Resistance to sliding of orthodontic appliances in the dry

and wet states: Influence of archwire alloy, interbracket

distance, and bracket engagement

Robert P. Kusy,1–4 John Q. Whitley1

1
Dental Research Center, University of North Carolina, Building 210H, Room 313, CB#7455,
Chapel Hill, North Carolina 27599-7455
2
Department of Orthodontics, University of North Carolina, Chapel Hill, North Carolina 27599-7450
3
Department of Biomedical Engineering, University of North Carolina, Chapel Hill, North Carolina 27599-7575
4
Curriculum in Applied and Material Sciences, University of North Carolina, Chapel Hill, North Carolina 27599-3290

Received 5 November 1999; revised 23 May 2000; accepted 16 June 2000

Abstract: Having established dimensional and mechanical identified the boundary between classical friction and bind-
characteristics, the resistances to sliding (RS) were measured ing phenomena. Stiffer archwires and shorter interbracket
in vitro for various archwires against stainless steel brackets. distances exacerbated binding, whereas, once corrected for
Using stainless steel ligatures, a constant normal force (300g) differing bracket engagement, RS was independent of slot
was maintained while second-order angulation (␪) was var- dimension. Unlike earlier results in the passive configura-
ied from −12° to +12°. Using miniature bearings to simulate tion, in the active configuration couples comprised of tita-
contiguous teeth, five experiments each were run in the dry nium alloys (NiTi and (␤-Ti) had higher RS values in the wet
or wet states with human saliva at 34°C as a function of four versus the dry state. For those archwire alloys evaluated,
archwire alloys, five interbracket distances, and two bracket two empirical expressions were adduced that comprise the
engagements. Outcomes were objectively analyzed to estab- binding component, the yield strength or elastic limit, and
lish when ␪=0, and the relative contact angles (␪r) were re- the beam length, which implicitly represent the stiffness,
plotted. Critical contact angles (␪c) that were determined via flexibility, and interbracket distance. © 2000 John Wiley &
experimentation were in good agreement with theory. Sons, Inc. J Biomed Mater Res, 52, 797–811, 2000.
Slopes and y-intercepts were tabulated from linear regres-
sion equations of RS against ␪ plots in both the passive (␪ ⱕ Key words: binding; friction; orthodontics; sliding mechan-
␪ c) and active (␪ ⱖ ␪c) configurations, for which ␪ = ␪c ics; tribology

INTRODUCTION brackets. From these results (see also Fig. 10–32 of

reference 4), we learned that sliding is dependent
Sliding mechanics is just in its nascent stages of sci- upon a couple’s geometry and its inherent material
entific development. In 1997, Nanda and Ghosh1 pre- characteristics. More recent in vitro studies on binding
sented an informative review of the last decade that have considered the behavior of preadjusted brackets
highlighted our current knowledge of classical friction subjected to angulation,5 alignment wires against
and implicitly underscored our limited knowledge of stainless steel brackets,6 and the results of reducing
binding. Among the early literature on binding,2,3 An- the ligature force by changing the material’s surface
dreasen’s group probably best expressed the essence chemistry.7
of binding when frictional resistance versus angula- Recently Kusy and Whitley8,9 presented a theory of
tion was plotted for couples comprised of nickel tita- sliding mechanics that partitioned the resistance to
nium or stainless steel wires against stainless steel sliding into three components [classical friction (FR),
elastic binding (BI), and physical notching (NO)] and
Correspondence to: R.P. Kusy; e-mail: rkusy@bme.unc. had as a boundary state the critical contact angle (␪c) at
edu which classical friction gives way to binding. This ␪c
Contract grant sponsor: The American Association of was calculated from the relative geometry of the
Orthodontics Foundation (AAOF) archwire-bracket couple—that is, the archwire dimen-
No benefit of any kind will be received either directly or
indirectly by the authors.
sion that contacts the floor of the slot (SIZE), the cor-
responding bracket dimension at the floor of the slot
© 2000 John Wiley & Sons, Inc. (SLOT), and the medial-distal dimension of the
798 KUSY AND WHITLEY

bracket (WIDTH) (Fig. 1). In closed-form solution this saliva is present as the third body) may revert to the
theory may be stated as: dry state as the saliva is literally squeezed out from
between the archwire-bracket contacts. This transfor-
(SIZE)2 − (WIDTH)2 mation from the wet to the dry state has been pre-
␪c = cos−1
(SIZE) (SLOT) ± ((WIDTH)2 (−(SIZE)2 sumed but never proven, although archwire-bracket
+ (SLOT)2 + (WIDTH)2))0.5 performance of the dry and wet states differ when
(1) ␪<␪c.11 Since recent in vivo work has shown that bench
experiments can be valid representations of clinical
which can be readily simplified to:
behavior,12 the present in vitro study investigates the
57.32 (1 − (SIZEⲐSLOT)) influence of archwire alloy, interbracket distance, and
␪c = (2) bracket engagement on the resistance to sliding, when
(WIDTHⲐSLOT)
tested against stainless steel brackets in the dry and
where the (SIZE/SLOT) equals the ‘ENGAGEMENT wet states.
INDEX’ and the WIDTH/SLOT equals the ‘BRACKET
INDEX.’ This angle is important for the practitioner to
know because sliding mechanics must ideally occur at MATERIALS AND METHODS
or below ␪c. Knowing that wire-bracket geometry dic-
tates ␪c places new importance on knowing the exact Brackets, archwires, and ligatures
dimensions of archwires and brackets.
Even if the practitioner knows ␪c, however, the rate Two bracket sizes (nominally a 0.018⬙ and a 0.022⬙
at which binding occurs is not specifically known. SLOT, Table I) having bracket WIDTHs of 0.1008⬙ and
That is a complex function of the material parameters 0.1037⬙, respectively, were selected without any angu-
of the archwire and bracket [e.g., stiffness (EI), hard- lation or pretorque. Four archwire alloys having the
ness (KHN or VHN), and yield strength (YS)], as well same nominal dimensions (0.016⬙ × 0.022⬙, Table I)
as geometric parameters [e.g., interbracket distance were selected that clinicians generally use for sliding
(IBD, Fig. 1), the ENGAGEMENT INDEX, and the with either bracket SLOT.8 Archwire-bracket couples
BRACKET INDEX]. For example, high EI wires should were secured using 0.010⬙ stainless steel (SS) ligature
have a more difficult time negotiating greater second- wires.
order angulations (␪), reductions in the IBD, and/or
increases in the ENGAGEMENT INDEX. Conse-
quently, the wire that traditionally has been the “gold Dimensional analysis
standard” of friction, stainless steel,10 may be among
the worst to slide upon whenever binding occurs. And Prior to testing, the actual SLOTs of the brackets
if the ␪ is large enough, the wet state (in which human were measured (±0.0001⬙) four times on each side for
a total of eight measurements per bracket using the
optics of a Kentron microhardness tester (Kent Cliff
Labs, Peekskill, New York), a Panasonic CCTV camera
(Matsoshita Electric Corporation of America, Secau-
cus, New Jersey), and a Burle TC1119 video monitor
(Burle Security Products Division, Lancaster, Pennsyl-
vania). The actual WIDTHs of these brackets were
measured (±0.0001⬙) three times using Starrett calipers
(L. S. Starrett Co., Athol, Massachusetts). The actual
SIZEs of the selected wires were measured (±0.00005⬙)
at five locations using a Sony ␮-Mate micrometer
(Sony Magnescale America, Inc., Orange, California).

