- Email: [email protected]
Bending properties of superelastic and nonsuperelastic nickel-titanium orthodontic wires
Bending properties of superelastic and nonsuperelastic nickel-titanium orthodontic wires Salwa E. Khler,* William A. Brantley,** and Raymond A. Fournelle*** Mihvaukee, Wis. Cantilever bending properties were evaluated for several clinically popular sizes of three superelastic and three nonsuperelastic brands of nickel-titanium orthodontic wires in the as-received condition, and for 0.016-inch diameter wires after heat treatment at 500 ° and at 600 ° C, for 10 minutes and for 2 hours. A torque meter apparatus was used for the bending experiments, and the specimen test-span length was 1/4 inch (6 mm). In general, the bending properties were similar for the three brands of superelastic wires and for the three brands of nonsuperelastic wires. For the three brands of superelastic wires, heat treatment at 500 ° C for 10 minutes had minimal effect on the bending plots, whereas heat treatment at 500 ° C for 2 hours caused decreases in the average superelastic bending moment during deactivation; heat treatment at 600 ° C resulted in loss of superelasticity. The bending properties for the three brands of nonsuperelastic wires were only slightly affected by these heat treatments. The differences in the bending properties and heat treatment responses are attributed to the relative proportions of the austenitic and martensitic forms of nickel-titanium alloy (NiTi) in the microstructures of the wire alloys. (AM J ORTHOD DENTOFACORTHOP 1991;99:310-8.)
N i c k e l - t i t a n i u m orthodontic wires have been of considerable interest to the specialty since the introduction of the original t Nitinol alloy somewhat more than a decade ago. The very low modulus of elasticity, considerable elastic force delivery range, and high springback of this alloy provide the orthodontist with unique advantages compared with the stainless steel, cobalt-chromium-nickel, and B-titanium wires. Within the last few years many new nickel-titanium orthodontic wire alloys have been introduced, and some of these new brands possess the property of superelasticity. Superelastic behavior was first specifically noted in the orthodontic literature for the Japanese 2NiTi alloy. When the wire is subjected to tensile loading, appreciable activation and deactivation take place at nearly constant values of stress. Under bending conditions, this superelastic behavior is less evident, although there is a substantial region of nearly constant bending moment during deactivation. Although the term superFrom Marquette University. Based on a dissertation submitted by Dr. Khier in partial fulfillment of the requirements for the PhD degree at Marquette University. *Former graduate student in materials science program, College of Engineering. Now at Mansoura University, Mansoura, Egypt. **Formerly Professor and Chairman, Department of Dental Materials, School of Dentistry. Now Professor, Section of Restorative and Prosthetic Dentistry, College of Dentistry, The Ohio State University, Columbus, Ohio. ***Professor, Department of Mechanical and Industrial Engineering, College of Engineering. 811119762
310
elasticity was not explicitly used in the article introducing the Chinese 3 NiTi wire alloy, the bending test plots for Chinese and Japanese NiTi were very similar, indicating that the former alloy also showed superelastic behavior. The superelastic property of some nickel-titanium wire brands has been attributed to a phase ti'ansformation from the body-centered cubic austenitic form to the hexagonal close-packed martensitie form of NiTi when the stress reaches a certain level during activation. z Upon deactivation, the reverse-phase transformation from the martensitic to the austenitic structure takes place when the stress is decreased to an appropriate level, which is somewhat less than that required to cause the forward transformation. It is thus necessary for the proprietary wire manufacturing processes to leave the nickel-titanium alloys largely in the austenitic structure for superelastic behavior to occur, whereas the original Nitinol alloy and other nonsuperelastic nickeltitanium wires have principally a work-hardened martensitic structure. A clinically useful consequence of superelastic behavior is that variations in heat treatment by the manufacturer can result in differing stress levels to initiate the phase transformations in the same nickel-titanium wires. For Japanese NiTi,Z heat treatment at 400 ° C had no affect on the bending plots. However, heat treatment at 500 ° C for periods of 5 minutes to 2 hours caused considerable differences in the constant force levels for q
Volume99 Number4
Bending properties of superelastic and nonsuperelastic NiTi wires 311
superelastic behavior, and the superelastic property of this alloy was lost after heat treatment at 600 ° C. By using appropriate heat treatments, the manufacturer is able to offer the Japanese NiTi alloy (commercially marketed as Sentinol) in three different superelastic force ranges of light, medium, and heavy for individual wire sizes. This responsiveness to heat treatment appears to be possible for other superelastic wire brands, in principle, but vacuum or inert atmosphere conditions are required because the nickel-titanium alloys react quickly with air at elevated temperatures. This investigation compares the bending properties of several superelastic and nonsuperelastic nickeltitanium wire brands in the size range of principal clinical interest, and it investigates the effects of heat treatment on the bending plots and superelastic behavior. Extensive studies of the metallurgical structure of the alloys have been performed by means of x-ray diffraction, 4 and these results will be presented in a separate article.
Table I. Summary of orthodontic wires used in
present investigation
Wire brand [ Nitinol SE
Sentinol (medium) Ni-Ti Nitinol
Titanal
Orthonol
Size (inch) 0.016, 0.018, 0.018 × 0.025, 0.021 × 0.025 0.016, 0.018, 0.018 × 0.025, 0.0215 x 0.028 0.016; 0.018" 0.016, 0.018 0.021 0.016, 0.018 0.021 0.016, 0.018 0.021
0.018, × 0.025, × 0.025 0.018, × 0.025, x 0.025 0.018, × 0.025, x 0.025
[
Manufacture~--'~ Unitek Monrovia, Calif. GAC International Central Islip, N.Y. Ormco/Div. of Sybron Glendora, Calif. Unitek Monrovia, Calif. Lancer Orthodontics Carlsbad, Calif. Rocky Mountain Orthodontics Denver, Colo.
*Only these two round wire sizes were available at the time of this investigation.
MATERIALS AND METHODS The six brands of nickel-titanium orthodontic wires and the sizes selected for this investigation are summarized in Table I. From product information literature and private communication with the manufacturers, it was found that the Nitinol SE, Sentinol, and Ni-Ti alloys show superelastic behavior, whereas the Nitinol, Titanal, and Orthonol alloys are not superelastic. The sizes of round and rectangular wires listed in Table I encompass a variety of the most common clinical appliances. A torque meter apparatus previously used for several studies 5-s in our laboratory provided accurate and reproducible measurements of the bending moment and angular deflection. A cantilever test span of 0.25 inch (6 mm) was selected, and the bending apparatus was based on the design recommended in the original version 9 of American Dental Association specification No. 28. Two torque meters (Models 783-C-2 and 783C-10, Power Instruments, Skokie, Ill.), with ranges of 0.05 to 2 inches • ounces and 0.5 to 10 inches per ounce, were used, depending on the maximum moment levels developed by a given group of specimens. The torque meter was operated manually, and the specimens were bent at room temperature (22 ° C --- 2 ° C) to angular deflections of approximately 80 ° and then unloaded. The rectangular wire specimens were subjected to second-order activation because of greater convenience with the design 9 of the specimen-gripping fixture and the direction of bending in the horizontal plane with the apparatus. The pointers on the torque meters obscured the position of zero bending moment and the
initial ranges to 0.05 or 0.5 inch • oun.ce, and it was not possible to establish the portions of the bending deformation plots near the origin. Consequently, the graphic plots in the following section are presented as relative values of angular position or bend angle, and the horizontal axes have been shifted and labeled so that the bending curves begin approximately at the origin. These torque meters are reported by the manufacturer to be accurate to within 2%, and values of bending moment were obtained at 5 ° increments of angular position. Values of the relative angular deflection could be read to approximately the nearest 0.25 ° to 0.5 ° from a protractor mounted on the base of the test apparatus. There were three replications for each wire brand-size combination, and the three bending moment values at each increment of angular position were averaged and converted from inches • ounces to grams • millimeters in the preparation of a single bending plot. Heat treatments were performed in a dental furnace (Big Brute, K.H. Huppert, South Holland, I11.) at 500 ° and 600 ° C for 10 minutes and 2 hours on 0.016-inchdiameter segments of the six wire brands. The specimens were sealed in evacuated (approximately 0.01 to 0.1 torr) quartz capsules and placed in ceramic boats. Each capsule also contained a small piece of pure titanium that served as a "getter" of residual atmospheric gases to suppress any oxidation of the wire or incorporation of other impurities from the air. After completion of the heat treatments, the quartz capsules were
Am. J. Orthod. Dentofac. Orthop. April 1991
312 Khier, Brantley, and Fournelle Ni-Ti
NITINOL SE WIRES (As-received)
• 0.016 in. • 0.018 in.
