American Journal of ORTHODONTICS

Foutded in 1915 Volume 86, Number 5 September, 1984

Copyright 0 1984 by The C. V. Mosb~ Cotnputr>

ORIGINAL ARTICLES

Elastic property ratios of a triple-stranded

stainless steel amh wire
R. P. Kusy, Ph.D.,* and G. J. Dilley, D.D.S., M.S.D., M.S. Dr. Kusy
Chtrpel Hill, N.C.

The general elastic property ratio equations for nth-stranded wires are derived and then specified for the case of
a triple-stranded arch wire. Several parameters are defined, including the modulus of elasticity (E) and the helical
spring (K) and bending plane (A) shape factors. Thereafter, the elastic property ratios of a wide range of
compositional/configurational combinations are determined, using a representative triple-stranded 0.0175 inch
(3 x 0.008 inch) stainless steel wire as the base line. These results show that the particular 3 x 0.008 inch wire
studied possesses the stiffness of an 0.010 inch stainless steel wire but has at least 20% more strength and
range. Furthermore, the stiffness of the 3 x 0.008 inch multistranded wire is similar to an 0.016 inch.
nickel-titanium wire but only 40% that of an 0.016 inch beta titanium wire. When these elastic property ratios are
compared with the previous results reported by Thurow, Burstone, and Kusy, differences are noted which can be
explained on the basis of the mechanical property values and/or the geometric modeling assumed.

Key words: Arch wires, elastic property ratios, stainless steel wires, nickel titanium wires, beta titanium
wires, multistranded wires

Since the mid-1960s, the materials used for

arch wire mechanics have been markedly expanded.
Today, not only stainless steel (S.S.) but also cobalt-
chrome (Co-Cr), nickel-titanium (Ni-Ti), and beta
titanium (P-Ti) wires are available.‘-’ In addition to
these different alloys, the number of wire configura-
tions has increased the wire inventory significantly.
Besides single-stranded round, square, and rectangular
wires, there are now multistranded wires of varying
strand size, shape, and number. Therefore, it is im-
perative that the characteristics of the various wires be
known if one is to select the appropriate wire for a
particular clinical situation.
From the Department of Orthodontics/Dental Research Center, University of
North Carolina.
Presented in part at the annual session of the American Association for Dental
Research, Cincinnati, Ohio, March, 1983 (Abstract 139).
This investigation was supported by National Institutes of Health Research
Grants DE02668 and RR05333.
*Associate Professor.
To facilitate the clinician’s use of this information,
several authors have devised tables and/or nomograms
to compare the physical properties of many of the
available arch wires.7P’” However, these efforts have
only estimated the properties of the multistranded or
twist wires, because the wire geometry was handled too
simplistically and/or because the modulus of elasticity
of orthodontic wires was assigned questionable values.
This latter case was in contrast to reports in the metal-
lurgic literature on stainless steel wires.‘“-‘5 As a re-
sult, there exists no unified theory which can describe
even the elastic property ratios of all types of wires:
single-stranded, simple multistranded, and complex
multistranded (that is, multiconfigurational or soldered
wires).
This study defines the general elastic physical
properties of strength, stiffness, and range for a wide
selection of single- and multistranded wires and spe-
cifically for the 0.0175 inch triple-stranded twist wire
177
 Am J. &thud.
178 Kusy and Dille) September 1984

Fig. 1. Typical geometry of a simple multistranded wire. Shown

is a wire of diameter 0 composed of three wire strands, each of
diameter d. The axial distance which a wire strand traverses per
rotation equals 5’. The helix angle, (Y, which a wire strand
makes with the normal to the wire axis may be described in
terms of d, D, and P’ (see equation 5).

Beam Type Geometry Moment Diagram

Fig. 3. Stress-strain diagram for a wire in tension and com-

pression. In addition to the point of failure (X), the proportional
P limit ((Tag,*), and the yield point (0) are marked. The slope of
the linear portion of these curves is denoted as the modulus of
Cantilever: elasticity, or simply as Young’s modulus (E). Strictly speaking,
the theory presented herein is applicable only within the elastic
I l3angerou; region of the stress-strain curves (the shaded areas). Under
Section certain circumstances, however, it may be extended to the yield
point. The shaded area is commonly referred to as the modulus
of resilience.

Simple: D- The diameter of the wire.

E- Young’s modulus of elasticity.
h- In a rectangular wire strand, the side nor-
mal to the neutral axis (n.a.).
i- An index used to define the limits of nu-
Fig. 2. Two common beam configurations and their correspond- merical integration.
ing moment diagrams. The dangerous section (that is, the I- Rectangular moment of inertia.
transverse plane in the beam at which the moment is maximum)
K- Load/support constant.
is located at the wall and at the midspan for a cantilever and a
simple beam, respectively. Since alternative wires are always I! - Distance from the end of the beam to the
evaluated for the same clinical situation, the force applied to the dangerous section.
beam, P, will persist only to equations 18-23. 1*- Axial displacement per twist of a wire
strand.
in the bending mode of stress. These results are then L- Beam’s span length.
compared to other wires via tables and nomograms as M- Moment applied to a beam.
well as to other results published in the literature. n- Number of wire strands in a wire.
n.a.- The line that results from the intersection of
NOTATION
any wire cross section with the neutral sur-
The following symbols will be used throughout this face which is neither under tension nor
text: compression (that is, the neutral axis).
A- The cross-sectional area of a wire strand. P- Force applied to a beam.
b- In a rectangular wire strand, the side paral- PL- On a stress-strain diagram, the point at
lel to the neutral axis (n.a.). which E is no longer invariant.
C- Distance from a neutral axis of a wire to the s- The side of a square wire strand.
outermost fibers of a strand. X- A wire of arbitrary composition/cross
Cmax- The maximum value which c attains in a section.
multistrand wire. Y- A wire of arbitrary composition/cross
d- The diameter of a wire strand. section.
d*- The fundamental wire dimension normal to (Y- Helix angle of a wire strand.
the neutral axis (n.a.). S- Deflection of a beam under load.
Volume 86 Elastic property ratios of triple-stranded stainless steel wire 179
Number 3

