Three-point bending tests were conducted in the Orthodontic Measurement and Simulation System (OMSS). This device comprises two motor-driven positioning tables and two force/moment sensors capable of three-dimensionally registering forces and moments. The sensors are connected to a personal computer with a custom-developed control program to record the movements and simultaneously measure the multidimensional force/deflection (N/mm) or moment/degree curves (Nmm/degree). The measuring ranges amount to ± 15.00 N for forces and ± 450 Nmm for moments. The OMSS is installed in a temperature-controlled chamber during the measurement cycles for a constant temperature of 37 ± 1°C (VEM 03/400; Vötsch Heraeus, Hanau, Germany).²⁵

The anterior (midline region) and the posterior (molar region) segments of the wires were cut off in 20 mm pieces and prepared for the three-point bending test. According to DIN EN ISO 15841,²⁶ the distance between the support points was 10 mm.²⁶ The wire pieces were deflected to a maximum of 3.3 mm by a thrust die, adjusted to the center of the support points (Figure 1). Care was taken to avoid slipping/flipping. This cycle was run twice with each of the 24 wire pieces and the mean value for each specimen was considered for statistical analysis. Wires were oriented edgewise (the smaller dimension was the height of the wire in the direction of displacement). The resulting force/deflection diagram was recorded by the OMSS. The forces on the unloading part of the force/deflection curve were registered and analyzed at deflections of 3.1 mm, 3.0 mm, 2.0 mm, and 1.0 mm.

Bending moduli (Eb) were calculated from the linear part of the three-point bending force-deflection curves with the formula:

where, ΔF and Δd are the differences in force and deflection, respectively, between the unloading points; L is the support span (10 mm); w is the width (in mm); and h is the archwire height (in mm). Deflection loads were obtained from the data at the linear part of the unloading curve between 0.1–0.5 mm.

The critical transformation temperature points of the austenite start (As) and Af were determined with BFR tests. The specimens were cooled in a custom-made, double-walled glass container “to its nominally fully martensitic phase”, deformed and placed on two support pins, and heated back (F12-ED; Julabo, Seelbach, Germany) to their fully austenitic phase (Figure 2).²⁷ The bath of the cold mix (Centro Kühlerschutz, Centro, Karlsruhe, Germany, > 50% ethanediol) was cooled down to a minimum of –20°C and the wire piece remained in the bath for at least 3 minutes prior to testing. Subsequently, the measurement process was started by initializing the heating of the thermostat. During the gradual temperature increase, specimen displacement was monitored by a laser triangulation sensor (RF603; Riftek Sensors & Instruments, Minsk, Belarus) and plotted versus its temperature. Accordingly, the As and Af temperatures were determined from the temperature−displacement graph. The data acquired during heating and recovery (displacement and temperature) were saved as text files and imported into a Microsoft^® Excel (Microsoft, Redmond, WA, USA) spreadsheet for plotting. A temperature/displacement graph was created for each wire piece using the spreadsheet tools to draw tangent lines to the linear portions of the curve.

The surface morphology and elemental composition of new specimens were studied with a scanning electron microscope (SEM, Quanta 200; FEI, Hillsboro, OR, USA) and by EDS microanalysis employing a spectrometer (XFlash®6–10 Silicon Drift Detector and a slew window; Bruker, Berlin, Germany) under high vacuum, 20 kV accelerating voltage, 108 μA beam current, area scan mode (250 μm × 220 μm), 200 seconds acquisition time, and 1% dead time. Three measurements were taken and the results were averaged. Quantitative analysis of collected spectra was carried out by the dedicated software (ESPRIT ver.1.9; Bruker) employing atomic number, absorbance, and fluorescence correction factors.

The sample size was calculated based on a recent in vitro experiment, which measured the differences in the force magnitude between the posterior and the anterior segments of similar wires.¹⁹ The results showed that a sample of four pairs of measurements would be required for the test to detect the resulting difference with a power of 80%. However, the present study measured a total number of 12 retrieved and new wires. This sample size was additionally assessed using the software GPower 3.1 for Windows (https://www.psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und-arbeitspsychologie/gpower), which determined a power of over 95%.

Statistical analysis was performed using STATA 13.1 (StataCorp LLC, College Station, TX, USA). To investigate possible differences between the measurements by the type of wire (new or retrieved) and segment (posterior or anterior), a multiple linear regression model with a random intercept at the wire level was applied, using the type of wire and location of measurements as covariates. Interactions between the two covariates were also tested; the model with the interactions was selected when the interaction term was statistically significant. The level of statistical significance was set to 0.05.

Forces recorded from the posterior segment were on an average larger than from the anterior segment, regardless of activation; the forces were larger by 0.73 N (95% confidence interval [CI]: 0.67, 0.79) at 1 mm, 0.80 N (95% CI: 0.77, 0.84) at 2 mm, and 1.78 N (95% CI: 1.69, 1.86) at 3 mm, on new wires and by 1.19 N (95% CI: 0.99, 1.40) on retrieved wires; at 3.1 mm the forces were larger by 1.87 N (95% CI: 1.78, 1.96) larger on new wires and 1.32 N (95% CI: 1.11, 1.53) on retrieved wires (Figure 3, Tables 1 and 2).

