0099-2399/97/2302-0077503.

00/0
JOURNAL OF ENDODONT{CS
Copyright 1997 by The American Association of Endodontists
SCIENTIFIC ARTICLES
Pri nt ed in U.S.A.
VOL. 23, No. 2, FI~BRUARY1997
Cyclic Fatigue Testing of Nickel-Titanium
Endodontic Instruments
J o h n P. P r u e t t , D D S , Da v i d J. C l e m e n t , D D S , a n d D a v i d L. C a r n e s , Jr . , P h D
Cycl i c fati gue of ni ckel -ti tani um, engi ne-dri ven in-
st rument s was st udi ed by determi ni ng the effect of
canal curvature and operati ng speed on the break-
age of Li ghtspeed i nstruments. A new method of
canal curvature evaluation t hat addressed bot h an-
gl e and abrupt ness of curvature was i ntroduced.
Canal curvature was si mul ated by const ruct i ng six
curved stai nl ess-steel gui de t ubes wi t h angl es of
curvature of 30, 45, or 60 degrees, and radii of
curvature of 2 or 5 mm. Size #30 and #40 Li ght-
speed i nst rument s were pl aced t hrough the gui de
t ubes and the heads secured in the col l et of a
Magtrol Dynamometer. A si mul ated operati ng load
of 10 g-cm was applied. I nst rument s were abl e to
rotate freel y in the t est apparat us at speeds of 750,
1300, or 2000 rpm until separati on occurred. Cy-
cl es t o failure were determi ned. Cycl es t o fai l ure
wer e not affected by rpm. I nst rument s di d not sep-
arate at t he head, but rather at the poi nt of maxi-
mum flexure of the shaft, correspondi ng t o the
mi dpoi nt of curvature wi thi n the gui de tube. The
i nst rument s wi t h l arger di amet er shafts, #40, failed
after si gni fi cantl y f ewer cycl es than di d #30 i nstru-
ments under i denti cal t est condi ti ons. Multivari-
able anal ysi s of variance i ndi cated t hat cycl es t o
failure si gni fi cantl y decreased as t he radi us of cur-
vature decreased f rom 5 mm t o 2 mm and as t he
angl e of curvature i ncreased great er t han 30 de-
grees (p < 0.05, power = 0.9). Scanni ng el ectron
mi croscopi c evaluation revealed ducti l e f ract ure
as t he fati gue failure mode. These resul ts i ndi cate
that, f or ni ckel -ti tani um, engi ne-dri ven rotary in-
struments, the radi us of curvature, angl e of curva-
ture, and i nstrument size are more i mport ant than
operati ng speed f or predi cti ng separati on. This
st udy support s engi neeri ng concept s of cycl i c fa-
ti gue failure and suggest s t hat standardi zed fa-
ti gue t est s of ni ckel -ti tani um rotary i nst rument s
shoul d i ncl ude dynami c operati on in a flexed state.
The resul ts al so suggest t hat t he effect of t he ra-
di us of curvature as an i ndependent variable
shoul d be consi dered when evaluating st udi es of
root canal i nstrumentati on.
77
Instrument separation is a serious concern in endodontics. Because
stainless-steel instruments usually deform before they separate,
dentists can inspect them for visible signs of instrument deforma-
tion. A deformed instrument usually shows severe bending or
unwinding of the flutes, indicating that the elastic limit of the metal
has been exceeded and that the instrument should be discarded.
Nickel-titanium (Ni-Ti) endodontic instruments were intro-
duced to facilitate instrumentation of curved canals. Ni-Ti instru-
ments are superelastic and will flex far more than stainless-steel
instruments before exceeding their elastic limit (1, 2). This flexi-
bility is an important property that allows preparation of curved
canals while minimizing transportation (3). Despite this increased
flexibility, separation is still a concern with Ni-Ti instruments, and
they have been reported to undergo unexpected fracture (4). Sep-
aration can occur without any visible signs of previous permanent
deformation, apparently within the elastic limit of the instrument.
Visible inspection, therefore, is not a reliable method for evalua-
tion of used Ni-Ti instruments.
The American Dental Association (ADA) Specification No. 28
(5) stipulates the design requirements, testing procedures, and
physical properties for endodontic instruments. Tests detailed in
ADA Specification No. 28 are conducted in a static mode. Instru-
ments are fixed at the tip and are twisted until failure to determine
instrument resistance to fracture under torsional loading. Angular
deflection is used to detertnine stiffness when instruments are bent
to predetermined angles of 30, 45, and 90 degrees. Tests described
in ADA Specification No. 28 are appropriate for establishing
minimum strength requirements of Ni-Ti hand instruments placed
under a static load (I, 2, 6, 7), but ADA Specification No. 28 is not
an appropriate test of the dynamic characteristics of engine-driven
rotary instruments. All Ni-Ti, engine-driven rotary systems require
that the instrument be activated at a predetermined rpm before
insertion into the canal. Therefore, to test rotary instruments, the
78 Pr ue t t et al .
tip should not be statically locked, but al l owed to rotate freely.
Because ADA Speci fi cat i on No. 28 does not consi der canal ge-
ometry, it also does not consi der fat i gue and breakage of rotary
instruments operated in fl exed condi t i ons whi l e preparing curved
canals. The phenomenon of repeat ed cycl i c met al fatigue caused
by canal curvatures may be the most i mport ant fact or in instrument
separation (2, 8).
Ni - Ti instruments exhi bi t superelastic behavi or and remai n
wi t hi n their elastic l i mi t in situations that woul d cause permanent
plastic deformat i on in stainless-steel instruments (1, 2). Met al
fatigue (cyclic fatigue) l eadi ng to fracture and separation can occur
bel ow the elastic l i mi t of t he i nst rument (wi t hout permanent de-
format i on) through known mechani sms cal l ed slip bands (9). Such
fatigue mechani sms occur mi croscopi cal l y and are not visible to
the eye. Cycl i c fat i gue is caused by repeat ed t ensi l e-compressi ve
stress. Rotation subjects an endodont i c instrument to both tensile
and compressi ve stress in the area of the curve. Instruments placed
in curved canals deform, creating stress wi t hi n the instrument. Hal f
of the instrument shaft on the outside of the cur ve is in tension, and
that hal f of the instrument shaft on the inside of the curve is in
compression. Each rotation within a cur ved canal causes an instru-
ment to undergo one compl et e t ensi on-compressi on stress cycle.
Thi s is the most dest ruct i ve form of cycl i c l oadi ng (9).
