A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
981–988
CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
VERTIKALNO KONTINUIRNO LITJE NiTi ZLITINE
Ale{ Stamboli}1,2, Ivan An`el3, Gorazd Lojen3, Aleksandra Kocijan1,
Monika Jenko1,2, Rebeka Rudolf3,4
1Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia
2Jo`ef Stefan International Postgraduate School, Jamova 39, 1000 Ljubljana, Slovenia
3University of Maribor, Faculty of Mechanical Engineering, Smetanova 17, 2000 Maribor, Slovenia
4Zlatarna Celje d.d., Kersnikova 19, 3000 Celje, Slovenia
ales.stambolic@imt.si
Prejem rokopisa – received: 2016-06-17; sprejem za objavo – accepted for publication: 2016-06-27
doi:10.17222/mit.2016.111
In this paper we present research that is connected to the performance of a series of experiments combined with the
vacuum-induction melting and continuous vertical casting of a NiTi alloy in order to produce the strand. The theoretical chosen
parameters made it possible to obtain a continuously cast strand with a diameter of 11 mm. The strand microstructures were
investigated with a light and scanning electron microscope, while the chemical composition of the single phase was identified
with the semi-quantitative micro-analysis energy-dispersive X-ray spectroscopy and inductively coupled plasma – optical
emission spectrometry. The research showed that the microstructure is dendritic, where in the inter-dendritic region the eutectic
is composed of a dark NiTi phase and a bright TiNi3–x phase. In some areas we found Ti carbides and phases rich in Fe. The
micro-chemical analysis of the NiTi strand showed that the composition changed over the cross and longitudinal sections, which
is proof that the as-cast alloys are inhomogeneous. In the final part, the electrochemical behaviours of NiTi strand samples were
compared to a commercially available NiTi cast alloy with the same composition.
Keywords: NiTi alloy, continuous vertical casting, microstructure, potentiodynamic and impedance test
V tem prispevku predstavljamo raziskavo, ki je povezana z izvedbo niza preizkusov vakuumskega pretaljevanja in so~asnega
kontinuirnega vertikalnega litja NiTi zlitine s ciljem odliti palico. Teoreti~no izbrani parametri so omogo~ili, da smo uspeli
kontinuirno odliti NiTi palico s premerom 11 mm. Dobljeno mikrostrukturo palice smo raziskali s svetlobnim in vrsti~nim
elektronskim mikroskopom, kemijsko sestavo posameznih faz pa smo identificirali s semi-kvantitativno mikro-kemi~no analizo
Energijsko disperzijsko spektrometrijo in z opti~nim emisijskim spektrometrom z induktivno sklopljeno plazmo. Preiskave so
pokazale, da je mikrostruktura dendritska, medtem ko s v meddendritskem prostoru nahaja evtektik, sestavljen iz temne NiTi
faze in svetle TiNi3–x faze. Mestoma smo identificirali tudi Ti karbide in fazo bogato s Fe. Mikro-kemi~na analiza NiTi palice je
odkrila, da se sestava spreminja po prerezu in po dol`ini, kar nakazuje, da je zlitina po strjevanju nehomogena. V zaklju~nem
delu smo primerjali elektrokemijsko obna{anje vzorcev NiTi palice s komercialno dostopno valjano NiTi zlitino enake sestave.
