Parametrization of optical properties of indium–tin–oxide thin films by spectroscopic ellipsometry: Substrate interfacial reactivity

Parametrization of optical properties of indium–tin–oxide thin filmsby spectroscopic ellipsometry: Substrate interfacial reactivity

M. Losurdo,a) M. Giangregorio, P. Capezzuto, and G. BrunoPlasma Chemistry Research Center—CNR, via Orabona, 4-70126 Bari, Italy

R. De Rosa and F. RocaCR ENEA Portici, Loc. Granatello, 80055 Napoli, Italy

C. Summonte, J. Pla, and R. RizzoliLamel Institute—CNR, via Gobetti 101, I-40129 Bologna, Italy

~Received 30 March 2001; accepted 1 October 2001!

Indium–tin–oxide~ITO! films deposited by sputtering and e-gun evaporation on both transparent~Corning glass! and opaque (c-Si, c-Si/SiO2) substrates and inc-Si/a-Si:H/ITO heterostructureshave been analyzed by spectroscopic ellipsometry~SE! in the range 1.5–5.0 eV. Taking the SEadvantage of being applicable to absorbent substrate, ellipsometry is used to determine the spectraof the refractive index and extinction coefficient of the ITO films. The effect of the substrate surfaceon the ITO optical properties is focused and discussed. To this aim, a parametrized equationcombining the Drude model, which considers the free-carrier response at the infrared end, and adouble Lorentzian oscillator, which takes into account the interband transition contribution at theUV end, is used to model the ITO optical properties in the useful UV–visible range, whatever thesubstrate and deposition technique. Ellipsometric analysis is corroborated by sheet resistancemeasurements. ©2002 American Vacuum Society.@DOI: 10.1116/1.1421596#

drsnd

le

-iea

-

ngngceond

anubIT

u-soanis

hei-

onpa-

is-

telyingalsoge

ringp-ssedthe

ed

bytures

forheaf-inimi-rate/. Ad acalachthetion

ma

I. INTRODUCTION

Indium–tin–oxide ~ITO! thin films are widely used astransparent electrodes in a large variety of optoelectronicvices including flat-panel displays, thin-film transisto~TFTs!, and solar cells, because of ITO high conductivity atransparency in the visible range.1 Therefore, accurateknowledge of ITO optical properties, such as its compdielectric function and, hence, its refractive index,n, andextinction coefficient,k, is critical for the design and performance of optical devices. Although a large number of studdeal with the effect of deposition conditions on the opticproperties of ITO films,2–6 mainly determined by transmittance measurements in the near-infrared region~0.3–2.2eV!,6,7 some important aspects have been neglected. Chaobserved in ITO transmittance curves can be due to chain film thickness, surface roughness, and to the appearana substrate/ITO interface, rather than to the ITO optical cstants themselves. Moreover, the common approach useinvestigate the ITO optical properties assumes that the trmission curves of the ITO film on transparent glass sstrates are also representative of optical properties offilms deposited on device structures, i.e., ona-Si or c-Silayers in flat-panel displays and solar cells. This is becathe absorption of thea-Si or c-Si films precludes transmission measurements. Nevertheless, this approach neglectthe different thermal coefficient and surface reactivityvarious substrates can alter the ITO optical propertiescan result in absorbing substrate/ITO interface layers. Thparticularly true for ITO deposited ona-Si:H films for solar

a!Author to whom correspondence should be addressed; [email protected]

37 J. Vac. Sci. Technol. A 20 „1…, Jan ÕFeb 2002 0734-2101Õ20

e-

x

sl

esesof-tos--O

se

thatfdis

cells and TFTs, where out-diffusion of hydrogen from ta-Si:H into the ITO film can occur during the ITO depostion. Hence, the measurement of the ITO optical responseopaque substrates and in device configuration, and therametrization of ITO optical properties are important to dcuss the effect of ITO deposition/treatment parameters.

In this context, spectroscopic ellipsometry~SE! is a non-destructive and noninvasive technique, which can complecharacterize ITO films, including optical constants, gradprofile, film thickness, and surface roughness, and cangive information on the ITO resistivity. SE has the advantaof being applicable to opaque substrates. Pioneeworks8–10 have shown the capability of spectroscopic ellisometry to measure ITO optical constants and have discuthe use of different dispersion equations, such asCauchy,9 the Lorentzian oscillator,9 and, lately, theHerzinger–Johs,10 to describe the ITO optical properties.

