Full text of DOI:10.1093/ejo/22.3.327

European Journal of Orthodontics 22 (2000) 327–334
2000 European Orthodontic Society
The force levels required to mechanically debond
ceramic brackets: an in vitro comparative study
Selim Arici and Chris Minors
Department of Child Dental Health, School of Clinical Dentistry, University of Sheffield, UK
SUMMARY The in vitro force levels generated by four differing methods of mechanical
debonding techniques for ceramic brackets, using debonding pliers, were measured. The
forces generated using wide (method W) and narrow blades (method N) were compared
with those generated using a diagonally opposite corner application of the wide blades
(method C) and incisal-gingival application of a pair of pointed blades (method P).
Chemically retained ceramic brackets (Transcend) were bonded to bovine teeth using a
filled, two-paste, chemically cured composite (Concise). After 24 hours storage at 37°C in
water, each specimen was subjected to one of the four mechanical debonding methods
in a custom-built jig, simulating the clinical application of conventional debonding pliers.
A one-way ANOVA with a Tukey’s honestly significant difference test revealed statistically
significant differences in debonding strengths between the four methods at the 0.05 level
of significance. The mean debonding strength generated by method C was 40 and 25 per
cent lower than that for methods W and N, respectively. Scoring of the adhesive remnant
index (ARI) revealed that the predominant bond failure site was at the bracket/adhesive
interface for all groups. Macroscopically, no enamel damage or bracket fractures were
observed.
Introduction
bracket base design (chemical, mechanical,
or mechanical/chemical bonding);
The ceramic orthodontic bracket, when introduced,
was considered a viable alternative to stainless
steel, particularly since it offered a significant
improvement in aesthetics. Unfortunately, unlike
stainless steel, the material cannot be flexed
slightly to aid debonding and the high initial
forces that are necessarily applied in the removal
of ceramic brackets at the end of orthodontic
treatment have become an area of concern.
Several complications have been encount-
ered during debonding of these brackets, such as
enamel tear-outs, fractures, cracks, and bracket
failure (Machen, 1990; Redd and Shivapuja,
1991; Gibbs, 1992). The reported complications
have been attributed to:
3. The method employed for debonding
(Bishara and Trulove, 1990a; Chaconas et al.,
1991; Winchester, 1992).
The various manufacturers have therefore
been continuously modifying the base design of
ceramic brackets and introducing new debonding
techniques specifically for their particular brand
of bracket. Although new techniques, including
electrothermal, ultrasonic, and laser debonding,
have been advocated, mechanical debonding of
ceramic brackets, either with sharp-edged pliers
or custom-built lifting tools engaging under the
bracket wings, remains the technique of choice
(Swartz, 1988; Bishara and Fehr, 1993).
Several articles in the orthodontic literature
have reported on the behaviour of dental enamel
during the in vitro use of sharp-edged debonding
pliers to debond ceramic brackets (Bishara and
Trulove, 1990b; Storm, 1990; Redd and
Shivapuja, 1991). There have, however, been
1. The mechanical properties (brittleness and
low fracture toughness) of the ceramic
materials;
2. The high bond strength generated due to
bonding characteristics of the ceramic

328
S. ARICI AND C. MINORS
relatively few studies reporting the actual force
levels generated during this procedure (Bishara
and Fehr, 1993; Bishara et al., 1994, 1995).
Bishara and Fehr (1993), in an in vitro study,
showed that the use of debonding pliers with
narrow blades effectively debonded ceramic
the bracket bases were used in the study. The
brackets were bonded to bovine enamel using
a
chemically-cured, two-paste, highly-filled
(75 per cent quartz), composite resin (Concise,
3M Dental Products, St Paul, Minnesota, USA).
Extracted primary bovine mandibular incisor
teeth were obtained and stored at room tem-
perature in 70 per cent ethyl alcohol. The teeth
were taken from Yerli-Kara cattle, approximately
14–18 months old, raised and slaughtered in the
same farm complex. The labial enamel surfaces
of these teeth were visually inspected, using a
magnifying glass, to exclude those with defects
and/or caries in the labial enamel. Eighty teeth
selected in this manner were used for this
project.
brackets with
a significantly lower mean
debonding force than pliers with wider blades.
They also stated that the relatively smaller
contact area of the narrow blades (2 mm) was
sufficient to initiate and propagate a crack in the
adhesive. This was claimed to reduce the trauma
of debonding due to the reduced stress on the
enamel surface.
A logical development of this is to reduce the
contact area to the minimum practicable, which
is, in effect, a point, and to measure the force
levels created by the use of a pair of pointed
debonding blades on the adhesive in the enamel/
bracket base interface.
