Formation of Al2O3-Ta2O5 double-oxide thin films by low-pressure MOCVD and evaluation of their corrosion resistances in acid and alkali solutions

Article abstract of DOI:10.1149/1.1391636

Al₂O₃-Ta₂O₅ double-oxide thin films with different cationic mole fractions of Ta, X_Ta, were formed by metallorganic chemical vapor deposition using aluminum tetraisopropoxide [Al(O-i-C₃H

Full text of DOI:10.1149/1.1391636

5
10
Journal of The Electrochemical Society, 146 (2) 510-516 (1999)
S0013-4651(98)04-098-1 CCC: $7.00 © The Electrochemical Society, Inc.
Formation of Al O –Ta O Double-Oxide Thin Films by Low-Pressure
2
3
2
5
MOCVD and Evaluation of Their Corrosion Resistances in Acid and
Alkali Solutions
Nobuyoshi Hara,* Shintaro Nagata, Noboru Akao, and Katsuhisa Sugimoto*
Department of Metallurgy, Graduate School of Engineering, Tohoku University, Sendai 980-8579, Japan
Al O –Ta O double-oxide thin films with different cationic mole fractions of Ta, XTa, were formed by metallorganic chemical
2
3
2 5
vapor deposition using aluminum tetraisopropoxide [Al(O-i-C H ) ] and tantalum pentametoxide [Ta(OCH ) ] as source gases and
3
7 3
3 5
oxygen as a reaction gas. The corrosion resistances of the films were examined in 6 and 12 M HCl and 1 M NaOH at 298 K by
the ellipsometric measurement of thinning rates of the films. The Al O –Ta O films having X values between 0.0 and 1.0 showed
2
3
2
5
Ta
homogeneous amorphous structures. The thinning rate of the films in HCl solutions decreased with increasing XTa value and
reached the analytical limit of ellipsometry. The films with XTa values larger than 0.35 hardly dissolved in 6 M HCl and those with
XTa values larger than 0.45 in 12 M HCl. The corrosion resistance of the Al O –Ta O films in 1 M NaOH increased significant-
2
3
2 5
ly with an increase in XTa: the dissolution rate of the Al O –Ta O film with X value of about 0.5 is lower than that of pure Al O
2
3
2
5
Ta
2 3
film by six orders of magnitude.
1999 The Electrochemical Society. S0013-4651(98)04-098-1. All rights reserved.
©
Manuscript submitted April 29, 1998; revised manuscript received October 16, 1998.
rate in 1 M HCl.17,18 In the present study we examined the corrosion
resistance of Al O –Ta O films in HCl and NaOH solutions as a
function of film composition. The results provide a better under-
standing of the passivation behavior of Al–Ta alloys in HCl solutions
and give a clue to the mechanism responsible for the improved pit-
ting resistance of the alloys in neutral chloride solutions.
In an earlier series of papers, thin films of double oxides having
two or more different kinds of cations were synthesized by metallor-
ganic chemical vapor deposition (MOCVD) and utilized as artificial
passivation films for comparison to passive films formed on corro-
2
3
2 5
1-9
sion-resistant alloys. Since the artificial passive films are prepared
on an inert substrate, it is possible to elucidate the quantitative rela-
tionship between the corrosion resistance and the composition of the
films without any influence of substrates. For example, Fe O –Cr O
Experimental
2
3
2 3
Synthesis of Al O –Ta O films.—Al O –Ta O films were pre-
and Fe O –Cr O –NiO films, which simulate passive films on ferrit-
2
3
2
5
2
3
2 5
2
3
2 3
pared by MOCVD. Aluminum tetraisopropoxide [Al(O–i–C H ) ,
ic and austenitic stainless steels, respectively, were formed on Pt sub-
strates by MOCVD and their dissolution rates in acid solutions were
3 7 3
9
9.999 wt %] and tantalum pentametoxide [Ta(OCH3)5,
99.999 wt %] were employed as vapor sources and heated at 383 and
13 K, respectively. Nitrogen (99.999 vol %) was used as a carrier
gas of each vapor source. The flow rate of carrier gas for each vapor
1-3
examined as a function of film composition and applied potential.
