ONLINE ONLY

Mechanical effects of third-order movement

in self-ligated brackets by the measurement
of torque expression
Thomas W. Major,a Jason P. Carey,b David S. Nobes,c Giseon Heo,d and Paul W. Majore
Edmonton, Alberta, Canada

Introduction: Axial rotation of orthodontic wire produces buccal or lingual root movement and is often referred to
as third-order movement or “torque expression.” The objective of this study was to quantify torque expression in
3 self-ligation bracket systems (Damon Q, Ormco, Orange, Calif; In-Ovation R, GAC, Bohemia, NY; and Speed,
Strite Industries, Cambridge, Ontario, Canada) during loading and unloading. Methods: A stepper motor was
used to rotate a wire in a fixed bracket slot from –15
to 63
in 3
increments, and then back to –15
. The bracket
was mounted on top of a load cell that measured forces and moments in all directions. Results: Damon’s and
In-Ovation’s maximum average torque values at 63
were 105 and 113 Nmm, respectively. Many Speed
brackets experienced premature loss of torque between 48
and 63
, and the average maximum was 82
Nmm at 54
. The torque plays for Damon, In-Ovation, and Speed were 11.3
, 11.9
, and 10.8
, respectively.
Conclusions: Generally, In-Ovation expressed the most torque at a given angle of twist, followed by Damon
and then Speed. However, there was no significant difference between brackets below 34 Nmm of torque.
From a clinical perspective, the torque plays between brackets were virtually indistinguishable. (Am J Orthod
Dentofacial Orthop 2011;139:e31-e44)

E
dgewise orthodontic treatment involves rectan-
gular wires placed into rectangular bracket slots.
Axial rotation of the orthodontic wire in the
bracket creates a force couple that produces buccal or
lingual root movement relative to the tooth crown.1 In
the orthodontic literature, this type of tooth movement
is often called “root torque” or third-order movement.
Also, the terms “torque” and “torque expression” refer
to the physical moment generated in the bracket in
newton millimeters (Nmm).
From the University of Alberta, Edmonton, Alberta, Canada.
a
Research associate, Department of Mechanical Engineering, Faculty of
Engineering.
b
Associate professor, Department of Mechanical Engineering, Faculty of
Engineering.
c
Assistant professor, Department of Mechanical Engineering, Faculty of
Engineering.
d
Assistant professor, Orthodontic Graduate Program, Department of Dentistry,
Faculty of Medicine and Dentistry.
e
Professor, and program director, Orthodontic Graduate Program, Department of
Dentistry, Faculty of Medicine and Dentistry.
The authors report no commercial, proprietary, or financial interest in the prod-
ucts or companies described in this article.
Reprint request to: Paul W. Major, Room 4051, Dentistry/Pharmacy Centre, Fac-
ulty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada
T6G 2N8; e-mail, major@ualberta.ca.
Submitted, December 2009; revised and accepted, April 2010.
0889-5406/$36.00
Copyright Ó 2011 by the American Association of Orthodontists.
doi:10.1016/j.ajodo.2010.04.029
Torque expression is a function of wire properties,
bracket slot dimension and bracket design, archwire
dimension, and degrees of wire twist relative to the
bracket slot.1-8 The angle in degrees that the wire is
twisted is called the “angle of twist” or the “torque
angle.” The “zero position” is the position defined as
having an angle of 0
where the wire must twist an
equal angle in the positive and negative directions to
engage the bracket slot walls. The term “torque play”
is the angle at which the wire just engages the slot,
meaning that the “torque play region” is the range of
angles at which there is approximately no torque
expression. This is also called the “engagement angle.”
Torque expression is also influenced by the ligation
method.9 In passive ligation, the wire is free to rotate
in the slot until the edges of the rectangular wire contact
the sides of the bracket slot. As the wire is twisted to the
limit of the torque play region, a force couple is gener-
ated. In active ligation, the wire is pressed against the
base of the slot. The interaction of the active ligation
method (wire, ligature, or active bracket door) creates
a second force couple. In the case of self-ligation
brackets where a clip presses against the wire, the force
can act on the edge of the wire and alter the zero
position. As the wire rotates, the interaction of the clip
against the wire might contribute to torque expression.
e31
e32 Major et al

Previous studies have evaluated torque expression

and torque play for a variety of bracket and wire types.3,9
Some authors have acknowledged the potential role of
bracket deformation, but to date the effects of wire or
bracket elastic and plastic deformations have not been
quantified.2,4,10 Elastic deformation is a nonpermanent
deformation. Plastic deformation occurs under higher
loads than elastic deformation and is permanent. To
date, no published studies have evaluated both the
loading and unloading torque values associated with
archwire twist and return to zero position.
The objective of this study was to quantify torque
expression in 3 self-ligation bracket systems during
loading (“increasing angle”) and unloading (“decreasing
angle”). The loading and unloading curves will be used
to characterize the combined bracket and wire defor-
mations.