Resistance to sliding (RS)

Figure 1. Important geometric parameters for sliding The RS values were determined using a custom-built,
mechanics: the size of an archwire (SIZE), the width of a friction-testing apparatus (Fig. 2) that was mounted to
bracket (WIDTH), the slot dimension of a bracket (SLOT), a mechanical testing machine (Instron Model TTCM,
the critical contact angle (␪c) for second-order angulation (␪),
the interbracket distance (IBD), and the beam length be- Instron Corp., Canton, Massachusetts). The general
tween contiguous teeth (L) in which L = IBD − (WIDTH/2)1 operation of this apparatus has been detailed else-
− (WIDTH /2)2. where in the literature.13,14 Using an in-line force
RESISTANCE TO SLIDING OF ORTHODONTIC APPLIANCES 799

TABLE I
Materials Evaluated
Brackets
Slot Width
Dimension, Dimension,
Alloy Code SLOT (⬙)a,b WIDTH (⬙)a,b Product Manufacturer
Stainless steel SS 0.0189 ± 0.0001 0.1008 ± 0.0006 Ultra-Minitrim Dentaurum (Pfzorheim, Germany)
Stainless steel SS 0.0232 ± 0.0002 0.1037 ± 0.0003 Ultra-Minitrim Dentaurum (Pfzorheim, Germany)

Archwires (all nominally 0.016⬙ × 0.022⬙)

SizeBase Size Height
Dimension, Dimension
Alloy Code SIZE (⬙)a,c (⬙)a,c Product Manufacturer
Stainless steel SS 0.01605 ± 0.00010 0.02210 ± 0.00015 Standard Edgewise American Orthodontics (Sheboygan, WI)
Cobalt chromium CoCr 0.01595 ± 0.00005 0.02195 ± 0.00005 Elgiloy Blue RMO Corp. (Denver, CO)
Beta-titanium ␤-Ti 0.01625 ± 0.00010 0.02185 ± 0.00005 TMA Ormco Corp. (Glendora, CA)
Nickel titanium NiTi 0.01575 ± 0.00010 0.02180 ± 0.00010 Nitinol Classic Unitek/3 M (Monrovia, CA)

Ligatures (nominally 0.010⬙)

Alloy Code Product Manufacturer
Stainless steel SS Preformed GAC (Commack, NY)
a
The SLOT, WIDTH, and SIZE dimensions may be converted to mm by multiplying by 25.4 mm/inch.
b
Dimensions were reported as the means ± standard deviations of twenty 0.018⬙ brackets or fifty-four 0.022⬙ brackets.
c
Dimensions were measured five times on three wires of each alloy and reported as the means ± standard deviations of 15
measurements.

transducer in conjunction with a computer-controlled via a saliva injector (not shown). All archwire and
motor, a constant normal force (N) of 300g was main- bracket materials were washed in 95% ethanol and air
tained throughout all tests. This N was applied di- dried prior to testing. For each determination, a
rectly to the archwire (a) by two 0.010⬙ SS ligature bracket was translated along virginal sections of
wires (b) (Table I and Fig. 2). The bracket (c) and the archwire at a sliding velocity of 10 mm/min.
ligatures were rotated in discrete intervals (±0.1°) via
a vernier scale. Sets of four frictionless roller bearings
(d), each having a BEARING DIAMETER = 4.00 mm, Data analyses
were placed above and below the archwire-bracket
engagement to simulate the SLOT edges of adjacent Each measurement was recorded as a drawing force
brackets. These bearings were positioned to simulate (P) against distance (␦) and plotted by a second com-
five discrete IBDs, which ranged from 18 to 8 mm puter. Such plots recorded some 60 to 100 data points
(Table II). The drawing force (P) that was required to within the kinetic region of interest [Fig. 3(a)]. After
translate the bracket relative to the archwire was mea- these values were averaged, they were converted to a
sured by the load cell (e) of the Instron machine. singular RS value by dividing the mean value by two.
For each experimental parameter (see headings in All RS values were plotted against ␪, and the best
Table II) 31 drawing forces were measured. First, the second order polynomial regression was fitted, which
drawing force was measured as the ␪ was evaluated in always had p < 0.05 [Fig. 3(b)]. Because of inevitable
the following order: −12.0, −10.0, −8.0, −6.0, −5.0, −4.5, manufacturing variations, bracket positioning, and
−4.0, −3.5, −3.0, −2.5, −2.0, −1.5, −1.0, −0.5, and 0°. machining tolerances, the mounting of brackets at ␪ =
Then the process was reversed as ␪ was evaluated in 0° is all but impossible. Consequently, the first deriva-
the same increments in the positive direction from 0 to tive of this regression line was calculated and solved
+12.0°. Finally, the drawing force at 0° was measured for ␪ to determine the actual location of the minimum
again. point of the polynomial curve, which represented the
Five experimental sequences were run in the dry true position of ␪ = 0°. Accordingly, the true experi-
and/or wet states at an ambient oral temperature of mental values of ␪s were corrected by systematically
34°C (Table II). Wet tests were run using fresh, well- shifting all measured ␪s. The RS values were replotted
characterized whole human saliva that was obtained against these corrected values of ␪. These RS values
without additional stimulation.15,16 Operating in a could be dominated by FR or BI, the former of which
closed-loop system, a peristaltic pump provided a occurred at ␪ < ␪c and the latter of which occurred at
flow rate of 3 cc/min to the archwire-bracket couple16 ␪ > ␪c [Fig. 3(c)]. The intersection of these extrapolated
800 KUSY AND WHITLEY

shifted such that ␪c was set to 0°, thereby defining the

relative contact angles (␪r) as ␪r = ␪ − ␪c [Fig. 3(d)].