4
The results showed that the dissolution rates of the films under
cathodic polarization and open-circuit conditions decreased expo-
Ϫ6
3 Ϫ1
nentially with increasing amount of Cr3ϩ cations.1-3 Similar compo-
source was changed in the range 0 to 6.7 ϫ 10 m s , while the
total flow rate of the carrier gases was kept constant at 6.7 ϫ
sition-dependent changes in the dissolution rate were found on real
passive films on ferritic stainless steels, while the absolute dissolu-
tion rate of the real passive films was always higher than that of the
Ϫ6
3 Ϫ1
1
0
m s . Oxygen (99.9 vol %) was used as a reactant gas and its
Ϫ6
3 Ϫ1
flow rate was kept constant at 8.3 ϫ 10 m s . The formation of
the films was performed at a substrate temperature of 623 K. At this
temperature, homogeneous amorphous Al O and Ta O films are
1
artificial passivation films. The other studies on sputter-deposited
10
11
12,13
Fe O , Cr O , and Fe O –Cr O
thin films demonstrated
2
3
2 5
2
3
2
3
2
3
2
3
20,21
deposited from Al(O–i–C H ) and Ta(OCH ) , respectively.
that the reductive dissolution of Fe O and the oxidative dissolution
3
7 3
3 5
2
3
Plates of polycrystalline Pt and single-crystal Si(100) in dimen-
sions of 25 ϫ 15 ϫ 0.5 mm were used as substrates. The surface of
the Pt plates was finished with 1 ␮m diamond paste and degreased
ultrasonically in acetone. Native oxide films on the Si(100) plates
were removed by dipping in 2 wt % HF just before each experiment.
Figure 1 shows the schematic of the MOCVD system used.
Source gases were evaporated in two evaporators and carried by the
carrier gas. After the source gases were mixed with oxygen in a
pipeline they were introduced into a cold wall CVD reactor. The tub-
ing between the evaporators and the reactor was heated at 413 K to
prevent the recondensation of source gases. The total pressure of the
CVD reactor was kept at 2 kPa. The substrate was mounted on a
holder with a built-in heater. Three baffle plates were used to get
homogeneous gas flow. Two small square holes on a cover of the
holder and two optical windows of the rector cell allowed optical
access to a substrate during the deposition under vacuum. The com-
position of the films was controlled by the flow rate of the carrier gas
for each vapor source. The thickness of the films was adjusted to an
appropriate value by monitoring the film growth and controlling the
deposition time.
of Cr O were suppressed by the presence of Cr O and Fe O in the
2
3
2
3
2 3
films, respectively. It is noteworthy that the corrosion resistance does
not vary linearly with film composition: even small amounts of a cor-
rosion-resistant oxide component in the films can lead to a signifi-
cant improvement of the corrosion resistance. In the case of
4
5
6-8
ZrO –Ta O , ZrO –TiO , and Fe O –TiO films, relatively
2
2
5
2
2
2
3
2
small amounts of Ta O and TiO improve the corrosion resistance in
2
5
2
9
strong acid solutions. The presence of ZrO in Al O –ZrO films is
2
2
3
2
effective in improving corrosion resistances in both acid and alkali
solutions. The objective of the present work is to extend these stud-
ies to obtain Al O –Ta O double-oxide films having high corrosion
2
3
2 5
resistances against both acid and alkali solutions.
The study on the corrosion resistance of Al O –Ta O thin films
2
3
2 5
will also provide new insight into the passivity of Al–Ta alloys. In
recent years there has been much interest in nonequilibrium solid-
solution Al–Ta alloys prepared by sputter deposition, because the
alloys show an excellent corrosion resistance in solutions containing
1
4-19
chloride ions.
atom % Ta alloy in 0.1 M KCl at pH 7 is higher than that of pure
The pitting potential of a sputter-deposited Al-
8
1
6
Al by about 0.8 V. Sputter-deposited Al–Ta alloys with about
0 atom % Ta behave like pure Ta, showing no pitting up to 1.6 V
vs. SCE) in 0.1 M NaCl at pH 10 and a negligibly small corrosion
4
Ellipsometric monitoring of film growth.—A rotating analyzer
1
5
(
ellipsometer was used to monitor the film formation in situ. The
9
detail of the ellipsometer is described in the previous paper. Briefly,
*
Electrochemical Society Active Member.
monochromatic and linearly polarized light of wavelength 546.1 nm
Downloaded on 2015-05-05 to IP 140.115.130.238 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 146 (2) 510-516 (1999)
511
S0013-4651(98)04-098-1 CCC: $7.00 © The Electrochemical Society, Inc.
dissolved by distilled water to separate the films. The separated films
were picked by Cu mesh (180 mesh) and served for observation.