Al t hough some dynami c testing has been performed on Ni -Ti
and stainless-steel endodont i c instruments (2, 8, 10- 13) , fatigue
testing remains undevel oped as an endodont i c instrument testing
met hodol ogy (2, 10, 11). Stress l evel s during cycl i c loading are
general l y dependent on the actual shape of the curvature and the
applied loads. This stress is greatest in the area of curvature (14).
Mor e severe bends create greater stress (14); and larger, stiffer
instruments will exper i ence greater stress than smal l er instruments
when confi ned to the same curved canal shape.
The purpose of this study was 2-fol d; first to introduce a new
met hod of det ermi ni ng canal curvature. A new met hod was nec-
essary because canal curvat ure had to be descri bed in parameters
known to determine i nst rument stress and t herefore affect instru-
ment fatigue life. Second, cycl i c fatigue was exami ned in a com-
merci al l y available Ni - Ti instrument system. The study exami ned
the effect of instrument operat i ng rpm, shaft di amet er, and canal
curvature (including angl e of curvat ure and radius of curvature) on
the fatigue life and resulting separation of Ni - Ti al l oy endodontic
instruments.
MAT E R ~ L S AND ME T HODS
Thi s study required defi ni ng the paramet ers of canal curvature
in a more exact manner than general l y used in endodont i c research.
Canal curvature is usual l y depi ct ed by a single parameter, an
arbitrary angle measured in degrees as descri bed by Schnei der (15)
(Fig. 1). To det ermi ne the degree of root curvature, he scribed a
line parallel to the l ong axis of the canal. A second line was drawn
from the apical foramen to intersect wi t h the first line at the point
wherei n the canal began to l eave the l ong axis of the canal. The
acute angl e formed ( a) was defi ned as the degree of root curvature.
The shape of any root canal curvat ure is more accurately de-
scribed using two parameters: angle of curvat ure ( a) and radius of
curvat ure (r) (Fig. 2). To det ermi ne t hese parameters, a straight line
is drawn along the l ong axis of the coronal portion of the canal. A
second line is drawn al ong the long axis of the apical port i on of the
canal. There is a poi nt on each of these lines at whi ch the canal
deviates to begin (point a) or end (point b) the canal curvature. The
J our na l of E n d o d o n U c s
30
A B
FIG 1. Degree of root canal curvature. The met hod descri bed by
Schnei der f or det ermi ni ng canal curvat ure uses onl y one paramet er
t o defi ne t he angl e. Tt =, t wo simulated teeth demonst rat e how a
di fference in abruptne~,, of t he curve (radius of curvature) wi l l vary
t he degree of curvat ure as measured by thi s met hod. A has an angl e
of 43 degrees and B has an angle of 52 degrees, even t hough bot h
angl es measured accordi ng t o our met hod equal ed 60 degrees.
Locati on of t he curve al ong t he canal wi l l al so change t he measured
angle.
curved port i on of the canal is represented by a ci rcl e with tangents
at points a and b. The angl e of curvature is the number of degrees
on the arc of the ci rcl e bet ween points a and b. Angl e of curvat ure
can also be defi ned by the angle formed (c~ 1 and c~2) by perpen-
dicular lines drawn from the points of devi at i on (a and b) that
intersect at t he cent er of the circle. The l engt h of these lines is t he
radius of the ci rcl e and defines the radius of the canal curvature.
The radius of curvat ure (r I and r 2) is the length of the radius of the
ci rcl e measured in mi l l i met ers. The radius of curvat ure represents
how abruptly or severel y a specific angl e of curvat ure occurs as the
canal devi at es from a straight line. The smal l er the radius of
curvature, t he more abrupt the canal deviation. The paramet ers of
angl e of curvat ure and radius of curvat ure are i ndependent of each
other. Canal s can have the same angle of curvat ure whi l e havi ng
different radii of curvature, resulting in more abrupt curves.
Because cont rol l ed cycl i c fatigue studies on Ni -Ti , engi ne-
dri ven endodont i c instruments had not pr evi ousl y been performed,
pi l ot studies were requi red to devel op a test apparatus and speci fy
the test parameters. The parameters of radius of curvat ure and
angl e of curvat ure needed to be defi ned narrowl y enough to al l ow
accurate data col l ect i on. Extracted teeth with severe root curva-
tures suggest i ng possi bl e high-stress condi t i ons l eadi ng t o instru-
ment separation were radiographed. Radi ographs were also made
of teeth treated cl i ni cal l y in which separation of Ni - Ti Li ght speed
instruments had occurred. The radius of curvat ure of the canals in
these teeth was det ermi ned by using a ci rcl e gauge al i gned over the
i mage of t he canal profile. Additionally, size #30 Li ght speed
instruments were operated through bent needles si mul at i ng in-
creasi ng canal curvatures. Instruments tested in canals havi ng a
5- mm radius of curvature wi t h an angl e of curvat ure of 30 degrees
separated as a result of cycl i c fatigue. Stress l evel s i nduced by
l esser curvat ures did not result in instrument separation (occurri ng
in over 30 mi n at a speed of 2000 rpm) wi t hi n the t i me f r ame for
the test apparatus or data collection. Thus, angles of curvat ure of
30, 45, and 60 degrees were chosen for eval uat i on at radii of
curvature of 2 and 5 mm.
To si mul at e instrumentation conditions during use, a cl i ni cal l y
Vol. 23, No. 2, February 1997
o~=60 o~z =60
r2=2rnrn
' A B
FIG 2. Radius of curvat ure and angl e of curvature. New met hod of
descri bi ng canal geomet r y using t wo parameters. Radius of curva-
ture (r) and angl e of curvat ure (e) are det ermi ned on t he same teeth
as in Fig. 1. These teeth accurat el y depi ct t he gui de t ubes used in
thi s st udy t hat had a 60-degree angl e ((~1 = (x2). Angl e of curvat ure
(c~) is det ermi ned by t he angl e f ormed by t he lines t hat i ntersect at
t he ci rcl e' s center. These t wo lines are perpendi cul ar t o t he lines
drawn al ong t he l ong axes of t he coronal and api cal porti ons of t he
root canal space. Poi nts a and b are t he poi nts wherei n t he canal
devi at es from t he strai ght lines and ei ther begi n or end t he curved
porti on of t he root canal space. Angl e ((~) is t aken t o be t he angl e
f ormed by t he arc in degrees bet ween poi nts a and b. The ar c lies
on a ci rcl e whose size is speci fi ed by it' s radius, and t he ci rcl e' s
radi us is taken t o be t he radi us of curvat ure of t he canal in t hat local
area. The ci rcl e' s radius is t he radi us (rl and r2) of t he curved porti on
of t he root canal space and defi nes how abrupt l y t he canal curves.