• 0.016 in. • 0.018 in. span: 6mm * 0 . 0 1 8 x 0 . 0 2 5 in. ~,0.021 x0.025 in.
~.~
I
I
I
i
I
I
30
i
w
I
60
SENTINOL WIRES (As-received) • 0.016 in. • 0.018 in. span: 6mm • 0 . 0 1 8 x 0 . 0 2 5 in. vO.0215 x0.028 in.
o1 v
0
q
I
o} ¢'1o e-
° 0
rn
o
i
t
I 30
I
T
Angular Position
I 60
I
60
90
Fig. 3. Bending plots for as-received Ni-Ti wires.
0
E
I 30
Angular Position (deg)
Fig. 1. Bending plots for as-received Nitinol SE wires.
r-
2
0
90
Angular Position (deg)
E E. IE 2 -
span: 6mm
"~
o
1~
WIRES (As-received)
t
I
90
(deg)
Fig. 2. Bending plots for as-received Sentinol wires.
immersed in water at room temperature and broken to quench the specimens. RESULTS
The bending plots for the as-received superelastic alloys, Nitinol SE, Sentinol, and Ni-Ti, are shown in
Figs. 1 to 3, and the bending plots for the as-received nonsuperelastic alloys, Nitinol, Titanal, and Orthonol, are shown in Figs. 4 to 6. Examination of these two sets of figures reveals that the bending curves for the three superelastic alloys were similar and that the three nonsuperelastic alloys also had similar bending curves. However, it is evident that there were considerable general differences in the bending deformation behavibr for the superelastic and nonsuperelastic wires. Although it was not as apparent in the loading or activation por-; tions of the curves, the unloading portions of the bending plots for the superelastic wires generally contained a nearly horizontal region or plateau where deactivation took place at almost constant moment values. An exception was found for the two rectangular sizes, of Sentinol (Fig. 2), where the major portion of each deactivation plot had a linear region with a relatively small slope. In contrast, the activation and deactivation curves shown in Figs. 4 to 6 for the nonsuperelastic wires had much g~'eater slopes compared to the bending plots for the superelastic wires. Another distinguishing feature of superelastic and nonsuperelastic alloys was the difference in elastic springback, which follows from the differences for the residual permanent deformation after unloading. The permanent set for the 1/4-inch test span specimens was approximately 10° to 15° for the three superelastic wires, whereas approximate values of 35 ° to 40 ° permanent deformation for Nitinol and Titanal and 20 ° to 30 ° for Orthonol correspond to much less springback for nonsuperelastic wires.
Volume 99 Number 4
Bending properties of superelastic and nonsuperelastic NiTi wires 313
NITINOL WIRES (As-received) • 0 . 0 1 6 in. • 0 . 0 1 8 in. span: 6mm • 0 . 0 1 8 x 0 . 0 2 5 in. ~, 0.021 x0.025 i n . / f ~
E E E
/
3
TITANAL WIRES (As-received) in. in.
• 0 . 0 1 8 x 0 . 0 2 5 in. span: 6mm ,, 0.02 l x 0 . 0 2 5 in.
f
E E E
tO
0
3
tO
0
'T"
t'-"
• 0.016 • 0.018
2
tO
o
E
E
o
z
o 121 t.-.
f-
"o
to
I
t--
rn
m v
t
I 30
i
I
I 60
v
f
I 90
v
I
I
30
Angular Position (deg)
i
v
I
.v
v
60
I
90
Angular Position (deg)
Fig. 4. Bending plots for as-received Nitinol wires. Fig.
The maximum bending moment at 80 ° activation for the three superelastic alloys ranged from about 1200 to 2000 gm • mm as the diameter varied from 0.016 inch to rectangular specimens of either 0.021 × 0.025 inch or 0.0215 × 0.028 inch (Figs. l to 3). During deactivation, the bending moment for the region of superelasticity correspondingly ranged from about 400 to 1200 gm • ram. For the three nonsuperelastic alloys, the maximum bending moment at 80 ° activation ranged from nearly 2000 gm • mm for the 0.016-inch-diameter specimens to approximately 3800 g m . m m for the 0.021 x 0.025-inch rectangular specimens. While there were only small differences in the bending plots for the Nitinol (Fig. 4) and Titanal (Fig. 5) specimens of the same wire size, the Orthonol specimens (Fig. 6) displayed greater differences with respect to the other two nonsuperelastic alloys. The maximum bending moment delivered by the 0.021 x 0.025-inch Orthonol specimens was about 3000 gm • mm, a value considerably less than that for the corresponding Nitinol and Titanal specimens. In addition, the maximum moment for the 0.018-inch-diameter segments of Orthonol exceeded that for the 0.018 × 0.025-inch rectangular
5. Bending plots for as-receivedTitanal wires.
segments of this alloy in second-order activation; the maximum moment was greater for these rectangular segments of Nitinol and Titanal, compared to Orthonol. For the 0.016-inch-diameter specimens, the maximum bending moment was similar for the three nonsuperelastic alloys. The effects of the heat treatments on the bending properties of the 0.016-inch-diameter specimens of the three superelastic alloys ~.vere very similar, as shown in Figs. 7 to 9. There was little difference between the bending plots for these wires in the as-received condition and after heat treatment at 500 ° C for 10 minutes. Heat treatment at 500 ° C" for 2 hours resulted in decreases in both the maximum moment at 80 ° activation and the average moment for the central or superelastic portion of the deactivation curve; there was little change in the value of springback. The superelastic behavior was lost for all three alloys after the heat treatments at 600 ° C for 10 minutes or for 2 hours. The criteria for the loss of superelasticity were the disappearance of the plateau region in the deactivation curve and the decrease in springback. For the heat treatments of 10 minutes and 2 hours at 600 ° C, the permanent set for
314
Am. J. Orthod. Dentofac. Orthop. April 1991
Khier, Brantley, and Fournelle ORTHONOL
WIRES (As-received)
SENTINOL
• 0 . 0 1 6 in.
_
--
= in. s p a n : t)mm
v 0.021 x0.025
i
^
~
WIRE
500°C,
(O.016in.,Heat-treated)
lOmin
___.___ ~oooc,12omin
"~. ~:~
• 0 . 0 1 8 in. • 0.018x0.025
--e---
• --.e---
_
dOg°C. 10min . 600Oc.120min
span =~mm
2 -
3 -
z
=
-'7,
,
30 Angular
I 60
,
,
9
Position (deg)
Fig. 8. Bending plots for 0.016-inch-diameter Sentinol wires after heat treatments.
l
i
I
i
30
l
I
i
t
60
J 90
Ni-Ti W I R E (0.016in.,Heat-treated) Angular
Position (deg) E E E
Fig. 6. Bending plots for as-received Orthonol wires.
•- e - - - 5 0 0 ° C , 1 0 r a i n - - ...e- - - 5 0 0 ° C , 1 2 0 m i n • .600°C, 10rain --
span =dram
o*---- • 6 0 0 ° C , 1 2 0 r a i n
2-
to
o
E E o
T" V
2i
N I T I N O L S E W I R E (O.016in.Heat-treated) -e- -500°C, 10rain -- - .e----SO0°C,12Omin • 6O0°C, lOmin
v r"
E
span =dram
0
• -e--- - - 6 0 0 ° C , 1 2 0 r a i n .
