Biomaterials 23 (2002) 2887–2894

On the properties of two binary NiTi shape memory alloys.

Effects of surface finish on the corrosion behaviour and
in vitro biocompatibility
Mohammed Es-Sounia,*, Martha Es-Sounib, Helge Fischer-Brandiesb
a
University of Applied Sciences, Surface and Thin Film Technology, Materials Testing and Joining, Grenzstr. 3, D-24149 Kiel, Germany
b
Clinic of Orthodontics, Christian-Albrecht-University, Arnold Heller Str., D-24113, Kiel, Germany
Received 30 August 2001; accepted 5 December 2001

Abstract

The present paper compares the transformation behaviour and mechanical properties of two orthodontic wires of close chemical
compositions. The effects of surface topography and surface finish residues on the potentiodynamic corrosion behaviour and
biocompatibility are also reported. The cytotoxicity tests were performed on both alloys in fibroblast cell cultures from human
gingiva using the MTT test. It is shown that the surface finish and the amounts of surface finish residues affect dramatically the
corrosion resistance. Bad surface finish results in lower corrosion resistance. The in vitro biocompatibility, though not affected to
the extent of corrosion resistance, is also reduced as the surface roughness and the amounts of residues increase. This is thought to
be due to surface effects on corrosion and metallic ions release. r 2002 Published by Elsevier Science Ltd.

1. Introduction Ni to Ti ratio but more easily, starting from a common

NiTi-based shape memory alloys constitute an inter-
esting group of smart alloys which enjoy an ever
increasing market share as biomaterials. They are most
widely used as orthodontic arch wires, stents, etc. The
mechanisms of the shape memory effect are based on the
thermoelastic martensitic transformation which lead, in
these alloys, to highly reversible strains either under the
effects of stress (superelasticity through stress-induced
martensite formation) or temperature (thermal shape
memory effect). Details about the mechanisms of the
shape memory effect and the potential applications of
NiTi alloys may be found in numerous review articles,
e.g. [1–3]. The NiTi shape memory alloys usually consist
of binary alloys, with Ni and Ti concentrations near the
equiatomic composition. Their microstructures are
generally processed using complex thermomechanical
treatments in order to obtain suitable properties, e.g.
thermal shape memory, superelasticity, all-round mem-
ory effect, etc. Contrary to a general view, these
properties are usually not obtained via changes in the
*Corresponding author. Tel.: +49-431-210-2660; fax: +49-431-210-
2661.
E-mail address: mohammed.es-souni@fh-kiel.de (M. Es-Souni).
ingot composition, through the selection of appropriate
thermomechanical treatment sequences which can be
better controlled than the chemical composition. The
transformation and superelastic hysteresis properties as
well as the long-term stability are generally positively
influenced by the addition of suitable amounts of the
alloying elements Cu and Fe in the range from 5 to
10 wt% and from 1% to 2%, respectively [1].
The (thermo)mechanical properties of NiTi-base
alloys certainly constitute the outstanding and most
perceptible attribute for their use. However, in biome-
dical applications involving contact with live tissues,
biocompatibility issues may prove as crucial (if not
more) as mechanical behaviour. In a recent work [4–5],
the in vitro biocompatibility of a binary NiTi and a
ternary NiTiCu alloys has been investigated using
epithelial [4] and fibroblast [5] cell cultures. It has been
shown that biocompatibility is negatively influenced by
the presence of Cu which has been explained in terms of
Cu ions release. Usually the biocompatibility of NiTi
alloys has been inferred from their high corrosion
resistance [6–10]. The corrosion resistance may, how-
ever, depend on factors such as surface finish quality,
the amount of residues and the degree of homogeneity
of microstructures. All these factors are known to affect

0142-9612/02/$ - see front matter r 2002 Published by Elsevier Science Ltd.

PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 4 1 6 - 1
2888 M. Es-Souni et al. / Biomaterials 23 (2002) 2887–2894
the corrosion resistance of metals via establishing
corrosion elements which may then lead to higher metal
ions release and consequently to higher cytotoxicity.
Due to the complex thermomechanical processes the
NiTi alloys have to undergo and to the compelling cost
reduction, it is thought that the surface finish and
microstructure may not always be of good quality. In
fact, in a broad survey of 25 commercial arch wires from
different suppliers, bad surface finish has been found in
the majority of them, with deep grooves, surface finish
residues like silicon and aluminium oxides, wear residue
from drawing dies, etc. [11]. This gave the impetus for
this investigation which compares the properties,
including corrosion behaviour and biocompatibility, of
two NiTi arch wires of close chemical compositions but
different surface finish.
2. Materials and methods
The alloys studied are ‘‘Neosentalloy F 80’’ (GAC),
hereafter designated Neosent, and SE NiTi (G&H),
hereafter designated SeNiTi. The samples were in the
form of arch wires with rectangular sections of
0.57
0.42 mm2. For the different investigations, the
straight ends of the wires were used in the as-received
states. The chemical compositions of the alloys were
determined in an analytical scanning electron micro-
scope (SEM) (Philips XL 30, EDAX SUTW Spahire
detector) using energy-dispersive spectroscopy (EDS)
analysis. The following compositions (in wt%), which
have been determined with an accuracy of 70.1% using
a standard NiTi alloy of known composition, were
found: Ni57.6Ti42.4 for Neosent, and Ni57.8Ti42.2 for
SeNiTi. Metallographic investigations were conducted
on electrolytically polished (Lectropol-5, Streuers) speci-
mens in a solution of 5% HClO4 in ethanol under 38 V
and 1.1 A. To reveal microstructure, the specimens were
etched in a solution of 10 ml HF, 20 ml HNO3 and 30 ml
H2O.
The transformation behaviour was investigated using
differential scanning calorimetry (DSC, Perkin Elmer
Pyris I) in the temperature range from 801C to +801C
at a rate of 10 K/min. The samples weighing approxi-
mately 10 mg were gently cut and ultrasonically cleaned
in acetone before testing. A base line was recorded
before each test run. Three batches were tested for each
alloy with, however, reproducible results in the range of
accuracy of the system (71 K).
The mechanical properties were investigated in 3-
point bending tests (Synergie, MTS, 10 N load cell)
using a beam length of 12 mm. The load vs. deflection
curves were recorded at 371C. Heating of the specimens
was conducted in an oil bath of constant viscosity in the
range of the temperature investigated. The temperature
was controlled to 70.51C using a closed-circuit cooler/
heater (hetofrig, Heto). In order to show the effects of
loading and unloading cycles on the mechanical proper-
ties, 5 cycles were generally performed.
The corrosion properties were investigated using
potentiodynamic testing in a Ringer solution (Merck,
pH 7) at 221C and 371C, and a voltage scan rate of
10 mV/s. A saturated calomel electrode was used as a
reference. The solution was changed for every test run.
For each alloy, 3 specimens were tested to verify the
reproducibility of the results.
The biocompatibility of the chosen alloys was
characterized by means of the in vitro MTT cytotoxicity
test on cultured fibroblast cells from explants of human
gingiva. Healthy gingival tissue explants from 5 subjects
(not treated for orthodontic failures) aged between 33
and 59 years, have been used for this purpose. Right
after excision, the explants were cut into cubes of about
1 mm3, and cultured in a-MEM (supplemented with
10% FCS, 2.5 mg/ml amphotericin B, 100 U/ml penicillin
and 100 mg/ml streptomycin, BIOCHROM) in sterilized
non-treated culture dishes (FALCON), at 371C, and 5%
CO2. The medium was changed every 72 h beginning
with the fifth day after seeding. Approximately 11 days
after seeding, the fibroblasts were subcultured. The cell
monolayers were removed by treatment with 1.5 ml
Trypsin/EDTA (0.05/0.02%, BIOCHROM) per dish for
10 min at 371C and 5% CO2. The trypsin activity was
blocked by an addition of 1.5 ml a-MEM (containing
10% FCS) per dish. Subsequently, the suspension was
centrifuged for 7 min at 120g: The supernatant was
carefully aspirated and the cells resuspended in fresh
RPMI 1640 (Gibco BRL) (supplemented with 10%
FCS, 2.5 mg/ml amphotericin B, 100 U/ml penicillin and
100 mg/ml streptomycin, BIOCHROM).
Used in the MTT test, 104 cells were seeded per
microwell.
When used in the MTT test, the tested alloys were cut
into pieces of 5 mm length, 0.42 mm thickness and
0.57 mm height, cleaned ultrasonically in absolute
alcohol and then heat sterilized at 1201C for 20 min.
For the cytotoxicity determination of the tested
alloys, the method of Tada et al. [12] was applied. The
test was performed in sterile 96-microtiter plates
(COSTAR) at 371C and 5% CO2 equipped with 2
pieces (10.38 mm3 active surface) of the tested alloys.
The filtered MTT solution (100 ml/well, 5 mg/ml) was
added to the cell cultures (i) 24 h after seeding, (ii) 48 h
after seeding and (iii) 72 h after seeding. Four hours
after the MTT addition, cell lysis and formazan
solubilization were started by adding 100 ml of the lysing
buffer (10% SDS in 0.01 n HCl pH 3.5) and incubating
the plates for 20 h at 371C and 5% CO2. Wells
containing cells but no alloys were used for negative
controls (NCs). The positive controls (PCs) consisted of
cells containing wells, where the addition of the MTT
solution and lysing buffer were inverted. For each alloy
 M. Es-Souni et al. / Biomaterials 23 (2002) 2887–2894 2889

blanks containing the metal, but no cells, were prepared. On cooling, only the R-phase transformation is
All experiments were carried out in triplicate. Prior to observed in the temperature range investigated, and
the determination of the formazan formation by the martensitic transformation is shifted towards tem-
absorption measurements at 550 nm in a microplate peratures below 801C (although the beginning of the
reader (DYNATECH, MR 5000), 175 ml supernatant of martensitic transformation may be seen in the DSC
each well were transferred into a new plate. The curve with an Ms temperature of approximately 721C).
incoherent background was determined by absorption
measurements at 630 nm. Background, due to the
absorption of the medium and MTT was corrected 3.2. Mechanical properties
by subtracting each blank from the corresponding
sample. The stress–strain curves determined from 3-point
bending tests at 371C are illustrated in Fig. 2. Both
alloys show the superelastic hysteresis characteristic for
the austenitic microstructure. A peak stress, which
denotes the beginning of the stress-induced martensite
3. Results transformation, can be seen in both curves. The
appearance of this peak stress on loading and unloading
3.1. Transformation behaviour at temperatures above As suggests that for both alloys
a critical stress has to be overcome before the
Fig. 1 illustrates the DSC curves of the alloys austenite"martensite transformations take place, and
investigated, and Table 1 shows the transformation .
is in many aspects similar to the Luders deformation
temperatures. The Neosent is characterized by the R- [14,15]. The flexure modulus is practically the same for
phase transformation on cooling with the typical low both alloys, the strength properties are, however, very
hysteresis temperature taking into account the peak to different. Comparing the loading and unloading plateau
peak temperatures of the austenite and the R-phase stresses given by s1:0 offset stresses, higher values are
peaks. The martensitic transformation is shifted towards obtained for SeNiTi because of the thermal stabilization
lower temperatures. The endothermic peak denotes the of the austenite phase. It can also be seen that repeated
austenitic transformation with an austenite finish loading and unloading cycles result in a gradual
temperature of 331C. The transformation behaviour of softening of SeNiTi in contrast to Neosent, where
SeNiTi offers a more complex picture. Two well- practically no strength loss occurs.
separated endothermic peaks are observed, which
suggests that the specific thermomechanical treatments
used for this arch wire might have resulted in hetero- Table 1
geneous microstructures with different transformation Transition temperatures determined from the DSC curves
behaviours. Similar results have been reported by
Wire Mf (1C) Ms (1C) Rf (1C) Rs (1C) As (1C) Af (1C)
Ohakata and Tamura [13] who show that heat
treatments below 753 K subsequent to cold working Neosent 59 23 15 21 22 33
SeNiTi o80 E72 1 15 8a 18.5a
lead to a double endothermic peak in the DSC curve.
Unfortunately, no attempt was made to explain these a
Only the high temperature endothermic peak is considered. A:
results in their work. austenite, R: R-phase, M: martensite. Start: s, finish: f.

22 22
(a)
(b)
21 GAC / Neo Sentalloy F80 G+H / SE NiTi

As = 22 ˚C 21
Heat Flow [mW]
Heat Flow (mW)

20 Af = 33 ˚C

19

20
18
Ms = -23 ˚C Rs = 21 ˚C Rf
17 : Cooling
: Heating Mf = -59 ˚C Rf = 15 ˚C Rs (heating)
: Base-line 19
16 (baseline)
(cooling)

-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60

Sample Temperature [˚C]
Sample Temperature (˚C)
Fig. 1. The DSC curves of Neosent (a) and SeNiTi (b).
2890 M. Es-Souni et al. / Biomaterials 23 (2002) 2887–2894

1200 (a) 1200 (b)

Bending Stress (MPa) Neosentalloy, 37 ˚C G+H/SE NiTi, 37 ˚C

Bending Stress [MPa]

1000 1000
σ1.0 max = 786 MPa
800 800

σ1.0 max = 485 MPa Eb = 43 GPa

600 Eb = 43 GPa 600

400 400
σ1.0 min = 456 MPa

200 200
σ1.0 min = 117 MPa
0 0
0 1 2 3 4 0 1 2 3 4

Bending Strain [%] Bending Strain [%]

Fig. 2. Bending properties at 371C of the alloys investigated: (a) Neosent and (b) SeNiTi. The loading–unloading cycles were repeated 5 times.

Fig. 3. BSE micrographs showing the surface topography of SeNiTi (a) and Neosent (b).

Fig. 4. BSE micrograph of the SeNiTi surface and corresponding elemental EDS mappings of oxygen Ka, titanium Ka and silicon Ka.

3.3. Surface topography and microstructures grooves over the whole length of the wire. The dark
contrast of these grooves suggests that they may be
Figs. 3a and b show back-scattered electron (BSE) oxidized areas. Fig. 4 which illustrates elemental EDS
micrographs of the surface morphology of the alloys mappings of SeNiTi reveals in fact that the surface is
investigated. Neosent shows a quite smooth surface with badly oxidized (see the oxygen elemental mapping) in
a two-phase microstructure consisting of feathery, the grooves and contains many surface finish residues
martensitic, and blocky, austenitic areas. Homoge- such as silicon oxide.
neously distributed second-phase particles (dark dots Metallographic investigations of electrolytically
in Fig. 3b) which have been proven via EDS analysis to polished and etched specimens reveal a duplex micro-
be Ti-carbides and nitrides can also be seen. In contrast, structure of recrystallized (smooth) and non-recrystal-
the surface topography of SeNiTi exhibits longitudinal lized (feathery), elongated grains, see Fig. 5. The general

Fig. 5. SEM micrographs of the duplex microstructure of SeNiTi showing strongly deformed (feathery) and recrystallized (smooth) areas. Second-
phase particles of titanium carbide and titanium nitride can also be seen.
0.6
0.4
Current density (A/cm²)
0.2
0.0
-0.2
-0.4
-0.6
-1.0 -0.5 0.0
Voltage (V)
Fig. 6. Current density vs. voltage for Neosent and SeNiTi at 221C
and 371C in the Ringer solution.
aspect of the microstructure suggests that the feathery
appearance might be due to heavy deformation with a
high density of glide lines and deformation substructures
(the feathery appearance is not due to martensite nor
due to the R-phase because the alloy has an Af
temperature of 18.51C). Furthermore, the presence of
numerous elongated second-phase particles can be seen.
All these particles show tails denoting stress accumula-
tion and void formation at the matrix–particle interface
during deformation. As for Neosent, EDS analysis of
the particles shows that they consist of titanium carbide
and titanium nitride.
3.4. Potentiodynamic corrosion properties
The potentiodynamic corrosion properties are shown
in Fig. 6 which reveals unambiguously a higher corro-
sion resistance of Neosent at both temperatures. While
the pitting corrosion potentials of Neosent are approxi-
mately 1.4 V at 221C and 1.2 V at 371C, they reach 0.5
and 0.43 V in the case of SeNiTi.
Since the alloys investigated have almost the same
composition, the different corrosion behaviour results
M. Es-Souni et al. / Biomaterials 23 (2002) 2887–2894 2891
obviously from the difference in surface finish. SEM
micrographs of the corrosion tested samples reveal for
SeNiTi a severe corrosion attack with large pits which
predominantly follow the patterns of the above-men-
tioned surface grooves and surface finish residues, see
Fig. 7.
In contrast, the Neosent alloy surface shows only
isolated areas with moderate corrosion attack and small
pits which seem to be preferentially located along the
Neosent 22˚C rolling direction. The second-phase particles constitute
Neosent 37˚C
SeNiTi 22˚C also the loci of small pits, see Fig. 8.
SeNiTi 37˚C The difference in corrosion behaviour observed is not
surprising since it is well known that the quality of
0.5 1.0 1.5
surface finish and the density of embedded surface finish
residues (grinding and blistering particles) determine the
resistance of a material of a given composition against
corrosion [16].
3.5. Cytotoxicity tests
Fig. 9 illustrates the difference in the biocompatibility
of the tested samples. Statistical evaluation of the results
using the one-way Anova test yields a significant
difference between both populations with p ¼ 0:0255
at a significance level of 0.05. As can be seen, Neosent
seems to be better tolerated than SeNiTi. The survival
rate of the cells cultured in the presence of Neosent
amounts to approximately 9075% for each incubation
period, whereas cell cultures in the presence of SeNiTi
exhibit an apparently time-dependent biocompatibility.
In the latter case, a survival rate of about 7773% is
observed for a first incubation period of 24 h, which
raises to 8173% after 48 h and finally reaches 8573%.
Since the chemical composition is the same, the net
difference in the biocompatibility is thought to arise
from the difference in the corrosion behaviour of the
alloys investigated.
4. Discussion
The alloys investigated in this work, although
characterized by very close chemical compositions, show
completely different transformation behaviour and
2892 M. Es-Souni et al. / Biomaterials 23 (2002) 2887–2894