The canal geomet r y of these t wo teeth differs onl y in t he radius of
curvat ure (r~ and r 2 ) , whereas t he angl e of curvat ure ((x 1 and (x2)
equal s 60 degrees. A represents a sweepi ng canal curvat ure havi ng
a 5- mm radi us of curvat ure (rl). B represents an abrupt canal cur-
vat ure havi ng a 2- mm radius of curvature (r2). Clinically, t he radi us of
curvat ure and angl e of curvat ure of any root canal space coul d be
measured using thi s t echni que wi th t he aid of a circle gauge.
appl i cabl e head t orque load was determined. Data were obtained
by measuri ng the resultant t orque when hi ghl y cal ci fi ed canals in
l ower central incisors were i nst rument ed whi l e mount ed in a dy-
namomet er. Canal preparation with successi vel y larger Li ght speed
instruments was performed. When small instruments di d not ac-
t i vel y engage and cut the dentin walls during the preparation of
l ower central incisors, no percept i bl e t orque val ues coul d be re-
peatably recorded. When act i ve dentin cutting occurred, recorded
t orque val ues ranged from 6 to 14 g-cm. Therefore, a preset t orque
of 10 g- cm was applied to the test instruments duri ng all of the test
conditions.
Si ze #30 and #40 Li ght speed instruments (Li ght speed Technol -
ogies, Inc., San Ant oni o, TX) were evaluated. Li ght speed instru-
ments were chosen because their noncut t i ng shafts offer the ad-
vant ages of predi ct abl e shaft diameters. This also el i mi nat ed the
vari abl es of instrument taper and flute desi gn on met al fatigue and
breakage. Si ze #40 was sel ect ed because it is t he mi ni mum mast er
apical size advocat ed for the Li ght speed technique. Si ze #30 was
selected because it is comparat i vel y smal l er in size and has a head
design that is nearly parallel. In addition, Li ght speed instruments
were chosen because their use has been advocat ed over a wi de
range of operat i ng speeds. Ten instruments of each size were tested
in each of the six di fferent experi ment al canal shapes (3 angles c~
each of 2 radii r) at three di fferent speeds (750, 1300, and 2000
rpm). This produced 18 test groups for each i nst rument size. Each
instrument was operated to failure.
Cyclic Fatigue of Ni-Ti Rotary Instruments 79
Ni-Ti Lightspeed Instrument
Test Support Block
Guide Tube
Holddown
. ~ " , . Clamp
I I
Dynamometer Collet
LJ
Input shaft t o dynamometer, dat a aquisition computer
FIG 3. Test apparatus. Cut - away vi ew of a Li ghtspeed i nstrument
inserted through t he l umen of a stai nl ess-steel gui de t ube cl amped
in t he dynamomet er collet. The i nstrument shank butts up agai nst
t he end of t he gui de tube, ensuring t hat each i nstrument was pl aced
t o t he same dept h wi thi n t he artificial canal. The dynamomet er
al l owed free rotati on of t he i nstrument duri ng t he cycl i c fati gue test,
yet was capabl e of appl yi ng torsi onal prel oad t o si mul ate clinical
u s e ,
Bot h radius of curvat ure and angl e of curvat ure were used to
defi ne accurat el y the curvat ures of artificial canals used in this
study. Art i fi ci al canals were fabricated from 18-gauge, stainless-
steel needl es havi ng an internal di amet er of 0.83 ram. Ref er ence
radii were measured to the i nner aspect of the cur ve of the gui de
tubes. A 2- mm radius represent ed an abrupt curvat ure; and a 5 mm
radius represent ed a sweepi ng curvat ure (Fig. 2), each with angles
of curvat ure of 30, 45, and 60 degrees. This resulted in six artificial
canals. One end of the artificial canal butted against the flat shank
of the instrument, and the cent er of radius in the curved segment
of the canal was 7 mm f r om the tip of the i nst rument (Fig. 3). The
head of the instrument prot ruded - 2 mm from the artificial canal.
A test stand (Fig. 3) was const ruct ed that al i gned the i nner aspect
of the artificial canals with the cent erl i ne of the dynamomet er and
hel d t hem securel y in place. Instruments were inserted through the
canals, and the heads were cl amped in the col l et of t he Magt rol
Dynamomet er (Magtrol, Inc., Buffal o, NY). Cl amped instruments
coul d rotate freel y wi t hi n the tube assembly. An Asept i co electric
handpi ece (Asept i co model AEU- 17, Asept i co, Inc., Woodi nvi l l e,
WA) was attached to the instrument shank. The experi ment al
operat i ng speed of 750, 1300, or 2000 rpm was adjusted usi ng t he
dynamomet er readout that pr ovi ded a more accurate rpm det ermi -
nation than did the readout on t he Asept i co control box. Rot at i on
at the desi red speed was achi eved by act i vat i ng the Asept i co
electric handpiece.
This study used i nst rument at i on equi pment speci fi ed in ADA
Speci fi cat i on No. 28 for torsional testing: the Magt rol Dynamom-
80 Pruett et al.
eter. The dynamometer allows the torque applied to the instrument
to be preset and measured while the instrument rotates. All instru-
ments were operated with a head load of 10 g-cm torque. A
software program (Sunset Resources, Inc., San Antonio, TX) was
written for an IBM-compatible computer to permit data collection
while the instrument was rotating. An 18-rain time limit for data
collection was set as limited by the data collection program, but
instrument separation always occurred within 5 min of beginning
the test. Data collection from the dynamometer consisted of con-
tinuous torque, and rpm readouts were comprised of 10 averaged
readings/s. The time of instrument failure was determined when
the torque values suddenly dropped, indicating instrument separa-
tion. The number of cycles to failure was calculated from the rpm
data by multiplying the rpm by the time to failure. Data analyzed
were the number of cycles to failure for each instrument tested
under the specified angle of curvature, radius of curvature, rpm,
and instrument size conditions. Data were analyzed by analysis of
variance (ANOVA), with the Neuman-Kuels procedure serving as
the post-hoc test using a 95% confidence level.
Fractured instrument fragments were collected by group. Rep-
resentative samples of the fracture surface from the various groups,
and sizes were examined under scanning electron microscopic
magnification with a JSM-840 scanning electron microscope
(Japanese Electronic and Optical Ltd., Japan). Another group of
instruments was partially cycled (75 to 80% to failure) using mean
cycles-to-failure data from the matching experimental groups.
