Passive film on titanium

Electrochimica
Pergamon
ho.
Vol. 41, Nos. 7/S, pp. 1143-l 153. 1996
Copyright c 19% Ekvier Science Ltd.
Gnat Britain. All righa reserved
Printedin
0013~4686/96
$15.00
+ 0.00
oo13-46gqa5)oo465-3
ELECTROCHEMICAL
IMPEDANCE SPECTROSCOPY
STUDY OF THE PASSIVE OXIDE FILM ON TITANIUM
FOR IMPLANT APPLICATION
J. PAN,*? D. THIERRY$ and C. LEYGRAFt
t Department of Materials Science and Engineering, Royal Institute of Technology, Dr Kristinas vlg 51,
S-100 44 Stockholm, Sweden
$ Swedish Corrosion Institute, Roslagsvagen 101, Hus 25, S-104 05 Stockholm, Sweden
(Received 8 Mny 1995)
Abstract-The surface oxide film on titanium and its long-term stability in biological environments play
a decisive role for the biocompatibility of titanium implants. In this study, the passive oxide film formed
on titanium and its natural growth in a phosphate buffered solution with and without an H,O, addition
have been investigated by electrochemical impedance spectroscopy (EIS) measurements over a period of
several weeks. In the absence of H,O,, the impedance response indicated a stable thin oxide film on
titanium. However, the introduction of H,O, into the solution resulted in significant changes in the
EIS-spectra, which varied with exposure time. The interpretation of results is based upon a two-layer
model of the oxide film, consisting of a thin barrier-type inner layer and a porous outer layer. The H,O,
addition in the solution led to a significant decrease in corrosion resistance of titanium and also to a
thickening of the porous outer layer. The observations may provide an explanation of the unexpected in
uivo titanium oxide growth and ion incorporation into titanium implant oxide surfaces.
Key words: impedance spectroscopy, titanium, oxide film, H,O,,
INTRODUCTION
Titanium is one of the most important materials for
biomedical and dental implant applications. This is
partly due to the excellent corrosion resistance of
titanium and its alloys in many aqueous environments provided by a most protective passive film,
which spontaneously forms on titanium. The passive
film is normally a few nm (1 nm = lOa9 m) thick and
consists mainly of amorphous titanium dioxide[ 11.
The physiochemical and electrochemical properties
of the oxide film and its long-term stability in biological environments play a decisive role for the biocompatibility
of titanium implantsC2, 33. Many
investigations have shown that this type of natural
surface film is essentially a thin layer of TiO,[4, S],
which seems very stable in vitro[6, 73, and it is commonly believed that titanium surgical alloys have a
high corrosion resistance[8]. However, a marked difference has been observed between in vitro and in
vioo behaviors of titanium. It has been reported that
the surface film formed on titanium implants in a
human body after some years can reach a thickness
far beyond the nm-range. Under these circumstances
the thickness and composition of the oxide layer
changed with implantation time, and some incorporation of mineral ions occurred[9, lo]. Despite a high
corrosion resistance of titanium in vitro, there is also
increasing evidence showing that titanium is released
into and accumulated in tissue adjacent to titanium
* Author to whom correspondence should be addressed.
biomaterials.
implants[l l-133. When examined with proton
induced X-ray emission (PIXE), extensive titanium
release was observed[14]. With prolonged implantation time, the titanium release was observed to reach
a steady state rate[15]. To explain the observed high
in vivo oxidation/corrosion
rates, it has been suggested that H,O, generated in biological systems
plays an important
role. An increase in the
oxidation/corrosion
rate of titanium
has been
observed with increased H,O, concentration in a
phosphate-buffered saline (PBS) solution[16, 171.
Moreover, different models have been proposed to
describe titanium oxide films and to account for the
titanium release in viuo[18-211. However, little is
known about the in situ change of the titanium oxide
film due to H,O,.
During the last years, various oxide films on metal
surfaces have been characterized by electrochemical
impedance spectroscopy (EIS), eg anodic oxides on
aluminum[22, 231 and zirconium[24], as well as
passive films on titanium[25-271. The spectra are
frequently interpreted in terms of an “equivalent
circuit” based on a plausible physical model with the
circuit elements representing electrochemical properties of the metal and its oxide film. To investigate the
influence of H,O, on the oxidation/corrosion behaviour of titanium, EIS has been used in the present
study. The passive oxide film formed on titanium
exposed in a PBS solution has been characterized in
situ by the EIS measurement, and the progress of
film formation due to an H,O, addition has been
monitored by EIS during a period of several weeks.
The main emphasis of this paper is on the interpreta-
1143
J.
1144
PAN et al.
tion of the EIS results, which utilizes an equivalent
circuit approach. In related papers, the main results
from EIS were found to be in general agreement with
the observations from surface analytical techniques,
namely X-ray photoelectron spectroscopy (XPS) and
scanning tunneling microscopy (STM)[28-301.
EXPERIMENTAL
General aspects and details of experimental procedures have been described previouslyC28, 291. To
investigate the effect of H,Oz, two sets of experiments were performed in parallel. In one set, titanium was exposed to phosphate-buffered
saline
(PBS) solution without H,O, addition, while in the
other set 1OOmM H,O, was introduced into the
PBS solution at the beginning of the exposure. EIS
was recorded once a day during the exposures,
which lasted for a period of several weeks. The
experiments with H,O, addition were terminated
after 1 or 2 weeks when the titanium surface
appeared blue due to an interference effect. This indithat
the
oxide film had
thickened
cates
considerably[29]. Three parallel runs of each set of
experiments were conducted.
The EIS measurements were made using an
Impedance/Gain-Phase
Analyzer (Solarton 1260)
coupled to a Potentiostat-Galvanostat
System
(EG&G Part, Model 273A), which was connected to
a three-electrode electrochemical cell. Spectra were
obtained at open-circuit potential of the titanium
sample in PBS solution, with an amplitude of 10mV.
The frequency span was normally from 1 kHz down
to 5 mHz, and sometimes extended to 1 mHz. The
impedance data at frequencies above 1 kHz seemed
to be affected by manually setting the current range
of the potentiostat at 1 mA (corresponding to a
current measuring resistor of 1 KQ), and therefore no
attempt was made to include data at higher frequencies. Data registration and analysis were performed with a computer.
The spectra were
interpreted using the nonlinear least square fitting
procedure developed by Boukamp[31]. Because of
the distributed relaxation feature which is commonly
observed in titanium oxide films, a constant phase
element (cpe) was utilized for data fitting instead of
an ideal capacitor. For simplicity reasons, the value
obtained from data fitting was taken as the capacitance in the forthcoming discussion, The quality of
fitting to the equivalent circuit was judged firstly by
the chi-square value, and secondly by the error dis-
1"
10’
i
n,mmmmmmmmmmmmmmmm
mm
;ms
+++
v
'rn++
n
++r
m
‘f
++
n
++
++
l
++
++
(A)
m
++
f+
n
++
n
++++
0
10'
10’
0.V
0
lOi
10*
10-l
10-l
10’
10’
13
ld
Frequency, Hz
Fig. 1. Bode plots for titanium exposed in the PBS solution without H,O, addition. (A): 1 day; (B): 30
days of exposure.
Passive oxide film on titanium for implant application
tribution vs the frequency comparing
with simulated data[31].
by a phase angle close to -90”
over a wide frequency range. Furthermore, this does not change
with exposure time (see Fig. 1) indicating that the
oxide film is rather stable in the absence of H,O, .
However, when titanium is exposed to PBS with
an addition of H,O, , the spectrum appears very different and varies significantly with exposure time. A
set of spectra at different exposure times is shown in
Fig. 2. The evolution of the spectra may be divided
into an earlier and a later stage, distinguished by the
appearance of a blue colour on the sample surface.
experimental
RESULTS AND DISCUSSION
EIS
1145
spectra
When titanium is exposed to PBS solution
without any addition of H,O,,
its EIS spectra
exhibit behaviour typical of a thin passive oxide film
on titanium, ie, a near-capacitive response illustrated
“““mm
mm
n
I.-
m
n
n
n
n
++++
n
m
-++++p
’
n
nn
m
++++
++
nn
n
m
++
I
+++
m
++ L
r+
++
++++t
L
(A)
10’ r
lo6
nBmmDWmrn
mm
r
n
n
mm
++
3.f
‘m
m-60
n
++++
+++
10'
i!
or
n
n
101
m
10'
a
n
9
n
P
n
+++
m
E
-30
I
102
(JN
t
90
60
30
0
10=
10-1
lo.’
10°
10’
102
Id
Frequency,Hz
Fig. 2. Bode plots for titanium exposed in the PBS solution with H,O, addition having an initial concentration of 1OOmM.(A): 1 day; (B): 15 days; (C): 30 days of exposure.
J. PAN et al.
Equivalent Circuit I
Ti
TiO,
h
ii-J_
ChO
cb
RP
Rho
Ti
TiOf
ates or
precipitates
%W Go)
Equivalent Circuit II
Fig. 3. Equivalent circuits used for the two-layer oxide film on titanium. (I): porous layer unsealed; (II):
porous layer sealed, and schematic representation of the oxide film on titanium under different exposure
conditions in the PBS solution. (A): without H,O, ; (B): with H,O,, earlier stage; (C): with H,O,, later
stage, when the titanium surface appears blue. Notations: R, is the solution resistance; C,, R, are the
inner layer capacitance and resistance; C,, R, the outer layer capacitance and resistance (or the electrolyte resistance inside pores); C,,, R,, represent the capacitance and resistance of hydrates/precipitates
inside pores, respectively.
During the earlier stage, the spectrum varies slowly
with exposure
time, as shown in Fig. 2(A)-(B).
However, the spectrum is clearly different when the
blue colour is seen on the surface during the later
stage of exposure (see Fig. 2(C)). It is interesting to
note that the remarkable change in the spectrum
coincided with the appearance of the blue colour,
although the length of time it took for the colour to
appear varied significantly (eg, from 1 to 4 weeks
between repeated experiments). These observations
clearly show that the introduction of H20, into the
solution modifies the passive oxide film on titanium.
As a result, the electrochemical properties of the film
also change with exposure time.
Selection of the equivalent circuit
Figure 3 shows the two equivalent circuits, both
based on a two-layer model of an oxide film, which
can be satisfactorily used for fitting the spectra
obtained at different exposure conditions. The
spectra for the exposure without H,O,, and during
the earlier stage of exposure with H,O, , can be well
fitted to circuit I. During later stages of the exposure,
when the sample surface appears blue, the spectra
can no longer be fitted to circuit I with a reasonable
fitting quality. Instead, they can be well fitted to
circuit II. Figure 4 provides a comparison of fitting
quality for the two circuits when applied to a spectrum obtained after 30 days of exposure. From the
simulation, the chi-square value and error distribution vs frequency, it seems clear that circuit II can
better describe the titanium oxide film when it
appears blue. Previously, circuits I and II have been
proposed
by other authors
to represent an
unsealed[22], and sealed1321 anodic oxide films on
aluminum, respectively. For the unsealed anodic
oxide film, it is generally agreed that the pores in the
outer porous layer are filled with electrolyte, while
Passive oxide film on titanium for implant application
10‘
Fitto drcult
(A)
r
I
1”
d
3
1147
60
10’
!
a
g
2
f
ld
30
%
31
d
102
0
mclsurement
l
simulation
10'
0
1"
10‘
r
l
10’
10J
Fittocircuit11
@I
simulation
'
10-l
10'
Frequency,
I
do
D-W
Fig. 4. Comparison of fitting quality of the two equivalent circuits for a spectrum obtained after 30 days
of exposure, when the titanium surface appears blue. (A)-(B): measured and simulated spectra; (C)-(D),
next page: chi square value and error distribution vs frequency comparing the measured and simulated
data for circuit I and II, respectively.
