Spectroscopy study of anodic growth of zirconium oxide film in NaOH medium

Article abstract of DOI:10.1002/pssa.200461162

The growth of anodic oxide films on zirconium metal has been followed up to 300 V by electrochemical impedance spectroscopy and scanning electron microscopy. The maximum layer thickness is 720 nm, the dielectric constant of the film is measured at 19.5 and the growth constant is 2.4 nm V^-1. Above 50 V, the presence of two impedance relaxations between 1 Hz and 200 kHz reveals a bilayered structure. This may be a consequence of a lower resistivity of the outer layer induced by some electrolytic solution infiltration into film defects.

Full text of DOI:10.1002/pssa.200461162

phys. stat. sol. (a) 202, No. 8, 1502–1507 (2005) / DOI 10.1002/pssa.200461162
Impedance spectroscopy study of anodic growth
of zirconium oxide film in NaOH medium
T. Pauporté*, J. Finne, and D. Lincot
Laboratoire d’ Électrochimie et Chimie Analytique, UMR 7575, École Nationale Supérieure de Chimie,
11 rue P. et M. Curie, 75231 Paris cedex 05, France
Received 29 July 2004, revised 17 November 2004, accepted 27 January 2005
Published online 8 June 2005
PACS 68.37.Hk, 81.07.Bc, 82.45.Yz, 82.80.Fk
The growth of anodic oxide films on zirconium metal has been followed up to 300 V by electrochemical
impedance spectroscopy and scanning electron microscopy. The maximum layer thickness is 720 nm, the
dielectric constant of the film is measured at 19.5 and the growth constant is 2.4 nm V–1. Above 50 V, the
presence of two impedance relaxations between 1 Hz and 200 kHz reveals a bilayered structure. This may
be a consequence of a lower resistivity of the outer layer induced by some electrolytic solution infiltration
into film defects.
© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction
ZrO2 is an insulator compound with a large bandgap at 5 eV [1]. This oxide present potential applications
in various fields such as a protective coating [2], dielectric compound in MOS devices or as high tem-
perature O2– conducting electrolyte, important in sensors and fuel cell applications [3]. The growth of
thick oxide films can be obtained by sputtering, pulsed laser deposition, chemical vapor deposition,
atomic layer epitaxy [3] and also by wet methods such as anodization in which the metal, used as an
anode, is oxidized electrochemically in an electrolytic solution [4, 5]. In the case of dielectric oxide
films, a high electric field must be maintained in the layer in order to control the kinetic of the growth by
anodization. It is well-documented that in the case of zirconia the growth rate is almost exclusively lim-
ited by O2– migration from the film/electrolyte interface towards the metal/oxide interface where O2–
reacts with the oxidized zirconium, Zr4+. The transport number of Zr4+ is almost null [4, 5].
Zr ⇒ Zr4+ + 4e– ,
Zr4+ + 2O2– ⇒ ZrO2 .
(1a)
(1b)
In the present paper we take the case of the anodic growth of zirconia in NaOH since this medium is
known to give rise to non-contaminated oxide films [6]. We describe a non-destructive procedure, based
on electrochemical impedance spectroscopy (EIS) measurements, to follow the variation of layer thick-
ness and microstructure with the applied potential.
2 Experimental
The electrodes were zirconium rods of purity 99.8% (Goodfellow), 1.5 mm in diameter and 5 to 10 mm
2
in length (the electrode surface areas exposed to the electrolyte were ranging between 0.25 and 0.5 cm ).
* Corresponding author: e-mail: pauporte@ext.jussieu.fr, Phone: 331 55 42 63 75, Fax: 331 44 27 67 50
© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Original
Paper
phys. stat. sol. (a) 202, No. 8 (2005) / www.pss-a.com
1503
They were sealed in a glass tube with an epoxy resin. Several sealing modes were tested but this configu-
ration gave the best results in terms of highest voltage attained. After a first mechanical polishing with
P400 and P1200 abrasive paper they were finally polished with a 3 µm diamond powder on a metallurgi-
cal cloth. Then the electrodes were degreased in acetone and ethanol 5 minutes each in an ultrasonic bath
before to be chemically etched 2 minutes in a solution of (1:15:34) volume mixture of reagent grade
48% HF:68% HNO3 :Millipore quality water. After the etching, an impedance spectrum was recorded at
a potential close to the rest potential, that is –1 V vs. MSE (for mercurous sulfate electrode, with a poten-
tial of +0.65 V vs NHE) in the anodization medium in order to control the initial electrode surface state.
