Colloids and Surfaces B: Biointerfaces 84 (2011) 400405

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces

journal homepage: www.elsevier.com/locate/colsurfb

In vitro corrosion study of different TiO2 nanotube layers on titanium in solution

with serum proteins
Wei-qiang Yu, Jing Qiu, Fu-qiang Zhang
Department of Prosthodontics, School of Stomatology, Shanghai Ninth Peoples Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai Research Institute of Stomatology,
639 Zhizaoju Road, Huangpu District, Shanghai 200011, China

a r t i c l e

i n f o

Article history:
Received 10 October 2010
Received in revised form 18 January 2011
Accepted 22 January 2011
Available online 5 March 2011
Keywords:
Titanium
TiO2 nanotube layer
Corrosion
Serum proteins

a b s t r a c t
Titanium oxide nanotubes prepared by anodization have received considerable attention in the biomaterials domain. The objective of this study was to demonstrate the electrochemical behavior of different
diameter TiO2 nanotube layers on titanium in phosphate buffered saline (PBS) and Dulbeccos minimum
essential medium + 10% fetal calf serum (D-FCS) using open circuit potentials (OCP), electrical impedance
spectroscopy (EIS), and a potentiodynamic polarization test. The results showed that the nanotubes had
higher OCP, higher resistance of the inter barrier layer (Rb ), and lower Ipass in the two test solutions
compared to the smooth Ti, especially the 30 nm diameter nanotubes. The corrosion resistance of the
nanotubes in D-FCS was higher than in PBS because of protein adsorption from the D-FCS solution as
suggested by scanning electron microscope (SEM) images. In addition, protein aggregates of 30 nm diameter nanotubes caused the model of EIS spectra to transform from two-layer to three-layer. The corrosion
behavior of the nanotubes for use as a dental implant material is discussed.
Crown Copyright 2011 Published by Elsevier B.V. All rights reserved.

1. Introduction
Titanium is currently one of the most important materials in
biomedical and dental implants [1]. Chemical composition and
topography of Ti-implant were modied by various methods to
improve osseointegration of endosseus implants, which could
eventually play an important role in clinical outcomes [2]. Recently,
the researchers began to modify the implant surface in nanometer
regime because of the in vivo nanometer physiological environment that bone cells are accustomed to [3].
Many studies have shown that the nano-Ti implant can enhance
osteoblast adhesion and function more signicantly than what
the conventional Ti implant dose [413]. Corrosion studies are
also essential because higher corrosion rate means more ion
release, which may interfere with cell metabolism in tissues around
implants, even induces implant failure [14,15]. Furthermore, the
greater real surface area in nano-Ti implants may exhibit a lower
corrosion resistance than the conventional Ti implant [5,6,16].
However, up to now only a few corrosion studies have been carried
out on nano-implants [1719]. Zheng et al. studied the corrosion behavior of TiTiCTiC/diamond-like carbon (DLC) gradient
nano-composite lms on NiTi alloy in hanks solution. They found
that nano composite lms coated NiTi showed excellent corrosion
resistance properties than uncoated NiTi [17]. Karpagavalli et al.

Corresponding author. Tel.: +86 021 23271699 5694; fax: +86 021 63136856.
E-mail addresses: fredzc@online.sh.cn, qiang1951@126.com (F.-q. Zhang).

