Antireflective coating for ITO films deposited on glass substrate

J O U R N A L O F M AT E R I A L S S C IE N C E: M ATE R IA LS I N EL E C T RO N I C S 1 0 (1 99 9 ) 49 1± 49 5

Antire¯ective coating for ITO ®lms deposited on

glass substrate

BI-SHIOU CHIOU, JEN-HUAN TSAI

Department of Electronics Engineering and Institute of Electronics, National Chiao Tung

University, Hsinchu, Taiwan

E-mail: bschiou@cc.nctu.edu.tw

The refractive index n of radio-frequency (r.f.) magnetron sputtered indium tin oxide (ITO)

®lms varies with sputtering parameters, such as sputtering power and oxygen percentage in

the sputtering ambient. In this study, the feasibility to fabricate multilayer antire¯ective (AR)

coating with a single ITO target by controlling the sputtering conditions is explored.

Reduction in the re¯ectance can be achieved by using a one-quarter-wavelength inner layer

ITO with a refractive index n ˆ 1:87 and a one-quarter-wavelength outer layer ITO with

n ˆ 2:17. Hence, a single ITO target suf®ces in the preparation of multilayer AR coating. This

simpli®es the deposition processes and equipment for the fabrication of AR coating. Surface

corrugation, another approach to the reduction of re¯ectance, is also discussed.

1. Introduction

Indium tin oxide (ITO) is an In2O3based material that has been doped with Sn to improve electrical conductivity. Tin acts as a cationic dopant in the In2O3 lattice and substitutes on the indium sites to bind with interstitial oxygen. The presence of SnO2would result in n doping of the lattice because the dopant would add electrons to the conduction band [1]. The ITO ®lm, with a band gap of approximately 3.8 eV, is highly transmissive in the visible region [2]. Hence, transparent conductive ITO ®lm has many applications, such as in transparent electrodes, anti-re¯ection coatings, display devices, and photoelectronic devices [3±5]. In the design of anti-re¯ection (AR) coatings, materials with low refractive indices such as

SiO₂ …n ˆ 1:46† and MgF₂ …n ˆ 1:38† are used as

coating for ITO ®lms to reduce unwanted re¯ections at the surface of the optical elements [6±8]. However, because MgF₂ and SiO2 are electrical insulators this exhibits a limitation of ITO ®lm in the application of display devices due to the high resistivity of the SiO2(or MgF₂) coating. In addition, multiple source materials are required in the preparation of multi-layer AR coatings, this complicates the deposition processes and equipments for the fabrication of AR coating.

Previous studies reveal that the electro-optical proper-ties of radio frequency (r.f.) sputtered ITO ®lms are sensitive to sputtering parameters, such as r.f. power [9± 11], oxygen content in the sputtering ambient [1], substrate roughness [12], and annealing conditions [13]. It is anticipated that the refractive index of the ITO ®lms can be varied by changing the sputtering parameters during deposition of ITO. Hence, the objective of this research is to explore the feasibility of the multilayer antire¯ective coating design with ITO

®lms of various refractive indices obtained by varying the sputtering conditions.

2. Experimental procedures

ITO ®lms were prepared by using a commercial r.f. magnetron sputtering system (Ion Tech, UK). The sputtering target was a 1 inch hot-pressed oxide ceramic

(90 wt % In2O3 and 10 wt % SnO2, 99.99% purity)

supplied by Cerac, Inc., USA. The substrates employed were Corning 7059 glass, degreased ultrasonically in a dilute detergent solution, rinsed ultrasonically in

deio-nized water and blown dry in N2 gas before they were

introduced into the chamber. The substrate holder was ®xed directly above the target and was rotated at * 10 r.p.m. by a motor. The target-to-substrate distance was 5 cm and a mechanical shutter was attached to the target.

The vacuum chamber was a stainless steel bell jar pumped by a conventional oil diffusion pump (Diffstak

250, Edwards, UK). High-purity Ar (99.999%) and O₂

(99.5%) were introduced through a mass ¯ow controller after the vacuum chamber was evacuated to about

3:1610ÿ4_{Pa. The gas pressure was monitored with a}

precision ionization gauge and was kept at 1 + 0.03 Pa during deposition. The r.f. power (13.56 MHz) was introduced through an r.f. power supply (RF Plasma Products, Inc., USA) with an automatic matching network which could be tuned for minimum re¯ected power. Before deposition, the target was presputtered to remove any contaminants and eliminate any differential sputtering effects. The presputtering time was 20 min for pure Ar and was increased to 30 min as O₂was added to the sputtering ambient.