The surface capacitance of the etched electrode was determined from this spectrum. An etched electrode
was used only once and each new experiment was performed with a new rod. The aqueous electrolytic
solution was 0.1 M NaOH (reagent grade, Prolabo) (pH 13). The anodization was performed in a two
electrodes configuration, the counter electrode being a cylindrical titanium grid covered by an IrO2 coat-
ing. The Zr rod was placed in the center of the cylinder. The applied potential during anodization, noted
E, will be referred to the IrO2 electrode. The solution was stirred with a magnetic barrel. The potential
sweep was started at 0 V vs. IrO2 and performed at a constant sweep rate, noted vb, classically 100 mV s–1
or 25 mV s–1 with a Keitley 2410 sourcemeter. The voltage sweep was stopped at different potentials and
an impedance spectrum was recorded at 0 V vs. MSE. The voltage sweep was then continued and so on.
The AC measurements were done with an EG & G PAR 283 potentiostat combined with a Solartron
1255 model frequency response analyser both monitored by the Zplot-2 software from Scribner Associ-
ate. The frequency range was 1 or 10 Hz to 200 kHz with 10 points per decade. The AC amplitude was
30 mV. The spectra were fitted and analysed with help of the Zview-2 software package. The layer cross
sections were observed with a scanning electron microscope (SEM) (Stereoscan 440 from Leica).
3 Results and discussion
In Fig. 1 is presented the variation of the current density with the potential for two different scan rates,
25 mV s–1 and 100 mV s–1. The breakdown occurs at 300 V where a dramatic increase of the current den-
sity is observed and the anodisation can not be continued. The breakdown is not due to film homogene-
ous failure but to an intense corrosion localized near the seal (seal failure). The electrolytic solution is
brought in contact with the bare metal and a violent current leakage is produced. During this process a
flame is observed, located in the vicinity of the epoxy seal.
A large current peak (noted A) is observed at low voltage which can be ascribed to the oxygen evolu-
tion reaction occurring before the passivation of the surface. Afterwards, the current density is not con-
stant but slowly increases with applied voltage. The maximum current density, j, at 100 mV s–1 before the
8
A
6
Breakdown
4
2
Fig. 1 Variation of the current density during a
potential scan at 100 mV s–1 (+) and 25 mV s–1 (×).
0
0
50
100
150
200
250
300
350
E / V vs IrO2
© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1504
T. Pauporté et al.: Impedance spectroscopy study of anodic growth of zirconium oxide film
106
105
104
1000
100
10
b
as-etched
Porous
Oxide
film
Dense
oxide
film
Metal
50V
100V
200V
295V
CPE
2
1
1
10
100
1000
104
105
106
frequency / Hz
-100
-80
-60
-40
-20
0
R
1
CPE1
R
2
c
as-etched
50V
R1
CPE
CPE3
1
100V
200V
295V
1
10
100
1000
104
105
106
frequency / Hz
Fig. 2 (a) Bode representation of the effect of the increasing applied voltage (etched, 50 V, 100 V,
200 V and 295 V) on the impedance; (b) equivalent circuit used for the analysis of the EIS response of the
two layered film; (c) high frequency approximation.
breakdown is 2.5 mA cm–2. Figure 1 also reveals that j is not proportional to the scan rate as expected in
the case of a high field [7] or point defect [8] processes, but only decreases slightly at the lower sweep rate
(25 mV s–1). This assumes the involvement of side reactions and low faradaic efficiency of the process.
The variation of the electrical ac-response of the electrode has been followed with increasing the ap-
plied voltage applied (Fig. 2a). At 50 V two relaxations appear which become better-defined at higher
voltage. We can notice that at frequencies above 100 Hz the impedance increases with E. This suggests
that the growing anodic oxide layer is observed in this frequency range. Whereas below 100 Hz, the
impedance modulus poorly changes with the voltage up to 295 V. This behaviour can be explained by a
bi-layer model with dense inner layer and porous outer layer of lower resistivity. This may be due to
infiltration of the electrolytic solution at the grain boundaries. R2 is the resistance of the outer layer, and
R1 is a series resistance that contains the metal contribution, electrical contacts and electrolytic solution
resistances.
