J Dent Res 74(10): 1679-1688, October, 1995

Tensile Bond Strength and SEM Evaluation

of Caries-affected Dentin
Using Dentin Adhesives
M. Nakajimal, H. Sanol, M.F. Burrowl, J. Tagamil, M. Yoshiyama2, S. Ebisu2, B. Ciucchi3, C.M. Russell4,
and D.H. Pashley5t
'Department of Operative Dentistry, Tokyo Medical and Dental University, 5-45, 1-chome, Yushima, Bunkyo-ku, Tokyo 113, Japan;
2Department of Conservative Dentistry, Tokushima University, 3-18-15, Kuramoto-cho, Tokushima 770, Japan; 3Department of Restorative
Dentistry and Endodontics, University of Gen6ve, 19, rue Barthelemy-Menn, CH-1211 Geneve 4, Switzerland; 4Division of Biostatistics and
Department of Oral Diagnosis and Patient Services and 5Department of Oral Biology/Physiology,School of Dentistry, Medical College of
Georgia, Augusta, Georgia 30912, USA; *to whom reprint requests should be addressed; tto whom correspondence should be addressed

Abstract. Tensile bond strength measurements are commonly Introduction

used for the evaluation of dentin adhesive systems. Most tests
are performed using extracted non-carious human or bovine
dentin. However, the adhesion of resins to caries-affected
dentin is still unclear. The objectives of this study were to test
the hypothesis that bonding to caries-affected dentin is inferior
to bonding to normal dentin, and that the quality of the hybrid
layer plays a major role in creating good adhesion. We used a
micro-tensile bond strength test to compare test bond
strengths made to either caries-affected dentin or normal
dentin, using three commercial adhesive systems (All Bond 2,
Scotchbond Multi-Purpose, and Clearfil Liner Bond II). For
scanning electron microscopy, the polished interfaces between
the adhesive bond and dentin were subjected to brief exposure
to 10% phosphoric acid solution and 5% sodium hypochlorite,
so that the quality of the hybrid layers could be observed.
Bonding to normal dentin with either All Bond 2 (26.9 8.8
MPa) or Clearfil Liner Bond 11 (29.5 10.9 MPa) showed
tensile bond strengths higher than those to caries-affected
dentin (13.0 3.6 MPa and 14.0 4.3 MPa, respectively). Thle
tensile bond strengths obtained with Scotchbond Multi-
Purpose were similar in normal and caries-affected dentin
(20.3 5.5 MPa and 18.5 4.0 MPa, respectively). The hybrid
layers created by All Bond 2 in normal dentin and by Clearfil
Liner Bond II in normal or caries-affected dentin showed
phosphoric acid and sodium hypochlorite resistance, whereas
the hybrid layers created by All Bond 2 in caries-affected
dentin and those created by Scotchbond Multi-Purpose to
normal and caries-affected dentin showed partial susceptibility
to the acid and sodium hypochlorite treatment. The results
indicate that the strength of adhesion to dentin depends upon
both the adhesive system used and the type of dentin.
Moreover, the quality of the hybrid layer may not always
contribute significantly to tensile bond strength.
Key words: caries, dentin bonding, dentin, resins, sclerotic
dentin.
Received February 20, 1995; Accepted August 11, 1995
Although most bond strength testing is done on normal
dentin for convenience, clinically most bonding substrates
are not normal dentin, but rather are caries-affected dentin
or sclerotic cervical dentin. Unfortunately, little work has
been published on these clinically-relevant substrates.
Harnirattisai et al. (1992, 1993) reported SEM studies on the
interface of resins bonded to either caries-affected dentin or
sclerotic cervical dentin, respectively. They reported that
both bonding substrates exhibited occluded dentinal
tubules which prevented resin penetration, and they found
very shallow resin penetration into cervical root dentin
relative to caries-affected dentin. Van Meerbeek et al. (1994)
reported similar observations of very shallow resin
infiltration into sclerotic cervical root dentin. Neither
group, however, measured bond strengths of resins to
these substrates.
Several studies have recently attempted to simulate the
adhesive properties of sclerotic or caries-affected dentin in
vitro (Nakajima, 1991; Perdigao et al., 1994; Ehudin and
Thompson, 1994). The bond strengths to these modified
dentin substrates were very low (i.e., 2 or 3 MPa) and are
difficult to interpret in view of the few reports available on
bond strength measurements made to natural caries-
affected dentin (Fusayama et al., 1979). The limited size and
shape of caries-affected dentin creates technical difficulties
when investigators attempt to measure resin composite
bond strengths using conventional testing methods. Since
the sizes and shapes of these clinically-relevant lesions are
often irregular, there is the danger of including normal
dentin in the bonded surface. Recently, Sano et al. (1994)
have introduced a micro-method to evaluate bond strengths
in specimens with small (ca. 1 mm2) surface areas. Using
this method, they found that it is possible to assess the
adhesion strength of resin composites to caries-affected vs.
normal dentin.
The purposes of this study were to test the hypotheses
that bonding to caries-affected dentin will yield bond

