Composite wires and bracket materials

The composite wires were made from S2-glass®, continuous-fiber yarns (493 S2 CG150 1/0 1.0Z, Owens Corning Corp, Toledo, Ohio) and a glassy copolymer matrix. Each yarn was composed of 204 glass filaments (9 μm diameter) and was characterized by 1 axial rotation (twist) per inch of yarn length. The comonomer used for the matrix contained 61% (wt/wt) bisphenol-A diglycidyl methacrylate (Nupol 046-4005, Cook Composites and Polymers Co, North Kansas City, Mo) and 39% (wt/wt) triethylene glycol dimethacrylate (TEGDMA, Polysciences Inc, Warrington, Pa). Benzoin ethyl ether (BEE, Aldrich Chemical Co Inc, Milwaukee, Wis) was added (0.4% (wt/wt)) as the ultraviolet light–sensitive initiator for polymerization. Straight wire specimens, having round cross-sectional profiles, were formed using a photo-pultrusion manufacturing technique.9,10 The stiffnesses of the wires were varied by changing the volume fraction of reinforcement (V_f) that was pultruded into the composite profiles. Wires having 3 different V_f values (Table 1) were fabricated with nominal cross-sectional diameters of 0.020 inches (0.51 mm) each. The V_f values were calculated using the number of reinforcement yarns and the cross-sectional areas of the precoated wires, the latter of which were determined from the mean of 8 diameter measurements (Sony μ-mate Digital Micrometer, Sony Magnescale America Inc, Orange, Calif).

TABLE 1.  Archwire and Bracket Materials

View Fullscreen

The stainless steel (SS), polycrystalline alumina (PCA), and single-crystal alumina (SC) brackets each had a 0.022 in. (0.56 mm) slot width and no preangulation (Table 1). Two of the brackets (SS and PCA) had −7° of pretorque and the other bracket (SC) had 0° of pretorque.

Surface coating

Prior to coating, all archwire materials were washed with 95% ethanol to remove surface debris and air-dried in a class 100 clean room. Specimens were cut from the bulk material in 75 cm lengths and mounted in a rack that maintained a minimum separation of 5 mm between wires. The rack was sealed in a plastic wrap prior to removal from the clean room and shipped to an outside provider to be coated with a 10 μm layer of poly(chloro-p-xylylene) (Parylene C—Specialty Coating Systems Inc, Indianapolis, Ind). After coating, all wire specimens were stored in the dark until testing to ensure that no ultraviolet degradation would occur.

The Parylene coating was applied using a 3-stage deposition procedure.28 In the first stage, Parylene dimer was heated to approximately 150°C, which caused controlled sublimation to occur. The dimer vapor was transported by a vacuum pump to a second chamber and heated to approximately 680°C. At this temperature, the dimer broke down into Parylene monomer. The monomer was then moved into a deposition chamber that was kept at room temperature (25°C). Here, the monomer polymerized upon contact with any exposed surface, thereby coating the wire specimens.

Frictional testing

Frictional characteristics were measured using a device that was designed to simulate orthodontic sliding mechanics (Figure 1). This device and the associated testing procedure have been described elsewhere7,27 and are only briefly mentioned here. Each archwire-bracket combination (couple) was tested by mounting the bracket in the frictional testing device and connecting the wire to the load cell (50 kg scale) of a mechanical testing machine (Instron Model TTCM, Instron Corp, Canton, Mass). A computer-controlled normal force of ligation (N_FR) was applied through two 0.010 inch (0.25 mm) SS wires (Item PL 1010 Ligature Wire, GAC International, Commack, NY) to retain the archwire in the bracket slot (Figure 1c). (In previous work, the normal force of ligation was referred to as the “normal force” and was designated by N. The “FR” subscript has been added to this and other parameters to delineate between those that are associated with classical friction (FR) and those that are associated with binding (BI). See the first section of the Discussion.)

View Figure

• Download as Powerpoint
• Download Figure

Bracket angulation (θ) was obtained by rotating the bracket and ligature wires with respect to the archwire (Figure 2). The transverse beam of the mechanical testing machine, to which the device was mounted, created a relative motion between the archwire and bracket.