These instruments were then mounted in both unflexed and flexed
positions with the area of interest visible in side view, and exam-
ined under scanning electron microscopy (SEM). For comparison,
new same-sized instruments were mounted in the same manner and
examined under SEM.
RESULTS
Operating rpm
Multivariable ANOVA indicated that operating rpm was not a
significant factor affecting cycles to failure (p < 0.10 for size #30;
p < 0.06 for size #40) (Fig. 4). No difference in cycles to failure
was recorded between the 750, 1300, and 2000 rpm groups within
the two instrument sizes. This was further verified by one-way
ANOVA, wherein only 2 of the possible 36 groups in the size #30
and #40 instrument groups had cycles to failure that varied when
rpm was the only independent variable evaluated.
Instrument Shaft Diameter
Shaft diameter had a significant effect on cycles to failure (p <
0.00001). Cycles to failure significantly decreased with an increase
in instrument size. At all radii and angles tested, the larger size #40
instruments (0.269 mm shaft diameter) failed at significantly fewer
cycles to failure than the size #30 (0.223 nun shaft diameter),
regardless of rpm (Fig. 4).
Radius of Curvature
Cycles to failure significantly decreased as the radius of curva-
ture decreased (Fig. 5). The more abrupt 2-ram radius groups had
significantly fewer cycles to failure than the 5-mm radius groups
(p < 0.00001). This was consistent for both instrument sizes.
Journal of Endodonti cs
800
700
600
500
400
300
200
100
0
750 rpm
I I size #30
l
1300 rpm
J
2000 rpm
Size #40
FiG 4. Effect of operating rpm and shaft diameter. Cycles to failure
is depicted for each instrument size at operating speeds of 750,
1300, and 2000 rpm. The asterisk and its absence indicates statis-
tically similar groups, which shows clear statistical differences ac-
cording to size. Cycles to failure was not affected by operating
speed within the size #30 and #40 instrument groups. The size #40
instruments had significantly fewer cycles to failure than size #30
instruments.
Radius of curvature was the most consistent parameter describing
root canal curvature that predicted instrument separation occurring
as a result of cyclic fatigue.
Angle of Curvature
Instruments operated in 45- and 60-degree angles of curvature
were not significantly different and failed at significantly fewer
cycles than instruments operated at 30 degrees (p < 0.00001).
When the effect of angle was examined by analyzing the data from
four subgroups of 2 and 5 mm radius, for each instrument size, this
same trend was apparent (Fig. 6). The exception was the size #40,
2-mm radius group. Each successive increase in angle of curvature
(30-, 45-, and 60-degree groups) resulted in significantly decreased
cycles to failure.
Instrument Separation
Instrument separation always occurred at the point of maximum
flexure within the artificial canals, the midpoint of the curved
segment of the tubes. Separation never occurred immediately be-
hind the head of the instrument (Fig. 7).
SEM Analysis
Partially fatigued instruments seemed normal both to the naked
eye (Fig. 8A) and when viewed under SEM in an unflexed state
(Fig. 8B). Under flexure (Fig. 8C), opening of the fractures was
evident under SEM and generally seemed to be associated with
flaws on the instrument surface (Fig. 8D). These initiated fractures
were located at the point of maximum curvature within the tube
along the instrument shaft.
The fatigue failure of the instrument can be broken down into
stages (Fig. 9). Evidence of these stages was visible upon higher
magnification (1900) under SEM examination (Fig. 9B). Stage I,
crack initiation and growth, was characterized by the smooth,
almost featureless area at the periphery of the fracture face. Stage
Vol. 23, No. 2, February 1997 Cyclic Fatigue of Ni-Ti Rotary Instruments 81
A
1000
900
800
700
600
500
400
O
300
O
200
100
0
30 degrees 45 degrees 60 degrees
SIZE #30
5mm Radius ~ 2ram radius
A
1000
gO0
8OO
~ 700
~ 600
O 500
~ 400
~ 300
O
2O0
100
0
5mm Radius 2ram Radius
SIZE #30
30 degrees ~ 45 degrees ~ 60 degrees
B
1000
900
800
700
600
500
400
300
200
100
0
30 degrees 45 degrees 60 degrees
SIZE #40
] 5rnm Radius ~ 2ram Radius
FIG 5. Effect of radius of curvature. The effect on cycles to failure in
the size #30 (A) and #40 (B) instruments is depicted in accumulative
rpm groups. The more abrupt 2-ram radius curvatures demon-
strated significantly fewer cycles to failure than the 5-ram radius
curvatures (*), regardless of angle.
II, crack propagation, was typified by striations. Each striation
represents the progression of the crack caused by tension during
one rotation of the instrument. Fractures propagated from the
periphery of the instrument toward the center. Stage ]II, ultimate
ductile fracture, was seen in the center of the fracture surface (Fig.
9A). Microvoid formation and dimpling typically characterized
ductile fracture.
DI SCUSSI ON
This is the first comprehensive study of cyclic fatigue that
incorporates canal geometry as a factor in the breakage of Ni-Ti
rotary instruments. All instruments were allowed to rotate freely
under a 10 g-cm torsional load through artificial curved canals until
separation occurred. Instruments were not statically loaded until
failure as described in ADA Specification No. 28 (5), and instru-
ment separation did not occur because of static torsional overload-
ing. The results indicate that cyclic fatigue is an important factor
in the separation of Ni-Ti rotary instruments used clinically, and
that specific measurable parameters of canal geometry and instru-
B
=.
O
I -
J
o
>.
o
1000
900
800
700
600
500
400
300
200
100
0
i
5ram Radius 2ram Radius
SIZE #40
30 degrees ~ 45 degreee ~ 60 degrees
FIG 6. Effect of angle of curvature. Angle subgroups are arranged
according to radius of curvature for size #30 (A) and #40 (B) instru-
ments. Forty-five degree or 60-degree angles in both instrument
size groups had significantly fewer cycles to failure than the 30-
degree angle groups (*). The size #40, 2-mm radius subgroup varied
from this pattern, with all angles having significantly different cycles
to failure (** and ***).
merit size exist, which significantly reduce instrument cyclic fa-
tigue life.
There is a clear need for the development of a test protocol for
the cyclic fatigue of Ni-Ti, engine-driven rotary instruments.
Brantley et al. (10) and Luebke et al. (11), in studies on stainless-
steel rotary instruments, have also advocated the development of a
cyclic fatigue test protocol that simulates clinical conditions of
instrument use. Serene et al. (2) published a short volume on Ni-Ti
instruments, including results from a preliminary study on rotation
to breakage (cyclic fatigue) of hand instruments. A significant
difference was found between conventional stainless-steel files and
all Ni-Ti file groups. It was also noted that in rotation-to-breakage
studies of Ni-Ti files that allowed free tip rotation, the range of
results within each same-size group was large and unpredictable.