1679
 1680 Nakalima et al.
composite resin
canagentn
rn
cc,ot ,
normal dentin
itar corious dentin
(stain.bic by CDS)
inier ea, 'us denfin
(no sta nlable by CDS)
1 mlnl
0 I
ca, es-aflecled
dentin/nes n
slices
Micro-tensile test
Figure 1. Schematic showing how carious teeth were prepared for
bonding and how they were prepared for microtensile testing. CDS =
caries detector stained material. Note that caries-affected and normal
dentin sites can be tested in the same tooth. Outer carious dentin is
sometimes referred to as infected carious dentin. Inner carious dentin
is sometimes referred to as caries-affected dentin and includes the
transparent layer.
strengths that are lower than bonds made to normal dentin
and that the quality of the hybrid layer is correlated with
bond strength.
Materials and methods
Specimen preparation
The teeth used in this study were erupted carious third molars
without opposing occlusion. They were obtained with a
protocol that was reviewed and approved by an appropriate
institutional review board with the informed consent of the
donors.
Forty-seven extracted human molars with coronal carious
lesions, stored at 4C in isotonic saline containing 0.2".. sodium
azide, were used in this study. The inclusion criteria were that
the caries be limited to the occlusal surface, that it extend at
least half the distance from the enamel-dentin junction to the
pulp chamber, and that there be enough surrounding normal
dentin to serve as a control bonding site (although the latter
criterion was not always possible). If, after being bonded and
sectioned, the carious lesion was found to be too shallow or
too deep, the tooth was excluded from the study. We used
#320 silicon carbide paper under running water to grind the
occlusal surface perpendicular to the long axis of the carious
lesion to expose and flatten the lesion, after grinding off the
coronal enamel with a model trimmer. To obtain caries-
affected dentin, we performed the grinding using the
combined criteria of visual examination, surface hardness
using a dental explorer, and staining by a caries detector
solution (CDS, Kuraray Co., Ltd., Osaka, Japan). Using this
procedure, we removed all soft, stainable, carious dentin.
What was left was the relatively hard, caries-affected non-
staining dentin on the experimental side and normal dentin at
] Dciit Rcs 7400) 1995
the same level on the control side of the same tooth (Fig. 1).
Bonding procedure
After the flat dentin surface was polished with #600 silicon
carbide paper, each tooth was randomly assigned to one of three
commercial adhesive systems: All Bond 2 (Bisco Inic., Itasca, IL),
Scotchbond Multi-Purpose (3M Dental Plroducts, St. Paul, MN),
or Clearfil Liner Bond 11 System (Kuraray Co., Ltd., Osaka,
Japan). The components used for bonding are shown in Table 1.
These adhesive systems were applied to the dentin surface only.
v For the All Bond 2 adhesive system, the dentin was kept moist
after being acid-etched and rinsed as recommended by the
manufacturer. Although moist dentin is reported to be an
applicable substrate for All Bond 2 (Gwinnett, 1992; Kanca,
1992), for Scotchbond Multi-PuLpose, we used the "dry
technique" according to the manufacturer's instructions. The
norrmal
dent n/resir Clearfil Liner Bond Il system requires no water rinse after dentin
sI ces
conditioning/priming with its self-etching primer. After each
adhesive resin was light-cured according to the manlufactLrer's
instructions, a resin composite crown (Fig. 1) was built up
incrementally in four layers with Silux Plus (3M Dental
Products, St. Paul, MN) resin composite to a height of 3 to 5 mm
(Sano et al., 1994). Each increment was cured for 20 sec. The
bonded specimens were then stored in water at 37C for 24 hrs.
Bond strength testing
From 4 to 6 vertical slices, approximately 1 mm thick, were
made through the caries-affected and normal portions of each
tooth perpendicular to the bonded surface (Fig. 1). These slices
were then trimmed and shaped, by meaiis of super-fine
diamond burs (c-16 ff, GC Ltd., Tokyo, Japan), into a gentle
curve along the adhesive interface from both sides, to form a
square cross-section (approximately 1 mm wide at its
narrowest). At the time of trimming, selection of the testing
region, such as caries-affected or normal dentin, was carefully
performed by both visual and light microscopic observations.
The sclerotic dentin that makes up much of the caries-affected
dentin is transparent in transmitted light and light-brown in
reflected light. Normal dentin is opaque in transmitted light and
white in reflected light. However, since this was done visually
and at low magnification (from 10 to 20x), it was not always
possible to select regions that were 100", caries-affected or
normal dentin. Thus, we re-evaluated all debonded specimens
by scanning electron microscopy to confirm the type of bonded
dentin before the samples were included in the final data sets
for statistical analyses. Bonded areas that showed mixtures of
normal and sclerotic dentin were excluded from further
consideration. Not all teeth yielded control and experimental
sections, so the data could not be analyzed as paired data. Mean
values were calculated for each caries-affected or normal dentin
area. The values listed in Table 2 are the means of these
individual mean values. We calculated the surface area for
bonding before testing by measuring the width and thickness of
the narrowest portion with digital calipers and a light
microscope. Since the bond strength of some bonding systems
has been reported to vary directly with dentin thickness (Suzuki
and Finger, 1986), the dentin thickness was measured between
the bonded surface and the nearest pulp horn or region of the
pulp chamber by means of a digital micrometer. The amount of
dentin remaining between the bonded dentin surface and the
j Dent Res 74(10) 1995 Bonding to Caries-affected Dentin 1681

