ORIGINAL ARTICLE

Complete orthodontic load systems on teeth in

a continuous full archwire: The role of
triangular loop position
Jie Chen,a Irina Bulucea,b Thomas R. Katona,c and Susan Ofnerd
Indianapolis, Ind, and Cupertino, Calif

Introduction: A novel approach to characterizing orthodontic spring-generated force and moment systems
has been developed. This method allows simultaneous measurement of all 6 force and moment components
acting on a tooth. Methods: A continuous full archwire space-closure technique was simulated, and the
complete force and moment systems acting on the teeth adjacent to the extraction space were measured.
Results and Conclusions: The data showed that, in addition to the intended forces and moments, there are
nontrivial activation-dependent interactions with the other load components, and these complex relation-
ships are affected by the position of the triangular loop. (Am J Orthod Dentofacial Orthop 2007;132:
143.e1-143.e8)

C
rown and root positioning are essential for
achieving esthetic, functional, and stable orth-
odontic results. The necessary tooth movement
and control are derived from the clinician’s ability to
manipulate the force-moment systems produced by
orthodontic appliances. However, the traction gener-
ated by actual clinical appliances has never been
quantified. Thus, in practice, control is achieved empir-
ically and qualitatively.
Experimental1-8 and computational3,9-12 studies, in-
cluding finite element analyses,5,10 have demonstrated
that load control is affected by wire material or cross
section,4,7,8 loop shape and size,2,4,5,9-11 gable bends,1-3,8
interbracket positioning,3,6 and attachment/ligation
method.12,13 Unfortunately, the applicability of these
studies is partly compromised. For example, in many
experimental3-5,8 and most analytical studies, rigid
attachments (clamping) are used. In contrast, in clinical
a
Professor, Mechanical Engineering, Purdue University School of Engineering
and Technology; Orthodontics, Indiana University School of Dentistry, India-
napolis.
b
Private practice, Cupertino, Calif.
c
Associate professor, Orthodontics, Indiana University School of Dentistry;
Mechanical Engineering, Purdue University School of Engineering and Tech-
nology, Indianapolis.
d
Biostatician II, Biostatistics, Indiana University School of Medicine, India-
napolis.
Supported by NIH-R41 DE017025, Indiana University School of Dentistry
Student Research Fund, and the American Association of Orthodontists
Foundation.
Reprint requests to: Thomas R. Katona, Indiana University School of Dentistry,
Department of Oral Facial Development, 1121 West Michigan St., Indianap-
olis, IN 46202-5186; e-mail, tkatona@iupui.edu.
Submitted, June 2006; revised and accepted, October 2006.
0889-5406/$32.00
Copyright © 2007 by the American Association of Orthodontists.
doi:10.1016/j.ajodo.2006.10.016
situations, archwires are secured firmly, not rigidly,
with stainless steel or elastomeric ligature ties. This has
an effect on the generated load system.14 Furthermore,
with few exceptions, previous studies focused primarily
on 2-dimensional (2D) (plane) segmental springs usu-
ally located between 2 isolated teeth,1,2,4,15,16 But,
because clinical appliances are often attached to groups
of teeth via noncoplanar brackets, activation inevitably
creates unintended and (most likely) undesirable load
components.
Our objectives were to demonstrate the instrument
by quantifying the effects of triangular loop activation
and location on the generated load system. There are 4
important departures from typical studies: (1) rather
than a 2D plane simulation, an actual 3-dimensional
(3D) clinical strap-up is simulated; (2) all 6 load
components (vs only 2 force components and 1 moment
component in plane simulations) are measured, (3)
simultaneously (vs independently); and (4) all previous
elements were incorporated into 1 experiment.
MATERIAL AND METHODS
For each tooth, the buccolingual, mesiodistal, and
occlusogingival axes were the local x, y, and z axes,
respectively (Table and Fig 1). Thus, the distal and mesial
brackets adjacent to the extraction space each has a unique
coordinate system aligned so that the x-axes (buccal) are
perpendicular to the plane of their bracket bases, and the
z-axes are in the occlusal direction. Therefore, the paths of
space closure are aligned with their respective ⫹y and ⫺y
axes, approximately 15° apart (only the z-axes of the 2
teeth are parallel).
An instrument (Fig 2) based on a microscope frame
143.e1
143.e2 Chen et al American Journal of Orthodontics and Dentofacial Orthopedics
August 2007

Table. Force and moment nomenclature (Fig 3) and

effect on crown movement (Fig 1)
Mx (second order)
⫹Mx Distal
⫺Mx Mesial
My (third order)
⫹My Buccal
⫺My Lingual
Mz (first order)
⫹Mz Distal crown out
⫺Mz Distal crown in

Fig 2. Instrument designed and built to simultaneously

measure all force and moment components acting on
tooth in a complete arch.