At frequencies above 5 kHz R2 contribution becomes negligible. If we assume that the dielectric be-
haviour of outer layer is the same as that of the inner one (they are made of the same material, zirconia),
we can simplify the equivalent circuit into that presented in Fig. 2c. The problem here is to extract the
film capacitance, Cf and to get rid of the dispersion phenomena. It is important to note that the approach
used in a previous paper to analyse layers obtained at high temperature and high pressure and demon-
strating a variation of ε with the frequency, failed to describe the present EIS data [2]. This is shown in
Fig. 2a where the slope of the high frequency part of the impedance modulus does not vary with the
frequency.
Therefore we have used an approach first proposed by Brug et al. [8] in the case of a capacitance af-
fected by a.c. frequency. In the present case, we have supposed that the dielectric layer is not the double
layer as in Ref. [8] but the oxide film. The impedance of CPE is written Z = [A3 ( jw)n ]- 1 , with
3
0.5 < n < 1. A is a constant and the CPE acts like a capacitor when n is equal to 1. Value of n varies
3
slightly with E, ranging between 0.94 and 0.98. The total film capacitance is obtained by:
1- n
1/n
C = [A R
(
1
]
)
.
(2)
f
3
© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Original
Paper
phys. stat. sol. (a) 202, No. 8 (2005) / www.pss-a.com
1505
Fig. 3 Cross-sectional SEM view of a 500 nm thick
film grown at 225 V (sweep rate 100 mV s–1)
Oxide
Metal
The film thickness is given by:
ee0S
d =
.
(3)
Cf
Where ε0 is the permittivity of vacuum (8.85 × 10–14 F cm–1), ε the dielectric constant of the oxide and
S the real surface area which includes the roughness factor. SEM views of the electrode surface show
that the surface roughness of the samples was low and we retain a value of 1 in the following.
However, in Eq. (3) ε is an unknown parameter. The values reported in the literature vary in a very
large range domain, from 21 [4, 10] to 38.3 [11] and are described to depend upon the growing electro-
lyte used since some anions can be inserted in thick zirconium oxide layers [3, 12, 13]. Anodization in
NaOH produces a non contaminated zirconium oxide film [13]. ε has been determined from the film
thickness determined by direct SEM observation of anodised rod cross-sections (Fig. 3). The samples
have been observed in the backscattering mode and the superficial oxide layer appears darker if com-
pared to the metal (Fig. 3). As shown in Fig. 4, the best fit between EIS and SEM data has been obtained
for ε = 19.5. We can note that high calculated dielectric constant seems sometime due to an overestima-
tion of the film thickness measured from the charge exchanged (e.g. [7]) whereas side reactions
can occur and the faradaic efficiency of oxide film growth is generally lower than 1 as shown below.
Figure 4 shows that there is no influence of vb on both ε and the growth rate. The film thickness varies
linearly with the applied voltage and the growth constant is 2.4 nm V–1. The thickness of the dense layer
measured at 50 V is thus equal to 110 nm (Fig. 4) and increases slightly at higher voltage when the porous
layer is present. Since the breakdown occurs at 300 V, the maximum total film thickness in NaOH is about
720 nm. The electric field in the film during the growth is very high and constant, measured at 4.2 × 106 V cm–1
if we neglect the potential drops at the interfaces. The Faradaic efficiency of the growth measured at
290 V and 100 mV s–1 is 23% with a theoretical thickness at 3.1 µm and an experimental one at 0.72 µm.
800
700
600
500
400
breakdown
300
Fig. 4 Variation of film thickness with applied voltage,
200
as deduced from EIS measurements: (+) vb
=
25 mV s–1; (×) vb = 100 mV s–1 and SEM observation:
(♦) vb = 100 mV s–1; (■) vb = 25 mV s–1.