Table 1. Adhesive systems used for bonding

System Ingredients Code/Lot# Proceduresa

All Bond 2: Etchant 10% phosphoric acid semigel 129083 a (15 s), b (15 s)
Primer A NTG-GMAb 019104 d, e, f, g
Primer B BPDM 129093 h (20 s)
Adhesive Bis-GMA 019134
UDMA
HEMA

Scotchbond Multi-Purpose:
Etchant maleic acid 4EB a (15 s), b (15 s)
Primer HEMA 4DK c (10 s), f, g
polyalkenoate copolymer h (20 s)
Adhesive Bis-GMA 4DB
HEMA

Clearfil Liner Bond 2:

Primer A Phenyl-P 006 e, f (30 s), g
5-NMSA h (20 s)
CQ
Primer B HEMA 007
water
Clearfil LB Bond MDP 0009
HEMA
Bis-GMA
microfiller
a
Procedures: a = acid-etching, b rinse, c = air-dry, d blot-dry, e = mix primer, f = apply primer, g = apply= adhesive, and h = light-cure.
=
= 2-
b Abbreviations: Bis-GMA = bisphenyl-glycidyl-methacrylate; BPDM = biphenyl-dimethacrylate; CQ camphorquinone; HEMA acid;
hydroxyethyl-methacrylate; MDP = 10-methacryloyloxydecyl-dihydrogen phosphate; NMSA = N-methacryloyl-5-aminosalicylic
NTG-GMA = N-tolylglycine-glycidyl-methacrylate; Phenyl-P = 2-methacryloyloxyethyl-phenyl hydrogen phosphate; and UDMA =
urethane dimethacrylate.