The ®lm thickness was measured with a stylus surface pro®ler. The optical transmittance of the ®lms was measured with an ultraviolet-visible-near infrared spec-trophotometer (Hitachi U-3410, Japan).

3. Results and discussion

The refractive index n…l† and the extinction coef®cient k…l† of the ITO ®lm are calculated from transmission spectrum of the ®lm. According to Swanepoel's method [14], which is based on the idea of Manifacier et al. [15] of creating the envelopes of interference maxima and minima in the transmission spectra, an approximate initial value of the refractive index of the ®lm n₀ in the spectral region of medium and weak absorption, can be calculated by the expression

n₀ˆ N ‡ …Nh 2_{ÿ s}2_†1=2i1=2 _1† where N ˆ 2sTM_Tÿ Tm MTm ‡ s2_{‡ 1} 2 …2†

here T_M and T_mare the transmission maximum and the

corresponding minimum at a certain wavelength l, one being measured and the other calculated; s is the refractive index of the Corning 7059 glass substrate and a value of 1.53 is used. The basic equation for interference fringes is

4nd ˆ ml …3†

where m is an even integer for maxima and an odd integer for minima, and d is the ®lm thickness measured with a stylus surface pro®ler. The order of a given extremum m₀ can be estimated from Equation 3 using d and the corresponding n0. In addition, the values of m0 can be determined by a simple graphical method based on Equation 3. This expression can be rewritten for that purpose as

` ˆ 4d n_l0 ÿ m1 …4†

where ` ˆ 0; 1; 2; . . . and m₁ is the ®rst extremum. Therefore, plotting l against n₀=l yields a straight line with slope 4d and intercept on the y axis of ÿm₁. From this plot the values of m₁ and hence each corresponding

order of a given extremum m₀ can be estimated. The

orders of m of the neighboring extrema are in fact consecutive integers, even for the maxima and odd for the minima of the transmission. The ®nal value of n for each extremum is obtained by substituting ®lm thickness d and the corresponding exact integer values of m associated with each extreme point in Equation 3. The value of the extinction coef®cient k is obtained by the following equations E_Mˆ8n_T2s M ‡ …n 2_{ÿ 1†…n}2_{ÿ s}2_{† …5†} x ˆEMÿ E 2 Mÿ …n2ÿ 1†3…n2ÿ s4† h i1=2 …n ÿ 1†3…n ÿ s2_† …6† x ˆ exp ÿ4pk_l d   …7† The refractive index and extinction coef®cient as functions of wavelength for the ITO ®lms deposited at different sputtering powers are shown in Fig. 1. The refractive index of the ITO ®lm decreases with the wavelength l, i.e. dn=dl50. This is consistent with what one would expect from a Kramers±Kronig analysis [16]. It is seen that the ITO ®lms deposited at high sputtering power have large values of refractive index and extinction coef®cient. The n value at 550 nm ranges between 1.97 (20 W) and 2.17 (100 W). Fig. 2 gives the refractive index and extinction coef®cient versus

Figure 1 (a) Refractive index and (b) extinction coef®cient versus wavelength for the as-deposited ITO ®lms prepared at various sputtering power, ®lm thickness 600 nm, sputtering ambient 0% oxygen.

wavelength for ®lms prepared under various oxygen percentage. At 20 W, decreases in n and k are observed when the oxygen content increased from 0% to 2%. However, no apparent difference is found in k between samples prepared under 2% O₂ and those under 8% O₂. The n value at 550 nm ranges from 1.97 (0% O₂) to 1.87 (8% O₂).

When a light wave is incident on a system of thin ®lms consisting of an assembly of l layers ( j ˆ 1 to l, with j ˆ 1 as the outermost layer) on a substrate of refractive index nS, the characteristic matrix of the optical system can be given by [17]: B C   ˆ Yl jˆ1 cos d_j i sin d_j=Z_j iZ_jsin d_j cos d_j   ( _) 1 Z_s  …8† where djˆ …2p=l†…Njtjcos yj† …9†

is the effective phase thickness of the jth layer; l is the wavelength of the incident radiation in vacuo, and y_j is the angle of refraction in the jth layer and is related to the angle of incidence y by Snell's law:

N0sin y ˆ Njsin yj …10†

N_jt_jand N_jt_jcos y_jare called the optical thickness and the effective optical thickness, respectively.

The equivalent optical admittance Y, the re¯ectance R, transmittance T and absorption A of the system can be given by Y ˆ C=B …11† R ˆ n0ÿ Y n0‡ Y  ₂ …12† T ˆRen0…1 ÿ R† Re…BC† …13† A ˆ 1 ÿ R ÿ T …14†

where C_{is the complex of C element, n}

0is the refractive index of incident medium (n0ˆ 1 for air) and Reis the real part of R.