In this article, the optical constants of ITO films depositon different substrates, i.e., Corning glass,c-Si, c-Si/SiO2 ,and c-Si/a-Si:H heterojunctions have been investigatedspectroscopic ellipsometry. These substrates and strucare of great technological importance because of their usethe fabrication of thin-film transistors and solar cells. Tfocus is on the substrate/ITO interface reactivity whichfects the ITO optical and electrical response. Additionally,the case of ITO stacked in heterostructures, SE can discrnate, through a multilayer analysis, the presence of substITO interface layers, because of its monolayer sensitivityparametrized equation combining the Drude model andouble Lorentzian oscillator is used to derive the ITO opticonstants in the energy range 1.5–5.0 eV. This approgives a detailed description of the ITO optical response inuseful range, which extends from the fundamental absorpil:

3702Õ20„1…Õ37Õ6Õ$19.00 ©2002 American Vacuum Society

yedn

gribda

alr

a-

l

ec

0erm

un

dtiogblso

–

el-pine

th

m-on

thecs.the

-fitlmre-

theheownm-,

rn-

with

38 Losurdo et al. : Parametrization of optical properties of indium 38

edge on the UV end to the plasma absorption frequencthe IR end, where maximum ITO transparency is requirThe ellipsometric analysis is corroborated by sheet resistameasurements.

II. EXPERIMENT

ITO thin films with a thickness of;90 nm were depositedon Corning Glass 7059, Si~100!, SiO2-overcoated Si~100!substrates, andc-Si/a-Si:H heterojunctions by sputterinand e-gun evaporation according to the procedures descin Refs. 11, and 7, respectively, starting from a self-matarget of indium and tin alloy at a temperature as low260 °C. The optimization of ITO films and their electricand crystalline characterizations have been extensivelyported previously.7,11,12The c-Si/a-Si:H heterojunction wasrealized starting from ap-type Si~100! wafer on which ann-typea-Si:H layer;25 nm thick was deposited by plasmenhanced chemical-vapor deposition~PECVD! at 240 °C, us-ing SiH4 as a precursor. This heterostructure was usefuexamine the effect of hydrogen at thea-Si:H/ITO interfaceon the ITO optical response.

A phase-modulated spectroscopic ellipsometer~UVISELJobin-Yvon! is used to acquire spectra of the pseudodieltric function, ^«(v)&5^«1(v)1 i ^«2(v)&5@n(v)1 ik(v)#2 and, hence, of the refractive index,n(v), andextinction coefficient,k(v), in the spectral range of 1.5–5.eV at the incidence angle of 70.57°. Ellipsometric data wanalyzed by models based on the Bruggeman effectivedium approximation~BEMA!.13 A linear gradient of the op-tical properties was also considered to account for nonformity throughout the ITO bulk.

For a description of the ITO optical response, the Drutheory was applied to consider the free-electron contributo the ITO optical properties in the near-infrared enerrange. The Drude model was combined with a douLorentzian oscillator, which considers the interband trantion contribution to absorption, to describe the dispersionITO dielectric function«~v! in the whole energy range 1.55.0 eV, according to the equation

«~v!5«1~v!1 i«2~v!

5«`2vp

2

v21 ivtv1(

i 51

2 f iv i2

v i22v21 ig iv

, ~1!

where«` is the high-frequency dielectric constant, andvp

andvt are the plasma and collision frequency, respectivThe parametersv i , f i , and g i refer to the oscillator resonance frequency, strength, and damping constant, restively. In the fit analysis,«` was assumed to be equal to 1all cases according to the discussion reported by Palik in R14 in order to reduce the number of model parameters.Independent x-ray photoelectron spectroscopy~XPS!, sheetresistance~by the four-probe technique!, and thickness~byalpha-step! measurements were carried out to corroborateSE analysis.

J. Vac. Sci. Technol. A, Vol. 20, No. 1, Jan ÕFeb 2002

at.ce

edes

e-

to

-

ee-

i-

enyei-f

y.

ec-

f.

e

III. RESULTS AND DISCUSSION

The validity of Eq.~1! has been tested through the paraetrization of the optical properties of ITO films depositedCorning glass substrates by different techniques~e-gun andsputtering! and of a commercial ITO sample~provided byCentral Research Laboratories, Ltd., U.K.!.

