The application of the barrier-type anodic oxidation method to thickness testing of

aluminum films
Jianwen Chen, Manwen Yao, Ruihua Xiao, Pengfei Yang, Baofu Hu, and Xi Yao

Citation: Review of Scientific Instruments 85, 094101 (2014); doi: 10.1063/1.4894525

View online: http://dx.doi.org/10.1063/1.4894525
View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/85/9?ver=pdfcov
Published by the AIP Publishing

Articles you may be interested in

Resistive switching in aluminum/anodized aluminum film structure without forming process
J. Appl. Phys. 106, 093706 (2009); 10.1063/1.3253722

In situ detection of porosity initiation during aluminum thin film anodizing

Appl. Phys. Lett. 94, 074103 (2009); 10.1063/1.3081014

Observation of isolated nanopores formed by patterned anodic oxidation of aluminum thin films
Appl. Phys. Lett. 88, 233112 (2006); 10.1063/1.2212535

Effects of starting material of aluminum doped zinc oxide underlayer on the electric properties of palladium doped
silver film
J. Vac. Sci. Technol. A 21, 1389 (2003); 10.1116/1.1560716

Bimodal spatial distribution of pores in anodically oxidized aluminum thin films

J. Appl. Phys. 88, 6875 (2000); 10.1063/1.1321780

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 11:46:56
 REVIEW OF SCIENTIFIC INSTRUMENTS 85, 094101 (2014)

The application of the barrier-type anodic oxidation method to thickness

testing of aluminum films
Jianwen Chen, Manwen Yao,a) Ruihua Xiao, Pengfei Yang, Baofu Hu, and Xi Yao
Functional Materials Research Laboratory, Tongji University, Shanghai 200092, China
(Received 20 May 2014; accepted 20 August 2014; published online 9 September 2014)
The thickness of the active metal oxide film formed from a barrier-type anodizing process is directly
proportional to its formation voltage. The thickness of the consumed portion of the metal film is
also corresponding to the formation voltage. This principle can be applied to the thickness test of the
metal films. If the metal film is growing on a dielectric substrate, when the metal film is exhausted
in an anodizing process, because of the high electrical resistance of the formed oxide film, a sudden
increase of the recorded voltage during the anodizing process would occur. Then, the thickness of
the metal film can be determined from this voltage. As an example, aluminum films are tested and
discussed in this work. This method is quite simple and is easy to perform with high precision.
© 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4894525]