reported that Ti6Al4 V with nanostructured TiO2 lms deposited

eletrolytically could enhance the corrosion resistance compared to
the bare Ti6Al4V [18]. But some reports exhibited contrary corrosion performance. For example, Garbacz et al. found that the
nano-Ti by hydrostatic extrusion demonstrated lower corrosion
resistance than the micro-Ti in a NaCl solution did [19]. So the
corrosion behaviors of nano-Ti in physiological environment need
further investigation.
Recently, the biological characteristics of TiO2 nanotube layers by anodization have been extensively studied and shown
promising results in biomaterial application [2026]. The titanium with nanotube layers can not only signicantly increase
osteoblast or bone-forming cell adhesion and function in vitro
[2224], but also promote in vivo bone formation around implants
compared to their unanodized counterparts [25,26]. However, up
to now, few articles about corrosion behavior of nanotubes have
been published [27,28]. Saji et al. studied the corrosion behavior
of Ti35Nb5Ta7Zr alloy with amorphous nanotubular oxide in
Ringers solution at 37 C. Their results showed that this nanotube
layer had good corrosion resistance due to forming an immediate and effective passivation [27]. Yu et al. tested the corrosion
characteristics of titanium with amorphous and anatase nanotubes
in Hanks solution and found that this nanotube layer, especially
the anatase nanotubes, exhibited better corrosion resistance than
smooth-Ti [28]. The Ringers and Hanks solution only mimic inorganic components of body uids [29]. Many studies have shown
that the serum proteins of body uids can play an important role
in corrosion of the implant materials [3034]. So it is more rep-

0927-7765/$ see front matter. Crown Copyright 2011 Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.colsurfb.2011.01.033

W.-q. Yu et al. / Colloids and Surfaces B: Biointerfaces 84 (2011) 400405

resentative of physiological environment to test the corrosion of

TiO2 nanotube layers in solution with serum proteins. Furthermore,
the diameter of nanotubes has been proved to affect the adsorb
of serum proteins, which can inuence cell behavior [35,36]. It is
not known whether the adsorb of serum proteins can change the
corrosion of nanotubes on titanium.
The present work was to test corrosion behavior of titanium
with different diameter nanotube layers in PBS solution and uids
with serum proteins and detected by scanning electron microscope (SEM), electrochemical impedance spectroscopy (EIS) and
potentiodynamic polarization. This study aims to enhance the
understanding of corrosion characteristics of nanotubes on titanium in physiological environment.
2. Materials and methods
2.1. TiO2 nanotube layer preparation
Ti thin foils (0.25 mm thick, 99.5%, Alfa Aesar) were used to prepare the nanotubes. Firstly, titanium foils were immersed in an
mixture (2 mL 48% HF, 3 mL 70% HNO3 (both reagent grade chemicals) and 100 mL deionized water) for 3 min to remove the natural
oxide layer, then ultrasonically washed in deionized (DI) water for
5 min and dried by nitrogen stream. The nanotube layers were prepared by anodization in 1 M H3 PO4 and 0.5 wt% HF solution for 3 h
at different voltages (5 V and 15 V) to obtain different diameter
nanotubes [35]. The counter electrode was platinum plate. After
anodization, the samples were ultrasonically washed in DI water
and dried by nitrogen stream. The Ti foil was polished with SiC
emery papers No. 240, 400, 600, 800, 1000 and 1200 grit sizes in
series for use as a control group sample.
2.2. Surface characterization
Some samples were immersed in 30 mL test solutions at 37 C for
half an hour. Then the samples were picked up and gently washed
with DI water. Their microstructures were observed by Scanning Electron Microscopy (Philips, The Netherlands, Sirion200). All
specimens were sputter-coated with Au before imaging using a
HUMMER I Sputtercoater for 3 min.
2.3. Preparation of corrosion test
The three kinds of test specimens used in this research were (a)
smooth Ti (S), (b) the nanotube layers at 5 V (N-5), and (c) the nanotube layers at 15 V (N-15). The mechanical and electrical contact
between the specimen and the cupper wire were obtained by electric conduction paste with copper. The cold epoxy resin was used to
carefully mount the samples. The exposed test surface (area:1 cm2 )
was ultrasonically washed in deionized water, ethanol, and then
deionized water for about 5 min, respectively. Three specimens
were prepared for each condition to corrosion test.
Two synthetic solutions simulating physiological environment
were used in this study. That is phosphate buffered saline (PBS)
and Dulbecco minimum essential medium + 10% fetal calf serum
(D-FCS). The solution was prepared from deionized water, analytic
grade agents, Dulbecco minimum essential medium, and fetal calf
serum. The main components of solution are shown in Table 1 [32].
2.4. Corrosion test
The apparatus for electrochemical measurement consisted of
a computer-controlled potentiostat (PARSTAT 2273 Advanced
Electrochemical System) with research corrosion software (PowerSuite), an Ag/AgCl with a saturated potassium chloride as a
reference electrode, a platinum plate as a counter electrode, and a