Figure 1 shows calculated and experimental data ofreal,^«1&, and imaginary, «2&, parts of the pseudodielectrifunction of different ITO films deposited on Corning glasThe calculated spectra are the result of the simulation bytwo-layer model shown at the bottom of Fig. 1. The bestDrude and Lorentzian oscillator parameters of each ITO fiare reported in Table I. An oscillator with a resonance fquency (v2) outside the measurement range~at about 7 eV!has to be introduced to fit the measured dispersion ofdielectric function. A sensitivity test to the parameters of tsecond oscillator based on a multisample analysis has shthatv2 and f 2 are far less sensitive than the other fit paraeters. Hence,v2 and f 2 have been fixed to 7.1 eV and 2.3respectively, since they changed from 7.0460.09 eV to7.2160.13 eV and from 2.2060.04 to 2.3860.08 for vari-

FIG. 1. Calculated and experimental spectra of the real,^«1&, and imaginary,^«2&, parts of the pseudodielectric function of ITO films deposited on Coing glass 7059 by The Central Research Laboratories Ltd., U.K. No.~No. 1!,by e-beam~No. 2!, and by sputtering~No. 3!. A two-layer BEMA modelsketched at the bottom and consisting of a homogeneous bulk ITOsurface roughness has been used to fit the spectra.


TABLE I. Best-fit parameters~reported inside the black box! of the Drude and double Lorentzian oscillator@defined in Eq.~1! in the text#, which describes thedispersion law of ITO films of Fig. 1. The value of the optical gap,Eg , and of the sheet resistance,Rh ~measured with the four-probe technique!, are alsoreported. The table also includes the compositional ratio In/Sn and O/In1Sn determined by XPS analysis. In the fit analysis, thev2 and f 2 values have beenfixed to 7.1 and 2.3, respectively.

No.dB

~Å!dS

~Å!vp

~eV!vt

~eV!v1

~eV! f 1 g1

Eg

~eV!Rh

~V/h! In/Sn O/~In1Sn!

1a 8656

2446

11.6260.06

0.0560.01

4.8260.02

0.4760.03

0.8260.03

4.3560.02

21.7 2.360.2

1.26

0.2

2b 9546

4676

31.4260.04

0.2660.04

4.6360.02

0.2160.03

0.8660.02

4.0960.02

45.7 7.060.2

0.96

0.2

3c 7886

2576

21.6760.04

0.1560.04

4.8860.01

0.3460.02

0.8460.02

4.3860.02

23.1 8.060.2

1.66

0.2

aNo. 1 is a commercial sample supplied by Central Research Laboratories, Ltd., U.K.bNo. 2 is deposited by e-beam.cNo. 3 is deposited by sputtering.

rheonhaooorcfiee

m

n

oy

ly

ti-d

trent

et

e.

f

o. 2

y

sitye-eases

ri-

c-

fornd

ous ITO samples. This does not affect the shape of thesulting fitting curves. In addition, the layer thickness in tfirst step of the analysis has been used as a fit variable;the thickness best-fit values have been established, theybeen fixed in a successive fit analysis and the same fit gness is achieved. Hence, the thickness values are not clated to oscillator fit parameters. Also, in order to redufurther the number of fit parameter, it has been tried toeither vp or vt at a reasonable value and fitting the othparameters in the model to test the correlation among thparameters. It has been found that fixing thevp parameterdoes not affect the values of the Lorentzian oscillator paraeter, but a correlation betweenvp andvt exists. Probably, agreater sensitivity and decorrelation to the amplitude abroadening of the Drude oscillator~plasma and collision fre-quency, in this work! can be achieved by measurementsthe Drude tail farther into the infrared, which are limited bthe present ellipsometer. Therefore, the regression anawas for a five-parameter model fit~vp , vt , v1 , f 1 , andg1!. The calculated spectra of both refractive indexn andextinction coefficientk are shown in Fig. 2. The differenoptical curves correspond to ITO films with different chemcal compositions, as detected by XPS analysis, and withferent values of sheet resistance, as shown in Table I.

A correlation between the optical parameters and eleccal data can be found. In particular, the extinction coefficispectra show clearly two absorption edges, one at phoenergies below 2 eV and one above 4 eV, which are dufree carriers and valence-band electrons, respectively;former correlates tovp and vt , the latter to thev1 value,mainly. The lower value of sheet resistance~21.7 V/h! isfound for the ITO film~No. 1!, which has a high value of theplasma frequency, i.e., the carrier concentration~being vp

2

}N!,1 and the lowest value of the collision frequency, i.the highest carrier mobility~given by m5et/mc* ; t51/vt

is the relaxation time andmc* is the effective mass!. Lower-ing vt , i.e., increasingt, implies film of high quality. Forsample No. 3, the increase of thevt value is consistent withthe increase of the sheet resistance. Both the decrease oplasma frequency and the increase of thevt correlate with

JVST A - Vacuum, Surfaces, and Films

e-

ceved-re-exrfit

-

d

f

sis

if-

i-t

ontohe

,

the

the highest value of the sheet resistance of sample N~45.7V/h!.