1
. . . j / ~
e.0 e-
r-
El
E O
30
60
90
Angular Position (deg) C3~
.=_
Fig. 9. Bending plots for 0.016-inch-diameter Ni-Ti wires after heat treatments,
X3 e-.
30 Angular
60
90
Position (deg)
Fig. 7. Bending plots for 0.016-inch-diameter Nitinol SE wires after heat treatments.
Nitinol SE, Sentinol, and Ni-Ti increased to values of approximately 20-35 °. For the three nonsuperelastic alloys, these heat treatments had only minor effects on the bending plots, compared to the as-received condi-
tion, as shown in Figs. 10 to 12 for Nitinol, Titanal, and Orthonol. DISCUSSION The results of this investigation agree with the previous studies by Miura et al. 2 for Japanese NiTi and Burstone et al. 3 for Chinese NiTi. The 6 mm cantilever test spans used in the present study were chosen to be intermediate in length between the equivalent 7 mm cantilever specimens used'- to evaluate Japanese NiTi,
Volume 9 9 Number 4
Bending properties of superelastic and nonsuperelastic NiTi wires 315 NITINOL WIRE --
__
E E E E~
..-e-- -e,- • - - . ~ • -
-
T I T A N A L WIRE (0.01Sin..Heat-,eated)
(0.016in.,Heat-treated}
500°C. 10min 500°C, 120min 60O°C, 10min 60O°C, 120min
B span =6ram
E
~
500°C, 10min 500°C,120min 600°C, lOmin
• --
span =6mm
• 600°C. 120rain
E
E to
o
to
o
----e---e--•
2
2
v
e-
cO
E
E
0
O
1 t..m '13 to
e" .m e"
,
,~.7
Ixl
1:13 i
30
60
I
30
90
Angular Position (deg) Fig. 10. Bending plots for 0.016-inch-diameter Nitinolwires after heat treatments.
where three-point bending and 14 mm test spans were employed, and the 5 mm cantilever test spans used 3 to evaluate the Chinese alloy. The bending plots for the three superelastic .wire alloys in Figs. 1 to 3 have the distinctive appearance originally observed for the Chinese and Japanese alloys, and the values of bending moment and permanent deformation after unloading for 0.016-inch-diameter specimens of the superelastic wires and Nitinol were very similar to those reported by Burstone et al. 3 For the superelastic wires, short cantilevered specimen lengths are necessary to achieve adequately high bending-moment values during testing. 3 Such test spans provide the additional advantage of relevance to clinical interbracket distances; in contrast to the longer 0.5-inch and l-inch test spans used in some earlier studies '''°''' of the bending properties of Nitinol cantilevered wires. The bending curves for the superelastic alloys have pronounced nonlinear appearances, and it is not possible to define a single value of stiffness or slope of the bending plot for a particular wire. 3 The dramatic effects of decreasing the cantilever test-span length from 0.5 inch (12.5 mm) to 5 or 6 mm for the nonsuperelastic Nitinol can be seen in a comparison between the original bending plots charted by Andreasen and Morrow' for 0.018-inch-diameter wires and the results Burstone et al.3 found for 0.016-inch-diameter wires and the present study. With the shorter test spans, the bending curves for Nitinol were considerably more nonlinear,
60
9O
Angular Position (deg) Fig. 11. Bending plots for 0.016-inch-diameter'Titanal wires after heat treatments.
O R T H O N O L WIRE (O.016in.,Heat-treated) -~ - 500°C, lOmin -- - -.e--- - 500°C, 120rain __ .,..~., 600°C, 10rain
E
span =6mm
- 600°C. 120rain
E to
c> "" 2
v .i.-, c-
i
O
E O
""
I
t
I
30
-I
i
I
T
60
I
I
90
Angular Position (deg) Fig. 12. Bending plots for 0.016-inch-diameter Orthonol wires after heat treatments.
and the permanent deformation after unloading was 30 ° to 35 °, compared with about 5 ° for the 0.5-inch (12.5 mm) test spans. Under the well-defined and relatively simple loading conditions of the tension test, the existence of su-
31 6
Khier, Brantley, and Fournelle
perelastic behavior for an orthodontic wire is proved unambiguously when extensive horizontal regions of constant elastic strain occur on the loading and unloading portions of the stress-strain curve. 2 It can be seen from Figs. 1 to 3 that similar superelastic regions of nearly constant bending moment were more evident for the deactivation portions of the bending plots than during activation. For the rectangular wire sizes of two superelastic alloys, these horizontal regions or plateaus were not as evident or present on the deactivation curves, but the slopes of the bending plots were still much less than those for the-nonsuperelastic alloys in Figs. 4 to 6. The value of springback provides a complementary and probably superior criterion for assessing the presence of superelasticity in bending. For the present 6 mm cantilever test spans and approximately 80 ° activation, the as-received superelastic alloys and wire sizes had springback values ranging from about 65 ° to 70 °, whereas the as-received nonsuperelastic wire specimens had values of springback ranging from about 40 ° to 60 ° . (We obtained these data by subtracting the values of permanent set given in the previous section from the maximum activation of 80 ° before unloading.) In principle, it would be possible theoretically to derive the cantilevered bending moment-angular deflection plot from the experimental tensile stress-strain curve. '2 However, this is a formidable task for the nickeltitanium wires because of several factors: the variation in strain across the specimen cross section during bending; the occurrence of both elastic deformation and permanent deformation, where the former may include superelastie strain; and the complexity of the bending mechanics analysis for large deflections) 3 which would be particularly important with the use of short test span lengths. Because of these considerations, we anticipate problems in the determination of traditional ~ mechanical properties of elastic modulus and yield strength for the nickel-titanium wires when short cantilevered test spans are used. Following the approach of Burstone et al.,3 researchers can make quantitative measurements of the moment at yielding, the maximum h,oment developed at the greatest angular deflection, the average moment for the superelastic region of deactivation, and the permanent deformation that remains after unloading. The average unloading stiffness during deactivation is an important characteristic for the nonsuperelastic wires, and values of unloading stiffness that correspond to the different regions on the bending curves for the superelastic wires can be determined. With the exception of the moment at yielding, all of these foregoing measurements can be obtained with facility from the cantilevered bendin~ plots. Determination of the too-
Am. J. Orthod. Dentofac. Orthop. April 1991
merit value for some small amount of permanent deformation-e.g., in the range of 1° to 3°-requires additional experiments in which nominally identical test segments are subjected to increasing amounts of angular deflection and then unloaded. Thus the standardized testing procedures, which appear in ADA specification No. 32 for orthodontic wires) 4 should be modified for the nickel-titanium alloys. This specification was originally developed for longer l-inch test spans of high-stiffness alloys (stainless steel, cobalt-chromiumnickel) where the cantilevered bending plots are well behaved and very similar for loading and unloading. Our complementary research with the techniques of x-ray diffraction (XRD), differential scanning calorimetry (DSC), and optical microscopy has shown that the metallurgical structures of the as-received superelastic wires are complex. Although the XRD spectra confirm that these wires consist largely of an austenitic matrix of nickel-titanium, some martensitic NiTi is also present? The DSC thermograms ~5 provide clear evidence of an additional metastable rhombohedral structure (termed the R structure), ~6which is an intermediate phase for the austenite-to-martensite transformation in NiTi. The microstructural phases present in the superelastic wires depend on the transformation temperatures for the austenitic phase. For example, if the temperatures for the transformations from austenite to the R structure and martensite are below room temperature, the superelastic wires will have an austenitic matrix at room temperature. In contrast, the as-received nonsuperelastie wires have microstructures consisting of coldworked martensitie NiTi and R structure. 4'~5 The heat treatment temperatures and times used in this study were chosen on the basis of the previously published results reported by Miura et al. 2 for Japanese NiTi. The objectives were to ascertain whether the heattreatment responses of the other commercial superelastic alloys would be similar, as confirmed in Figs. 7 to 9, and to investigate the effects of such heat-treatment conditions for the nonsuperelastic alloys (Figs. 10 to 12). The heat treatments at 500 ° and 600 ° C result in some loss of internal stresses in the wires and cause transformatiofi of the martensitic NiTi and R phase in the microstructure to austenitic NiTi. 4 During subsequent cooling to room temperature, the austenitic NiTi may be retained or may undergo transformation. The relative proportions of austenitic NiTi, R structure, and martensitic NiTi in the wire microstructures depend on the phase-transformation temperatures, which, in turn, depend on both the heat-treatment temperature and the amount of cold work. ~7'~sA more complete description of these relatively complex metallurgical processes will be published separately.