401

Table 1
The component (mol/l) of solutions.
Component
NaCl
KCl
CaCl2
Na2 HPO4
KH2 PO4
MgSO4
NaHCO3
Amino acid
Proteins (g/l)

PBS

D-FCS
1

1.37 10
2.68 103
8.10 103
1.47 103

1.16 101
5.36 103
1.80 103
8.98 104
8.11 104
2.382.62 102
5.5 103
<10

standard 3-electrode microcell. After the specimen was immersed

in PBS and D-FCS solution, the open circuit potential (Eocp ) was
measured for half an hour at 4 s intervals. The electrochemical
impedance spectra (EIS) measurements were taken at the Eopen
achieved in the E vs. t test using a frequency response analyzer
coupled with the potentiostat. The frequency span analysed ranged
from 102 to 105 Hz with a perturbing signal of 10 mV. The EIS data
tting was done using the suitable equivalent circuit by the ZSimpWin 3.21 software. The potentiodynamic polarization behavior of
specimens was recorded after immersed for 90 min in test solution.
The scan range of the potentiodynamic polarization was 400 mV
to +3000 mV (vs. open circuit potential) at a scanning rate of 1 mV/s.
All tests were maintained at 37 C. All the tests were repeated three
times with each sample and solution. The reproducibility of the data
by repeating measurements on a series of electrodes was within 5%.
All experiment data were statistically analysed using ANOVA and
Scheffes test ( = 0.05) of SPSS 11.0.
3. Results
3.1. The morphology of surfaces
The morphology of nanotube layers in different test solutions
was shown by Fig. 1. The N-5 and N-15 is the nanotube layers with
approximately 30 nm and 70 nm diameter by anodized at 5 V and
15 V (Fig. 1A). The morphologies of all the surfaces after immerged
in PBS were the same as before (Fig. 1B). However, SEM images
revealed that protein aggregates deposited from D-FCS settle on
the surfaces after immerged in D-FCS for half an hour. Few protein
aggregates were found in S and N-15 surfaces, but lots of protein aggregates were deposited on N-5 surface so that the pores
of nanotubes were sealed.
3.2. Eocp in different solutions
Open circuit potential (Eocp ) data after half an hour immersion
in PBS and D-FCS solutions were reported in Fig. 2. The Eocp of N-5
was signicantly higher than that of S and N-15 in two test solutions
(P < 0.05). The Eocp of S and N-15 in D-FCS became lower than in PBS
test solutions (P > 0.05). But the Eocp of N-5 in D-FCS became higher
than in PBS test solutions (P > 0.05).
3.3. EIS analysis
Fig. 3 shows the electrochemical impedance data for S, N-5, and
N-15 in PBS and D-FCS solutions for half an hour. The Bode-phase
plots showed clear two time constants in the N-15 surfaces. But
only one time constant was seen in the S and N-5 surfaces.
The spectra obtained for S was interpreted by using one time
constant due to passive oxide layer. The N-5 in PBS and N-15 in PBS
and D-FCS solutions were interpreted by using a model with two
constants because of the nanotubes with an outer tube layer and
an inner barrier layer [27] (even though only one time constant

402

W.-q. Yu et al. / Colloids and Surfaces B: Biointerfaces 84 (2011) 400405

Fig. 1. High magnication SEM images of the samples in different test solutions ((A) unimmersed, (B) immersed in PBS, and (C) immerged in Dulbecco minimum essential
medium + 10% fetal calf serum and (a) smooth Ti, (b) the nanotube layers at 5 V, and (c) the nanotube layers at 15 V, 150,000).

in N-5 shown by Bode-phase plots). Furthermore, the N-5 in DFCS solution was interpreted by using three time constant because
protein aggregates were deposited on nanotubes surface. In our
studies, a constant-phase element (CPE) was use to replace the pure
capacitance in the tting of the EIS because of the rough surface
acquired by 1200-SiC-abrase paper on S and the porous surface on
N-5 and N-15. The impedance of CPE is dened as [31],
n 1

ZCPE = [Y0 (jw) ]

(1)

where n is the empirical exponent of the CPE, 1 < n < 1, if n = 1,

CPE is the pure inductance; if n = 1, CPE is the pure capacitance; if
n = 0, CPE is the pure resistance.