MATERIAL AND METHODS

The torque measurement device, previously used by
Badawi et al9 and modified for this study, is shown in
Figure 1. The key assemblies for measuring torque and
angle are the wire and motor assembly and the load
cell assembly. The modified load cell assembly consists
of 2 translational stages and 1 rotational stage (Thor-
labs, Newton, NJ), a 6-axis load cell (ATI Industrial
Automation Nano 17 Multi-Axis force/torque trans-
ducer, Apex, NC), and the bracket mount on which sits
the orthodontic bracket. Through the mount, the load
cell measures forces and moments at the bracket slot.
By using a FaroArm (FARO USA, Lake Mary, Fla) to de-
termine the location of the bracket slot in relation to
the designed load cell origin, a transformation is applied
to the data so that the final data set of forces and
moments are those at the slot of the bracket. By using
the left-hand rule for torque (T) direction, the transfor-
mation equation is

tor was controlled by custom software, and the angle
Tx 5Tx 0  Fy0 3Dz 1ðFz0 3DyÞ
where the x direction is the parallel to the long axis of the
wire and bracket slot, the z direction is vertical against
the bracket base, and the y direction is perpendicular
to both x and z, with the origin at the bracket slot center.
X0 , y0 , and z0 are the equivalent coordinates, but with the
origin located at the load cell. The measured distances
between the load cell origin and the slot center, Dx,
Dy, and Dz, are shown in Figure 2. The left-hand rule
is used for torque direction in the equation to match
the coordinate calibration output from the load cell.
The 3-dimensional Cartesian (x0 , y0 , and z0 ) force and
moment data are collected through a data acquisition
card (DAC 16-bit E series NI PCI-6033E, National
Fig 1. Apparatus showing the bracket and load cell.
Instruments, Austin, Tex) and logged with commercial
software (LabWindows CVI, National Instruments). The
data are exported to a spreadsheet, and the transforma-
tion is applied so that the final data output of torque is at
the slot center.
The motor and wire assembly consists of 2 dies with
rectangular slots that clamp tightly onto 0.019 3 0.025-
in stainless steel orthodontic wire (Ormco, Orange, Calif).
The distance between the 2 fixing dies was 15 mm. The
entire wire assembly was rotated by a worm gear stepper
motor (Cool Muscle CM1-C-11L30, Myostat Motion
Control, Newmarket, Ontario, Canada). The stepper mo-
was accurately measured by using the motor’s internal
control loop at any time.
After the wires were tightened into the dies, they were
dropped into the bracket slot at the initially estimated
zero position. By using the translation and rotation
stages, the forces were zeroed within 0.01 N in the y
and z directions by using the 2 translational stages,
and the moments were zeroed to within 0.08 Nmm in
the z direction by using the rotational stage. Preload
forces in the x direction cannot be easily controlled
and reach a maximum of 0.3 N. However, the x-direction
forces have no direct impact on Tx, as can be seen in
the equation. Preload torque in the y direction cannot
be controlled, because it depends on the bracket

January 2011  Vol 139  Issue 1 American Journal of Orthodontics and Dentofacial Orthopedics
Major et al
Fig 2. Coordinate systems and origins of the load cell and bracket slot.
manufacturing tolerances and how the adhesive sets
between the bracket and the mounting cylinder. Since
Ty does not appear in the equation, it is assumed to
have little impact on Tx. Because it is not possible to ac-
curately determine the zero position, the wire is rotated
into the starting position, –15
from the estimated zero
position, followed by a 78
rotation forward from this
position to a maximum of 63
and then returned to
the starting position. Force and moment data are logged
every 3
during the tests; data are thus collected
between –15
and 63
.
Three types of 0.022-in slot self-ligating brackets were
used: Damon Q (Ormco), In-Ovation-R (GAC, Bohemia,
NY), and Speed (Strite Industries, Cambridge, Ontario,
Canada). The sample size was determined from previous
data published by Badawi et al9 to be 30 for each bracket
type. The brackets were numbered and randomly tested.
A new section of wire was used for each test.
In this article, we report standard deviations and use
descriptive statistics in the discussion. The use of these
statistics is supported by testing the data for normality
with the Kolmogorov-Smirnov test. To determine
whether there was a statistically significant difference be-
tween brackets, a statistical package (version 17.0, SPSS,
Chicago, Ill) was used to carry out analysis of variance
(ANOVA). Since the equal variance assumption for
ANOVA was violated, the test statistics were obtained
by using the Brown-Forsythe and Welch methods at
American Journal of Orthodontics and Dentofacial Orthopedics
e33
each angle. If the P value based on the Brown-Forsythe
and Welch methods was greater than 0.05, all compari-
sons among the brackets at that angle were considered
to have no statistically significant difference. The post-
hoc multiple comparisons were performed at the angles
where the statistical difference was detected from ANOVA
at the significance level of 0.05. There were 5 post-hoc
tests to make pairwise comparisons among 3 types of
brackets, and the significance level was determined as
0.05/5 5 0.01. That is, comparisons with P values less
than 0.01 were considered statistically significant.
RESULTS
Torque vs wire twist angles for the Damon, In-
Ovation R, and Speed brackets are shown in Figure 3.
Each experiment started at –15
on the line labeled “in-
creasing angle.” At each 3
increment, a new data point
was taken as the angle was increased. At 63
, the angle
reached a maximum and then was decreased, with tor-
que measurements taken every 3
. The line of decreasing
angle, which for all 3 brackets had a lower torque than
the increasing line, also took points every 3
from 63

to –15
. The error bars represent 6 1 SD (68% CI). The
shape of the curves was similar between Damon (Fig 3,
a) and In-Ovation R (Fig 3, b). The average torque
expression for Speed brackets reached a plateau at ap-
proximately 50
of positive wire twist and then declined.
January 2011  Vol 139  Issue 1
e34 Major et al

120

100
Torque (Nmm) 80

60
40
20
0
-20 -10 0 10 20 30 40 50 60 70
-20
-40

-60
Angle (Degrees)
Increasing Angle Decreasing Angle

(a)
120
100

80
Torque (Nmm)

60

40
20
0
-20 -10 0 10 20 30 40 50 60 70
-20
-40
-60
Angle (Degrees)
Increasing Angle Decreasing Angle

(b)
120
100
80
Torque (Nmm)

60
40
20
0
-20 -10 0 10 20 30 40 50 60 70
-20
-40
-60
Angle (Degrees)
Increasing Angle Decreasing Angle

(c)

Fig 3. Torque expression vs angle of wire twist with 1 SD error bars for (a), Damon Q; (b), In-Ovation R;
and (c), Speed.