I. INTRODUCTION controlled solely by the formation voltage in a barrier-type

Thickness is a basic characteristic of thin films. Thick-
ness test is getting more and more importance in material
science and technology.1 There are many different ways to
test the thickness of the film. Each of them has their own
advantage and disadvantage. The testing thickness range,
accuracy, and application are quite different. For example,
the ellipsometry2 or the white-light interferometer3 is very
good for the thickness measurement of transparent or semi-
transparent thin films, while they are not suitable for the
opaque metal films. The x-ray reflectivity technique4 and the
direct observation method by FE-SEM5 or TEM6 are two very
powerful thickness test methods for metal films. However, the
sophisticated instruments used in the two methods are ex-
pensive and not easy to perform. They are not inconvenient
in practical applications, especially for the thickness tests of
metal films in the nanometer scale.
In this work, a simple thickness test method is proposed
for the metal film in the nanometer scale. The principle of the
method is based on the proportionality relationship between
the thickness of the formed oxide film and the formation volt-
age in a barrier-type anodizing process.7 No special and ex-
pensive instruments are required in this method. It is also very
easy to perform. As an example, the thickness measurement
of aluminum films is tested by this method. Some important
technical issues are discussed in this paper.
II. PRINCIPLE OF ANODIZING METHOD
As a proven technology, the anodic oxidation has been
widely applied to industry and daily life.8–10 There are two
types of anodic aluminum oxide (AAO) films, barrier type and
porous type. The barrier-type AAO film is formed in neutral
electrolyte,11, 12 while the porous type AAO film is formed in
corrosive electrolyte.7, 13, 14 The thickness of an AAO film is
a) Author to whom correspondence should be addressed. Electronic mail:
yaomw@tongji.edu.cn
anodizing process.7
There are two different control ways in a barrier-type an-
odizing process.7 The changes in voltage and ionic current
density with respect to anodizing time would show the behav-
ior illustrated in Fig. 1. In a current control step (region 1),
the current is maintained at a constant value; thus the voltage
and the thickness of formed oxide film are increasing linearly
with time. In a voltage control step (region 2), the voltage is
reached and maintained at a constant value. Thus, the current
and the growth rate of the oxide film decrease rapidly with
time.
Because the current efficiency is relatively invariant in a
given anodizing condition, the molar amount of a formed ox-
ide film is proportional to the consumed metal film. For exam-
ple, the thickness ratio of an evaporated aluminum film to the
formed oxide film is 0.725, which is anodized in 3% ammo-
nium tartrate.10 Hence, the formation voltage also determines
the thickness of the metal film. If the formation voltage can be
found out when the metal film is exhausted in a barrier-type
anodizing process, the thickness of the metal film is deter-
mined and tested.
Indeed, if the metal film is growing on a dielectric sub-
strate, the recorded voltage can be easy to find out when the
metal film is exhausted. As shown in Fig. 2, the anodizing pro-
cess consists of three regions. At the first step (region AB in
Fig. 2), the recorded voltage increases linearly with increas-
ing time, which is the same as the current control step in a
barrier-type anodizing process. Because of the electrical insu-
lation of the formed film and the substrate, the specimen sys-
tem becomes an electrical insulation system when the metal
film that immersed in the electrolyte is exhausted. At this step
(region BC in Fig. 2), the voltage suddenly increases to the
setting upper limit voltage. Thus the recorded current is found
to decrease to about zero. At last (region CD in Fig. 2), the
voltage is maintained at the setting upper limit voltage, and
the recorded current is slowly decreased almost to zero.
Therefore, the thickness of the metal film is solely deter-
mined by the voltage VB . It can be calculated by the following