for the sealed oxide, the pores are filled with
hydrated compounds. In the latter case, the hydrates
inside pores need to be taken into account in the
equivalent circuit for fitting the spectra and the inner
layer may be approximated by a capacitor[32]. The
present observations of the evolution of the oxide
film on titanium due to an H,O, addition may be
compared to a sealing process of the anodic oxide
film. The presence of some hydrates/precipitates
inside the oxide film has been verified by XPS
showing that considerable amount of ions had been
incorporated
into the outer part of the oxide
film[29]. The oxide model corresponding to different
exposure conditions is schematically represented in
Fig. 3, together with the equivalent circuits I and II
found suitable for representing the impedance characteristics.
Although the passive film which forms on titanium
in aqueous solutions is frequently described as a
single Ti02 layer, there is substantial evidence that
this film in many exposure conditions exhibits a twolayer structure, ie, a dense inner layer and a porous
outer layer[33-351.
By using surface sensitive
angular-resolved XPS, we also observed the two-
layer feature of the oxide film formed on titanium
exposed to PBS solution, and the porous layer had
thickened due to the introduction of H,0z[28, 291.
The titanium oxide was found to be essentially
TiOz , with a transient region between the inner and
outer layers[28]. In addition, upon termination of
the exposure with H,O, , ions from the PBS solution
were found to be incorporated into the oxide
frlm[29]. It seems likely that the outer layer basically
consists of the same oxide as the inner layer, but possesses microscopic pores which may be filled by
either the solution or some hydrated/precipitated
compounds depending on the exposure conditions.
In summary, the two-layer model of the oxide film
used for the impedance data fitting is supported by
the results from independent techniques.
Fitting results and interpretation
No H,02 addition. The EIS spectra can be well
fitted to circuit I. As an example, the fit results of R,
and R, (ie, the resistance of the barrier layer and
porous layer/the electrolyte inside the pores,
respectively), Cb and C, (ie, the capacitance of the
J. PAN et al.
1148
o
03
Fit to drcuit I,
&I squam= 2 I 10’
AA
old
"brginw
A
(D)
-10
I
1Od
m
Fit to drcuit II, chi squarer; 6 x 10d
-l
’
10*’
101
Fwueney.
-10
lol
WI
Fig. 4 (Continued)
barrier and the porous layer, respectively) have been
plotted vs exposure time in Fig. 5. It can be seen in
Fig. 5(A) that R, is very high, around 5 MRcm’. In
addition, C, is relatively low and decreases slightly
with exposure time, reaching a steady state value
about lO~F/cn?. The slight decrease of C, may correspond to a slow growth of the titanium oxide film,
indicating a long-term stability of the thin passive
film in PBS solution without H,Oz. On the other
hand, as shown in Fig. 5(B), R, is low and increases
slightly with exposure time (roughly from 100 to
200Rcm2). This indicates that the pores are probably filled only with the solution. C, is around
25pF/cm2. Because of the open porous structure, it
is difftcult to calculate the thickness of the outer
layer from the C,r value; however, the XPS data and
independent optical measurements indicate the outer
layer could not be thick, owing to a small total
thickness[29]. In this case without H202, the corrosion of titanium is prohibited mainly by the nonporous inner layer. The very high resistance R,
implies a high corrosion resistance, ie, a low rate of
titanium release and oxide growth.
Since the outer layer seems to be thin and the
pores are filled with electrolyte, the contribution
from this porous layer to the electrochemical proper-
ties is rather small, and the impedance response is
dominated by the inner layer. As shown in Fig. 6,
when a single layer model with one time constant is
assumed, the fitting quality is just slightly decreased,
compared with the fitting to the above two-layer
model. This also indicates that the parameters of the
outer layer may not be accurately determined from
the spectra.