Evaluation of corrosion resistance.—The corrosion resistance of
Al O –Ta O films formed on Pt substrates was examined in 6 and
2
3
2 5
1
2 M HCl and 1 M NaOH solutions at 298 K. After the films were
immersed in the test solutions for a given time, the decrease in film
thickness was measured by ellipsometry. Ellipsometric measure-
ments were performed under the same conditions as described pre-
viously. The composition of the test solutions was analyzed by ICPS
to examine the dissolved composition of the films. Before the ICPS
Ϫ6
3
analysis, 46 wt % HF solution of 1 ϫ 10 m volume was added to
Ϫ6
3
the test solution of 4 ϫ 10 m volume in order to dissolve Ta O
2
5
groups, which may be detached from the films through the dissolu-
tion of Al O .
2
3
Results and Discussion
Ellipsometric analysis of the film formation process.—Figure 2
illustrates the experimental ⌬ vs. ⌿ plots obtained during the depo-
sition of Al O –Ta O films with X values of 0.16, 0.46, and 0.75.
Figure 1. Schematic of MOCVD system.
2
3
2
5
Ta
was used as incident light and the angle of incidence was set at
Solid lines in this figure indicate theoretical ⌬ vs. ⌿ curves calcu-
lated for the growth of films having different optical constants, N2 ϭ
n2 Ϫ k2i. The experimental plot for the film of XTa ϭ 0.16 fits well
the theoretical curve for N2 ϭ 1.51 Ϫ 0.00i, indicating no change in
optical constants of the film during the deposition. Similarly, the
experimental plots for the films of XTa ϭ 0.46 and 0.75 fit the theo-
retical curves calculated for the growth of films with N2 ϭ 1.66 Ϫ
0.00i and N2 ϭ 1.83 Ϫ 0.00i, respectively. The real part of optical
constant N2, i.e., refractive index n2, of the film increases with
6
0.00Њ. The intensity of light, which was reflected from the speci-
men surface and then passed through the rotating analyzer, was
detected and digitized as a function of the angle of the analyzer by
using a photodiode and related circuits. Computer-aided Fourier
analysis of this signal provided two ellipsometric parameters, the
relative phase retardation, ⌬, and the arc tangent of the relative
amplitude reduction, ␺. The measured ellipsometric parameters can
2
2
be related to the complex reflectance ratio given by
increasing XTa value, while the imaginary part of N , i.e., extinction
2
␳
ϭ R /R ϭ tan ⌿иexp(i⌬)
[1]
coefficient k , stays zero, indicating that films formed are optically
p
s
2
transparent. From the fact that n values stay constant during the
2
where R and R are the complex amplitude reflection coefficients for
p
s
deposition of the films, it is suggested that the in-depth composition
of the films is homogeneous. The ellipsometric monitoring of film
growth allowed control of film thickness to a given value. In the
experiments shown in Fig. 2, the supply of source gases was stopped
when the film thickness reached ϳ80 nm.
light polarized parallel (p) and perpendicular (s) to the plane of inci-
dence. The thickness and optical constants of deposited films were de-
termined using a theoretical ⌬ vs. ␺ curve which fits the experimental
data (⌬, ␺) with minimal error. The theoretical ⌬ vs. ␺ curves were
calculated using Drude’s exact optical equations which express Rp
and R for a three-medium (ambient-film-substrate) model.
Micro- and crystal structures.—Figure 3 shows TEMs and ED
s
patterns for Al O –Ta O films having different X values, which
2
3
2
5
Ta
ICPS analysis.—The chemical composition of oxide films was
determined by chemical analysis. The oxide films deposited on the
Si(100) plates were dissolved into 46 wt % HF solution of 1 ϫ
were formed at 623 K. All the films with XTa values between 0 and
.0 have no crystal grains and show homogeneous structures. The
ED patterns show unclear ring patterns, indicating that every film
1
Ϫ5
3
Ϫ5
3
1
0
m volume and then the solution was diluted to 5 ϫ 10
m
2
0
has an amorphous structure. It has been reported that CVD-Al O
2
3
by adding distilled water. This diluted solution was analyzed by
inductively coupled plasma-emission spectroscopy (ICPS). The
results of ICPS analysis give the mass of Al and Ta in the films, WAl
2
1
and Ta O films, which are deposited at 523-773 K from Al(O–i–
2
5
and W . The cationic mole fraction of Ta in the film, X , was cal-
Ta
Ta
culated from WAl and WTa using the following equation
XTa ϭ (W /M )/(W /M ϩ W /M )
[2]
Ta
Ta
Al
Al
Ta
Ta
where MAl and MTa are atomic weight of Al and Ta, respectively.