0034-6748/2014/85(9)/094101/5/$30.00 85, 094101-1 © 2014 AIP Publishing LLC

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 11:46:56
 094101-2 Chen et al.
FIG. 1. Formation of a barrier-type anodic oxide film.
formula:
dmetal = a(bVB ) = cVB ,
where a is the thickness ratio of metal film to the formed oxide
film; bis the anodizing ratio at the current control step in a
barrier-type anodizing process; c is the ratio of the metal film
to the formation voltage; and VB is the recorded voltage at the
point B shown in Fig. 2.
In addition, a natural oxide film will be formed on the
surface of metal films in air. This natural oxide film is repre-
sented as the initial measurement voltage VA in the anodizing
process; thus, the formula (1) should be rewired as:
dmetal = ab(VB − VA ) = c(VB − VA ).
III. THE SETUP
The experimental setup of the test method is a typical
anodizing system.15, 16 A general computer, a common source
meter, a container, and clamps are used in this test system. The
clamps are used to fix the sample and cathodes into the con-
tainer. The source meter should be with real time communi-
cation, control and data processing functions, which make the
test work automatically and avoid human disturbance. Also,
the upper limit voltage of the source meter should be large
enough to make sure that the voltage VB is existed. Source
meter (Keithley 2410, USA) is used in this study.
The good conductivity and chemical inert of the given
electrolyte should be considered as the selection of cathodes’
material. Platinum and graphite would be suitable materials
in some anodizing systems. Besides, the same metal material
of the test sample would be also a suitable cathodal material
FIG. 2. Principle of the thickness test method in a barrier-type anodizing
process.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 11:46:56
Rev. Sci. Instrum. 85, 094101 (2014)
since no additional ions are added to the anodizing system.
A pure Al plate is used as the cathodal material in this test
example. The detailed study of the electrochemical electrode
is referred to Ref. 17.
IV. EXPERIMENTAL
As a test example, the evaporated aluminum films were
tested by this method. The aluminum films were used in this
work. The aluminum films were deposited on polished quartz
substrate of 10 × 10 × 1 mm3 by vacuum evaporation equip-
ment (ZHD-400, Technol Science, Beijing, China). During
the deposition, the substrates were held in a vacuum cham-
ber with the pressure of lower than 2 × 10−4 Pa at 353 K. The
rotation of the substrates was 10 revolutions per minute. 3%
(1) ammonium tartrate aqueous solution is used as an electrolyte
in this study. The aluminum films are tested at 293 K.
On account of electrical insulation of the quartz substrate,
the tested voltage VB can be easy to find out in the anodizing
process. The thickness of the aluminum film can be calculated
by the formula (2), since the variables are related through
a × b = c and only 2 of them have to be determined.
The thickness ratio of the oxide film to the aluminum
layer replaced, which is anodized in 3% ammonium tartrate
aqueous solution, has been measured by Hass,10 that is
dAl
(2) a= = 0.725. (3)
dAl O
2 3
By measuring the thickness of 10 different AAO films
that formed in 3% ammonium tartrate aqueous solution by
the interferometer (Filmetrics F20, San Diego, CA), the ratio
of formation alumina film to formation voltage at the current
control step in the anodizing process is given in:
dAl O3
b= 2
= (1.25 ± 0.03) nm/V. (4)
Formation voltage
This ratio is consistent with the literature reports.7, 10
Then, the ratio of aluminum film to the formation voltage that
anodized in 3% ammonium tartrate aqueous solution is given
by:
c = ab = 0.725 × (1.25 ± 0.03) nm/V
≈ (0.91 ± 0.02) nm/V. (5)
So, the formula (2) in this application can be expressed
as:
dAl = [(0.91 ± 0.02) · (VB − VA )] nm/V. (6)
V. RESULTS AND DISCUSSION
Setting the upper limit voltage of 200 V and the current
density of 1.25 mA/cm2 , the test voltage and current has been
automatically recorded with the time by computer. As shown