Klju~ne besede: NiTi zlitina, vertikalno kontinuirno litje, mikrostruktura, potenciodinami~ni in impedan~ni test
1 INTRODUCTION
NiTi alloys are an attractive group that also include
nitinol. Nitinol is a group of nearly equiatomic alloys of
nickel and titanium which is located in the central region
of the NiTi phase diagram and bounded by the Ti2Ni and
TiNi3 phases.1 It exhibits a unique combination of good
functional properties and a high mechanical strength,
such as super-elasticity and a shape-memory effect, good
corrosion resistance, an unusual combination of strength
and ductility and excellent biomechanical compati-
bility.2,3 This alloy was developed in the 1970s and its
properties have enabled its use especially for biomedical
purposes, first in orthodontic treatments, and later on in
cardiovascular surgery for stents, guide wires, filters,
etc., in orthopaedic surgery for various staples and rods,
and in maxillofacial and reconstructive surgery.4 In
addition to bio-engineering, nitinol has been used in
aerospace, automotive, civil and structural engineering.5
Super-elastic NiTi is capable of recovering large inelastic
strains spontaneously upon unloading. On the other
hand, shape memory is exhibited when NiTi recovers
large strain deformation upon heating. Both the super-
elasticity and shape-memory effect are induced in nitinol
by reversible, displacive, diffusionless, solid–solid phase
transformations from a high-temperature parent phase
(austenite) with a highly ordered crystal structure to a
low temperature, stress-free martensite that has a less or-
dered structure. Nitinol is hysteretic, and there are seve-
ral transformation temperatures, including the austenite
start temperature (As), the austenite finish temperature
(Af) during heating and the martensite start temperature
(Ms) and the martensite finish temperature (Mf) during
cooling. Super-elastic behaviour will only occur if the
material is loaded above its Af temperature.6–8
The common production route for a NiTi alloy with a
shape-memory effect is known and has been experi-
mented on laboratory equipment with the technological
aspects of vacuum induction melting, hot and cold work-
ing operations. The process is still being optimized with
a particular focus on obtaining a small dimension in the
cross-section and with stabilisation of its functional
properties over its lifetime.9 Vacuum induction melting
Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988 981
UDK 621.74.047:669.24:669.295 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 50(6)981(2016)
(VIM) is often used as the first technique in the prepa-
ration of a melt. Basically, it is a typical melting tech-
nique for the production of different NiTi-based alloys.
This is appreciated particularly for NiTi alloy due to the
strong influence of the chemical composition on the
reactivity with oxygen and other elements, leading to
oxidation of the NiTi melt. In the second step, such a
prepared melt is cast, which enables pouring the molten
metal into a mould of the desired shape, and allowing it
to solidify. When the molten metal is poured into the
mould, chill crystals nucleate on the cold walls of the
mould and grow inwards. Conventional casting is a batch
process that produces large ingots requiring significant
subsequent processing. Large mechanical equipment
with high construction and operational costs is necessary
to break down most ingots. These problems can be
solved by using continuous vertical casting (CVC).10–13
With CVC the raw material is placed into a VIM furnace,
in which the material melts. After melting, the melt is,
based on gravity force, moved against the nozzle, which
adjusts the rate and direction of the melt flow. The melt
flows through the nozzle into a water-chilled mould,
where the melt is solidified, and obtains the final strand
shape.
Nitinol is often subjected to deformations or stresses
that result in some kinds of mechanical failures. Two
very important factors must be considered when using
various materials in medicine, i.e., the toxicity of the
material and the failure of material. The main problem of
NiTi alloys is the high Ni content. Ni releasing can in-
duce toxic, allergic and hypersensitive reactions or tissue
necrosis after long-term implantation. To prevent failure
and Ni release, a coating of appropriate thickness must
be formed on the NiTi surface. Titanium oxide coatings
effectively suppress the nickel ions outleaching. The niti-
nol surface is spontaneously covered by Ti dioxide
because of the gain in free energy of formation for this
oxide compared to the Ni oxides. However, the oxides
formed on the nitinol surface always contain a certain
fraction of Ni.14–19
The main goal of this work was the performance of
the series of experiments combined with vacuum-induc-
tion melting and continuous vertical casting of NiTi alloy
in order to produce the strand. This was followed by the
characterization of the obtained microstructure and
finally we compared the electrochemical behaviour bet-
ween a NiTi strand and commercially available nitinol.
2 EXPERIMENTAL PART
2.1 Continuous vertical casting of NiTi alloy
The NiTi alloy composed of 50 % of amount frac-
tions of Ni and 50 % of amount fractions of Ni was
prepared with the combination of techniques: VIM and
CVC. A clay-graphite crucible was filled up to 2/3 of its
volume due to the high metallostatic pressure (pressure
that occurs within a molten metal) with Ti pellets
(99.99 % purity) and Ni tablets (99.99 % purity). By
remelting the NiTi alloy with VIM at a temperature of
about 1450 °C a pressure lower than 10–2 mbar was
achieved in the system. The induction power during
heating was for first 10 min 10 kW, then next 10 min
20 kW and in final 5 min 30 kW, while during casting it
was between 25 and 30 kW. Continuous casting was
operating in the mid range frequency (4 kHz). In the
experiments a Cu-mould (Figure 1), a ZrO2 nozzle
stabilized with Y2O3 and an Fe starter bar were applied.