RESULTS
As Table II indicates, the results from these five ex-
periments are summarized in Tables III–VII as a func-
tion of archwire alloy (SS, Co-Cr, ␤-Ti, or NiTi), IBD,
state (dry or wet), and SLOT dimension (0.018⬙ or
0.022⬙) for a constant normal force (N = 300g) at 34°C.
Having calculated the theoretical values of ␪ c based
on Equations 1 or 2, and having corrected values of ␪,
the true experimental values of ␪ c were compared and
differed by less than 1° (columns 3 and 4 of Tables
III–VII). In both the passive and active configurations,
linear correlations between RS and ␪ or ␪ r were ob-
served (Fig. 4), from which slopes (m) and y-intercepts
(b) were determined.
In the passive configuration [Fig. 4(a)] the slopes
equaled zero, the highest magnitude never exceeding
0.005 kg/° (column 5 of Tables III–VII). Any other
outcome would have indicated an experimental error
because, until the wire contacts opposing sides of the
bracket SLOT, friction is independent of contact area17
and angulation.18,19 The y-intercepts were quantities
that varied with wire alloy (column 6 of Tables III or
VI) or state (Tables V and VII) but not with bracket
SLOT size (Tables IV and V), nor with IBD (Tables IV,
V, or VII).
In the active configuration [Fig. 4(b)] the slopes dif-
fered depending on how much the RS increased as a
consequence of BI. Now the wire alloy (column 7 of
Tables III or VI), the state (Tables V and VII), or the
IBD (Tables IV, V, or VII) could profoundly influence
BI, whereas the bracket SLOT size (Tables IV and V)
did not. As expected, the y-intercepts at ␪ r = 0° [see
Fig. 4(b); or equivalently at ␪ = ␪ c in Fig. 4(a)] closely
mirrored the outcomes and the observations of the
y-intercepts in the passive configuration (cf. columns 8
with columns 6 of Tables III–VII).
When the RS against ␪ r plots were made for two
parameters at a time, some further insights were ob-
tained (Fig. 5–7). When bracket SLOT size and IBD
were grouped for an SS-SS couple in the dry state (Fig.
5), binding increased about 2.5 times as the IBD de-
creased from 18 mm to 8 mm. As has been seen in
earlier work in the passive configuration,11,20 SLOT
size wasn’t a factor. Nonetheless, note that because the
Figure 2. Present frictional apparatus indicating the posi- ENGAGEMENT INDEX of an 0.016⬙× 0.022⬙ archwire
tion of the archwire (a), the stainless steel (SS) ligature wires in an 0.022⬙ bracket SLOT is less than this same wire in
(b), the SS bracket (c), the adjustable frictionless roller bear- an 0.018⬙ SLOT, the value of ␪ c is greater in the former
ings (d), and the drawing force load cell (e).
than the latter (cf. Fig. 3 of reference 8). Consequently,
binding will indeed be more severe in the latter couple
linear regression lines identified each true experimen- for equivalent values of ␪, once each ␪ exceeds its
tal ␪c, which could be compared to each correspond- value of ␪c.
ing theoretical ␪ c [cf. Eq.(1) or (2)]. To focus solely on Having eliminated SLOT size from further analysis,
the binding aspects at ␪ > ␪c, the data was further the remaining three parameters were considered two
RESISTANCE TO SLIDING OF ORTHODONTIC APPLIANCES 801

TABLE II
Summary of Experimental Parameters Investigated Against SS Brackets
Archwire Alloy
Bracket Nominal
Slot Size IBD SS CoCr ␤-Ti NiTi
(⬙) (mm) Dry Wet Dry Wet Dry Wet Dry Wet
a
18 3, 4 3 3 3
14 4
0.018 12 4
10 4
8 4
18 5 6, 7 6, 7 6, 7 6, 7
14 5 7 7 7 7
0.022 12 5 7 7 7 7
10 5 7 7 7 7
8 5 7 7 7 7
a
Table numbers in which the results of an experimental sequence may be found.

Figure 3. Details of the procedure involved in the data reduction of frictional measurements. Typical drawing force (P)
against distance (␦) plot for a constant normal force (N) and ␪, highlighting the kinetic region (a). From a series of plots such
as (a) at different values of ␪, a plot of resistance to sliding (RS) against ␪ at a constant N is generated, from which the first
derivative of a polynomial regression line identifies the small error in ␪ that is associated with manufacturing and/or
mounting (b). After correcting for this error, linear regression lines are determined for the data above and below the
theoretical value of ␪c, and RS is plotted against the absolute value of ␪ (c). The intercept of the two lines at ␪ = ␪c represents
the experimental value of ␪c. The ␪ data at or above this experimental ␪c is shifted by ␪c so that the relative contact angle, ␪r
= (␪ − ␪c), may be replotted versus RS and the elastic binding (BI) studied (d).
802 KUSY AND WHITLEY

TABLE III
Results for 0.018ⴖ Stainless Steel Brackets Tested in the Dry Statea (cf. Table II)
␪c (°) Passive Configuration Active Configuration
c
Regression Equation Regression Equationc
IBD Slope, m y-intercept, b Slope, m y-intercept, b
Wire Alloy (mm) Theoreticalb Experimental (kg/°) (kg) (kg/°) (kg)
SS 18 1.7 1.8 −0.001 0.033 0.019 0.032
CoCr 18 2.1 2.6 0.001 0.039 0.030 0.041
␤-Ti 18 1.9 1.7 0.005 0.065 0.011 0.074
NiTi 18 2.1 2.8 0 0.055 0.008 0.056
a
N = 300g.
b
Based on equation (1) or (2).
c
y = mX + b.

TABLE IV
Results for 0.018ⴖ Stainless Steel Brackets Tested in the Dry Statea (cf. Table II)
␪c (°) Passive Configuration Active Configuration
c
Regression Equation Regression Equationc
IBD Slope, m y-intercept, b Slope, m y-intercept, b
Wire Alloy (mm) Theoreticalb Experimental (kg/°) (kg) (kg/°) (kg)
SS 18 1.7 1.8 −0.001 0.033 0.019 0.032
14 1.7 1.9 0 0.033 0.024 0.034
12 1.6 1.8 0 0.031 0.030 0.030
10 1.6 2.1 0.001 0.032 0.035 0.034
8 1.6 2.5 0 0.035 0.047 0.030
a
N = 300g.
b
Based on equations (1) or (2).
c
y = mX + b.

TABLE V
Results for 0.022ⴖ Stainless Steel Brackets Tested in the Dry Statea (cf. Table II)
␪c (°) Passive Configuration Active Configuration
c
Regression Equation Regression Equationc
IBD Slope, m y-intercept, b Slope, m y-intercept, b
Wire Alloy (mm) Theoreticalb Experimental (kg/°) (kg) (kg/°) (kg)
SS 18 4.1 4.6 0 0.035 0.018 0.034
14 4.1 4.4 −0.001 0.035 0.024 0.031
12 4.0 4.7 0.001 0.031 0.031 0.036
10 4.1 4.5 0 0.033 0.038 0.032
8 4.0 4.6 0 0.034 0.047 0.035
a
N = 300g.
b
Based on equations (1) or (2).
c
y = mX + b.