A- Difference between the centroidal distances

of a wire strand and its wire.
K- Helical spring shape factor.
A- Bending plane shape factor.
v- Poisson’s ratio.
19- Angle of rotation of the bending plane.
cr- The internal resistance of a material to an
externally applied force or load (that is,
stress).
~max- Maximum elastic stress. Fig. 4. Schematic definition of the helix angle (a). If one revolu-
tion of a wire strand is unfurled and its base length [n(D-d)] and
THEORY corresponding distance traversed along the original wire axis
(R*) are determined (Fig. 1) then a ratio of these two distances
The general multistranded arch wire case
equals tan CL Everything else being equal, the greater n(D-d) or
Classic mechanics theory shows that as the diame- the less II’ is, the more compliant a wire will be.
ter of a wire strand is reduced, the stiffness decreases as
a function of the fourth power, and the range increases In equation 3, K is the overall load/support arrange-
proportionally. In terms of performance, the wire is ment constant of the beam (for example, % for the
delivering lighter forces per unit of activation over a cantilever beam and l/4x for the simple beam”), L is
greater distance. These trends would be acceptable if it the length of the beam (Fig. 2), I is the rectangular
were not for the accompanying strength reductions to moment of inertia, E is the modulus of elasticity, and K
the third power which permit the distortion and even is designated the helical spring shape factor. This last
failure of a light single-stranded arch wire. (Recall term may be derived as
that strength = stiffness x range.‘“) To improve the 2sina
strength and, at the same time, to maintain the desirable K= (4)
2 + vcos’)a .-I-

stiffness and range properties, many small wires are

twisted together and even swagged or spot-welded. The
result is an inherently high-elastic modulus material
behaving as a low-stiffness member because of its
coaxial springlike nature.
If in a .symmetricul wire of diameter D there are n
frictionless stmnds, each of diameter d (Fig. l), then
the moment (M) exerted on each strand at its dangerous
section may be written as
(1)
in which P is the applied force and ! is the measured
distance (Fig. 2).
At this section the maximum stress in each strand
will equal the familiar formula for flexure,“’
MC
mIlla\ = - I >
Here v represents the ratio of the lateral to the axial
strain (that is, Poisson’s ratio)‘” and cyis the helix angle
of a wire strand or.
a = tan-’ 1*
-ir(D - d) ’
(5)
in which j* is the distance through which a wire strand
translates along the wire axis per revolution and T(D-d)
is the length of that wire per revolution (Figs. 1 and 4).
For an arbitrary wire, X, the strength, stiffness, and
range in the elastic region (shaded areas of Fig. 3) may
be determined as follows: By substituting equation 1
into equation 2 and solving for P,
Strength expression: P , = ((I)( f )(n)( a, (6)
By rearranging equation 3 in terms of P/6,

in which crnax = u = cpL is defined as the limits of the Stiffness expression: (i)x = @)(1)(m)(&) (7)
elastic region (Fig. 3), and c is the distance from the
outermost reaches of a wire strand to the neutral axis By substituting equations 1 and 2 into equation 3 and re-
(n.a.) of the wire. arranging in terms of 6,
Recognizing that the deflection of a wire (6) will
Range expression: Sy = ($($)(i)(y) (8)
diminish proportionally as either P is reduced or n is
increased, a general beam formula for a variable-strand
In equations 6 to 8 the first three terms on the right-
wire may be described,
iThere are specific assumptions implicit in equation 4 which are beyond the
fi = KPL” scope of this text,‘*-“” but which could lead to misleading results. One such case
(3)
KnEI could occur if wires which had asymmetrical wire strands were considered.
180 Kusy and Dille!

nd4
; Ina.= ; d*=d

TiziG
1
n.a. ; In.a.=-
4

;2
’
’
d*=s

bh’
-i n.a. ; In.a.=12 ; d*=h
I IfiF
-4bk

bh3
; In.a.=F ; d* = h

Fig. 5. Geometry of three common wire configurations. Shown

are the critical dimensions (d, s, b, and h) and the definitions of
their geometric parameters, the moment of inertia, /n.a. (I G/C,
in which i = 7 for single-stranded wires; see equation 13) and
the fundamental wire dimension perpendicular to the neutral
axis (ma.), d’. Compared to b, h has a much greater influence
on the geometric stiffness of a wire and reduces to the square
wire values for b = h. When d = s, a square wire will be 70%
stiffer than a circular wire; that is, 1.7 times more force will be
delivered per unit of activation.
Section A-A

