps in the process are the reactive evaporation of vanadium oxide at ambienttemperature followed by anneal in flowing oxygen. Optical and microstructural data for material depositedon sapphire and fused quartz substrates will be compared to standard vanadium dioxide.

1. Introduction

Vanadium dioxide is a phase transition materialwhich displays significant changes in infrared opticalproperties when heated through 670C.1-3 For exam-ple, Fig. 1 compares semiconductor and metal spectralcurves in the 2.5-7-jum range for a V0 2 film preparedby reactive evaporation onto an A1203 substrate held athigh temperature (480-520'C). Note that the appar-ent decrease in both semiconductor-state transmissionT and substrate-side (back) reflectance RB for X > 4.5,gm is actually due to the onset of high absorption in thesapphire substrate. Although the changes in V02 op-tical properties with temperature are quite striking forinfrared wavelengths, the material does not exhibitsuch pronounced contrast in the visible. A standardV02 film processed by thermal oxidation or reactiveevaporation appears bronze in transmitted visiblelight and any color change on heating is slight.4 Ny-berg and Buhrman5 reported recently on V0 2 filmsreactively deposited in stringent conditions at highsubstrate temperature. The films exhibit largechanges in total transmittance (directional and scat-tered) from blue to red in the visible while the reflec-tance response is relatively small. In our investigationa new technique for depositing blue V0 2 was devel-

The author is with LTV Aerospace & Defense Company, VoughtMissiles & Advanced Programs Division, P.O. Box 650003, Dallas,Texas 75265-0003.

Received 21 November 1986.0003-6935/87/081550-04$02.00/0.© 1987 Optical Society of America.

oped by which ambiently deposited vanadium oxidefilms are taken to the modified form by anneal inflowing oxygen. These films exhibit large differencesin microstructure, infrared, and visible optical proper-ties relative to those of standard V0 2; most significant-ly, there is a striking change in transmission from blueto red on heating a modified film through the phasetransition.

II. Experimental

VIOY films were reactively evaporated onto fusedquartz and randomly oriented single crystal sapphiresubstrates in an Edwards electron beam coating sys-tem using a vanadium deposition rate of 4 A/s and apartial pressure of oxygen of 1.5 X 10-3 Torr. Sub-strates were obtained from Valpey-Fisher and CrystalSystems, respectively. Pressure was controlled via anMKS pressure/flow controller. These as-depositedfilms were then analyzed with a Perkin-Elmer model580 infrared spectrophotometer. Resistivity mea-surements were made with an Alessi four-point probein conjunction with a Keithley model 225 constantcurrent source. X-ray diffraction data were obtainedusing a Siemens x-ray diffractometer equipped with aHuber grazing-angle thin-film arrangement. Thisconfiguration samples a greater effective film thick-ness and provides signal enhancement compared witha standard diffractometer. Film thicknesses weremeasured with a Tencor Alpha Step profilometer.Semiconducting state optical constants for the V0 2film were determined by a closed-form solution tech-nique which calculates n and k consistent with direc-tional front reflectance (filmside), back reflectance,transmission data, and measured film thickness. 6

Color contrast was observed with an Olympus model

1550 APPLIED OPTICS I Vol. 26, No. 8 / 15 April 1987

2.50 3.50 4.50 5.50 6.50

WAVELENGTH (MICRONS)

Fig. 1. Infrared spectral response of a standard V0 2 thin film. RF

and RB denote reflectance measured from the filmside (front) andsubstrate side (back) of the sample, respectively.

BH-2 metallurgical microscope, and film surface mi-crostructure was examined with a JEOL model 35CFscanning electron microscope. Following these pre-liminary analyses, the films were annealed in from 8 X10-4 to 1.2 X 10-3 Torr of oxygen at a temperature of575-5901C for 1-1.5 h. The annealed samples werethen reexamined using the same techniques describedabove but also including additional spectral data from870 to 250 nm using a Perkin-Elmer model 330 UV-visible spectrophotometer.