The second aspect of debonding considered
was the site of application of the force.
Conventionally, debonding pliers are applied
to the adhesive layer on opposing faces of the
bracket (i.e. mesially and distally, or sometimes
incisally and gingivally). It is postulated, however,
that applying the force across the diagonally
opposite corners of a bracket with conventional
debonding pliers blades might be an alternative
method of reducing the pliers/adhesive contact
area.
The purpose of this study therefore was to
determine the in vitro force levels required to
achieve debonding by these different methods of
achieving minimal contact between the debond-
ing blades and the adhesive resin, and to compare
the results with those generated by wide and
narrow blades in the mechanical debonding
of standardized chemically bonded ceramic
brackets. The amount of adhesive remaining on
the tooth surface and visible enamel damage
were also investigated.
Substrate preparation and bonding
The selected teeth were randomly assigned to
one of the four test groups. Before bonding, the
labial surfaces of the crowns were polished using
a pumice and water slurry in a rubber cup for
10 seconds. They were then rinsed with water
for 15 seconds and blown dry with oil-free
compressed air. A 37 per cent solution of liquid
phosphoric acid (Concise etching agent) was
applied to the labial surface for 60 seconds.
Finally, the teeth were washed with water for
30 seconds to remove the orthophosphoric acid
and dried with compressed air. The labial
surfaces of the teeth appeared chalky white in
colour, as is normal after etching. The brackets
were then bonded to the teeth at room tem-
perature in accordance with the manufacturer’s
suggested procedure.
After bonding, excess adhesive resin around
the bracket base was removed with a dental
scaler. The bracketed teeth were left undisturbed
to air dry for 10 minutes until the adhesive was
sufficiently set. They were then stored in distilled
water at 37°C for 24 hours prior to testing.
Materials and methods
Test apparatus
Polycrystalline aluminium oxide (Al
2O3) ceramic
In this study, debonding of ceramic brackets
were carried out using debonding pliers fitted
with one of three differing types of paired blades.
The first pair of blades used were 3.2 mm wide
brackets for upper central incisors (Transcend,
Unitek Corp., Monrovia, California, USA)
with silane chemical coatings for retention on

MECHANICAL DEBONDING OF CERAMIC BRACKETS
329
stainless steel blades produced as replaceable
substitutes for ETM 345–6 RT direct bond
remover pliers (ETM Corporation, Monrovia,
California, USA). The second pair had narrow
blades (2 mm). The third pair were also the wide
type, but their tips were modified in order to give
a pointed edge.
A special compression jig (Figures 1 and 2)
was constructed and attached to the jaws of a
Lloyd M 5K testing machine (Lloyd Instruments
Plc., Fareham, Hampshire, UK), in order to
simulate the movement of the debonding plier
blades, and to enable accurate and direct measure-
ment of the forces created during mechanical
debonding. For each test, the selected pair of
plier blades were secured at the centre of each
of the opposing steel cylinders by means of set-
screws. The steel cylinders were removable from
the compression jig to enable easy replacement
of the blades. The compression jig provided a
controllable and mechanically sound oppositional
movement in the vertical axis of the steel
cylinders carrying the blades.
Figure 1 Diagrammatic illustration (side view) of the test
apparatus.
Methods of debonding force application
The four methods of debonding force application
compared in the study were:
1. Wide blades (method W) in the inciso-
gingival plane (Figure 3a).
2. Narrow blades (method N) in the inciso-
gingival plane (Figure 3b).
3. Pointed blades (method P) in the inciso-
gingival plane (Figure 3c).
4. Wide blades (method C) applied across the
mesio-incisal and disto-gingival (diagonally
opposite) corners (Figure 3d).
For the tests of each method, the specimens were
positioned freely between the two blades by
the same operator, until the blades touched the
adhesive layer from both sides. Twenty samples
were debonded by each method using a new pair
of blades.
Figure 2 Test apparatus (frontal view).
During testing, the increasing force levels were
monitored on the digital display on the machine.
When the bond failed, the force level was auto-
matically recorded and presented in Newtons
which was later converted into MPa by dividing
by the bracket base area. The crosshead speed of
the machine was 5 mm per minute.

330
S. ARICI AND C. MINORS
Figure 3 Methods of debonding force application (a = method W, b = method N, c = method P, d = method C).
The bonding surface area of the bracket base
was measured to the nearest 0.01 mm with a
reflex microscope connected to a computerized
video image analysis system. The bases of 10
brackets were measured and the mean nominal
surface area was calculated as 11.32 mm2.