hand side of each expression are the material, the Fig. 6. Variation of c in a simple multistranded wire for which
geometric, and the dimensionless constants, respec- n = 3. When the idealized cross section is viewed (section A-A),
the threefold symmetry is obvious. If the wire traversed along
tively. Note that since these elastic property expres- its wire axis to another section (for example, section B-B), a
sions will be used to evaluate alternative wires under different orientation of the wires relative to the neutral axis
the same clinical situation, the parameters, K, 1, and L (ma.) would occur. Indeed, the same result may be attained if
will ultimately disappear from the subsequent elastic only section A-A is retained and the ma. is rotated through any
property ratio expressions. For the moment, they 0 of 180”/n. In this case, then, 0 = 60” (see equations 9 and
14).
comprise the fourth term, although knowledge of their
exact values will not be necessary.
While Fig. 5 defines the geometric constants for
elementary cross sections, the determination of both c Combining equation 9 with equations 6 to 8,
and I is, in principle, more complex for multistranded
arch wires. This occurs in the case of c (Fig. 6) because Strength expression: Ps = (a) ( zI )(f)(f) . (10)
the maximum perpendicular distances from the neutral
axis of the wire to the outermost fibers of a strand (one Stiffness expression: (i), = (R)(I)(Kn)(&j (11)
on the compression side and one on the tension side)
vary along the wire’s axis. Fortunately, for nth-fold Range expression: 6 y = (;)(&)(&)(yq. (12)
symmetry, c need only be evaluated through an angle
of rotation of the bending plane (0) equal to rrln radians Particularly note two items: That the factor of 2 will
in order to determine the maximum value. For any disappear in all elastic property ratio expressions
variable-stranded arch wire, (hence its placement in the fourth term) and that the
stiffness expression is unaffected by the distance c
C=hd* (hence equation 11 is a recapitulation of equation 7.
2 ’
What functional relationship A will have, however, will
in which d* is defined as the fundamental wire dimen- be dependent upon the geometry of the wire under in-
sion normal to the neutral axis (n.a.) (Fig. 5) and h is vestigation .
designated as the bending plane shape factor. Likewise, the moment of inertia (I) of a wire may
Volume 86 Elastic property ratios of triple-stranded stainless steel wire 181
Number 3

P--
Centroidal -
Distance of
Strand C

Centroidal- Distance of
Distance of the Wire
Strand B (Strands A.&C)

Fig. 7. A simplified illustration showing the centroidal distances

of a triple-stranded wire and its individual wire strands, when
measured from an arbitrary base line. The difference between
the former and each of the latter equals A (see equation 13).
When each of these quantities is squared and then multiplied by
the cross-sectional area of its respective wire strand, the sum- Bending Plane Angle, 8
mation of these n values equals the error which previous mod-
els accepted (see equations 13 and 24).‘-‘O Fig. 8. Dependence of the bending plane shape factor (A) on
the bending plane angle (0) for a triple-stranded wire (n = 3)
(see equation 14). As discussed in Fig. 6, 0 need be evaluated
be dependent not only upon the individual wire strand only over a 60” increment in order to determine Amax which, in
geometry but also upon the overall configuration of the turn, establishes cmax via equation 9. Consequently, the de-
wire itself. This dual dependency is seen in the follow- grees noted on the abscissa are arbitrary. Similar 0-A plots for
any value of n may be obtained by evaluating a geometric con-
ing formula, which describes the moment of inertia of struction like the one depicted in the Appendix.
the multistranded wire around its neutral axis (I,,,,):
Table 1. Wire inventory investigated
I,,,, = I = ig, [ICI + A,Af]
Geometry
Here, the moment of inertia for a wire strand about its
Material Round Square Rectangular*
centroid ($) plus the product of the area of this wire
strand (AJ times the squared difference (as measured s.s 3 x 0.008” 0.016 x 0.016 0.016 x 0.022
from an arbitrary base line) between the centroidal dis- 0.010 0.017 x 0.017 0.018 x 0.022
0.012 0.017 x 0.025
tance of this wire strand and the centroid of the wire
0.014 0.018 x 0.025
itself (A:) are summed over n (Figs. 5 and 7). For a 0.016 0.019 x 0.026
symmetrical wire, I,.,. (or simply I) is independent of 0.018 0.021 x 0.025
the bending plane chosen, regardless of the number of 0.020
different wire strand geometries present. Since in the Ni-Ti 0.016 0.018 x 0.018 0.016 x 0.022
present derivation the summation is explicit in n and 0.018 0.017 x 0025
0.020 0.019 x 0.025
AiKf is implicit in K, only Ici must be evaluated for
0.021 x 0.025
equations 10 and 11 (Fig. 5).
p-Ti 0.016 0.0175 x 0.0175 0.016 x 0.022
0.018 0.017 x 0.025
A specific triple-stranded arch wire case
0.019 x 0.025
The preceding theory is applicable to many simple 0.021 x 0.025
or complex multistranded wires and, in its degenerative
*Evaluated both edgewise (e) and Aatwise (f).
form, to single-stranded wires. Since the absence of
friction and the presence of wire symmetry have been
assumed, soldered wires of any construction and mul- When c is evaluated over any 0 = 60”) from Figs. 6 and
tistranded wires of any asymmetrical geometry are 8 and equations 9 and 14,
excluded.
To use equations 10 to 12, the material parameters
((T, E), the geometric factors (I,d*), and the dimension- in which h = 2.155 and d* = d.
less constants (n, K, and A) must be determined for the Determination of the remaining parameters (o, E,
arbitrary wire, X. If X is defined as any frictionless, I, d, and K) requires a specific choice of a triple-
symmetrical wire having n = 3 and cy Z 65”, then for stranded wire. Quite recently, a nominal 0.0175 inch
its threefold symmetry (see Appendix), (3 X 0.008 inch) triple-stranded stainless steel wire
~coso ) (Wildcat wire, GAC International, Inc., Commack,
A = (1 + sin0 + -
3 N.Y.) was studied in great detail.‘O Having a d =
182 Kusy and Dilley Am J. Orthod.
Sqmnber 1984