With H,O, addition. In the earlier stage of exposure, the spectra can also be well fitted to the circuit
I. During the first few days after introduction of
H,O, into solution, the value of R, is low and C, is
high. These data are not plotted, because during this
initial period the concentration of H,O, is high and
decreases quickly, as measured by a VACUette
ampoule test[29] and illustrated in Fig. 7, which
may lead to some uncertainties in the interpretation.
Nevertheless, the low frequency limit of the impedance modulus is around 100KRcm2 during this
period, which is a significant decrease as compared
to before the introduction of H,O,. This indicates
that the native dense oxide layer may be partially
dissolved and becomes more defective, due to the
H,O, addition. The decreased value implies a relatively low corrosion resistance. Thus, the dissolution/
oxidation rate of titanium can be expected to
Passive oxide film on titanium for implant application
1149
Inner layer
--A-Rb
ID,
,,’ - -.
0
‘.
=8
*-0..
“s
,
40
I
20
J
20
I
__-- _- a .
‘VB
Outerlayer
(B)
“s
..___----
--.-cp
--a-Rp
10
0
0
5
10
20
15
Time ot exposure,
25
20
days
Fig. 5. Fit parameters (capacitance and resistance) for titanium exposed in PBS without H,O,
function of exposure time. (A): inner layer; (B): outer layer.
increase upon introduction of H,O, , and as a result,
a considerable amount of titanium may be released
into the solution.
After this initial period, the H,O, concentration
becomes stable at a low level (Fig. 7). The fitting
results from the spectra have been plotted vs exposure time and an example is given in Fig. 8. It can be
seen from Fig. 8(A) that C, seems to reach a steady
Vahe,
while &, gradually increases with exposure
time and approaches a high value of several
MQcm’, which is about the same level as without
H20,. This increase of R, may be attributed to a
regrowth of the inner layer or to a self-rehealing
process. These processes are known to occur in
passive films on titaniurr and imply that the film
becomes protective again. From Fig. 8(B) it is also
evident that C, remains stable over a certain introductory period and shows a decreasing trend during
later exposure times. The relatively high C, value
may indicate the enhanced porous feature of the
outer layer. Meanwhile, R, is in the order of several
tens of Kncm2 and increases with exposure time.
The R, value seems to be quite high for an electrolytic resistance inside pores and probably incorporates a contribution
from some hydrates and/or
precipitates generated inside the pores.
as a
When the sample surface appears blue, circuit I is
no longer suitable for the fitting, as demonstrated in
Fig. 4. Instead the spectra in the later stage readily
fit circuit II, in which C, is clearly resolved, and
values of C,,, and R,, are obtained. An example of
the fit parameters is plotted in Fig. 9. During this
stage, C, seems to be quite stable (see Fig. 9(A)). Its
low value of about 3pF/cmZ indicates a thick outer
layer. Moreover, R, is about 50KRcm2, and ChO
decreases with exposure time (see Fig. 9(B)). These
two circuit elements probably correspond to a
gradual hydration or to some precipitation process.
Based on the increased low frequency limit of the
impedance modulus, the corrosion resistance of the
film could be expected to reach a high level again
during prolonged exposure. When the sample was
taken out from the solution, dried in air for 1 hour
and exposed again to the same solution, R, was
found to increase remarkably and reach a high level
of around 400 KRcm2 while other parameters
remained almost unchanged, as illustrated in Fig.
9(A) and 9(B). This observation confirms that R, is
indeed somehow related to the hydrated and precipitated compounds, which can easily lose water and
become dried in air. The interpretation is in accordance with a lower amount of water or OH groups
J.
1150
PAN et al.
10
10
(A)
Two-layermodel, chl square = 8 I
10’
A imaginary
ORal
5
*
i
0
f
F
‘6b
-5
1
-10
-10
10
10
09
Single layer model, chi square - 4 I: IO”
A imagh~
oreal
A
AA
AA
5
dr
.