After testing, the separated assemblies were
recovered and examined under a light micro-
scope at ×20 magnification in order to classify
the enamel surfaces according to the adhesive
remnant index of Årtun and Bergland (1984).
by this procedure were further investigated using
Tukey’s honestly significant difference (HSD)
test with a 95 per cent confidence interval.
Results
The descriptive statistics of the debonding
strengths and the grouping of the mean values
according to Tukey’s HSD test for the methods
used are given in Table 1. While the wide blade
pairs applied from the diagonally opposite corners
(method C) generated the lowest debonding
strength, the wide debonding blades (method W)
showed the highest mean debonding strength,
followed by the narrow blades (method N) and
pointed blades (method P), respectively.
Statistical analysis
The differences in the debonding strengths
were investigated statistically using an analysis of
variance (ANOVA). Any differences revealed

MECHANICAL DEBONDING OF CERAMIC BRACKETS
331
Table 1 Descriptive statistics for each method and grouping of the mean debonding strengths according to
Tukey’s HSD test.
Debonding
method
Size (n)
Mean
(MPa)
SD
Min.
value
Max.
value
Tukey’s HSD*
grouping
W
N
P
20
20
20
20
14.51
10.77
10.05
8.37
3.57
2.90
2.77
2.72
9.06
5.99
7.00
5.17
20.63
16.78
14.45
16.49
A
B
B
B
C
*Grouping with different letters are significantly different at the 0.05 significance level.
Table 2 Frequency and percentage occurrence of the adhesive remnant index (ARI) for each method tested.
Method
ARI = 0
ARI = 1
ARI = 2
ARI = 3
W
N
P
3 (15%)
1 (5%)
0 (0%)
1 (5%)
4 (20%)
7 (35%)
3 (15%)
3 (15%)
9 (45%)
7 (35%)
11 (55%)
8 (40%)
4 (20%)
5 (25%)
6 (30%)
8 (40%)
C
Adhesive remnant index (ARI) scores: 0 = no adhesive left on the tooth. 1 = less than half of the adhesive left on the
tooth. 2 = more than half of the adhesive left on the tooth. 3 = all adhesive left on the tooth.
The one-way analysis of variance, which was
used to test the hypothesis that there was
no significant difference between the debonding
strengths of the four methods, showed a highly
significant difference (P < 0.001) between the
methods. A further examination of the results
by Tukey’s HSD test indicated that the mean
debonding strength for method W (14.51 MPa)
was significantly higher than those for method N
(10.77 MPa), method P (10.05 MPa), and method
C (8.37 MPa) at the 5 per cent significance level
(Table 1).
The adhesive remnant index (ARI) scores
2 and 3 applied to the majority of the teeth
indicate that the predominant failure site was the
bracket/adhesive interface for all the methods
(Table 2). In method C, 40 per cent of the
specimens had an ARI score of 3 (all adhesive
resin remained on the enamel surface). None of
the teeth debonded with the pointed blades
(method P) exhibited an ARI score of 0.
Discussion
Sources of error
Although technique inconsistencies are minimized
by the use of standardized bonding and repro-
ducible direct methods of testing, subtle differences
in enamel prism micro-morphology, thickness of
adhesive layer, and porosities within the adhesive
could cause variation in the debonding strengths
for each group (Regan and van Noort, 1989;
Lew et al., 1991). There are two further sources
of error, which might have had an effect on the
results.
Since the thickness of the adhesive layer is
small, the tips of the blades could not be
accurately placed on it when the force was
applied. However, even though the tips of the
blades may have deviated towards either the
joint between the adhesive and the bracket
base, or that between the adhesive and the
enamel, these would not significantly affect the
results in a clinical application, as long as the tips
are kept away from the ceramic bracket itself.
None of the specimens showed any macro-
scopic evidence of bracket fracture or visible
enamel damage.

332
Blunting during use of the blades, particularly
the pointed ones, would have an increasing effect
on the force level applied to later specimens.
However, because the number of the specimens
tested (20) with each type of blade was not
excessive (Bishara and Fehr, 1993), this was
ignored.
S. ARICI AND C. MINORS
finding is supported by Storm (1990) who also
found, in most cases, that failure occurred at
the bracket/adhesive interface when debonding
Transcend ceramic brackets with a hard wire
cutter.
Although previous authors (Gwinnett, 1988;
Ødegaard and Segner, 1988) found bond
failure for Transcend ceramic brackets to be
more prevalent at the enamel/adhesive inter-
face, others (Baron et al., 1990; Forsberg and
Hagberg, 1992; Sam et al., 1993) reported that
the predominant failure site for the same bracket
was at the bracket/adhesive interface. However
it is important to note that the type of force
(shear) applied to the brackets in these studies is
not normally used for the debonding of ceramic
brackets and differs from those used in this
investigation.