60

OL
Modulus of Elasticity.
E ( I O’psi)
1
0 32

(4
0 30 60
Fig. 10. Histograms comparing the modulus of elasticity (f, of
Helix Angle, a an 0.0175 inch (3 x 0.008 inch) wire (black bars) to the
grouped data of two single-stranded wires (open bars), an
Fig. 9. Dependence of the helical spring shape factor (K) on the O.OlO-inch stainless steel (Unitek Corporation, Monrovia, Calif.)
helix angle (a) for any wire material having a Poisson’s ratio,
and an 0.01.8-inch stainless steel (GAC International, Inc.,
v = 0.28 (see equation 4). Although small variations in v are not Commack, N.Y.), in three-point (a) and four-point (b) bending.
expected to influence the functional relationship significantly, The results for this simple multistranded wire (28.6 t 0.6 x lo6
the present curve should be used only for ferrous alloys, includ-
and 29.6 ? 0.9 x 106 psi, respectively), compare favorably to
ing the stainless steels. Note that K changes very slightly as the
the overall values (27.9 i- 0.8 x IO6 and 28.4 -+ 1.3 x 106 psi,
wire transforms from a straight strand ((Y = 90”) to one having a respectively), and the metallurgic literature for stainless steel
step pitch (LY = 80”). When a 5 50”, however, a nearly linear
(28.0-30.0 x lo6 psi13m15 ). For further details of these experi-
relationship is established for which equation 4 may be approx-
ments, see Kusy and Dilley.zo
imated by K = 0.0148 0~.

Table II. Updated physical properties of set equal to 315 X 10:’ psi$ and E equalled 27.9 ?
modulus of elasticity (E) and stress at 0.8 x lo6 psi and 28.4 i 1.3 x lo6 psi in three- and
proportional limit (opL) for stainless steel, four-point bending, respectively (Fig. lo).‘“. %OThe
nickel titanium, and beta titanium wires data which served as a backdrop represented the pro-
Material &IO’ psi)* a,,,(lO~ psi)
jected stiffness and bracket engagement analogs, that
I I
is, an 0.010 inch and an 0.018 inch S.S. arch wire.
S.S. 27.9 315 With this information, all of the parameters for this
Ni-Ti 4.8t I80t triple-stranded arch wire may be substituted into equa-
P-Ti 10.ot 170t
tions 10 to 12 and the elastic properties calculated. In
*To convert psi to Pa, multiply by 6.895 x IO’. practice, however, clinicians are not concerned with
t See Kusy and Greenberg. I’) what the absolute properties of an arch wire are; rather,
they are concerned with how an alternative wire com-
0.007709 & 0.000014 (5) inch, the value for I equalled pares to a wire that they now use. Therefore, on the
1.734 x 10-l’ in4 and basis of equations 10 to 12, the concept of elastic
0.07300 property ratios is used to compare one wire, X, with an
= 68.1” , (16) alternative wire, Y.
n(0.01705-0.007709)
from which RESULTS

2sin(68.1”) The wire inventory (Table I) that will be compared

= 0.910 (17)
K = 2 + (0.28)cos”(68.1”) against the triple-stranded arch wire is essentially the

This and other CT-K combinations are shown in Fig. 9

J;Thls value was actually determined with a similar size of smgle-stranded arch
for the Poisson’s ratio of stainless steel (0.28).‘“, ” wire because the braided or twisted wire strands could not be sufficiently
With regard to the material properties, cr = opL was straightened to provide suitable test specimens.
 Eltrstic property rdos of‘ triple-.strardrd sttrinlrss .rtwl wirr 183

Bending Bending
Stiffness Strength Range Stiffness Strength Range

IO
9
e3

I\

,016

,014

,012

.OlO
3x.m
LO175)

Fig. 11. Nomogram comparing the elastic properties of a Fig. 12. Nomogram comparing the elastic properties of a
triple-stranded 0.0175 inch arch wire to other stainless steel triple-stranded 0.0175 inch stainless steel arch wire to selected
round arch wires. nickel-titanium (Ni-Ti) arch wires.

same expanded one that recently appeared in an article
illustrating the use of nomograms to determine the
elastic property ratios of orthodontic arch wires.‘”
Using the updated physical properties of S.S., Ni-Ti,
and p-Ti wires shown in Table II, all forty-three wire
permutations (designated wire ‘ ‘Y “) were compared
against the triple-stranded arch wire (designated wire
“X”) in the following generalized elastic property ratio
expressions:
Whereas in the denominator n, K, and X equal 3,
0.911, and 2.155, respectively, for all of the present
cases the numerator degenerates to the conventional
single-stranded formulas in which n = K = A = 1 (equa-
tions 5 to 7 of Kusy and Greenberg”‘). Thus, equations
18 to 20 become
Strength ratio: 2 = (21)