That preliminary information on rotation-to-breakage of hand in-
struments stated that the instruments were rotated in a 90-degree
metal tube until they broke. Defining the test parameters in such a
limited manner, without strictly controlling the variables of curve
geometry, may explain the variability in rotation-to-breakage (cy-
82 Pruett et al. Journal of Endodonti cs
FIG 7. Location of separation. Randomly selected size #40 Light-
speed instruments that were cycled to failure. All instruments tested
showed the same pattern of separation, with fractures occurring
within the guide tube at the midpoint of the curved segment.
A
D
FIG 8. Partially separated instruments. Partially separated instru-
ments appear normal to the eye (A), and under SEM evaluation in an
unflexed state at x150 original magnification (B). Arrows indicate
the location of the fracture in-the unflexed instrument, which is
closed and not apparent. Upon mounting the instrument shaft in a
flexed (stressed) position (C), incomplete fractures became appar-
ent under SEM examination at x150 original magnification (D).
Arrows indicate the same site of examination of the instrument shaft,
which correlated to the midpoint of the curved segment of the guide
tube.
cles to failure) results and the conclusion drawn concerning Ni-Ti
instruments. Parameters including radius of curvature, angle of
curvature, instrument size, and the point of maximal instrument
flexure were all found to have a significant effect on the number of
cycles to failure and location of breakage in our study. All of these
factors should be considered when developing a test protocol for
the examination of cyclic fatigue.
Mathematics and engineering principles (14) suggest that a
system is needed that describes canal geometry in terms that
consistently describe the state of stress and cyclic fatigue on a
flexed instrument shaft. The independent variables that describe
canal shape are radius of curvature and angle of curvature. The
usual method of measuring canal curvature published by Schneider
(15) in 1971 provides a means of estimating canal shape using a
single parameter: an estimated arbitrary angle (Fig. 1). This
method does not consider radius of curvature as a measurable
aspect of canal geometry. Using only one parameter to describe
canal shape does not differentiate the canal shape by abruptness or
radius of curvature. Two canals measured at the same angle in
FIG 9. Stages of instrument fracture. A shows the fracture surface of
a size #30 instrument at 300 original magnification. The boxed area
is enlarged in B having a x1900 original magnification. The asterisk
indicates the area of crack initiation on the periphery. Crack prop-
agation is seen in B as striations marked by the arrows. Each
striation represents a "jump" in the fracture surface that occurred
under tension in that area during rotation. Ultimate ductile fracture is
seen in the central area of A, characterized by the presence of
microvoid formations and dimpling.
degrees by the Schneider method could have very different radii or
abruptness of curvatures, thus having a very different impact on the
difficulty of canal instrumentation and instrument fatigue.
The more exact method of determining canal curvature and
geometry described in this study considers the radius of curvature
(r) and the angle of curvature (a) separately (Fig. 2). These are
terms that mathematically specify the shape factors of a curved
root canal system. The radius of curvature (in millimeters) is the
radius of a circle that coincides with the path taken by the canal in
the area of the most abrupt curvature. A more abrupt curve cor-
responds to a smaller radius of curvature. The angle of curvature
is the degrees arc formed between the points of deviation on the
circle, or the angle formed between the perpendicular lines drawn
from the tangents intersecting at the center of the circle (Fig. 2).
The angle of curvature is independent of the radius. Thus, two
canals with the same degree of curvature can have radically dif-
ferent radii. Pilot studies were required to define systematically the
range of parameters for this study on cyclic fatigue. The radii of
curvature and angles of curvature chosen could be considered some
of the more significant that are encountered in clinical practice.
When the simulated canals in this study were measured using the
Schneider method, the 2-mm radius, 30-, 45-, and 60-degree angle
groups measured 26, 40, and 52 degrees, respectively. The 5-mm
radius, 30-, 45-, and 60-degree angle groups measured 24, 35, and
43 degrees, respectively. Authors have cited average canal curva-
Vol. 23, No. 2, February 1997
tures for mandi bul ar molars measured using the Schnei der met hod
bet ween 20 to 30 degrees, with ranges bet ween 6 to 48 d e g r e e s
(16). Our 30-degree angle groups are wel l within the averages of
canal measurement to be cl i ni cal l y relevant, whereas the angles
l arger than 30 degrees woul d be expect ed to be encount ered less
frequently.
Of the t wo canal -shape parameters, angl e and radius of curva-
ture, radius of curvat ure was the most si gni fi cant fact or of canal
shape affect i ng the number of cycl es to failure of Ni -Ti , engi ne-
dri ven rotary instruments. Consi st ent l y, a decrease in the radius of
curvat ure from 5 mm to 2 mm radius resulted in a si gni fi cant
decrease in the number of cycl es to failure (Fig. 6). Expl anat i on of
the effect of radius is compl ex, but can be grossly si mpl i fi ed by
stating that stress on the instrument is i nversel y proport i onal to the
radius of curvature. Therefore, as the radius of curvat ure decreases,
i nst rument stress and strain increases, and the fatigue life de-
creases. To our knowl edge, radius of curvat ure has not been
previ ousl y consi dered as a separate i ndependent vari abl e in any
i nst rument at i on study.
Cl i ni cal l y, the sweepi ng curvat ure of the 5- mm radius groups,
al t hough chal l engi ng, is a much less di ffi cul t canal curvat ure to
manage than the abrupt curvat ure represent ed by the 2- mm radius
groups. Abrupt curvat ures predomi nat e in the apical third of the
canal or in di l acerat ed roots. Bot h condi t i ons are associ at ed with an
i ncreased i nci dence of i nst rument separation. Degr ee of root cur-
vat ure and instrument stiffness have been i mpl i cat ed in canal
preparation errors, such as i nst rument breakage, l edgi ng, bl ock-
ages, zipping, and perforation. However , radius of curvat ure with
its resultant i ncreased stress on endodont i c instruments may also be
a si gni fi cant factor cl i ni cal l y cont ri but i ng to i nst rument breakage
and canal transportation. These areas should be studied more
t horoughl y, and the clinical effect of radius of curvat ure on instru-
ment breakage and root canal preparation shape should be more
cl earl y defined.
The 30-degree angl e groups had si gni fi cant l y great er cycl es to
failure than the 45- and 60-degree angle groups, regardless of size
(Fig. 6). Thi s was an unexpect ed result when the engi neeri ng
literature was revi ewed, wherei n the effect of angl e al one s e e ms
negl i gi bl e or uncertain. It was suspected that the 0.83 mm di amet er
stainless-steel tubes used in this study, al t hough al l owi ng free
rotation of the instruments, did not suffi ci ent l y constrain the in-
strument shafts in the 30-degree angle groups. Thi s l ack of con-
straint, al t hough not affect i ng the overal l angle, did slightly in-
crease the ef f ect i ve radius of curvature, al l owi ng the i nst rument
shaft to fol l ow a larger radius path within the gui de tube. Depend-
i ng on the angl e of the tube and the instrument shaft di amet er, the
i nfl uence of angl e of curvat ure was general l y consistent in both the
2- and 5-ram groups. However , in the most hi ghl y stressed states
studied (size #40 instruments in the 2-ram radius), i ncreasi ng angl e
resulted in a si gni fi cant decrease in cycl es to failure bet ween all
three angle groups. Future studies should address this pot ent i al
variable, more preci sel y cont rol l i ng i nst rument constraint through-
out the defi ned curvature.
The Li ght speed i nst rument is one of several Ni - Ti systems
avai l abl e to the practitioner. These instruments have a small cut-
ting head, with a noncut t i ng pi l ot tip. The shaft of the i nst rument
i s smooth, noncutting, and has a di amet er that is smal l er than the
head. These features make Li ght speed instruments good experi -
ment al instruments in several respects. In particular, the desi gn of
the apparatus used for cycl i c fatigue testing coul d be si mpl i fi ed,
al l owi ng the apparatus to create test stresses similar to t ypi cal
Cyclic Fatigue of Ni-Ti Rotary Instruments 83
fatigue-test equi pment commonl y used in ot her met al l urgi c appli-
cations. The sl ender shaft al l ows qui t e severe bends wi t hout per-
manent deformat i on, thus l endi ng i t sel f to a wi der range of cur-
vat ure conditions. A predictable met al shaft di amet er el i mi nat ed
havi ng to account for additional variables, such as i nst rument
flutes and taper. These characteristics, however, do not prevent the
separation of Li ght speed instruments. Al l instruments tested in this
study separated within a 5-mi n t i me limit.
Anot her advant age of usi ng the Li ght speed i nst rument in this
study was that operat i on has been advocat ed over a wi de range of
speeds. To i nvest i gat e the effect of r pm on cycl i c fatigue, testing
was performed at 750, 1300, and 2000 rpm. The three rpm groups
wer e not statistically di fferent from one anot her (Fig. 4). Upon
rotation, ri si ng stress l evel s cause movement of di sl ocat i on defect s
and breaks at omi c bonds wi t hi n the cryst al l i ne mat ri x of the alloy,
l eadi ng to crack initiation and propagation. The sharpness of the
crack tip causes ext r eme stress concent rat i on exceedi ng the
strength of the met al at the crack tip in tension. The fracture
propagates a short di st ance under tension, then duct i l e deformat i on
blunts the crack tip, stopping the propagation. In the ot her hal f of
the cycl e, under compressi on, the alloy buckl es mi croscopi cal l y
and the crack resharpens (9). Thi s mi nut e propagat i on occurs on
the order of 10- 6 s or approxi mat el y at the speed of sound in metal
(l 7), and the propagat i on event is compl et ed much fast er than the
instrument is rotating. Thus, cycl i c fatigue and the ul t i mat e number
of cycl es to failure are unaffect ed by the rpm.
It is i mport ant to r ecogni ze that an i nst rument has a mean
number of cycl es to failure that is det ermi ned by speci fi c param-
eters of canal radius, canal angle, and i nst rument diameter. A
hi gher rpm will consume the useful life of the i nst rument much
faster than a l ower rpm. Thus, a l ower rpm woul d be benefi ci al and
pr ovi de a greater clinical life, more sl owl y using the finite number
of cycl es to failure available. This is especi al l y true when a Ni -Ti
rotary instrument is used in severe curvat ure condi t i ons suspected
to i nduce stress fatigue. Al t hough a wi de range of operat i ng speeds
has been advocat ed for Ni -Ti , engi ne- dr i ven systems, those rec-
ommendat i ons must take into account more than j ust cycl i c fatigue
as a cause of instrument separation. Nonet hel ess, the effect of rpm
when studying the breakage of Ni - Ti instruments has not been
previ ousl y exami ned and, in the present test conditions, rpm had no
ef f ect on cycl es to failure.
The results of our i nvest i gat i on conf i r med that instrument life
span is i nversel y proport i onal to i nst rument size 0 2 , 13) and that
met al fatigue is deepl y i mpl i cat ed in fi l e breakage (2, 8). Larger
di amet er instruments had a l ower fat i gue life, resul t i ng in signif-
icantly f ewer cycl es to fai l ure than smal l er di amet er instruments
(Fig. 5). The di fference in average shaft di amet ers bet ween t h e s e
t wo instruments is onl y 0. 046 ram. The larger di amet er i nst rument
was under more stress wi t hi n the confi nes of a defi ned curvature.
Therefore, consi deri ng cycl i c fat i gue as a cont ri but or to i nst rument
failure, larger instruments should not be consi dered safer or stron-
ger in clinical practice. In fact, rotary Ni -Ti i nst rument s wi t h l arger
di amet ers woul d be expect ed to separate at f ewer cycl es than
smal l er di amet er instruments.
The results of our study should not be interpreted as a fi xed
mean number of cycl es to failure for ei t her size #30 or #40
Li ght speed instrument during cl i ni cal use. Two i mport ant points of
expl anat i on are required. First, the poi nt of maxi mal fl exure in the
curvat ure and, therefore, the stress on the i nst rument duri ng this
study was in a const ant fi xed location until separation occurred.
Cl i ni cal l y, axial mot i on (in and out movement s) duri ng i nst rumen-
8 4 Pr u e t t et al.
tation with Ni-Ti, engi ne-dri ven rotary i nst rument s woul d be ex-
pected, as concluded by Dederich et at. (12) exami ni ng stainless-
steel instruments, to si gni fi cant l y extend i nst rument life. Ni-Ti
i nst rument manufacturers suggest that these i nst rument s be used
with cont i nuous axial mot i on in the canal. Li ngeri ng at a single
depth i n the canal woul d closely duplicate the test condi t i ons of
this study, significantly reduci ng the fatigue life of an i nst rument
at a specific point along the i nst rument shaft. Thi s mode of oper-
ation is discouraged and should be avoided.