No statistically significant differences were found between new and retrieved specimens for the unloading points of 1 mm and 2 mm. Contrarily, at 3 mm and 3.1 mm, statistically significant differences were demonstrated between retrieved and new wires (Tables 1 and 3). These differences depend on the location of measurement according to results from models with interaction. At 3 mm, the measurement on the anterior segment of the retrieved wires was on an average 0.18 N less than on the new wires and this difference ranges between 0.07–0.29 N (95% CI); on the posterior segment, the measurement of the retrieved wires was on an average 0.76 N less than on the new wires (95% CI: 0.53, 0.99). Results of similar magnitude and direction were obtained at 3.1 mm (Tables 1 and 3).

The Eb of the retrieved wires were lower by 2.67 (95% CI: 1.22, 4.11) compared with the Eb of new wires in their anterior segment, and by 0.67 (95% CI: 0.94, 3.83) in the posterior segment (Tables 1 and 3). The Eb of the posterior segments was on an average 2.17 (95% CI: 0.95, 3.38) larger than on the anterior segments in new wires, while the respective difference on the retrieved wires was 4.17 (95% CI: 1.23, 3.66) new wires (Tables 1 and 2).

The transformation temperature As obtained from the anterior segments was on average 3.86°C (95% CI: 2.88, 4.83) higher than from the posterior segments. Af in the anterior segments was on average 6.83°C (95% CI: 5.92, 7.73) higher than the posterior Af (Table 4).

No statistically significant difference was found in As and Af temperatures between retrieved and new wires, adjusting for the location of the measurements. Figure 4 depicts representative temperature−displacement graphs of anterior and posterior wire segments.

Backscattered electron image from the anterior segment of a new specimen without mean atomic number, implied a homogeneous structure (Figure 5). In the EDS spectrum, only Ni and Ti were detected (Figure 6). The mean elemental composition was determined as follows (percent weight): Ni = 52.6 ± 0.5 and Ti = 47.4 ± 0.5, which is a typical NiTi formulation.

The load-deflection properties of the orthodontic archwires are important parameters that determine the biological nature of tooth movement. The classic three-point bending test is a suitable method for determining these properties²⁸ and is used for distinguishing superelastic wires.²⁹ In all subgroups of wires, higher force magnitudes and Eb were observed in the posterior segments compared with the anterior counterparts (150–250% higher, depending on the deactivation point). A recent study demonstrated the existence of multiple force and torque zones in multizone NiTi wires.¹⁹ Similarly, another study reported that Bioforce archwires exert lower forces for a given amount of deflection in the anterior segment when compared to conventional NiTi archwires. These differences may be attributed to differences in wire composition or cold-working process during manufacturing.³⁰

There is still no consensus in orthodontic literature on the optimal force to be applied during orthodontic tooth movement. A recent systematic review concluded that a force between 50 cN and 100 cN seems optimal.³¹ Different forces have been suggested to bring about the different types of orthodontic movements, such as 0.10–0.20 N for tooth intrusion and 0.35–0.60 N for tooth inclination, extrusion, or rotation.³² However, the recorded force magnitudes in the present experimental setup were higher, even at 2 mm deflections, and may lie beyond biologically safe limits. This finding was expected, since a rectangular NiTi archwire exerts high force magnitudes even with small deflections.³³

The present configuration of the three-point bending test is similar to a previous investigation facilitating direct comparison of results.¹⁹ They reported similar force levels for other multiforce NiTi products of the same cross section and lower levels for multistrand products without force zones. Interestingly, a 0.017 × 0.025 nine-strand NiTi archwire presented biologically safer forces, despite its large cross section. However, the authors did not recommend the use of the tested multizone wires for tooth displacements more than 1.5 mm.¹⁹ Another study demonstrated a force difference between the posterior and anterior segments of multizone archwires to be at least 45%.³⁴

Additionally, the present investigation compared new and retrieved wire pieces. The latter presented lower forces in all deactivation points, except for the last stages of deactivation. Specifically, the mean difference in force magnitude between new and retrieved archwires at the anterior segment and at the 3 mm unloading point was 0.18 N (7% of the force produced). The respective difference for the posterior segment was 0.76 N (17% of the force exerted by a new archwire).