A number of Speed (Fig 3, c) brackets (16 of 30)
reached maximum torque before 63
, usually followed
within a few degrees by a sudden loss of torque.
Figure 4 shows an example of this phenomenon for 1
bracket. Table I shows the value of maximum torque
for each bracket that had a premature torque peak.
This was the reason that the standard deviation
increased substantially in Figure 3, (c), at a torquing
angle of 48
to 63
. Figure 5 shows the Speed torque
expression for the 14 brackets that did not have a prema-
ture torque peak and, presumably, whose bracket doors
did not open. Speed brackets that do not have a prema-
ture maximum torque have similar torque curves as
Damon and In-Ovation R brackets. Unless otherwise

January 2011  Vol 139  Issue 1 American Journal of Orthodontics and Dentofacial Orthopedics
Major et al e35

90
80
Torque (Nmm) 70
60
50
40
30
20
10
0
-20 -10 -10 0 10 20 30 40 50 60 70

-20
Angle (Degrees)
Increasing Angle Decreasing Angle

Fig 4. An individual test on a Speed bracket that expressed a premature and sudden loss of torque.

from the initial Welch or Brown-Forsythe test were re-

Table I. Speed premature torque peaks
moved from the comparison graphs. The significant cut-
Torque off for the individual comparisons was P 5 0.01, and the
Bracket Angle peak line was again shown as a horizontal, with anything
number (
) (Nmm)
above the line having no statistically significant mean
3 54 81.5
7 57 83.9 difference.
8 54 91.8 To determine torque play for each bracket type,
9 45 57.6 a line was fitted by using linear regression. The angle
11 57 92.3 of intersection of the line with the x-axis was calcu-
15 48 68.4
lated for both lines. Subtracting these intersection
18 48 69.2
19 51 78.4 points and dividing by 2 gave the calculated torque
20 48 78.7 play in degrees. For example, Figure 9, a, shows a fitted
21 51 81.4 line for selected points of the Speed bracket tests. The
22 48 80.6 intersections with the x-axis were calculated at –11.2

23 54 87.0
and 10.3
. The difference between these points divided
24 45 64.5
25 48 72.7 by 2 was 10.8
, which is the torque play for Speed.
28 60 88.2 Similarly, Figure 9, b and c, illustrates the torque
29 51 78.3 play, which for Damon can be calculated as 11.3

and for In-Ovation as 11.9
. The theoretical torque
play was calculated by using the formulas from Meling
indicated, the 14 Speed brackets with no premature et al5 as 7.2
, assuming a 0.022-in slot and an ideal-
torque peak were called “Speed-NTP,” and the full ized rectangular 0.019 3 0.025-in wire.
sample of 30 brackets was called simply “Speed.”
Figure 6 compares the averages of all 3 bracket types DISCUSSION
and includes both Speed and Speed-NTP. A complete This study was designed to evaluate torque expres-
listing of the numeric torque values and standard devi- sion during loading and unloading of third-order move-
ations for each bracket type is presented in Table II. ment (rotation of the wire in the bracket). Previous
Figure 7 shows the results from the Welch and published studies focused on the loading torque expres-
Brown-Forsythe statistical tests for increasing and de- sion and did not report the unloading torque expression.
creasing angles. The 0.05 statistically significant differ- In clinical practice, the wire is engaged into the bracket,
ence cutoff is shown by a horizontal line; anything and the unloading torque expression characteristics re-
above 0.05 indicates that the mean difference compari- sult in tooth movement. At the start of the experiment,
sons were not statistically significant. This was used as it was unknown whether the force moment would be
an initial screening of statistically relevant mean differ- zero when the wire was rotated back to the estimated
ences. Presented in Figure 8 are the individual compar- zero position. For this reason, the data are presented
isons between brackets for increasing and decreasing with start and end points of –15
from the estimated
angles. Any points that were not statistically significant zero position.

American Journal of Orthodontics and Dentofacial Orthopedics January 2011  Vol 139  Issue 1
e36 Major et al

120
100
Torque (Nmm) 80
60
40
20
0
-20 -10 0 10 20 30 40 50 60 70
-20
-40
-60
Angle (Degrees)
Increasing Angle Decreasing Angle

Fig 5. Torque vs angle of wire twist for the Speed data set with no premature torque peak.

120

100

80
Torque (Nmm)

60

40

20

0
-20 -10 0 10 20 30 40 50 60 70
-20

-40
Angle of Twist (Degrees)
Damon In-Ovation Speed Speed-NTP

Fig 6. Total torque vs angle of wire twist comparisons between all bracket types.