Fig. 7. SEM micrographs at two different magnifications of corrosion-tested SeNiTi at 371C. The pits (oxidized) follow the groove patterns.

Fig. 8. BSE micrographs of the Neosent alloy corrosion tested at 371C. The corroded areas show a dark contrast. Fig. 8b is a higher magnification of
(a). Notice the pit morphology in Fig. 8b.

Controles: regions of different dislocation densities (e.g. Fig. 5)

NC= negative controle
PC= positive controle
which might have resulted from low annealing tempera-
Neosent tures, heterogeneous grain microstructures and/or non-
100 SeNiTi
optimized cold drawing conditions may lead to the
rel.OD550/630 nm

80 splitting of the endothermic peak where the lower

60
40
20
0 on the thermomechanical processing, it is possible to
NC PC 24h 48h 72h 24h 48h
Fig. 9. Relative OD550/630 nm of formazan formation for both alloys
for different incubation periods of 24, 48 and 72 h. The figure also
shows NC and PC. The error bars denote the standard error.
microstructures. This result clearly shows that the
microstructure and the phase transformation tempera-
tures of shape memory NiTi alloys can be well
controlled by suitable thermomechanical treatments.
The complex processing steps which introduce disloca-
tion substructures in the material may often result in a
complex transformation behaviour with the appearance
of the trigonal R-phase and the shift of the monoclinic
B190 martensite transformation to lower temperatures.
Both alloys investigated show this behaviour. For
SeNiTi, the presence of an additional endothermic peak
complicates this figure still more. A plausible explana-
tion of the presence of this peak requires a closer insight
into the processing steps of the SeNiTi wire which is
unfortunately not possible. Nevertheless, the presence of
temperature peak would correspond to the regions of
higher dislocation densities (the austenite phase is
stabilized by work hardening). The transformation
behaviour and the ability to control it is of prime
interest for the practical use of NiTi alloys. Depending
72h
control the austenite finish temperature, Af ; in a wide
range and consequently the mechanical properties.
Martensitic wires, e.g. Neosentalloy, have an Af of
approximately 35–371C, and can be easily deformed at
room temperature, thus allowing a comfortable hand-
ling. They develop the thermal shape memory effect at
body temperature and exert a gentle load on the teeth
which prevents undesirable pains for the patient, and
make them the wires of choice for levelling treatments
(first orthodontic treatment stages). The austenitic wires
are characterized by a higher strength (compare Figs. 2a
and b) and cannot be deformed permanently without
destroying the superelastic effect based on the stress-
induced martensite transformation. These wires develop
higher loads on the teeth, and are used in later treatment
stages.
The surface finish is expected to strongly influence the
functionality of the wire and its lifetime. Rough surfaces
may develop non-permissible friction forces between
wire and bracket leading to limited development of the
shape memory effect and load transmission. Further-
more, as clearly shown in the present study, the
 M. Es-Souni et al. / Biomaterials 23 (2002) 2887–2894
corrosion resistance properties are badly affected by
poor surface finish. The surface imperfections and
impurities as they are present on SeNiTi are thought
to negatively affect the properties of the passive
layer thus leading to high local corrosion rates in these
areas as they are evidenced by Fig. 7. These results
are well in agreement with previous results reviewed
by Ryh.anen [17]. Trepanier et al. [18] report on the