100
0
0
50
100
150
200
250
300
350
E / V vs IrO2
© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1506
Porous
T. Pauporté et al.: Impedance spectroscopy study of anodic growth of zirconium oxide film
Fig. 5 Schematic view of the microporous layer (arbitrarily scale).
Oxide
film
Electrolyte
Dense
oxide
film
s
dm
By further analyzing the EIS spectra, we can get information on the porous layer characteristics. Fig-
ure 5 shows a schematic cross-sectional view of the porous layer. The porosity will be estimated consid-
ering that the film contains pores which are oriented perpendicular to the electrode surface and filled
with the electrolytic solution. The mean pore section area is s and the number of pores per square centi-
meter is n. dm is the thickness of the microporous layer. The volume of pore per square centimetre is ns
dm and the electrical resistance of the pores is given by: R2 = dm/σ ns, with σ the conductivity of the elec-
trolytic solution which fills the pores. The fraction of surface and volume occupy by the pores, η, is
given by:
dm
h = ns =
.
(4)
s R2
In the case of a film grown in 0.1 M NaOH at 250 V, we have R2 = 23 kΩ cm–2 (the plateau observed on
the modulus curve in Fig. 2a around 100 Hz corresponds to R2 + R1), σ = 0.02 S cm–1 and dm = 580 – 110
= 470 nm. We obtain a very low value, η ≈ 10–7, in agreement with the fact that the bilayered structure is
not observed on SEM views. Accordingly, the pores may be very thin and their density very low. Further
study of the film structure, by transmission electron microscopy for instance, would be necessary to
validate the present model.
In summary, we have studied by impedance spectroscopy and SEM the growth of anodic films of
zirconia in 0.1 M NaOH. In spite of several attempts to improve the sealing by testing some other sealing
modes, we have not been able to go beyond 300 V. By taking 19.5 for the dielectric constant of ZrO2 we
have found a very good correlation between the EIS thickness measurements (using the method de-
scribed by Brugg et al. to determine the film capacitance) and direct observation of film cross-section.
EIS also shows that the film is composed of two layers with different electric behaviour. These results
suggest that some very fine porosity is present in the outer layer and that the electrolytic solution could
be infiltrated into defects present in the film such as micropores. However, this hypothesis remains to be
validated by the direct film observation.
Acknowledgements The authors are grateful to Dr. J. Schefold and Dr. A. Ambard (EDF Research and develop-
ment, Département Matériaux et Mécanique des Composant, Morêt sur Loing, France) for fruitful discussions.
References
[1] P. Meisterjahn, H. W. Hoppe, and J. W. Schultze, J. Electroanal. Chem. 217, 15 (1987).
[2] J. Schefold, D. Lincot, A. Ambard, and O. Kerrec, J. Electrochem. Soc. 150, B 451 (2003).
[3] M. Cassir, F. Goubin, C. Bernay, P. Vernoux, and D. Lincot, Appl. Surf. Sci. 193, 120 (2002).
[4] F. Di Quarto, S. Piazza, and C. Sunseri, J. Electrochem. Soc. 130, 1014 (1983).
[5] N. Khalil, A. Bowen, and J. S. L. Leach, Electrochim. Acta 33, 1721 (1988).
© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Original
Paper
phys. stat. sol. (a) 202, No. 8 (2005) / www.pss-a.com
1507
[6] E. M. Patrito, R. M. Torresi, E. P. M. Leiva, and V. A. Macagno, J. Electrochem. Soc. 137, 524 (1990).
[7] L. Zhang, D. D. Macdonald, E. Sikora, and J. Sikora, J. Electrochem. Soc. 145, 898 (1998).
[8] G. J. Brug, A. L. G. Van den Eeden, M. Sluyters-Rehabach, and J. H. Sluyters, J. Electroanal. Chem. 176, 275
(1984).
[9] J. A. Bardwell and M. C. H. MacKubre, Electrochim. Acta 36, 647 (1991).
[10] J. L. Ord and D. J. Smet, J. Electrochem. Soc. 142, 879 (1995).
[11] J. S. L. Leach and B. R. Pearson, Electrochim. Acta 29, 1263 (1984).
[12] M. A. A. Rahimand and M. W. Khalil, J. Appl. Electrochem. 26, 1037 (1996).
© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim