dental

materials
Dental Materials 18 (2002) 227±238
www.elsevier.com/locate/dental

The use of ¯owable composites as ®lled adhesives

Roland Frankenberger a,*, Manuela Lopes b, Jorge PerdigaÄo c, Wallace W. Ambrose d,
Bruno T. Rosa e
a
Policlinic for Operative Dentistry and Periodontology, University of Erlangen-Nuremberg, Glueckstrasse 11, D-91054 Erlangen, Germany
b
Minnesota Dental Research Center for Biomaterials and Biomechanics, University of Minnesota, 16-212 Moos Tower B 515 Delaware St SE, Minneapolis,
MN 55455, USA
c
Division of Operative Dentistry, Department of Restorative Sciences, University of Minnesota, 8-450 Moos Tower B 515 Delaware St SE, Minneapolis,
MN 55455, USA
d
School of Dentistry, Dental Research Center, University of North Carolina at Chapel Hill CB no. 7455, Chapel Hill, NC 27599-74550, USA
e
Private practice, Londrina, PR, Brazil
Received 17 October 2000; received in revised form 23 March 2001; accepted 9 April 2001

Abstract
Objective: The effect of ®lled adhesives on bonding resin composites to dentin has not been fully understood. Due to their ®ller content,
®lled adhesives may act as stress breakers. The aim of this in vitro study was to evaluate the use of ¯owable composites of different viscosities
on bonding to enamel and dentin without the use of an intermediate bonding resin.
Materials and methods: Enamel and dentin bond strengths of OptiBond FL, Syntac Classic, and EBS Multi combined either with their
proprietary bonding agent or a ¯owable resin composite (Ultraseal XT Plus or Revolution) were measured. The tests were carried out with a
microtensile device at a crosshead speed of 1 mm/min after 24 h of storage at 378C in water. Mean bond strengths were analyzed using the
Wilcoxon test and multiple comparisons according to the Mann-Whitney U-test. The micro-morphology of corresponding resin±dentin
interfaces of the same teeth were analyzed using SEM and TEM.
Results: The control groups with adhesive systems used as per manufacturers' protocol showed bond strengths of 38.9±41.1 MPa to
enamel and 28.8±33.4 MPa to dentin. With respect to bond strength to etched enamel, only Ultraseal XT Plus as bonding resin reached the
level of the control groups. When used as bonding agents on dentin, both ¯owable composites produced lower microtensile bond strengths to
etched and primed dentin than did the control groups. Micro-morphological analysis using SEM and TEM resulted in hybrid layer formation
for both control and experimental groups. However, many areas of the resin±dentin interface showed insuf®cient penetration of the ¯owable
composites at the top of the hybrid layer as well as numerous tubules obstructed by ®ller particles.
Conclusions: The ¯owable composites tested in this study should not be used to replace bonding agents. Flowable composites of thinner
viscosity, such as Ultraseal XT Plus, may bond to enamel adequately without the requirement of an intermediate bonding resin. q 2002
Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved.
Keywords: Enamel bonding; Dentin bonding; Hybrid layer; Resin composites; Flowable composites; Filled adhesives