Voh~me 99 Number 4
Bending properties of superelastic and nonsuperelastic NiTi wires 317
In closing, one must be careful not to draw unwarranted conclusions for orthodontic practice from this laboratory study o f bending properties, in particular about the relative merits o f superelastic versus nonsuperelastic wires,-which can be established only after clinical research. It may also be noted that the minor effects of the 500 ° and 600 ° C heat treatments on the bending properties o f the nonsuperelastic wires are consistent with the results reported by Mayhew and Kusy. ~9 Th~se investigators found that the mechanical properties o f Nitinol and Titanal arch wires were not affected by approved heat-sterilization methods performed at temperatures below 180 ° C. Future research in this area has been recommended ~9 for the superelastic NiTi wires.
siderable response to heat treatment. For all three superelastic brands, the maximum moment at 80 ° activation and the average superelastic moment during deactivation were decreased by heat treatment at 500 ° C for 2 hours, although there was little change in springback. H e a t treatment at 500 ° C for 10 minutes caused minimal cfianges in the bending plots for the three superelastic alloys, while heat treatment at 600 ° C for as little as 10 minutes resulted in complete loss of superelastic behavior. The differences in the bending properties and heat-treatment responses for the superelastic and nonsuperelastic wire brands are largely attributed to the relative proportions o f the austenitic and martensitic forms o f NiTi in the alloy microstructures.
CONCLUSIONS
We gratefully acknowledge the generous support of the Unitek Corporation, Lancer Orthodontics, Rocky Mountain Orthodontics, and GAC International for their extensive contributions of materials to our study. We also acknowledge the expert graphics assistance of Mary Kastem at Marquette University and Kevin Taylor at The Ohio State University in preparing the figures.
From an extensive series of cantilever bending experiments performed on several popular sizes o f six brands of nickel-titanium orthodontic wires, the following conclusions are drawn: 1. The bending deformation plots for the three brands of superelastic wires were similar, but differed substantially from the bending plots for the three brands o f nonsuperelastic wires, which were also similar. For example, the m a x i m u m bending moment at 80 ° activation for l / 4 - i n c h (6 mm) cantilevered test spans o f the superelastic wires ranged from about 1200 to 2000 gm • m m when the cross-section varied from 0.016 inch in diameter to 0.021 × 0.025 o r 0 . 0 2 1 5 × 0.028 inch. The corresponding maximum moment values for the nonsuperelastic wires were nearly twice as great, ranging from about 2000 to 3800 g m • mm. Permanent deformation o f the test spans after unloading from 80 ° activation was about 10 ° to 15 ° for the superelastic wires and at least twice as great for the nonsuperelastic wires. 2. A superelastic region o f nearly constant bending moment was observed for deactivation o f the round wires but was less evident during deactivation for the rectangular superelastic wires tested. The slopes o f the nonlinear bending plots (stiffness values) were always considerably less for the superelastic wires, compared to the nonsuperelastie wires. The present results indicate that the value o f springback in bending is an excellent criterion to distinguish between the existence or absence o f superelastic behavior for these nickeltitanium alloys. Unambiguous proof of superelastic behavior requires use o f the tension test where extensive horizontal regions appear on both the loading and unloading portions o f the stress-strain curve. 3. While heat treatment at 500 ° and 600 ° C caused only small changes in the bending plots for the nonsuperelastic wires, the superelastic wires showed con-
REFERENCES 1. Andreasen GF, Morrow RE. Laboratory and clinical analyses of nitinol wire. Ar,i J ORTHOD1978;73:142-51. 2. Miura F, M~i M, Ohura Y, Hamanaka H. The super-elastic property of the Japanese NiTi alloy wire for use in orthodontics. AM J ORTtIOODENTOFACORTHOP1986;90:1-10. 3. Burstone CJ, Qin B, Morton JY. Chinese NiTi wire--a new orthodontic alloy. AM J ORTHOD1985;87:445-52. 4. Khier SE. Structural characterization, biomechanicalproperties, and potentiodynamic polarization behavior of nickel-titanium orthodontic wire alloys [Thesisl. Milwaukee, Wisconsin: Marqueue University, 1988. 5. Hixson ME, Brantley WA, Pincsak JJ, Conover JP. Changes in bracket slot tolerance followingrecyclingof direct-bond metallic orthodontic appliances. AM J ORmOrt 1982;8h447-54. 6. Sebenc J, Brantley WA, Pincsak JJ, Conover JP. Variabilityof effective root torque as a function of edge bevel on orthodontic arch wires. A,',IJ OR'rHOD1984;86:43-51. 7. Krupp JD, Brantley WA, Gerstein H. An investigation of the torsional and bending properties of seven brands of endodontic files. J Endod 1984;10:372-80. 8. WaliaH, Brantley WA, Gbrstein H. An initial investigation of the bending and torsional properties of Nitinol root canal files. J Endod 1988;14:346-51. 9. New American Dental Association specificationNo. 28 for endodontic flies and reamers. J Am Dent Assoc 1976;93:813-8. 10. DrakeSR, Wayne DM, Powers JM, Asgar K. Mechanical properties of orthodontic wires in tension, bending and torsion. AM J OR'ntOD 1982;82:206-10. i I. Asgharnia MK, Brantley WA. Comparison of bending and tension tests for orthodontic wires. AM J ORTHOO1986;89:228-36. 12. Popov EP. Introduction to mechanics of solids. Englewood Cliffs, NJ: Prentice-Hall, 1968: chapter 6. 13. Yoshikawa DK, Burstone CJ, Goldberg AJ, Morton J. Flexure
318
14.
15.
16. 17.
18.
Am. J. Orthod. Dentofac. Orthop. April 1991
Khier, Brantley, and Fournelle
modulus of orthodontic stainless steel wires. J Dent Res 1981;60:139-45. New American Dental Association specification No. 32 for orthodontic wires not containing precious metals. J Am Dent Assoc 1977;95:1169-71. Leu LH~ Effect of annealingon the tensile properties and transformation behavior of a 50.8 at.% Ni - 49.2% at.% Tj alloy [Thesis]. Milwaukee, Wisconsin:Marquette University, 1989. Goldstein D, Kabacoff L, Tydings J. Stress effects on nitinol phase transformations. J Metals 1987;39:16-26. TodorokiT, TamuraH. Effect of heat treatmentafter cold working on the phase transformationin TiNi alloy. Trans Japan Inst Metals 1987;28:83-94. Fariabi S, Thoma PE, AbujudomDN. The effect of cold work
and heat treatment on the phase transformationsof near equiatomic NiTi shape memory alloy. Proc ICOMAT-1989. 19. MayhewMJ, Kusy RP. Effects of sterilizationon the mechanical properties and surface topographyof nickel-titaniumarch wires. AM J ORTHODDENTOFACORTHOP 1988;93:232-6.
