Surface preparation for orthodontic bonding

to porcelain
Yngvil ~rstavik Zachrisson, DDS," Bj6rn U. Zachrisson, DDS, MSD, PhD,b and
Tamer B~y(ikyilmaz, DDS"

Oslo, Norway
This study evaluated the effect of various porcelain surface treatments on the tensile strength of
orthodontic brackets bonded to a feldspathic metal ceramic porcelain. The porcelain was fused to
flat gold alloy tabs and divided into six groups that were subjected to sandblasting, silane
application, intermediate resin, or etchants (9.6% hydrofluoric acid or 4% APF gels). Two brackets
were bonded onto each porcelain/metal tab (n = 60) with Bis-GMA resin (Concise, 3M Corp., St.
Paul, Minn.) or 4-META resin (MCP-bond, Sun Medical Co. Ltd., Tokyo, Japan). The samples were
stored in 37 C water, thermocycled 1000 times from 5C to 55C, and tested in tension. Alignment
and uniform loading during testing were secured by engaging a hook in a circular ring soldered
onto the bracket slot before bonding. Similar control brackets (n = 12) were bonded with Concise
to extracted caries-free mandibular incisors. Bond failure sites were classified according to a
modified Adhesive Remnant Index (ARI) system. Silane application to the sandblasted porcelain
surface significantly increased the bond strengths according to analysis of variance and Duncan's
multiple range test. The quality of the bonds was further enhanced by the addition of the
intermediate resin. Etching the porcelain with 9.6% hydrofluoric acid provided similar bond
strengths, but the 4% APF gel was less effective. The MCP-bond was not significantly better than
Concise in bond strength to sandblasted porcelain. Several difficulties associated with the clinical
interpretation of laboratory data on bonding to dental porcelains are discussed, and clinical trials
are necessary for final evidence of efficacy. (AMJ ORTHOD DENTOFACORTHOP 1996;109:420-30.)

C e r a m i c and metal-ceramic restorations

are now widely used for restoring damaged or
missing teeth to enhance the esthetics of the natural dentition. Particularly in adults, there is an
increased likelihood that orthodontic brackets and
retainer wires have to be fitted to patients who
have porcelain surfaces on some crowns or veneers.
Many ceramics are marketed for use as dental
crown and bridge materials, a4 The types range
from traditional hand-condensed jacket crown porcelains with or without alumina reinforcement to
porcelains for metal-bonding, pressure-formed ceramics, castable glasses and new all-ceramic single
crowns and anterior 2- or 3-unit fixed prostheses.
Because classifications may vary between reports,
the characterization of dental ceramics as feldspathic porcelains, aluminous porcelains, and glass
ceramics (Table I) would appear useful for practical orthodontic purposes. The conventional feldspathic porcelains are made of the mineral feldspar
with some additions for color and translucency and
From the Department of Orthodontics, University of Oslo.
aGraduate student.
bprofessor n.
Copyright 1996 by the American Association of Orthodontists.
0889-5406/96/$5,00 + 0 8/1/62497

420

contain silica (SiO2) and alumina (A1203), with

small amounts of KzO and Na20 to control expansion. 2 The aluminous porcelains are mainly core
materials, and not applicable for orthodontic bonding. The glass ceramics appear by a secondary
crystallization process of a glass matrix.4
There are only a few reports on bonding to
porcelain surfaces for orthodontic purposes. 516
Most studies have used silane as a coupling agent
to increase the bond strength to either glazed or
roughened dental porcelains. The silane reacts with
the silica within the ceramic and the organic groups
of the bonding resin. 4 However, some of these
studies have used denture teeth, 6-9 which are
burned at considerably higher temperatures and
have properties that differ significantly from the
feldspathic porcelains commonly used for porcelain
crowns, bridges, and v e n e e r s , m'12'17 Furthermore,
most earlier studies have not included water storage and a thorough thermocycling regimen before
the testing. 5-9'~8Thermal cycling simulates the temperature fluctuations in the mouth by subjecting
the bonds to alternating hot and cold water baths.
The differences in thermal expansion coefficients
between porcelain, resin, and metal result in
stresses that cause fatigue. Recent experiments

AmericanJournalof OrthodonticsandDentofacialOrthopedics
Volume109, No. 4

Zachrisson et al. 421

Table I. Characterization of dental procelains as feldspathic porcelains, aluminous porcelains, and glass
ceramics, and commonly studied brands of the different types.* Information regarding the
percentage of aluminum oxide has been provided by the manufacturers

Feldspathicporcelains

Aluminousporcelains

Glassceramics

HiCeram (Vident)- 52%

Vitadur N Core (Vident)-50%
NBK 1000core (DeTrey/Dentsply)--50%
In-Ceram (Vident)
before glassinfiltration-100%
after glass infiltration--80%

Ceramco lI (J&J)--10%-15%
Vitadur N (Vident)- 14%
G-Cera (GC Int.)-12%-16%
Biodent (DeTrey/Dentsply)- 14%-16%
Optec (Jeneric/Pentron)- 17%-19%
VMK 68 (Vident)- 18.5%

Dicor (DeTrey/Dentsply)-< 2%
Empress (IPS)- 19%-23.5%

*For details, see r e f e r e n c e s l , 3, 4, 17, 18, 36, 39, 42, 43.