View Figure

• Download as Powerpoint
• Download Figure

In order to validate the frictional testing apparatus and its operator, a couple was tested and compared with prior results. Using the procedure that follows, an SS archwire (Standard Edgewise 0.021 inch × 0.025 inch [0.53 mm × 0.64 mm], American Orthodontics, Sheboygan, Wis) was tested at θ = 0° and N_FR = 0.2, 0.4, 0.6, 0.8, and 0.95 kg against a SS bracket. (Although SI units of force are conventionally expressed in newtons (N), grams (g) or kilograms (kg) are used almost exclusively in the orthodontic literature. Since 1 kg is approximately equal to 10 N, all measurements of force reported in this study can be thought of equivalently as either kg or dekanewtons (daN)). Analysis of the data, as described later, gave a kinetic coefficient of friction (μ_k-FR) equal to 0.12, which was comparable to the results from previous frictional studies.27,29

Prior to testing, all materials were rinsed in 95% ethanol to remove surface debris. Specimens of each composite archwire were tested against each of the brackets (Table 1) in the dry state at 34°C. When necessary to ensure that N_FR was applied normal to the floor of the bracket slot, brackets were mounted in the frictional testing apparatus on 7° inclines. Drawing forces (P) were measured with the Instron load cell as each bracket was translated a distance (δ) of 5 mm along an archwire at a sliding velocity of 10 mm/min. Each couple was sequentially tested at 12 different N_FR values (N_FR = 0.05, 0.10, 0.15, 0.20, 0.25, 0.30, 0.075, 0.125, 0.175, 0.225, 0.275, and 0.40 kg) and 6 different θ values (θ = 0, 2.5, 5.0, 7.5, 10.0, and 12.5°) for a total of 648 individual tests. Virginal sections of archwire were used for each test, and new brackets and ligature wires were used for each unique value of θ.

Frictional data were analyzed using techniques described in the literature.7,27,29,30 The P data in the kinetic region of each test (Figure 3) were averaged and divided by 2 to obtain the resistances to sliding (RS) of the 9 archwire-bracket couples at each N_FR and θ. These data were plotted as a function of N_FR, and linear-regression coefficients were calculated for each RS-N_FR plot in the passive or active configurations (Figure 4). The μ_k-FR value for each couple at each θ was given by the slope of the regression line, and the portion of RS that was due to binding (BI) was given by the y-intercept.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Microscopic examination

After testing, selected archwire specimens were examined with a scanning electron microscope (SEM, Model JSM-6300FV, JEOL USA, Peabody, Mass) at 15 keV for any surface effects that were attributable to sliding. In addition, all tested wire specimens were examined under a stereo microscope (Spencer Stereoscopic Microscope, American Optical Company, Buffalo, NY) to evaluate the post test integrity of the Parylene C coating. This evaluation was made for each specimen on a pass or fail basis. Specimens that had no penetrations through the coating that exposed the underlying composite substrate were assigned a value of 1, and specimens with penetrations were assigned a value of 0.

Statistical analysis

All regression lines were evaluated using the probabilities (P) associated with the correlation coefficients (r) and the number of data points (n). The μ_k-FR and BI data were evaluated to indicate differences with respect to bracket material, V_f, and θ using multiple analysis of variance (MANOVA; SYSTAT Version 5, SYSTAT Inc, Evanston, Ill). Significant factors and interactions that were revealed using MANOVA were also examined using Tukey-Kramer pairwise comparisons. Contingency tables of the coating-integrity data were created for all combinations of each of the test factors: bracket material, V_f, N_FR, and θ. Independence was established using the Pearson chi-square statistic.31

Friction

The P-δ traces were often characterized by large P variations in the kinetic data region, especially as N_FR (Figure 3a) and θ (Figure 3b) were increased. However, several specimens tested at θ < 5° had lower and sometimes less varied P magnitudes (Figure 3c). RS values were progressively less scattered with decreasing N_FR and θ (Figure 4a). As θ decreased, increases in the r values of the RS-N_FR regression lines were observed (Table 2). The slopes of these regression lines, which defined μ_k-FR, varied from 0.10 to 0.65. The y-intercepts, which defined BI, varied from −0.04 kg to 0.24 kg. Increases in θ corresponded to proportional increases in BI (Table 2; Figure 4). All but 2 of the RS-N_FR correlations were statistically significant (P < .05), and most were highly significant (P < .001, Table 2).