effects of surface treatment on the corrosion properties
of NiTi stents and show that electropolished surfaces
are characterized by a higher corrosion resistance in
Hank’s solution at 371C. Similar results have been also
reported by Oshida et al. [19]. However, a direct
comparison of the pitting potentials obtained in the
present work with those of the references above is not
possible because of different alloy compositions and
testing conditions.
The in vitro biocompatibility of NiTi alloys has been
also reviewed by Ryh.anen [17]. It seems that depending
on the test conditions, contradictory results have been
reported. The general trend is, however, in favour of a
high biocompatibility of NiTi alloys. In particular, Rose
et al. [20] investigated the effects of NiTi on fibroblast
cell cultures using the MTT test and concluded that
NiTi had no effects on cell proliferation. The results
reported in the present work show that mitochondrial
activity is influenced by the corrosion properties.
Although the alloys investigated have the same compo-
sition, the higher corrosion rate associated with the bad
surface finish of SeNiTi leads to diminished in vitro
biocompatibility probably due to the release of metallic
ions. Gil and Planell [6] and Gil et al. [7] report on the
ion release of NiTi arch wires in artificial saliva and
show that the release of Ni and Ti follow a parabolic
law, and lead at the end to concentrations below the
daily take-up. Unfortunately, no indications were given
as to the surface topography and chemistry of their
wires. Wever et al. [8] also show a parabolic law for Ni
ions release with an initial release rate of 14.5
107 mg/
cm2 s which then decreases rapidly due to the build-up of
a TiO2 passive layer. For their investigations, they used
a stoichiometric NiTi binary alloy with a well-polished
surface. Based on these studies, it is thought that Ni ions
release may be higher for SeNiTi compared with
Neosent, particularly in the initial stages which then
would lead to the observed lower biocompatibility, and
the results of Fig. 9 which show a lower dehydrogenase
activity for the first 24 h incubation time points to the
validity of this assumption. However, more studies are
needed, particularly using atomic adsorption spectro-
scopy, in order to correlate the corrosion behaviour with
an eventual enrichment of the nutritive medium with
such ions. We may conclude from the present study that
more attention should be devoted to the surface finish
quality of the NiTi orthodontic products and that
stringent specifications, as they already exist for metallic
2893
implant materials [21], should also be worked out for
these products.
5. Conclusions
The present work reports a comparative study of the
properties, including corrosion and biocompatibility, of
two binary NiTi42 shape memory alloys of almost the
same composition. The following conclusions may be
inferred:
* The alloys show different transformation behaviour
and mechanical properties due to their different
thermomechanical treatments.
* The alloys are characterized by a different surface
finish. SeNiTi shows numerous longitudinal grooves
partly oxidized, while Neosent exhibits a smooth and
homogenous martensitic/austenitic surface structure
which leads finally to a better corrosion resistance.
* This study shows also a significant difference in the
in-vitro biocompatibility. The presence of the rough-
er and oxidized surface of SeNiTi and its lower
corrosion resistance are thought to induce a net
reduction of the dehydrogenase activity of the
fibroblast cell culture.
References
[1] Saburi T. Ti–Ni shape memory alloys. In: Otsuka K, Waymann
CM, editors. Shape memory materials. NY: Cambridge Uni-