1. Introduction materials is fundamentally dependent on effective and

The increasing attractiveness of tooth-colored restora-
tions has promoted research in this particular area of
operative dentistry during the last few years [1±3].
For Class II restorations, resin-based restorative materials
are now being used instead of amalgam [4,5]. Due to the
characteristic polymerization shrinkage of resin-based
composites, clinical success with composite restorative
* Corresponding author. Tel.: 149-9131-853669; fax: 149-9131-
853603.
E-mail address: frankbg@dent.uni-erlangen.de (R. Frankenberger).
0109-5641/02/$22.00 + 0.00 q 2002 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved.
PII: S0109-564 1(01)00040-9
durable adhesion to enamel and dentin [6±8].
Bonding to enamel is now accepted as clinically reliable,
because acidic etchants, such as 30±40% phosphoric acid, are
capable of creating microporosities on the enamel surface
which enable the penetration of polymerizable monomers
to provide micromechanical retention [9,10]. However,
dentin still remains an unpredictable substrate for adhesion.
The wet tubular microstructure, combined with the high
organic content, is responsible for bonding to dentin being
far more dif®cult to perform [11±14]. Although different
approaches of surface pre-treatment are available, acid-
etching and subsequent penetration of reactive primer
molecules into the decalci®ed dentin surface is the most
228 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238
recommended [2,14,15]. While modern one-bottle adhesives
combine primer and bonding resin in one solution, conven-
tional two-bottle systems rely on a low-viscosity primer
acting exclusively for penetration and impregnation of the
collagen network exposed by acid etching [13,16,17]. High
vapor-pressure solvents, like acetone or ethanol, are
commonly used to facilitate the penetration and carry the
monomers into a direct contact with collagen ®bers [12,17±
19]. The primer is followed by a hydrophobic bonding resin
designed for ®lling the inter®brillar spaces as a backbone and
mechanically stabilizing the established interface generating
a mixed zone of resin-entangled collagen ®brils, the hybrid
layer [20]. The use of these hybridizing adhesive systems has
been reported to produce high bond strengths, primarily when
the dentin was left moist instead of being dried after etching
and rinsing [6,12,18].
A modi®cation was introduced when ®lled bonding
resins, such as OptiBond FL (Kerr, Orange, CA, USA)
and Permaquik (Ultradent Products, Inc., Salt Lake City,
UT, USA) were marketed [21,22]. After air thinning, the
high viscosity of these adhesives results in thicker layers at
both the cavity ¯oor and margin [6]. It has been repeatedly
reported to be promising regarding marginal and internal
seal of resin composite restorations [23±26].
This phenomenon characterized by a thicker adhesive
layer between restorative material and tooth substrates
was described as elastic cavity wall concept [26]. Using
un®lled adhesives, extensive air thinning due to different
cavity geometries might result in thin layers which could
not be polymerized due to the polymerization inhibition by
oxygen [2,12]. Thicker layers of un®lled resins may there-
fore be as advantageous for marginal adaptation as the
application of ®lled adhesives, but it cannot be recom-
mended clinically, because these materials usually provide
no radiopacity [27]. Therefore, this modi®ed use could
mislead clinicians to interprete the adhesive radiotran-
sparency as gap formation or recurrent caries at the margin
of the restoration [27].
Flowable resin composites have been reported to adapt
well to the cavity wall [5]. This optimal adaptation may
result in an improvement of the adhesive performance of
resin composites [5,28,29]. Unterbrink and Liebenberg
suggested using a ¯owable resin composite as bonding
agent to solve this problem by forming a strong and radio-
paque material at the resin±dentin interface [30,31].
Due to the fact that ®lled adhesives and ¯owable
composites appear to have similar viscosities, this tech-
nique may have a strong impact for the restoration of
carious lesions with direct resin composites. Nevertheless,
the effect of ¯owable composites on bond strengths and
interfacial micro-morphology has not been fully studied.
Therefore, the null hypothesis to be tested in this project
was that ¯owable composites would be as effective as
un®lled bonding resins when used as part of a bonding
system, by using a microtensile bond strength test (m-
TBS) combined with scanning electron microscopy
(SEM) and transmission electron microscopy (TEM)
evaluation.
2. Materials and methods
2.1. Specimen preparation and microtensile testing
A total of thirty six caries-free human third molars were
used in this investigation. The teeth were stored in 0.1%
thymol solution at ambient temperature for less than 4
weeks after extraction and were consecutively debrided
and examined to ensure that they were free of defects. For
bond strength testing, the occlusal enamel was removed and
700±900 mm thick enamel±dentin disks were cut from the
mid-coronal level of the tooth, perpendicular to the tooth
axis by slow-speed diamond-saw sectioning (Isomet,
Buehler, Lake Bluff, IL, USA) under continuous water cool-
ing. A standardized smear layer was created in both corre-
sponding surfaces by wet-sanding with 600-grit sandpaper
for 60 s [17]. Eighteen teeth were randomly assigned to
three dentin adhesives. Further, the teeth were subdivided
into three groups with a ¯uid resin and two ¯owable resin
composites acting as bonding agent. Both surfaces were
treated with the adhesive systems (Table 1) OptiBond FL
(Kerr), Syntac Classic (Vivadent, Schaan, Principality of
Liechtenstein), and EBS Multi (ESPE, Seefeld, Germany)
according to the manufacturer's instructions except for
Syntac classic which was applied upon etching dentin for
15 s like alternatively recommended in the manufacturer's
protocol. The experimental groups were treated with the
primers like in the control groups but afterwards bonded
with the ¯owable resin composites Ultraseal XT Plus (Ultra-
dent Products, Inc., Salt Lake City, UT, USA) and Revolu-
tion (Kerr, Orange, CA, USA) replacing the manufacturers'
bonding resins and scrubbed continuously for 15 s onto the
primed surface. The ¯owable composites acting as bonding
agent were thinned with a brush but not air-thinned and
consecutively light-cured for 40 s. The crowns of the ¯at-
tened teeth were then reconstructed with four 1 mm-layers
of a mini®lled hybrid resin composite (Tetric Ceram,
Vivadent) each layer being light-cured for 40 s with an
Elipar Highlight curing unit (without softstart). The intens-
ity of the light was checked periodically with a radiometer
(Demetron/Kerr, Danbury, CT, USA) to ensure that an
intensity 400 mW/cm 2 was exceeded. The disks made for
micromorphological investigations were pretreated simul-
taneously and covered with resin composite in one layer
of 1 mm thickness, then cut in half and stored in distilled
water for making the SEM and TEM specimens (Fig. 1).
The m-TBS specimens were stored in distilled water for
24 h at 378C and then sectioned. The peripheral areas of the
tooth were removed resulting in a 5 £ 5 mm square central
part. The remaining specimen was sectioned into ®ve slices,
which were sectioned again to receive 25 resin±dentin
beams. The saw was adjusted to steps of 1 mm, due to the
 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238 229

Table 1
Investigated adhesive systems, ¯owable resin composites replacing the bonding resins, and manufacturers

Adhesive Etchant Primer Bonding resin Manufacturer

system (code)

OptiBond FL Kerr Etchant OptiBond Primer OptiBond Adhesive Kerr, Orange,

CA, USA
37.5% phosphoric acid HEMA, GPDM, MMEP, ethanol, Bis-GMA, HEMA, GPDM, barium-
water, initiators aluminum borosilicate glass, disodium
hexa¯uorosilicate, fumed silica
(total ˆ 48% ®ller)
Ultraseal XT Plus (Ultradent)
Bis-GMA, TEGDMA, 53% ®ller
Revolution (Kerr)
Bis-GMA, TEGDMA, 53% ®ller
Syntac classic Total Etch Syntac classic primer Heliobond Vivadent, Schaan,
Liechtenstein
36% phosphoric acid Maleic acid, TEGDMA, water, acetone Bis-GMA, TEGDMA, UDMA
Syntac classic adhesive Ultraseal XT Plus
(2nd primer) Revolution
PEGDMA, glutaraldehyde, water
EBS Multi MiniTip Etch Gel EBS Multi Primer EBS Multi Bond ESPE, Seefeld,
Germany
32% phosphoric acid HEMA, MMC, water Bis-GMA, HEMA, TEGDMA, MAM
Ultraseal XT Plus
Revolution

Fig. 1. (a) Set up of the study divided in SEM/TEM investigation and m-TBS measurement. (b) Set up for m-TBS measurement on ¯attened enamel.
230 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238
thickness of the blade (300 mm) resulting in sticks with a
cross-sectional area of 700 £ 700 mm (0.5 mm 2). From the
resulting 25 central sticks of each tooth, ten were selected
revealing a distance to the pulp of 2.0 ^ 0.5 mm resulting in
20 microtensile specimens for each group (n ˆ 20; Fig. 1a).
For evaluating enamel bond strength, 18 third molars
were cut longitudinally and the buccal and lingual aspects
of the teeth were ¯attened on an area of 4 £ 4 mm without
dentin exposure using 600 grit sandpaper under continuous
water cooling. Then the same specimen preparation proce-
dure was carried out like in the dentin specimens without
rubbing of adhesive components. From the resulting 64
sticks of each tooth, ten beams with an enamel thickness
of 0.5 ^ 0.1 mm were selected resulting in 20 microtensile
beams in each group (n ˆ 20; Fig. 1b). For the simulation of
mixed cavities, also during enamel bonding procedures the
primers were applied according to the manufacturer's
recommendations for use.
The sticks were mounted in a Bencor Multi-T device
(Danville Engineering, Danville, CA, USA) with a special
cyanoacrylate glue (Zapit, Dental Ventures of America,
Corona, CA, USA) and debonded using an universal testing
machine model 4411 (Instron Co., Canton, MA, USA) with
a 50-N load cell traveling at a crosshead speed of 1 mm/min.
m-TBS was determined by computing the quotient of maxi-
mum load (N) and adhesion area. After the m-TBS test, the
dentin sides of the fractured interfaces were analyzed under
a stereo light microscope at £ 40 magni®cation to deter-
mine the fracture mode.
2.2. Statistical analysis
The statistical analysis was performed using SPSS 8.0 for
Windows (SPSS Inc., Chicago, IL, USA). The mean bond
strength and margin analysis were non-normally distributed
(as shown by the Kolmogorov-Smirnov test). Therefore,
non-parametric tests were used to compute differences
between dependent groups (Wilcoxon matched-pairs signed-
ranks test) and among independent groups (Mann-Whitney
U-test, correction method according to Bonferroni-Holm)
for pairwise comparisons at the 0.05 level of signi®cance
(a ). To assess the in¯uence of the different bonding resins,
the levels of signi®cance were adjusted to a * ˆ 1-(1 2 a ) 1/k
(k ˆ number of performed pairwise tests) [3].
2.3. SEM analysis
The disk halves for the resin±dentin interfacial SEM
investigation (n ˆ 2) were ®xed in 2% paraformaldehyde,
dehydrated in ascending concentrations of ethanol (50±70±
95±100%) for 1 h, immersed in hexamethyldisilazane
(HMDS, Ted Pella, Redding, CA, USA) and ®nally
embedded in epoxy resin (Epo-Thin, Buehler) exposing
the adhesive interfaces of the middle of the tooth. The
exposed interfaces were polished with wet silicon carbide
paper of decreasing abrasiveness (up to 1200 grit) and soft
tissue with increasingly ®ne diamond suspension (Buehler
Inc.) to a particle size of 1 mm. Finally the blocks were
treated with 6N hydrochloride acid for 30 s followed by a
10 min immersion in 2.5% sodium hypochlorite to enable
the visualization of a three-dimensional interface relief.
One further specimen for each enamel group was made
by etching ¯attened buccal aspects of third molars, applying
the same protocol as in the m-TBS study and dissolving the
thinned enamel by use of HCl for 3 days resulting in resin
replicas of the etch pattern. The ®nished SEM specimens
were mounted on aluminum stubs (Ted Pella) using adhe-
sive carbon sticks and a quick drying silver paint (Ted
Pella).
The specimens were sputter-coated (E 5100, Polaron
Instruments Inc., VG Microtech, Watford, England),
mounted on aluminum specimen holders to be observed
under a SEM (4000 £ magni®cation at 15 kV, JSM
6300 V scanning electron microscope, JEOL, Peabody,
MA, USA). Furthermore the thickness of the adhesive/¯ow-
able composite layer above the hybrid layer was measured
on ten points in the dentin specimens (TiffMes 1.9, Univer-
sity of Erlangen).
2.4. TEM analysis
For TEM evaluation of the resin±dentin interfaces, the
resin composite layers of the corresponding disk halves
were thinned and six small disks with a cross-section area
of 1.0±1.5 mm 2 were cut from the central area using a slow-
speed diamond saw under continuous water cooling, to
expose the opposite sides of the interfaces viewed by SEM.
Three of the resulting six sticks in each group were then
demineralized in 10% buffered EDTA (pH 7.4) for 72 h and
consecutively ®xed using 2.5% glutaraldehyde and 2%
paraformaldehyde in 0.1 M sodium cacodylate buffer at
pH 7.4 for 12 h at 48C. After the ®xation process, the speci-
mens were rinsed with 0.1 M sodium cacodylate buffer at
pH 7.4 for 2 h at 48C. Post-®xation was performed with 2%
osmium tetroxide for 1 h followed by washing in 0.1 M
sodium cacodylate buffer at pH 7.4 for 1 h. After rinsing
in deionized water for 20 min, both the demineralized and
the non-demineralized sticks were dehydrated in ascending
grades of ethanol in analogy to the SEM specimens. The
specimens were then immersed in propylene oxide for
20 min and embedded in a 50/50 mix of propylene oxide
and epoxy resin (Eponate 12, Ted Pella) in a rotator (Pelco
In®ltrator, Ted Pella) for 6 h at 6 rpm. Finally, the speci-