Fig 1. Local coordinate axes defined for 2 teeth in left

maxillary quadrant. Only z-direction (occlusal) is the
same.

was designed and built to measure all 6 load compo-

nents (Fig 3) acting on the tooth of interest (TOI) in a
typodont (US patent 6,120,287). It consists of a load
cell attached to the frame, an x-y-z adjustment mecha-
nism (microscope stage), and a 3D coordinate measuring
system (not shown in Fig 2). The TOI was separated from
the typodont. Then, by using a full-arch impression to
align it to the original occlusion, the TOI was cemented to
the adapter. The state-of-the-art gamma load cell (ATI
Industrial Automation, Apex, NC) can simultaneously
measure, with minute deflections, 3 force (0-65 ⫾ 0.2 N) Fig 3. Spring dimensions and forces (Fx, Fy, Fz) and
and 3 moment (0-5 ⫾ 0.0009 Newton-meter [N-m]) moments (Mx, My, Mz) acting on bracket. Moment
components. By using the measuring system to determine directions (Table) given by right-hand rule convention.
the relative positions of the load cell and the TOI bracket,
the instrument was calibrated to report moment and force D85SDP-INU.1, Kilgore International, Coldwater,
values at the bracket. These readings were displayed Mich). The maxillary canines were removed to simu-
simultaneously on a computer with an ISA bus interface. late incisor segment closure. The typodont base was
For illustrative purposes, a clinical case was repli- fastened to the microscope stage, and the premolar
cated on the maxillary arch of a typodont (model (TOI) was attached to the load cell (Fig 2).
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 132, Number 2
Fig 4. Four loop positions in 12.7-mm interbracket
distance.
Stainless steel triangular closing loops were fabri-
cated from 0.016 ⫻ 0.022-in maxillary small size
preformed archwire (Ormco, Glendora, Calif). Dupli-
cating previous studies, loop geometry consisted of an
equilateral triangle with 8-mm sides (Fig. 3).1 All
springs were preactivated with gable bends at the base
of the triangle. To minimize residual stresses, the
gables were overbent and then adjusted to the desired
15°. The wires were inspected for unwanted third-order
bends, and a template was used to assess symmetry.
With a modified protocol, the springs were heated to
700°F for 10 minutes and then bench cooled.17
The teeth were bonded with 0.018-in Ormco brack-
ets or banded with molar bands. To enable a passive
buccolingual fit of the wire into the brackets, composite
offsets were built onto the buccal surface of teeth as
needed. Elastomeric ties were used to hold the wire in
the brackets. Four loop locations in the 12.7-mm
interbracket distance were considered (Fig 4).
Before engaging the wire, the instrument was ze-
roed. Then, a P1 wire (Fig 4) was ligated into the arch
with the loop legs separated by a 0.07-mm shim, and
both ends were cinched back 90° against the distal
molar tubes (Fig 2). Then, with the shim removed, a set
of recordings, corresponding to zero activation, was
taken. Force and moment recordings were then made at
1.6 and 3.3 mm activations. Activation was accom-
plished by removing the ties distal to the loop, short-
ening the distal leg of the wire by pulling it back, and
then placing appropriate spacers (1.6 or 3.3 mm)
between the cinch bend and the distal side of the tube.
New elastomeric ligature ties were placed, and the
forces and moments on the TOI were recorded. Nine
additional P1 loops were then tested. Then, 10 springs
each in the P2, P3, and P4 groups were measured. Thus,
a total of 40 springs was tested.
By using a full-arch impression of the original
Chen et al 143.e3
occlusion, the premolar and its socket were reattached
with resin into their original positions in the typodont.
The lateral incisor’s socket was sawed off, and the
incisor assumed the role of the TOI. The sequence
described above was then repeated with a new set of 40
springs, 10 each in the P1, P2, P3, and P4 groups.
Statistical analysis
For each tooth type, an analysis of variance
(ANOVA) model with fixed effects for loop location,
activation distance, and the 2-way interaction was used
to model the absolute value of each force. A random
effect for each tooth was included in the model to allow
for the correlation induced by repeated sampling of the
same tooth. Residual plots were examined for violation
of model assumptions of normality and homogeneity of
variance. The Sidak adjustment was used for all pair-
wise comparisons for each force/moment and each
interproximal surface to control type 1 statistical error
(the probability of seeing a significant difference when,
in truth, there is no difference). To compare the
magnitudes of forces/moments between premolars and
incisors, an ANOVA model with terms for tooth type,
loop location, activation distance, the 2-way interac-
tions of these terms, and the 3-way interactions was
used to model the absolute value of force. Residual
plots were examined to assess possible violation of
model assumptions. A Sidak adjustment was used to
control for type 1 error.
RESULTS
The 6 force and moment components (Fig 3) as
functions of activation for the 4 families of loop
position (P1, P2, P3, and P4; Fig 4) were obtained for
the lateral incisor and the premolar, the teeth adjacent
to the extraction space. In general, all 6 components
changed with activation and loop location. The forces
on the brackets in the buccolingual direction, Fx, are
shown in Figure 5, A (positive Fx is in the buccal
direction). Similarly, the mesiodistal forces, Fy, are
shown in Figure 5, B. The positive and negative Fy force
components on the premolar and lateral incisor, respec-
tively, move the teeth into the extraction space (Fig 1).
Changes in the apical (⫺Fz, intrusive) and occlusal (⫹Fz,
extrusive) forces are shown in Figure 5, C.
The concomitant moments about the buccolingual
(Mx), mesiodistal (My), and occlusogingival (Mz)
directions are shown in Figure 5, D-F, respectively.
The positive directions of Mx, My, and Mz correspond
to the buccal, mesial, and occlusal directions according
to the right-hand rule convention.
The Figure 5 layout emphasizes the role of activa-
tion. To highlight the effects of loop location, the same
143.e4 Chen et al American Journal of Orthodontics and Dentofacial Orthopedics
August 2007