Second, engi neeri ng applications of Ni -Ti alloys have been
ext ensi vel y investigated ( 18- 20) , and the fatigue behavi or of Ni -Ti
alloys has been shown to be related to the static yield stress of the
alloy. At low static stress, the alloy exhibits l i near strain. Once the
stress rises above the yi el d stress, small increases i n the stress will
gi ve rise to large increases in strain that, unl i ke most metals, are
reversible. Thi s is the so-called superelastic characteristic. Ni-Ti
fatigue tests conducted bel ow the yield stress give results similar to
ferrous alloys, such as stainless steel. Fat i gue tests conducted
above the yield stress i n the superelastic range gives rise to distinct
low-cycle fatigue behavi or (18), which results in very rapid reduc-
tions i n fatigue life with i ncreasi ng stress. Thi s l ow-cycl e fatigue
has been shown to be more closely related to the amount of strain
or flexure of the speci men t han to the stress level and is largely a
crack propagation process. This study was conducted i n the range
of superelastic rather t han elastic state of deformat i on of the Ni-Ti
instruments. Clinically, the fatigue life of an i nst rument can be
related to the degree that it is flexed when pl aced in a curved canal,
with greater flexures havi ng a shorter expected life. The pilot
studies used to determine the test parameters suggested that the
state of superelastic deformation, when stresses rose above the
yield stress, occurred bet ween angles of 25 to 30 degrees when a
size #30 i nst rument passed through a 5-ram radius curvature. At 30
degrees, low-cycle fatigue behavi or resulted. At 25 degrees and
below, the fatigue life was indefinitely long, suggesting elastic
deformat i on of the alloy. Bei ng tested in the superelastic state also
helps expl ai n the si gni fi cant differences in the test groups when
stresses were increased by decreasing the radius of curvature or
increasing the angle of curvature; both resulted in increased strain.
Most cl i ni cal applications of Ni -Ti i nst rument s will at least peri-
odically i nvol ve the superelastic flexure of these instruments, thus
the need to understand this behavior. It is difficult to estimate the
stress level in an i nst rument duri ng use, but the degree to which it
is strained or flexed can be monitored through knowl edge of the
canal curvature determine~t radiographically. Flexed conditions
resulting in characteristics of relatively low stain and elastic de-
format i on woul d be expected to exhibit increased number of cycles
to failure several orders of magni t ude hi gher than those determined
in this study (18).
Al l instruments tested separated at the poi nt of maxi mum flex-
ure wi t hi n the curved segment of the tube, wherei n the stress on the
i nst rument was greatest (Fig. 8). The fractured surfaces seemed
si mi l ar to those previously observed for fatigue testing (1 i ). Un-
flexed, partially fractured Ni-Ti i nst rument s seemed normal with-
out per manent plastic deformation. Under flexure, openi ng of the
fractures was evident and generally seemed to be associated with
flaws on the i nst rument surface directly i n the area of maxi mum
flexure. No instrument speci men separated at or j ust behi nd the
head of the i nst rument i n a successful test to its cyclic fatigue limit.
Thi s contradicts reports suggesting that the Canal Master instru-
ment desi gn is prone to separation j ust behi nd the head or in the
apical third. Static torsional studies have shown that separation of
J o u r n a l of E n d o d o n t i c s
Ni-Ti Lightspeed i nst rument s always occurred --2 mm behi nd the
i nst rument head (7). That failure mode is limited to static testing.
I f i nst rument separation occurs behi nd the head of the i nst rument
clinically, two expl anat i ons are possible. First, separation occurred
as the result of l ocki ng or torsional overloading of the i nst rument
head duri ng rotation. Second, stress in a curved canal can cause
fatigue failure directly behi nd the head. The head is i nfl exi bl e and
must pass t hrough any curvature that the shaft experiences. Ex-
cessive apically directed force on the i nst rument wilt cause the
shaft to flex j ust behi nd the head. Both circumstances lead to stress
concent rat i on causi ng fatigue damage, accelerating separation
caused by bi ndi ng and fatigue. Separation of unbound i nst rument s
in the area of the most severe canal curvature should always be
considered a result of cyclic fatigue with any i nst rument system.
Furt her study is requi red to improve the underst andi ng of cri-
teria that may be used to predict Ni-Ti i nst rument fatigue l eadi ng
to separation. The fatigue of an i nst rument can be broken down
into stages as evi dent on SEM examination (Fig. 9). Stage I, crack
initiation, is characterized by the smooth, al most featureless area at
the surface periphery. This area of fracture initiation maybe re-
versible through proper heat treatment (9). Sterilization has been
suggested to enhance the fatigue life of Ni -Ti files and reverse the
stress-induced martensite transformation to the parent austenite
phase (2). Furthermore, sterilization may reverse the crack initia-
tion stage of fracture in Ni -Ti alloys, but this is an area for future
investigation. Torsi onal l oadi ng during rotational use is another
variable to consi der that may dramatically i nfl uence the cyclic
fatigue life of instruments. Engaged i nst rument s that are actively
cutting dent i n woul d be expected to have different fatigue charac-
teristics. The cyclic fatigue phenomenon may be further com-
pounded by the addition of flutes to the shaft and by i nst rument
taper. Flutes coul d act as stress concentrators on the i nst rument
shaft, pot ent i al l y resulting i n more rapid crack initiation. Inst ru-
ment taper enl arges the di amet er of the i nst r ument i n the area of a
coronal curvature, compared with the apical area, thus resul t i ng i n
higher stress caused by a larger cross-sectional diameter. The
effect of i nst rument flutes and cross-sectional diameters is an area
of concern, particularly with the introduction of nonst andardi zed
tapered Ni -Ti instruments. These areas should be addressed in the
design of Ni -Ti , engi ne-dri ven rotary i nst rument s and warrant
further study.
Specific recommendat i ons for prevent i ng breakage of Light-
speed i nst rument s cannot be drawn from this study. However,
because partially fatigued i nst rument s did not demonst rat e vi si bl e
deformation, yet did exhibit crack formation, a general concl usi on
can be made. To prevent the breakage of Ni -Ti , engi ne-dri ven
i nst rument s secondary to cyclic fatigue, it seems that the best
recommendat i on is to discard them after a specified use period as
recommended for stainless-steel instruments. That use period
would be expected to vary widely dependi ng on i nst rument design,
size, and stresses pl aced on it because of the manner of use, and the
canal curvature prepared. In some severe angles havi ng a small
radius of curvature and a large angle of curvature, discarding an
i nst rument after a single use may be the safest advice.
