INSTITUTE OF PHYSICS PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS

J. Phys. D: Appl. Phys. 39 (2006) 54–60 doi:10.1088/0022-3727/39/1/010

Refractive index gradients in TiO2 thin

films grown by atomic layer deposition
A Kasikov, J Aarik, H Mändar, M Moppel, M Pärs and T Uustare
Institute of Physics, University of Tartu, Riia 142, 51014 Tartu, Estonia
E-mail: [email protected]

Received 16 September 2005, in final form 28 October 2005

Published 15 December 2005
Online at stacks.iop.org/JPhysD/39/54

Abstract
Transmission spectra of TiO2 films grown by atomic layer deposition on
fused silica have been analysed by using Lorentz dispersion and a model
consisting of two sublayers inside a film. It has been shown that the
deposition process parameters significantly influenced the refractive index
gradient in the film growth direction. The films grown at a lower flow of the
carrier gas showed a 35 nm thick sublayer with low refractive index at the
silica substrate and a layer with a higher refractive index at the film surface.
The films deposited at higher carrier gas flow had a higher refractive index at
the substrate and a lower refractive index at the film surface. The results also
demonstrate that by using a parameter modelling one can obtain some
information about the internal structure of a film even if there are no clearly
defined interference fringes in a transmission spectrum.

1. Introduction anatase is preferentially formed in thinner films and the initial

stage of deposition, while the growth of rutile starts dominating
Titanium dioxide (TiO2 ) is a material that has a relatively in thicker films [7].
wide band gap [1–7] and a high refractive index [2–13]. It is obvious that the optical properties of TiO2
For these reasons TiO2 can be applied as a high-refractive- significantly depend on the phase composition. For instance,
index material in the optical coatings designed for visible the refractive index has the lowest values for amorphous
and infrared spectral ranges [12–16]. At normal or low films [2,4,5,7] and higher values for crystalline anatase [2, 3, 7–
pressure, stoichiometric TiO2 is usually formed in thin films 10] and, particularly, rutile [3, 4, 6, 7, 10] films. Unfortunately,
in amorphous [2–5, 7, 8, 11, 12, 17], anatase [1–10, 12, 17, 18] no data on the refractive index of other crystalline phases of
or rutile [1, 3, 4, 6–8, 10, 18] phases. In addition to these TiO2 have been accessible to us in the literature. However, the
phases, brookite has also been observed in thin films [19]. very high density of TiO2 -II phase [20] allows a conclusion that
Orthorhombic TiO2 -II [20–23] and monoclinic baddeleyite- the refractive index of this phase is comparable to or higher than
type [22,23] modifications of TiO2 have mainly been obtained that of rutile. In the films with composition gradients in the
at high pressures, while the TiO2 -II phase is metastable growth direction, the corresponding refractive index gradients
at normal pressure [20, 22, 23]. Sometimes the formation can also be expected. The spectral ellipsometry method has
of TiO2 -II has also been observed at normal and low mainly been used to study the refractive index gradients in
pressures from Ti3 O5 [24], at the interface of rutile and non- TiO2 thin films [9, 10, 17, 18, 30, 31]. However, transmission
stoichiometric titanium oxide [25] and in thin films [26]. In spectroscopy has also been applied for this purpose [5, 17].
these cases, the deficiency or a limited supply of oxygen during Very little information has been published about the
the material synthesis is important in the stabilization of this refractive index homogeneity in the TiO2 thin films grown
phase. by the atomic layer deposition (ALD), although this method
A general trend is that an amorphous phase is seems to be promising for the preparation of high-quality
preferentially formed at lower temperatures and in the initial optical coatings [11, 12]. Furthermore, we have not found
stage of the thin film deposition [8, 27–30]. With the increase any information about the refractive index of the TiO2 films
of the growth temperature, first anatase and then rutile is containing TiO2 -II phase. This kind of film can be grown
formed in the films. If the films are grown at the temperatures by using the ALD method, but the amount of TiO2 -II phase
where the formation of anatase as well as rutile is possible, then starts to decrease when the film thickness exceeds a critical

0022-3727/06/010054+07$30.00 © 2006 IOP Publishing Ltd Printed in the UK 54

 Refractive index gradients in TiO2 thin films

Table 1. Process parameters used for growing TiO2 thin films.