The results of no visible enamel damage
observed in this study is in line with the findings
of Storm (1990), Redd and Shivapuja (1991),
and Bishara et al. (1995) who also found no
visual enamel damage after debonding different
types of ceramic brackets. However, under light
microscopy two specimens, one from method
P, and one from method N, were found to exhibit
slight surface damage in the form of surface
roughening.
Debonding strengths
No attempt was made to compare the debonding
strengths of the groups of chemically bonded
Transcend ceramic brackets by in vitro tensile
and shear bond strengths as reported by previous
studies because, as Bishara and associates (1995)
emphasized, the forces generated by debonding
pliers applied to both sides of the bracket are not
directly comparable to the shear and/or tensile
forces.
The mean applied debonding strength for the
narrow blade pairs (method N) was approximately
25 per cent lower than that of the wide blade
pairs (method W). This essentially corroborates
the findings of Bishara and Fehr (1993). It should,
however, be noted that they used a different
adhesive and a different type of ceramic bracket
(which had a combination of mechanical and
chemical retention on the base) to those in this
study.
Method P caused a 30 per cent lower mean
debonding strength than method W. The mean
debonding strength for method C was about 40
and 25 per cent lower than those of method W
and method N, respectively. This means that
when a pair of debonding pliers, fitted with
wide (or even narrow) blades, is applied to the
diagonally opposite corners of chemically bonded
ceramic brackets, the stresses on the enamel
surface may be expected to be lower than those
occurring when both wide and narrow flat blades
are applied, during debonding, in an inciso-
gingival plane.
Because of the increased ARI score of 3, it is
reasonable to suggest that the diagonally opposite
corner application of the debonding blades
(method C) might reduce the probability of bond
failure at the enamel/adhesive interface, further
minimizing the risk of enamel damage.
Clinical implications
Extrapolation of laboratory data to the clinical
situation should always be undertaken with care
because of the complex oral environment. The
changes in temperature, humidity, and acidity
(pH), and the mechanical and masticatory stresses
placed on a bracket in the oral cavity have a
deteriorating effect on the adhesive bond, and
are impossible to simulate in a laboratory (Øilo,
1993). Furthermore, moisture control in vitro is
much more accurate than that pertaining during
orthodontic bonding in vivo. For the above
reasons it is conceivable that clinical debonding
Failure sites
The results from this study show that the
predominant bond failure site for the chemically
bonded ceramic brackets under test was at the
bracket/adhesive interface in all groups. This

MECHANICAL DEBONDING OF CERAMIC BRACKETS
333
strength values would be lower than those
reported in this study. Nevertheless, it may
reasonably be suggested that comparison
between standardized in vitro studies may be
extrapolated to predict clinical results and that
laboratory testing can be used as a screening
mechanism for predicting clinical performance.
It has been stated that bond strengths larger
than 138 kg/cm2 (13.53 MPa) during removal of
brackets with conventional debonding pliers
should be avoided (Bishara et al., 1994). This
might be interpreted as implying that applying
significantly lower forces would be a safer and
more satisfactory method of debonding. This
effect may be achieved by conventional debonding
pliers placed on the diagonally opposite corners
of ceramic brackets since this rarely produces a
force higher than 13 MPa. It is also worthy of
note that, when used for the removal of canine
and premolar ceramic brackets (which have rela-
tively more curved bases than incisor brackets),
this application method would eliminate the
blades inevitably resting in part on the brittle
ceramic bracket (Figure 4).
Figure 4 Diagrammatic illustration of the impingement of
the wide blades on curvature of bracket base (Thick arrows
indicate the possible impingement points).
instead of relying on hand pressure on a pair of
levers, as with conventional debonding pliers, the
force should be applied by means of a screw (as
in a hand vice). This would mean the degree of
‘excessive force application’, which is unpleasant
for the patient, may be significantly reduced.
During debonding, the continued presence
of the archwire is a useful safety factor in
controlling the bracket subsequent to release.
Applying the pliers across the bracket corner
obviates the need to remove the archwire before
debonding, as is normally necessary if pliers are
used in a mesio-distal application.