versity Press, 1998. p. 49–96.
[2] Melton KN. General applications of SMA’s smart materials. In:
Otsuka K, Waymann CM, editors. Shape memory materials. NY:
Cambridge University Press, 1998. p. 220–39.
[3] Duerig TW. Present and future applications of shape memory and
superelastic materials. In: George EP, Takahashi S, Trolier-
McKinstry S, Uchino K, Wun-Fogle M, editors. Materials for
smart systems. Pittsburgh: Materials Research Society, 1995.
p. 497–506.
[4] Es-Souni M, Es-Souni M, Fischer-Brandies H. On the transfor-
mation behaviour, mechanical properties and biocompatibility of
two NiTi-based shape memory alloys: NiTi42 amd NiTi42Cu7.
Biomaterials 2001;22:2153–61.
[5] Es-Souni M, Es-Souni M, Fischer-Brandies H. Wirkung von
NiTi-Legierungen auf Epithelzell- und Fibroblastenkulturen aus
humaner Gingiva. Kieferorthop 2001;15:171–6.
[6] Gil FJ, Planell JA. Effect of copper addition on the superelastic
behaviour of NiTi shape memory alloys for orthodontic applica-
tions. J Biomed Mater Res 1999;48(5):682–8.
[7] Gil FJ, Manero JM, Planell JA. Relevant aspects in the clinical
applications of NiTi shape memory alloys. J Mater Sci: Mater
Med 1996;7:403–6.
[8] Wever DJ, Veldhuizen AG, Sanders MM, Schakenrad JM, Horn
v JR. The cytotoxic, allergic and genotoxic activity of a nickel–
titanium alloy. Biomaterials 1997;18:1115–20.
[9] Ryh.anen J, Kallioinen M, Tuukkanen J, Lehenkari P, Junila J,
Niemel.a E, Sanvik P, Serlo W. Bone modelling and cell–material
interface response induced by nickel–titanium shape memory
alloy after periosteal implantation. Biomaterials 1999;20:1309–17.
2894 M. Es-Souni et al. / Biomaterials 23 (2002) 2887–2894
[10] Wever DJ, Veldhuizen AG, de Vries J, Busscher HJ, Uges DRA,
van Horn JR. Electrochemical and surface characterization of a
nickel–titanium alloy. Biomaterials 1998;19:761–9.
[11] Es-Souni M, Fischer-Brandies H, Kock N, Bock O, R.atzke K.
Chemische Zusammensetzung, Umwandlungsverhalten und me-
chanische Biegeeigenschaften ausgew.ahlter kieferorthop.adischer
.
NiTi-Drahtbogen. IOK 2001;33:87–106.
[12] Tada H, Shibo O, Kuroshima K, Koyama M, Tsukamoto K. An
improved colorimetric assay for interleukin-2. Immunol Meth
1986;93:157–65.
[13] Ohakata I, Tamura H. The R-phase transformation in the Ti–Ni
shape memory alloy and its application. In: George EP,
Takahashi S, Trolier-McKinstry S, Uchino K, Wun-Fogle M,
editors. Materials for smart systems. Pittsburgh: Materials
Research Society, 1995. p. 345–55.
[14] Duerig TW, Melton KN. Diffuse yield drop and snap action in a
NiTi alloy. In: George EP, Takahashi S, Trolier-McKinstry S,
Uchino K, Wun-Fogle M, editors. Materials for smart systems.
Pittsburgh: Materials Research Society, 1995. p. 159–64.
[15] Orgeas L, Favier D. Stress induced martensitic transformation of
a NiTi alloy in isothermal shear, tension and compression. Acta
Mater 1998;46:5579–91.
[16] Wintermantel E, Suk-Woo Ha. Biokompatible Werkstoffe und
Bauwesen. Heidelberg: Springer, 1998. p. 172–94.
[17] Ryh.anen J. Biocompatibility Evaluation of nickel–titanium shape
memory metal alloy. PhD thesis, University of Oulu, 1999.
p. 31–43.
[18] Trepanier C, Tabrizian M, Yahia LH, Bilodeau L, Piron DL.
Effect of modification of oxide layer on NiTi stent corrosion
resistance. J Biomed Mater Res 1998;43:433–40.
[19] Oshida J, Sachdeva RC, Miyazaki S. Microanalytical chracteriza-
tion and surface modification of TiNi orthodontic archwires.
Bio-Med Mater Eng 1992;2:51–69.
[20] Rose EC, Jonas IE, Kappert HF. In vitro investigation into the
biological assessment of orthodontic wires. J Orofac Orthop
1998;59:253–64.
[21] Standard practice for surface preparation, marking of metallic
surgical implants. ASTM standard F 86–91, 1991.