N2 flow, TiCl4 TiCl4 pulse Number of Thickness, Growth rate,
Sample µmol/s pressure, Pa duration, s cycles nm nm/cycle
1 20 7 0.5 700 61 0.087
2 20 7 1 1000 125 0.125
3 20 7 0.5 2000 153 0.077
4 50 3 2 1000 51 0.052
5 50 3 0.5 3000 151 0.050
6 50 3 2 3000 159 0.053

value that is usually around 100–200 nm [26]. Therefore, 4500

to obtain reliable information about the effect of this phase
on optical properties, relatively thin films should be studied. 4000
In this report, we are making an attempt to demonstrate that R 101
A 004
even in the case of inhomogeneous films of such relatively R 211
3500 S6
small thickness some information about their structure can be
obtained by the optical transmission spectroscopy method.
3000 S5

Intensity, counts
2. Experimental details
2500 S4
R110
The TiO2 thin films studied in this work were grown on fused
silica (SiO2 ) substrates by using the ALD method. To deposit 2000
R 101
the films the substrates were alternately exposed to volatilised R 211
TiCl4 and H2 O as the precursors of titanium and oxygen, 1500 S3
respectively. The precursors were carried to the substrates
in the flow of N2 gas, which was also used to purge the reactor A 200
1000 S2
between precursor pulses. The duration of H2 O pulses and
purge periods was 2 s, while the TiCl4 pulse duration was varied A 101
500 S1
from 0.5 to 2 s (table 1). In addition, we varied the flow rate
of the carrier gas and the TiCl4 pressure in the reaction zone.
The films listed in table 1 were grown at 425 ◦ C. 0
The structure of the films was studied by using the 20 30 40 50 60 70
x-ray diffraction (XRD) and Raman spectroscopy methods. 2 , deg
According to the XRD studies anatase was the main crystalline Figure 1. XRD spectra of TiO2 films deposited at low (S1–S3) and
phase in samples 1 and 2, while rutile dominated in sample 3 high (S4–S6) carrier gas flow. A and R denote the Miller indices of
(figure 1). On the basis of these data, one could conclude anatase and rutile, respectively.
that anatase was formed in the initial stage of deposition, but
after achieving a certain critical thickness, the film material of the TiO2 -II phase formed in the films was much smaller
crystallized in the rutile phase. No marked amounts of than that of anatase and/or rutile. The reflection high-energy
crystalline phases were observed by XRD in the films of a electron diffraction (RHEED) studies made for the samples
similar thickness grown at a higher flow rate of the carrier grown simultaneously with samples 1–3 on silicon substrates
gas (samples 4–6). The Raman spectroscopy measurements confirmed the formation of TiO2 -II in samples 2 and 3.
showed, however, that all the films were crystalline. Anatase Moreover, only TiO2 -II was detected by RHEED in these
dominated in samples 1, 2, 4, 5 and 6 (figures 2(a) and (b)), samples and also in sample 1. Very similar results were
i.e. in thinner films grown at a lower flow rate of the carrier gas obtained by RHEED on SiO2 substrates, although these results
and in the films grown at a higher flow rate of the carrier gas. were less reliable, because the conductivity of these samples
The rutile phase was observed in samples 2, 3 and 6, whereas was insufficient to get high-quality RHEED patterns. Anyway,
in sample 3 the dominating phase was rutile. These results it was possible to conclude that the thin-film structure observed
confirm that at the deposition temperature used the rutile phase by RHEED was similar on silicon and silica substrates. Thus,
can be formed only in sufficiently thick films and preferentially the reason for the difference in the Raman and RHEED data
at lower flow rates of the carrier gas. The films obtained at might be the formation of TiO2 -II only in a very thin surface
these process parameters obviously contain an anatase-rich layer and/or an additional formation of TiO2 -II during the
layer at the substrate-film interface and a rutile-rich layer at RHEED measurements due to the influence of electron-beam
the outermost surface. As a result, refractive index gradients irradiation of the samples [25]. It was revealed, however,
can be expected in these films. that no TiO2 -II was formed from the phase-pure anatase and
In addition to the Raman peaks of anatase and rutile, the phase-pure rutile during the RHEED measurements. Our
the peaks attributable to TiO2 -II were observed in the Raman RHEED studies also demonstrated that TiO2 -II phase was
spectra of samples 2 and 3 at 280 and 310 cm−1 (figure 2(a)). obtained in the films only when soda lime glass substrates
The low intensity of these peaks indicated that the amount were together with silica and silicon substrates in the reactor