Reprint requests to: Dr. William A~ Brantley Section of Restorative and Prosthetic Dentistry College of Dentistry The Ohio State University 305 W. 12th Ave. Columbus, OH 43210-1241
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the AMERICAN JOURNAL OF ORTHODONTICS AND DENTOFACIAL ORTHOPEDICS are available to subscribers (only) for the 1991 issues from the Publisher, at a cost of $48.00 ($59.00 international) for Vol. 99 (January-June) and Vol. 100 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, v o l u m e number, and year stamped in gold on the spine. Payment must accompany all orders. Contact M o s b y - Y e a r Book, Inc., Circulation Department, 11830 Westline Industrial Drive, St. Louis, MO 63146-3318, USA; telephone (800)325-4177, ext. 4351. S u b s c r i p t i o n s m u s t be in force to qualify. B o u n d volumes a r e n o t available in place of a r e g u l a r J o u r n a l s u b s c r i p t i o n .
N i c k e l - t i t a n i u m orthodontic wires have been of considerable interest to the specialty since the introduction of the original t Nitinol alloy somewhat more than a decade ago. The very low modulus of elasticity, considerable elastic force delivery range, and high springback of this alloy provide the orthodontist with unique advantages compared with the stainless steel, cobalt-chromium-nickel, and B-titanium wires. Within the last few years many new nickel-titanium orthodontic wire alloys have been introduced, and some of these new brands possess the property of superelasticity. Superelastic behavior was first specifically noted in the orthodontic literature for the Japanese 2NiTi alloy. When the wire is subjected to tensile loading, appreciable activation and deactivation take place at nearly constant values of stress. Under bending conditions, this superelastic behavior is less evident, although there is a substantial region of nearly constant bending moment during deactivation. Although the term superFrom Marquette University. Based on a dissertation submitted by Dr. Khier in partial fulfillment of the requirements for the PhD degree at Marquette University. *Former graduate student in materials science program, College of Engineering. Now at Mansoura University, Mansoura, Egypt. **Formerly Professor and Chairman, Department of Dental Materials, School of Dentistry. Now Professor, Section of Restorative and Prosthetic Dentistry, College of Dentistry, The Ohio State University, Columbus, Ohio. ***Professor, Department of Mechanical and Industrial Engineering, College of Engineering. 811119762
310
elasticity was not explicitly used in the article introducing the Chinese 3 NiTi wire alloy, the bending test plots for Chinese and Japanese NiTi were very similar, indicating that the former alloy also showed superelastic behavior. The superelastic property of some nickel-titanium wire brands has been attributed to a phase ti'ansformation from the body-centered cubic austenitic form to the hexagonal close-packed martensitie form of NiTi when the stress reaches a certain level during activation. z Upon deactivation, the reverse-phase transformation from the martensitic to the austenitic structure takes place when the stress is decreased to an appropriate level, which is somewhat less than that required to cause the forward transformation. It is thus necessary for the proprietary wire manufacturing processes to leave the nickel-titanium alloys largely in the austenitic structure for superelastic behavior to occur, whereas the original Nitinol alloy and other nonsuperelastic nickeltitanium wires have principally a work-hardened martensitic structure. A clinically useful consequence of superelastic behavior is that variations in heat treatment by the manufacturer can result in differing stress levels to initiate the phase transformations in the same nickel-titanium wires. For Japanese NiTi,Z heat treatment at 400 ° C had no affect on the bending plots. However, heat treatment at 500 ° C for periods of 5 minutes to 2 hours caused considerable differences in the constant force levels for q
Volume99 Number4
Bending properties of superelastic and nonsuperelastic NiTi wires 311
superelastic behavior, and the superelastic property of this alloy was lost after heat treatment at 600 ° C. By using appropriate heat treatments, the manufacturer is able to offer the Japanese NiTi alloy (commercially marketed as Sentinol) in three different superelastic force ranges of light, medium, and heavy for individual wire sizes. This responsiveness to heat treatment appears to be possible for other superelastic wire brands, in principle, but vacuum or inert atmosphere conditions are required because the nickel-titanium alloys react quickly with air at elevated temperatures. This investigation compares the bending properties of several superelastic and nonsuperelastic nickeltitanium wire brands in the size range of principal clinical interest, and it investigates the effects of heat treatment on the bending plots and superelastic behavior. Extensive studies of the metallurgical structure of the alloys have been performed by means of x-ray diffraction, 4 and these results will be presented in a separate article.
Table I. Summary of orthodontic wires used in
present investigation
Wire brand [ Nitinol SE
Sentinol (medium) Ni-Ti Nitinol
Titanal
Orthonol
Size (inch) 0.016, 0.018, 0.018 × 0.025, 0.021 × 0.025 0.016, 0.018, 0.018 × 0.025, 0.0215 x 0.028 0.016; 0.018" 0.016, 0.018 0.021 0.016, 0.018 0.021 0.016, 0.018 0.021
0.018, × 0.025, × 0.025 0.018, × 0.025, x 0.025 0.018, × 0.025, x 0.025
[
Manufacture~--'~ Unitek Monrovia, Calif. GAC International Central Islip, N.Y. Ormco/Div. of Sybron Glendora, Calif. Unitek Monrovia, Calif. Lancer Orthodontics Carlsbad, Calif. Rocky Mountain Orthodontics Denver, Colo.
*Only these two round wire sizes were available at the time of this investigation.
MATERIALS AND METHODS The six brands of nickel-titanium orthodontic wires and the sizes selected for this investigation are summarized in Table I. From product information literature and private communication with the manufacturers, it was found that the Nitinol SE, Sentinol, and Ni-Ti alloys show superelastic behavior, whereas the Nitinol, Titanal, and Orthonol alloys are not superelastic. The sizes of round and rectangular wires listed in Table I encompass a variety of the most common clinical appliances. A torque meter apparatus previously used for several studies 5-s in our laboratory provided accurate and reproducible measurements of the bending moment and angular deflection. A cantilever test span of 0.25 inch (6 mm) was selected, and the bending apparatus was based on the design recommended in the original version 9 of American Dental Association specification No. 28. Two torque meters (Models 783-C-2 and 783C-10, Power Instruments, Skokie, Ill.), with ranges of 0.05 to 2 inches • ounces and 0.5 to 10 inches per ounce, were used, depending on the maximum moment levels developed by a given group of specimens. The torque meter was operated manually, and the specimens were bent at room temperature (22 ° C --- 2 ° C) to angular deflections of approximately 80 ° and then unloaded. The rectangular wire specimens were subjected to second-order activation because of greater convenience with the design 9 of the specimen-gripping fixture and the direction of bending in the horizontal plane with the apparatus. The pointers on the torque meters obscured the position of zero bending moment and the
initial ranges to 0.05 or 0.5 inch • oun.ce, and it was not possible to establish the portions of the bending deformation plots near the origin. Consequently, the graphic plots in the following section are presented as relative values of angular position or bend angle, and the horizontal axes have been shifted and labeled so that the bending curves begin approximately at the origin. These torque meters are reported by the manufacturer to be accurate to within 2%, and values of bending moment were obtained at 5 ° increments of angular position. Values of the relative angular deflection could be read to approximately the nearest 0.25 ° to 0.5 ° from a protractor mounted on the base of the test apparatus. There were three replications for each wire brand-size combination, and the three bending moment values at each increment of angular position were averaged and converted from inches • ounces to grams • millimeters in the preparation of a single bending plot. Heat treatments were performed in a dental furnace (Big Brute, K.H. Huppert, South Holland, I11.) at 500 ° and 600 ° C for 10 minutes and 2 hours on 0.016-inchdiameter segments of the six wire brands. The specimens were sealed in evacuated (approximately 0.01 to 0.1 torr) quartz capsules and placed in ceramic boats. Each capsule also contained a small piece of pure titanium that served as a "getter" of residual atmospheric gases to suppress any oxidation of the wire or incorporation of other impurities from the air. After completion of the heat treatments, the quartz capsules were
Am. J. Orthod. Dentofac. Orthop. April 1991
312 Khier, Brantley, and Fournelle Ni-Ti
NITINOL SE WIRES (As-received)
• 0.016 in. • 0.018 in.