XPS analysis.—The chemical states of constituent elements of
Al O –Ta O films deposited on Pt plates were analyzed by X-ray
2
3
2 5
photoelectron spectroscopy (XPS). A Mg K␣1,2 X-ray source
1253.1 eV) was operated at 5 kV and 30 mA. Photoelectron spectra
(
of Al 2s, Ta 4d, O 1s, and C 1s levels were measured. The measure-
Ϫ5
ment of spectra was performed at a pressure of 5 ϫ 10 Pa and at
room temperature. For calibration of the photoelectron binding ener-
gy, the C 1s peak appearing at 285.0 eV in the background spectra
was used. To examine chemical states and quantities of film con-
stituent elements, measured spectra of O 1s and C 1s levels were
approximated using the Gauss-Lorentz mixed function and those of
Al 2s and Ta 4d levels were compared with the reference spectra
taken from reagent-grade Al O and Ta O powder.
2
3
2 5
Transmission electron microscopy observation.—Micro- and crys-
tal structures were examined by transmission electron microscopy
Figure 2. Experimental ⌬-⌿ loci for formation of Al O –Ta O films and
2
3
2 5
(
1
TEM) and electron diffraction (ED), respectively. Oxide films ca.
50 nm thick were deposited on KBr disks and then the disks were
theoretical ⌬-⌿ curves for growth of films with different optical constants,
N (ϭn Ϫ k i).
2
2
2
Downloaded on 2015-05-05 to IP 140.115.130.238 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

5
12
Journal of The Electrochemical Society, 146 (2) 510-516 (1999)
S0013-4651(98)04-098-1 CCC: $7.00 © The Electrochemical Society, Inc.
C H ) and at 473-773 K from Ta(OCH ) , respectively, have an
Al O and Ta O . Good coincidence between the measured and the
2 3 2 5
3
7 3
3 5
amorphous structure.
reference spectra indicates that Al and Ta in the film exist as triva-
lent and pentavalent ions, respectively.
Oxidation states of film constituent elements.—Figure 4 shows
XPS spectra of Al O –Ta O films with different X values. The
Figure 5c shows the decomposition of the O 1s spectrum for the
Al O –Ta O film with X ϭ 0.19. The O 1s spectrum consists of
2
3
2
5
Ta
2
3
2
5
Ta
Al 2s spectra of the films exhibit a peak at 119.3 Ϯ 0.1 eV (Fig. 4a),
and the Ta 4d spectra two peaks at 230.7 Ϯ 0.1 eV and 242.4 ±
contributions from three species having peaks at 531.10, 532.35, and
533.80 eV. The binding energies of these peaks are close to those for
oxides (Al O : 530.95 eV, Ta O : 530.50 eV), a hydroxide
0
.1 eV (Fig. 4b). The binding energies of these peaks are almost in-
2
3
2
5
dependent of XTa value. On the other hand, the peak energy of the
O 1s spectrum shifts from 531.7 eV at XTa ϭ 0.00 (Al O ) to
(Al(OH) : 532.30 eV), and adsorbed H O (533.65-533.80 eV),
3
2
2
Ϫ
respectively. Therefore, the oxygen species in the film are O
,
2
3
Ϫ
Ϫ
5
31.0 eV at XTa ϭ 1.00 (Ta O ) with increasing X (Fig. 4c). The
OH , and H O. The ratio of the peak area for OH species to that
2
5
Ta
2
Ϫ
2Ϫ
2Ϫ
shift of the O 1s peak can be ascribed to the change in the OH /O
ratio of the films.
for O species, S /S , depends on the film composition: the values
OH
O
of S /S for the films with X ϭ 0.00, 0.18, 0.48, and 1.00 are 0.76,
OH
O
Ta
In Fig. 5a and b, the Al 2s and Ta 4d spectra for the Al O –Ta O
film with XTa ϭ 0.19 are compared with reference spectra from pure
0.54, 0.15, and 0.14, respectively. In the previous study using
infrared spectroscopy, it has been demonstrated that CVD-Al O
2
3
2 5
2
3
Figure 3. TEM (left) and ED patterns
(
right) for Al O –Ta O films of XTa ϭ
2 3 2 5
0
1
.00 (Al O ) (a), 0.40 (b), 0.68 (c), and
2 3
.00 (Ta O ) (d).