mens were immersed in 100% epoxy resin and set under
vacuum for 12 h for resin in®ltration into the sticks. After
positioning of the specimens in rubber molds, the sticks
were embedded in fresh resin exposing the central aspects
of the interfaces and dried in an oven for 12 h at 658C. The
blocks were sectioned in an ultra-microtome (MT 2-B Ultra-
microtome, Ivan Sorvall Inc., Norwalk, CT, USA) equipped
with a material science diamond knife (Micro Star Tech-
nologies Inc., Huntsville, TX, USA). The demineralized
sections were additionally stained with 2% uranyl acetate
 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238
Table 2
Results of dentin microtensile bond strength [MPa] (SD) for different restorative combinations. Groups with same superscript letters are not statistically
different (Wilcoxon and Mann-Whitney U-test, P . 0.05)
Adhesive system Manufacturer's adhesive
OptiBond FL 33.4 (7.2) A
Syntac classic 28.8 (8.9) A
EBS Multi 29.5 (7.7) A
for 10 min and 3% lead citrate for 5 min. After drying at
room temperature, the sections were analyzed using a CM-
12 LaB6 Transmission Electron Microscope (Philips Elec-
tronic Instruments, Mahwah, NJ, USA) at an accelerating
voltage of 100 kV.
3. Results
An overview of the results is shown in Tables 2 and 3.
Both ¯owable resin composites resulted in lower mean
dentin m-TBS when used as bonding resin (11.6±
15.9 MPa for Ultraseal XT Plus vs 7.1±13.3 MPa for
Revolution; Table 2) compared with the control groups
(28.8±33.4 MPa; P , 0.05). The different adhesive systems
tested exhibited no signi®cant differences in mean m-TBS
(P . 0.05). The failure modes were exclusively adhesive
within the resin±dentin interface.
For enamel m-TBS, the control groups exhibited no statis-
tical differences and were between 38.9 and 41.1 MPa
(Table 3). The use of Ultraseal XT Plus as a bonding resin
resulted in similar mean m-TBS (35.4±40.6 MPa) to the
control groups (P . 0.05). The application of Revolution
without an intermediate bonding resin resulted in statisti-
cally lower mean m-TBS to etched enamel (20.8±24.6 MPa;
P , 0.05) than did the original bonding agents and Ultraseal
XT Plus. The failure modes detected in the enamel groups
were adhesive between resin and enamel for the experi-
mental groups. Within the control groups, cohesive failures
in resin were detected (OptiBond FL: 20%, Syntac Classic:
10%, EBS Multi: 30%) (Table 3).
SEM analysis of the dentin specimens showed a well-
de®ned interface of the control groups (Figs. 2a, 3a, 4a),
but also penetration of the ¯owable composites into super-
®cially acid-demineralized and primed dentin, because a
hybrid layer of similar thickness as in the control groups
was clearly evident (Figs. 2b,c, 3b,c, 4b,c). However, closer
investigation exhibited distinct zones of ®ller accumulation
Table 3
Results of enamel microtensile bond strength [MPa] (SD) and percentages of cohesive failures in enamel for different restorative combinations. Groups with
same superscript letters are not statistically different (Wilcoxon and Mann-Whitney U-test, P . 0.05)
Adhesive system Manufacturer's adhesive
OptiBond FL 40.4 (8.6) A 20%
Syntac classic 38.9 (9.2) A 10%
EBS multi 41.1 (10.1) A 30%
231
Ultraseal XT Plus Revolution
15.9 (3.4) B 13.3 (4.0) B
11.6 (3.5) B 7.1 (4.1) C
12.5 (5.0) B 8.0 (3.0) C
(Fig. 2c) as well as insuf®cient hybridization of the peritub-
ular and intertubular dentin at the top of the hybrid layer
(Figs. 2b, 3b,c, 4b,c). The resin replicas of etched enamel
exhibited a less pronounced formation of resin tags extend-
ing into acid-conditioned enamel when Revolution was used
for bonding (asterisks in Fig. 5c) compared with Ultraseal
XT Plus and the control specimens (Fig. 5a,b).
The mean layer thicknesses of the adhesives in the control
group have been 1.7 ^ 1.6 mm for EBS Multi, 3.2 ^ 1.9 mm
for Syntac Classic, and 13.3 ^ 5.7 mm for OptiBond FL.
The layer thicknesses of the ¯owable composites exhibited
no statistically signi®cant difference and have therefore been
pooled revealing a mean thickness of 101.8 ^ 44.9 mm for
Ultraseal XT Plus and 173.3 ^ 57.1 mm for Revolution. All
groups have been signi®cantly different from each other
(layer thickness: EBS , Syntac , OptiBond , Ultraseal ,
Revolution; P , 0.05).
TEM investigation revealed total-etch interfaces for the
control groups (Fig. 6a) as provided by the literature
and also penetration of the ¯owable composites into acid-
demineralized dentin (Fig. 6b,c). In analogy to the SEM
observations, ®ller accumulation near the top of the hybrid
layer was clearly evident in the TEM (Fig. 6b,c). Filler
particles of the diameter of the dentinal tubule entrance
often tended to form plugs and consequently insuf®cient
resin tag formation (Fig. 6b,c) like displayed in the SEM
®gures (Figs. 2b, 4b,c).
4. Discussion
The objective of this in vitro study was to clarify whether
¯owable resin composites could be used as ®lled adhesives
for bonding of resin composites. The method utilized for
bond strength testing was the microtensile bond test
reported to be well-suited for the evaluation of bond
strengths to enamel and dentin [32±36]. This methodology
provides the possibility to investigate interfacial bond
Ultraseal XT Plus Revolution
37.4 (7.1) A 0% 24.6 (5.9) B 0%
35.3 (6.0) A 0% 25.1 (4.5) B 0%
40.6 (9.2) A 0% 20.8 (5.8) B 0%
232 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238