• 0.016 in. • 0.018 in. span: 6mm * 0 . 0 1 8 x 0 . 0 2 5 in. ~,0.021 x0.025 in.
~.~
I
I
I
i
I
I
30
i
w
I
60
SENTINOL WIRES (As-received) • 0.016 in. • 0.018 in. span: 6mm • 0 . 0 1 8 x 0 . 0 2 5 in. vO.0215 x0.028 in.
o1 v
0
q
I
o} ¢'1o e-
° 0
rn
o
i
t
I 30
I
T
Angular Position
I 60
I
60
90
Fig. 3. Bending plots for as-received Ni-Ti wires.
0
E
I 30
Angular Position (deg)
Fig. 1. Bending plots for as-received Nitinol SE wires.
r-
2
0
90
Angular Position (deg)
E E. IE 2 -
span: 6mm
"~
o
1~
WIRES (As-received)
t
I
90
(deg)
Fig. 2. Bending plots for as-received Sentinol wires.
immersed in water at room temperature and broken to quench the specimens. RESULTS
The bending plots for the as-received superelastic alloys, Nitinol SE, Sentinol, and Ni-Ti, are shown in
Figs. 1 to 3, and the bending plots for the as-received nonsuperelastic alloys, Nitinol, Titanal, and Orthonol, are shown in Figs. 4 to 6. Examination of these two sets of figures reveals that the bending curves for the three superelastic alloys were similar and that the three nonsuperelastic alloys also had similar bending curves. However, it is evident that there were considerable general differences in the bending deformation behavibr for the superelastic and nonsuperelastic wires. Although it was not as apparent in the loading or activation por-; tions of the curves, the unloading portions of the bending plots for the superelastic wires generally contained a nearly horizontal region or plateau where deactivation took place at almost constant moment values. An exception was found for the two rectangular sizes, of Sentinol (Fig. 2), where the major portion of each deactivation plot had a linear region with a relatively small slope. In contrast, the activation and deactivation curves shown in Figs. 4 to 6 for the nonsuperelastic wires had much g~'eater slopes compared to the bending plots for the superelastic wires. Another distinguishing feature of superelastic and nonsuperelastic alloys was the difference in elastic springback, which follows from the differences for the residual permanent deformation after unloading. The permanent set for the 1/4-inch test span specimens was approximately 10° to 15° for the three superelastic wires, whereas approximate values of 35 ° to 40 ° permanent deformation for Nitinol and Titanal and 20 ° to 30 ° for Orthonol correspond to much less springback for nonsuperelastic wires.
Volume 99 Number 4
Bending properties of superelastic and nonsuperelastic NiTi wires 313
NITINOL WIRES (As-received) • 0 . 0 1 6 in. • 0 . 0 1 8 in. span: 6mm • 0 . 0 1 8 x 0 . 0 2 5 in. ~, 0.021 x0.025 i n . / f ~
E E E
/
3
TITANAL WIRES (As-received) in. in.
• 0 . 0 1 8 x 0 . 0 2 5 in. span: 6mm ,, 0.02 l x 0 . 0 2 5 in.
f
E E E
tO
0
3
tO
0
'T"
t'-"
• 0.016 • 0.018
2
tO
o
E
E
o
z
o 121 t.-.
f-
"o
to
I
t--
rn
m v
t
I 30
i
I
I 60
v
f
I 90
v
I
I
30
Angular Position (deg)
i
v
I
.v
v
60
I
90
Angular Position (deg)
Fig. 4. Bending plots for as-received Nitinol wires. Fig.
The maximum bending moment at 80 ° activation for the three superelastic alloys ranged from about 1200 to 2000 gm • mm as the diameter varied from 0.016 inch to rectangular specimens of either 0.021 × 0.025 inch or 0.0215 × 0.028 inch (Figs. l to 3). During deactivation, the bending moment for the region of superelasticity correspondingly ranged from about 400 to 1200 gm • ram. For the three nonsuperelastic alloys, the maximum bending moment at 80 ° activation ranged from nearly 2000 gm • mm for the 0.016-inch-diameter specimens to approximately 3800 g m . m m for the 0.021 x 0.025-inch rectangular specimens. While there were only small differences in the bending plots for the Nitinol (Fig. 4) and Titanal (Fig. 5) specimens of the same wire size, the Orthonol specimens (Fig. 6) displayed greater differences with respect to the other two nonsuperelastic alloys. The maximum bending moment delivered by the 0.021 x 0.025-inch Orthonol specimens was about 3000 gm • mm, a value considerably less than that for the corresponding Nitinol and Titanal specimens. In addition, the maximum moment for the 0.018-inch-diameter segments of Orthonol exceeded that for the 0.018 × 0.025-inch rectangular
5. Bending plots for as-receivedTitanal wires.
segments of this alloy in second-order activation; the maximum moment was greater for these rectangular segments of Nitinol and Titanal, compared to Orthonol. For the 0.016-inch-diameter specimens, the maximum bending moment was similar for the three nonsuperelastic alloys. The effects of the heat treatments on the bending properties of the 0.016-inch-diameter specimens of the three superelastic alloys ~.vere very similar, as shown in Figs. 7 to 9. There was little difference between the bending plots for these wires in the as-received condition and after heat treatment at 500 ° C for 10 minutes. Heat treatment at 500 ° C" for 2 hours resulted in decreases in both the maximum moment at 80 ° activation and the average moment for the central or superelastic portion of the deactivation curve; there was little change in the value of springback. The superelastic behavior was lost for all three alloys after the heat treatments at 600 ° C for 10 minutes or for 2 hours. The criteria for the loss of superelasticity were the disappearance of the plateau region in the deactivation curve and the decrease in springback. For the heat treatments of 10 minutes and 2 hours at 600 ° C, the permanent set for
314
Am. J. Orthod. Dentofac. Orthop. April 1991
Khier, Brantley, and Fournelle ORTHONOL
WIRES (As-received)
SENTINOL
• 0 . 0 1 6 in.
_
--
= in. s p a n : t)mm
v 0.021 x0.025
i
^
~
WIRE
500°C,
(O.016in.,Heat-treated)
lOmin
___.___ ~oooc,12omin
"~. ~:~
• 0 . 0 1 8 in. • 0.018x0.025
--e---
• --.e---
_
dOg°C. 10min . 600Oc.120min
span =~mm
2 -
3 -
z
=
-'7,
,
30 Angular
I 60
,
,
9
Position (deg)
Fig. 8. Bending plots for 0.016-inch-diameter Sentinol wires after heat treatments.
l
i
I
i
30
l
I
i
t
60
J 90
Ni-Ti W I R E (0.016in.,Heat-treated) Angular
Position (deg) E E E
Fig. 6. Bending plots for as-received Orthonol wires.
•- e - - - 5 0 0 ° C , 1 0 r a i n - - ...e- - - 5 0 0 ° C , 1 2 0 m i n • .600°C, 10rain --
span =dram
o*---- • 6 0 0 ° C , 1 2 0 r a i n
2-
to
o
E E o
T" V
2i
N I T I N O L S E W I R E (O.016in.Heat-treated) -e- -500°C, 10rain -- - .e----SO0°C,12Omin • 6O0°C, lOmin
v r"
E
span =dram
0
• -e--- - - 6 0 0 ° C , 1 2 0 r a i n .
1
. . . j / ~
e.0 e-
r-
El
E O
30
60
90
Angular Position (deg) C3~
.=_
Fig. 9. Bending plots for 0.016-inch-diameter Ni-Ti wires after heat treatments,
X3 e-.
30 Angular
60
90
Position (deg)
Fig. 7. Bending plots for 0.016-inch-diameter Nitinol SE wires after heat treatments.