closest pulp horn of each specimen was defined as the
remaining dentin thickness. These specimens were then
attached to a Bencor-Multi-T testing apparatus (Danville
Engineering Co., Danville, CA) with a cyanoacrylate adhesive
(Zapit, DVA, Anaheim, CA) which, in turn, was placed in a
universal testing machine (AG 500 B, Shimadzu Co., Kyoto,
Japan) for tensile testing at a cross-head speed of 1 mm/min
(Sano et al., 1994).
After specimens were tested, the fracture modes of each
specimen were determined by means of a low-power (8x) light
microscope. The modes of fracture were designated as adhesive
if 100% of the bonded interface failed between the dentin and
the bonding resin, cohesive if 100% of the failure was in the
resin composite or dentin, or mixed if the failures were partially
adhesive and partially cohesive.
Micro-hardness measurement and SEM observation
of resin/dentin interface
After the tensile testing procedure, each fractured specimen was
fixed in 10% neutral buffered formalin (Fusayama et al., 1966;
Sano, 1987), then embedded in epoxy resin (Epon 812, Nisshin
EM Co., Ltd Tokyo, Japan) for the measurement of subsurface
micro-hardness. Each embedded specimen was polished with
diamond pastes down to a particle size of 1 pm (Struers,
Copenhagen, Denmark). Knoop hardness was measured on the
dentin side of specimens 50 pm below the bonded surface, by
means of a micro-hardness tester (Akashi MVK-E hardness
tester, Akashi Co., Tokyo, Japan) under a load of 50 g and
duration of 15 s (Perinka et al., 1992). A mean Knoop Hardness
Number for each specimen was calculated as an average of
three Knoop Hardness Numbers obtained from each specimen.
Generally, hardness measurements were obtained from both
caries-affected dentin and normal dentin from the same teeth at
the same depth or remaining dentin thickness.
After the dentin hardness was measured, specimens were
lightly polished with diamond pastes (1 ,um, Struers,
Copenhagen, Denmark), followed by overnight desiccation at
room temperature. Specimens were gold-sputter-coated and
observed with an SEM (JXA-840, JEOL, Tokyo, Japan) for
assessment of the morphology of the caries-affected or normal
dentin that was used for microhardness testing. At this time,
those specimens exhibiting a combination of normal and caries-
affected dentin beneath the bonded area were excluded. Due to
such exclusions, the specimen numbers for each group were
reduced to ten. Scanning electron microscopy was used to
characterize both the bonding substrates and the interface
between the bonded resins and the normal or caries-affected
dentin. Since most of the modes of bond failure showed both
adhesive and cohesive failure in resin composite, there was
usually some resin composite on the debonded surface which
facilitated polishing of the bonded interface. We used both
debonded and non-stressed specimens for SEM observations to
1682 Nakajima et al. J Dent Res 74(10) 1995
Table 2. Tensile bond strength, remaining dentin thickness, and Knoop Hardness of bond groups
Tensile Bond Strength Remaining Dentin Knoop Hardness
(MPa)a (mm)a (kg mm-2)a
Normal Dentin
All Bond 26.90 ( 8.83)b 2.09 (0.56) 62.48 (3.57)
Scotchbond 20.32 (5.55) 2.28 (0.43) 59.87 (6.46)
Clearfil 29.52 (10.90) 2.17 (0.49) 62.70 (3.15)

Sclerotic Dentin
All Bond 13.01 (3.64)1 1.91 (0.65)1 33.58 (17.22)1
Scotchbond 18.49 (4.04) 2.51 (0.50) 22.17 (8.33)
Clearfil 13.97 (4.30)1 1.72 (0.67)1 29.05 (19.10)11
a Values are
X (SD); N = 10 in each group.
b Groups connected by vertical lines in the same plane are not
significantly different (p > 0.05). Groups not connected by lines in the same
plane are significantly different (p < 0.05).