Ill. Results and Discussion

The VOy ambiently as-deposited material is char-acterized in the infrared by the spectrum shown in Fig.2. The structure observed near 3.0-Am wavelength isattributable to water absorption. Neither XRD signa-tures nor discernible surface microstructure was ob-served for these films prior to anneal. The films ap-peared smooth under SEM analysis and the lack ofXRD structure is characteristic of an amorphous ma-terial. The films are charcoal in color and exhibit highresistivities of the order of 90-120 Q cm. Comparisonto published resistivity data places the film stoichio-metry near V601 3 or V307.7

During anneal at 575-590° C in from 8 X 10-4 to 1.2 X10-3 Torr of oxygen for 1-1.5 h, the smooth, continuousfilm coalesces into large grains of the order of 1-2 um in

diameter, separated by wide grain boundaries whichappear to be devoid of material. A typical SEM micro-graph of the modified material is shown in Fig. 3 andcompared to standard V0 2 microstructure. Theemergence of a small Al signal from both AES andESCA analyses of selected samples on A12 03 substratesindicates that, at least for the thinner films of the orderof 1500 A, the grain boundaries following anneal areindeed void of evaporated material. The films nowexhibit XRD structure with characteristic V02 peaks.Figure 4 compares semiconductor and metal direction-al transmission in the 850-550-nm range for a standardand a modified 1500-A V0 2 film. An interesting pointto note from this data is that, while there is similarityin the behavior of the semiconducting-state transmis-

0.0 T

0.70

0.60

0.50

0.40

0.30

0.20 SU RF (fllmslde)

0.10 - X B

2.5 3.5 4.5 5.5 6.5WAVELENGTH (MICRONS)

Fig. 2. Infrared response for ambiently deposited vanadium oxide

film prior to anneal. The substrate material is A120 3; this material

absorbs radiation for wavelengths greater than -4.5 im.

STANDARD MODIFIED

1 pm

Fig. 3. SEM micrograph comparison of standard and modified V0 2

films at 20,000X.

60 l ,

-- STANDARD

- MODIFIED

40 _- _

20

850 750 650 550

X (NANOMETERS)

Fig. 4. Comparison of directional transmission for standard and

modified V0 2 films at 24 0C and 80'C.

sion for both film types, a striking difference is ob-served in metal-state responses. The standard mate-rial shows a drop in transmission through the phasetransition of a few percent while the transmission ofthe modified material rises by a factor of 5 at 650 nm.

The formation temperature for the modified materi-al is very critical, as the data shown in Fig. 5 illustrates.Transmission hysteresis loops and SEM micrographswere taken on a sample which experienced a thermalgradient during anneal. The film has two zones, onebronze and the other modified blue when observed at

15 April 1987 / Vol. 26, No. 8 / APPLIED OPTICS 1551

I - ---- - -

I------ - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

N1 - X BRONZE

~is; kREGION

MODIFIED

TURE SAC) i

as a function of temperature at 3.4 um fordual-zone V0 2 film.

PARTIAL ANNEAL

- i L r i

i2'. , Y i an

FULL ANNEAL

SAMPLE 1

NA1203l

20,000 X SAMPLE 2ON AJ203

SAMPLE 3ON FUSED QUARTZ

1 pm

Fig. 7. Grain development progression, partial to complete anneal.

3.50 4.50 5.50 6.50

WAVELENGTH (MICRONS)

Fig. 6. Infrared spectral response of a modified V0 2 film. RF andRB denote reflectance measured from the filmside (front) and sub-

strate side (back) of the sample, respectively.

room temperature. At 1000C the blue region hasswitched to red, while the bronze is relatively un-changed. Infrared spectral data for the blue region areseen in Fig. 6. Both semiconductor and metallic statetransmission of the blue film is characteristicallygreater than that for the standard film in Fig. 1 andmay be due to the presence of large grain boundaries.Metallic state reflectance of the blue film is reducedover that of standard metallic-phase V02. Again theA12 03 substrate cutoff is observed for X > 4.5 m.

Figure 7 compares microstructure for a film follow-ing partial anneal and full anneal times. The incom-plete anneal did not take the material to the modifiedform, and the microstructure and grain boundaries aremuch smaller than those for the final modified bluefilm. Following complete anneal, modified sampleswere examined with an optical microscope. At lowmagnification (50X) the contrast in transmission be-tween the semiconductor (blue) and metallic states(red) is striking. At 1000X magnification blue grainsare observed; as the sample is heated, the color densityof these grains changes and shifts toward a mixture ofblue and red. At 1000C the bulk, but not all the grains,appear red. These visual observations of large con-trast are not supported by UV-visible spectrophoto-metric measurements. However, the specular trans-

1 Pm

24°C 100oCFig. 8. SEM micrograph comparison of a modified V0 2 film above

and below the transition temperature at 10,OOOX.

mitted light detected by the spectrophotometer doesnot include contributions to the total transmittancefrom scattered radiation. It is likely then that thescattered component plays a dominant role in the visu-al color change, and a closer correlation between spec-tral measurements and visual observations will resultfrom planned integrated sphere measurements.