Nitinol SE, Sentinol, and Ni-Ti increased to values of approximately 20-35 °. For the three nonsuperelastic alloys, these heat treatments had only minor effects on the bending plots, compared to the as-received condi-
tion, as shown in Figs. 10 to 12 for Nitinol, Titanal, and Orthonol. DISCUSSION The results of this investigation agree with the previous studies by Miura et al. 2 for Japanese NiTi and Burstone et al. 3 for Chinese NiTi. The 6 mm cantilever test spans used in the present study were chosen to be intermediate in length between the equivalent 7 mm cantilever specimens used'- to evaluate Japanese NiTi,
Volume 9 9 Number 4
Bending properties of superelastic and nonsuperelastic NiTi wires 315 NITINOL WIRE --
__
E E E E~
..-e-- -e,- • - - . ~ • -
-
T I T A N A L WIRE (0.01Sin..Heat-,eated)
(0.016in.,Heat-treated}
500°C. 10min 500°C, 120min 60O°C, 10min 60O°C, 120min
B span =6ram
E
~
500°C, 10min 500°C,120min 600°C, lOmin
• --
span =6mm
• 600°C. 120rain
E
E to
o
to
o
----e---e--•
2
2
v
e-
cO
E
E
0
O
1 t..m '13 to
e" .m e"
,
,~.7
Ixl
1:13 i
30
60
I
30
90
Angular Position (deg) Fig. 10. Bending plots for 0.016-inch-diameter Nitinolwires after heat treatments.
where three-point bending and 14 mm test spans were employed, and the 5 mm cantilever test spans used 3 to evaluate the Chinese alloy. The bending plots for the three superelastic .wire alloys in Figs. 1 to 3 have the distinctive appearance originally observed for the Chinese and Japanese alloys, and the values of bending moment and permanent deformation after unloading for 0.016-inch-diameter specimens of the superelastic wires and Nitinol were very similar to those reported by Burstone et al. 3 For the superelastic wires, short cantilevered specimen lengths are necessary to achieve adequately high bending-moment values during testing. 3 Such test spans provide the additional advantage of relevance to clinical interbracket distances; in contrast to the longer 0.5-inch and l-inch test spans used in some earlier studies '''°''' of the bending properties of Nitinol cantilevered wires. The bending curves for the superelastic alloys have pronounced nonlinear appearances, and it is not possible to define a single value of stiffness or slope of the bending plot for a particular wire. 3 The dramatic effects of decreasing the cantilever test-span length from 0.5 inch (12.5 mm) to 5 or 6 mm for the nonsuperelastic Nitinol can be seen in a comparison between the original bending plots charted by Andreasen and Morrow' for 0.018-inch-diameter wires and the results Burstone et al.3 found for 0.016-inch-diameter wires and the present study. With the shorter test spans, the bending curves for Nitinol were considerably more nonlinear,
60
9O
Angular Position (deg) Fig. 11. Bending plots for 0.016-inch-diameter'Titanal wires after heat treatments.
O R T H O N O L WIRE (O.016in.,Heat-treated) -~ - 500°C, lOmin -- - -.e--- - 500°C, 120rain __ .,..~., 600°C, 10rain
E
span =6mm
- 600°C. 120rain
E to
c> "" 2
v .i.-, c-
i
O
E O
""
I
t
I
30
-I
i
I
T
60
I
I
90
Angular Position (deg) Fig. 12. Bending plots for 0.016-inch-diameter Orthonol wires after heat treatments.
and the permanent deformation after unloading was 30 ° to 35 °, compared with about 5 ° for the 0.5-inch (12.5 mm) test spans. Under the well-defined and relatively simple loading conditions of the tension test, the existence of su-
31 6
Khier, Brantley, and Fournelle
perelastic behavior for an orthodontic wire is proved unambiguously when extensive horizontal regions of constant elastic strain occur on the loading and unloading portions of the stress-strain curve. 2 It can be seen from Figs. 1 to 3 that similar superelastic regions of nearly constant bending moment were more evident for the deactivation portions of the bending plots than during activation. For the rectangular wire sizes of two superelastic alloys, these horizontal regions or plateaus were not as evident or present on the deactivation curves, but the slopes of the bending plots were still much less than those for the-nonsuperelastic alloys in Figs. 4 to 6. The value of springback provides a complementary and probably superior criterion for assessing the presence of superelasticity in bending. For the present 6 mm cantilever test spans and approximately 80 ° activation, the as-received superelastic alloys and wire sizes had springback values ranging from about 65 ° to 70 °, whereas the as-received nonsuperelastic wire specimens had values of springback ranging from about 40 ° to 60 ° . (We obtained these data by subtracting the values of permanent set given in the previous section from the maximum activation of 80 ° before unloading.) In principle, it would be possible theoretically to derive the cantilevered bending moment-angular deflection plot from the experimental tensile stress-strain curve. '2 However, this is a formidable task for the nickeltitanium wires because of several factors: the variation in strain across the specimen cross section during bending; the occurrence of both elastic deformation and permanent deformation, where the former may include superelastie strain; and the complexity of the bending mechanics analysis for large deflections) 3 which would be particularly important with the use of short test span lengths. Because of these considerations, we anticipate problems in the determination of traditional ~ mechanical properties of elastic modulus and yield strength for the nickel-titanium wires when short cantilevered test spans are used. Following the approach of Burstone et al.,3 researchers can make quantitative measurements of the moment at yielding, the maximum h,oment developed at the greatest angular deflection, the average moment for the superelastic region of deactivation, and the permanent deformation that remains after unloading. The average unloading stiffness during deactivation is an important characteristic for the nonsuperelastic wires, and values of unloading stiffness that correspond to the different regions on the bending curves for the superelastic wires can be determined. With the exception of the moment at yielding, all of these foregoing measurements can be obtained with facility from the cantilevered bendin~ plots. Determination of the too-
Am. J. Orthod. Dentofac. Orthop. April 1991
merit value for some small amount of permanent deformation-e.g., in the range of 1° to 3°-requires additional experiments in which nominally identical test segments are subjected to increasing amounts of angular deflection and then unloaded. Thus the standardized testing procedures, which appear in ADA specification No. 32 for orthodontic wires) 4 should be modified for the nickel-titanium alloys. This specification was originally developed for longer l-inch test spans of high-stiffness alloys (stainless steel, cobalt-chromiumnickel) where the cantilevered bending plots are well behaved and very similar for loading and unloading. Our complementary research with the techniques of x-ray diffraction (XRD), differential scanning calorimetry (DSC), and optical microscopy has shown that the metallurgical structures of the as-received superelastic wires are complex. Although the XRD spectra confirm that these wires consist largely of an austenitic matrix of nickel-titanium, some martensitic NiTi is also present? The DSC thermograms ~5 provide clear evidence of an additional metastable rhombohedral structure (termed the R structure), ~6which is an intermediate phase for the austenite-to-martensite transformation in NiTi. The microstructural phases present in the superelastic wires depend on the transformation temperatures for the austenitic phase. For example, if the temperatures for the transformations from austenite to the R structure and martensite are below room temperature, the superelastic wires will have an austenitic matrix at room temperature. In contrast, the as-received nonsuperelastie wires have microstructures consisting of coldworked martensitie NiTi and R structure. 4'~5 The heat treatment temperatures and times used in this study were chosen on the basis of the previously published results reported by Miura et al. 2 for Japanese NiTi. The objectives were to ascertain whether the heattreatment responses of the other commercial superelastic alloys would be similar, as confirmed in Figs. 7 to 9, and to investigate the effects of such heat-treatment conditions for the nonsuperelastic alloys (Figs. 10 to 12). The heat treatments at 500 ° and 600 ° C result in some loss of internal stresses in the wires and cause transformatiofi of the martensitic NiTi and R phase in the microstructure to austenitic NiTi. 4 During subsequent cooling to room temperature, the austenitic NiTi may be retained or may undergo transformation. The relative proportions of austenitic NiTi, R structure, and martensitic NiTi in the wire microstructures depend on the phase-transformation temperatures, which, in turn, depend on both the heat-treatment temperature and the amount of cold work. ~7'~sA more complete description of these relatively complex metallurgical processes will be published separately.