The magnitude of third-order movement used in this
experiment exceeds what would be used for a 0.019 3
0.025-in stainless steel wire in clinical practice. The
magnitude of wire rotation was chosen to fully investi-
gate the mechanical effects of third-order movement.
Torque expression of approximately 100 Nmm was re-
corded at 63
of wire rotation. Future research is needed
to evaluate loading and unloading torque expression in
a clinically relevant range of third-order movement.
Graph shape and standard deviations
In-Ovation and Damon brackets both have a similar
shape and magnitude. Both show standard deviations
with a maximum between 33
and 39
. This corresponds
to 56.5 to 75.4 Nmm for Damon and 59.0 to 79.3 Nmm
for In-Ovation. It is possible that, at approximately 33
,
the wire begins to plastically yield and strain harden.
However, the predictability and nature of strain harden-
ing depends on the initial wire alignment in the slot and
the manufacturing tolerances in wires and brackets. The
degree of variation in these factors has an especially
large impact on the torque expression while the wire is
strain hardening. The stainless steel wire yielded at ap-
proximately 3.2 radians in the study by Vena et al11;
this corresponds to 0.29 radians or 17
with the sample
length used in our study. Adding 11
on this for torque
play, the estimated angle of wire deformation was ap-
proximately 28
. The difference between this and the
observed yield angle of 33
would be largely due to
the bracket’s elastic deformation. After 39
, it could be
assumed that the wire has strain hardened enough that
its material characteristics are more predictable, and so
the standard deviation decreases, even though torque
continues to increase. At angles above 40
, torque
does not maintain its linear trend with the angle of twist.
This is due to the plastic deformation of the bracket, ad-
ditional wire deformation, edge beveling, and many
other factors.

January 2011  Vol 139  Issue 1 American Journal of Orthodontics and Dentofacial Orthopedics
Major et al e37

Table II. Torque data (SD)

Torque (Nmm)


Angle ( ) Damon In-Ovation Speed Speed-NTP
15 6.5 (2.9) 7.7 (2.3) 9.1 (4.1) 8.4 (4.7)
12 0.9 (2.1) 1.3 (2.2) 1.9 (2.4) 1.5 (2.6)
9 1.5 (1.0) 1.6 (1.6) 2.5 (1.3) 2.6 (0.9)
6 0.7 (0.7) 1.2 (1.0) 2.3 (0.8) 2.2 (0.7)
3 0.3 (0.4) 0.0 (0.9) 0.8 (0.7) 0.7 (0.6)
0 0.2 (0.4) 0.2 (0.9) 0.2 (0.7) 0.1 (0.6)
3 0.2 (0.3) 0.4 (0.7) 0.0 (0.6) 0.2 (0.5)
6 0.2 (0.5) 1.4 (0.6) 0.6 (0.6) 0.4 (0.5)
9 0.8 (1.2) 1.9 (0.8) 1.5 (0.9) 1.1 (0.6)
12 4.6 (1.6) 3.8 (1.1) 4.4 (2.1) 3.6 (1.5)
15 9.6 (2.3) 8.0 (1.7) 9.7 (3.4) 8.5 (3.2)
18 16.7 (3.3) 15.7 (2.0) 16.6 (4.5) 15.0 (5.0)
21 24.2 (4.3) 24.1 (2.2) 23.3 (5.2) 21.6 (6.4)
24 33.4 (4.9) 33.8 (2.3) 31.2 (5.7) 29.7 (7.4)
27 41.9 (5.3) 42.8 (2.5) 38.2 (6.2) 37.1 (8.1)
30 50.1 (5.5) 51.7 (2.6) 44.9 (6.6) 44.3 (8.7)
33 56.5 (5.8) 59.0 (2.7) 49.8 (6.8) 49.5 (9.1)
36 64.9 (5.8) 68.3 (2.7) 56.7 (7.1) 57.0 (9.3)
39 75.4 (5.6) 79.3 (2.7) 65.6 (7.4) 66.8 (9.3)
42 81.5 (5.3) 86.3 (2.5) 70.4 (7.4) 72.3 (8.8)
45 86.9 (5.0) 92.7 (2.4) 74.7 (7.5) 77.4 (8.0)
48 92.3 (4.6) 98.9 (2.2) 79.4 (8.6) 83.4 (7.4)
51 95.5 (4.2) 102.8 (2.1) 81.4 (10.7) 87.5 (6.8)
54 98.3 (3.9) 106.0 (2.0) 81.5 (15.2) 91.5 (6.2)
57 100.5 (3.6) 108.5 (2.0) 79.6 (21.7) 95.0 (5.7)
60 103.1 (3.4) 111.1 (2.0) 73.3 (32.9) 99.1 (5.4)
63 104.5 (3.2) 112.7 (2.1) 59.6 (43.4) 101.5 (6.0)
63 102.8 (3.2) 111.0 (2.1) 58.7 (43.0) 100.4 (5.8)
60 91.0 (3.3) 99.1 (1.9) 49.1 (39.8) 87.7 (5.7)
57 78.7 (3.3) 86.7 (1.8) 38.8 (37.0) 74.9 (5.6)
54 67.6 (3.3) 75.1 (1.7) 31.2 (32.2) 63.0 (5.4)
51 57.2 (3.2) 64.2 (1.7) 25.6 (26.5) 51.9 (5.3)
48 47.7 (3.0) 53.8 (1.7) 20.8 (20.8) 41.6 (5.1)
45 38.5 (3.1) 43.7 (1.8) 16.7 (15.8) 32.5 (4.7)
42 30.3 (2.9) 34.7 (1.8) 13.2 (11.3) 24.4 (4.5)
39 23.3 (2.7) 26.5 (1.7) 10.1 (7.7) 17.4 (4.1)
36 14.5 (2.5) 16.78 (1.8) 5.8 (4.3) 9.4 (3.3)
33 8.4 (2.6) 9.3 (1.7) 3.9 (2.6) 5.4 (2.7)
30 6.7 (1.8) 6.5 (1.5) 5.9 (2.2) 7.0 (2.3)
27 2.8 (1.3) 2.0 (1.3) 5.1 (1.6) 5.4 (2.0)
24 1.0 (1.1) 1.0 (1.2) 4.8 (1.6) 4.8 (2.1)
21 0.3 (0.7) 0.1 (1.2) 3.7 (1.5) 3.6 (2.0)
18 0.0 (0.5) 0.6 (1.3) 3.5 (1.6) 3.1 (2.1)
15 0.1 (0.4) 1.1 (1.3) 2.8 (1.6) 2.2 (2.0)
12 0.1 (0.3) 1.0 (1.4) 2.6 (1.7) 2.3 (2.2)
9 0.1 (0.4) 1.6 (1.6) 1.9 (1.7) 1.8 (2.2)
6 0.3 (0.7) 0.6 (1.5) 1.9 (1.7) 2.1 (2.2)
3 1.2 (0.8) 0.1 (1.5) 2.3 (1.6) 2.3 (2.3)
0 0.5 (1.5) 0.4 (2.2) 3.2 (1.5) 3.2 (2.1)
3 5.1 (2.8) 6.1 (2.5) 3.1 (1.5) 2.8 (2.1)
6 11.4 (3.7) 13.0 (2.6) 2.8 (2.0) 2.1 (2.6)
9 18.6 (4.9) 20.9 (2.7) 0.4 (4.1) 1.8 (4.8)
12 26.5 (5.9) 29.5 (2.8) 5.6 (5.9) 8.8 (7.0)
15 34.3 (6.8) 37.8 (2.8) 13.6 (7.1) 16.9 (8.9)