The components of this equivalent circuit are electrolyte resistance (Rs ), resistance of outer tube layer (Rt ), resistance of inter
barrier layer (Rb ), and resistance of the protein aggregates (Rp ). The
representative graphic examples of the excellent tting accuracy
obtained with simulation of model in Fig. 4 were shown in Fig. 5.
The 2 -values 103 for S and 104 for N-5 and N-15 have shown
the good t quality between experimental points and tted points
(Table 2).
The resistance, constant-phase element, and n values obtained
by choosing suitable model of equivalent circuit for all samples
were given in Table 2. The higher barrier layer resistance (Rb )
suggested better corrosion resistance of titanium with 30 nm and
70 nm diameter nanotubes compared to the smooth Ti in two test

Fig. 2. (A) Representative E vs. T-curves and (B) Open circuit potential (Eocp ) on surfaces in the test solutions. (PBS: phosphate buffered saline, D-FCS: Dulbecco minimum
essential medium + 10% fetal calf serum, S: smooth Ti, N-5: the nanotube layers at 5 V, and N-15: the nanotube layers at 15 V; mean (standard deviation), n = 3, *P < 0.05).

W.-q. Yu et al. / Colloids and Surfaces B: Biointerfaces 84 (2011) 400405

403

Fig. 3. The Bode-phase plots for smooth Ti and nanotubes in PBS (phosphate
buffered saline) and D-FCS (Dulbecco minimum essential medium + 10% fetal calf
serum) solutions (S: smooth Ti, N-5: the nanotube layers at 5 V, and N-15: the
nanotube layers at 15 V).

Fig. 4. Model of equivalent circuit obtained from the EIS-results: (A) Smooth Ti
in phosphate buffered saline and Dulbecco minimum essential medium + 10% fetal
calf serum solutions. (B) The nanotube layers at 5 V in phosphate buffered saline and
the nanotube layers at 15 V in phosphate buffered saline and Dulbecco minimum
essential medium + 10% fetal calf serum solution. (C) The nanotube layers at 5 V in
Dulbecco minimum essential medium + 10% fetal calf serum solution.

solutions. The highest Rb is shown in N-5 surface (33  105 ) in

PBS solutions, but in N-15 surface (56  105 ) in D-FCS solutions
(P < 0.05). Furthermore, the Rb values of all the surfaces in PBS are
lower than those in D-FCS (P < 0.05). The n values of barrier layers
in all samples are about 0.9. The higher Rt of N-15 is shown compared to N-5, but the Y0 -CPEt tendency is opposite (P < 0.05). The
Rp , Y0 -CPE, and n results of N-5 have also shown that proteins of
D-FCS could aggregate on the 30 nm nanotubes after immerged in
D-FCS for half an hour.

Fig. 5. Bode plot for representative samples ((A) smooth Ti in phosphate buffered
saline and Dulbecco minimum essential medium + 10% fetal calf serum solutions, (B)
the nanotube layers at 5 V in phosphate buffered saline and the nanotube layers at
15 V in phosphate buffered saline and Dulbecco minimum essential medium + 10%
fetal calf serum solution, and (C) the nanotube layers at 5 V in Dulbecco minimum
essential medium + 10% fetal calf serum solution). EIS-result with the experimental
points of phase () and module |Z| (Msd.) and the tted points (Calc.) obtained with
the simulation of the model shown in Fig. 4.