55
A Kasikov et al

(a) 300 measurements the spectrophotometer was equipped with

additional diaphragms and an integrating sphere to ensure the
accuracy of the obtained transmission values [32].
250 R
TiO 2-II,
R
B A 3. Computations
A
Intensity, arbitrary units

200
R R R A
The experimental data were analysed by using a method
S3 proposed by Dobrowolski et al [33]. A coated material was
150 A characterized by a one-band Lorentz dispersion:

B · λ2 · (λ2 − C 2 )
A S2 n2 (λ) = A + k 2 (λ) + ,
100 (λ2 − C 2 )2 + D 2 · λ2
(1)
1 B · D · λ3
k(λ) = · 2 .
50 2 · n(λ) (λ − C 2 )2 + D 2 · λ2
S1
x 0.2 The calculations showed that k  n in the investigated
0 wavelength region when some data points were excluded at the
200 300 400 500 600 700 800 absorption edge. This enabled the neglecting of the term k 2 (λ)
Raman shift, cm
-1 in the expression for n2 (λ) and straightforward calculations of
optical constants without introducing significant errors.
150 The theoretical transmission spectra were fitted to the
(b) experimental ones by using a Nelder–Mead (simplex) method.
A The theoretical spectra were obtained by using five free
125 A parameters (A, B, C and D for dispersion and an additional
parameter for the film thickness) for a homogeneous-layer
A (one-layer) model. The effect of a substrate, caused by
Intensity, arbitrary units

100 an additional reflectance from its backside, was taken into

R S6 A account by using the refractive index value determined for
A a bare substrate from its transmission spectrum. As there
75
were no well-defined minima or maxima that could give us
an initial approximation for fitting, the fitting procedure was
S5 started with the wide variation ranges of variable parameters.
50
The parameters obtained in this procedure were then used
as the first approximation for another fit where a possible
25 inhomogeneity of the refractive index in the growth direction
S4 was included. The model taking into account the refractive
x 0.25 index variation assumed that a film should consist of two
0 sublayers with different refractive indices. To limit the number
200 300 400 500 600 700 800
-1
of variable parameters the differences in the dispersion and the
Raman shift, cm absorption of these layers were neglected. In this simplified
Figure 2. The Raman spectra of TiO2 films deposited at (a) low and case, seven parameters (A, B, C and D for one sublayer,
(b) high carrier gas flow. A, R and B indicate the Raman bands of thickness values for both sublayers and a refractive index step
anatase, rutile and brookite, respectively. at the interface of two sublayers) were needed to calculate a
transmission spectrum. The transmission of the whole system
during the film growth. Consequently, the impurities that was obtained as that of the multilayer interference stack with
out-diffused from the soda-lime glass probably stimulated internal reflections.
the growth of TiO2 -II. It is worth noting, however, that the
Auger electron spectroscopy studies did not reveal any sodium 4. Results and discussion
impurities in these films. So, the concentration of these did not
exceed the sensitivity of the method. The RHEED studies also The spectra of all specimens were fitted independently and
demonstrated that the TiO2 -II phase was preferentially formed the results are represented in figures 3–6. Figure 7 shows
at the lower flow rates of the carrier gas. Hence, to grow films refractive index variations for different samples in a set of films
not containing TiO2 -II phase (samples 4–6) we used a higher deposited at a lower carrier gas flow. The numerical values for
flow rate of the carrier gas and only silicon and pure silica as the sample parameters, obtained in two-layer inhomogeneous
the substrates placed into the reactor. approximation, are presented in table 2. The first three rows
To reveal the possible refractive index variations, the describe the TiO2 films grown at a lower flow rate of the
ALD-coated films were analysed by using a transmission carrier gas. Comparing the fits in one-layer and two-layer
spectrometry data. The transmission spectra of the samples approximations with the experimental curves, one can see that
on fused SiO2 substrates were measured with a dual- the merit function for the two-layer approximation is lower
beam SPECORD M40 spectrophotometer. During the by 0.5–2 orders of magnitude, indicating a significantly better