To determine if the transmittance change throughthe phase transition of modified films is due to a grossmodification in the shape or size of the large grainsinherent in the film, SEM photos were obtained of thesame region of film in both states at 20,OOOX and10,OOOX magnifications. The latter micrographs areshown in Fig. 8. No significant change in film micro-structure is observed on heating the sample through itsphase transition, and the color change does not likelyresult therefore from shifting microstructural charac-teristics but is related instead to changes in the indexof the material itself. An examination of semiconduc-tor-state spectral data at 3.4 m wavelength for bothstandard and modified V02 films results in values of n= 3.0, k = 0.03 and n = 2.4, k = 002, respectively.

IV. Summary

Modified blue vanadium dioxide has been formed bythe high temperature anneal (575-590'C) of reactivelyevaporated vanadium oxide films deposited at ambi-ent temperature. These annealed films are generallycharacterized by large grains and grain boundaries anddisplay a striking blue-to-red change in transmittancewhen heated through the phase transition. The films

1552 APPLIED OPTICS / Vol. 26, No. 8 / 15 April 1987

Fig. 5. Transmission

1.0 -

0.8 _

0.6 _

0.4 _

0.2 -

0.0 -2.50

exhibit contrast at infrared wavelengths, albeit re-duced compared to standard V0 2. The exact natureof the fundamental differences between these twoforms of the material is yet to be determined. It ishoped that detailed integrated sphere analysis andcomparison of XRD signatures will identify some ofthese differences.

The author would like to express appreciation to H.E. Ratzlaff for his many technical contributions whichhave made this work possible. Special thanks are duealso to W. A. Patey for careful sample preparation, C.J. Day for extensive optical measurements, and to J. S.Smotherman for XRD measurements.

References1. F. J. Morin, "Oxides which show Metal-to-Insulator Transitions

at t.9 Neel Temperature," Phys. Rev. Lett. 3, 34 (1959).2. J. B. Goodenough, "The Two Components of the Crystallograph-

ic Transition in V0 2," J. Solid State Chem. 3,490 (1971).3. F. C. Case, "Modifications in the Phase Transition Properties of

Pre-Deposited V0 2 Films," J. Vac. Sci. Technol. A 2,1509 (1984).

4. H. W. Verleur et al., "Optical Properties of V0 2 between 0.25 and

5 Ev," Phys. Rev. 172, 788 (1968).5. G. A. Nyberg and R. A. Buhrman, "Preparation and Optical

Properties of Reactively Evaporated V0 2 Thin Films," J. Vac.Sci. Technol. A 2, 301 (1984).

6. W. E. Case, "Algebraic Method for Extracting Thin-Film Optical

Parameters from Spectrophotometer Measurements," Appl.Opt. 22, 1832 (1983).

7. S. Kachi et al., "Electrical Conductivity of Vanadium Oxides,"

Phys. Soc. Jpn. 18, 1839 (1963).

MAPPING INFRARED ENERGY

M51 is a spiral galaxy with a twin star in its northwest region. The graceful band of stars is

serving as a test model for a process that will make it possible for the Electro-Optics

Technology Center at Tufts University to map infrared energy in the sky.

The map, which will be developed with digital image processing equipment similar to that

used by the Woods Hole Oceanographic Institute to explore the remains of the Titanic, may

point out stars and planets that have never been seen because they produce light invisible to

the human eye. Working with Mission Research Corp. the Air Force Geophysical Research

Laboratory to figure out how to create the celestial chart, Robert Gonsalves of The Tufts EO

Technology Center will use the images gathered by the Infrared Astronomical Satellite (IRAS)

in 1983.

Dutch physicists who initiated the IRAS project wanted to gather information about celestial

sources of infrared energy, which has wavelengths longer than the red end of the visible

spectrum but shorter than radio waves. The Netherlands Agency for Aerospace Programs

joined NASA and the British Science & Engineering Research Council to launch the infrared-

detecting satellite, which picked out hot spots of infrared energy in the sky. But the images it

recorded were extremely blurred. Scientists now believe the data IRAS gathered can be used

to produce a detailed image of galaxies and individual stars, and Tufts and Mission Research

Corp. will conduct research over the next three years to create that map, using a mathematical

formula Gonsalves created through a form of image processing called maximum entropy

restoration.To apply their techniques to the IRAS data, Gonsalves and his graduate students distort

known pictures the same way the IRAS detector distorted the images it collected.

Then, working backward they use a mathematical formula to change the blobs they have

created into more recognizable images. They check the images against the originals and

repeat the process. Eventually, the researchers should come up with a formula that can be ap-

plied to all the IRAS data to create a chart that will look much like a map of the constellations.

15 April 1987 / Vol. 26, No. 8 / APPLIED OPTICS 1553