Table 2
Values for the elements of the equivalent circuit (R, CPE, n) shown in gure and 2 .
Group

Rb

Y0 -CPE

PBS: S
PBS: N-5
PBS: N-15
D-FCS: S
D-FCS: N-5
D-FCS: N-15

7 (1.3)
33.2 (2.3)
25.1 (6.4)
26.7 (2.9)
42.1 (3.3)
56.5 (5.0)

4.4 (0.21)
2.3 (0.41)
1.9 (0.61)
3.4 (0.23)
4.4 (0.43)
8.8 (1.67)

0.91 (0.002)
0.91 (0.002)
0.93 (0.015)
0.90 (0.004)
0.99 (0.004)
0.93 (0.032)

Rt

Y0 -CPE

42.8 (2.2)
264.0 (108.0)

59.2 (14.3)
19.5 (6.2)

0.53 (0.013)
0.57 (0.97)

18.4 (2.4)
231.6 (90.1)

194.5 (52.1)
21.9 (5.3)

0.53 (0.042)
0.59 (0.029)

Values: mean (standard deviation); Rt (); Y0 -CPE (S cm2 sn 105 ); Rb and Rp ( 105 ), n = 3.

Rp

2.7 (1.3)

Y0 -CPE

7.5 (1.53)

0.746 (0.032)

2
103
104
104
103
104
104

404

W.-q. Yu et al. / Colloids and Surfaces B: Biointerfaces 84 (2011) 400405

Fig. 6. Typical diagrams of the potentiodynamic polarization for the samples (PBS:
phosphate buffered saline, D-FCS: Dulbecco minimum essential medium + 10% fetal
calf serum, S: smooth Ti, N-5: the nanotube layers at 5 V, and N-15: the nanotube
layers at 15 V).

3.4. Potentiodynamic polarization

Fig. 6 shows the typical potentiodynamic polarization curves
for all surfaces in the two test solutions. The corrosion parameters about Ipass , Ebreak , and Icorr were acquired by potentiodynamic
polarization curve analysis [29,37]. The Ipass which is passive current density was obtained from the potentiodynamic polarization
curve at 1 V. The Ebreak is the potential where the current density
remarkably increases with increasing potential. Icorr means the corrosion rate. The corrosion parameters are showed in Table 3. The
value trends of Ipass are N-5 < N-15 < S acquired by one-way ANOVA
( = 0.05) in the two test solutions. Moreover, the Ipass values of all
the surfaces in PBS are higher than those in D-FCS, but not statistically signicant. The Ebreak values for N-15 were highest among
the surfaces in the two test solutions. The N-15 Icorr values for PBS
and D-FCS were 145 and 142 nA/cm2 , respectively, and were signicantly higher than that of S and N-5 (P < 0.05). The Icorr values
of S and N-5 in PBS are signicantly higher than those in D-FCS
(P < 0.05).
4. Discussion
In this study, the OCP of all the surfaces in the two test solutions rose rapidly and then gradually reached a stable value. This
is because a titanium oxide lm began to grow and thicken after
the samples were immerged in the solutions, which resulted in
increased corrosion resistance and subsequent a relatively stable
state [30,38]. Our results have shown that the OCP of S and N-15
in PBS is higher than that in D-FCS at 37 C. Contu et al. [39] suggested that the slope of the oxygen cathodic reduction in PBS was
lower than that in the solution containing proteins. The anodic and
Table 3
Corrosion parameters tested in this research. The parameters are mean (standard
deviation), n = 3.
Group

Ipass (nA/cm2 )

Ebreak (mV)

Icorr (nA/cm2 )

PBS: S
PBS: N-5
PBS: N-15
D-FCS: S
D-FCS: N-5
D-FCS: N-15

6380 (81)
1473 (419)
4390 (331)
5840 (65)
578 (99)
3920 (127)

1276 (80)
1580 (79)
1643 (15)
1293 (28)
1386 (51)
1650 (60)