determine if there were any differences in appearance. Selected strength. The two-way ANOVA was significant for both
specimens from each substrate and bonding material group (at dentin substrate differences and dentin-material interaction.
least three specimens from each group) were subjected to 10% This phenomenon may be seen in Table 2, which shows that,
phosphoric acid treatment for 3 to 5 sec (Gwinnett and Kanca, for All Bond 2 and Clearfil Liner Bond II, the tensile bond
1992; Sano et al., 1995), followed by 5% sodium hypochlorite strength for normal dentin was about twice that for caries-
immersion for 5 min (Wang and Nakabayashi, 1991). After affected dentin (All Bond 2 - 26.90 vs. 13.01 MPa; Clearfil
being extensively rinsed in water, the treated specimens were Liner Bond II - 29.52 vs. 13.97 MPa). Scotchbond Multi-
air-dried, gold-sputter-coated, and observed with the SEM. Purpose had tensile bond strength values which were not
much different for normal and caries-affected dentin (20.32
Statistics vs. 18.49 MPa). This makes the rank order of increasing
The data were collected in a two-way analysis of variance tensile bond strength means different for materials in the
(ANOVA) design, with bonding material and type of dentin two dentin substrate groups: caries-affected - All Bond 2,
substrate as the two factors. In addition, the interaction of the Clearfil Liner Bond II, Scotchbond Multi-Purpose vs. normal
factors was tested. Response variables were tensile bond dentin - Scotchbond Multi-Purpose, All Bond 2, Clearfil
strength, Knoop hardness, and remaining dentin thickness. The Liner Bond II (Table 4). In addition, Table 4 gives the least-
two-way ANOVA for tensile bond strength was also done, with squares means comparisons by vertical line designations.
values adjusted for remaining dentin thickness as a covariate. Among the materials applied to normal dentin, there are no
Group comparisons were done by the least-squares means significant differences. Among the materials applied to
method. Least-squares means are the expected value of group caries-affected dentin, Scotchbond Multi-Purpose and All
or subgroup means that one expects for a balanced design Bond 2 are significantly different. Since the dentin may have
involving the group variable with all covariates at their mean different properties based on depth (inversely proportional
value. Statistical significance was set in advance at the 0.05 to remaining dentin thickness), we adjusted the tensile bond
level. All analyses were conducted with SAS software for the
personal computer (SAS Institute, Cary, NC).
Results
Table 3. Analysis of variance of the results
The means and standard deviations of the groups are given
in Table 2 for the variables tensile bond strength, remaining p-values
dentin thickness, and Knoop hardness values. Each of these Tensile Bond Remaining Dentin Knoop
response variables was examined in a two-way ANOVA, the Strength Thickness Hardness
results of which are given in Table 3. The effect of the dentin
substrate was the only factor that correlated with the Two-way
differences in the means of KHN. For each material, the ANOVA
magnitude of hardness values for normal dentin was Material 0.5239 0.0496 0.1548
approximately twice that of caries-affected dentin. There Dentin 0.0001 0.4255 0.0001
was no difference in the hardness values for specimens that Material * Dentin 0.0041 0.2228 0.4881
were debonded compared with those that were not stressed
to failure. The remaining dentin thickness was different Two-way (adjusted)
among the three materials tested; that of the Scotchbond RDT 0.4803
Multi-Purpose caries-affected dentin was higher than that of Material 0.4480
the other groups. Dentin 0.0001
The main response variable of interest was tensile bond Material * Dentin 0.0081
j Dent Res 74(10) 1995 Bonding to Caries-affected Dentin
strength data for the remaining dentin thickness and
repeated the two-way ANOVA. These results are also
shown in Table 4. In this case, both the normal and caries-
affected dentin groups for Scotchbond Multi-Purpose are
significantly different from the caries-affected dentin groups
of bond strengths produced by All Bond 2 and Clearfil Liner
Bond II. The order of the adjusted means did not differ from
that of the original means. Thus, the overall inferences are