TABLE 2.  Regression Analysis Summary From the Resistance to Sliding (RS) Versus Normal Force of Ligation (N_FR) Data of the Coated Composite Archwires^a

View Fullscreen

Microscopy

Microscopic examinations of the tested archwire specimens revealed abrasive wear patterns in areas that had contacted the bracket or the ligature wires. Many of the specimens were characterized by complete delamination of the Parylene C coating at the points of contact (Figure 5a). However, others exhibited discrete penetrations that were separated by sections of undamaged coating (Figure 5b), and several specimens were undamaged. These later specimens were the same wires that had been observed to have lower and less varied P magnitudes during frictional testing (Figure 3c). The severity of the coating damage increased with increasing N_FR and θ, and wear was greater at the wire-bracket interface than at the wire-ligature interface. No damage to the composite substrate, ie, no release of reinforcement fibers (Figure 5c), was observed on any of the Parylene C–coated wire specimens.

View Figure

• Download as Powerpoint
• Download Figure

Archwire-bracket configuration

In studies of archwire-bracket sliding mechanics where θ is an independent variable, the data can usually be better examined by partitioning RS into several components. Looking first at the passive configuration, which has been defined as the condition where θ is less than some critical angle for binding (θ_C),27,30,32,33,34 RS results only from sliding contact between the wire and the floor and/or a side of the bracket slot (Figure 2a). In this case, RS is exclusively caused by classical friction (FR), which increases proportionally with N_FR at a rate described by μ_k-FR (Figure 4b). When θ equals or exceeds θ_C, however, RS is complicated by the addition of BI, which results from sliding contact between the wire and both sides of the bracket slot (Figure 2b). In this active configuration, the magnitude of BI increases proportionally with the normal force of binding (N_BI), which increases with θ. Ultimately, at higher values of θ, the wire is permanently deformed, and relative sliding motion may cease as notching (NO) occurs.35,36 Since NO represents an undesirable clinical condition, RS should generally be described in terms of FR and BI.

Statistical analyses

Statistical analyses of μ_k-FR revealed no significant differences with respect to bracket or V_f. Among all specimens tested, a highly significant (P < .001, n = 54) difference was only observed for θ = 12.5°. Since this represented an extreme value of θ, at which evidence of NO behavior was observed, the μ_k-FR values for θ = 12.5° were omitted. The remaining data (Table 2) were grouped (n = 45) to obtain an overall mean μ_k-FR of 0.43 ± 0.10 for the Parylene C–coated composite (CC) wire couples.

Statistical analysis of the BI data revealed a highly significant (P < .001, n = 54) interaction between θ and V_f. Pairwise comparisons showed highly significant (P < .001, n = 54) differences for all values of θ tested except between values of 0° and 2.5° where BI values were consistently near zero (Table 2). This corresponded to the passive configuration at approximately θ ≤ 2.5°. Differences in BI were also observed with respect to V_f for the C49 wires (P < .001). Since no other differences were observed, and since previous work with uncoated composite (UC) wires had shown that BI was affected by both θ and V_f,7 the data were grouped accordingly. The BI data were plotted as a function of θ for θ ≥ 2.5° (Figure 6), and regression analysis gave a highly significant (P < .001) line for each V_f. A hierarchical pattern was observed where the slope of the lines increased from 0.017 to 0.022 and 0.023 kg per degree as V_f increased for the C49, C59, and C70 wires, respectively.

View Figure

• Download as Powerpoint
• Download Figure

Significant interactions were shown between coating integrity and both N_FR (P < .002) and θ (P < .001) by the Pearson chi-square statistical analysis (Table 3). The coating integrity was independent of both bracket and V_f.