2
5
Downloaded on 2015-05-05 to IP 140.115.130.238 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 146 (2) 510-516 (1999)
513
S0013-4651(98)04-098-1 CCC: $7.00 © The Electrochemical Society, Inc.
Figure 4. XPS spectra for Al O –Ta O films deposited at 623 K.
2
3
2 5
films deposited from Al(O–i–C H ) at 623 K had little or no Al–OH
Corrosion resistance against HCl solution.—Figure 6 illustrates
changes in ellipsometric parameters, ⌬ and ⌿, with immersion time
for an Al O –Ta O film of X ϭ 0.12 in 6 M HCl at 298 K. Theo-
3
7 3
2
0
bond. Therefore, hydroxyl groups detected by XPS may arise from
Ϫ
adsorbed OH species on the surface of films, which is formed by
2
3
2
5
Ta
water adsorption after the film formation.
retical ⌬-⌿ curves calculated for the dissolution of films with dif-
Figure 5. Comparison of Al 2s and Ta 4d spectra between Al O –Ta O film of (a, b) X ϭ 0.19 and standard materials, Al O and Ta O , and (c) decomposi-
2
3
2
5
Ta
2
3
2 5
tion of O 1s spectrum of the film.
Downloaded on 2015-05-05 to IP 140.115.130.238 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

5
14
Journal of The Electrochemical Society, 146 (2) 510-516 (1999)
S0013-4651(98)04-098-1 CCC: $7.00 © The Electrochemical Society, Inc.
ferent optical constants, N ϭ n Ϫ k i, are also given in Fig. 6. The
2
2
2
experimental ⌬-⌿ locus moves from a theoretical curve for a high
optical constant, N ϭ 1.68 Ϫ 0.00i, to that for a low optical con-
2
stant, N ϭ 1.55 Ϫ 0.00i. This decrease in the optical constant may
2
be due to an increase in the surface roughness of the film with the
progress of dissolution. A two-layer model, in which the film is
assumed to consist of an outer rough layer and an inner compact
layer, is better than a single, uniform layer model for determining the
thickness changes during the dissolution. However, there are three or
more unknown parameters in the two-layer model, and these param-
eters cannot be determined from two measured quantities, ⌬ and ⌿.
The following method was therefore employed to estimate the film
thickness changes. First, an appropriate theoretical ⌬ vs. ␺ curve
which provides the best fit to the experimental (⌬, ␺) data through
the whole test duration was searched. In the case of Fig. 6, the theo-
retical curve for the optical constant of N ϭ 1.63 Ϫ 0.00i was
2
selected as the best fit curve. Then the thickness of the film remain-
ing at each stage of dissolution was determined by drawing a per-
pendicular to the theoretical curve from the experimental plot. This
method of film thickness determination leads to a small overestima-
tion of the thickness of the film at the initial stage of dissolution. For
the film at t ϭ 0 s in Fig. 6, the thickness determined using the the-
oretical curve for N ϭ 1.68 Ϫ 0.00i, which is the optical constant
2
of as-formed film, is 85.5 nm; the thickness estimated using the the-
oretical curve for N ϭ 1.63 Ϫ 0.00i, which is the optical constant
2
selected for the thickness analysis during the dissolution, is 89.5 nm.
However, the error due to the overestimation never exceeded 5%.
Figure 7 shows the change in film thickness, ⌬d, in 6 M HCl at
Figure 7. Decrease in film thickness, ⌬d, as a function of time, t, for
2
98 K as a function of immersion time, t, for Al O –Ta O films of
2 3 2 5
Al O –Ta O films in 6 M HCl at 298 K.
2
3
2 5
XTa ϭ 0.0 Ϫ 1.0. The thickness of the films with XTa values of 0.0-
.25 decreases linearly with time. One can determine the thinning
0
rate of film thickness, Ϫ⌬d/⌬t, from the gradient of ⌬d vs. t curves.
The thinning rates thus determined are plotted against XTa in
Fig. 8 where the corresponding results obtained in 12 M HCl at
To examine whether the selective dissolution of film constituent
elements occurs, the amount of metal ions dissolved in the test solu-
tion was measured by ICPS. Figure 9 exhibits the amount of dis-
2
98 K are also displayed. When the value of XTa exceeds about 0.1,
solved Al and Ta ions, WAl and W , as a function of the exposure
Ta
the thinning rate of film thickness starts to decrease markedly with
time, t, for the Al O –Ta O film of X ϭ 0.25 in 6 M HCl at 298 K.