Fig. 2. Scanning electron micrographs of resin±dentin interfaces bonded with OptiBond FL. (®nal magni®cation: 4000 £ ). (a) (as per manufacturer's
instructions. The ®lled adhesive (A) penetrates well into the dentinal tubules forming resin tags (T) and various lateral branches (asterisks). The hybrid
layer (H) is approximately 4 mm thick (region between arrowheads). (b) Specimen etched with phosphoric acid, primed with OptiBond Primer and bonded with
Ultraseal XT Plus replacing the OptiBond adhesive (A). Note the porous upper half (hand) of the hybrid layer (H, region between arrowheads). The resin tags
(T) are approximately 10 mm long and show thin endings (asterisks). (c) Specimen etched with phosphoric acid, primed with OptiBond Primer and bonded with
Revolution instead of the OptiBond adhesive (A). The resin tags (T) show lateral branches (asterisk), a hybrid layer (H, between arrowheads) is clearly evident.
Note the accumulation of ®llers and polishing particles at the top of the hybrid layer (right hand) and the gap (left hand) resulting from insuf®cient peritubular
penetration at the tubule entrance.

strengths on small areas below 1 mm 2. Conventional shear
and tensile bonding procedures reveal much larger bonding
areas with diameters between 3 and 5 mm [12,33]. When
dentin is used as a bonding substrate in combination with
recently developed adhesives reaching bond strengths over
15 MPa, these tests often produce non-uniform stresses
resulting in pulling out dentin [37±39]. The observations
of the present study indicate that this problem can be
prevented with microtensile testing because only adhesive
failures were observed for the dentin groups. This con®rms
the ®ndings of other studies dealing with microtensile test-
ing [32±35]. As shown by Shono et al. in recent investiga-
tions using the same methodology used in the present study,
bonding to ¯attened dentin surfaces at variable distances
from the pulp, results in regional differences in dentin
bond strength [40,41]. Therefore, only the very central
areas of the specimens were used to obtain a reliable ran-
domization of test specimens. Compared with the results of
Shono et al. having involved 50 specimens of different
occlusal regions per group, the present standard deviations
are in a comparable range [41]. This consistency demon-
strates the reliability of the modi®ed methodology using 20
central specimens in each group providing equal distances
from the bonded surface to the pulp chamber. According to
 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238 233

Fig. 3. Scanning electron micrographs of interfaces bonded with Syntac Classic (®nal magni®cation: 4000 £ ). (a) as per manufacturer's instructions. The
interface shows the adhesive (A), hybrid layer (H, between arrowheads), resin tags (T) and various lateral branches (asterisks). (b) Specimen etched with
phosphoric acid, primed with Syntac Classic Primer/Adhesive and bonded with Ultraseal XT Plus as adhesive instead of the Heliobond adhesive. Note the gap
(asterisks) between hybrid layer (H) and ¯owable resin composite acting as adhesive resin (A). T: resin tags. (c) Specimen etched with phosphoric acid, primed
with Syntac Classic Primer/Adhesive and bonded with Revolution instead of the Heliobond adhesive. Between the top of the hybrid layer (H) and the ¯owable
resin acting as adhesive (A) a distinct gap is detectable (asterisks). Multiple voids (small asterisks) are visible within the applied ¯owable resin composite
forming the tags (T).

the suggestion of Pashley et al., statistics have been carried
out with n ˆ 20 as number of specimens under investigation
per experimental group [35,35].
Variation of the bonding substrate is less evident in enamel
bonding procedures. Nevertheless, the lingual and buccal
aspects of the teeth were ¯attened in order to achieve similar
bonding areas. Due to the method involving sticks with a
cross-sectional area of 0.5 mm 2, the recorded values for
both enamel and dentin bonding are higher compared with
conventional shear or tensile studies [17,32,33,41]. For the
evaluation of enamel bond strengths, the primers were also
applied to simulate the situation in cavities, when dentin and
enamel are treated simultaneously, but without rubbing,
because detrimental effects of scrubbing motions during appli-
cation on enamel have been reported in the literature [42].
The bond strength investigation was accompanied with
micro-morphological analyses of the resin±dentin interface.
SEM analysis of the interfaces displays the resin-side of the
interface. TEM evaluation of both demineralized and non-
demineralized specimens represents a complement of the
investigation at an ultra-morphological level of the
previously tested interfacial characteristics [2,43].
The bonding systems used different priming philosophies
according to different solvents. OptiBond FL primer
contains water and ethanol, Syntac Classic contains acetone
and water, and EBS Multi contains water as solvent. Syntac
classic was used with phosphoric acid etching despite being
a self-etching primer when used on dentin [44,45]. As repre-
sentative ¯owable resin composites two materials were
selected because of their different viscosities [46] and
234 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238

Fig. 4. Scanning electron micrographs of interfaces bonded with EBS Multi (®nal magni®cation: 4000 £ ). (a) as per manufacturer's instructions. The hybrid
layer (H) is approximately 4 mm thick (arrowheads), the resin tags (T) exhibit various lateral branches (asterisks). Due to the thin layer of adhesive, the resin
composite (RC) has contact to the hybrid layer (H). (b) Specimen etched with phosphoric acid, primed with EBS Multi Primer and bonded with Ultraseal XT
Plus replacing the EBS Bond. Note the short resin tags (T) and the gap formations at the top of the hybrid layer (H, asterisk) and peritubular at the tubule
entrances (asterisks). (c) Specimen etched with phosphoric acid, primed with EBS Multi Primer and bonded with Revolution as Bonding Agent (A). Note the
partially short resin tags (T, lower asterisk) and the gaps (upper asterisk) at the top of the hybrid layer (H, between arrowheads).

similar radiopacity [47]. Therefore, Ultraseal XT Plus (¯ow:
534 mm 2/30 s at 0.5 MPa [46]) was used as thinner material
and Revolution (¯ow: 243 mm 2) was used providing a
higher viscosity although it is equally ®lled (both materials
53% by weight [46]) like Ultraseal XT Plus. This way, the
effect of the consistency was also investigated as there are
characteristic differences among the numerous ¯owable
resin composites on the market.
Another important point when discussing bond strengths
to dentin is the thickness of the bonding agent
[13,17,21,23,26]. The evaluation of the SEM specimens,
however, clearly demonstrated that thicker layers did not
automatically lead to higher microtensile bond strength
results. In contrast, the ¯owable composites revealed the
thickest layers (up to 173 mm due to not possible air-
thinning) and lowest dentin bond strengths. However, the
thicker adhesive layers in the OptiBond FL control groups
may have contributed to slightly higher dentin bond
strengths, but this particular difference was not statistically
signi®cant (P . 0.05; Table 2).
Unterbrink and Liebenberg describe the use of Syntac
Single Component with the ¯owable resin composite Tetric
Flow [31]. However, the adhesive system Syntac Single
Component is a one-bottle adhesive combining features of
primer and bonding resin (which is light-cured), leading to
hybridization itself. To act as a real ®lled adhesive like
OptiBond FL, the dentin should have been acid-etched
and primed with a primer of a two-bottle system (which is
not light-cured). The ¯owable resin applied after priming
would play the role of a bonding resin. However, due to its
higher viscosity compared with un®lled resins, OptiBond
FL adhesive may behave as a ¯owable resin composite.
 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238 235