in Fig. 3, the initial voltage VA and the tested voltage VB of
sample one are 4.0 V and 180.0 V, respectively. The thickness
of oxide layers formed in air for sample one is:
dalumina = 1.25 × 4.0 nm = 5.0 nm. (7)
 094101-3 Chen et al.
FIG. 3. Selecting the upper limit voltage at 200 V and the current density at
1.25 mA/cm2 , the initial voltage and the recorded voltage VB of Sample one
is 4.0 V and 180.0 V, respectively.
This thickness of oxide layers formed in air is consis-
tent with the value tested by Hass.10 So, the thickness of alu-
minum film of sample one can be calculated by the following
formula (2):
dAl = (0.91 ± 0.02) × (180.0 − 4.0) nm
≈ (159.5 ± 3.8) nm.
As shown in Fig. 4, this value is consistent with the value
determined by FE-SEM within the error range. The thickness
of the aluminum film of sample one is about 157 nm, where
the error rate is about 5%.
VI. TECHNICAL CONSIDERATIONS
Since the current efficiency of reaction occurring, the
molecular weight of substance, the density of the metal film,
and the density of formed oxide film are invariable in a given
anodizing system, and the thickness of a formed oxide film
is only dependent upon the amount of charge passed by Fara-
day’s laws.7 In other words, the current density value is less
sensitive for the thickness test method.
As shown in Fig. 5, selecting the current density of
1.25 mA/cm2 and 0.625 mA/cm2 , the thicknesses of the nat-
FIG. 4. FE-SEM of cross-section of sample one, thickness of the aluminum
film is about 157 nm.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 11:46:56
Rev. Sci. Instrum. 85, 094101 (2014)
FIG. 5. Selecting current density of 1.25 mA/cm2 and 0.625 mA/cm2 , the
initial voltages are 3.4 V and 3.4 V and the recorded voltages VB are 102.9
and 102.9 V, respectively.
ural oxide films are 4.3 nm and 3.8 nm, respectively. The
thicknesses of the aluminum films are 90.2 nm and 90.6 nm,
respectively. The thickness tested method with the different
current density is consistent with each other within the mar-
(8) gin of error. Thus it consumes a great deal of time in a low
current density and the anodizing system is unstable at a too
large current density. Therefore, the current density of about
1 mA/cm2 is a suitable value in this test method for aluminum
film in 3% ammonium tartrate aqueous solution.
Anderson and Devereux18 have discussed the importance
of anodizing temperature in the anodizing process. In their
study, the current is dependent upon the anodizing tempera-
ture at a given overpotential. The current efficiency decreases
with the increasing of anodizing temperature, and therefore
the anodizing ratio is solely dependent upon the anodizing
voltage. As discussed above, the formation voltage is inde-
pendent on the current density. The thickness test method is
less sensitive for the anodizing temperature. Thus, the anodiz-
ing system would be unstable at a high temperature.19 Gener-
ally, the selection of anodizing temperature should be consis-
tent with the temperature that is applied to testing the ratio c.
The room temperature is suitable for anodizing temperature
except in special conditions.
The electrolyte applied to this thickness test method
should be proven as a steady, reliable, and high current ef-
ficient electrolyte. Some useful electrolytes are shown in
Table I. The anodizing ratio of aluminum or aluminum alloy
with the different electrolyte is from 1.2 to 1.35. These differ-
ent anodizing ratios are mainly dependent upon the different
anodizing time at a given anodizing voltage.
Hass10 has extensively discussed the relationship be-
tween anodizing ratio and anodizing time. Because the cur-
rent efficiency is relatively invariant in a given anodizing con-
dition, thickness of anodic oxide film is increasing linearly
with the formation voltage and time at the current control
step in the anodizing process. In this study, the anodizing ra-
tio of pure aluminum is 1.25 nm/V at the current control step
in a 3% weight ammonium tartrate, which is tested by the
interferometer (Filmetrics F20, San Diego, CA). The anodiz-
 094101-4 Chen et al. Rev. Sci. Instrum. 85, 094101 (2014)