2.2 Preparing of the samples for further investigation
The samples for characterization were cut longitu-
dinally (according to the direction of casting) and across
the cross-section. For this purpose, an Accutom 50 elec-
tronic saw was used for precision cutting. The grinding
was performed with 320 grit SiC abrasive paper,
mechanical polishing with MD-Largo discs with 9-μm
diamond suspension and with peroxide grains in a
chemically aggressive suspension – OP-S (colloidal
silica). The sample was then etched with Kroll’s reagent
(3 mL HF, 6 mL HNO3 and 100 mL of distilled water).
2.3 Analytical techniques
The microstructure was investigated with a light
microscope – Microphot FXA, Nikon 3CCD-Hitachi
Camcorder HV-C20A and Thermal Field Emission SEM
JEOL JSM-6500F equipped with energy-dispersive
X-ray spectroscopy (EDS) analytical technique. Chemi-
cal analyses were performed by inductively coupled
plasma – optical emission spectrometry ICP-OES (Agi-
lent 720).
Potentiodynamic polarisation measurements and
electrochemical impedance spectrometry (EIS) have
been used to study the electrochemical behaviour of
A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
982 Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988
Figure 1: Schematic presentation of copper mould with cooling sys-
tem at the Faculty of Mechanical Engineering, Maribor, Slovenia
Slika 1: Shema bakrene kokile s hladilnim sistemom na Strojni fakul-
teti v Mariboru, Slovenija
samples. All the measurements were recorded by
BioLogic Modular Research Grade Potentiostat/Galva-
nostat/FRA Model SP-300 with an EC-Lab Software and
a three-electrode cell. In this cell, the sample was the
working electrode, saturated calomel electrode (SCE,
0,242 V vs. SHE) was used as reference electrode and
the counter electrode (CE) was a platinum net. The expe-
riment was held in simulated physiological Hank’s
solution, containing 8 g/L NaCl, 0.40 g/L KCl, 0.35 g/L
NaHCO3, 0.25 g/L NaH2PO4×2H2O, 0.06 g/L
Na2HPO4×2H2O, 0.19 g/L CaCl2×2H2O, 0.41 g/L
MgCl2×6H2O, 0.06 g/L MgSO4×7H2O and 1 g/L gluco-
se, at pH = 7.8 and 37 °C. All the chemicals were from
Merck, Darmstadt, Germany. The potentiodynamic
curves were recorded after 1 h of sample stabilisation at
the open-circuit potential (OCP), starting the measure-
ment at 250 mV vs. SCE more negative than the OCP.
The potential was then increased, using a scan rate of
1 mV s–1, until the transpassive region was reached.
Long-term open circuit potentiostatic electrochemical
impedance spectra were obtained for the investigated
samples. The impedance was measured at the OCP, with
sinus amplitude of 5 mV peak to peak and a frequency
range of 65 kHz to 1 mHz, in the sequence of directly
after immersion after 1h, 2 h, 6 h, 12 h, 24 h, 48 h, 72 h,
96 h, 120 h, 144 h, 168 h and 192 h. The impedance data
are presented in terms of Nyquist plots. For the fitting
process Zview v3.4d Scribner Associates software was
used.
3 RESULTS AND DISCUSSION
3.1 Continuous vertical casting of a NiTi alloy
The CVC of a NiTi alloy is a complex process that
requires precise process parameters. Accurate measure-
ment and regulation of temperature was very difficult
because the thermocouple was not in constant contact
with the melt due to the potential contamination of the
melt and the temperature at the crucible wall is quite
different from the actual temperature of the melt. The
frequency of induction is also very important for the
casting, as a high frequency enables the temperature to
rise and low frequency means more intensive stirring. In
this case the casting was operated at a mid-range fre-
quency of induction that does not provide adequate
mixing power, causing an undesirable chemical compo-
sition in some places of the strand. The drawing of the
strand was carried out in the sequence of pull – pause, as
this reduces the possibility of a reaction between the
alloy and the mould, as well as the porosity of the
material or the occurrence of cracks in the material. The
drawing stroke had a length of between 0 and 10 mm and
the pause lasted between 0 and 1 s. The drawing rate is
also an important factor. When the drawing is too slow,
the temperature decreases, which leads to solidification
of the alloy in the nozzle and retraction of further
drawing. This leads to fracture of the strand and the
process ends without the desired result. The strand also
breaks when the drawing rate is too fast due to the
adhesion to the mould and the weakness of the thin
solidified skin.