TABLE VI
Results for 0.022ⴖ Stainless Steel Brackets Tested in the Wet Statea (cf. Table II)
␪c (°) Passive Configuration Active Configuration
c
Regression Equation Regression Equationc
IBD Slope, m y-intercept, b Slope, m y-intercept, b
Wire Alloy (mm) Theoreticalb Experimental (kg/°) (kg) (kg/°) (kg)
SS 18 3.9 4.2 0.001 0.040 0.022 0.045
CoCr 18 3.9 4.3 0 0.046 0.029 0.047
␤-Ti 18 4.0 4.2 −0.002 0.082 0.015 0.073
NiTi 18 4.2 4.7 0 0.070 0.010 0.070
a
N = 300g.
b
Based on equations (1) or (2).
c
y = mX + b.
RESISTANCE TO SLIDING OF ORTHODONTIC APPLIANCES 803

TABLE VII
Results for 0.022ⴖ Stainless Steel Brackets Tested in the Wet Statea (cf. Table II)
␪c (°) Passive Configuration Active Configuration
c
Regression Equation Regression Equationc
IBD Slope, m y-intercept b Slope, m y-intercept b
Wire Alloy (mm) Theoreticalb Experimental (kg/°) (kg) (kg/°) (kg)
SS 18 3.9 4.2 0.001 0.040 0.022 0.045
14 3.9 4.3 −0.001 0.048 0.030 0.046
12 4.1 4.5 0 0.042 0.034 0.042
10 4.2 4.7 0.001 0.051 0.046 0.056
8 4.1 4.9 0.001 0.056 0.071 0.059
CoCr 18 3.9 4.3 0 0.046 0.029 0.047
14 3.9 4.6 0.002 0.035 0.037 0.044
12 4.1 4.5 0 0.040 0.042 0.041
10 4.0 4.7 0 0.050 0.052 0.050
8 4.1 4.2 0.001 0.042 0.060 0.045
␤-Ti 18 4.0 4.2 −0.002 0.082 0.015 0.073
14 4.1 4.6 −0.002 0.077 0.022 0.068
12 4.0 4.1 0.003 0.064 0.029 0.075
10 4.0 4.4 0.003 0.063 0.035 0.075
8 4.0 4.4 0 0.064 0.050 0.064
NiTi 18 4.2 4.7 0 0.070 0.010 0.070
14 4.2 4.0 0 0.066 0.016 0.068
12 4.3 4.8 0.002 0.064 0.023 0.074
10 4.0 4.9 0.001 0.086 0.031 0.093
8 4.4 5.0 0 0.073 0.038 0.072
a
N = 300g.
b
Based on equations (1) or (2).
c
y = mX + b.

at a time (Fig. 6 and 7). While maintaining IBD at a slightly greater in the wet state for three of the wire
constant 18 mm and using SLOTs of 0.018⬙and 0.022⬙ alloys, with the fourth alloy (CoCr) being collinear.
for the dry and wet states, respectively, the archwire When archwire alloy and IBD were grouped as pa-
alloy and state were grouped as parameters (Fig. 6). rameters (Fig. 7) while maintaining the archwires in
Alloy had the greatest influence on RS, the values in- 0.022⬙ SLOTs in the wet state, stiffer wires and smaller
creasing as the stiffness (or, for wires having the same IBDs resulted in higher values of RS. Here the absolute
dimensions, the modulus of elasticity) increased in magnitudes of the RS values at ␪r = 0° (or equivalently
this order: NiTi, ␤-Ti, SS, and CoCr.21 The dry or wet at ␪ = ␪c) are small and correspond to RS = FR. Once
state had little affect on these outcomes, RS being binding dominates, those RS values pale relative to

Figure 4. Influence of ␪ on RS in the passive

configuration (a) and ␪r on RS in the active
configuration (b) for a 0.016⬙× 0.022⬙ SS
archwire-0.022⬙ SS bracket in the wet state at
an IBD = 10 mm. Using that couple’s data, the
details of the constructions are illustrated to
determine the slopes (m), y-intercepts (b),
and average of all data points from which the
coefficients of friction (␮k) are calculated (cf.
Tables VIII and IX). Note that the ␪ = ␪c in
part (a) is equivalent to the ␪r = 0° in part (b)
because ␪r = ␪ − ␪c.
804 KUSY AND WHITLEY

more scattered. Specifically, the difference between

couples bearing against SS or CoCr wires were imma-
terial, and the same was true for couples bearing
against NiTi or ␤-Ti wires. In all couples the values of
␮k in the wet state were larger than the values ob-
tained in the dry state, which suggested that saliva, as
the third body, was behaving more like an adhesive
than a lubricant. Earlier work suggested that there was
a slight adhesive effect in SS-SS couples, and a lubri-
cating effect in couples comprised of Ti-bearing wires
(NiTi and ␤-Ti).11