56
 Refractive index gradients in TiO2 thin films

1.0 (a) S1; 1-layer

4.0
S1; 2-layer
3.8 S2; 1-layer
0.8 S2; 2-layer
3.6

Refractive index
S3; 1-layer
3.4
Transmission

S3; 2-layer
0.6 3.2
3.0
Quartz
0.4 2.8
Measured
1-layer 2.6
2-layer
0.2 2.4
2.2

0.0 2.0
3.5 3.0 2.5 2.0 1.5 1.0 3.0 2.5 2.0 1.5 1.0
-1 -1
Wavenumber, m Wavenumber, m

Figure 3. Transmission of a 127 nm thick TiO2 sample containing (b)

3.6
anatase, rutile and TiO2 -II phases modelled in 1- and 2-layer
approximations. 3.4

Refractive index 3.2

1.0 1-layer
2-layer
3.0

0.8 2.8

2.6
Transmission

0.6 2.4

2.2
Quartz
0.4
Measured 2.0
1-layer 3.0 2.5 2.0 1.5 1.0
-1
0.2
2-layer Wavenumber, m

Figure 5. Refractive index dispersions calculated in 1- and 2-layer

0.0
approximations for TiO2 films deposited at (a) low and (b) high
3.0 2.5 2.0 1.5 1.0 carrier gas flow. The shifts of n values indicate different types of
-1 inhomogeneity in the two sets.
Wavenumber, m

Figure 4. Transmission of a 153 nm thick TiO2 sample containing layers in the films with different thickness (figure 7). The same
anatase, rutile and TiO2 -II phases modelled in 1- and 2-layer
must be said about the thinnest sample in which the refractive
approximations.
index of the upper layer seems to be physically unsound.
When comparing the dispersion curves calculated for films
fit to the measured spectrum. All three independent fits of deposited at low carrier gas flow (figures 5(a) and 6(a)), one
samples 1–3 demonstrate that there is a layer with a lower can see that the curves obtained from the two-layer model
refractive index at the substrate–film interface. The thickness show more similarities with each other than those obtained
of this layer is approximately 35 nm. According to the results from the one-layer model. This result, together with the
of the fitting, the outermost layer has a higher refractive index better merit function of the two-layer model, demonstrates that
and its thickness increases with the increase of the total film the latter model more reliably describes the samples than the
thickness. homogeneous (one-layer) model does. We believe that in the
As the interfering components of the reflected light in the homogeneous case, the contribution of the reflectance from
layered thin film systems are set by the Fresnel coefficients the inner structure of a material is partially transferred into
on the borders between layers, the main contribution to the dispersion information. The decrease of the refractive indices
reflection in films with a high index of refraction comes from obtained for two spectra at the highest wavenumber values
the film–air interface. On the other hand, the light is reflected (figure 5(a)) is probably related with the significant increase
because of the refractive index variation inside the film and at of the absorption. In this region, λ < C and formulae (1) for
the film–substrate interface. These reflections are, however, the one-oscillator approximation [33] are not valid any more.
weaker than those at the film–air interface and for this reason Figure 6(a) shows that the absorption index versus wavelength
they influence the measured spectra to a lesser extent. This, dependences calculated from a two-layer model, are steeper
together with the inevitable dispersion effects that we have not for all TiO2 coatings when compared with the corresponding
taken into account, makes the obtained refractive index of the dependences obtained from the one-layer model. Thus, the
interface layer less reliable. These effects probably lead to a simple model could not present the absorption dispersion well
large scatter of refractive index values obtained for interface enough. A rise of k with the material thickness indicates a