E
0
%
P
-5
-la
10’
lo-’
ld
1
-10
Frequency, Hz
Fig. 6. Frequency error distribution for fitting of the spectrum of titanium exposed in PBS without H,O,
addition. (A): fit to the two-layer model (circuit I); (B) fit to a single layer model with one time constant.
found on the surface by two independent surface
analytical methods, XPS and IR spectroscopyC361.
From the fitting results it may be noticed a considerable difference in the capacitance value of the
outer layer between the earlier and the later stages of
exposure. To some extent the difference may be due
to the different equivalent circuits used. In practice,
different equivalent circuits are commonly employed
for representing unsealed and sealed porous layers of
anodic oxide films with their different microstructures. To reveal the capacitive behaviour of the
porous layer on anodized aluminum, it has been
demonstrated by other authors that the pores need
to be sealed[32]. The results from the present study
also indicate that the outer porous layer of the
passive oxide film on titanium can be better characterized by EIS when the pores are filled by some
hydrated/precipitated compounds.
Efect of H,O, on the passive oxidefilm on titanium
AS shown above, the interpretation of the EIS
data provides parameters of the passive oxide film
on titanium, such as resistance and capacitance
values. From the variation of these parameters with
exposure time, it seems that the passive oxide film on
titanium can be attacked by H,O, in the solution.
As a result, the film becomes more defective, probably porous and exhibits a relatively low corrosion
resistance. However, as a result of prolonged exposure and probably also a decreased H20, concentration, the oxide can regrow to a larger thickness. The
porous oxide may become hydrated, and ions may
easily be incorporated into the pores and further precipitate. When the pores are filled by hydrated or
precipitated compounds, the thickened oxide film
can become highly protective again, ie, the oxide film
is self-rehealed. The evolution of the spectra reflects
the pronounced two-layer feature of the oxide tilm
and the development of a thick outer porous layer
which may facilitate the incorporation of ions from
the PBS solution. It should be stressed that the time
for the blue colour to appear on the sample surface
may vary significantly, and the same is true for the
evolution of the EIS spectra, Therefore, these observations should be regarded as qualitative rather than
giving an exact time-dependent behaviour of the
passive oxide film on titanium under present exposure conditions.
The study may provide an explanation for the
unexpectedly high rate of in vivo titanium release and
Passive oxide film on titanium for implant application
1151
120
Ti exposed to the PBS solution
100
Inittat H,O, concentration: 100 mM
80
60
40
20
0
0
10
5
15
20
Time of exposure, days
Fig. 7. Concentration of H,O, in PBS vs the time of exposure.
facilitate the incorporation of mineral ions such as
phosphate groups from biological fluids. When the
oxide film becomes thickened and the pores are
sealed, the corrosion resistance can reach a high level
again and prevent further dissolution. This may also
oxide thickening on titanium implants. The results
and interpretation suggest that a H,O, addition can
lead to an increased corrosion rate (titanium release)
and also an enhanced film thickening. The pronounced porous outer layer may be expected to
00
Inner layer
(4
--&-Rb
60
40
a0
.-A
_-
,
,’
aa*,
x
PaA
a’
0
40
Outerlayer
(W
--•-cp
--6.Rp
5s
*
,oo*
0
so
“*oo*
-_a.,
e
I
-*
,A,>:‘.-
-.
4s
40
Fig. 8. Fit parameters (capacitance and resistance) for titanium exposed in PBS with H,O, as a function
of exposure time during the earlier stage of exposure. (A): inner layer;(B): outer layer.
J. PAN et
1152
loo
al.