On the basis of Equations 8 to 12, when the optical thickness of a single-layer coating is adjusted so that n₁d₁ˆ l=4, zero re¯ectance occurs if n₁ˆ pn₀n_s. However, the substrate employed in this study has an n_Sof 1.53 which suggests an n₁(1.24) much smaller than those of ITO ®lms. The re¯ectance of a l/4 ITO coating as a function of wavelength is shown in Fig. 3. The R values increases with the increase of sputtering power and, hence, the refractive index.

When the thickness of a single layer coating is

adjusted so that n1d1 ˆ l=2, the re¯ectance can be

reduced to

R ˆ n0ÿ ns n0‡ ns

!2

…15† which is equivalent to an uncoated substrate at a wavelength l, as shown in Fig. 4 for l ˆ 550 nm. However, the re¯ection is appreciable and, due to the absorption of the coating, the transmittances are larger than those of a bare substrate.

According to Equations 8 to 12, the re¯ectance of a non-absorbing double-layer …l=4 ÿ l=4† coating can be expressed as;

Figure 2 (a) Refractive index and (b) extinction coef®cient versus wavelength for the as-deposited ITO ®lms prepared under various

R ˆ n0n22ÿ n21ns n2n22‡ n21ns

 2

…16† A zero re¯ectance requires:

n1 n₂ˆ  n0 n_s r ˆ 0:808 …17†

thus, the refractive index of the inner layer …n₂† should be larger than that of the outer layer …n₁†. The n values of ®lms in this study range from 1.97 to 2.17. Setting n₂ equal to 2.17, one ®nds that R decreases with n₁, as obtained from Equation 16 and shown in Fig. 5. The re¯ection spectra of three double-layer …l=4 ÿ l=4† coatings prepared under various conditions and one single layer …l=2† coating are exhibited in Fig. 6. The minimum re¯ectances are 7.45%, 6.46%, and 5.71% for

samples l/4 …40 W†* _{± l/4 …100 W†, l=4 …20 W†}

ÿ l=4 …100 W†, and l=4 …20 W; 8%O2†{ÿ l=4 …100 W†, respectively. The n values for the inner layers

l=4 …40 W†, l/4 (20 W), and l/4 (20 W, 8% O2) are

2.03,1.97, and 1.87, respectively. The R does decrease with the increase of n₁, as suggested in Fig. 5. It is also noted that the double layer ITO coatings …l=4 ÿ l=4† have lower re¯ectance than the single layer …l=2† one. Hence, it is feasible to design multilayer AR coatings with a single ITO target by adjusting the sputtering parameters, such as oxygen concentration in the sputtering ambient and sputtering power. The presence of oxygen during sputtering enhances the crystallization of the ®lm and increases the ®lm grain size. The addition of oxygen reduces the oxygen-de®cient region of the ®lm and, consequently, the optical transmittance of the ®lm is improved while the ®lm conductivity decreases [1]. The structure and orientation of ITO ®lms also strongly depends on the energy of the sputtered particles arriving at the substrate. The preferred orientation of the ITO ®lms changes from (2 2 2) to (4 0 0) as the sputtering power increases. A high sputtering power causes the increase of oxygen vacancies in the ®lms and results in the loss of optical transmittance of the ®lms [9±11]. Hence, scrupulous care is needed in the design of multilayer AR coatings.

The n₁and n₂for a l=4 …20 W; 8%O₂† ÿ l=4 …100 W† coating at 550 nm are 1.87 and 2.17, respectively. The re¯ectance of the coating is * 0.41%. The backside re¯ectance of the substrate is * 4.20%. These add up to a total re¯ectance of * 4.61%, as compared to 5.71% of the measured datum. The discrepancy between the experimental data and calculated data may be due to factors such as accuracy in thickness measurement, statistical deviation of the refractive index, and absorp-tion of the ®lm materials.

An alternative approach to single layer AR coating is to employ a high spatial-frequency rectangular-groove surface structure that behaves optically as an antire¯ec-tion coating. Enger and Case [18] reported a signi®cant reduction in R by etching linear fringe patterns on quartz. Motamedi et al. [19] generated pillar arrays on silicon substrate to simulate a single homogeneous AR layer Theoretical derivations by Gaylord et al. [20]

demon-Figure 4 The transmission and re¯ection spectra of l/2 ITO coating prepared at various sputtering powers; l ˆ 550 nm.

Figure 5 Calculated re¯ectance as a function of outer layer refractive index …n† for a two layer AR coating. The n of the inner layer is 2.17.

Figure 6 Re¯ection spectra of four coating systems. Note (20 W, 8%) means ®lms sputtered at 20 W with 8% O2.