The optical gap,Eg ~see Table I! has been determined bplotting the square of the absorption coefficient,a2, versusthe photon energy. The higherEg andv1 values for sampleNos. 1 and 3 are consistent with an increase of carrier den~Borstein–Moss shift15! and, hence, with the lower sheet rsistance values. Moreover, as the sheet resistance decrandEg ~and, hence, the carrier concentration! increases, theheight of the interband absorption increases~see Fig. 2! and,correspondingly, thef 1 values.

Thus, from the parameters of Eq.~1!, which describe theoptical response of the ITO films, information on the electcal properties can be inferred.

The same approach by Eq.~1! has been used to chara

FIG. 2. Spectra of the refractive index and extinction coefficient derivedthe three ITO films of Fig. 1 by using the BEMA model shown in Fig. 1 athe ITO dispersion law of Eq.~1! defined in the text.

tes

nn

TO

esae

eeio.

iddnd

ofwet

eaterate

ier

ive

is

ion.TOentlsomix-lyeine tothecalnt

a.t theTOonrre-

-on-

bon

-bying


terize ITO films deposited on opaque substrates and hestructures. Figure 3 shows the results of the SE analysiITO films deposited by rf sputtering on glass,c-Si ~100!, andSiO2-overcoated Si~100! substrates, in the same experimetal run. For the ITO film deposited over thermally growSiO2 , the SiO2 thickness has been measured before Ideposition by both alpha-step~210 nm! and SE (211365 Å) and fixed during the analysis of the Si–SiO2– ITOstructure in order to reduce the number of fitting variablDifferent best-fit models apply for the three samples,shown in Fig. 3. In particular, a homogeneous ITO bulk laywith surface roughness~modeled by a BEMA mixture of50% ITO150% voids! forms directly on glass and on thSiO2-overcoated Si~100!. In contrast, the deposition of thITO film on the Si substrate is accompanied by the formatof a thin SiO2 layer of about 20 Å at the Si/ITO interfaceThis oxide thickness is higher than the 2–6 Å native oxlayer that may form on the Si wafer surface when exposeair in the brief time interval between the HF cleaning a

FIG. 3. Spectra of the refractive index and extinction coefficient derivedusing the BEMA model shown in the inset and the ITO dispersion lawEq. ~1! defined in the text for ITO films deposited by rf sputtering oCorning glass 7059, Si~001!, and Si overcoated with a SiO2 film ;2100 Åthick.


ro-of

-

.sr

n

eto

loading into the sputtering reactor. In fact, the SE spectrathe Si substrate have been measured soon after the HFetching and modeled assuming the presence of 661 Å offreshly formed SiO2 . Figure 4 shows the comparison of thfit quality obtained for the ITO/Si sample using as substrthe experimental spectra of the freshly etched Si subst~spectra A! and a thicker~20Å! SiO2 underlayer~spectra B!that significantly improves the data fit. Besides, Fourtransformed infrared transmittance~FTIR! measurements~which have been subtracted by any contribution of natoxide on the Si wafer! have confirmed the presence of SiO2

through the stretching of Si–O–Si bonds at 1080 cm21.Therefore, the surface reactivity of Si in contact with ITOhighlighted.16 This interface layer stems from thec-Si sur-face bombardment at the very early stage of ITO depositTherefore, the optimization of the sputtering process for Ideposition should consider not only the oxygen contand/or temperature to improve the ITO layer itself, but athe substrate bias and pressure to minimize the interfaceing. In addition, Fig. 3 shows that although ITO is effectivegrowing on the SiO2 surface after all, different spectra of thextinction coefficient are found. Spectra differ especiallythe low-energy region, where the SE measure is sensitivthe ITO/substrate interface. Consequently, an effect ofunderlying surface structure and morphology on ITO optiabsorption is deduced. The different extinction coefficiespectra~and, hence, optical absorption! found by ellipsom-etry for the ITO films agree with the different electrical datIn particular, the data of the sheet resistance reported abottom of Fig. 3 show that the sheet resistance of the Ideposited on Si is slightly higher than that depositedglass. Moreover, the higher sheet resistance of ITO/Si cosponds to the lower value of plasma frequencyvp and thehigher value of collision frequencyvt . Hence, the SE analysis can also be used to infer the electrical properties in cnection with the optical properties.