Fig 5. Measured force and moment components acting on incisor and premolar with 4 families of
loop position: A, buccolingual (palatal) force component; B, mesiodistal force component;
C, intrusive-extrusive force; D, second-order moment; E, third-order moment; F, first-order
moment.
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 132, Number 2
data are graphed in Figure 6 as functions of loop
location with 3 families of activation.
Fy and Mx play the dominant roles in space closure
because they produce crown movement in the mesiodistal
direction. For all loop locations, |Fy| (the magnitude of Fy)
on the incisor (solid symbols, Fig 5, B) is low (⬍1.0N; ⬃0
with the loop at P4) and relatively insensitive to activation
and loop position (Fig 6, B). However, the force on the
premolar is about twice as high, and it increases with
activation. The highest and lowest forces on the premolar
occur with the P2 spring and in the most anterior (P1)
locations, respectively.
The most mesially located loop (P1) results in an |Mx|
that is about 3 times higher on the incisor than on the
premolar (Fig 6, D) (at 3.2 mm activation, P ⬍.0001). The
most distally located loop (P4) shows the opposite. With
a centered loop, P3, the |Mx| on the 2 brackets are nearly
equal (at 0, 1.6, and 3.2 mm activation, P ⫽ .8538, .0121,
and ⬍.0001, respectively). Furthermore, with the spring at
P4, Mx on the incisor changes very little with activation
(Fig 6, D); it is similar for P1 on the premolar. All x-axis
moments on the incisor and premolar are negative and
positive, respectively.
Relatively little attention has been focused on forces
in the buccolingual direction, Fx. For all loop locations,
Fx exists on both teeth, and it is negative (palatally
directed) except for location P4, which produces a small
buccally directed force on the premolar (Figs 5, A, and 6, A).
|Fx| on the lateral incisor is higher than on the premolar.
(For P4, at each activation distance, P ⬍.0001. For P2, at
0 activation, P ⫽ .0024; at 1.6 mm activation, P ⫽ .9999;
at 3.2 mm activation, P ⫽ .0138. For P3, the correspond-
ing P values are .0102, ⬍.0001, and ⬍.0001. P ⬍.0001
for all activations of P1.) The largest palatal force on the
incisor corresponds to the most distal loop position, P4.
Placing the loop mesially reduces the palatal force on the
incisor, but, even so, its magnitude remains about the
same as the closing force (Fy) on the premolar, and
substantially higher than the closing force on the incisor
itself (Fig 5, B). (On the premolar, P ⫽ .9959, .9812, and
.0512 for activations 0, 1.6, and 3.2 mm, respectively. On
the incisor, the corresponding P values are .0003, ⬍.0001,
and ⬍.0001.)
My has a more consistent pattern (Fig 5, E). On the
incisor, |My| increases with activation, but loop loca-
tion has little effect (Fig 6, E). On the premolar, |My| is
smaller, and activation has less influence, and there is a
direction reversal of the small magnitude My with P1.
All moments on the incisor (⫹My) produce rotation
tending to move the crown buccally or the root pala-
tally. The premolar experiences smaller and, except for
the P1 loop, opposite rotations.
The only effects of activation are with P3 and P4 on
Chen et al 143.e5
incisor intrusion and premolar intrusion/extrusion; all
other intrusive/extrusive forces are either small or
relatively constant with respect to activation (Fig 5, C).
In contrast, the magnitudes and senses (extrusion vs
intrusion) of Fz are strikingly sensitive to loop location
(Fig 6, C). With the exception of Mz on the incisor with
P3 and P4, |Mz| does not change with activation (Fig 5, F).
The largest effect of loop position is on the incisor Mz
at maximum activation (Fig 6, F).
DISCUSSION
Studies have traditionally looked at the force-moment
systems produced by various loop designs and the effects
of appliance material, shape, and position. Those studies
generally modeled (experimentally, analytically, or nu-
merically) a loop acting in a plane consisting of 2 isolated
teeth adjacent to an extraction space. This study is funda-
mentally different in 2 ways: it is a highly statically
indeterminate system because the loop is part of a contin-
uous full archwire, and arch curvature introduces ex-
tremely complex 3D interactions. As a consequence, the
results obtained with this model manifest in ways that
challenge some accepted notions.
For example, the measured force-moment system
experienced by the bracket is not the same as the
system generated by the spring per se. This is in
contrast to the aforementioned studies in which the
loop-generated system is identical to what is applied to
the bracket. This critical distinction is due to the
archwire connection to the proximal tooth. That is, it is
impossible to separate the individual contributions (to
the measured force-moment system) of the spring on 1
side of the bracket from that of the wire segment that
connects it to the adjacent tooth.
For a specific tooth, the sole determinant of a
particular translational and rotational movement direc-
tion is the force-moment system that acts on the tooth.
For convenience, that load system is generally defined
relative to the tooth’s center of resistance. It does not
matter how the load is applied to the crown as long
as the equivalent force-moment system is produced
at the center of resistance. It matters not whether the
tooth is free-standing or part of a segment; appliance
design is irrelevant; treatment philosophy is imma-
terial.
Another distinction is that, in this model, space clo-
sure requires movement of the lateral incisor in its distal
(⫺y) direction or movement of the premolar in its mesial
(⫹y) direction (Fig 1). These paths are not coincident, and
there are out-of-plane forces and moments, so equilibrium
principles do not require the spring-generated closing
forces on the 2 teeth to be the same magnitude. Further-
more, even if the arch were straightened out into a plane
143.e6 Chen et al American Journal of Orthodontics and Dentofacial Orthopedics
August 2007