American Journal of Orthodontics and Dentofacial Orthopedics January 2011  Vol 139  Issue 1
e38 Major et al

0.6

0.5

0.4

0.3 Welch
p-value

Brown-Forsyth

0.2 0.05 Significant Cut-off

0.1

0
-15 -5 5 15 25 35 45 55

-0.1
Angle (Degrees)

0.35

0.3

0.25

0.2

Welch
p-value

0.15 Brown-Forsyth
0.05 Significant Cut-off

0.1

0.05

0
-15 -5 5 15 25 35 45 55

-0.05
Angle (Degrees)

Fig 7. Welch and Brown-Forsythe significance of torque while the angle is increasing and while the
angle is decreasing.

When the angle is decreasing, the torque expression is
substantially less than when the angle is increasing, as
seen in Figures 3 and 6. The permanent plastic
deformation of the wire into a new effective angle of
twist accounts for most of the difference. The plastic
deformation of the bracket, although probably not as
substantial as wire deformation due to the wire’s smaller
cross-sectional area, could additionally reduce the torque
expression while the angle is decreasing. Since the wire
was permanently twisted, the range of torque play was
also different. While the angle was decreasing, torque
play was between 0
and 24
, even though at 24
while
the angle was increasing it expressed 33 to 34 Nmm of
torque in both the Damon and In-Ovation brackets. The
wire produced negative torque in the brackets at all angles
below 0
while the angle was decreasing. At the final angle
of –15
, the decreasing angle data point was expressing
–34 to –38 Nmm of torque in the Damon and In-
Ovation brackets, even though the first data point when
the angle was increasing at –15
expressed only –6 to
–8 Nmm. This is also explained by the wire’s permanent
deformation into a different twist angle. Even though

January 2011  Vol 139  Issue 1 American Journal of Orthodontics and Dentofacial Orthopedics
Major et al e39

1.2
Damon - In-Ovation
Damon - Speed
1
Damon - SpeedNTP
In-Ovation - Speed
0.8
In-Ovation - SpeedNTP
0.01 Significance Cutoff
0.6
p-value

0.4

0.2

0
-15 -5 5 15 25 35 45 55

-0.2
Angle (Degrees)

1.200
Damon - In-Ovation
Damon - Speed
1.000 Damon - SpeedNTP
In-Ovation - Speed
In-Ovation - SpeedNTP
0.800
0.01 Significance Cutoff

0.600
p-value

0.400

0.200

0.000
-15 -5 5 15 25 35 45 55

-0.200
Angle (Degrees)

Fig 8. Significance comparison of torque expression between individual bracket types with (A),
increasing angle, and (B), decreasing angle.

the maximum angle of twist used in this experiment pro-
duced deformation far beyond clinical relevance, it would
still be prudent for a clinician to be aware of the potential
impacts and losses of torque expression when the wire or
bracket is plastically deformed.
On average, the standard deviation for Damon was
1.4 times higher than that for In-Ovation. Damon’s
maximum standard deviation was 5.8 Nmm, and In-
Ovation’s was 2.7 Nmm; they were both in the same
range of angles of 33
to 36
while the angle was in-
creasing. The difference between the brackets could be
due to a number of factors. Differences in slot shape
and overall bracket dimensions affect stress distributions
and the nature of engagement between the slot and
wire; these directly impact torque expression.10 There-
fore, discrepancies between the manufacturing repeat-
ability of In-Ovation and Damon brackets could
account for part of the difference in standard deviations.
Also in the experimental procedure, all brackets were
initially zeroed at a point estimated to be in the torque
play region. Due to In-Ovation’s active ligation system,
there was high repeatability of the zero force-moment

American Journal of Orthodontics and Dentofacial Orthopedics January 2011  Vol 139  Issue 1
e40 Major et al