CnpaeItnnce of the outer and Inner layers
69
I
*.-_*@_*_______*__+
____--
_a*@
0
30
35
45
40
Time of exposmt,
30
I
50
55
days
400
Hydratesor pnclpRata tnstde potw
00
--&-Rho
”
_A-----__
400
‘*A
A
I
200
0
,,
Q-A-
30
.A.‘,
d’ :
35
a
%
40
45
50
55
0
Time of exposure, days
Fig. 9. Fit parameters for titanium exposed in PBS with H,O, as a function of exposure time during the
later stage. (A): capacitance for the inner and outer layers; (B): capacitance and resistance for the
hydrates or precipitates inside pores.
the fact that the amount
biological surroundings can reach
a long period of implantation[lS].
discussed in more detail in another
explain
of titanium in the
a stable level after
These aspects are
publication[29].
CONCLUSIONS
From this electrochemical impedance spectroscopy (EIS) study the following conclusions may be
drawn :
1. EIS is a powerful method for in situ characterization of the passive oxide film on titanium. The
evolution of the film due to an introduction of H,02
can be monitored by EIS measurements as a function of exposure time.
2. In a phosphate-buffered saline solution without
H202, the oxide film on titanium exhibits a high
corrosion resistance and a long-term stability. The
introduction of H202 into the solution results in a
decreased corrosion resistance and an enhanced
dissolution/oxidation rate.
3. The EIS spectra can be interpreted in terms of
a two-layer model of the oxide film, which consists of
a thin barrier-type inner layer and a porous outer
layer. The parameters obtained and their variation
with exposure time indicate that the oxide film can
be attacked by H,O,, which results in more defective inner layer and in a thicker and more porous
outer layer. After prolonged exposure, the oxide film
may reheal itself and exhibit a significantly higher
corrosion resistance again due to some sealing processes inside pores in the oxide film.
Acknowledgements-We would like to thank MS Ingegerd
Annergren (Dept. of Materials Science and Engineering,
Royal Institute of Technology, Sweden) for most valuable
comments on the interpretation of EIS results and MS
Sandra Brunsberg (Royal Institute of Technology, Sweden)
for checking the English.
Passive oxide film on titanium for implant application
REFERENCES
1153
17. P. Tengvall, I. Lundstriim, L. Sjiiqvist, H. Elwing and
L. M. Bjursten, Biomuterials IO, 166 (1989).
1. R. W. Schutz and D. E. Thomas, Metals Handbook,
Vol. 13: Corrosion, 9th edn, p. 669. ASM International
(1987).
2. D. F. Williams, Biocompatibility
of Clinic Implant
Materials (Edited by D. F. Williams), Chapter 2, CRC
Press, Boca Raton (198 1).
3. B. Kasemo and J. Lausmaa, Surface Characterization of
Eiomalerials (Edited by B. D. Ratner), p. 1. Elsevier,
Amsterdam (1988).
4. J. Lausmaa, L. Mattsson, U. Rolander and B. Kasemo,
Biomedical Materials, Mat. Res. Sot. Symp. Proc. Vol.
55 (Edited by J. M. Williams, M. F. Nichols and W.
5.
6.
7.
8.
9.
10.
Il.
Zingg), p. 351. Materials Research Society, Pittsburgh,
Pennsylvania (1986).
J. Lausmaa, B. Kasemo and H. Mattsson, Applied Surf
Sci. 44, 133 (1990).
T. Hanawa and M. Ota, Biomaterials 12,767(1991).
T. Hanawa and M. Ota, Applied SurjI Sci. 55, 269
(1992).
R. J. Solar. Corrosion and Dearadation of Imolant
Materials, STP 684 (Edited by B. C. Syreit anh A.
Acharya), p. 259. American Society for Testing and
Materials, Philadelphia (1979).
D. McQueen, J-E. Sundgren, B. Ivarsson, I. Lundstriim,
B. af Ekenstam, A. Svensson, P-I. Brinemark and T.
Albrektsson, Clinical Applications
of Biomaterials
(Edited by A. J. C. Lee, T. Albrektsson and P-I.
Branemark), p. 179. Wiley, New York (1982).
J-E. Sundgren, P. Bodii and I. Lundstriim, J. Colloid.
Interface Sci. 110,9 (1986).