Relative importance of geometry in determining ␪c

but not in determining elastic binding (BI)
Figure 5. Influence of SLOT and IBD on binding as seen in Equations (1) and (2) showed that the archwire-
plots of RS against ␪r for 0.016⬙× 0.022⬙ SS archwire-SS bracket dimensions, which constitute the ENGAGE-
bracket couples in the dry state (cf. Tables IV and V).
MENT INDEX (= SIZE/SLOT) and the BRACKET IN-
DEX (= WIDTH/SLOT), determine the boundary, ␪c.
their absolute magnitudes, for example, at ␪r = 10°. That parameter alone partitions the end of the classical
Even when the most benign circumstance is com- friction region from the beginning of the elastic bind-
pared, namely that of a NiTi-SS archwire-bracket ing region [Fig. 3(c)].
couple at an IBD = 18 mm (Fig. 7), RS at least doubles The present work shows that neither SLOT dimen-
at ␪r = 10° relative to its value at ␪r = 0°. sions nor dry/wet states control ␪r beyond ␪r = 0° (that
is, ␪ beyond ␪ = ␪c), and ultimately the RS that is
associated with the onset of binding (cf. Fig. 5 and 6).
DISCUSSION Instead, binding is profoundly influenced by the di-
mensions of an archwire and its brackets that consti-
tute a beam and its supports. Consequently, binding is
Computations of the kinetic coefficients of friction dependent upon the anatomy of the contiguous tooth
(µk) structure and the inherent material characteristics of
the appliance. The influence of changing IBD and the
From plots such as Figure 4 and their subsequent archwire alloy are but two manifestations of that de-
constructions, results from columns 6 and 8 of Tables pendence (Fig. 7). As IBD decreases, the interaction of
III–VII can be used to calculate ␮k. From the passive contiguous teeth increases, and this causes the values
configuration two computations may be gleaned [see of RS to increase. This outcome is further exacerbated
Fig. 4(a)]: one from the average of all data points and as the stiffness of the archwire alloy increases (column
one from the y-intercept at ␪ = 0°. In the active con- 3, Table X). For each alloy, the slopes of ␪r versus RS at
figuration a third computation may be obtained from each IBD extrapolate to a common value of RS at ␪r =
the y-intercept at ␪r = 0°, which also coincides with the 0°. Each of those values is small relative to the mag-
rightmost boundary of the passive configuration at ␪ = nitude of RS at ␪r = 8°, for example, as binding easily
␪c. In all three cases, values of ␮k are obtained by overwhelms any effects from classical friction.
dividing the y-intercept at ␪r = 0° by the normal force Note that only in one of these tests (SS archwire, IBD
of the test, which in this study is 300g. When these = 8 mm, and ␪r = 8°; Fig. 7) was there any indication
calculations are tabulated for each wire-bracket-IBD that the elasticity of the bearing surfaces had been
combination (Tables VIII and IX), the three values gen- exceeded, as evidenced by the departure from linear-
erally agree within ±0.01. ity and subsequent physical notching. Thus, although
In the dry state (Table VIII), the ␮k values for the this data represents a 12-fold range of RS over which
SS-SS couple varies from 0.10 to 0.12, with the CoCr-SS sliding mechanics can operate, albeit with increasing
couple a close second at 0.13; the NiTi-SS couple binding, a practitioner should intuitively attempt to
follows at 0.19, with the ␤-Ti-SS couple last at minimize RS and ␪r.
0.23. 1,11,20,22–24 Among these materials, the ␤-Ti-SS
couple has shown the greatest improvement over the
Influence of classical friction (FR) on the
years, although this couple still has about twice the normalized RS values of elastic binding (BI)
friction of an SS-SS couple of the same dimensions. As
alluded to before, the SLOT dimension has no effect As implied in Figure 4(b), the RS in the active con-
on ␮k.20,21 figuration (i.e., for ␪r ⱖ 0°) is comprised of FR plus BI,
In the wet state (Table IX), the data was somewhat when NO = 0. Thus, the BIs of each couple may be
RESISTANCE TO SLIDING OF ORTHODONTIC APPLIANCES 805

Figure 6. Influence of
archwire alloy and dry/wet
state on binding as seen in plots
of RS against ␪ r for 0.016⬙×
0.022⬙ archwire-SS bracket
couples (cf. Tables III and VI).
Although IBD = 18 mm for all
couples, SLOTs were 0.018⬙ and
0.022⬙ for the dry and wet
states, respectively.

calculated for each IBD and SLOT size in the dry and/ The slopes of the resulting linear regression lines
or wet states via the equation,8,21 were generally rank-ordered for all IBDs as follows
(Fig. 8): CoCr, SS, ␤-Ti, and NiTi. In other investiga-
BI = RS − FR, (3)
tions involving SS brackets, the CoCr and SS archwires
in which RS may be obtained from Figures 5–7 and FR that had the lowest values of ␮k also had the highest
may be obtained from the y-intercepts of plots like values of elastic moduli21,25,26 (Table X) and hard-
Figure 4(b) , or equivalently, from the FR values of RS ness.21,27 Unfortunately, they were also the most un-
at ␪ = ␪c in Figure 4(a) (cf. columns 5 and 6 or column forgiving at ␪r > 0° (i.e., at ␪ > ␪c). This observation
8 of Tables III–VII). underscores why practitioners need to know what the

Figure 7. Influence of
archwire alloy and IBD on BI as
seen in plots of RS against ␪r for
0.016⬙× 0.022⬙ archwire-0.022⬙
SS bracket couples in the wet
state (cf. Table VII).
806 KUSY AND WHITLEY

TABLE VIII
Frictional Coefficients for Stainless Steel Brackets Tested in the Dry State
Passive Configuration Active Configuration
a b c b
IBD Mean RS Coef. frict. y-intercept Coef. frict. y-interceptd Coef. frict.b
Wire Bracket (mm) (kg) (␮k) (kg) (␮k) (kg) (␮k)
SS 0.018 18 0.033 0.11 0.033 0.11 0.032 0.11
14 0.033 0.11 0.033 0.11 0.034 0.11
12 0.030 0.10 0.031 0.10 0.030 0.10
10 0.032 0.11 0.032 0.11 0.034 0.11
8 0.034 0.11 0.035 0.12 0.030 0.10
CoCr 0.018 18 0.039 0.13 0.039 0.13 0.041 0.14
␤-Ti 0.018 18 0.068 0.23 0.065 0.22 0.074 0.25
NiTi 0.018 18 0.056 0.19 0.055 0.18 0.056 0.19
SS 0.022 18 0.035 0.12 0.035 0.12 0.034 0.11
14 0.034 0.11 0.035 0.12 0.031 0.10
12 0.033 0.11 0.031 0.10 0.036 0.12
10 0.032 0.11 0.033 0.11 0.032 0.11
8 0.034 0.11 0.034 0.11 0.035 0.12
a
Mean of all RS values in the passive configuration [cf. Fig. 4(a)].
b
Kinetic coefficient of friction = ␮k = (mean RS/normal force) = (y-intercept/normal force), where here N = 300g.
c
RS values at ␪ = 0° [cf. Fig. 4(a)] which correspond to column 6 of Tables 3, 4, and 5.
d
RS values at ␪ = ␪c [cf. Fig. 4(a)] or at ␪r = 0° [cf. Fig. 4(b)] which correspond to column 8 of Tables 3, 4, and 5.

actual ␪c is that differentiates classical friction from the binding data is complete. Now the y-inter-
elastic binding, since failure to do so can result in sub- cepts equal zero, in effect partitioning BI from
stantial BI, or worse yet, in increases in RS which re- the other components of RS.8,21 Moreover, comparison
sult from permanent damage to the wire via physical of the relative contact angles (␪ r ) show the rate
notching (NO).28,29 at which a practitioner pays for additional second-
Having modified both axes, the practical utility of order angulation in terms of enhanced sliding resis-