70

60

50

40
Torque (Nmm)

30

-11.2° 20
10.3°

10

0
-20 -10 0 10 20 30 40
-10

-20 Angle (Degrees)

(a)

70

60

50

40
Torque (Nmm)

30

20

10

0
-20 -10 0 10 20 30 40
-10

-20
Angle (degrees)
(b)

80

70

60

50
Torque (Nmm)

40

30

20

10

0
-20 -10 0 10 20 30 40
-10

-20 Angle (degrees)

(c)

January 2011  Vol 139  Issue 1 American Journal of Orthodontics and Dentofacial Orthopedics
Major et al
In-Ovation
Active door puts
a force on wire
Wire will always
go to this
location in the
slot due to the
active force from
door
Fig 10. In-Ovation and Damon preloading scenarios.
position. However, the Damon bracket is a passive liga-
tion system; therefore, the wire is free to move in the
bracket slot. The forces and moments can read zero
even though the position of the wire in the bracket
slot is not necessarily as repeatable. This concept is illus-
trated in Figure 10.
Since 16 of 30 samples of Speed brackets had prema-
ture loss of torque expression, the standard deviation
between Speed tests rose substantially above 48
of
twist. Presumably, this was because the bracket door
partially opens at the peak and then completely opens
at the sudden drop. The standard deviations for the
full sample of 30 and for the smaller sample of 14
were within approximately 20% of each other until
48
. When the 14 brackets that did not have a torque
peak were averaged, the torque vs angle graph had a sim-
ilar shape as those of Damon and In-Ovation. However,
its standard deviations were on average 1.8 times those
of Damon and 2.2 times those of In-Ovation. Consider-
ing the higher standard deviation and lower torque ex-
pression of the Speed and Speed-NTP brackets
through all angles, it could be assumed that these
brackets are made of lower modulus of elasticity mate-
rial, or their design lends itself more easily to plastic
deformation.
All experiments were carried out at room tempera-
ture. Nickel-titanium has a higher modulus of elasticity
=
Fig 9. Torque play calculations for (a), Speed; (b), Damon Q; and (c), In-Ovation R.
American Journal of Orthodontics and Dentofacial Orthopedics
e41
Damon
Passive door does not
put a force on wire
Since passive door
does not put load
on wire, it can have
a zero preload
while moving
anywhere in the
slot space.
at higher temperatures, meaning that the door of the
Speed brackets might react differently at typical mouth
temperatures.12,13 The door provides only a small force
to the top of the wire. The actual torque expression
comes almost entirely from the engagement between
the wire and the slot walls. Therefore, since the door is
taking only a minimal part of the load, the door
material will have a small effect on torque expression,
especially at higher angles. The primary purpose of the
active door is to decrease torque play at lower angles.9
As for the effect this could have on the door opening
prematurely, this depends on the door-opening mecha-
nism. If the bracket door opens as a direct result of the
deformation of the bracket wings, the door material is
irrelevant. Future studies are required with the Speed
bracket at mouth temperature to see whether this has
an impact on its torque expression.
Badawi et al9 measured the torque expression of 2
active ligation and 2 passive ligation brackets from
0
to 57
with increasing angles. Included in that study
were the measurements of the same models of Speed
and In-Ovation R brackets as we used, as well as an older
model of Damon brackets. Badawi et al9 reported the
standard deviations at 12
, 24
, 36
, and 48
. The Da-
mon and In-Ovation brackets they used had standard
deviations 2 to 8 times larger than in our study. Speed’s
standard deviations from their study were comparable
January 2011  Vol 139  Issue 1
e42
with the Speed standard deviations in our experiment at
12
, 24
, and 36
, but, at 48
, the standard deviations
were 33% worse than our Speed data (with 30 samples)
and 20% worse than our Speed-NTP data.
The improvements in standard deviations between
this study and that of Badawi et al9 could be partially
accounted for by the recent addition of translation and
rotation stages to improve the pretorque zeroing proce-
dure. In addition, Badawi et al did not indicate the
method of measuring the 3-dimensional distance be-
tween the bracket and the load cell origin, and the sub-
sequent mathematical adjustment of torques from
internal forces over that distance. This would result in
substantially different magnitudes and also presumably
contribute to the variations between the findings. We
used the FaroArm to measure the distances and apply
the required mathematical translation to determine
and report the torque at the bracket.
Comparison of torque magnitudes
The Welch and Brown-Forsythe tests showed that the