Fig 6. Measured force and moment components acting on incisor and premolar with 3 families of
activation: A, buccolingual (palatal) force component; B, mesiodistal force component; C, intrusive-
extrusive force; D, second-order moment; E, third-order moment; F, first-order moment.
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 132, Number 2
to resemble previous models, the 2 teeth would still not
experience the same closing forces because of the above-
mentioned load sharing with their respective neighbors. In
addition, in plane problems, the closing force, Fy, domi-
nates, and the buccolingual force, Fx, is zero. This is in
contrast to our measurements—the buccolingual forces
(Figs 5, A, and 6, A) can exceed the closing force
magnitudes (Figs 5, B, and 6, B).
Further examination of the results shows important
contradictions to traditional views on loop mechanics.
Activation used to increase the closing force (⫹Fy) on
the premolar (open symbols in Fig 5, B) also changes,
to various degrees, the other load components on that
tooth and on the incisor (Fig 5). The closing force
(⫺Fy) on the incisor (solid symbols in Fig 5, B) also
increases, but not at the same expected rate as on the
premolar. It is the concomitant palatal-directed force on
the incisor (⫺Fx, Fig 5, A) that increases the most with
activation.
Three dimensionality and connections to proximal
teeth wreak similar havoc on the measured moments.
For example, in the centered loop position (P3) at
placement (zero activation), |Mx| on the 2 teeth (solid
squares for the incisor, open squares for the premolar,
Fig 5, D) are approximately equal (20 N-mm) even
though their respective x-axes are not parallel. With
activation, however, |Mx| on the premolar increases
more rapidly. This can be partly attributed to the loss of
symmetry caused by activation displacing the loop
distally. In general, on the incisor, |Mx| is about the
same as |My| (solid symbols in Figs 5, D and E), not
|Mx| much greater than |My|, which is near 0, as is the
case with traditional 2D models.
Loop position is a means to control the M/F ratio.
Often, the goal is to obtain the necessary Mx/Fy for
tooth translation. As can be seen in Figure 6, D, spring
location has an influence on Mx, particularly at high
activations, but less so on Fy (Fig 6, B). However,
the buccal/lingual (Fx, Fig 6, A) and the intrusive/
extrusive (Fz, Fig 6, C) forces on the teeth are more
affected.
Thus, load coupling becomes obvious when all 3 force
and 3 moment components on a tooth in a 3D simulation
are simultaneously measured. It was demonstrated that the
extremely complex interrelationships between the load
components depend on activation and loop location.
These side effects, unintended and difficult to predict or
identify, are likely to be detrimental to treatment progress.
Interestingly, such load coupling was not detected in
segmented triangular springs with various combinations
of first- and second-order gable bends.1
Although we understand that many factors affect
the load system, we focused only on loop location.
Chen et al 143.e7
Other factors were fixed. For example, different
results would have been obtained if the loop had
been placed in an arch, or in an arch location with a
different curvature—ie, with a different angle be-
tween the y-axes (Fig 1). Although triangular and
T-loops are generally considered interchangeable, it