(18) (El),
Stiffness ratio: - = ___-. (22)
(2.73EI),

(19)
(23)

These last three equations define the elastic property

(20) ratios of any single-stranded wire (Y) against a specific
triple-stranded wire (X), the 0.0175 inch Wildcat wire.
 Am. J. Orrhad.
.Septrmher 1984

Table III. Elastic property ratios of stainless

Bending
steel wires versus a 3 X 0.008 inch (0.0175
Stiffness Strength Range
inch) stainless steel wire
X Y Ben&q

S.S. S.S. Strength Sriffness Range

3 x 0.008” 0.010” 0.78 I .o 0.76

(0.0175”) 0.012 1.4 2.7 0.63
0.014 2.2 4.0 0.54
0.016 3.2 h.X 0.47
0.018 4.6 1 I .o 0.42
0.020 6.3 17.0 0.3X
0.016 x 0.016 5.5 12.0 0.47
0.017 x 0.017 6.5 15.0 0.44
0.016 x 0.022 (e)* 10.0 30.0 0.34
(t-1; 7.5 16.0 0.47
0.018 x 0.022 (e) 12.0 34.0 0.34
(f) 9.5 23.0 0.42
0.017 x 0.025 (e) 14.0 47.0 0.30
(f) 9.6 22.0 0.44
0.018 X 0.025 (e) 15.0 50.0 0,.X)
(f) I I.0 26.0 0.32
0.019 x 0.026 (e) 17.0 59.0 0.39
(f) 13.0 3I 0 0.40
0.021 X 0.025 (e) 17.0 58.0 0.30
(f) 15.0 41 .o 0.36

*(e) = Edgewise.
f)-flatwire t(f) = Flatwise.

Fig. 13. Nomogram comparing the elastic properties of a
triple-stranded 0.0175 inch stainless steel arch wire to selected
beta titanium (/3-Ti) arch wires.
From the material properties of Fig. 10 and Table II
and from the geometric parameters stipulated in Fig. 5
and specifically defined in Table I, the elastic property
ratios of S.S., Ni-Ti, and P-Ti arch wires are shown in
Tables III to V and, for selected wires, in the nomo-
grams of Figs. 11 to 13.
Table III and Fig. 11 show that a 3 X 0.008 inch
S.S. arch wire has the stiffness of an 0.010 inch S.S.
arch wire. Consistent with the concept of the mul-
tistranded arch wire, the strength and the range of the
3 x 0.008 inch arch wire are 20% and 25% greater than
that of the 0.010 inch alternative. Note that a single-
stranded arch wire with the same bracket engagement
as the triple-stranded wire presently under considera-
tion would be more than an order of magnitude stiffer.
When the same 3 x 0.008 inch S.S. arch wire is
compared to the Ni-Ti wire inventory (Table IV and
Fig. 12), only the 0.016 inch Ni-Ti arch wire has simi-
lar stiffness. In this case, however, the Ni-Ti alterna-
configurations shown on the nomogram, the 0.0 16 inch
P-Ti arch wire is closest to the 3 X 0.008 inch arch
wire alternative, but in no respect is this wire compa-
rable. When Tables III and V or Figs. 11 and 13 are
analyzed, however, the elastic properties of this /3-Ti
arch wire (Y) are more like those of an 0.012 inch S.S.
(X) wire. These elastic property ratios are as follows:
stiffness, 1.1; strength, 1.2; and range, 1.1.4
DISCUSSION
The results of the present study may be compared
to the previous results of Thurow, Burstone, and
Kus~.~-‘”
The earliest analysis
In the first attempt to elucidate the differences be-
tween the elastic properties of multistranded and
straight wires, Thurow’ suggested that the properties of
a strand be multiplied by the number of strands. In
terms of I, this was equivalent to
I= i I,., , ( 241
i=l

tive has about twice the strength and over 50% more in which (Ai . A’$ = 0 (see equation 13). Using this
range than the multistranded wire. *A good cross-reference for these and other wires of equivalent stiffness is
Of the three alloys, the fi-Ti wires demonstrate the Table IV of Kusy and Greenberg. ” For an in-depth explanation that shows how
least compatibility (Table V and Fig. 13). Of the five to compare arch wires of different composition and configuration. see Kusy I”
 Elastic properq rutios qf triple-strunded stuinless steel wire 185
Number 3

TableIV. Elasticpropertyratiosof nickel TableV. Elasticpropertyratiosof betatitanium

titaniumwires versusa 3 x 0.008inch wiresversusa 3 x 0.008inch (0.0175inch)
(0.0175 inch) stainless steel wire stainless steel wire
X Y Bending X Y Bending

S.S. Ni-Ti Strength Stiffness Range S.S. /3-Ti Strength St$%ness Rmge

3 x 0.008” 0.016” 1.8 1.2 1.6 3 x 0.008” 0.016” 1.7 2.4 0.71
(0.0175”) 0.018 2.6 1.9 I .4 (0.0175”) 0.018 2.5 3.9 0.63
0.020 3.6 2.9 1.3 0.017 x 0.017 3.5 5.3 0.67
0.018 x 0.018 4.4 3.2 I .4 0.016 X 0.022 (e)* 5.6 Il.0 0.52
0.016 x 0.022 (e)* 5.9 5.2 1.1 (f)i 4.0 5.7 0.71
(fV 4.3 2.7 1.6 0.017 X 0.025 (e) 7.6 17.0 0.45
0.017 X 0.025 (e) 8.1 8.1 1.0 (f) 5.2 7.8 0.67
(f) 5.5 3.7 1.5 0.019 X 0.025 (e) 8.5 19.0 0.45
0.019 X 0.025 (e) 9.0 9.0 1.0 (f) 6.5 Il.0 0.60
(f) 6.9 5.2 1.3 0.021 X 0.025 (e) 9.4 21.0 0.45
0.021 X 0.025 (e) 10.0 10.0 1.0 (f) 7.9 15.0 0.54
(f) 8.4 7.0 1.2
*(e) = Edgewise.
*(e) = Edgewise. t(f) = Flatwise.
t(f) = Flatwise.