mean differences between data sets were generally not
as statistically significant at angles less than 25
than
between 25
and 63
while the angle was increasing.
Therefore, up to approximately 40 Nmm of torque, all
brackets statistically had no difference. Starting at 27
,
the P values for comparisons between Speed and other
brackets rapidly decreased, and this corresponded to
Speed’s expressing notably less torque. By 38
, all
brackets had statistically significant differences. The
comparison between Damon and Speed-NTP lost statis-
tically relevant significance as the P values rose to 0.65,
largely due to the increasing standard deviations of the
Speed-NTP brackets at higher angles of twist. This was
likely because of the greater amount of plastic deforma-
tion in the Speed brackets. Because the cutoff P value for
each test was largely arbitrary, the general trend was that
the brackets did not have statistically significant differ-
ences below 25
of twist (approximately 35 Nmm).
Up to 33
, the magnitude of In-Ovation’s torque ex-
pression was generally within 1 SD of Damon’s. From
33
to 63
, In-Ovation brackets had on average 7%
greater torque expression than Damon. When the angle
was decreasing, In-Ovation had 10% to 15% greater
torque expression than Damon; this was probably the
result of less plastic deformation to the In-Ovation
bracket. Up to 24
, the torque expression of Speed-
NTP brackets was also within 10% of the magnitude
of In-Ovation and Damon. By using descriptive statistics,
at 33
, only 15% of Speed-NTP brackets expressed more
torque than the average In-Ovation bracket, and 33%
expressed more torque than the average Damon bracket.
At 63
(maximum angle), Damon produced 93% and
January 2011  Vol 139  Issue 1 American Journal of Orthodontics and Dentofacial Orthopedics
Major et al
Speed-NTP 90% of In-Ovation’s torque. Also at this
angle, only 3% of Speed-NTP brackets expressed more
torque than the average In-Ovation brackets, and only
29% expressed more torque than the average Damon
bracket. Essentially, In-Ovation had the greatest torque
expression, followed by Damon and then Speed and
Speed-NTP. But Speed and Speed-NTP had high
variability between brackets.
The torque expression magnitudes of Speed brackets
reported by Badawi et al9 were substantially different
than presented here. At 36
, for example, the magnitude
found by Badawi et al was 22 Nmm; in our experiment, it
was 57 Nmm, a discrepancy of a factor of 2.6. The In-
Ovation and Damon comparisons between data collec-
tions had a comparable discrepancy. These substantial
differences in magnitude were primarily because Badawi
et al9 reported torque values at the load cell, and not at
the bracket, as discussed above.
Morina et al4 also measured third-order torque ex-
pression in newton millimeters, but only at 20
. They
reported torque values of 7.8 and 8.0 Nmm and stan-
dard deviations of 7.6 and 8.4 Nmm for Damon2 and
Speed brackets, respectively. In our experiment, the
Damon Q and Speed brackets expressed 21.7 and
21.1 Nmm of torque with standard deviations of 4.0
and 5.0 Nmm, respectively. The data in our experiment
had lower standard deviations. However, the difference
in torque magnitudes was again a result of the point of
torque measurement. Morina et al measured torque at
an estimated center of resistance 10 mm from the
bracket, whereas we measured the torque at the
bracket slot.
Huang et al14 generated a finite element model of
Damon MX and Speed brackets in torque. At 20
of
twist with 0.019 3 0.025-in stainless steel wire, they
reported torque values of approximately 55 and 15
Nmm for Damon MX and Speed, respectively. In our
study, Damon Q and Speed had 21.7 and 21.1 Nmm,
respectively, at 20
of twist (by interpolation). How-
ever, statistically, Damon Q and Speed were not differ-
ent at this angle. Finite element models are sensitive to
boundary conditions, dimensions, material properties,
and convergence methods. A small incorrect assump-
tion can lead to substantially different results com-
pared with what is physically observed. The model of
Huang et al was the first of its kind, and, like most
novel finite element models, experimental validation
and possibly further refinement are necessary to repre-
sent physical reality.
Torque play
The torque play of all 3 brackets was within 1.1
.
However, there were limitations in the methodology of
Major et al
calculating torque play. Points that are not part of the
linear range were discarded from the line generated by
linear regression. This was particularly noticeable be-
tween 6
and 12
with In-Ovation and Speed. In-
Ovation used points in the 15
to 33
range to develop
the line above 0
. However, In-Ovation produced 1 to
4 Nmm of torque in a nonlinear pattern between 6