57
A Kasikov et al

(a) 1.4 3.5

S1
S1; 1-layer
1.2 TiO 2 S2
S1; 2-layer 3.0
S2; 1-layer S3
Absorption index

1.0
S2; 2-layer

Refractive index
0.8
S3; 1-layer 2.5
S3; 2-layer
0.6
2.0
Silica
0.4
1.5
0.2 Air

0.0 1.0
3.0 2.5 2.0 1.5 1.0
-1
-100 -50 0 50 100 150 200
Wavenumber, m
Thickness, nm
(b) 0.10 0.10 Figure 7. Refractive index structures of the inhomogeneous TiO2
S1; 2-layer
0.09 S1; 1-layer samples deposited at lower carrier gas flow determined at the
0.09 S2; 2-layer
0.08
0.07
S2; 1-layer wavelength of 500 nm.
0.08 S3; 2-layer
0.06 S3; 1-layer
Absorption index

0.07 0.05

0.06
0.04 model, because the values for the upper sublayer, i.e. the layer
0.03
0.02 with a lower refractive index, are depicted in this figure. It
0.05 0.01 is also worth noting that the inhomogeneity insignificantly
0.04 0.00
3.0 2.5 2.0 1.5 1.0 changes the absorption spectra of samples 4–6, and thus the
0.03 results of the two-layer fitting are not shown for these samples.
0.02 The figures demonstrate that the films, containing
0.01 significant amounts of rutile and TiO2 -II, have refractive index
0.00 values of about 2.65 at the surfaces of the films (at the
3.0 2.5 2.0 1.5 1.0 wavelength of 500 nm). However, the absorption is also higher
Wavenumber, m
-1 in these films than in the anatase films, especially at higher
wavenumber values. This increase of the absorption index, k,
Figure 6. Absorption index dispersions calculated in (a) 1- and is evidently related to a lower band gap of rutile as compared
2-layer approximations for TiO2 films deposited at low carrier gas with that of anatase [1, 7]. The increasing amount of rutile
flow and (b) 1-layer approximations for TiO2 films deposited at high
carrier gas flow. The insert shows the low-absorption range of the can also explain the increase of k with the film thickness,
absorption spectra of TiO2 films deposited at low carrier gas flow. which is most significant for the films grown at a lower flow
rate of the carrier gas and which contain larger amounts of
rutile. A very low value of the absorption index for the thinnest
higher defect concentration and/or a lower homogeneity of
film deposited at a low carrier gas flow can at least partially
thicker films.
result from a difference in the ratio of the upper and lower
The samples grown at higher flow rates of the carrier
sublayer thicknesses, if we assume that most of the absorbance
gas demonstrate inhomogeneity that is typical of evaporated
occurs in the upper TiO2 layer. During a transmission fitting
materials. In this case, the two-layer fitting procedure gives an
procedure, the structure of absorption is not revealed and the
improvement in the merit function when lower refractive index
absorption is equally distributed over all film thicknesses.
values are attributed to the outermost sublayer. The decrease
of the refractive index at the film–air interface can be related
to the polycrystalline nature of the film material resulting in 5. Conclusions
a rough surface layer, which can be described as a porous
substance of a mixture of the film material and air. Still, this Transmission spectra were measured for two sets of TiO2 films
model does not seem very plausible, because the fitting gives grown on silica substrates by the atomic layer deposition at
as high thickness values as 40–50 nm for the outer sublayer. different process parameters. The first set of films, which
Thus, a more complicated model is needed to describe these were of a higher degree of crystallinity and contained anatase,
films. marked amounts of rutile and minor amounts of TiO2 -II at
The optical dispersion curves obtained from the fitting the outermost surface, was prepared at a relatively low flow
of the transmission curves of samples 4–6 are presented in of the carrier gas. Another set consisting of the films that
figures 5(b) and 6(b). Figure 5(b) shows that the introduction of contained mainly the anatase phase was deposited at a higher
a negative inhomogeneity, i.e. the attributing of lower refractive flow of the carrier gas. The transmission spectra were analysed
index values to the outermost layers of films, does not change by fitting by using a two-layer model for a film, the Lorentz
much the shape of the refractive index dispersion obtained. The dispersion and a constant refractive index difference between
refractive index values calculated from the two-layer model two sublayers. Optical studies also demonstrated that the two
are somewhat lower than the ones obtained from the one-layer sets of samples had clearly different structures.