is unlikely that the 2 springs would have similar 3D
characteristics. These are obvious questions for fur-
ther investigations.
CONCLUSIONS
The complete load system acting on a single orth-
odontic bracket can be measured simultaneously by
using the newly developed technology. The instrument
enables improved verisimilitude with measurements
under simulated clinical settings.
Our results contradict several important notions
about spring mechanics because previous studies were
based primarily on plane models with 1 tooth on either
side of an extraction space. Although these 2D models
are statically indeterminate, they lack the additional
indeterminacy due to connections to adjacent teeth.
Furthermore, arch curvature introduces tremendously
confounding 3D effects.
REFERENCES
1. Katona TR, Le YP, Chen J. The effects of first- and second-order
gable bends on forces and moments generated by triangular
loops. Am J Orthod Dentofacial Orthop 2006;129:54-9.
2. Chen J, Markham DL, Katona TR. Effects of T-loop geometry on
its forces and moments. Angle Orthod 2000;70:48-51.
3. Mazza D, Mazza M. Specialized spring design in segmented
edgewise orthodontics: further verification of dedicated software.
Angle Orthod 2000;70:52-62.
4. Menghi C, Planert J, Melsen B. 3-D experimental identification
of force systems from orthodontic loops activated for first order
corrections. Angle Orthod 1999;69:49-57.
5. Chen J, Chen K, Katona TR, Baldwin JJ, Arbuckle GR. Non-
linear large deformation FE analysis of orthodontic springs.
Biomed Mater Eng 1997;7:99-110.
6. Kuhlberg AJ, Burstone CJ. T-loop position and anchorage
control. Am J Orthod Dentofacial Orthop 1997;112:12-8.
7. Odegaard J, Meling T, Meling E. The effects of loops on the
torsional stiffnesses of rectangular wires: an in vitro study. Am J
Orthod Dentofacial Orthop 1996;109:496-505.
8. Manhartsberger C, Morton JY, Burstone CJ. Space closure in
adult patients using the segmented arch technique. Angle Orthod
1989;59:205-10.
9. Raboud D, Faulkner G, Lipsett B, Haberstock D. Three-
dimensional force systems from vertically activated orthodon-
tic loops. Am J Orthod Dentofacial Orthop 2001;119:21-9.
10. Siatkowski RE. Continuous arch wire closing loop design,
optimization, and verification. Part I. Am J Orthod Dentofacial
Orthop 1997;112:393-402.
11. Lipsett AW, Faulkner MG, el-Rayes K. Large deformation analysis
of orthodontic appliances. J Biomech Eng 1990;112:29-37.
12. Koenig HA, Burstone CJ. Force systems from an ideal arch—
large deflection considerations. Angle Orthod 1989;59:11-6.
143.e8 Chen et al
13. Meling TR, Odegaard J, Holthe K, Segner D. The effect of
friction on the bending stiffness of orthodontic beams: a theo-
retical and in vitro study. Am J Orthod Dentofacial Orthop
1997;112:41-9.
14. Bednar JR, Gruendeman GW. The influence of bracket design on
moment production during axial rotation. Am J Orthod Dento-
facial Orthop 1993;104:254-61.
American Journal of Orthodontics and Dentofacial Orthopedics
August 2007
15. Rinaldi TC, Johnson BE. An analytical evaluation of a new
spring design for segmented space closure. Angle Orthod 1995;
65:187-98.
16. Gjessing P. Biomechanical design and clinical evaluation of a
new canine-retraction spring. Am J Orthod 1985;87:353-62.
17. Marcotte MR. Optimum time and temperature for stress relief
heat treatment of stainless steel wire. J Dent Res 1973;52:1171-5.