Fig. 5. Resin replicas of composite bonded with EBS Multi (®nal magni®cation 5000 £). (a) as per manufacturer's instructions. The microporosities and
channels provided by acid etching are poured out. (b) Specimen etched with phosphoric acid, primed with EBS Multi Primer and bonded with Ultraseal XT
Plus replacing the EBS Bond. Also all structures of etched enamel are indirectly visible. (c) Specimen etched with phosphoric acid, primed with EBS Multi
Primer and bonded with Revolution as Bonding Agent. Distinct ®ller accumulation prevents the complete penetration of the ¯owable composite into the acid-
conditioned enamel on the left resulting in short resin tags providing less retetion (asterisks).

Therefore, a ¯owable resin composite may have the capabil-
ity to act in the same way as a ®lled adhesive replacing the
bonding agent of two-bottle systems. The results of the
present study clearly demonstrate that the ¯owable compo-
sites tested are not capable of hybridizing etched and primed
dentin as ef®ciently as the commercially available bonding
agents recommended by the manufacturers.
In contrast, bonding to enamel shows far less complexity-
only the penetration into the porosities of acid-etched enamel
is an prerequisite for successful adhesion [9,10]. The micro-
tensile bond strength results reveal that Ultraseal XT Plus
may act as an enamel bonding agent due to its lower viscos-
ity. The viscosity represented by the material Revolution
seems to be too high for complete penetration into etched
enamel, as demonstrated by the signi®cantly lower bond
strengths. Although there is not much noticeable difference
between the viscosity of the ®lled adhesive OptiBond FL and
the ¯owable composites during application, there seems to
exist a borderline of viscosity avoiding penetration into
etched and primed dentin. This point of view is supported
by the fact that the ®lled one-bottle adhesive OptiBond Solo
(Kerr, Orange, CA, USA) is recently marketed in a lower
®lled formula by the manufacturer. A further explanation
for the differences in bond strength could be the different
chemical composition of the ®lled adhesive vs the ¯owables
marketed primarily as restorative materials.
Penetration of liquids into narrow capillaries, such as the
microporosities of etched enamel, is in¯uenced by proper-
ties of the liquid, such as viscosity and the surface free
energy of the capillary wall [48]. Additionally, viscosity
236 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238

Fig. 6. Transmission electron micrographs of interfaces bonded with OptiBond FL. (a) As per manufacturer's instructions (non-decalci®ed specimen, ®nal
magni®cation: 6300 £ ). H: hybrid layer; Note the complete penetration of the hybrid layer (H) and the `shag carpet'-extension of collagen ®bers into the
adhesive layer (black hand). The small ®llers of the ®lled adhesive penetrate to the entrances of lateral branches (white hand). (b) With Ultraseal XT Plus
replacing the EBS Bond (non-decalci®ed specimen, magni®cation 6,300 £). Note the ®ller particle obstructing the tubule entrance (white hand) and resulting in
short tag formation (black hand). This corresponds to the observations made in Fig. 4b. H: Hybrid layer. (c) With Revolution as bonding agent (decalci®ed
specimen, magni®cation 5000 £ ). Note the ®ller particle clogging the dentinal tubule (arrow) and preventing complete resin tag formation (asterisk) and
hybridization (H) of the demineralized peritubular dentin. The arrow shows a slight `shag carpet' appearance of the top of the hybrid layer resulting from
scrubbing the ¯owable resin composite.