procedure, three strands of an unspecified 0.007 inch
S.S. arch wire had the strength of an 0.010 inch wire,
slightly more stiffness than an 0.009 inch wire, and
three times the range of an 0.021 inch wire. Fortui-
tously, these figures paralleled the results obtained in
the present study, except that a triple-stranded 0.008
inch S. S. wire was used as a reference instead. This
simple multistranded wire had the approximate strength
of an 0.011 inch wire, the stiffness of an 0.010 inch
wire, and three times the range of an 0.020 inch wire.
Nowhere in Thurow’s pioneering work was there any
attempt to account for the helical spring shape factor
(K), (see equation 19); that is, all wires with n = 3 had
the same elastic properties, regardless of their helix
angle, (Y.The same was true of the bending plane shape
factor (A), when strength and range properties were
compared (see equations 18 and 20).
Wire-stiffness numbers
Many years passed until, in 1981, Burstone’ intro-
duced the concept of the wire-stiffness numbers (W,).
The idea was quite straightforward: Via the product of
a material and geometric property, designated the
material-stiffness number (M,) and the cross-sectional
stiffness number (C,), a wire-stiffness number (W,)
was defined. According to Tables IV and V, a simple
triple-stranded 0.0175 inch Hi-T Twist-Flex wire
(Unitek Corporation, Monrovia, Calif.) had a W, num-
ber of 62 (that is, W, 62), while single-stranded 0.009
inch and 0.012 inch S.S. wires equalled W, 26 and W,
81, respectively. Among these products, the triple-
stranded arch wire most closely approximated the
0.012 inch wire. Had a complex multistranded 0.0175
inch Respond wire (Ormco Corp., Glendora, Calif.)
been compared instead (W, 25), the 0.009 inch wire
would have been the better match. Although the present
Wildcat wire was not investigated in Burstone’s study,
the difference between the other two simple mul-
tistranded wires suggests the range of stiffnesses pos-
sible. Actually, the helix angle of the Twist-Flex wire
is so great ((Y > 85”, see Fig. 25 of Dusenberryz2) that
this simple multistranded wire probably represents the
high side of W, for a nominal 3 x 0.008 inch S.S. arch
wire (K = 1, Fig. 9).
With regard to the two titanium alloys, wire-
stiffness numbers were available for both the 0.016
inch Ni-Ti (W, 67) and /3-Ti wires (W, 108), respec-
tively. Here the 0.0175 inch Twist-Flex wire closely
simulated the former. In fact, the stiffness interrelation-
ship among these three wires was similar to that ob-
served in the present work for a Ni-Ti, /3-Ti, and
Wildcat wire. Whether this apparent correlation is sub-
stantive or merely a coincidence is not known at this
time, although the latter is suspected.
In contrast to the present method of analysis,
Burstone’s wire-stiffness numbers differ most from the
advocated method in their semantics. Comparison of
the present Table II to Table III of the previous work
indicates that the moduli of elasticity (E) for stainless
steel differ by nearly 3 x lo6 psi. Actually in Bur-
stone’s article on “Variable-Modulus Orthodontics, “r
values are cited which vary from 23 to 28 X lo6 psi in
bending, whereas a more recent article’” cites a value of
21.5 & 0.8 x 10” psi in tension. Other absolute values
of E are included in this last article, too. There the
Ni-Ti modulus equals 6.55 -+ 0.19 X 10” psi in bend-
ing versus the present 4.8 x 10” psi, whereas the P-T1
modulus is within 5% of the present value. In a second
Table VI. Comparison of the elastic property ratios of two 0.016 inch titanium wires versus a
3 X 0.008 inch (0.0175 inch) stainless steel wire
X Y
S.S. Ni-Ti .slrPll,~th
Present work* 1.x
Earlier estimate’ “I ( I .X)
S.S. /3-T; SWC’,l,&
Prebent work‘: 1.7
Earlier estimate”’ (I 6)
*See Table IV or Fig. 12.
-ISee Table V or Fig. I?.
table, Goldberg also reports the apparent flexure mod-
uli of elasticity of multistranded orthodontic wires
which, when divided by the modulus of elasticity of
S.S., generate the values of M, listed in Table V of
Burstone.’ When this normalization is done, the overall
difficulty with these measurements is apparent: That the
M, of all the wires is ultimately dependent upon a wire
for which consistent moduli have not been obtained,
namely, for S.S. Thereby, systematic errors may be
introduced into all of the M, values.
The other parameter incorporated into W, is C,. For
single-stranded wires of simple cross sections, C,
equalled the ratio of two wires, one of which (X) was
designated to be the 0.004 inch wire, via equation 24.
From such calculations, the W, which appear in Table
IV of Burstone’ may be generated. Note that W, is
equivalent to equation 19 for which
w,= gg, (25)
\ article which followed,“’ a simplification was made
M=EL
\ 0% * (26)
and
(27)
in which all K ‘s and n’s equal 1.
When the braided wires of Table V are studied,
however, the values for C, are not calculated as though
there were even n strands of diameter d but, rather, as
though there were only one strand of diameter D. This
may be seen when the C, values of Tables I and II are
compared to the C, values of Table V. Although the C,
values appear to represent an oversimplification, any
error of interpretation is circumvented by associating
the dimensionless parameters, K and n, with the M,
terms rather than with the C, terms (see equation 19 and
equations 25 to 27). Since K and n are clearly related to
Brrldirl~
Srif-fnrs.s
I.2
(0.X8)
Sl~ffwss
3.4
(1.X)
wire geometry, one might argue that, by incorporating
them into equation 26, the cross-sectional stiffness
number looses some meaning. Experimentally, how-
ever, the apparent flexure modulus measurements
shown in Table IT of the article by Goldberg and col-
IeagueP must, by definition, reflect the variations of
E, K. and n in the M, term. As a result, the indices (W,)
generated in both Tables IV and V of Burstone’ offer
an empirical solution to one of the three theoretical
expressions summarized in equations 18 to 20.
Elastic property ratios
At the same time that Burstone’s work was being
published, there appeared a paper which introduced the
concept of elastic property ratios.” As was the case for
wire-stiffness numbers, these ratios were a function of
both E and I. In Tables 1 and II, the same Wildcat
triple-stranded S.S. wire was reported alongside any
titanium wires of equivalent stiffness. In this and the
(ultimately the difference between equations 13 and
24), so that the elastic properties of a cable could be
treated as three independent strands of 0.008 inch S. S.
wire. This was the same assumption that ThurowX had
made previously; in the present case, however, the ap-
proximate nature of the data was highlighted by enclos-
ing all of these computations in parentheses. By using
Tables IV and V of Kusy and Greenberg”’ the elastic
strength, stiffness, and range ratios of an 0.012 inch
S .S. arch wire (Y) were compared to a 3 x 0.008 inch
S.S. wire (X) and equalled (1.8111.52) = 1.2, (0.8831
0.523) = 1.7, and (2.05/2.91) = 0.70 in bending. Al-
ternatively, the same information could have been cal-
culated from the general bending ratio formulas listed
in Figs. 2 to 4 of Kusy and Greenberg”’ or, more con-
veniently, from the nomograms of Fig. 4 of KusY.‘~
Using the revised physical properties of Table II, and
either equations 21 to 23 (Table III), or Fig. 11, the
Volume 86 Elastic property ratios of triple-stranded stainless steel wire 187
Number 3