Table II. Mean tensile bond strengths and mode of failure in pilot study 2 (i.e., silane application with all

combinations and not thermocycling). Note abundance of cohesive failures (ARI score 3 and
porcelain fracture)

~
Sandblasting +
All-Bond 2 Primers A + B
Sandblasting
Sandblasting + APF-gel
H y d r o f l u o r i c acid
H y d r o f l u o r i c acid +
All-Bond 2 Primers A + B

Dislodged* ~
17.1

1.7

16.3
15.9
15.8
14.4

2.6
2.8
2.6
2.7

Porcelainfracture
1
.

5
.
.

.
.

1
.
.
.

1
.

.
.

.
8
7

*During thermocyclingor testing.

have shown that rigorous thermocycling (500 times
or more) of the bonds to porcelain is necessary to
approximate the clinical reality (see Discussion). If
thermocycling is not performed, high laboratory
strengths may be found, which may not correspond
to the chairside experiences. 13'~9'z Therefore the
clinical relevance of some of the previous studies
appears limited.
However, more recent reports have used adequate thermocycling. Eustaqio et alJ 2 studied the
tensile bond strength to metal-ceramic crowns
(Ceramco II, Johnson & Johnson, Skillman, N.J.)
with water storage at 37 C for 24 hours and 2500
thermocycles from 16 C to 56 C. They found bond
strengths from 5 to 7 MPa for Scotchprime +
Concise (3M Corp., St. Paul, Minn.) and Ormco
porcelain primer with System 1 + (Ormco Corp,
Glendora, Calif.). The bond strengths to glazed
and deglazed porcelains were not significantly different, and these authors therefore recommended
to bond to glazed porcelain to minimize surface
damage. Most failures occurred in the adhesiveforacket interface, and these values may therefore not reflect the real adhesive strength to

porcelain. The debonding strengths will represent

the true adhesive force only if the failure is of an
adhesive nature, i.e., is located in the adhesive
interface, and not if they are cohesive fractures, i.e.,
failure in one of the materials to the side of the
interface.
Aggressive thermocycling was also performed by
Kao et al. ~5 in shear studies of brackets bonded to
veneers made of two types of feldspathic porcelains
(Ceramco II, and Vitadur-N, Vident, Baldwin Park,
Calif.). The results showed increased bond strengths
with the application of a silane primer (Scotchprime), and stronger bonds to roughened than to
glazed surfaces. A large filler particle adhesive
(Concise) gave stronger bonds than a microfilled
resin (Unite, 3M/Unitek, Monrovia, Calif.). They
also discussed the risk existing for the different porcelains to fracture during debonding.
Zelos et al. 16 investigated the bond strength
of ceramic brackets bonded to two different feldspathic porcelains (Ceramco and VMK 68, Vident,
Baldwin Park, Calif.). The bonded samples were
cycled 500 times between 8 C and 54 C. The
results showed stronger bonds to Ceramco than to

422

Zachrisson et aL

Arneffcan Journal of Orthodontics and Dentofacial Orthopedics

April 1996

MATERIAL AND METHODS

Fig. 1. Cohesive failure within porcelain. Sample of pilot study

2 (Table II). Per, porcelain; Go, gold alloy supporting porcelain.

V M K 68 porcelain, and the O r m c o prime + Concise

bonds were significantly stronger than those obtained with Scotchprime + Trans-bond
(3M/
Unitek, Monrovia, Calif.). The tensile type debonding methods were the least likely to damage the
porcelain surface, with a majority of failures at the
bracket/composite interface. On the contrary,
samples debonding with a shear m o d e experienced a
high percentage of porcelain destruction.
O t h e r common procedures than silane coupling
to increase the bond strength to porcelain may
include etching with hydrofluoric acid 4'19 and
roughening by sandblasting. 4 T h e etching of dental
ceramics with hydrofluoric acid or other acids provides a clean, microretentive surface before bonding or repair of ceramic restorations2 Roughening
of ceramic materials by sandblasting has been used
as a substitute for etching. 4 Silane is generally
applied after p r e t r e a t m e n t in the form of etching
and sandblasting. O u r clinical experiments since
1991 that use a 2-minute etching with 9.6% hydrofluoric acid gel, followed by silane application and
bonding with Concise, have provided excellent
bond strengths. ~ However, there are also drawbacks with the clinical use of hydrofluoric acid. It is
a very strong acid that requires careful isolation of
the working area for several minutes.
The purpose of this study was to evaluate the
bond strengths of orthodontic brackets to a feldspathic dental porcelain, after sandblasting and silane application, with or without additional use of a
bond reinforcing intermediate resin, and compare
these bond strengths with those obtained with porcelain etchants. Orthodontic bonding to aluminous
porcelains and glass ceramics is not addressed in
this study.

Before the main study could be made, several pilot

studies were needed to find an experimental model that
adequately tested the tensile bond strength of composite
resins to the porcelain surface and avoided porcelain
fractures.
Pilot study 1. Initial experiments were performed on
pure porcelain pieces. Flat rectangular tabs of a commonly used porcelain for metal bonding (Biodent) were
produced and fired to a standardized size of 9 5 x 1.5
mm according to the manufacturer's recommendations.
Six groups of 12 brackets each were bonded with Concise. After sandblasting to remove the glaze, surface treatment comprised etching with 9.6% hydrofluoric acid, 4%
APF gel, additional silane priming (Scotchprime); all with
and without intermediate application of All-Bond 2 Primers A + B before bonding. After water storage at 37 C for
24 hours and thermocycling from 5 C to 55 C and back
1000 times, tensile tests were performed. However, in the
vast majority of cases, the porcelain tabs broke into two
pieces before the bonds failed, and this approach had to
be abandoned.
Pilot study 2. The second experiment was performed
on gold alloy tabs supporting the porcelain samples. Flat
rectangular tabs of a high-noble-metal alloy, Esteticor
Plus (Cendres & Mdtaux SA, Biel-Bienne, Switzerland),
were cast, as described elsewhere, 21 and the Biodent
porcelain were fired onto the metal pieces in a porcelainfiring oven. Five groups of 12 brackets each were used:
(1) 9.6% hydrofluoric acid gel for 4 minutes; (2) d:o plus
All-Bond 2 Primers A + B (Bisco Dental, Itasca, Ill.); (3)
sandblasting with 50 txm aluminum oxide for 5 seconds;
(4) d:o plus 4% APF gel (Chameleon Dental Products,
Inc., Kansas City, Kan.) for 2 minutes; and (5) sandblasting plus All-Bond 2 Primers A + B . A silane agent
(Scotchprime) was applied to all specimens in this series.
After water storage at 37 C for 24 hours, but without any
thermocycling, the tensile bond strengths were examined.
As shown in Table II, the bond strengths thus
achieved were quite high for all combinations and generally exceeded the cohesive strength of the porcelain.
Only in five cases did the failure occur in the adhesive
interface. All other fractures were cohesive and were
located in one of the materials to the side of the
porcelain, either within the porcelain (Fig. 1) or in the
composite/bracket interface.
Final experimental design. Flat rectangular gold alloy
tabs (Esteticor Plus) of a standardized size of
9 6 x 1.5 mm were cast, as described elsewhere, 21 and
a commonly used feldspathic porcelain for metal bonding
(Biodent) was then fused to the metal as recommended
by the manufacturer, with the final glaze being fired at
945 C for 2 minutes without vacuum? One hundred and
twenty GAC microarch mandibular brackets (GAC International Inc., Central Islip, N.Y.) with a base area of
9.4 mm ~ were bonded to each of 60 porcelain-fused-tometal tabs. Before the bonding procedure, a circular ring

American Journal of Orthodontics and Dentofacial Orthopedics

Volume 109, No. 4

Z a c h r i s s o n et aL

423

I/
Fig. 2. SEM photomicrographs of sandblasted (A, B) and hydrofluoric acid etched (C, D) feldspathic
porcelain surfaces, showing increased micromechanical retention. Original magnifications: A and C
x 300, B and D x 2000. Bar is 10 ixm.

was soldered onto the bracket slot to control the type of

loading stress and to reduce the risk of introducing
peripheral stress concentrations along the bond plane, as
discussed earlier. 2''22
The porcelain surfaces were deglazed by aluminum
oxide sandblasting with 50 txm abrasive powder with a
Microetcher erc (Danville Engineering, San Ramon,
Calif.) at 7 kg/cm2 of air pressure for 5 seconds from
a distance of 10 ram. The specimens were rinsed with
water and air spray for 15 seconds, and air dried. The
sandblasted samples were divided into six groups of 20
brackets each and subjected to the following treatments
(Table III): (1 and 2) controls, no further surface
treatment; (3) silane application (three coats of
Scotchprime, dull surface); (4) silane application plus the
intermediate resin All-Bond 2 Primers A + B (three
coats, shiny surface); (5) etched with 9.6% hydrofluoric
acid gel for 4 minutes; and (6) etched with 4% APF gel
for 2 minutes.
The silane, intermediate resin, and adhesive materi-

als were manipulated in strict accordance with the directions of the manufacturers or as discussed previously. 2~'22
Selected specimens were prepared for scanning electron
microscopy (SEM) and placed on SEM studs, coated
with a conductive layer of gold and palladium (about
300 ~ ) and examined in a Philips SEM 515, operated at
14.8 to 18.1 kV. Fig. 2 shows the microscopic appearance
of sandblasted and hydrofluoric acid-etched porcelain
surfaces, respectively.
Two brackets were bonded at right angles to one
another onto each porcelain sample (Fig. 3, A). The
specimens in groups 1 and 2 were bonded with a modified Concise composite resin 23 and MCP-bond (Sun
Medical Co. Ltd., Tokyo, Japan), respectively. The latter
is a new all-purpose 4-META adhesive derived from
Superbond C&B (Sun Medical Co. Ltd., Tokyo, Japan)
marketed in the United States as C&B Metabond
(Parkell, Farmingdale, N.Y.). The brackets in groups 3-6
were bonded with modified Concise. The excess adhesive
outside the brackets was left to allow undisturbed setting.

424

Zachrisson et al.

American Journal of Orthodontics and Dentofacial Orthopedics

April 1996

Table III. Final experimental design

Groups
1,2
3
4
5
6

Bonding adhesive

Surface treatment
Sandblasting
Sandblasting +
Sandblasting +
Sandblasting +
Sandblasting +

silane
silane + All-Bond 2 Primers A + B
9.6% hydrofluoric acid gel
4% APF-gel

After polymerization, all specimens were transferred to a

water bath at 37 C for short-term storage for 24 hours, z4
Thereafter, they were thermocycled from 5 C to 55 C
and back to 5 C 1000 times (Fig. 3, B ) . 24 The exposure to
each bath was 20 seconds, and the transfer time between
baths was 4 seconds.
When removed from the water container, the excess
adhesive outside the bracket bases were carefully removed with a small round tungsten-carbide (TC) bur to
standardize the surface area during the bond strength
testing (Fig. 3, C). Next, the specimens were mounted
in a specially made holding device on a Lloyd 1000R
testing machine (Lloyd Instruments Ltd., Fareham,
Hants, England). The tensile load was applied perpendicularly to the porcelain surface through a stainless steel
hook (0.8 mm thick), engaging the circular ring of the
bracket (Fig. 3, D). A crosshead speed of I ram/rain was
used until failure occurred. The force required to dislodge the bracket was recorded electronically on a graph
and measured in Newtons (Fig. 3, E). The force per unit
area required for breakage was calculated and reported
as the tensile bond strength in megapascals (MPa). The
fracture sites were examined to determine the location
of failure during debonding, and classified according to
the modified Adhesive Remnant Index (ARI) system of
~rtun and Bergland. 25
For reference, 12 caries-free mandibular incisors extracted for periodontal reasons were used. After storage
in distilled water, these control teeth were embedded in
acrylic blocks and prepared for bracket placement. After
conditioning with 37% phosphoric acid for 60 seconds,
identical brackets with soldered rings were bonded to the
enamel surface with Concise. The tensile bond strength
testing was performed after water storage at 37 C for 24
hours. They were not thermocycled because, compared
with specimens stored at 37 C, thermocycling apparently
has a marginal effect on the bond strength of composite
resins to etched enamel, z6
The significance of the differences between the results presented in Table IV was estimated by analysis of
variance (ANOVA) and by Duncan's Multiple-Range
test at a 5% level of significance. 27
RESULTS

Table I V presents the m e a n tensile bond

strengths for orthodontic steel brackets bonded to

Concise, MCP-bond
Concise
Concise
Concise
Concise

porcelain subjected to different types of surface

treatment, that is, silane primer and intermediate
resin, etching, and sandblasting only. The A N O V A
and Duncan Multiple-Range methods showed no
significant differences in bond strength between the
hydrofluoric acid etched porcelain and the silanetreated and silane plus intermediate resin-treated
specimens. Neither of these means were significantly different from the bond strength of Concise
to etched enamel. On the other hand, the APFgel etched porcelain and the sandblasted only
samples showed significantly lower tensile strength
values. There was a significant (P < 0.05) increase in bond strength with the application of
silane on a sandblasted porcelain surface. The
bond strength obtained with MCP-bond to sandblasted porcelain was higher than that achieved
with Concise, but the difference was small and not
statistically significant.
Table V is a tabulation of the mode of failure
for the samples tested. The majority of failures for
the specimens treated with sandblasting plus silane
plus All-Bond 2 Primers A + B occurred in the
composite/bracket interface (Table V, Fig. 4). In
contrast, less than half of the sandblasted plus
silane treated samples had such fractures, and most
failures with this combination were located in the
porcelain/composite interface. All the specimens
with sandblasting and etching, and sandblasting
only, had the majority or all fractures occurring in
the porcelain/composite interface. Notably, there
were no cohesive failures within the porcelain in
this series.
DISCUSSION

The results of this in vitro study indicates that

adequate bond strength to porcelain crowns and
veneers should theoretically be possible with different approaches. After deglazing the porcelain surface by sandblasting, either (1) etching with hydrofluoric acid, or (2) application of silane plus AllBond 2 Primers A + B can be used. As mentioned,

Zachrisson et al.

American Journal of Orthodontics and Dentofacial Otlhopedics

Volume 109, No. 4

425

Fig. 3. Experimental design. After soldering steel ring (SR) in slot of mandibular incisor brackets, two
brackets (Br) were bonded to each porcelain-fused-to-gold alloy tab at right angles to one another
(A). After short-term water storage, specimens were thermocycled 1000 times in water baths of 5 C
and 55 C (B). Excess adhesive (Adh) was removed with small, round TC bur (C), and brackets were
subjected to tensile load testing (open arrow in D indicates director of pull) in Lloyd 1000R machine
(E). D represents boxed area in E (at open arrow). W, Wheel guiding travel of sample-containing
basket (Ba) from bath to bath; H, hook engaging ring of bracket; S, slot of holding device; CC, control
console; Mo, television monitor; Pr, printer; K, keyboard.

however, there are several problems associated

with the clinical interpretation of laboratory data
on bond strengths of composite resin to dental
porcelains. The uncertainties relate to (1) the special characteristics of different porcelains; (2) the
difficulty in duplicating in vivo conditions in the
laboratory; and (3) the question of the clinical
significance of gradually increasing tensile or shear
load testing until failure occurs. 2' It is emphasized
therefore that clinical tests are needed to verify the
validity of the present in vitro results. The following
discussion will first address several theoretical considerations, before dealing with some clinical reflections regarding risks in debonding.

Porcelain Varieties

One problem in vivo when bonding brackets to

porcelain crowns and veneers is the multitude of
dental porcelains available (Table I). When bonding
to most porcelain-fused-to-metal crowns and bridges,
and conventional jacket crowns, the porcelain is most
probably feldspathic with low or medium aluminum
oxide content. For veneers, both conventional porcelains and castable glass ceramics (Dicor, Dentsply
GmbH, Dreieich, Germany) are currently used.-" It is
virtually impossible for an orthodontist to differentiate between various types and brands of porcelain in
a clinical situation. Also, many commercially available porcelains despite being similar in chemical

426

Zachrisson et al.

AmericanJournalof Orthodontics"andDentofacialOrthopedics
April1996

Table IV. Mean tensile bond strengths of 120 brackets bonded with Concise to a feldspathic dental
porcelain (Biodent) after various surface treatments. All samples were stored in water at 37 C for
24 hours and then thermocycled 1000 times from 5 C to 55 C and back

[
Sandblasting +
Sandblasting +
Sandblasting +
Sandblasting 4Sandblasting?
Sandblasting

MPa

silane + All-Bond 2 Primers A + B

silane
hydrofluoric acid
APF-gel

For reference:
Concise to etched enamel

SD

Duncangrouping*

11.7
11.6
11.5
3.4
2.8
2.5

2.8
2.9
2.8
1.2
0.7
0.7

A
A
A
B
B
B

13.2

4.4

*Means with the same letter are not significantly different (p > 0.05).
?Bonded with MCP-bond.

Table V. Mode of failure of 120 brackets bonded with Concise to a feldspathic dental porcelain (Biodent)

after various surface treatments. All samples were stored in water at 37 C and then thermocycled
1000 times from 5 C to 55 C and back. Note absence of cohesive porcelain failures

Indexscore

Dislodged* ~
Sandblasting +
Sandblasting +
Sandblasting +
Sandblasting +
Sandblasting?
Sandblasting

silane + All-Bond 2 Primers A + B

silane
hydrofluoric acid
APF-gel

1
5
4
6

]
Porcelainfracture

2
10
16
15
16
14

2
2
1

1
-

16
7
2
-

*Adhesive R e m n a n t Index ( A R I )
score 0 = no composite left on porcelain surface.
score 1 = less than half of composite left.
score 2 = more than half of composite left.
score 3 = all composite left on porcelain surface, with distinct i mpre s s i on of bracket mesh.
*During thermocycling or testing.
?Bonded with MCP-bond.

formula may still represent distinct differences in

constituents, crystalline structure, particle size, sintering behavior, and microtopography by etching.2s
In their report on shear strengths of composite resin
to nine different feldspathic porcelains subjected to
sandblasting only, sandblasting combined with silane
application, etching with hydrofluoric acid, or etching plus silane application, Sorensen et al. z8 found
that although the hydrofluoric acid etching significantly increased the bond strength of composite
resin to eight of the nine porcelains, there were quite
considerable individual variations. For example, one
porcelain (G-Cera, GC Int. Tokyo, Japan) had shear
bond values after sandblasting only that were similar
to the etched porcelain shear bond values of many of
the other porcelains tested. Two others (VMK 68

and Mirage, Chameleon Dent. Prod., Placentia,

Calif.) had a significant increase in bond strength
with the application of silane on their sandblasted
surfaces, in contrast to a negligible effect in the
remaining seven porcelains. Therefore it should be
kept in mind that simple generalized statements regarding bonding to porcelains cannot be established.
Removal of Surface Glaze?

Deglazing the porcelain or not before orthodontic bonding remains controversial. Several laboratory studies indicate that it is possible to achieve
adequate bond strengths to silane-treated glazed
porcelain 12'14"16and for safety reasons they have recommended not to remove the glaze by grinding before the bonding. However, depending on the choice

American Journal of Orthodontics and Dentofacial Orthopedics

Volume 109, No. 4

Zachrisson et aL

427

Fig. 4. SEM photomicrographs in one sample where different amounts of adhesive (Adh) remained
on porcelain (Por) surface after debonding two brackets (A). In B (left side), all composite remains,
and there are distinct impressions of bracket mesh (ARI score 3). In C (right side, turned 90),
however, less than half of composite is left (score 1).

of individual laboratory technicians, the feldspathic

porcelains may sometimes have an alumina overglaze, which cannot be distinguished. The silane will
then be ineffective, because it will not enhance the
bond to porcelains, which contain only a small
amount of silica. 4 Both self-glazes and overglazes
may be effective in strengthening the porcelain,
thereby reducing crack propagation, 2 but careful
polishing may also help improve the strength of the
porcelain restoration. 2 Fracture resistance is improved by etching and bonding techniques? A careful removal of the glaze in the immediate area of the
bond is probably of little harmful consequence, 4"25
and it will increase the surface area available for
chemical or mechanical retention. 9'11'~6 The deglazing should preferably be done with a sandblaster and
not with a green stone, which may produce micro
cracks? 2'~4'15After the orthodontic treatment is finished, the original smoothness of the porcelain surface may be restored with slow speed polishing rubber wheels, and enamel-like gloss can be created by
diamond polishing paste. 2~ Further discussions on

regimens for polishing porcelain intraorally are presented elsewhere. 9'' 1-15,19.30-32
Effect of Hydrofluoric Acid Etching

Hydrofluoric acid etching has been shown to

significantly improve the bond strengths for different feldspathic porcelains, but not for high-aluminous porcelains, such as Hi-Ceram (Vident, Baldwin Park, Calif.). 2~'33'34
Apparently, increasing aluminum oxide content
although increasing the strength of the porcelain
also makes it more resistant to chemical attack and
reduces the effect of hydrofluoric acid etching. The
finding in this study that hydrofluoric acid is more
effective than APF is in agreement with others. ~7'3s-39
The SEM photomicrographs (Fig. 2, C and D) revealed surfaces very similar to those in previous studies, 37-3~ and showed that the hydrofluoric acid etching provides excellent micromechanical retention for bonding to feldspathic
porcelains. The depth of porcelain etching was estimated by Yen et al. 3~ to be in the range of 5 to 7/xm

428

Zachrisson et al.

American Journal of Orthodontics and Dentofacial Orthopedics

April 1996

Fig. 5. Gentle, yet effective debonding may be achieved with 095 anterior bond removing plier
(Orthopli), by using peeling tensile pull in direction indicated by white arrow in A. Indent (In) of plier
grips gingival tie-wing and plastic cover (PC) rests on incisal edge. Peripheral stress concentration is
induced so that bracket comes off at low force level, with all composite being left on tooth (open
arrows in B).

only, with no reduction in flexural strength occurring. Interestingly, considerable microretention was
apparently also achieved with the sandblasting (Fig.
2, A and B).
Problems in Imitating the Oral Environment

The complex oral environment with variations in

temperature, stresses, humidity, acidity, and plaque
is not reproducible in the laboratory. It appears that
many previous in vitro studies on porcelain repairs
with composite resin and orthodontic bonding to
porcelain have provided high bond strength values
that probably do not correspond well with the clinical performance. The adversarial forces of polymerization shrinkage, thermal expansion coefficient
mismatch, and hygroscopic expansion of composite
greatly challenges the bonds to dental porcelains.
The storage conditions recommended in 1993 for
quality testing of adhesive materials by the International Organization for Standardization, the ISO
document,24 gives three possibilities: short- (24
hours) or long-term (6 months) storage in water at
37 C, and thermocycling 500 times from 5 C to
55 C and back. The results of this study strongly
indicate that the thermocycling is necessary for testing silane-coupled bonds to porcelain. A comparison
of Tables II and IV through V shows that, if water
storage only is used without thermocycling, the bond
strengths to porcelain as well as the incidence of
cohesive porcelain fractures are excessively high.
This corroborates recent findings by Sorensen
et al. a8'4 In contrast to observations in several
previous studies,35'36'41-43 in which the samples had
been submitted to no or limited thermocycling, they
found that the application of silane to an etched
porcelain surface was not improving bond strength.

It was claimed therefore that porcelain bonding

studies that do not store samples in 37 C water and
thermocycle the specimens 500 or more cycles,
probably offer figures that misrepresent the actual
contribution of silane coupling agents to the longterm in vivo bond strengths. Similarly, Pratt et al. 44
found that long-term water storage (3 months at
37 C) followed by thermocycling for 24 hours
significantly lowered the mean bond strengths of
different resins to feldspathic porcelain.
Cohesive Fractures

As mentioned, the debonding strength values

may represent the true adhesive force of composite
to porcelain only if cohesive fractures can be
avoided. A failure in one of the materials to the
side of the interface indicates that the physical
properties of that material limits the bond strength
of the assembly.45 In orthodontic bond strength
testing, cohesive fractures in the composite resin
(ARI score 3) reflect the internal strength of the
composite rather than the actual adhesion to the
surface under study. In most studies on porcelain
repair systems, particularly in those where thermocycling has not been used, many or even a
majority of fractures have been cohesive in the
porcelain, z~'33'35"36'42'44'46-4"~It is noteworthy therefore
that there were no cohesive porcelain fractures at
all in the main part of this study (Table V). This
corroborates the findings in the study of Eustaqio
et al., '~ in which a similar thermocycling regimen
was used, and reinforces the need for temperatureinduced stress in studies on bonding to porcelain.
Furthermore, in reviewing Table IV, it becomes
evident that using the All-Bond 2 Primers on a
sandblasted and silane treated porcelain surface

American Journal of Orthodontics and Dentofacial Orthopedics

Volume 109,No. 4

m a y be of merit. 49 T h e failure m o d e thus c h a n g e d

f r o m mostly adhesive (score 0) to p r e d o m i n a n t l y
cohesive in the composite (score 3). This change
reflects a higher quality of the b o n d to porcelain, even if the m e a n values for tensile strength
were not statistically significantly different (Table
III).
Bracket Removal/Clinical Considerations

O n the basis of findings of high laboratory

strength data w h e n no or limited thermocycling was
performed, and findings that cracks and porcelain
fractures o c c u r r e d during d e b o n d i n g in machines,
s o m e authors have w a r n e d against trying to obtain
too high b o n d forces to porcelain crowns and
v e n e e r s . 9"11"14"15 However, b e c a u s e the type of deb o n d i n g force in machines is not the same as the
force applied in careful clinical debonding, the risk
for d a m a g e during d e b o n d i n g n e e d n o t be a
p r o b l e m with gentle, still effective m a n u a l techniques. As r e c o m m e n d e d elsewhere, 13'2 a peeling
tensile pull will cause peripheral stress concentration in the composite, 5 and the bracket
will c o m e off at a low force, 45 usually with the
fracture at the m e s h interface ( A R I score 3). This
can be achieved b o t h with an anterior b o n d removing plier applied to a gingival tie-wing (Fig.
5, A), or the bracket wings m a y be squeezed with
a W e i n g a r t plierJ 9'2 If a ceramic bracket does not
loosen easily on a v e n e e r or crown, it should be
g r o u n d away with d i a m o n d and water cooling. 5~
T h e residual adhesive (Fig. 5, B) can be r e m o v e d
with a T C bur almost without touching the porcelain surface, which is t h e n polished as previously
discussed. 29-32
CONCLUSION

T h e continuously increasing load applied to

b o n d e d brackets in vitro is not the same type of
stimulus that occurs clinically. B o n d e d brackets in
the oral cavity are subjected to shear, tensile and
torsional forces, and to combinations of all these.
Except for traumatic incidents, brackets coming
loose in the m o u t h p r o b a b l y do so as a result of
r e p e a t e d stress producing micro cracks that p r o p a gate until b o n d failure occurs. Hence, load fatigue
testing, 33 in which a sample is repeatedly subjected
to a load well below the level that causes fracture in
static tensile or shear tests, and eventually fails,
m a y provide interesting information. Llobell et al. 33
recently r e p o r t e d on the load fatigue characteristics
of eight newer generation porcelain repair systems,
in which composite resins w e r e b o n d e d to feld-

Z a c h r i s s o n et aL

429

spathic porcelain (VMK). Only two consistently

survived the 2000.000 cycle limit, that is, the Clearfil Porcelain B o n d and All-Bond 2 systems.
We are grateful to Professor Jon Orstavik, Department of Prosthodontics, University of Oslo, Norway, for
his kind assistance associated with the fabrication of the
porcelain-fused-to-gold alloy specimens. The thermocyclings were performed at several occasions at the Scandinavian Institute of Dental Research (NIOM), Oslo,
Norway, through the courtesy of its Director, Professor
Gudbrand Oilo. The MCP-bond was developed and
placed at our disposal by the generosity of Dr. Morio
Takeyama, Director, Sun Medical Co., Ltd., Tokyo,
Japan. We also thank Donald S. McCauley, Vice President, GAC International Inc., N.Y., for the generous
support of brackets for our studies on bonding to
artificial tooth surfaces; and Mr. Jens Gundersen, Director, 3M Norge A/S, for supplying the Scotchprime
ceramic primer.
REFERENCES

1. Oilo G. Flexural strength and internal defects of some

dental porcelains. Acta Odontol Scand 1988;46:313-22.
2. Giordano RA, Campbell A, Pober R. Flexural strength of
feldspathic porcelain treated with ion exchange, overglaze,
and polishing. J Prosthet Dent 1994;71:468-72.
3. Castellani D, Baccetti T, Giovannoni A, Bernardini UD.
Resistance to fracture of metal ceramic and all-ceramic
crowns. Int J Prosthodont 1994;7:149-53.
4. Kern M, Thompson VP. Sandblasting and silica coating of a
glass-infiltrated alumina ceramic: volume loss, morphology,
and changes in the surface composition. J Prosthet Dent
1994;71:453-61.
5. Ghassemi-Tary B. Direct-bonding to porcelain: an in vitro
study. AM J ORTHOD 1979;76:80-3.
6. Johnson RG. A new method for direct bonding orthodontic
attachments to porcelain teeth using a silane coupling agent:
an in vitro study. AM J ORTHOD 1980;78:233-4.
7. Newman GV. Bonding to porcelain. J Clin Orthod 1983;17:
53-5.
8. Newman SM, Dressier KB, Grenadier MR. Direct bonding
of orthodontic brackets to esthetic restorative materials
using a silane. AM J ORTHOD 1984;86:503-6.
9. Wood DP, Jordan DC, Way DC, Galil KA. Bonding to
porcelain and gold. AM J ORTHOD 1986;89:194-205.
10. Stokes AN, Hood JAA, Tidmarsh BG. Surface preparation
for bonding to porcelain and gold. Austr Orthod J 1989;9:
321-3.
11. Smith GA, Mcfnnes-Ledoux P, Ledoux WR, Weinberg R.
Orthodontic bonding to porcelain - bond strength and
refinishing. AM J ORTHOD DENTOFAC ORTHOP 1988;94:
245-52.
12. Eustaquio R, Garner LD, Moore BK. Comparative tensile
strengths of brackets bonded to porcelain with orthodontic
adhesive and porcelain repair systems. AM J ORTHOD
DENTOFACORTHOP 1988;94:421-5.
13, Andreasen GF, Stieg MA. Bonding and debonding brackets
to porcelain and gold. AM J ORTHODDENTOFACORTHOP
1988;93:341-5.

430

Zachrisson et al.