TABLE IX
Frictional Coefficients for Stainless Steel Brackets Tested in the Wet State
Passive Configuration Active Configuration
a b c b
IBD Mean RS Coef. frict. y-intercept Coef. frict. y-interceptd Coef. frict.b
Wire Bracket (mm) (kg) (␮k) (kg) (␮k) (kg) (␮k)
SS 0.022 18 0.042 0.14 0.040 0.13 0.045 0.15
14 0.047 0.16 0.048 0.16 0.046 0.15
12 0.042 0.14 0.042 0.14 0.042 0.14
10 0.053 0.18 0.051 0.17 0.056 0.19
8 0.057 0.19 0.056 0.19 0.059 0.20
CoCr 0.022 18 0.047 0.16 0.046 0.15 0.047 0.16
14 0.038 0.13 0.035 0.12 0.044 0.15
12 0.041 0.14 0.040 0.13 0.041 0.14
10 0.050 0.17 0.050 0.17 0.050 0.17
8 0.043 0.14 0.042 0.14 0.045 0.15
␤-Ti 0.022 18 0.078 0.26 0.082 0.27 0.073 0.24
14 0.074 0.25 0.077 0.26 0.068 0.23
12 0.068 0.23 0.064 0.21 0.075 0.25
10 0.068 0.23 0.063 0.21 0.075 0.25
8 0.064 0.21 0.064 0.21 0.064 0.21
NiTi 0.022 18 0.070 0.23 0.070 0.23 0.070 0.23
14 0.067 0.22 0.066 0.22 0.068 0.23
12 0.068 0.23 0.064 0.21 0.074 0.25
10 0.089 0.30 0.086 0.29 0.093 0.31
8 0.073 0.24 0.073 0.24 0.072 0.24
a
Mean of all RS values in the passive configuration [cf. Fig. 4(a)].
b
Kinetic coefficient of friction = ␮k = (mean RS/normal force) = (y-intercept/normal force), where here N = 300g.
c
RS values at ␪ = 0° [cf. Fig. 4(a)] which correspond to column 6 of Tables 6 and 7.
d
RS values at ␪ = ␪c [cf. Fig. 4(a)] or at ␪r = 0° [cf. Fig. 4(b)] which correspond to column 8 of Tables 6 and 7.
RESISTANCE TO SLIDING OF ORTHODONTIC APPLIANCES 807

TABLE X
Mechanical Measurements of Archwires
Elastic Proportional Elastic Johnson elastic Yield Ultimate tensile
modulus, limit, limit, limit, strength, strength,
Alloy Code E (GPa)a PL (MPa)a EL (MPa)b JEL (MPa)a YS (MPa)a UTS (MPa)a
Stainless steel SS 182 758 1,060 1,520 1,370 1,990
Cobalt chromium CoCr 180 786 1,250 1,190 1,240 1,760
Beta-titanium ␤-Ti 69.3 469 1,130 1,060 835 1,410
Nickel titanium NiTi 43.5 234 1,090 1,022 332 1,660
a
Using a ½⬙ gage length, 10% extensometer and a ½⬙ gage length, 50% extensometer at a maximum load scale of 50 kg (1g
= 1cN), the E, PL, JEL, YS, and UTS values were obtained on three specimens from the slopes of the stress-strain plots (E),
the last stress values at which the quotients of stress and strain were constants during continuous loading (PL), the stresses
at which the slopes of the stress-strain plots were reduced to 50% of the initial slopes (JEL), the corresponding magnitudes
of stresses after offsetting the slopes by a strain of 0.001 (YS), and the maximum stresses that were observed (UTS).
b
Using a nip-to-nip distance of 4⬙ at a maximum load scale of 50 kg, the EL values were obtained on three specimens from
the first stresses at which a separation between replicate return cycles were observed as the load was increased in 2.5 kg
increments. The EL values reported are the averages between those last stresses that show no separation and the first stresses
that do show a separation.

tance. Perhaps this is best illustrated by a practical mately 10°. From columns 3 and 4 of Table VII, the
example. theoretical ␪c and the experimental ␪c equal 4.2° and
An orthodontist has activated a 0.016⬙× 0.022⬙ SS 4.7°, respectively. Averaging these values gives ␪c =
archwire in the arch of a patient, on whose teeth 0.022⬙ 4.5°. Since ␪ = ␪c+ ␪r, rearrangement and substitution
SS brackets have been cemented. The IBD (Fig. 1) is 10 gives ␪r = 10°− 4.5° = 5.5°. By referencing the lower
mm, and the ␪ between two bicuspids is approxi- left-hand frame of Figure 8 for SS at ␪r = 5.5° (BI = 0.24

Figure 8. Influence of IBD,

archwire alloy, dry/wet state,
and SLOT dimension on plots
of BI against ␪ r for 0.016⬙×
0.022⬙ archwire-SS bracket
couples. The BI values were ob-
tained for each plot by subtract-
ing its y-intercept at ␪ = ␪c [that
is, the value of classical friction
(FR) at ␪r = 0°; cf. column 8 of
Tables III–VII] from the value
of RS at each ␪r (cf. Fig. 5–7).
Since ␪r is such that physical
notching (NO) equals zero in
these experiments, BI = RS − FR
− NO = RS −FR.21
808 KUSY AND WHITLEY