approximately the same.
Fractography of debonded specimens following
microtensile testing failed to show exclusive adhesive failure
in many cases (Table 5). A high percentage of specimens
exhibited mixed failure, i.e., adhesive failure between the
dentin and bonding resin plus cohesive failure within the
composite.
Scanning electron microscopy
There were no consistent differences between the SEM
appearance of the bonded interfaces of specimens which had
not been stressed and that of those that had been debonded.
When caries-affected dentin was bonded with All Bond 2,
SEM examination of the polished cross-sections of the
interface (Fig. 2a) showed a resin-infiltrated demineralized
zone approximately 4 to 5 pm thick between the overlying
adhesive and the underlying demineralized dentin. The
orifices of the dentinal tubules were funnel-shaped due to
the dissolution of peritubular dentin in the top 3 to 4 pm of
the mineralized dentin zone. Although only two resin tags
are seen in Fig. 2a, the scalloped border between the
mineralized dentin and the resin-infiltrated dentin identifies
adjacent resin tags that were beyond the plane of the section
and indicates that this specimen was made from relatively
deep dentin. The resin-infiltrated dentin or hybrid layer
(Nakabayashi, 1985) appeared to be well-filled with resin. In
Fig. 2a, a vertical structure can be seen in the middle tubule,
passing from the mineralized dentin through the hybrid
layer to a point near the adhesive/hybrid-layer junction.
To determine the resistance of these structures to serial
acid/base treatments, we first treated the polished caries-
affected specimens with 10% phosphoric acid (5 sec), then
Table 4. Rank order of tensile bond strength means before and after
adjustment for remaining dentin thickness
Ranking of Tensile Bond Strength (MPa)
Unadjusted Adjusted
All Bond 2-
13.01 13.25
Caries-affected
Clearfil Liner Bond II-
Caries-affected 13.97 14.58
Scotchbond Multi-Purpose
Caries-affected 18.49 18.04
Scotchbond Multi-Purpose
20.32 20.19
Normal dentin
All Bond 2-
Normal dentin 26.90 26.58
Clearfil Liner Bond 2-
Normal dentin 29.52 29.09
1683
rinsed and immersed them in 5% NaOCl for 5 min. After
being rinsed, air-dried, and sputter- coated, the specimens
were viewed under SEM. Scanning electron microscopy of
sections of bonded caries-affected dentin treated in this
manner revealed a relatively thick hybrid layer (5 to 6 pm)
which appeared to be well-filled by acid/base-resistant resin
in the top and bottom regions. The middle region of the
hybrid layer was somewhat susceptible to the effects of
acid/base treatments. In some specimens, treatment with
acid/base removed the entire middle third of the hybrid
layer in caries-affected dentin (not shown). This
susceptibility of the middle of the hybrid layer was not seen
in normal dentin bonded with All Bond 2. At the junction
between the bottom of the hybrid layer and the underlying
dentin in caries-affected specimens, the resin seemed to
blend in with the mineralized matrix (Fig. 2b). The top of the
hybrid layer had a rougher texture than the bulk of the
resin-infiltrated zone. Two tube-like structures were seen
extending from the funnel-shaped resin tags to the top of the
hybrid layer. There were few lateral extensions of resin tags
into the lateral branches of the tubules (Fig. 2b) of caries-
affected dentin.
Examination of normal dentin treated with the All Bond
2 system followed by acid/base etching revealed three
distinct layers (Fig. 2c): The dark surface layer, from 5 to 7
pm thick, is the adhesive layer. Beneath it is a hybrid layer
from 3 to 4 pm thick, containing a granular-textured
intertubular region penetrated every 5 to 10 pm by
smoother-textured extensions of the adhesive layer to form
the tops of resin tags. Most of the resin tags in this section
appear to have been removed by the polishing step. The
hybrid layer created in normal dentin by the All Bond 2
system appears to be thinner (compare Figs. 2a and 2c) than
the hybrid layer formed in caries-affected dentin. Below the
remaining hybrid layer, funnel-shaped resin tags can be
Table 5. The distributions of fractures that occurred in the bonded
specimens.
Cohesive Failurec
Adhesivea Mixedb In Dentin In Composite
All Bond 2
normal dentin 1 9 0 0
0 0