76 (9)
45 (7)
145 (6)
23 (3)
20 (5)
142 (21)

cathodic reactions of electrode determine the OCP. In this respect,

the corrosion resistance of S and N-15 decreased due to proteins
contained in D-FCS. However, the OCP value of N-5 in PBS is lower
than in D-FCS, which may because a protective smear layer was
formed on N-5 surface by the protein aggregates deposited from DFCS (Fig. 1). Furthermore, the OCP values for N-5 and N-15 stabilized
in the two test solutions suggested the better corrosion resistance
properties than those for S, especially, there was statistically signicant difference in OCP values between N-5 and S. This tendency
lies in the fact that the N-5 and N-15 have not only TiO2 tube layer
but also TiO2 barrier layer, which is thicker than the native oxide
lm of S [29].
In our studies, a two-layer (barrier oxide layer and nanotube
layer) model was established to interpret the EIS spectra of N-5 in
PBS and N-15 in the two test solutions. This agrees with reports of
Saji and Yu et al. [28,29]. After the N-5 immerged in D-FCS for hour
an half, a three-layer (barrier oxide layer, nanotube layer, and protein layer) was used to interpret the EIS spectra. The protein layer
on N-5 was conrmed by SEM image in Fig. 1C-b. Oh et al. [36]
found that serum proteins proned to aggregate smaller nanotube
layer. We veried this tendency. At the same time, our results indicate that the two-layer model of EIS spectra can be transformed to a
three-layer one when the corrosion behaviors of smaller nanotubes
are investigated on solution with serum proteins.
The results of bode-phase spectra (phase of Z vs. frequency) suggested that a compact passive lm was present on the S and N-15
surfaces in the two test solutions. The cause may lie in the fact
that the phase angles of the S and N-15 surfaces are close to 90
over a wide frequency range, a characteristic of the compact passive lm reported by Delgado-Alvarado and Sundaram [40]. At the
same time, the surface of the N-15 may form the compact passive
lm in the two test solutions due to the phase angle about 90 at
high frequency range. The EIS data were analysed using the model
of Fig. 4 to determine all the R and CPE. The barrier layer resistance (Rb ) is an important parameter whose values are positively
proportional to corrosion resistance [18,29]. The N-5 and N-15 layers have higher Rb values compared to the S surfaces in the two
test solutions. On one hand, the barrier oxide layers of the N-5
and N-15 by anodization are thicker than that of the S surfaces. On
the other hand, the nanopore and nanotube structure could adsorb
more proteins compared with the S surfaces in the D-FCS solution
[46], which is highlighted by the physical evidence provided by
the SEM images in Fig. 1. Our results have shown that all the samples in D-FCS have higher Rb values than in PBS. It is possible that
the adsorbed proteins from the D-FCS increase the corrosion resistance of all the surfaces by covering the electrode surface [41]. The
protein adsorption is conrmed by SEM in Fig. 1. Furthermore, the
Rt values of the N-15 are higher than that of the N-5, and the reason may be that the longer nanotubes of the N-15 increases the
Rt value compared to the N-5. The length of nanotubes becomes
longer when the voltages of anodization increase from 5 V to 15 V
[42]. (Our results about nanotube length have been used in another
article.) The n values of barrier layers for all the samples are 0.9,
which indicates that the barrier layers of all the samples show a
similar pure capacitive behavior [43]. The n value of tube layers for
the N-5 and N-15 are about 0.55, which means that the two surfaces have a considerable deviation from ideal capacitor behavior
because of their porous surface [29,43].
As for corrosion data acquired from the potentiodynamic polarization curves, the N-5 and N-15 exhibited lower current density
of Ipass than the S surfaces. This may be due to an immediate and
effective passivation on the N-5 and N-15 surfaces [28]. But current
density of Icorr for nanotubes (N-15) is greater than for the smooth Ti
(S), which means lower corrosion resistance for the nanotubes compared to the smooth Ti. Saji et al. [27] also found this phenomenon
when studying corrosion characteristic of nanotubular oxide on