14. Kao EC, Boltz KC, Johnston WM. Direct bonding of orthodontic brackets to porcelain veneer laminates. AM
J ORTHOD DENTOFAC ORTHOP 1988;94:458-68.
15, Kao EC, Johnston WM. Fracture incidence on debonding of
orthodontic brackets from porcelain veneer laminates.
J Prosthet Dent 1991;66:631-7.
16. Zelos L , Bevis RR, Keenan KM. Evaluation of the ceramic/ceramic interface. AM J ORTHOD DENTOFAC ORTHOP 1994;106:10-21.
17. Morena R, Lockwood PE, Fairhurst CW. Fracture toughness of commercial dental porcelains. Dent Mater 1986;2:
58-62.
18. White GJ, Tyas MJ. The bond strength of orthodontic resins
to porcelain. Austr Orthod J 1993;16:8-12.
19. Zachrisson BU, Bnyukyilmaz T. Recent advances in bonding
to gold, amalgam, and porcelain. J Clin Orthod 1993;27:66175.
20. Zachrisson BU. Bonding in orthodontics. In: Graber TM,
Vanarsdall RL, eds. Orthodontics: current principles and
techniques. 2nd ed. St Louis: Mosby-Year Book, 1994:
542-626.
21. Buyukyilmaz T, Zachrisson YO, Zachrisson BU. Improving orthodontic bonding to gold alloy. AM J ORTHOD
DENTOFAC ORTHOP 1995;108:510-9.
22. Zachrisson BU, Buyukyilmaz T, Zacbrisson YO. Improving
orthodontic bonding to silver amalgam. Angle Orthod 1995;
65:35-42.
23. ~ r t u n J, Zachrisson BU. Improving the handling properties
of a composite resin for direct bonding. AM J ORTHOD
1982;81:269-76.
24. International Organization for Standardization. ISO TR
11405 dental materials - guidance on testing of adhesion to
tooth structure. Geneva, Switzerland: WHO, 1993.
25..3~rtun J, Bergland S. Clinical trials with crystal growth
conditioning as an alternative to acid-etch pretreatment.
AM J ORTHOD 1984;85:333-40.
26. Draughn RA. The effect of thermal cycling on retention of
composite restoratives. (IADR abstr 303.) J Dent Res
1976;55:B137.
27. Bruning JL, Kintz BL. Computational handbook of statistics. Glenview: Scott, Foresman, 1977.
28. Sorensen JA, Engelman MJ, T0rres TJ, Avera SP. Shear
bond strength of composite resin to porcelain. Int J Prosthod 1991;4:17-23.
29. Polishing ceramics (Editorial commentary). Reality Now
1994;56:31.
30. Patterson CJW, McLundie AC, Stirrups DR, Taylor WG.
Efficacy of a porcelain refinishing system in restoring surface finish after grinding with fine and extrafinc diamond
burs. J Prosthet Dent 1992;68:402-6.
31. Hulterstrom AK, Bergman M. Polishing systems for dental
ceramics. Acta Odontol Scand 1993;51:229-34.
32. Scurria MS, Powers JM. Surface roughness of two polished
ceramic materials. J Prosthet Dent 1994;71:174-7.
33. L10bell A, Nichols JI, Kois JC, Daly CH. Fatigue life of
porcelain repair systems. Int J Prosthod 1992;5:205-13.
34. Calamia JR, Vaidyanathan J, Vaidyanathan TK, Hirsch SM.
Effect of coupling agents on bond strength of etched porcelain. (1ADR abstr 1096.) J Dent Res 1985;64:296.

American Journal of Orthodontics and DentofaciaI Orthopedics

April 1996
35. Lacy AM, LaLuz J, Watanabe LG, Dellinges M. Effect of
porcelain surface treatment on the bond to composite.
J Prosthet Dent 1988;60:288-91.
36. Bertolotti RL, Lacy AM, Watanabe LG. Adhesive monomers for porcelain repair. Int J Prosthod 1989;2:483-9.
37. Al Edris A, A1 Jabr A, Cooley RL, Barghi N. SEM evaluation of etch patterns by three etchants on three porcelains.
J Prosthet Dent 1990;64:734-9.
38. Sheth J, Jensen M, Tolliver D. Effect of surface treatment
on etched porcelain bond strength to enamel. Dent Mater
1988;4:328-87.
39. Yen TW, Blackman RB, Baez RJ. Effect of acid etching on
the fiexural strength of a feldspathic porcelain and a
castable glass ceramic. J Prosthet Dent 1993;70:224-33.
40. Sorensen JA, Kang SK, Avera SP. Porcelain-composite
interface microleakage with various porcelain surface treatments. Dent Mater 1991;7:118-23.
41. Newburg R, Pameijer CH. Composite resins bonded to
porcelain with silane solution, J Am Dent Assoc 1978;96:
288-91.
42. Tjan AHL, Nemetz H. A comparison of the shear bond
strength between two composite resins and two etched
ceramic materials. Int J Prosthod 1988;1:73-9.
43. Nichols JI. Tensile bond of resin cements to porcelain
veneers. J Prosthet Dent 1988;60:443-7.
44. Pratt RC, Burgess JO, Schwartz RS, Smith JH. Evaluation
of bond strength of six porcelain repair systems. J Prosthct
Dent 1989;62:11-3.
45. Oilo G. Bond strength testing-what does it mean? Int Dent
J 1993;43:492-8.
46. Thomas CA, Hartsock TE, Dossett J, Mitchell RJ. Thermocycling, silanization, and tensile bond strength of resin to
porcelain. (IADR abstr 808.) J Dent Res 1987;66:207.
47. Suliman AHA, Swift EJ, Perdigao J. Effects of surface
treatment and bonding agents on bond strength of composite resin to porcelain, J Prosthet Dent 1993;70:118-20.
48. Appeldoorn RE, Wilwerding TM, Barkmeier WW. Bond
strength of composite resin to porcelain with newer generation porcelain repair systems. J Prosthet Dent 1993;70:6-11.
49. Kanca J. Dental adhesion and the All-Bond system. J Esthet
Dent 1991;3:129-32.
50. Katona TR, Moore BK. The effects of load misalignment on
tensile load testing of direct bonded orthodontic brackets a finite element model. AM J ORTHOD DENTOFAC ORTHOP
1994;105:543-51.
51. Vukovich ME, Wood DP, Daley TD. Heat generated by
grinding during removal of ceramic brackets. AM J ORTHOD
DENTOFAC ORTHOP 1991;99:505-12.
Reprint requests to:
Prof. Bjorn U. Zachrisson
Department of Orthodontics
Dental Faculty
PO Box 1109 Blindern
N-0317 Oslo, Norway