caries-affected 4 6
Scotchbond
Multi-Purpose
8 2 0 0
normal dentin
5 4 0 1
caries-affected
Clearfil
Linerbond II
0 0
normal dentin 7 3
5 5 0 0
caries-affected
a
Adhesive failures = failures between resin and dentin.
b Mixed failures = failures that were partially adhesive and partially
cohesive.
c Cohesive failures = failures that occurred 100% within the resin
composite or 100% in dentin.
1684 Nakajima et al.
Figure 2. Scanning electron micrographs of specimens bonded with
All Bond 2. (A) Caries-affected dentin sectioned through the
bonded interface and polished. R = resin adhesive, H = hybrid
layer, and D mineralized dentin. Funnel-shaped resin tags are
seen, one of which contains a non-resin, presumably mineralized
core. Hybrid layer (dark zone) is about 4 to 5 pm thick. (B) Caries-
affected dentini prepared as in A but then subjected to sequential
acid and base treatment to challenge the hybrid layer. Note that the
middle of the hybrid layer (H) was partially removed by acid/base
challenge. (C) Normal dentin. The hybrid layer (H) is 3 to 4 pm
thick and was resistant to acid/base challenige. R = resin, D
mineralized dentin.
seen extending into the mineralized dentin. Many resin tags
exhibited lateral extensions in dentinal tubules, seen in this
section as small holes in and around the tubule lumina.
Most of the hybrid layer and the resin tags in normal dentin
were resistant to the effects of acid/base etching. There were
no tube-like projections through the hybrid layer of normal
dentin bonded with All Bond 2.
Treatment of excavated carious dentin with the
Scotchbond Multi-Purpose bonding system revealed a
relatively thin hybrid layer (3 pm thick) with a funneled
tubule orifice containing a rod-like structure (Fig. 3a) which
projected from the peritubular dentin through the hybrid
layer and extended slightly into the overlying adhesive
layer. The junction between the top of the hybrid layer and
the bottom of the adhesive layer is indicated by a line of
white arrows because it was somewhat indistinct on some
prints. The hybrid layer appeared to be better infiltrated by
resin in the top two-thirds than in the bottom third, which
had a rougher texture (Fig. 3a). The rod-like structure
appeared to occlude the tubule lumen completely and
seemed contiguous with peritubular dentin.
When the polished cross-sections of caries-affected
dentin treated with Scotchbond Multi-Purpose were etched
with acid and basic solutions, the hybrid layer appeared to
be 3.5 to 4.5 pm thick, with very short (1 to 2 pm thick) resin
tags (Fig. 3b). No resin could be seen in the lateral branches
of the dentinal tubules.
Scanning electron microscopic observation of polished
cross-sections of normal dentin bonded with Scotchbond
Multi-Purpose and etched with acid/base treatment
revealed numerous, long, well-formed resin tags (not
J Dciit Rcs 7400) 1995
'-AN'
shown). The hybrid layer appeared to be thinnller (2 to 2.8
pm) than that seen in caries-affected dentin. The top third of
the hybrid layer seemed to resist acid/base attack better
than did the bottom two-thirds. The peritubular dentini was
removed by the acid/base treatment and revealed
numerous small lateral extensionis of "micro-tags' branchilng
off at right angles from the main resini tag. Most of these
lateral extensions appeared to have been broken during
sample preparation (not shown).
When excavated carious dentin was treated with Clearfil
Liner Bond 11, the hybrid layer that was produced was
between 3 and 5 pIM thick (Fig. 4a). The dentinial tubules
were funneled at the orifice to the mineralized dentin, and
each tubule contained tube-like extensions of material that
passed up from the tubule lumen into the top of the hybrid
layer. The high number of tubules per field at 5000x
magnification and their relatively large diameter identify the
bonded region as middle to deep dentini (Garberoglio and
Brannstrbm, 1976).
When similar sections of caries-affected dentin bonded
with Clearfil Liner Bond 11 were etched with acid followed
by sodium hypochlorite, numerous resin tags were seen
extending into the dentinl (Fig. 4b). Most of the tags were
smooth and were about I to 2 pm in diameter. There were
few lateral branches to the tubules of caries-affected dentin.
j Dciit Res 7400) 1995 Bontdinig to Caries-affected Dentin11 685