same elastic property ratios now equalled 1.4, 2.2, and

0.63, respectively. Two years ago the elastic property
ratios of an 0.0175 inch S.S. triple-stranded wire ver-
sus an 0.0 16 inch Ni-Ti wire in bending were estimated
to be 1.8 (strength), 0.88 (stiffness), and 2.0 (range)
(Table I of Kusys or Table IV of Kusy and Green-
berg’“). In the present study, the wires gave 1.8, 1.2,
and 1.6, respectively (Table VI). For the case of a
3 x 0.008 inch S.S. wire versus an 0.016 inch P-Ti
arch wire, the strength, stiffness, and range ratios were
1.6, 1.8, and 0.88 (Table IV of Kusy and GreenberglO)
versus the present 1.7, 2.4, and 0.71 (Table VI). For
each of these wire alloy combinations, some differ-
ences were seen in the elastic property ratios, in part as Fig. 14. Simplified geometry of a triple-stranded arch wire.
a result of the physical properties assumed and in part Shown are the constructions required to determine the dis-
as a result of the more precise theoretical model used. tances from the wire centroid to the outermost reaches of the
wire and the general bending plane shape factor, A.
As long as the foregoing theoretical relationships are
employed within the elastic region of the stress-strain
curves, any immediate improvements should be di- was considerably stronger but also could accept over
rected toward the mechanical property measurements 50% more activation. On the other hand, an 0.016 inch
themselves, most of which have been adopted provi- wire made from the other titanium alloy (P-Ti) was
sionally herein and are either poor or ill-defined. Only more than twice as stiff, making it really more compa-
as both experimental and theoretical progress is made rable to an 0.012 inch S.S. arch wire.
will optimal results be obtained.