2
3
2
5
Ta
increasing XTa value. The XTa values at which the thinning rate reach-
Ϫ6
Ϫ1
es an analytical limit of ellipsometry, 2.8 ϫ 10 nm s , are 0.35
and 0.45 in 6 and 12 M HCl solutions, respectively. The films hav-
ing XTa values larger than 0.45 show superior corrosion resistance
even in 12 M HCl at 298 K.
Figure 6. Experimental ⌬-⌿ locus for dissolution of Al O –Ta O film of
XTa ϭ 0.12 in 6 M HCl at 298 K and theoretical ⌬-⌿ curves for films with
different optical constants, N2.
2
3
2
5
Figure 8. Thinning rate of film thickness, Ϫ⌬d/⌬t, as a function of cationic
mole fraction of Ta, X , for Al O –Ta O films in 6 M HCl and 12 M HCl
at 298 K.
Ta
2
3
2 5
Downloaded on 2015-05-05 to IP 140.115.130.238 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 146 (2) 510-516 (1999)
515
S0013-4651(98)04-098-1 CCC: $7.00 © The Electrochemical Society, Inc.
The dashed curve in Fig. 9 indicates the WAl vs. t curve calculated
from the experimental WTa vs. t curve under the assumption that no
selective dissolution occurs. The experimental WAl values are about
three times the calculated ones, indicating that the selective dissolu-
3ϩ
tion of Al ions takes place during the immersion. The selective
3
ϩ
dissolution of Al ions probably produces a porous or rough sur-
face layer, where Ta O is enriched, in the surface region of residual
2
5
films. As a result, the optical constants of Al O –Ta O films change
2
3
2 5
with immersion time (Fig. 6). Under the assumption that a rough sur-
face layer is optically equivalent to a mixture of oxides and voids,
the refractive index of the surface layer, n , is given as a function of
s
the refractive indices of oxides and voids, nox and n (ϭ1), and the
v
volume fractions of these components, fox and f (ϭ1 Ϫ f ). As
v
ox
expected from Fig. 2, nox increases as Ta O is enriched by the selec-
2
5
tive dissolution of Al, which would result in an increase in n if there
s
is little or no surface roughening (f Ϸ 0). n is expected to decrease
v
s
when the dissolution of Al O –Ta O films is followed by signifi-
2
3
2 5
cant surface roughening (f > 0). A decrease in n leads to a decrease
v
s
in averaged optical constants of the films remaining at each stage of
dissolution. The observed decrease in optical constants with time
(Fig. 6) should show that the surface of the films becomes rough in
the progress of dissolution.
Corrosion resistance against NaOH solution.—Changes in the
ellipsometric parameters during the dissolution of Al O –Ta O
2
3
2 5
films in 1 M NaOH were similar to those observed in 6 M HCl
Fig. 6). Figure 10 shows the change in film thickness, ⌬d, in 1 M
NaOH at 298 K as a function of immersion time, t, for Al O –Ta O
films of XTa ϭ 0.0 Ϫ 1.0. The thickness of the films with XTa values
of 0.0-0.53 decreases linearly with time. The thinning rate of film
thickness, Ϫ⌬d/⌬t, for each film was determined from the gradient
of the corresponding ⌬d vs. t curve.
(
2
3
2 5
Figure 10. Decrease in film thickness, ⌬d, as a function of time, t, for
Al O –Ta O films in 1 M NaOH at 298 K.
2
3
2 5
Ta O double-oxide films with X values larger than 0.5 have excel-
2
5
Ta
Figure 11 shows the thinning rates of film thickness for Al O –
lent corrosion resistance against both the acid and alkali solutions.
2
3
Ta O films in 1 M NaOH at 298 K as a function of X . In 1 M
2
5
Ta
NaOH, the film with XTa ϭ 0.0 (Al O ) is very susceptible to corro-
Insight into passivity and its breakdown of Al-Ta alloys.—It has
2
3
sion, because the thinning rate in this solution is higher than that in
M HCl by a factor of 1,000. The thinning rate of the Al O –Ta O
been reported that sputter-deposited Al–Ta alloys containing more
1
4
6
than 14 atom % Ta passivate spontaneously in 1 M HCl at 303 K.
2
3
2 5
films, however, decreases significantly with increasing X , and at
Ta
XTa values larger than 0.5 it becomes six orders of magnitude lower
than that of the film with XTa ϭ 0.0 (Al O ) . The films with X val-
2
3
Ta
ues larger than 0.5 behave like pure Ta O film. Taking account of
2
5
the results obtained in HCl solutions, it is concluded that the Al O –
2
3
Figure 9. Amount of dissolved Al and Ta ions, WAl and W , as a function of
Ta
time, t, for Al O –Ta O film of X ϭ 0.25 in 6 M HCl at 298 K. Dashed
curve indicates WAl vs. t relationship calculated by assuming uniform (nons-
elective) dissolution of the film.