of the restorative resin has been described as a parameter
that in¯uences the penetration of restorative resins into acid-
etched enamel [49].
Depth of resin penetration into enamel decreases only
slightly with increasing viscosity [50]. Several studies
have reported that, when applied to acid-etched enamel,
there are no signi®cant differences between low-viscosity
resins and high-viscosity resins in terms of adaptation and
depth of penetration [50±52]. These studies were performed
by comparing early-generation high-viscosity composite
resins with low-viscosity un®lled resins. In fact, composite
resins seem to adapt themselves to etched enamel surfaces
equally well as do low-viscosity resins [51].
SEM and TEM evaluations corroborate the bond strength
results on a micro-morphological level. The SEM speci-
mens show an acid resistant hybrid layer formation in the
groups with ¯owable resin composites acting as ®lled adhe-
sive resins. The detection of accumulated ®llers as well as
polishing particles under the SEM alone would not primar-
ily be suspicious for inferior bonding to dentin, and incom-
plete resin penetration may not be visible due to the acid
treatment during the specimen production using hydro-
chloric acid and sodium hypochlorite. The distinct gap
formations at the top of the hybrid layer, however, indicate
the inferior dentin bonding perfomance of the experimental
groups (Figs. 2b, 3b,c, 4b,c). The additional TEM analysis
 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238
proved to be an important method to look closer at the
mechanisms of dentin hybridization. These more detailed
views received by TEM clearly indicate an insuf®cient tubu-
lar penetration of the ¯owables applied after etching and
priming (Fig. 6b,c).
Due to the similar results for the differently viscous ¯ow-
able resin composites when used as ®lled adhesives, the
microscopic results con®rm the hypothesis that viscosity
alone may not be the only crucial factor for the inferior
bonding behavior of the ¯owables. The chemical composi-
tion such as the addition of HEMA to un®lled resins (as an
hydrophilic agent) as well as the ®ller size and ®ller distri-
bution in ®lled resins could be another factor for the still
superior bonding performance of the adhesive resins tested
as controls.
The resin replicas of enamel produced for SEM investi-
gation show less pronounced penetration of the ¯owable
Revolution into acid-conditioned enamel when compared
with Ultraseal XT Plus or the bonding resins of the control
groups (Fig. 5). This explains the less predictable results
obtained with the more viscous material Revolution acting
as ®lled adhesive. However, only in the control groups
cohesive failures in resin occurred, while in the groups
bonded with ¯owables adhesive failures were observed
exclusively (Table 3). Nevertheless, bond strength results
clearly indicate that Ultraseal XT Plus bonded equally
compared to the control groups.
5. Conclusions
² The ¯owable composites tested in this study do not ful®ll
the requirements to act as ®lled dentin adhesives.
² Both ¯owable composites were unable to fully hybridize
etched and primed dentin.
² The ¯owable composite Ultraseal XT Plus may be used
as enamel bonding resin without an intermediary bonding
agent.
References
[1] Wilson NHF, Dunne SM, Gainsford ID. Current materials and tech-
niques for direct restorations in posterior teeth. Part 2: Resin compo-
site systems. International Dental Journal 1998;47:185±93.
[2] PerdigaÄo J, Lopes M. Dentin bonding-questions for the new millen-
nium. Journal of Adhesive Dentistry 1999;1:191±209.
[3] Friedl KH, Schmalz G, Hiller KA, Mortazavi F. Marginal adaptation
of composite restorations versus hybrid ionomer/composite sandwich
restorations. Operative Dentistry 1997;22:21±9.
[4] Opdam NJM, Roeters JJM, Peters TCRB, Burgersdijk RCW, Teunis
M. Cavity wall adaptation and voids in adhesive class I resin compo-
site restorations. Dental Materials 1996;12:230±5.
[5] Frankenberger R, KraÈmer N, Pelka M, Petschelt A. Internal adapta-
tion and overhang formation of direct posterior resin composite
restorations. Clinical Oral Investigations 1999;3:208±15.
[6] PerdigaÄo J, Lambrechts P, Van Meerbeek B, Braem M, Yildiz E,
237
Yucel T, Vanherle G. The interaction of adhesive systems with
human dentin. American Journal of Dentistry 1996;9:167±73.
[7] Swift EJ, PerdigaÄo J, Heymann HO. Bonding to enamel and dentin: A
brief history and state of the art, 1995. Quintessence International
1995;26:95±110.
[8] Davidson CL, de Gee AJ, Feilzer AJ. The competition between the
composite dentin bond strength and the polymerization contraction
stress. Journal of Dental Research 1984;63:1396±9.
[9] Buonocore MG. A simple method of increasing the adhesion of
acrylic ®lling materials to enamel surfaces. Journal of Dental
Research 1955;34:849±54.
[10] Gwinnett AJ, Matsui A. A study of enamel adhesives. The physical
relationship between enamel and adhesive. Archives of Oral Biology
1967;12:1615±20.
[11] Eick JD, Cobb CM, Chappel P, Spencer P, Robinson SJ. The dentinal
surface: its in¯uence on dentinal adhesion. Part I. Quintessence Inter-
national 1991;22:967±77.
[12] Eick JD, Gwinnett AJ, Pashley DH, Robinson SJ. Current concepts on
adhesion to dentin. Critical Reviews in Oral Biology and Medicine
1997;8:306±35.
[13] PerdigaÄo J. An ultramorphological study of human dentine exposed to
adhesive systems. PhD thesis, Catholic University of Leuven,
Belgium, 1995.
[14] Van Meerbeek B, PerdigaÄo J, Lambrechts P, Vanherle G. The clinical
performance of adhesives. Journal of Dentistry 1998;26:120.
[15] Nakabayashi N, Ashizawa M, Nakamura M. Identi®cation of a resin-
dentin hybrid layer in vital human dentin created in vivo: durable
bonding to vital dentin. Quintessence International 1992;23:135±41.
[16] Tay FR, Gwinnett JA, Wei SH. Micromorphological spectrum of
acidconditioned dentin following the application of a waterbased
adhesive. Dental Materials 1998;14:329±38.
[17] Pashley DH, Sano H, Ciucchi B, Yoshiyama M, Carvalho RM. Adhe-
sion testing of dentin bonding agents: a review. Dental Materials
1995;11:117±25.
[18] Kanca J. Resin bonding to wet substrate. I. Bonding to dentin. Quin-