and 12
, as can be seen in Figure 3, b. This could be con-
sidered the partially engaged range, where the wire was
not fully engaging the slot, but, because of the forces on
the actively ligating bracket door, the twist in the wire
was producing small magnitudes of torque. Although
not as pronounced, Speed had a similar profile with
a nonlinear torque region. At mouth temperatures, the
Speed door would apply more force to the wire in the
partially engaged region and likely have a nonlinear
range from 6
to 12
, which is similar to In-Ovation’s.
The door mechanism for both Speed and In-Ovation is
active on just 1 side of the brackets; this might partially
explain the nonlinear behavior. Damon is a passively
ligating door and appeared to immediately jump from
zero torque in the torque play region to a linearly in-
creasing torque. Therefore, considering the calculation
procedure, the reported torque play for Damon is likely
the most accurate of the 3 brackets.
Clinically significant minimum torque is often con-
sidered in the literature to be 5 Nmm.15 An argument
could be made that the partially engaged range should
be ignored, since it is not clinically significant torque.
Solving the linear equations in Figure 9, the angles at
which the bracket is 5 Nmm for Damon, In-Ovation,
and Speed, respectively, were 13.0
, 14.1
, and 12.6
.
These are all within 1.5
of each other; from a clinical
perspective, this is unnoticeable. Therefore, it could be
concluded that the torque play values for all 3 brackets
are essentially the same.
The theoretically calculated torque play was 7.2
. The
actual torque play was 10.8
to 11.9
for all tested
bracket types. This difference could be the result of sev-
eral factors, including rounding at the corners of the
wires and differences between nominal and actual
wire-slot dimensions.2,3,10
The measurements of torque play by Badawi et al9
were between 7
and 15
, whereas in our experiment
they were between 10.8
and 11.9
for all brackets. Ba-
dawi et al measured torque play by reading off the graph.
However, in our study, it was calculated by using lines
generated by linear regression. This, in addition to the
substantial improvements in the data collection methods
mentioned above, accounts for the differences in re-
ported torque play.
Sebanc et al3 reported the torque play for different
wires and bracket slot combinations. For the 0.019 3
American Journal of Orthodontics and Dentofacial Orthopedics
e43
0.025-in stainless steel wires in 0.022-in Ormco bracket
slots, the theoretical torque play of Sebanc et al was be-
tween 6.8
and 6.9
, whereas their measured torque play
was between 10.9
and 12.2
. In our experiment, the
theoretical torque was 7.2
. The difference between
these theoretical calculations was that Sebanc et al
used measured wire and bracket slot widths, and we
used the nominal values. Our measured torque play
was 10.8
to 11.8
, within the range of the measure-
ments of Sebanc et al.
CONCLUSIONS
The purposes of this experiment were to investigate
the mechanical effects of third-order torque among 3
bracket types and determine torque magnitudes over
a range of angles and torque play. We concluded the fol-
lowing.
1. The collected torque expression data of the different
brackets provided a clear graphic visualization of
a zero torque region, followed by linearly increasing
torque and nonlinearly increasing torque at angles
greater than 40
. Respectively, these presumably
correspond to a region of torque play, a region of
linear elastic deformation of the wire, and, above
40
, a region of plastic deformation of the wire or
bracket. When the angle was decreased from the
maximum, the magnitude for all brackets was sub-
stantially less, primarily due to the plastic deforma-
tion of the wire. Speed brackets presumably had the
bracket door partially or completely opening in 16
of 30 samples, resulting in a substantial loss of tor-
que at high angles of twist.
2. At angles of twist 24
and less, corresponding to 34
Nmm and less, all 3 brackets exhibited similar tor-
que magnitudes with no statistically relevant differ-
ence. At higher angles, In-Ovation had the greatest
torque expression, followed by Damon and then
Speed. Speed’s torque standard deviations were on
average 1.8 and 2.2 times those of Damon and In-
Ovation, respectively. This indicates that Speed has
greater variations in torque expression from 1
bracket to another.
3. The brackets had torque play values of 11.3
, 10.8
, and
11.9
for Damon, Speed, and In-Ovation, respectively.
From a clinical perspective, these differences are
nearly indistinguishable.
REFERENCES
1. Wagner JA, Nikolai RJ. Stiffness of incisor segments of edgewise
arches in torsion and bending. Angle Orthod 1985;55:37-50.
2. Fischer-Brandies H, Orthuber W, Es-Souni M, Meyer S. Torque
transmission between square wire and bracket as a function of
January 2011  Vol 139  Issue 1
e44
measurement, form and hardness parameters. J Orofac Orthop
2000;61:258-65.
3. Sebanc J, Brantley WA, Pincsak JJ, Conover JP. Variability of effec-
tive root torque as a function of edge bevel on orthodontic arch
wires. Am J Orthod 1984;86:43-51.
4. Morina E, Eliades T, Pandis N, Jager A, Bourauel C. Torque expres-
sion of self-ligating brackets compared with conventional metallic,
ceramic, and plastic brackets. Eur J Orthod 2008;30:233-8.
5. Meling TR, Odegaard J, Seqner D. On bracket slot height: a methodo-
logic study. Am J Orthod Dentofacial Orthop 1998;113:387-93.
6. Odegaard J, Meling T, Meling E. An evaluation of the torsional mo-
ments developed in orthodontic applications. An in vitro study.
Am J Orthod Dentofacial Orthop 1994;105:392–400; erratum in
Am J Orthod Dentofacial Orthop 1994;106:218.
7. Meling TR, Odegaard J, Meling EO. On mechanical properties
of square and rectangular stainless steel wires tested in torsion.
Am J Orthod Dentofacial Orthop 1997;111:310-20.
8. Siatkowski RE. Loss of anterior torque control due to variations
in bracket slot and archwire dimensions. J Clin Orthod 1999;33:
508-10.
9. Badawi HM, Toogood RW, Carey JP, Heo G, Major PW. Torque ex-
pression of self-ligating brackets. Am J Orthod Dentofacial Orthop
2008;133:721-8.
January 2011  Vol 139  Issue 1 American Journal of Orthodontics and Dentofacial Orthopedics
Major et al
10. Gioka C, Eliades T. Materials-induced variation in the torque
expression of preadjusted appliances. Am J Orthod Dentofacial
Orthop 2004;125:323-8.
11. Vena A, Carey J, Badawi H. Clinical variability in arch wires: a pre-
liminary study evaluating mechanical and surface characteristics of
two different sized rectangular stainless steel wires. Open Biomed
Eng J 2007;1:13-22.
12. MatWeb material property data (homepage on the Internet).
c1996–c2009 by Automation Creations. Nitinol-NiTi shape mem-
ory alloy: low-temperature phase. Available from: http://www.
matweb.com/search/DataSheet.aspx?MatGUID544afc7d3c6eb48
29bc2df27884fd2d6c. Accessed on October 13, 2009.
13. MatWeb material property data (homepage on the Internet).
c1996–c2009 by Automation Creations. Nitinol-NiTi shape mem-
ory alloy: high-temperature phase. Available from: http://www.
matweb.com/search/DataSheet.aspx?MatGUID5de9dd08433714
f698d513766dccea437. Accessed on October 13, 2009.
14. Huang Y, Keilig L, Rahimi A, Reimann S, Eliades T, Jager A, et al. Nu-
meric modeling of torque capabilities of self-ligating and conven-
tional brackets. Am J Orthod Dentofacial Orthop 2009;136:638-43.
15. Gmyrek H, Bourauel C, Richter G, Harzer W. Torque capacity of
metal and plastic brackets with reference to materials, application,
technology and biomechanics. J Orofac Orthop 2002;63:113-28.