2
3
2
5
Ta
Figure 11. Thinning rate of film thickness, Ϫ⌬d/⌬t, as a function of cation-
ic fraction of Ta, X , for Al O –Ta O films in 1 M NaOH at 298 K.
Ta
2
3
2 5
Downloaded on 2015-05-05 to IP 140.115.130.238 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

5
16
Journal of The Electrochemical Society, 146 (2) 510-516 (1999)
S0013-4651(98)04-098-1 CCC: $7.00 © The Electrochemical Society, Inc.
As a result, the corrosion rate of the alloys decreases significantly
with increasing Ta content, and the alloy containing 53 atom % Ta
0.35 and 0.45 exhibited superior corrosion resistance against 6 M
HCl and 12 M HCl solutions at 298 K, respectively.
1
7,18
hardly dissolves in 1 M HCl at 303 K.
The results of XPS analy-
sis showed that the fraction of Ta in the passive films formed in
M HCl is almost the same as that of Ta in the alloys.14 These obser-
vations agree with the results obtained in the present study: Fig. 8
4. The thinning rate of the Al O –Ta O films in 1 M NaOH
2
3
2 5
5
ϩ
decreased significantly with increasing XTa value and at XTa larger
than 0.5 it becomes six orders of magnitude lower than that of pure
Al O film.
1
2
3
3
ϩ
5ϩ
reveals that double-oxide films consisting of Al and Ta cations
Tohoku University assisted in meeting the publication costs of this article.
5ϩ
become protective with increasing Ta fraction, X , and the films
Ta
with XTa values of about 0.45 are immune to corrosion even in con-
centrated HCl solutions.
References
1. K. Sugimoto, M. Seto, S. Tanaka, and N. Hara, J. Electrochem. Soc., 140, 1586
(1993).
The pitting potential of sputter-deposited Al–Ta alloys in neutral
2
.
S. Tanaka, N. Hara, and K. Sugimoto, in Corrosion Protection by Coatings and
Surface Modification, M. W. Kendig, K. Sugimoto, and N. R. Soerensen, Editors,
PV 93-28, p. 54, The Electrochemical Society Proceedings Series, Pennington, NJ
(1993).
chloride solutions increases with increasing Ta content of the
alloys.1
5,16,19
Several mechanisms have been proposed to explain the
1
6
improved pitting resistance of Al–Ta alloys. Davis et al. found,
using XPS, that the passive film on Al–Ta alloys is enriched in
3
4
.
.
S. Tanaka, N. Hara, and K. Sugimoto, Mater. Sci. Eng., A198, 63 (1995).
K. Amano and K. Sugimoto, J. Jpn. Inst. Metals, 56, 1192 (1992).
Ta O . They proposed that the Ta O -rich passive film resists the
2
5
2
5
5. S. Kikkawa, N. Hara, and K. Sugimoto, Mater. Sci. Forum, 185-188, 497 (1995).
6. K. Sugimoto, H. Kim, N. Akao, and N. Hara, in Surface Oxide Films, J. A. Bard-
well, Editor, PV 96-18, The Electrochemical Society Proceedings Series, p. 195,
Pennington, NJ (1996).
Ϫ
Cl penetration through the film, improving the resistance to pitting
corrosion. Natishan et al. proposed that the enrichment of Ta O
1
6
23
2
5
in the passive film decreases the pH of zero charge of the film, in-
7
8
.
.
H. Kim, N. Hara, and K. Sugimoto, J. Electrochem. Soc., To be published.
H. Kim, N. Akao, N. Hara, and K. Sugimoto, J. Electrochem. Soc., 145, 2818
(1998).
Ϫ
hibiting the adsorption of Cl and thus enhancing the pitting resist-
2
4
25
ance. On the other hand, Smialowska and Frankel et al. suggest-
ed that the alloying elements in Al- transition metal alloys exert their
9. E. Takahashi, N. Akao, N. Hara, and K. Sugimoto, in Organic and Inorganic Coat-
ings for Corrosion Prevention-Research and Experiences, L. Fedrizzi and P. L.
Bonora, Editors, p. 299, The Institute of Materials, London (1996).