58
 Refractive index gradients in TiO2 thin films

Table 2. Internal structure of the TiO2 films and the results of transmission modelling.


Merit function (Ti )2
Thickness of Thickness of Mean T
Sample lower layer, nm upper layer, nm for 1-layer model for 2-layer model for 2 layers
1 39 22 1.1 × 10−3 2.2 × 10−5 1.0 × 10−3
2 35 92 1.0 × 10−3 3.8 × 10−4 4.3 × 10−3
3 35 118 5.0 × 10−3 8.1 × 10−5 2.0 × 10−3
4 10 42 1.3 × 10−4 1 × 10−4 2.3 × 10−3
5 97 48 3.45 × 10−3 1.05 × 10−3 7.9 × 10−3
6 75 84 3.75 × 10−3 6.1 × 10−4 6.0 × 10−3

The results for the first set show a pronounced

improvement of a merit function, when the two-layer model
was used instead of the model with a homogeneous film, and 0.96
indicate that a 35 nm sublayer of a material with a lower
refractive index is formed at the silica substrate. The second
set represents films, in which the refractive index is lower at the 0.95

Transmission
film surface than at the substrate–film interface. A relatively
good qualitative agreement between optical and structural
0.94
studies demonstrates that even if there are no well-developed
interference fringes in the transmission spectra, the internal
structure of the films can still be characterized through the 0.93
parameter modelling of their transmission spectra.
0.92
Appendix 300 350 400 450 500 550 600
Wavelength, nm
In our work, we have followed an approach measuring an
Figure 8. Modelling of a film with n = 1.60 on a semi-infinite
uncoated substrate and a coated sample together—a method not
substrate with n0 = 1.50. The introducing of a systematic
very widely used. Here we try to highlight a difference between transmission measurement error of −0.5% gives us either n = 1.64
the measurements with an uncoated sample and without it, on n0 = 1.54 or n = 1.60 − 0.0005i on n0 = 1.50 depending on
based on a simple example. whether the corresponding error is also taken into account for an
Every physical measurement contains some errors that uncoated substrate or not.
can be either noise-type or systematic. By smoothing we
can eliminate a random noise but not the systematic errors. absorption are influenced only a little. As we are interested
During the optical parameters computation the systematic in the effects of the refractive index inhomogeneity, i.e. the
errors are transferred into parameter errors. There is a simple differences of the index through a layer, the second approach is
model case presented in figure 8. For a substrate we take an preferred here.
index of refraction as 1.50, that for a coated layer, 1.60 − 0i
(no absorption). A small difference of the refractive indices References
ensures here that the errors will be more pronounced.
An upper horizontal line and an upper curve correspond to [1] Tang H, Prasad K, Sanjines R, Schmid P E and Levy F 1994
J. Appl. Phys. 75 2042–7
the theoretical transmission results for an uncoated substrate
[2] Aarik J, Aidla A, Kiisler A-A, Uustare T and Sammelselg V
and a substrate with 375 nm nonabsorbing layer on it. No 1997 Thin Solid Films 305 270–3
influence of the backside of a substrate is accounted for in [3] Bendavid A, Martin P J, Jamting A and Takikawa H 1999 Thin
this figure. If a measurement contains a systematic error of Solid Films 355–356 6–11
−0.5% T, all the measured values will accordingly be smaller. [4] Dannenberg R and Greene P 2000 Thin Solid Films 360 122–7
They are presented as a down horizontal line with the two [5] Rodrigues J, Gomez M, Ederth J, Niklasson G A and
Granqvist C G 2000 Thin Solid Films 365 119–25
curves close to each other. A horizontal line can be presented as [6] Won D-J, Wang C-H, Jang H-K and Choi D J 2001 Appl.
an equivalent to an uncoated substrate with the refractive index Phys. A 73 595–600
n0 = 1.54; the two curves present the cases of the absorbing [7] Aarik J, Aidla A, Mändar H, Uustare T, Schuisky M and
material n = 1.60 − 0.0005i on a substrate with n0 = 1.50 Harsta A 2002 J. Cryst. Growth 242 189–98
and a nonabsorbing material n = 1.64 on a substrate with [8] Aarik J, Aidla A, Uustare T and Sammelselg V 1995 J. Cryst.
Growth 148 268–75
n0 = 1.54. We obtain the first case, if transmission modelling [9] Kumagai H, Matsumoto M, Toyoda K, Obara M and Suzuki M
is made with the tabulated substrate refractive index values 1995 Thin Solid Films 263 47–53
without measuring it, and the second case, if the coated and [10] Leprince-Wang Y, Souche D, Yu-Zhang K, Fisson S, Vuye G
uncoated specimens are measured together. In the first case, and Rivory J 2000 Thin Solid Films 359 171–6
the errors of the transmission measurement are transferred into [11] Mitchell D R G, Attard D J, Finnie K S, Triani G, Barbe C J,
Depagne C and Barlett J R 2005 Appl. Surf. Sci. 243
the absorption; in the second case, the refractive indices of 265–77
both the substrate and the coated layer will take the values [12] Zaitsu S, Motokoshi S, Jitsuno T, Nakatsuka M and
different from the original ones, but their relationship and Yamanaka T 2002 Japan. J. Appl. Phys. 41 160–5