Voh~me 99 Number 4
Bending properties of superelastic and nonsuperelastic NiTi wires 317
In closing, one must be careful not to draw unwarranted conclusions for orthodontic practice from this laboratory study o f bending properties, in particular about the relative merits o f superelastic versus nonsuperelastic wires,-which can be established only after clinical research. It may also be noted that the minor effects of the 500 ° and 600 ° C heat treatments on the bending properties o f the nonsuperelastic wires are consistent with the results reported by Mayhew and Kusy. ~9 Th~se investigators found that the mechanical properties o f Nitinol and Titanal arch wires were not affected by approved heat-sterilization methods performed at temperatures below 180 ° C. Future research in this area has been recommended ~9 for the superelastic NiTi wires.
siderable response to heat treatment. For all three superelastic brands, the maximum moment at 80 ° activation and the average superelastic moment during deactivation were decreased by heat treatment at 500 ° C for 2 hours, although there was little change in springback. H e a t treatment at 500 ° C for 10 minutes caused minimal cfianges in the bending plots for the three superelastic alloys, while heat treatment at 600 ° C for as little as 10 minutes resulted in complete loss of superelastic behavior. The differences in the bending properties and heat-treatment responses for the superelastic and nonsuperelastic wire brands are largely attributed to the relative proportions o f the austenitic and martensitic forms o f NiTi in the alloy microstructures.
CONCLUSIONS
We gratefully acknowledge the generous support of the Unitek Corporation, Lancer Orthodontics, Rocky Mountain Orthodontics, and GAC International for their extensive contributions of materials to our study. We also acknowledge the expert graphics assistance of Mary Kastem at Marquette University and Kevin Taylor at The Ohio State University in preparing the figures.
From an extensive series of cantilever bending experiments performed on several popular sizes o f six brands of nickel-titanium orthodontic wires, the following conclusions are drawn: 1. The bending deformation plots for the three brands of superelastic wires were similar, but differed substantially from the bending plots for the three brands o f nonsuperelastic wires, which were also similar. For example, the m a x i m u m bending moment at 80 ° activation for l / 4 - i n c h (6 mm) cantilevered test spans o f the superelastic wires ranged from about 1200 to 2000 gm • m m when the cross-section varied from 0.016 inch in diameter to 0.021 × 0.025 o r 0 . 0 2 1 5 × 0.028 inch. The corresponding maximum moment values for the nonsuperelastic wires were nearly twice as great, ranging from about 2000 to 3800 g m • mm. Permanent deformation o f the test spans after unloading from 80 ° activation was about 10 ° to 15 ° for the superelastic wires and at least twice as great for the nonsuperelastic wires. 2. A superelastic region o f nearly constant bending moment was observed for deactivation o f the round wires but was less evident during deactivation for the rectangular superelastic wires tested. The slopes o f the nonlinear bending plots (stiffness values) were always considerably less for the superelastic wires, compared to the nonsuperelastie wires. The present results indicate that the value o f springback in bending is an excellent criterion to distinguish between the existence or absence o f superelastic behavior for these nickeltitanium alloys. Unambiguous proof of superelastic behavior requires use o f the tension test where extensive horizontal regions appear on both the loading and unloading portions o f the stress-strain curve. 3. While heat treatment at 500 ° and 600 ° C caused only small changes in the bending plots for the nonsuperelastic wires, the superelastic wires showed con-
REFERENCES 1. Andreasen GF, Morrow RE. Laboratory and clinical analyses of nitinol wire. Ar,i J ORTHOD1978;73:142-51. 2. Miura F, M~i M, Ohura Y, Hamanaka H. The super-elastic property of the Japanese NiTi alloy wire for use in orthodontics. AM J ORTtIOODENTOFACORTHOP1986;90:1-10. 3. Burstone CJ, Qin B, Morton JY. Chinese NiTi wire--a new orthodontic alloy. AM J ORTHOD1985;87:445-52. 4. Khier SE. Structural characterization, biomechanicalproperties, and potentiodynamic polarization behavior of nickel-titanium orthodontic wire alloys [Thesisl. Milwaukee, Wisconsin: Marqueue University, 1988. 5. Hixson ME, Brantley WA, Pincsak JJ, Conover JP. Changes in bracket slot tolerance followingrecyclingof direct-bond metallic orthodontic appliances. AM J ORmOrt 1982;8h447-54. 6. Sebenc J, Brantley WA, Pincsak JJ, Conover JP. Variabilityof effective root torque as a function of edge bevel on orthodontic arch wires. A,',IJ OR'rHOD1984;86:43-51. 7. Krupp JD, Brantley WA, Gerstein H. An investigation of the torsional and bending properties of seven brands of endodontic files. J Endod 1984;10:372-80. 8. WaliaH, Brantley WA, Gbrstein H. An initial investigation of the bending and torsional properties of Nitinol root canal files. J Endod 1988;14:346-51. 9. New American Dental Association specificationNo. 28 for endodontic flies and reamers. J Am Dent Assoc 1976;93:813-8. 10. DrakeSR, Wayne DM, Powers JM, Asgar K. Mechanical properties of orthodontic wires in tension, bending and torsion. AM J OR'ntOD 1982;82:206-10. i I. Asgharnia MK, Brantley WA. Comparison of bending and tension tests for orthodontic wires. AM J ORTHOO1986;89:228-36. 12. Popov EP. Introduction to mechanics of solids. Englewood Cliffs, NJ: Prentice-Hall, 1968: chapter 6. 13. Yoshikawa DK, Burstone CJ, Goldberg AJ, Morton J. Flexure
318
14.
15.
16. 17.
18.
Am. J. Orthod. Dentofac. Orthop. April 1991
Khier, Brantley, and Fournelle
modulus of orthodontic stainless steel wires. J Dent Res 1981;60:139-45. New American Dental Association specification No. 32 for orthodontic wires not containing precious metals. J Am Dent Assoc 1977;95:1169-71. Leu LH~ Effect of annealingon the tensile properties and transformation behavior of a 50.8 at.% Ni - 49.2% at.% Tj alloy [Thesis]. Milwaukee, Wisconsin:Marquette University, 1989. Goldstein D, Kabacoff L, Tydings J. Stress effects on nitinol phase transformations. J Metals 1987;39:16-26. TodorokiT, TamuraH. Effect of heat treatmentafter cold working on the phase transformationin TiNi alloy. Trans Japan Inst Metals 1987;28:83-94. Fariabi S, Thoma PE, AbujudomDN. The effect of cold work
and heat treatment on the phase transformationsof near equiatomic NiTi shape memory alloy. Proc ICOMAT-1989. 19. MayhewMJ, Kusy RP. Effects of sterilizationon the mechanical properties and surface topographyof nickel-titaniumarch wires. AM J ORTHODDENTOFACORTHOP 1988;93:232-6.
Reprint requests to: Dr. William A~ Brantley Section of Restorative and Prosthetic Dentistry College of Dentistry The Ohio State University 305 W. 12th Ave. Columbus, OH 43210-1241
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the AMERICAN JOURNAL OF ORTHODONTICS AND DENTOFACIAL ORTHOPEDICS are available to subscribers (only) for the 1991 issues from the Publisher, at a cost of $48.00 ($59.00 international) for Vol. 99 (January-June) and Vol. 100 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, v o l u m e number, and year stamped in gold on the spine. Payment must accompany all orders. Contact M o s b y - Y e a r Book, Inc., Circulation Department, 11830 Westline Industrial Drive, St. Louis, MO 63146-3318, USA; telephone (800)325-4177, ext. 4351. S u b s c r i p t i o n s m u s t be in force to qualify. B o u n d volumes a r e n o t available in place of a r e g u l a r J o u r n a l s u b s c r i p t i o n .