G. Meachim and D. F. Williams, J. Biomed, Mater.
Res. 7,6, 555 (1973).
12. R. J. Solar, S. R. Pollack and E. Korostoff, J. Biomed.
Mater. Res. 13, 217 (1979).
13. K. Merrit and S. A. Brown, Compatibility of Biomedical
Implants, Proc. 185th. Electrochemical Society Meeting,
Vol. 94-15 (Edited by P. Kovacs and N. S.
Istephanous), p. 14. The Electrochemical Society, Inc.
USA (1994).
14. A. M. Ektessabi, T. Otsuka, Y. Tsuboi, K. Yokoyama,
T. Albrektsson, L. Sennerby and C. Johansson, International .I. o!PIXE, 4,213, 81 (1994).
15. J. L. Woodman, J. J. Jacobs, J. 0. Galante and R. M.
Urban, J. Orthop. Res. 1,421 (1986).
16. P. Tengvall, H. Elwing, L. Sjiiqvist, I. Lundstrijm and
L. M. Bjursten, Biomaterials 10, 118 (1989).
18. R. J. Solar, S. R. Pollark and E. Korostoff, Corrosion
and Degradation ofImplant Materials, STP 684 (Edited
by B. C. Syrett and A. Acharya), p. 161. American
Society for Testing and Materials, Philadelphia (1979).
19. T. Hanawa, The Bone-Biomaterial Interface, Chapter 4
(Edited by J. E. Davies), p. 49. University of Toronto
Press (1991).
20. K. E. Healy and P. Ducheyne, J. Biomed. Mater. Res.
26,319 (1992).
21. K. E. Healy and P. Ducheyne, Biomaterials 13, 8, 553
(1992).
22. F. Mansfeld, Analysis and Interpretation of EIS Data
for Metals and Alloys, Chapter 4, Technical Report 26,
Solartron-Schlumberger (1993).
23. J. L. Dawson, G. E. Thompson and M. B. H.
Ahmadun, Electrochemical
Impedance:
Analysis and
Interpretation, STP 1188, (Edited by J. R. Scully, D. C.
Silverman and M. W. Kendig), p. 255. American
Society for Testing and Materials, Philadelphia (1993).
24. J. A. Bardwell and M. C. H. McKubre. Efectrochim.
Acta 36, 314,647 (1991).
25. T. P. Cheng, J. T. Lee and W. T. Tsai, Electrochim.
Acta 36, 14,2069 (1991).
26. W. A. Badawy, S. S. Elegamy and Kh. M. Ismail, Br.
Corros. J. 28, 2, 133 (1993).
27. D. G. Kolman and J. R. Scully, J. Electrochem. Sot.
141, 10.2633 (1994).
28. J. Pan, D. Thierry and C. Leygraf, J. Biomed. Mater.
Rex 28, 113 (1994).
29. J. Pan, D. Thierry and C. Leygraf, J. Biomed. Mater.
Res. in press (1996).
30. A. A. Ejov, S. V. Savinov, I. V. Yaminsky, J. Pan, C.
Leygraf and D. Thierry, J. Vat. Sci. Technol. B, 12, 3,
1547 (1994).
31. B. Boukamp, Proc. 9th. Euro. Congr. Corros., FU-252,
Utrecht. The Netherlands (1989).
32. A. Baltat-Bazia, N. Celati: M. ‘Keddam, H. Takenouti
and R. Wiart, Mater. Sci. Forum 1 I l-1 12, 359 (1992).
33. N. D. Tomashov, G. P. Chernova, Yu. S. Ruscol and G.
A. Ayuyan, Electrochim. Acta 19, 159 (1974).
34. M. M. Hefny, A. A. Mazhar and M. S. El Basiouny, Br.
Corros. J. 17, 1, 38 (1982).
35. M. S. El Basiouny and A. A. Mazhar, Corrosion-NACE
38, 5, 237 (1982).
36. J. Pan, unpublished data.