Canal curvature and angl e have been implicated in i nst rument
separation. Cycl i c met al fatigue has been further implicated, as an
important factor in i nst rument breakage. Thi s study clearly showed
that, i n cyclic fatigue testing, cycles to failure is inversely propor-
tional to i nst r ument stress. Smaller radius of curvature and i n-
creased i nst rument diameter resulted in fewer cycles to failure
caused by hi gher i nst rument stress. Cycles to failure was not
affected by operating rpm, and the effect of angl e al one on cycles
Vol. 23, No. 2, February 1997
to fai l ure was unclear. New Ni -Ti instrument syst ems are currently
being r ecommended for use in a cont i nuousl y rotary manner, and
a growi ng number of incidents of i nst rument separation are bei ng
i nformal l y discussed. Gi ven the i ncreased number of Ni -Ti end-
odont i c instrument systems on the market, these probl ems are
l i kel y to become more common. To date, there is no test prot ocol
that sets mi ni mum standards for cycl i c fatigue of these instru-
ments, nor are the operat i ng paramet ers of i mport ance yet fully
understood. Thi s study concl udes wi t h a call f or a compr ehensi ve
cycl i c fatigue test speci fi cat i on for all rotary endodont i c instru-
ments, i ncl udi ng Ni - Ti instruments desi gned to be engi ne-dri ven
and fully prepare curved canals. Furthermore, the parameters of
angl e of curvat ure and radius of curvature, and instrument cross-
sectional core di amet er must be i ncl uded in any such specification.
Wi t hout i nvest i gat i on of these parameters, recommendat i ons for
the safe use of Ni -Ti , engi ne-dri ven rotary instruments will remai n
speculative. The results also suggest that the effect of the radius of
curvat ure as an i ndependent variable should be consi dered in
studies eval uat i ng root canal instrumentation.
This study was supported in part by an Endodontic Student Award from
the Research and Education Foundation of the American Associ ati on of
Endodontists.
The opinions, assertions, materials, and methodologies herein are private
ones of t he authors and are not t o be construed as official or reflecting the
views of the American Association of Endodontists or t he Foundation.
We gratefully acknowl edge David S. Pittman of Sunset Resources, Inc.
(San Antonio, TX) f or modifying and writing the data collection program for the
Magtrol Dynamometer. We express our gratitude t o Drs. James A. Gilles,
William A. Walker Ill, and Carlos E. del Rio for their advice during article
preparation.
Drs. Pruett, Clement, and Carnes are affiliated wi th the Department of
Endodontics/Dental School, University of Texas Health Science Center at San
Antonio, San Antonio, TX. Address requests for reprints t o Dr. David L.
Carnes, Jr., Department of Endodontics/Dental School, University of Texas
Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, i X
78284-7898.
Cycl i c Fati gue of Ni -Ti Rotary I nst rument s 85
References
1. Walia H, Brantley WA, Gerstein H. An initial investigation of t he bending
and torsional properties of nitinol root canal files. J Endodon 1988;14:346-51.
2. Serene TP, Adams JD, Saxena A. Nickel-titanium instruments: appli-
cations in endodontics, St. Louis: Ishiyaku EuroAmerica, Inc., 1994.
3. GIosson CR, Hailer RH, Dove SB, del Rio CE. A comparison of root
canal preparations using Ni-Ti hand, Ni-Ti engine driven, and K-Flex end-
odonti c instruments. J Endodon 1995;21:146-51.
4. Cohen S, Burns RC. Pathways of t he pulp. 6th ed. St. Louis: Mosby-
Year Book, Inc., 1994:206.
5. Council on Dental Materials and Devices. New ADA Specification No.
28 f or endodonti c files and reamers. J Am Dent Assoc 1976;93:813-8.
6. Camps J, Pertot WJ. Torsional and stiffness properties of Canal Master
U stainless steel and nitinol instruments. J Endodon 1994;20:395-8.
7. Marsicovetere ES, Burgess JO, Clement D J, del Rio CE. Torsional
testi ng of the lightspeed NiTi instrument system. J Endodon (in press).
8. Sot okawa T. An analysis of clinical breakage of root canal instruments.
J Endodon 1988;14:75-82.
9. Dieter GE. Mechanical metallurgy. 3rd ed. New York: McGraw-Hill,
1986:119, 138, 185-8, 382-7, 394.
10. Brantley WA, Luebke H, Luebke FL, Mitchell JC. Performance of
engine-driven rotary endodonti c instruments with a superimposed bending
deflection. V. Gates-Glidden and peeso drills. J Endodon 1994;20:241-5.
11. Luebke NH, Brantley WA, Sabri ZI, Luebke FL, Lausten LL. Physical
dimensions, torsional performance, bending properties, and metallurgical
characteristics of rotary endodonti c instruments. VI. Canal Master drills. J
Endodon 1995;21:259-63.
12. Dederich DN, Zakariasen KL. The effects of cyclical axial motion on
rotary endodonti c instrument fatigue. Oral Surg 1986;61:192- 6.
13. Haikel Y, Gasser P, Allemann C. Dynamic fracture of hybrid endodon-
tic hand instruments compared with traditional files. J Endodon 1991 ;17:217-20.
14. Crandall SH, Dahl NC, Lardner TJ. An introduction t o the mechanics of
solids. 2rid ed. New York: McGraw-Hill, 1972:416-77.
15. Schneider SW. A compari son of canal preparations in straight and
curved canals. Oral Surg 1971;32:271-5.
16. Cunningham CJ, Senia ES. A three-dimensional study of canal curvatures
in the mesial roots of mandibular molars. J Endodon 1992;18:294-300.
17. Kanninen MF, Popelar CH. Advanced fracture mechanics. New York:
Oxford University Press, 1985:22, 1980.
18. Melton KN, Mercier O. Fatigue of NiTi thermoelastic marfensites. Act a
Met 1979;27:137-44.
19. Miyazaki S, Sugaya Y, Otsuka K. Effects of various factors on fatigue
life of Ti-Ni alloys. MRS Int Mtg Aclv Mater 1989;9:251-6.
20. Tobushi H, Iwanaga H, Tanaka K, Hod T, Sawada T. Stress-strain-
temperature relationships of TiNi shape memory alloy suitable for thermome-
chanical cycling. JSME Int J Series I 1992;35:271-7.
You Might Be Interested
A case of fraud was recently uncovered in whi ch an imposter had, over the past 25 years, managed to get
a number of essays and letters to the editor published in respected journals. The JAMA unmasked the rascal
because his "writing style had been recognized" (BMJ 311" 523, 1995).
Of course, there was another clue. The fi cti ci ous name signed t o all those published letters over all those
years was--you' l l love t hi s--T. H. Fiddleman.
Zachariah Yeomans