*_{Film sputtered at 40 W, pure Ar.} {_{Film sputtered at 20 W, 8% O}

strated that zero re¯ectivity can be achieved by

employing high-frequency rectangular grooves.

However, the dimensions of the grooves are around one to one-tenth of the wavelength, which are too small to be etched precisely in the case of ITO [21]. Nonetheless, a preliminary attempt was made by fabricating an ITO coating with an abrupt decrease in thickness, and a reduction in re¯ectance is observed [22].

4. Conclusions

ITO ®lms were deposited on a glass substrate with r.f. magnetron sputtering. The refractive indices of the ®lms are found to decrease with the decrease of sputtering power or increase of the oxygen content in the sputtering ambient. The refraction index n at 550 nm ranges from 1.87 for ®lms prepared at 20 W, 8% O₂to 2.17 for those prepared at 100 W, 0% O₂. A double layer coating was fabricated on a glass substrate with ITO ®lms of different refractive indices. It is found that the re¯ectance at 550 nm decreases from * 8.78% of the uncoated substrate to * 5.71% of a l=4 ÿ l=4 double layer coating with refractive index of inner ITO ®lm equals 1.87 and that of outer ITO ®lm equals 2.17. Besides, the double layer ITO coating …l=4 ÿ l=4† has lower re¯ectance than the single layer ITO coating …l/2† does. Hence, it is feasible to fabricate a multilayer antire¯ective (AR) coating with a single ITO target by controlling the sputtering parameters. Another approach to a single target ITO AR coating is through micro-structure engineering of the ITO surface. However, a breakthrough in the etching technique of the ITO ®lm is required before this approach can be realized. A decrease in re¯ectance is observed at wavelengths between 800 nm and 1500 nm.

Acknowledgement

This work is supported by National Science Council, Taiwan (NSC82-0417-E009-395) and partly supported

by the Chung-Shan Institute of Science and Technology (CS83-0210-D009-001).

References

1. W. F. W UandB . S . C H I O U, Semicond. Sci. Technol. 11 (1996) 196.

2. T. J . CO U T T S, Thin Solid Films 138 (1986) 65. 3. J . C . M A N I FAC I E R, ibid. 90 (1982) 297.

4. K . L . C H O P R A,S . M A J O RandD . K . PA N DYA, ibid. 80 (1981) 143.

5. F. L . B O U Q U E TandC . R . M A A G, IEEE Trans. Nucl. Sci. NS-33 (1986) 1408.

6. W. F. W UandB . S . C H I O U, Appl. Surf. Sci. 115 (1997) 96. 7. W. F. W UandB . S . C H I O U, Semicond. Sci. Technol. 11 (1996)

1317.

8. J . T. CO X,G . H A S SandA . T H E L E N, J. Opt. Soc. Am. 52 (1962) 965.

9. W. F. W U andB . S . C H I O U, Semicond. Sci. Technol. 9 (1994) 1242.

10. W. F. W UandB . S . C H I O U, Thin Solid Films 247 (1994) 201. 11. B . S . C H I O U,S . T. H S I E HandW. F. W U, J. Am. Ceram. Soc. 77

(1994) 1740.

12. B . S . C H I O UandS . T. H S I E H, Thin Solid Films 229 (1993) 146. 13. W. F. W UandB . S . C H I O U, Appl. Surf. Sci. 68 (1993) 497. 14. R . S WA N E PO E L, J. Phys. E, Sci. Instrum 16 (1983) 1214. 15. J . C . M A N I FAC I E R,J . G A S I OT andJ . P. F I L L A R D, ibid. 9

(1976) 1002.

16. A . F E L D M A N, Appl. Opt. 23 (1984) 1193.

17. K . L . C H O P R AandI . K A U R, ``Thin Film Device Applications'', (Pleum Press, New York, 1983) p. 55.

18. R . C . E N G E RandS . K . CA S E, Appl. Opt. 22 (3220) 1983. 19. M . E . M OTA M E D I,W. H . S O U T H W E L LandW. J . G U N N I N G,

ibid. 31 (1992) 4371.

20. T. K . G AY LO R D,W. E . B A I R DandM . G . M O H A R A M, ibid. 25 (1986) 4562.

21. B . S . C H I O UandJ . H . L E E, J. Mater. Sci., Mater. Electron. 7 (1996) 241.

22. B . S . C H I O U, ``Fabrication and Characterization of Antire¯ection and Hard Coatings (III)'', Technical report under contract No. CS83-0210-D-009-001, June 1995, Hsinchu, Taiwan.

Received 18 January 1999 and accepted 23 April 1999.