59
A Kasikov et al

[13] Chao S, Wang W-H and Lee C-C 2001 Appl. Opt. 40 [23] Lagarec K and Desgreniers S 1995 Solid State Commun. 94
2177–82 519–24
[14] Honda T, Yanashima K, Yoshino J, Kukimoto H, Koyama F [24] Grey I E, Li C, Madsen I C and Braunshausen G 1988 Mater.
and Iga K 1994 Japan. J. Appl. Phys. 33 3960–1 Res. Bull. 23 743–53
[15] Tappura K, Aarik J and Pessa M 1996 IEEE Photon. Technol. [25] McCartney M R and Smith D J 1991 Surf. Sci. 250 169–78
Lett. 8 319–21 [26] Aarik J, Aidla A, Sammelselg V and Uustare T 1997 J. Cryst.
[16] Battaglin C, Caccavale F, Menelle A, Montecchi M, Growth 181 259–64
Nichelatti E W, Niccoletti F and Polato P 1999 Thin Solid [27] Aarik J, Aidla A, Mändar H and Uustare T 2001 Appl. Surf.
Films 351 176–9 Sci. 172 148–58
[17] Mosaddeq-ur-Rahman Md, Yu G, Krishna K M, Soga T, [28] Aarik J, Karlis J, Mändar H, Uustare T and Sammelselg V
Watanabe J, Jimbo T and Umeno M 1998 Appl. Opt. 37 2001 Appl. Surf. Sci. 181 339–48
691–7 [29] Mitchell D R G, Attard D J and Triani G 2003 Thin Solid
[18] Mosaddeq-ur-Rahman Md, Yu G, Soga T, Jimbo T, Ebisu H Films 441 85–95
and Umeno M 2000 J. Appl. Phys. 88 4634–41 [30] Sun L and Hou P 2004 Thin Solid Films 455–456 525–9
[19] Moret M P, Zallen R, Vijay D P and Desu S B 2000 Thin Solid [31] Bhattacharyya D, Sahoo N K, Thakur S and Das N C 2000
Films 366 8–10 Thin Solid Films 360 96–102
[20] Simons P Y and Dachille F 1967 Acta Crystallogr. 23 334–6 [32] Kasikov A and Kuznetsov A S 1994 J. Phys. D: Appl. Phys. 27
[21] Nicol M and Fong M Y 1971 J. Chem. Phys. 54 3167–70 2470–4
[22] Sato H, Endo S, Sugiyama M, Kikegawa T and Shimomura O [33] Dobrowolski J A, Ho F C and Waldorf A 1983 Appl. Opt. 22
1991 Science 251 786–8 3191–200

60