TABLE 3.  Contingency Table of Coating Integrity for 648 Tests

View Fullscreen

Tribology of the passive configuration

Because the results of several previous studies were validated with the same aforementioned SS-SS couple,7,27 comparisons could be made with the present data. Of particular interest was how the present frictional properties compared to those of beta-titanium (β-Ti) and nickel titanium (NiTi) wires, since the CC wires have similar elastic moduli (E) and would likely be used in the same clinical situations. The properties of SS wires were also considered since SS-SS couples remain the low frictional standard of contemporary orthodontics.

The overall mean μ_k-FR of the CC wires was generally greater than the values reported for SS, β-Ti, and NiTi wires (Figure 7).27 MANOVA of the CC and UC data from a previous study revealed a significant (P < .001, n = 99) difference between μ_k-FR of the CC and UC wires.7 Assessment of the means showed that the coating caused a 72% increase in μ_k-FR, which was unexpected since Parylene C has been reported to form low-friction surfaces.25,28

View Figure

• Download as Powerpoint
• Download Figure

For sliding friction, FR is often defined in terms of 3 mechanisms: the frictional resistances caused by shearing (SH_FR) of surface asperities, interlocking (IN_FR) of surface roughnesses, and plowing (PL_FR) of one surface by the other.37–42 With precision-engineered surfaces, such as metallic orthodontic wires and brackets, IN_FR and PL_FR can often be neglected. If one surface is substantially harder than the other, however, PL_FR tends to dominate sliding resistance as protrusions of the hard surface cause abrasive wear of the soft material.37–40

Metals and ceramics are much harder than polymers.37 Consequently, for the CC wire-bracket couples, the bracket materials were all much harder than Parylene C. Analysis of the coating-integrity data indicated that as N_FR increased, the edges of the bracket slots tended to cut notches through the wire coating (Figure 5a,b; Table 3). Such wear patterns are characteristic of PL_FR.39,40 As NO behavior increased, more energy transfer was required to slide the wire as more of the coating was damaged or removed. This caused an increase in RS with N_FR, which in turn caused the higher-than-expected mean μ_k-FR value for the CC wires. Similar notching patterns have been observed on metallic wires,35,36 and NO-induced increases in sliding resistances have been demonstrated for epoxy-coated SS wires.43

The effect of NO on FR, and consequently on RS, was particularly discernible on the specimens that exhibited stick-slip characteristics.38,40,42 A representative example of this behavior was shown by the P-δ trace (Figure 3c) and the notched wire surface (Figure 5b) of the C59-SS couple that was tested at N_FR = 0.075 kg and θ = 0°. (Recall from the statistical analysis of coating integrity in Table 3 that this wire could just as easily have been undamaged, as were 4 of the 9 wires tested at the same N_FR and θ.) The notched sections on this wire corresponded with the stick regions on the P-δ trace, and the undamaged sections corresponded with the slip regions. When the data for these regions were independently examined, P equaled 0.090 kg for the stick regions and 0.048 kg for the slip regions (Figure 3c). Additionally, one of the undamaged specimens had a P (Figure 3c) that was 0.02 kg. These observations suggest that sliding resistance could be substantially reduced by minimizing NO.

Tribology of the active configuration

From the previous discussion of wire-bracket configuration, the relationship between N_BI and BI is analogous to the relationship between N_FR and FR. The primary difference is that instead of being controlled directly, the magnitude of N_BI is controlled, at least in part, by the stiffness of the wire (given by the product of E and the area moment of inertia, I), the interbracket distance (IBD), the bracket width (width), θ, and θ_C.30,44,45 This lack of direct control makes frictional property comparisons in the active configuration more difficult than they are in the passive configuration. While μ_k-FR indicates the relative magnitude of FR that can be anticipated from any archwire-bracket couple, no equivalent coefficient exists to characterize the magnitude of BI. However, a model of the archwire-bracket system can be used to derive a mathematical relationship for N_BI in terms of stiffness (E·I), IBD, width, θ, and θ_C (Appendix 1). Using this relationship (equation A.24), a kinetic coefficient of binding (μ_k-BI) can be determined from the slope of a BI vs N_BI plot. This outcome is analogous to the relationship among FR, N_FR, and their slope, μ_k-FR.