TABLE I. Anodizing ratio of aluminum or aluminum alloy in different electrolyte.

Aluminum or aluminum alloy Electrolyte (◦ C) Anodizing ratio (nm/V) and reference

Evaporated aluminum 3% weight ammonium tartrate aqueous solution 1.22 for 30 s,10 1.30 for 2 min,10 1.35 for 40 min10
Aluminium foil (trace impurities: 0.003% Fe; 0.1 mol/L ammonium pentaborate (20) 1.226, 20
0.004% Cu; 0.002 wt. % Si)
Al-1% Si-0.5% Cu thin film AGW electrolyte15 (3% aqueous solution of 1.30 for 30 min21
tartaric acid and propylene glycol at a volume
ratio of 1 to 4)
Aluminium foil (0.004 w/o Cu; 0.003 w/o Fe; Molybdate and Tungstate electrolyte (20) 1.222
0.002 w/o Si)
Pure aluminium sheet Aqueous borate solutions 1.3516

ing ratio of aluminum or aluminum alloy is about 1.25 nm/V
at the current control step in a barrier-type anodizing process.
This thickness test method is also suitable for those metal
films or semiconductors which can forma barrier-type anodic
oxide film in an appropriate electrolyte. What tabulated in
Table II are a number of anodizing ratios for several met-
als and semiconductors. The formation voltage is extremely
sensitive to the thickness variation of metal or semiconduc-
tor film, from the anodizing ratios with much more than
0.1 nm/V. If the surface of the metal film is smooth enough,
the test accuracy can be accurate to a single atomic layer level.
However, in the practical application, the accuracy of this
thickness test method is affected by the control of anodizing
conditions. The determination of the parameters c in a given
anodizing condition is critical in this thickness test method. In
other words, the accuracy of the measurement is affected by
the environmental uncertainties such as temperature, PH, and
so on. To improve the accuracy, a reproducible and reliable
preparation of the electrolyte under environmental uncertain-
ties is an essential requirement. The selection of electrolyte is
a critical issue in this method. The best selection of electrolyte
is the electrolyte, which is not sensitive to various conditions,
like the AGW electrolyte.5 The AGW electrolyte is a mixed
solution of glycol and water. Hasegawa and Hartnagel15 have
studied the AGW process well. In their study, the AGW pro-
cess has been found to be always very stable and reliable, and
the result is always re-producible even after many times of the
anodization. In addition, the application of standard sample
can effectively improve the testing accuracy, since the stan-
dard sample and the test sample are each anodizing in the
same anodizing condition.
Since this method is based on the theory of barrier-
type anodic oxidation, characteristics of this method are also
TABLE II. Anodizing ratio of some metals and semiconductors.
Metal Anodizing ratio (nm/V) and reference
Aluminum 1.30,10 1.3510
Tantalum 1.67
Niobium 2.27
Zirconium 2.0,7 2.1,7 2.4,7 2.77
Tungsten 1.87
Silicon 0.387
GaAs 1.3515
GaSb 2.523
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 11:46:56
inherited from the anodic oxidation. Because the thickness
of barrier-type oxide films is limited to <1 μm by dielec-
tric breakdown,24 the maximum film thickness of aluminum
tested by this method is less than 1 μm. In addition, since
thickness of the oxide layers of aluminum film formed in air
is about 4 nm,10 the minimum thickness of aluminum films
tested by this method should be more than 10 nm. There-
fore, the measuring range of the thickness test of aluminum
film is from 10 nm to 1 μm.This measuring range for the alu-
minum film is included by the direct observation method by
FE-SEM5 or TEM.6 However, no specific or expensive instru-
ments are necessary in this method and it is also very easy to
master. From this view, it is a beneficial complement to the
traditional thickness test method.
As a potential technology extension, this method can be
also applied in the thickness test of films growing on metal
or semiconductor substrates. The principle is similar to the
method discussed above. The basic issue of the technology
extension is to learn the recorded voltage VB when the metal
film is exhausted. The voltage change at point B should be
marked enough to easily find out the recorded voltage VB .
The following discussion is about this issue in detail.