W.-q. Yu et al. / Colloids and Surfaces B: Biointerfaces 84 (2011) 400405

Ti35Nb5Ta7Zr alloy in Ringers solution. They thought it was

the tube bottom interface shaped by distinctly separated barrier
oxide/concave that contributed to the lower corrosion resistance
of the naotubes. Our nanotubes also formed the similar interface,
which seem to be a reason for lower corrosion on N-15. The Ipass
and Icorr in D-FCS all became lower than in PBS, especially the Ipass
of the N-5 surfaces. This is because the proteins can deposit on the
surfaces to alleviate corrosion process. Therefore, the protein aggregates on the N-5 surface performed effective passivation, which
signicantly decreased the current dentisy of Ipass [30,41].
5. Conclusions
The nanotube layers of titanium with different diameter, especial at 30 nm, show better corrosion resistance than smooth Ti due
to higher OCP, Rb , and lower Ipass in PBS and D-FCS solution. The
corrosion resistance of nanotubes in a solution containing serum
proteins increased in comparison with that in PBS by analysing OCP,
EIS, and potentiodynamic polarization. Moreover, the model of EIS
spectra can transform from a two-layer to a three-layer one because
of proteins deposited from D-FCS solution, which indicates that it
should be careful to choose the EIS spectra model when investigating the corrosion behavior of nano-implant materials in solution
with serum protein. Our results exhibit that the titanium with
nanotubes can improve the corrosion resistance in physiological
environment.
Acknowledgements
The authors would like to thank Bang-shang Zhu (Instrumental Analysis Center, Shanghai Jiao Tong University) for helpful
assistance in experiments. This work was supported by Shanghai Leading Academic Discipline Project (Project No.: S30206) and
Science and Technology committee of Shanghai (08DZ2271100,
1052nm04300, and 10JC1408600) and Shanghai Leadind Academic
Discipline Project (T0202), and National Natural Science Foundation of China (81070866).
References
[1] T. Albrektsson, L. Sennerby, Int. J. Prosthodont. 3 (1990) 30.
[2] L. Le Guhennec, A. Soueidan, P. Layrolle, Y. Amouriq, Dent. Mater. 23 (2007)
844.
[3] G. Mendonca, D.B. Mendonca, F.J. Arago, L.F. Cooper, Biomaterials 29 (2008)
3822.
[4] T.J. Webster, R.W. Siegel, R. Bizios, Biomaterials 20 (1999) 12221227.
[5] T.J. Webster, C. Ergun, R.H. Doremus, R.W. Siegel, R. Bizios, Biomaterials 21
(2000) 1803.

405

[6] A. Bigi, N. Nicoli-Aldini, B. Bracci, J. Biomed. Mater. Res. A 82 (2007) 213.

[7] M. Sato, A. Aslani, M.A. Sambito, N.M. Kalkhoran, E.B. Slamovich, T.J. Webster,
J. Biomed. Mater. Res. A 84 (2008) 65.
[8] K.L. Elias, R.L. Price, T.J. Webster, Biomaterials 23 (2002) 3279.
[9] T.J. Webster, E.L. Hellenmeyer, R.L. Price, Biomaterials 26 (2005) 953.
[10] P. Li, J. Biomed. Mater. Res. A 66 (2003) 79.
[11] E.E. Swan, K.C. Popat, C.A. Grimes, T.A. Desai, J. Biomed. Mater. Res. A 72 (2005)
288.
[12] M. Kalbacova, B. Rezek, V. Baresova, C. Wolf-Brandstetter, A. Kromka, Acta
Biomater. 5 (2009) 3076.
[13] E. Lamers, X.F. Walboomers, M. Domanski, J. Riet, F.C. Delft, R. Luttge, L.A. Winnubst, H.J. Gardeniers, J.A. Jansen, Biomaterials 31 (2010) 3307.
[14] J.J. Jacobs, J.L. Gilbert, R.M. Urban, J. Bone Joint. Surg. Am. 80 (1998) 262.
[15] M. Hosoki, E. Bando, K. Asaoka, H. Takeuchi, K. Nishigawa, Biomed. Mater. Eng.
19 (2009) 53.
[16] C. Aparicio, F.J. Gil, C. Fonseca, M. Barbosa, J.A. Planell, Biomaterials 24 (2003)
263.
[17] Y. Zheng, D. Liu, X. Liu, L. Li, Biomed. Mater. 3 (2008) 044103.
[18] R. Karpagavalli, A. Zhou, P. Chellamuthu, K. Nguyen, J. Biomed. Mater. Res. A 83
(2007) 1087.
[19] H. Garbacz, M. Pisarek, K.J. Kurzydowski, Biomol. Eng. 24 (2007) 559.
[20] S.H. Oh, R.R. Finnes, C. Daraio, L.H. Chen, S. Jin, Biomaterials 26 (2005)
4938.
[21] A. Kodama, S. Bauer, A. Komatsu, H. Asoh, S. Ono, P. Schmuki, Acta Biomater. 5
(2009) 2322.