0. S. Virtanen, P. Schmuki, A. J. Davenport, and C. M. Vitus, J. Electrochem. Soc.,
144, 198 (1997).
2
4
influences at active pits rather than passive films. Smialowska pro-
posed that in the case of Al–Cr, Al–Zr, and Al–W alloys, the low sol-
ubility of oxidized alloying elements, CrOOH, ZrO , and WO , in an
acidified micropit impedes pit propagation. Frankel et al. suggest-
ed that alloying elements such as Nb and Mo reduce the ability of
micropits to maintain the critical environment necessary for growth.
As can be seen in Fig. 8, the dissolution rate of Al O –Ta O films
1
2
3
2
5
11. P. Schmuki, S. Virtanen, A. J. Davenport, and C. M. Vitus, J. Electrochem. Soc.,
43, 3997 (1996).
1
1
2. P. Schmuki, S. Virtanen, H. Böhni, H. S. Isaacs, and A. J. Davenport, in Surface
Oxide Films, J. A. Bardwell, Editor, PV 96-18, The Electrochemical Society Pro-
ceedings Series, p. 234, Pennington, NJ (1996).
2
3
2 5
in HCl solutions, which correspond to acidic micropit environments,
13. P. Schmuki, S. Virtanen, H. S. Isaacs, M. P. Ryan, A. J. Davenport, H. Böhni, and
5
ϩ
T. Stenberg, J. Electrochem. Soc., 145, 791 (1998).
decreases markedly with increasing Ta content of the films. It is at
least suggested that the repassivation ability of Al–Ta alloys in acidic
micropit environments increases significantly with increasing Ta
content of the alloys, even though the unique mechanism responsi-
ble for the improved pitting resistance of the alloys cannot be con-
clusively determined from these results.
1
4. H. Yoshioka, A. Kawashima, K. Asami, and K. Hashimoto, in Corrosion, Electro-
chemistry and Catalysis of Metallic Glasses, R. B. Diegle and K. Hashimoto, Edi-
tors, p. 242, The Electrochemical Society, Pennington, NJ (1988).
15. G. S. Frankel, M. A. Russak, C. V. Jahnes, M. Mirzamaani, and V. A.Brusic, J.
Electrochem. Soc., 136, 1243 (1989).
1
6. G. D. Davis, W. C. Moshier, T. L. Fritz, and G. O. Gote, J. Electrochem. Soc., 137,
22 (1990).
4
1
7. H. Yoshioka, Q. Yan, H. Habazaki, A. Kawashima, K. Asami, and K. Hashimoto,
Conclusions
Corros. Sci., 31, 349 (1990).
1
. Al O –Ta O double-oxide films having different cationic
2 3 2 5
18. K. Hashimoto, N. Kumagai, H. Yoshioka, H. Habazaki, A. Kawashima, K. Asami,
fraction of Ta, X , were formed at 623 K by MOCVD using Al(O–i–
and B.-P. Zhang, Mater. Sci. Eng., A133, 22 (1991).
Ta
1
2
2
2
9. C. C. Streinz, J. Kruger, and P. J. Moran, J. Electrochem. Soc., 141, 1126 (1994).
0. T. Go and K. Sugimoto, J. Jpn Inst. Metals, 56, 184 (1992).
1. C. H. An and K. Sugimoto, J. Electrochem. Soc., 139, 1956 (1992).
2. R. M. A. A. Azzam and N. M. Bashara, Ellipsometry and Polarized Light, North-
Holland, New York (1979).
3. P. M. Natishan, E. McCafferty, and G. K.Hubler, J. Electrochem. Soc., 135, 321
(1988).
4. Z. Szklarska-Smialowska, Corros. Sci., 33, 1193 (1992).
C H ) and Ta(OCH ) as source gases.
3
7 3
3 5
2
. The Al O –Ta O films having X values between 0.0 and 1.0
2 3 2 5 Ta
showed homogeneous amorphous structures. Aluminum, tantalum,
3
ϩ
5ϩ
2Ϫ
and oxygen in the films existed in the state of Al , Ta , and O
ions. In the surface region of Al O -rich films, oxygen existed also
2
2
3
as OHϪ ions.
. The thinning rate of the Al O –Ta O films in HCl solutions
2
3
2
3
2
5
25. G. S. Frankel, R. C. Newman, C. V. Jahnes, and M. A. Russak, J. Electrochem. Soc.,
140, 2192 (1993).
decreased with increasing XTa value. The films with XTa values of
Downloaded on 2015-05-05 to IP 140.115.130.238 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).