The thickness of formed oxide film can be written as the
following equation by Faraday’s laws:7
QM
doxide = , (9)
nρF S
where Q = i · t is the passage of charge; M is the molecular
weight of the oxide film; ρ is the density of the oxide film; n is
the charge to form an oxide molecule; F is the Faraday; and S
is an area of the specimen surface covered by the electrolyte.
Since the thickness of a formed oxide film is directly pro-
portional to the formation voltage in a barrier-type anodizing
process, doxide = bV, therefore,
doxide M
V = = · Q = γ · t, (10)
b bnρF S
where j is the current density, α = nρbF
M
, and γ = nρbF
M
· Si
= α · Si = α · j .
To make the voltage change at the point VB marked
enough, it should meet the following condition:
|γ1 − γ2 | = |(α1 − α2 ) · j |  0. (11)
In addition, the metal film should be compact enough to
ensure no breakdown in conductivity occurs during testing.
 094101-5 Chen et al.
Also, the substrates should not be contacted directly to the
electrolyte, since the anodizing of the metal substrates during
testing would influence the recorded voltage. Lastly, the pa-
rameters c of the film and substrates should be tested in the
same electrolyte in a given anodizing condition.
VII. CONCLUSIONS
The barrier-type anodic oxidation method can be effec-
tively applied to the thickness test of the metal film. The con-
trollable DC power is the only necessary equipment in this
method. The electrolyte required in this method is simple and
easy to prepare. The reliability of this method is from the
strong physical relevant between the formed voltage and the
thickness in a barrier-type anodizing process. This method is
cheap and reliable, which can be applied to industrial or sci-
entific fields.
ACKNOWLEDGMENTS
This work is supported by the National Science Founda-
tion of China (51272177) and International Science and Tech-
nology Cooperation Program of China (2013DFR50470).
1 M. Vyal’yas and R. Laaneots, Meas. Tech. 33(4), 313–315 (1990).
2 A.
Rothen, Rev. Sci. Instrum. 16(2), 26–30 (1945).
3 P. Flournoy, R. McClure, and G. Wyntjes, Appl. Opt. 11(9), 1907–1915
(1972).
4 D. Windover, E. Barnet, J. Summers, C. Gribbin, T.-M. Lu, A. Kumar, H.
Bakhru, and S. L. Lee, AIP Conf. Proc. 550, 243–248 (2001).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 11:46:56
Rev. Sci. Instrum. 85, 094101 (2014)
5 R. L. Chiu, P. H. Chang, and C. H. Tung, J. Electrochem. Soc. 143(3),
984–990 (1996).
6 P. Skeldon, K. Shimizu, G. Thompson, and G. Wood, Surf. Interface Anal.
5(6), 252–263 (1983).
7 J. W. Diggle, T. C. Downie, and C. W. Goulding, Chem. Rev. 69(3), 365–
405 (1969).
8 C. Lu and Z. Chen, in Encyclopedia of Nanoscience and Nanotechnology
(American Scientific Publishers, 2011), Vol. 11, pp. 235–259.
9 A. Nishino, J. Power Sources 60(2), 137–147 (1996).
10 G. Hass, J. Opt. Soc. Am. 39(7), 532–539 (1949).
11 M. W. Yao, J. W. Chen, P. F. Yang, W. Shan, B. F. Hu, and X. Yao, Ferro-
electrics 455(1), 21–28 (2013).
12 G. Alcalá, S. Mato, P. Skeldon, G. E. Thompson, A. B. Mann, H. Habazaki,
and K. Shimizu, Surf. Coat. Technol. 173(2–3), 293–298 (2003).
13 L. Vojkuvka, L. F. Marsal, J. Ferré-Borrull, P. Formentin, and J. Pallarés,
Superlattices Microstruct. 44(4–5), 577–582 (2008).
14 M. Hunter and P. Fowle, J. Electrochem. Soc. 101(9), 481–485 (1954).
15 H. Hasegawa and H. Hartnagel, J. Electrochem. Soc. 123(5), 713–723
(1976).
16 A. Harkness and L. Young, Can. J. Chem. 44(20), 2409–2413 (1966).
17 G. Inzelt, A. Lewenstam, and F. Scholz, Handbook of Reference Electrodes
(Springer, 2013).
18 P. G. Anderson and O. F. Devereux, J. Electrochem. Soc. 122(2), 267–273
(1975).
19 R.-L. Chiu, P.-H. Chang, and C.-H. Tung, Thin Solid Films 260(1), 47–53
(1995).
20 P. Skeldon, K. Shimizu, G. Thompson, and G. Wood, Surf. Interface Anal.
5(6), 247–251 (1983).
21 R. L. Chiu, P. H. Chang, and C. H. Tung, J. Electrochem. Soc. 142(2),
525–531 (1995).
22 G. Thompson, P. Skeldon, K. Shimizu, and G. Wood, Philos. Trans. R. Soc.
London, Ser. A 350(1692), 143–168 (1995).
23 O. Sulima, A. Bett, and J. Wagner, J. Electrochem. Soc. 147(5), 1910–1914
(2000).
24 F. Monfort, A. Berkani, E. Matykina, P. Skeldon, G. Thompson, H.
Habazaki, and K. Shimizu, J. Electrochem. Soc. 152(6), C382–C387
(2005).