tessence International 1992;23:39±41.
[19] PerdigaÄo J, Swift EJ, Denehy GE, Wefel JS, Donly KJ. In vitro bond
strengths and SEM evaluation of dentin bonding systems to different
dentin substrates. Journal of Dental Research 1994;73:44±55.
[20] Nakabayashi N, Kojima K, Masuhara E. The promotion of adhesion
by the in®ltration of monomers into tooth substrates. Journal of
Biomedical Materials Research 1992;16:265±73.
[21] Blunck U, Roulet J-F. Effect of one year water storage on the effec-
tiveness of dentin adhesives in class V composite resin restorations.
Journal of Dental Research 1999;78:444 (abstr. no. 2706).
[22] Van Meerbeek B, Yoshida Y, Snauwaert J, Hellemans L, Lambrechts
P, Vanherle G, Wakasa K, Pashley DH. Hybridization effectiveness of
a twostep versus a threestep smear layer removing adhesive system
examined correlatively by TEM and AFM. Journal of Adhesive
Dentistry 1999;1:7±23.
[23] da Cunha Mello FS, Feilzer AJ, de Gee AJ, Davidson CL. Sealing
ability of eight resin bonding systems in a Class II restoration after
mechanical fatiguing. Dental Materials 1997;13:372±6.
[24] Van Meerbeek B, Willems G, Celis JP, Roos JR, Braem M,
Lambrechts P, Vanherle G. Assessment by nano-indentation of the
hardness and elasticity of the resin±dentin bonding area. Journal of
Dental Research 1993;72:1434±42.
[25] Swift Jr EJ, Triolo Jr PT, Barkmeier WW, Bird JL, Bounds SJ. Effect
of low-viscosity resins on the performance of dental adhesives. Amer-
ican Journal of Dentistry 1996;9:100±4.
[26] Kemp-Scholte CM, Davidson CL. Marginal integrity related to bond
strength and strain capacity of composite resin restorative systems.
Journal of Prosthetic Dentistry 1990;64:658±64.
[27] Opdam NJ, Roeters FJ, Feilzer AJ, Smale I. A radiographic and
scanning electron microscopic study of approximal margins of
Class II resin composite restorations placed in vivo. Journal of Dentis-
try 1998;26:319±27.
238 R. Frankenberger et al. / Dental Materials 18 (2002) 227±238
[28] Ferdianakis K. Microleakage reduction from newer esthetic restora-
tive materials in permanent molars. Journal of Clinical Pediatric
Dentistry 1998;22:221±9.
[29] Behle C. Flowable composites: properties and applications. Practical
Periodontics and Aesthetic Dentistry 1998;10:350±1.
[30] Bouschlicher MR, Cobb DS, Boyer DB. Radiopacity of compomers,
¯owable and conventional resin composites for posterior use. Opera-
tive Dentistry 1999;24:20±5.
[31] Unterbrink G, Liebenberg H. Flowable resin composites as ®lled
adhesives: Literature review and clinical recommendations. Quintes-
sence International 1999;30:249±57.
[32] Sano H, Shono T, Sonoda H, Takatsu T, Chiucchi B, Carvalho RM,
Pashley DH. Relationship between surface area for adhesion and
tensile bond strength. Dental Materials 1994;10:236±40.
[33] Dickens SH, Liao NS, Reed BB. Comparison of a microtensile and
two shear dentin bond tests. Journal of Dental Research 1998;77:657
(abstr. no. 202).
[34] Kanemura N, Sano H, Tagami J. Tensile bond strength to and SEM
evaluation of ground and intact enamel surfaces. Journal of Dentistry
1999;27:523±30.
[35] Pashley DH, Carvalho RM, Sano H, Nakajima M, Yoshiyama M,
Shono Y, Fernandes CA, Tay F. The microtensile bond test. A review.
Journal of Adhesive Dentistry 1999;1:299±309.
[36] Carvalho RM, Santiago SL, Fernandes CAO, Suh BI, Pashley DH.
Effects of prism orientation on tensile bond strength of enamel.
Journal of Adhesive Dentistry 2000;2:251±7.
[37] Versluis A, Tantbirojn D, Douglas WH. Why do shear bond tests pull
out dentin? Journal of Dental Research 1997;76:1298±307.
[38] Van Noort R, Noroozi S, Howard IC, Cardew G. A critique of bond
strength measurements. Journal of Dentistry 1989;17:61±7.
[39] Van Noort R, Cardew GE, Howard IC, Noroozi S. The effect of local
interfacial geometry on the measurements of the tensile bond strength
to dentin. Journal of Dental Research 1991;70:889±93.
[40] Shono Y, Ogawa T, Terashita M, Carvalho RM, Pashley EL, Pashley
DH. Regional measurement of resin±dentin bonding as an array.
Journal of Dental Research 1999;78:679±705.
[41] Shono Y, Terashita M, Shimada J, Kozono Y, Carvalho RM, Russell
CM, Pashley DH. Durability of resin±dentin bonds. Journal of Adhe-
sive Dentistry 1999;1:211±8.
[42] Frankenberger R, KraÈmer N, Petschelt A. Long-term effect of dentin
primers on enamel bond strength and marginal adaptation. Operative
Dentistry 2000;25:11±9.
[43] Van Meerbeek B, Conn LJ, Duke ES, Eick JD, Robinson SJ, Guerrero
D. Correlative transmission electron microscopy examination of
nondemineralized and demineralized resin±dentin interfaces. Journal
of Dental Research 1996;75:879±88.
[44] Gwinnett AJ, Dickerson WG, Yu S. Dentin bond shear strength and
microleakage for Syntac/Heliomolar: a comparison between the
manufacturer's and total etch technique. Journal of Esthetic Dentistry
1992;4:164±8.
[45] Haller B, Trojanski A. Effect of multi-step dentin bonding systems
and resin-modi®ed glass ionomer cement liner on marginal quality of
dentin-bonded resin composite class II restorations. Clinical Oral
Investigations 1998;2:130±6.
[46] Bayne SC, Thompson JY, Swift EJ, Stamatiades P, Wilkerson M. A
characterization of ®rst-generation ¯owable composites. Journal of
the American Dental Association 1998;129:567±77.
[47] Murchison DF, Charlton DG, Moore WS. Comparative radiopacity of
¯owable resin composites. Quintessence Int 1999;30:179±84.
[48] Newman S. Kinetics of wetting of surfaces by polymers: Capillary
¯ow. Journal of Colloidal Interfaces Science 1968;26:209±13.
[49] Asmussen E. Penetration of restorative resins into acid etched enamel.
I. Viscosity, surface tension and contact angle of restorative resin
monomers. Acta Odontologica Scandinavica 1977;35:175±82.
[50] Asmussen E. Penetration of restorative resins into acid etched enamel.
II. Acta Odontologica Scandinavica 1977;35:183±91.
[51] Jùrgensen KD, Shimokobe K. Adaptation of resinous restorative
materials to acid etched enamel surfaces. Scandinavian Journal of
Dental Research 1975;83:31±6.
[52] Pahlavan A, Dennison JB, Charbeneau GT. Penetration of restorative
resins into acid-etched human enamel. Journal of the American
Dental Association 1976;93:1170±6.