kg) and comparing it to the classical friction (cf. col- equal to 18 mm versus each value of ‘n’. As Figure 9
umn 8 of Table VII; FR = y-intercept = 0.056 kg), this shows for the five L values of the 0.022⬙ SLOT in the
practitioner learns that BI is over four times greater wet state, this empirical expression adequately pre-
than the classical friction normally associated with dicted the behavior of the three archwire alloys (SS,
that archwire-bracket couple. Thus, friction could CoCr, and ␤-Ti) that obeyed linear elasticity. Only
have been reduced to 1/4 of its present value, if ␪ had NiTi did not comply because the YS was underesti-
been limited to an angle near its ␪c so that ␪r = 0°. Note mated by the linear construct to this superelastic alloy,
that even if a NiTi wire had been substituted, the bind- which follows nonlinear elasticity.26,38 Note that both
ing associated with this active configuration (BI = 0.16 the values of E and IBD are implicitly contained in
kg) would have exceeded the classical friction (FR = Equation (4) as YS≈(E)(flexibility) (see reference 26)
0.093 kg) associated with the passive configuration by and as L=IBD−(WIDTH/2)1−(WIDTH/2)2 (see Fig. 1),
over 70%. Apparently, no available materials can over- respectively.
come such an ill-conceived biomechanical solution. When the plots of Figure 9 are regrouped by alloy
using only the wet state and the 0.022⬙ data, the suc-
cess of the empirical model is evident for those alloys
that follow linear elasticity (Fig. 10). Here, even after
Wire stiffness (ExI) and interbracket distance (IBD)
in an empirical elastic binding (BI) model doubling the resolution of the y-axis, the function is
virtually independent of L. The absence of a system-
atic trend in L suggests that a random, experimental
When the BI is partitioned from RS, once again the error is present, which will decrease as instrumenta-
stiffest wire alloys and the smallest IBDs exacerbate tion and methodology improve.
the BIs of those archwire-bracket couples. But how are For the case of the NiTi alloy that obeyed non-linear
these related? So far, a comprehensive study has not elasticity, an alternative empirical approach was evalu-
produced sufficient data to adduce the fundamental ated in which EL was substituted for YS in Equation
relationship, although some theoretical approaches (4) so that:
have been suggested.19,30–32 The first two studies ex-
pressed the influence of a second-order force system 106(BIⲐEL)Ⲑ(LIBD=18ⲐLIBD=n), (5)
on the displacement of an archwire “beam” in terms of By definition, EL is the largest stress that an alloy can
stiffness [the mathematical product of the modulus of withstand without displaying any plastic deforma-
elasticity (E) and the area moment of inertia (I)], beam tion.39,40 This value is adduced by repeatedly loading
length (L), and WIDTH.30,31 The third derived the de- and unloading specimens to ever-increasing magni-
flection of an archwire in a bracket whether the SLOT tudes until the elastic recovery or springback is less
is filled or partially filled in terms of E, ␪, WIDTH, than 100% (Table X). Figure 11 shows the outcomes
IBD, SIZE, and yield strength (YS).32 The fourth study from Equation (5) against ␪r as a function of archwire
derived the normal force of binding in terms of E, I, BI, alloy. To facilitate comparison, the ranges of the data
␪r, WIDTH, and IBD.19 Other potential parameters of of Equation (4) (see Fig. 10) appear, too. Note that this
importance include N, KHN or VHN, proportional shaded region is coincident for CoCr alloys but
limit (PL), elastic limit (EL), and ultimate tensile slightly lower and higher for SS and ␤-Ti alloys, re-
strength (UTS). Fortunately, basic research has shown spectively. When the mean slope of all linear elastic
that simplified parametric interdependencies can ex- lines derived from either Equation (4) or (5) is super-
ist, e.g., between VHN and YS, or VHN and E, which posed as a dashed line, the NiTi alloy that caused a
should simplify any expression.33,34 Finally, coupled gross overestimation of values via Equation (4) now
phenomena, such as resilience [=(1/2)(PL)2/E] and causes an underestimation of values via Equation (5).
flexibility [=(PL)/E], may be important, too.26,35–37 Clearly such an empirical approach is less dependent
In the absence of a comprehensive study, an empiri- upon defining the elastic-plastic boundary in the three
cal approach was taken in which IBD and EI were linear elastic alloys (SS, CoCr, and ␤-Ti) than it is on
identified as essential parameters, although in this defining the nonlinear elastic alloy (NiTi) boundary.
study ‘I’ was maintained constant. The rationale was
simply that if BI could be reduced to a single-valued
function, then those would be the parameters of im-
portance. Among the many approaches that were CONCLUSIONS
based on various mechanical measurements (Table X),
one provocative outcome was the expression, Wire stiffnesses (EI, or when the wire dimensions
are constant, E) have profound influences on binding.
106(BIⲐYS)Ⲑ(LIBD=18ⲐLIBD=n), (4)
Stiffer wires (SS and CoCr) have a greater difficulty
in which YS is obtained at 0.1% offset (Table X) and negotiating greater angulation than do less stiff wires
LIBD = 18/LIBD = n is the ratio of beam lengths at IBDs (␤-Ti and NiTi).
RESISTANCE TO SLIDING OF ORTHODONTIC APPLIANCES 809

Figure 9. Outcomes when the

same data sets for BI (cf. Fig. 8)
are normalized (cf. Eq. 4) by the
yield strengths (YS) of their re-
spective archwire alloys (Table
X) and by the ratios of the L
values at an IBD = 18 mm ver-
sus an IBD = ‘n’ mm (cf. Fig. 1).
In the present experiment in
which WIDTH = 2.63 mm and
BEARING DIAMETER = 4.00
mm (cf. Fig. 2), the determina-
tion of L differs slightly, i.e., L =
IBD − (WIDTH/2) −(BEARING
DIAMETER/2) = IBD − 3.32
mm. Consequently, values of
IBD = 8, 10, 12, 14, and 18 mm
correspond to nominal values
of L = 5, 7, 9, 11, and 15 mm,
respectively.

Figure 10. Replot of Equation

(4) as a function of archwire al-
loy showing that the simple,
two parameter, empirical
model has merit for those
couples using archwires that
follow linear elasticity (SS, CoCr,
and ␤-Ti) in the wet state using
a 0.022⬙ SLOT. Having doubled
the resolution of the y-axis
here, the absence of any sys-
tematic behavior in L suggests
that the scatter is presumed to
be random, experimental error.
810 KUSY AND WHITLEY

Figure 11. Plot of Equation (5)

as a function of archwire alloy
to test whether EL can define
the elastic-plastic boundary
when nonlinear elasticity occurs.
For comparison the outcomes
of Equation (4) (cf. Fig. 10) are
designated by the shaded re-
gions. When the 30 slopes of
the three linear elastic alloys
are averaged and superposed
(the dashed lines), the nonlin-
ear elastic alloy now underesti-
mates the proposed empirical
model that the three linear elas-
tic alloys follow remarkably
well. Clearly the nonlinear elas-
tic alloy is more problematic.