An IBD of 19 mm was defined by the geometry of the frictional testing apparatus. For the CC wires, values of E were obtained from the literature,4 and the I value was calculated using nominal wire dimensions (Table 1, I = 3.32 × 10⁻¹⁵ m⁴). Using bracket and wire dimensions from the literature,27,46 a nominal width of 3.1 mm and a nominal θ_C of 1.1° were determined and used to calculate values of N_BI (Table 4). When the BI data were plotted with respect to N_BI, the 3 regression lines in Figure 6 collapsed into a single, highly significant (P < .001) line having a slope of 0.42 (Figure 8a), from which μ_k-BI was charted (Figure 8b). Values of N_BI and BI-N_BI linear regressions were similarly determined for UC (E and I were assumed the same as for CC wires) and metallic wires (E = 200, 68.9, and 33 GPa for SS, β-Ti, and NiTi wires, respectively; I = 8.03 × 10⁻¹⁵ m⁴).7,47 Because the investigators reported differences in the BI data for the various metal wire-bracket couples, each couple was analyzed separately. Although this limited the number of specimens for each regression (n = 4), all but 2 of the metal wire regression lines were statistically significant (Figure 8b).

TABLE 4.  Calculated Normal Force of Binding (N_BI) for the Coated Composite Archwires, Summarized by Wire Type^a

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

A comparison of the μ_k-BI data showed that the CC wire values fell within the values calculated for the other wire materials. When coupled with SS brackets, values of μ_k-BI for the β-Ti and NiTi wires were highest, perhaps because of cold welding.30,48 In contrast, the μ_k-BI of the SS-SS couple was the lowest, which only appeared to contradict the high binding sensitivity previously attributed to these wires.27,34,49 In fact, μ_k-BI was low for SS because equation A.24, which defined N_BI, was implicitly normalized with respect to wire stiffness. Thus, for a given θ, BI was highest for SS wires. For a given countervailing couple, however, the μ_k-BI data suggests that BI was lowest for SS wires.

The μ_k-BI values for the CC and UC wires were nearly identical, and no differences were shown by the statistical analysis of the BI data. Given the difference in the μ_k-FR values of these wires, the similarity in their μ_k-BI values was puzzling.

As with FR, BI can also be broken down into shearing (SH_BI), roughness interlocking (IN_BI), and plowing (PL_BI) components.37–42 Based on the previous discussion of tribological mechanisms in the passive configuration, PL_BI would be expected to dominate BI. Notches on the CC wires tested in the active configuration, which were like those observed in the passive configuration, supported this expectation.

For the UC wires, PL_BI-type behavior has been reported in the active configuration.7 The wear on these UC wire specimens, however, was different than the wear observed here for the CC wires in that NO occurred through the composite structure, rather than through a polymeric surface coating (Figures 5a, b vs Figure 5c). The differences in these 2 surfaces suggest some other mechanism, as perhaps differences in the depth of penetration, somehow equalized the energy transferred during sliding.

Observations of coating integrity indicated that the severity of plowing increased with θ (Table 3). The coating integrity data also showed an abrupt decrease to 0 for all θ ≥ 7.5°, whereas the decrease associated with N_FR was more gradual. Since N_BI was less than 0.2 kg for the C49 wires at θ = 7.5° (Table 4), the difference in this behavior could not be explained by differences in normal force magnitudes. However, N_BI was applied over a smaller area in the active configuration (part of the bracket slot side, Figure 2b) than N_FR was in the passive configuration (the entire bracket slot floor, Figure 2a). Consequently, the stress on the wire was greater, thereby increasing deformation of the coating. Thus, the sharp edges of the bracket slot dug further into the coating, increasing the severity of NO. Chamfering these slot edges could help reduce this behavior by making the load transition along the length of the wire more gradual,7,35,50 albeit at the expense of some loss of control.