yf

FIG. 4. Experimental~dots! and calculated~line! spectra of the pseudorefractive index and pseudoextinction coefficient for an ITO film depositedrf sputtering on Si~001!. The calculated spectra have been derived by usBEMA models that consider the formation of a 20-Å-thick SiO2 interfacelayer ~continuous line; model A!, the HF-etched Si substrate~dashed line;model B!, and the ITO dispersion law of Eq.~1! defined in the text. Good-ness fit,x2 is also reported for the two models.

ono

eheed,c

SEt t

o

h

erndpler

thn

ni-e

ex-po-o-inalure,s.iala-

era-of

e

of

er-Ot it

ont

byTOe


A set of ITO films deposited by e-gun evaporationglass, Si~100!, and thec-Si/a-Si:H heterojunction has alsbeen analyzed. Typical spectra of the real,^«1&, and imagi-nary, ^«2&, parts of the pseudodielectric function of thc-Si/a-Si:H/ITO heterostructure are shown in Fig. 5. In tcase of thec-Si/a-Si:H/ITO, SE spectra have been acquirbefore~on thec-Si/a-Si:H system! and after ITO depositionin order to discriminate any modification of the heterostruture only related to ITO deposition. In the analysis of thespectra of the heterostructure, BEMA models equivalenthat of the ITO/glass~model A, includes a homogeneous ITObulk layer and the surface roughness! and of ITO/Si~modelB, which considers the possible formation of a SiO2 interfacelayer due to oxygen atom in-diffusion at the early stageITO deposition! have been tested, as shown in Fig. 5~b!,where the goodness fit (x2) and 95% confidence limit of thefit parameters are also reported. However, the best fitbeen obtained by the two-layermodel C, which consists of agraded bulk ITO layer~linear gradient! with surface rough-ness. The derived~n,k! spectra are reported in Fig. 6 togethwith the best-fit BEMA models. Hence, for ITO on glass ac-Si, the same models of the sputtered films of Fig. 3 apFor thec-Si/a-Si:H/ITO heterostructure, a graded ITO laywith surface roughness is the only model fitting the data~thepresence of any SiO2 layer has been excluded by FTIR!. Inparticular, a less dense/ordered ITO layer forms ona-Si:H surface (76%ITO124%voids) and increases i

FIG. 5. ~a! Calculated and experimental spectra of the real,^«1&, and imagi-nary, ^«2&, parts of the pseudodielectric function of ITO films depositedthe c-Si/a-Si:H heterojunction.~b! BEMA model and their goodness fitested in the analysis of thec-Si/a-Si:H/ITO heterostructure.

JVST A - Vacuum, Surfaces, and Films

-

o

f

as

y.

e

density/order with thickness, reaching the~n,k! values typicalof ITO on glass. Consequently, ITO/a-Si:H interfacial reac-tivity is deduced. In particular, local diffusion of hydrogefrom the a-Si:H layer, whose surface is hydrogen termnated, into the ITO film, is likely to occur because of thsubstrate heating during the ITO deposition.

Thus, a substrate effect on the ITO optical propertiesists. This effect stems from two factors, whatever the desition technique. A factor could be the different thermal cefficients of various substrates, so that the same nomtemperature results in a different real surface temperatwhich induces growth of ITO films with different propertieIn fact, other authors previously showed that the initgrowth of ITO goes through the processes of island formtion, island coalescence, and film formation,17 whose crystal-line size and orientation strongly depend on surface tempture. Another factor to be considered is the occurrenceinterfacial reactions of ITO in contact with Si anda-Si:H,while a stable ITO/SiO2 interface is found. In the case of thc-Si substrate, a surface oxidation to SiO2 , assisted by oxy-gen ion bombardment, proceeds during the initial stagethe e-beam and sputtering process.10 Similarly, in the case ofa-Si:H, being surface SiH2 and SiH terminated, thehydrogen/oxygen interaction modifies the interfacial propties of ITO. An extensive study on the modification of ITproperties by hydrogen is reported in Ref. 18. The fact tha

FIG. 6. Spectra of the refractive index and extinction coefficient derivedusing the BEMA model shown at the bottom of the figure and the Idispersion law of Eq.~1! defined in the text for ITO films deposited bybeam on Corning glass, Si~001!, and a c-Si/a-Si:H heterojunction. Thethickness of thea-Si:H film deposited by PECVD is;240 Å.

heioS. 6

fi-omctiobret

peefnefoldth

be

I

d J.