SUMMARY AND CONCLUSIONS Appendix

The elastic property ratios were derived for nth-
stranded wires. Compared to the earlier derivation, the DETERMINATION OF THE BENDING PLANE
strength, stiffness, and range equations now contain the SHAPE FACTOR FOR A TRIPLE-STRANDED
number of wire strands (n), the helical spring shape ARCH WIRE
factor (K) (in which is contained the helix angle of the Fig. 14 illustrates the centroidal distances of a triple-stranded
coil, LY, and the Poisson’s ratio, v), and the bending wire as the bending plane rotates through an arbitrary angle 0 from
A-A to B-B. The idealized model assumes that each cross section of a
plane shape factor (h). A general K versus cy plot wire strand perpendicular to its helical axis approximates a circle
(v = 0.28) was determined, as well as a h versus 0 plot (valid if LY z 65”). The distance from the wire centroid to a plane is
for the case of a wire having threefold symmetry derived as follows:
(n = 3).
CR&R = “e” + “f” + “g”
Using equations 18 to 20, Figs. 8 to 10, and Tables
I and II. the elastic property ratios of stainless steel in which “e” = _d
2 '
(S.S.), nickel-titanium (Ni-Ti), and beta titanium
(P-Ti) arch wires (forty-three combinations in all) were “f” = 2” sin0 ,
!i
calculated using an 0.0175 inch (3 x 0.008 inch) S.S.
and “g” = y COSH.
arch wire as the base line. Both the tabular format
(Tables I11 to V) and the nomographic format (Figs. 11 Since y = z sin+ and z =$/ costi ,
to 13) showed that this multistranded wire bore no re-
“ >1 = z sin+ cosH
semblance to an 0.018 inch S.S. wire other than in its g
slot engagement, and that the 0.010 inch S.S. wire had d
1 sm+
similar stiffness but only 80% of the strength of this cosH
multistranded wire. This is consistent with the rationale cod L

for designing a multistranded wire. Only one other wire From the wire geometry, I/I = 30” so that
investigated was really comparable to the 0.0175 inch d 1
(3 x 0.008 inch) multistranded arch wire evaluated, “ 1. -2 i-)2 case
and that was the 0.016 inch Ni-Ti wire. With an equiva- g -
L5
lent stiffness, the 0.016 inch Ni-Ti arch wire not only 2
 Ani. J. Orthod.
.Srymmher 1984

10. Kusy RP, Greenberg AR: Effects of composition and cross sec-
tion on the elastic properties of orthodontic wires. Angle Orthod
51: 325-341, 1981.
11. Kusy RP, Greenberg AR: Comparison of elastic properties of
nickel-titanium and beta-titanium arch wires. AM J ORTHOD 82:
Therefore, cH-a = 4,2 199-205, 1982.
12. Kusy RP: On the use of nomograms to determine the elastic
property ratios of orthodontic arch wires. AM J OR~HOD 83:
374-381, 1983.
13. McClymonds NL: Forming of stainless steel bar and wire, which
in which appeared in Forming of StainlessSteels,Metals Park, 1968,
American Society for Metals, pp. 27-39.
c = CH-H 3
14. Lyman T (editor): Metals handbook. Vol. 1. Properties and se-
d* = d ,
lection of metals, ed. 8, Metals Park, 1969, American Society
and for Metals, pp. 408-431, 664.
15. Databook 1975. Metal Prog 108: 44-45, I1 8, 1975.
A = (1 + sin0 + 3 cosH) 16. Cemica JN: Strength of materials, New York, 1966, Holt,
Rinehart & Winston, Inc., pp. 41, 155.
(see equations 9 and 14, respectively) 17 Eshbach OW, Souders M (editors): Handbook of engineering
fundamentals, New York, 1975. John Wiley & Sons, pp. 518.
520.
REFERENCES 18. Timoshenko S, Lessells JM: Applied elasticity, East Pittsburgh,
1. Andreasen GF, Hilleman TB: An evaluation of 55 cobalt substi- 1925, Westinghouse Technical Night School Press, p. 241.
tuted nitinol wire for use in orthodontics. J Am Dent Assoc 82: 19 Dilley GJ: Elastic property ratios of a triple stranded twist ortho-
1373-1375, 1971. dontic wire, M.S. thesis, University of North Carolina School of
2. Andreasen GF, Barrett RD: An evaluation of cobalt-substituted Dentistry, 1983.
nitinol wire in orthodontics. AM J ORTHOD 63: 462-470, 1973. 20 Kusy RP, Dilley GJ: Elastic modulus of a triple-stranded stain-
3. Morrow RE: Stored energy in orthodontic wire. Presented at less steel arch wire in three- and four-point bending. J Dent Res
International Association for Dental Research, Washington, (in press).
D.C., 1978. 21 Nielsen LE: Mechanical properties of polymers, New York.
4. Goldberg AJ, Burstone CJ: An evaluation of beta titanium alloys 1967, Reinhold Publishing Corp., p. 7.
for use in orthodontic appliances. J Dent Res 58: 593-600, 1979. 22 Dusenberry GR: Elastic bending properties in multistranded
5. Burstone CJ, Goldberg AJ: Beta titanium: a new orthodontic wire, M.S. thesis, University of North Carolina School of Den-
alloy. AM J ORTHOD 77: 121-132, 1980. tistry, 1974.
6. Andreasen GF: A clinical trial of alignment of teeth using a 23 Goldberg AJ, Morton J, Burstone CJ: The flexure modulus of
0.019 inch thermal nitinol wire with a transition temperature elasticity of orthodontic wires. J Dent Res 62: 856-858, 1983.
range between 31” C and 45” C. AM J ORTHOD 78: 528-537.
1980. Reprint requests to:
Burstone CJ: Variable-modulus orthodontics. AM J ORTHOD 80:
Dr. Robert P. Kusy
1-16, 1981.
Department of Orthodontics
Thurow RC: Edgewise orthodontics, ed. 4, St. Louis, 1982, The
Dental Research Center
C.V. Mosby Company, pp. 33-37.
University of North Carolina
Kusy RP: Comparison of nickel-titanium and beta titanium wire
Chapel Hill, NC 27514
sizes to conventional orthodontic arch wire materials. AM J
OR~HOD 79: 625-629, 1981.