Conclusions
The results tend to confirm the hypothesis that
for debonding of ceramic brackets, because of
their rigid nature, a tenable method of initiating
the procedure is to produce a crack in the adhesive
at the bracket/adhesive/enamel interface and
then to extend this crack until the bracket is
free. The use of conventional debonding pliers
may in theory produce two cracks, one on either
side of the bracket, but in practice it may well
be that the first is propagated before the second
has time to form. The results support the
suggestion that the force required to initiate
debonding is directly related to the contact area
between the pliers tips and adhesive, and that
reducing the contact area to a minimum reduces
the force necessary to initiate debonding.
The forces required to initiate debonding of
ceramic brackets are related to the contact area
between the tips of the pliers and the adhesive.
This can be minimized by either using pointed
plier tips or placing a conventional pair of
debonding pliers diagonally opposite the corners
of the bracket. It is suggested that the latter
technique may also offer a safety bonus in that
the archwire may be left attached to the brackets
whilst they are removed.
Address for correspondence
Dr Selim Arici
Ondokuz Mayıs Üniversitesi,
Dis¸hekimlig˘i Fakültesi,
55139 Kurupelit/Samsun,
Turkey
It is additionally suggested that, in order to
make the applied force more easily controllable,

334
S. ARICI AND C. MINORS
Gibbs S L 1992 Clinical performance of ceramic brackets:
a survey of British orthodontists’ experience. British
Journal of Orthodontics 19: 191–197
References
Årtun J, Bergland S 1984 Clinical trials with crystal growth
conditioning as an alternative to acid-etch enamel
pretreatment. American Journal of Orthodontics 85:
333–340
Gwinnett A J 1988 A comparison of shear bond strengths
of metal and ceramic brackets. American Journal of
Orthodontics and Dentofacial Orthopedics 93: 346–348
Baron D I, Pontier J P, Guzman R C, Rudert M J 1990
Shear stress of ceramic brackets. Journal of Dental
Research 69: 313 (Abstract)
Lew K K K, Chew C L, Lee K W 1991 A comparison of
shear bond strengths between new and recycled
ceramic brackets. European Journal of Orthodontics 13:
306–310
Bishara S E, Fehr D E 1993 Comparisons of the effective-
ness of pliers with narrow and wide blades in debonding
ceramic brackets. American Journal of Orthodontics and
Dentofacial Orthopedics 103: 253–257
Machen D E 1990 Legal aspects of orthodontic practice:
risk management concepts. Ceramic bracket update.
American Journal of Orthodontics and Dentofacial
Orthopedics 98: 185–186
Bishara S E, Trulove T S 1990a Comparison of different
debonding techniques for ceramic brackets: an in vitro
study. Part I. Background and methods. American
Journal of Orthodontics and Dentofacial Orthopedics 98:
145–153
Ødegaard J, Segner D 1988 Shear bond strength of metal
brackets compared with a new ceramic bracket. American
Journal of Orthodontics and Dentofacial Orthopedics 94:
201–206
Bishara S E, Trulove T S 1990b Comparison of different
debonding techniques for ceramic brackets: an in vitro
study. Part II. Findings and clinical implications. American
Journal of Orthodontics and Dentofacial Orthopedics 98:
263–273
Øilo G 1993 Bond strength testing—what does it mean?
International Dental Journal 43: 492–498
Redd T B, Shivapuja P K 1991 Debonding ceramic brackets:
effects on enamel. Journal of Clinical Orthodontics 25:
475–481
Bishara S E, Fonseca J M, Fehr D E, Boyer D B 1994
Debonding forces applied to ceramic brackets simulating
clinical conditions. Angle Orthodontist 64: 277–282
Regan D, van Noort R 1989 Bond strengths of two integral
bracket-base combinations: an in-vitro comparison with
foil mesh. European Journal of Orthodontics 11: 144–153
Bishara S E, Fonseca J M, Boyer D B 1995 The use of
debonding pliers in the removal of ceramic brackets:
force levels and enamel cracks. American Journal of
Orthodontics and Dentofacial Orthopedics 108: 242–248
Sam W H, Chao S Y, Chung K H 1993 Shear bond strength
of ceramic brackets with two bonding adhesives. British
Journal of Orthodontics 20: 225–229
Storm E R 1990 Debonding ceramic brackets. Journal of
Clinical Orthodontics 24: 91–94
Chaconas S J, Caputo A A, Niu G S L 1991 Bond strength
of ceramic brackets with various bonding systems. Angle
Orthodontist 61: 35–41
Swartz M L 1988 Ceramic brackets. Journal of Clinical
Orthodontics 22: 82–88
Forsberg C M, Hagberg C 1992 Shear bond strength of
ceramic brackets with chemical or mechanical retention.
British Journal of Orthodontics 19: 183–189
Winchester L 1992 Methods of debonding ceramic brackets.
British Journal of Orthodontics 19: 233–237