2nder-

ni,

nk,

m.

M.


is an interface localized reactivity is also confirmed by tXPS analysis, which found the same chemical compositin the outmost 100 Å, which is the sampling depth of XPamong the ITO layers of the sets in Fig. 3 and that of Fig

IV. CONCLUSION

In conclusion, the refractive index and extinction coefcient of ITO films deposited on opaque substrates and cplex heterostructures have been determined in the sperange 1.5–5.0 eV by spectroscopic ellipsometry. A dispersequation, which combines the Drude model and a douLorentzian oscillator, is used to describe the ITO opticalsponse from the plasma frequency at the near-IR end tofundamental absorption edge at the near-UV end. The dedence of the ITO optical properties on the substrate has balso proven and discussed. In particular, the formation othin SiO2 interface layer occurs during the growth of ITO othe Si wafer, while H diffusion is the cause of the disorderand porous ITO/a-Si:H interface in the case of the growth oITO on c-Si/a-Si:H heterojunctions. Therefore, becausethis interface reactivity, the optical properties of ITO shoube measured in the device configuration, and to this aimspectroscopic ellipsometry analysis is particularly usefulcause it does not require transparent substrates.

ACKNOWLEDGMENTS

The authors acknowledge Progetto Finalizzato MSTAof CNR and ENEA-Portici~Contract No. 2000/50882! forfinancial support.


n,,.

-ralnle-hen-ena

d

f

e-

I

1T. J. Coutts, D. L. Yong, and X. Li, MRS Bull.25, 58 ~2000!, and refer-ence therein.

2I. Hamberg and C. G. Granqvist, J. Appl. Phys.60, R123~1986!.3K. Zhang, A. R. Forouhi, and I. Bloomer, J. Vac. Sci. Technol. A17, 1843~1999!.

4M. Bender, J. Trube, and J. Stollenwerk, Thin Solid Films354, 100~1999!.

5M. Bender, W. Seeling, C. Daube, H. Frankenberger, B. Ocker, anStollenwerk, Thin Solid Films326, 67 ~1998!.

6S. A. Knickerbocker and A. K. Kulkami, J. Vac. Sci. Technol. A14, 757~1996!.

7R. De Rosa, P. Grillo, G. Sinno, F. Roca, and M. Tucci, Proceeding ofWorld Conference and Exhibition on Photovoltaic Solar Energy Convsion ~1998!, p. 1583.

8J. A. Woollam, W. A. McGahan, and B. Johs, Thin Solid Films241, 44~1994!.

9R. A. Synowicki, Thin Solid Films313–314, 394 ~1998!.10J. A. Dobrowolski, L. Li, and J. N. Hilfiker, Appl. Opt.38, 4891~1999!.11J. Pla, E. Centurioni, R. Pinghini, C. Summonte, R. Rizzoli, F. Zigna

and A. Desalvo,Proceedings of the XVIth EC-PVSEC, Glasgow, UK, 1–5May 2000, edited by H. Scheer, B. McNelis, W. Palz, H. A. Ossenbriand P. Helm~James & James, 2000!, p. 597.

12G. Sinno, C. Minarini, S. Loreti, and G. Di Francia, Solid State Pheno67–68, 249 ~1999!.

13D. A. G. Bruggemann, Ann. Phys.~Leipzig! 24, 636 ~1965!.14E. D. Palik,Handbook of Optical Constants of Solids II~Academic, San

Diego, CA, 1991!, p. 163.15E. Burstein, Phys. Rev.93, 632 ~1954!.16C. W. Ow-Yang, Y. Shigesato, and D. C. Paine, J. Appl. Phys.88, 3717

~2000!.17X. W. Sun, H. C. Huang, and H. S. Kwok, Appl. Phys. Lett.68, 2663

~1996!.18M. Losurdo, M. Giangregorio, P. Capezzuto, G. Bruno, F. Varsano,

Tucci, and F. Roca, J. Appl. Phys.90, 6502~2001!.

Parametrization of optical properties of indium–tin–oxide thin films by spectroscopic ellipsometry: Substrate interfacial reactivity

Documents