A recent study demonstrated higher forces in new, non-multizone 0.016 NiTi archwires during the initial stages of deactivation, compared with retrieved specimens.³⁵ This is in agreement with the present findings. Although this phenomenon is of doubtful clinical significance, it is an interesting future direction for research, since the cause remains obscure. Furthermore, studies have reported alterations in the mechanical and thermal properties of NiTi after clinical use. The hysteresis energy of NiTi gradually diminished after clinical use and X-ray diffraction analysis has shown irreversible austenite to martensite transformations.³⁶

For the first time, the experiment results portray an intricate interplay between the mechanical and thermal properties of the Bioforce archwires. The BFR test is suitable for detecting the transformation temperatures of heat-activated NiTi archwires.³⁷ The present study detected statistically significant differences in the As and Af transformation temperatures between the anterior and the posterior segments of Bioforce archwires in new and retrieved specimens. However, aging did not significantly affect the transformation temperatures. A recent study that evaluated several heat-activated non-multizone NiTi archwires using BFR and DSC, reported differences in BFR As/Af between archwires of different cross sections even from the same manufacturer.³⁷ A further study conducted DSC for several rectangular multiforce NiTi wires, including Bioforce archwires. The recorded As temperatures ranged between 7.4–21.4°C in the anterior and 1.8–16.8°C in the posterior segments. Bioforce archwires demonstrated the highest As from the evaluated archwires. Af ranged between 22.7–38°C (30.2°C for Bioforce) in the anterior and 17.9–27.9°C (25.9°C for Bioforce) in the posterior segments.³⁸ These temperatures were lower than those found in the present experiment, especially for the anterior segments of the tested archwires, which may be attributed to the inherent differences between the two techniques or the different archwire dimensions of the specimens used in the two protocols. Another possible explanation could involve the applied strain level. Studies that compare BFR and DSC show differences that may range between 4–6°C based on the testing methods.³⁷ Similar results were demonstrated as long as the strain was ≤ 2.5% in the direction of the applied force;^39-41 however, the transformation temperatures shifted to higher levels on the increasing strain,⁴² as in the present simulation model.

To obtain full advantage of the superelastic properties of the NiTi archwires, the TTR should be below the usual resting intraoral temperature (34°C)⁴³ or even lower, since intraoral temperature fluctuates. Superelasticity is expected when the temperature rises above As and is fully expressed after the complete transformation from martensite to the austenite phase (Af).⁸ According to the present results, the Af temperature of these archwires during clinical use may lie higher than the average intraoral temperature, i.e., the wire segment ligated into the slots may not be fully austenitic after an increased activation. This is even more likely in the anterior wire sections, which showed increased Af. The clinical implication of this finding is related to the exerted force level that may fluctuate on consuming hot or cold beverages. In the BFR thermograms in the present experiment, the slope of the posterior segments was steeper compared with the slope of the anterior segments, in all subgroups, indicating a more rapid transformation during temperature fluctuations in this wire segment.

However, the transformation processes between austenite and martensite in NiTi orthodontic archwires is rather a complex mechanism.² The R-phase appears to be an intermediate phase in the martensite–austenite transformation path in some NiTi products or in both the martensite–austenite and austenite–martensite paths in others.⁴⁴ The austenitic transformations involving the R-phase begin below 0°C in superelastic NiTi alloys, but the nonsuperelastic NiTi is almost entirely martensite at room temperature and contains small amounts of austenite in the oral environment.² Butler et al.⁴⁰ did not report As values with DSC and stated that with the exact temperature at which austenite occurs is unclear with this method, due to overlapping between the completion of the R-phase and the onset of the austenitic phase.

The present experiment identified no differences in austenite transformation temperatures between new and retrieved specimens. Nevertheless, some irreversible transformation of austenite to martensite after the clinical use of NiTi archwires was demonstrated using X-ray diffraction analysis.³⁶ Another recent study evaluated the thermal behavior of intraorally exposed and as-received copper-nickel-titanium wires with DSC. No difference was reported for all parameters (heating onset, endset, and enthalpy and cooling onset, endset, and enthalpy), with the exception of the retrieved 27°C wires that exhibited a significantly decreased heating enthalpy associated with the martensite to austenite transition. This indicates a lessening of the extent of phase transformation; however, the clinical significance of this reduction is unclear.⁴⁵

Backscattered electron images showed the presence of elongated pores, a typical finding on orthodontic wire surface.^46-48 This pattern is appended to extensive cold working during wiring.⁴⁹ According to the present findings, the elemental composition (weight percentage) of the new Bioforce archwires was Ni, 52.6 ± 0.5 and Ti, 47.4 ± 0.5. Therefore, Bioforce archwires belong to the 55-Nitinol alloy family. These alloys present a higher shape memory effect and are comparatively easily processed by all forms of hot working than the 60-Nitinol forms.⁵⁰

Even though the present study strengthens the evidence of the behavior of new and retrieved NiTi archwires, it has some limitations. Sectioned wires were used for testing. This cold working of the material may have affected the transformation temperature.²⁷ Undoubtedly, differences exist in loading conditions between the oral cavity and the laboratory setup. Although highly reproducible, the three-point bend test is a laboratory simulation. The activation of the NiTi archwire, after its engagement into the bracket slot, is quite different than three-point bending tests in free NiTi wire segments. Moreover, the clinical extrapolation of the present findings on roughness is questionable. In vitro experiments may fail to reproduce intraoral factors that influence resistance to sliding, such as saliva, plaque, food debris, corrosion, temperature, and mastication.⁵¹