An inverse relationship between RS and IBD exists 3. Peterson L, Spencer R, Andreasen GF. A comparison of fric-
for all archwire alloys evaluated. tional resistance of Nitinol and stainless steel wires in Edge-
wise brackets. Quint Internat 1982;13:563–571.
The RS is independent of the SLOT once the second- 4. Proffit WR. Contemporary orthodontics, 2nd edition. St. Louis:
order angulation (␪) exceeds the critical contact angle Mosby; 1993. p 311.
(␪c), such that the relative contact angle, ␪r = ␪ − ␪c > 0°. 5. Sims AP, Waters NE, Birnie DJ. A comparison of the forces
For SS wires at all the IBDs studied, saliva acts like required to produce tooth movement ex vivo through three
an adhesive. At a single IBD (18 mm), no difference types of pre-adjusted brackets when subjected to determined
tip or torque values. Br J Orthod 1994;21:367–373.
was noted in the RS values for the dry and wet states 6. Dickson JA, Jones SP, Davies EA. A comparison of the fric-
of CoCr wires. For NiTi and ␤-Ti wires at the same tional characteristics of five initial alignment wires and stain-
IBD, the wet state had higher RS values than the dry less steel brackets at three bracket to wire angulations–an in
state. The results for the CoCr wire may be an indica- vitro study. Br J Orthod 1994;21:15–22.
tion that these wires are not wetted as readily as the 7. DeFranco DJ, Spiller RE Jr, von Fraunhofer JA. Frictional re-
sistances using Teflon-coated ligatures with various
other alloys are. bracket−archwire combinations. Angle Orthod 1995;65:63–74.
Once BI is partitioned from RS and normalized 8. Kusy RP, Whitley JQ. Influence of archwire and bracket di-
via the empirical expressions, 106(BI/YS)/(LIBD = 18/ mensions on sliding mechanics: Derivations and determina-
LIBD = n) or 106(BI/EL)/(LIBD = 18/LIBD = n), the data tions of the critical contact angles for binding. Eur J Orthod
generally consolidates. These outcomes suggest that at 1999;21:199–208.
9. Kusy RP, Whitley JQ. Assessment of second-order clearances
least one material parameter (YS and/or EL) and one between orthodontic archwires and bracket slots via the critical
dimensional parameter [the beam length between con- contact angle for binding. Angle Orthod 1999;69:71–80.
tiguous teeth (L)] play major roles. 10. Kusy RP. Materials and appliances in orthodontics: brackets,
archwires, and friction. Current Opinion in Dent 1991;1:634–
644.
The authors thank the Dynaroll Corporation for donating
11. Kusy RP, Whitley JQ, Prewitt MJ. Comparison of the frictional
the roller bearings used in the friction-testing apparatus. coefficients for selected archwire-bracket slot combinations in
the dry and wet states. Angle Orthod 1991;61:293–302.
12. Jost-Brinkmann P, Miethke R-R. Einfluß der physiologischen
Zahnbeweglichkeit auf die Friktion zwischen Bracket und Bo-
REFERENCES gen. Fortschr Kieferorthop 1991;52:102–109.
13. Articolo LC, Kusy RP. Influence of angulation on the resistance
1. Nanda R, Ghosh J. Biomechanical considerations in sliding me- to sliding in fixed appliances. Am J Orthod Dentofacial Orthop
chanics. In: Nanda R, editor. Biomechanics in clinical orth- 1999;115:39–51.
odontics. Philadelphia: Saunders; 1997. p 188–217. 14. Zufall SW, Kennedy KC, Kusy RP. Frictional characteristics of
2. Frank CA, Nikolai RJ. A comparative study of frictional resis- composite orthodontic archwires against stainless steel and ce-
tances between orthodontic bracket and arch wire. Am J Or- ramic brackets in the passive and active configurations. J Mater
thod 1980;78:593–609. Sci: Mater Med 1998;9:611–620.
RESISTANCE TO SLIDING OF ORTHODONTIC APPLIANCES 811

15. Kusy RP, Schafer DL. Rheology of stimulated whole saliva in a characteristics of ceramic brackets. Am J Orthod Dentofacial
typical pre-orthodontic sample population. J Mater Sci: Mater Orthop 1994;106:76–87.
Med 1995;6:385–389. 28. Hansen JD, Kusy RP, Saunders CR. Archwire damage from
16. Kusy RP, Schafer DL. Effect of salivary viscosity on frictional ceramic brackets via notching. Orthod Rev 1997;11:27–31.
coefficients of orthodontic archwire/bracket couples. J Mater 29. Articolo LC, Kusy K, Saunders CR, Kusy RP. Influence of ce-
Sci: Mater Med 1995;6:390–395. ramic and stainless steel brackets on the notching of archwires
17. Jastrezbski ZD. The nature and properties of engineering ma- during clinical treatment. Eur J Orthod 2000;22:409–424.
terials, 2nd edition. New York: John Wiley & Sons; 1976. p 30. Schlegel V. Relative friction minimization in fixed orthodontic
182–185. bracket appliances. J Biomech 1996;29:483–491.
18. Articolo LC, Kusy RP. Influence of angulation on the resistance 31. Meling T, Ødegaard J, Holthe K, Meling EØ, Segner D. A for-
to sliding in fixed appliances. Am J Orthod Dentofacial Orthop mula for the displacement of an archwire when subjected to a
1999;115:39–51. second-order couple. Am J Orthod Dentofacial Orthop 1998;
19. Zufall SW, Kusy RP. Sliding mechanics of coated composite 113:632–640.
wires and the development of an engineering model for bind- 32. Kusy RP. Unpublished work, as described in NIH-R01 grant
ing. Angle Orthod 2000;70:34–47. application.
20. Kusy RP, Whitley JQ. Coefficients of friction for arch wires in 33. Gilman JJ. Hardness-A strength microprobe. In: Westbrook JH,
stainless steel and polycrystalline alumina bracket slots. I: The Conrad H, editors. The science of hardness testing and its re-
dry state. Am J Orthod Dentofacial Orthop 1990;98:300–312. search applications. Metals Park, Ohio: American Society for
21. Kusy RP, Whitley JQ. Friction between different wire-bracket Metals; 1973. p 51–74.
configurations and materials. Sem Orthod 1997;3:166–177. 34. Rice RW. Correlation of hardness with mechanical effects in
22. Kusy RP, Whitley JQ. Effect of surface roughness on the coef- ceramics. In Westbrook JH, Conrad H, editors. The science of
ficients of friction in model orthodontic systems. J Biomech hardness testing and its research applications. Metals Park,
1990;23:913–925. Ohio: American Society for Metals; 1973. p 117–134.
23. Kapila S, Angolkar PV, Duncanson MG, Nanda RS. Evaluation 35. Drake SR, Wayne DM, Powers JM, Asgar K. Mechanical prop-
of friction between edgewise stainless steel brackets and orth- erties of orthodontic wires in tension, bending, and torsion.
odontic wires of four alloys. Am J Orthod Dentofacial Orthop Am J Orthod 1982;82:206–210.
1990;98:117–126. 36. Burstone CJ, Goldberg AJ. Beta titanium: a new orthodontic
24. Kusy RP, Saunders CR, Whitley JQ. Improving arch mechanics alloy. Am J Orthod 1980;77: 121–132.
through surface chemistry. In: Nanda R, editor. Biomechanics 37. Hill R. The mathematical theory of plasticity. Oxford, United
in clinical orthodontics. Philadelphia: Saunders; 1997. p 50–64. Kingdom: University Press; 1967. p 97–127.
25. Kusy RP, Greenberg AR. Effects of composition and cross sec- 38. Larson BE, Kusy RP. Torsional elastic property measurements
tion on the elastic properties of orthodontic archwires. Angle of selected orthodontic archwires. Clin Mater 1987;2:165–179.
Orthod 1981;51:325–341. 39. Shigley JE. Mechanical engineering design. New York: Mc-
26. Kusy RP, Dilley GJ, Whitley JQ. Mechanical properties of stain- Graw-Hill;1963. p 106.
less steel orthodontic archwires. Clin Mater 1988;3:41–59. 40. Greener EH, Harcourt JK, Lautenschlager EP. Materials science
27. Saunders CR, Kusy RP. Surface topography and frictional in dentistry. Baltimore: Williams & Wilkins;1972. p 47.