With CVC a strand with diameter of 11 mm was ob-
tained (Figure 2). ICP analysis for the first attempt of
CVC NiTi strand showed a constant material compo-
sition of 59.8 % of amount fractions of Ni, 38.9 % of
amount fractions of Ti and 0.3 % of amount fractions of
C, EDS analysis showed approximately 1 % of amount
fractions of Fe. Deviation from the desired value (50 %
of amount fractions of Ni) is probably caused by
complications with stirring of the melt (better mixing
takes place at a lower frequency induction, Ti is very
difficult to mix). The source of Fe could be attributed to
the Fe screw that was used as a starter bar. During fur-
ther attempts the chemical composition of the strand
varied during casting. At the beginning of drawing XRF
analysis showed that the strand was rich in nickel
(70.6 % of amount fractions of Ni; 27.1 % of amount
fractions of Ti) and with the increasing length of the
strand the nickel content decreased. Chemical compo-
sition during the fracture of the strand was 52 % of
amount fractions of Ni and 47 % of amount fractions of
Ti.
3.2 Microstructure
3.2.1 Continuously vertical cast NiTi alloy
The light microscopy of the strand cross-section
reveals the dendritic microstructure (Figure 3), where
inside the primary phase NiTi is located. This is accord-
ing to the Ni-Ti phase diagram where the first solidified
phase is NiTi. Dendrites grew in the direction from the
coldest location (from the walls of the nozzle) to the
middle of the strand. The orientation of dendrites is
random. These dendrites are arranged in the matrix of
eutectic (composed with NiTi eut + TiNi3–x). In the
microstructure there are no visible defects such as cracks
and porosity.
A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988 983
Figure 2: a) NiTi strand, produced at Faculty of Mechanical Engineer-
ing, Maribor, Slovenia and b) light microscope image of cross-section
of the strand
Slika 2: a) NiTi palica, lita na Strojni fakulteti v Mariboru, Slovenija
in b) posnetek pre~nega prereza palice, narejen s svetlobnim mikro-
skopom
The NiTi strand contains between 50 % and 60 % of
amount fractions of Ni. From the phase diagram (Figure
4) it is clear that this is a hypo-eutectic alloy (according
to the eutectic reaction at 1118 °C: L  NiTi + TiNi3).
With an ideal cooling the melt would begin to solidify in
the temperature range between 1310 °C and 1118 °C.
From the melt firstly the primary NiTi phase solidifies
that would be continuously generated and grew until the
eutectic temperature (1118 °C) is reached. At this tem-
perature, the remaining melt solidifies into a eutectic
structure composed of a NiTi phase and TiNi3 phase in
the form of lamellas.
In the real case, the cooling is non-equilibrium.
Solidifying rates are large, but the diffusion rates in the
solid state are too small to make it possible to achieve a
homogeneous solid phase. A backscattered electrons
image (Figure 5) shows a typical dendritic structure
(tree-like form) that are solidified primarily (NiTi phase).
At the eutectic temperature (1118 °C) solidifies typical
lamellar eutectic structure (NiTi + TiNi3–x) from the
residue of the melt. EDS analysis at 5 keV showed that
both the dendritic phase and the dark lamellas of
eutectic, have a composition of approximately 50 % of
amount fractions of Ni and 50 % of amount fractions of
A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
984 Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988
Figure 5: Backscattered-electron image of NiTi strand at 1000× mag-
nification
Slika 5: Posnetek povratno-sipanih elektronov NiTi palice pri 1000×
pove~avi
Figure 3: Light microscope image of NiTi strand at 100× magnifica-
tion
Slika 3: Posnetek NiTi palice na svetlobnem mikroskopu pri 100×
pove~avi
Figure 6: a) SE image of NiTi strand at 5000× magnification of area
where the TiC inclusions are present, b), c), d) and e) elemental mapp-
ing at the microstructural level by scanning electron microscopy
(SEM) with energy dispersive X-ray spectrometry (EDS) in the area
with TiC inclusions
Slika 6: a) SE-posnetek NiTi palice pri 5000× pove~avi v obmo~ju,
kjer so prisotni TiC vklju~ki, b), c), d) in e) elementna analiza na
mikrostrukturni ravni z vrsti~nim elektronskim mikroskopom (SEM) z
energijo disperzijsko rentgensko spektrometrijo (EDS) v obmo~ju s
TiC vklju~ki
Figure 4: Ni-Ti phase diagram
Slika 4: Fazni diagram Ni-Ti
Ti, while bright lamellas of eutectic have a composition
of 33 % of amount fractions of Ti and 67 % of amount
fractions of Ni.
The secondary electron (SE) image (Figure 6)
reveals in addition to the dendritic structure also the
presence of the individual inclusions. The EDS analysis
showed that the inclusions are titanium carbide (TiC).
Carbon originates from the clay-graphite crucible and
diffuses into the melt during the melting and reacts there
with the Ti. The Gibbs free energy for the formation of
TiC is very low, so the conditions for the formation of
TiC are very favourable. From the results of the EDS
analysis it appears that the carbon is located only in the
form of carbides, and there is none in the other phases.
Ni and Fe are located in the dendrites and the matrix, but
not in the carbides, while titanium is present in all the
phases. Another important fact is that, during CVC, there
was no contamination with oxygen because no dissolved
oxygen or oxides were observed in the strand. In this
manner it could be concluded that the vacuum was
appropriate.
So far several VIM + CVC experiments for the pro-
duction of NiTi strand were made. In the first attempt the
chemical composition of the strand was constant, but
incorrect. During further attempts it varied during draw-
ing in the direction of reducing the nickel content. It was
concluded that the mixing of the melt was inappropriate.
Insufficient stirring was attributed to the 4-kHz inductor.
To achieve better stirring, a low-frequency generator
should be modulated. Costs for something like that are
too high and therefore the remelting method will be fur-
ther used. CVC will be held with an in advance prepared
NiTi alloy. Instead of Fe starter bar, that probably intro-
duced Fe impurities in the alloy, a starter bar with a
Ti-tip will be applied. The vacuum by VIM was appro-
priate, because no oxygen or oxides were found in the
strand, but the crucible will also need to be modified due
to some concentration of TiC phase in the strand.
3.2.2 Commercially available NiTi alloy
The light microscope image (Figure 7a) reveals rela-
tively large grains (> 20 μm); the grain boundaries are
clearly noticeable and the grains have different shapes
and sizes.
The secondary-electron image made with SEM (Fig-
ure 7b) reveals that the commercially available NiTi
A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988 985
Figure 8: a), b) and c) Elemental mapping at the microstructural level by scanning electron microscopy (SEM) with energy-dispersive X-ray
spectrometry (EDS) of commercial NiTi alloy
Slika 8: a), b) in c) Elementna analiza komercialne NiTi zlitine na mikrostrukturni ravni z vrsti~nim elektronskim mikroskopom (SEM) z
energijsko disperzijsko rentgensko spektrometrijo (EDS)
Figure 7: a) Light microscope image of commercially available NiTi
alloy at 100× magnification and b) SE image of commercially
available NiTi alloy at 5000× magnification
Slika 7: a) Posnetek komercialno dostopne NiTi zlitine na svetlobnem
mikroskopu pri 100× pove~avi in b) SE slika komercialno dostopne
NiTi zlitine pri 5000× pove~avi
alloy consist of two phases. EDS analysis showed that
the prevailing phase is NiTi, containing 50 % of amount
fractions of Ni and 50 % of amount fractions of Ti. The
second phase is carbon rich phase (33.3 % of amount
fractions of C, 40.12 % of amount fractions of Ti,
26.59 % of amount fractions of Ni). Figure 8 shows the
distribution of elements in the individual phases.
3.3 Potentiodynamic test
Figure 9 shows the potentiodynamic curves for NiTi
strand and commercially available NiTi alloy, while
Table 1 contains the quantitative results of the measure-
ments. Corrosion potential and current, and breakdown
potential and current values were obtained by graphic
extrapolation.
The corrosion potential of the NiTi strand is 38 mV
higher than for the commercially available NiTi alloy,
which means that the passive layers spontaneously
developed on the NiTi strand are less affected by envi-
ronmental factors. The presence of a wider passivation
range was observed for the commercially available NiTi
alloy, while for the NiTi strand the passivity occurs in a
narrower range of potentials, indicating a higher ten-
dency for localized corrosion. On the surface of the
nitinol a double layer is formed. The outer layer is TiO2
and the inner layer is TiNi3. When the thickness of the
TiO2 layer increases, two phenomena play a competing
role. First, since Ni atoms are diffusing further away
from the surface, they accumulate in the region with the
lowest oxidation state (close to the oxide–metal inter-
face). Second, as TiNi3 appears as a line phase in the
Ni–Ti phase diagram, the amount of Ni in the inter-
metallic TiNi3 layer becomes saturated upon formation
of this layer. As a result it will be more energetically
favourable to form metallic particles within the TiO2
layer than increase the thickness of the intermetallic
layer.20 Breakdown of the passive film occurs as a result
of thickening of the oxide layer, leading to an increase in
the size of the nickel particles in the outer oxide layer.
These particles cause local stress, so the layer cracks,
which facilitates the progress of corrosion. The commer-
cially available NiTi alloy has higher breakdown po-
tential, meaning it will form thicker oxide layer before
the collapse. The corrosion rate of the commercially
available NiTi alloy is lower, so it is more corrosion
resistant.
3.4 Impedance test
Electrochemical impedance spectroscopy (EIS)
measurements were performed at open circuit potential
conditions in a simulated physiological fluid for 8 days.
Figure 10 shows the Nyquist impedance diagrams for
the NiTi strand and the commercially available NiTi
alloy. The analysed spectra proposed an equivalent
circuit, considering an outer titanium oxide layer with
A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
986 Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988
Figure 10: Nyquist diagrams for the NiTi strand and the commercially available NiTi alloy with corresponding fit after a) 12 h, b) 96 h, and c)
168 h of immersion
Slika 10: Nyquistovi diagrami NiTi palice in komercialno dostopne NiTi zlitine, z ustreznimi prilegajo~imi krivuljami po ~asu izpostavljenosti:
a) 12 h, b) 96 ur, in c) 168 h
Figure 9: Potentiodynamic curves for NiTi strand and commercially
available NiTi alloy
Slika 9: Potenciodinamske krivulje NiTi palice in komercialno do-
stopne NiTi zlitine
Table 1: Electrochemical parameters determined from the poten-
tiodynamic curves measured for the NiTi strand and the commercially
available NiTi alloy
Tabela 1: Elektrokemijski parametri, dolo~eni iz potenciodinamskih
krivulj, izmerjenih za NiTi palico in komercialno dostopno NiTi
zlitino
Ecorr
(mV)
Icorr
(μA)
Ebd
(mV)
Ibd
(μA)
vcorr
(mmpy)
NiTi strand -287.1 0.343 348.5 6.767 3.201·10–3
com. NiTi alloy -324.9 0.328 625.8 5.948 2.828·10–3
Ecorr – corrosion potential determined from potentiodynamic curves;
Icorr – corrosion current; Ebd – breakdown potential; Ibd – breakdown
current; and vcorr – corrosion rate
corrosion resistance R1 and an inner TiNi3 layer with
resistance R2 (Figure 11), where Rs is the resistance of
the solution. The use of a constant phase element (CPE)
was required to account for the non-ideal capacitive
response observed as a depressed semicircle when the
spectra were plotted in the corresponding Nyquist
diagrams. The CPE originates from the surface rough-
ness and inhomogeneities present in the titanium oxide
layers at the microscopic level.21
Table 2: Corrosion resistance of NiTi strand and commercially
available NiTi alloy in outer (R1) and inner (R2) oxide layer, and total
corrosion resistance Rp at certain time of immersion
Tabela 2: Korozijska odpornost NiTi palice in komercialno dostopne
NiTi zlitine v zunanji (R1) in notranji (R2) plasti oksida ter skupna
odpornost proti koroziji Rp pri dolo~enem ~asu izpostavljenosti
t/h R1com/
R2com
/
R1strand
/
R2strand
/
Rp, com
/
Rp, strand
/
1 180400 328850 10110 457060 509250 467170
2 170510 640510 10812 638590 811020 649402
12 265080 603270 11854 736850 868350 748704
24 333940 695490 12652 764400 1029430 777052
48 387740 1048800 12903 939500 1436540 952403
72 544540 1919700 28966 1206800 2464240 1235766
96 702330 3989100 9377 1470000 4691430 1479377
120 798180 5642300 5493 1500600 6440480 1506093
144 843170 6984000 25020 1511900 7827170 1536920
168 917140 8901100 96538 1526000 9818240 1622538
192 1018900 8101300 35321 1447100 9120200 1482421
As shown in Table 2, the resistances of the outer and
inner oxide layers in the commercial NiTi alloy are very
similar and very high, while the difference in resistance
between the outer and the inner layer by the NiTi strand
is very high. This means that the outer layer of the NiTi
strand has Ni particles, which are the weakest link in the
corrosion resistance of the NiTi alloy. Resistance values
in the outer layer of NiTi strand are so low (< 10000 ),
that they present no obstacle in the progress of corrosion
that can occur hazardous nickel ions outleaching from
this layer into the surrounding area. Corrosion has slower
progress in the inner TiNi3 layer.
Figure 12 represents the polarization or a totally
corrosion resistance Rp as a function of time. Rp can be
calculated according to Equation (1):
Rp = R1 + R2 (1)
as a function of time. The slope of the commercial NiTi
alloy increases rapidly with time, while the slope of
NiTi strand increases slightly with time. It is clear that
the corrosion resistance of the commercial NiTi alloy is
much greater than that of the NiTi strand at any time.
The main reasons for the poorer corrosion resistance of
the NiTi strand are a lower homogeneity and a lower
titanium content.
4 CONCLUSIONS
From this study the following conclusions can be
drawn:
• a dendritic microstructure of the NiTi strand was
formed while VIM+CVC,
• the chemical composition of the NiTi strand varied
through the cross and longitudinal sections, so the
drawing process by CVC is not optimal,
• TiC and Fe phases were identified in the NiTi strand,
• the commercially available NiTi alloy has a higher
breakdown potential than the NiTi strand, meaning it
will have thicker, more stable oxide layer before the
collapse,
• the corrosion resistance of the commercial NiTi alloy
is much greater than that of NiTi strand at any time,
• 10% deficit of titanium in NiTi strand is reflected in
poorer corrosion resistance properties,
• despite the fact that the corrosion resistance of the
NiTi strand is not sufficient, we have successfully
cast NiTi strand by VIM + CVC processes, so it is
evident that it is possible to produce such an alloy in
this way.
A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988 987
Figure 12: Rp vs time diagram for NiTi strand and commercially
available NiTi alloy
Slika 12: Diagram Rp proti ~asu NiTi palice in komercialno dostopne
NiTi zlitine
Figure 11: Equivalent circuit of two-layer model used for the
interpretation of the measured impedance spectra of NiTi alloy
Slika 11: Ekvivalentno vezje uporabljeno za razlago izmerjenih
impedan~nih spektrov NiTi zlitine na osnovi dvoslojnega modela
5 REFERENCES
1 A. Tuissi, P. Bassani, A. Mangioni , L. Toia, F. Butera, Fabrication
process and characterization of NiTi wires for actuators, SMST-
2004: Proceedings of the International Conference on Shape
Memory and Superelastic Technologies, Baden-Baden, 2004,
501–508
2 P. R. Halani, I. Kaya, Y. C. Shin, H. E. Karaca, Phase transformation
characteristics and mechanical characterization of nitinol synthesized
by laser direct deposition, Materials Science & Engineering A, 559
(2013), 836–843, doi:10.1016/j.msea.2012.09.031
3 J. Frenzel, Z. Zhang, K. Neuking, G. Eggeler, High quality vacuum
induction melting of small quantities of NiTi shape memory alloys in
graphite crucibles, Journal of Alloys and Compounds, 385 (2004)
1–2, 214–223, doi:10.1016/j.jallcom.2004.05.002
4 I. Milo{ev, B. Kapun, The corrosion resistance of Nitinol alloy in
simulated physiological solutions Part 1: The effect of surface pre-
paration, Materials Science and Engineering C, 32 (2012) 5,
1087–1096, doi:10.1016/j.msec.2011.11.007
5 C. H. Fu, M. P. Sealy, Y.B. Guo, X.T. Wei, Finite element simulation
and experimental validation of pulsed laser cutting of nitinol, Journal
of Manufacturing Processes, 19 (2015), 81–86, doi:10.1016/j.jmapro.
2015.06.005
6 B. Lin, K. Gall, H. J. Maier, R. Waldron, Structure and thermome-
chanical behavior of NiTiPt shape memory alloy wires, Acta Bioma-
terialia, 5 (2009) 1, 257–267, doi: 10.1016/j.actbio.2008.07.015
7 J. M. McNaney, V. Imbeni, Y. Jung, P. Papadopoulos, R. O. Ritchie,
An experimental study of the superelastic effect in a shape-memory
Nitinol alloy under biaxial loading, Mechanics of Materials, 35
(2003), 969–986, doi:10.1016/S0167-6636(02)00310-1
8 A. Saigal, M. Fonte, Solid, shape recovered “bulk” Nitinol: Part I –
Tension – compression asymmetry, Materials Science and Engineer-
ing A, 528 (2011) 16–17, 5536–5550, doi:10.1016/j.msea.2011.03.
060
9 K. Otsuka, X. Ren, Physical metallurgy of Ti-Ni based shape
memory alloys. Progress in Materials science, 50 (2005) 5, 511–678,
doi:10.1016/j.pmatsci.2004.10.001
10 F. M. Ashby, R. H. D. Jones, Engineering Materials 2: An Introduc-
tion to Microstructures, Processing and Design, Second Edition,
Butterworth-Heinemann, 1998
11 J. Beddoes, M. J. Bibby, Principles of Metal Manufacturing Pro-
cesses, Second Edition, Butterworth-Heinemann, 2003
12 H. Fredriksson, U. Åkerlind, Materials Processing during Casting,
First Edition, Wiley, 2006
13 M. Ferry, Direct Strip Casting of Metals and Alloys, First Edition,
Woodhead Publishing Limited and CRC Press LLC, 2006
14 M. J. Mahtabi, N. Shamsaei, M. R. Mitchell, Fatigue of Nitinol: The
state-of-the-art and ongoing challenges, Journal of the mechanical
behavior of biomedical materials, 50 (2015), 228–254, doi:10.1016/
j.jmbbm.2015.06.010
15 J. Fer~ec, R. Rudolf, Theoretical calculation of stress for the start of
stress induced martensitic phase transformation in the Shape
Memory Alloys NiTi, Anali Pazu, 3 (2013) 2, 75–78
16 T. Hu, C. Chu, L. Yin, Y. Pu, Y. Dong, C. Guo, X. Sheng, J.-CY
Chung, P.-K. Chu, In vitro biocompatibility of titanium-nickel alloy
with titanium oxide film by H202 oxidation, Trans. Nonferrous Met.
Soc. China, 17 (2007), 553–557
17 D. Vojtìch, M. Vodìrová, J. Fojt, P. Novák, T. Kubásek, Surface
structure and corrosion resistance of short-time heat-treated NiTi
shape memory alloy, Applied Surface Science, 257 (2010) 5,
1573–1582, doi:10.1016/j.apsusc.2010.08.097
18 S. A. Shabalovskaya, H. Tian, J. W. Anderegg, D. U. Schryvers, W.
U. Carroll, J. Van Humbeeck, The influence of surface oxides on the
distribution and release of nickel from Nitinol wires, Biomaterials,
30 (2009) 4, 468–477, doi:10.1016/j.biomaterials.2008.10.014
19 S. A. Shabalovskaya, G. C. Rondelli, A. L. Undisz, J. W. Anderegg,
T. D. Burleigh, M. E. Rettenmayr, The electrochemical characteris-
tics of native Nitinol surfaces, Biomaterials, 30 (2009) 22,
3662–3672, doi:10.1016/j.biomaterials.2009.03.034
20 H. Tian, D. Schryvers, D. Liu, Q. Jiang, J. Van Humbeeck, Stability
of Ni in nitinol oxide surfaces, Acta Biomaterialia, 7 (2011) 2,
892–899, doi:10.1016/j.actbio.2010.09.009
21 J. Izquierdo, M. B. González-Marrero, M. Bozorg, B. M. Fer-
nández-Pérez, H. C. Vasconcelos, J. J. Santana, R. M. Souto, Multi-
scale electrochemical analysis of the corrosion of titanium and
nitinol for implant applications, Electrochimica Acta, 203 (2016),
366–378, doi:10.1016/j.electacta.2016.01.146
A. STAMBOLI] et al.: CONTINUOUS VERTICAL CASTING OF A NiTi ALLOY
988 Materiali in tehnologije / Materials and technology 50 (2016) 6, 981–988