[22] S. Oh, C. Daraio, L.H. Chen, T.R. Pisanic, R.R. Finones,

S. Jin, J. Biomed. Mater. Res.
A 78 (2006) 97.
[23] K.C. Popat, L. Leoni, C.A. Grimes, T.A. Desai, Biomaterials 28 (2007) 3188.
[24] C. Yao, E.B. Slamovich, T.J. Webster, J. Biomed. Mater. Res. A 85 (2008)
157.
[25] C. von Wilmowsky, S. Bauer, R. Lutz, M. Meisel, F.W. Neukam, T. Toyoshima, P.
Schmuki, E. Nkenke, K.A. Schlegel, J. Biomed. Mater. Res. B: Appl. Biomater. 89
(2009) 165.
[26] L.M. Bjursten, L. Rasmusson, S. Oh, G.C. Smith, K.S. Brammer, S. Jin, J. Biomed.
Mater. Res. A 92 (2010) 1218.
[27] V.S. Saji, H.C. Choe, W.A. Brantley, Acta Biomater. 5 (2009) 2303.
[28] W.Q. Yu, J. Qiu, L. Xu, F.Q. Zhang, Biomed. Mater. 4 (2009) 065012.
[29] Y. Okazaki, T. Tateishi, Y. Ito, Mater. Trans. JIM 38 (1997) 78.
[30] X. Cheng, S.G. Roscoe, Biomaterials 26 (2005) 7350.
[31] G. Rondelli, P. Torricelli, M. Fini, R. Giardino, Biomaterials 26 (2005) 739.
[32] B. Bozzini, P. Carlino, L. DUrzo, V. Pepe, C. Mele, F. Venturo, J. Mater. Sci. Mater.
Med. 19 (2008) 3443.
[33] S. Hiromoto, E. Onodera, A. Chiba, K. Asami, T. Hanawa, Biomaterials 26 (2005)
4912.
[34] D.D. Zabel, S.A. Brown, K. Merrit, J.H. Payer, J. Biomed. Mater. Res. A 22 (1998)
31.
[35] J. Park, S. Bauer, K. vonder Mark, P. Schmuki, Nano Lett. 7 (2007) 1686.
[36] S. Oh, K.S. Brammer, Y.S. Li, D. Teng, A.J. Engler, S. Chien, S. Jin, Proc. Natl. Acad.
Sci. U. S. A. 106 (2009) 2130.
[37] S. Takemoto, M. Hattori, M. Yoshinari, E. Kawada, Y. Oda, Biomaterials 26 (2005)
829.
[38] S.L. De Assis, S. Wolynec, I. Costa, Electrochim. Acta 51 (2006) 1815.
[39] F. Contu, B. Elsener, H.B. ohni, J. Biomed. Mater. Res. A 62 (2002) 412.
[40] C. Delgado-Alvarado, P.A. Sundaram, Acta Biomater. 2 (2006) 701.
[41] R.L. Messer, G. Tackas, J. Mickalonis, Y. Brown, J.B. Lewis, J.C. Wataha, J. Biomed.
Mater. Res. B: Appl. Biomater. 88 (2009) 474.
[42] J.L. Zhao, X.H. Wang, T. Sun, L. Li, J. Alloys Compd. 434435 (2007) 792.
[43] M. Aziz-Kerrzo, K.G. Conroy, A.M. Fenelon, S.T. Farrell, C.B. Breslin, Biomaterials
22 (2001) 1531.