> 9 m
,~.., q
f:!, ,
.`
.:
..
4,_. .7_1.;I;,, ( "'_ ,7,'
I' 11-
- _.

LI
Figure 3. Scanning electron micrographs of specimens bonded with Scotclbonid Multi-Pui pose system. (A) Caries-affected dentiii sectioned
throLugh the bonded interface and polished. A single tubule is shown occupied by a mineralized cast which protrudes throuigh the hvbl-id
layer (H), which was about 3 pm thick. The junction between the top of the hybrid layer and the bottom of the adhesive layer (R) is indicated
by a line of white arrows. (B) Caries-affected dentin prepared as in A but then subjected to sequentiaL aciddbase challenige. Fewv resiln tags are
seen. There is a space between the bottom of the hybrid (H) layer and the mineralized dentin (D) whicih was created by the acid/base
treatment. The hybrid layer is 3 to 4 pm thick.

The hybrid layer was about 3 pm thick.
When normal dentin was bonded with Clearfil Liner
Bond II and the sections were exposed to phosphoric acid
followed by sodium hypochlorite, the hybrid layer was
about 2.5 pm thick (Fig. 4c), and the tubules were about 1
pm in diameter and regularly shaped. There were many
small protrusions from the sides of the resin tags, which
indicate that resin attempted to fill the many lateral canals
that were evident in the dentini matrix (Fig. 4c).
Discussion
The formation of the hybrid layer is due to the application of
acids or self-etching acidic primers to the dentin, followed
by adhesive resin penetration into the decalcified zone
(Nakabayashi, 1985; Wang and Nakabayashi, 1991; Inokoshi
et al., 1993). The complete penetration of resin monomers
into the demineralized dentin is essential to create strong
adhesion as well as a perfect seal of the enveloped collagen
fibers (Sano et al., 1995). For normal dentin, many reports are
available with regard to both the tensile properties and
interfacial morphology between the resin and the hybrid
layer. Little high-resolution SEM has been reported on1 the
interfacial morphology of the julnctioni between adhesive
resin and caries-affected dentin and/or sclerotic dentin.
Several studies have investigated the morphology of
sclerotic dentin (Gwinnett and Kanca, 1992; Harnirattisai et
al., 1992, 1993; Van Meerbeek et al., 1994). Harnirattisai et al.
(1992) showed morphological variations in the resin-
impregnated layer that were dependent on whether the
dentinal tubules were occluded or opened. The mineral
phase of carious dentin is remodeled by repeated sequences
of demineralization and remineralization, which usually
produce occlusion of the tubules with mineral crystals
(Fusayama et al., 1966).
The All Bond 2 all-etch technique, combined with moist
dentin bonding, has been reported to create a strong bond to
normal dentini, creating a relatively thick (4 to 5 pim) hybrid
layer (Tay et al., 1994). When 10%, phosplhoric acid is applied
to normal dentin, the smear layer and superficial deentini are
demineralized, and collagen fibers of the superficially
demineralized dentin are exposed. The acetone solvent of
the All Bond 2 primer is believed to act as a dehydratitng
agent, making the demineralized dentin less hydrophilic
and carrying resinous components of the primer into the
exposed collagen fibers. As a result, the application of the
All Bond 2 system to normal dentin creates a stronig
micromechanical interlocking between the resini and
superficially demineralized dentin (Kanca, 1992). The acid
and sodium hypochlorite resistance of the All Bond 2 hybr-id
layer created on normal dentini (Fig. 2c) suggests that the
adhesive monomers or primers were able to envelope the
exposed collagen fibers. However, Sano et al. (1995) repor-ted
the lateral diffusion of silver nitrate througlh the All Bond 2
hybrid layer. Nanometer-sized diffusioll channiiels or defects
in this layer probably exist due to the failure of the resin to
envelope all of the collagen fibers in the demineralized,
primed dentin completely and uniformly. These subsuLrface
porosities may have been obscured durinig specimen
polishing in the study.
Resin infiltration by All Bond 2 reagents into excavated
caries-affected dentin may be less complete, eveen thouglh the
hybrid layer created by All Bond 2 was deeper thani that
found in normal dentin (compare Figs. 2b and 2c). However,
portions of the hybrid layer of the caries-affected
intertubular dentin treated with All Bond 2 were susceptible
to the acid and base treatment (Fig. 2b), showing dissolution
within its central region. Caries-affected dentin may contain
substances that interfere with free radical generationi or
propagation, leading to poorer polymerization of resins in
such dentin. The peritubular dentin matrix of caries-affected
dentin has been reported by Shimizu et al. (1981) to take ulp
much more toluidine blue stain and to exhibit more intenise
1686 Nakajimna et a]. j Deitt Rcs 74(10) 1995

Figure 4. Scanning electron micrographs of specimens bonded with

Clearfil Liner Bond 11. (A) Caries-affected dentin sectioned through
the bonded surface and polished. The hybrid layer (H) is between 3
and 5 pm thick. Mineralized casts extended into the hybrid layer
from the peritubular dentin matrix of the mineralized dentin (D). R
resin adhesive layer. (B) Caries-affected specimen treated as in A
and then subjected to acid/base challenge. The hybrid layer (H) is
about 3 pm thick and shows numerous long smooth resin tags
measuring I to 2 pm in diameter. There is a large space between the
hybrid layer (H) and the mineralized dentin (D). There are few
lateral branches to the resin tags and few lateral branches to the
dentinal tubules seen in caries-affected dentin. R = adhesive resin
layer. (C) Normal dentin specimen bonded with Clearfil Liner Bond
11. The hybrid layer (H) is about 2.5 pm thick and exhibits long resin
tags with multiple extensions. The mineralized dentin (D) is very
porous where the acid/base treatment enlarged the sizes of lateral
branches to the dentinal tubules. R = adhesive layer. There is a large
space between the bottom of the resin and the mineralized dentin
created by the acid/base challenge.

metachromasia than normal peritubular dentin, suggesting
the presence of mucopolysaccharides or glycoproteins.
These molecules may interfere with resin wetting of fine
porosities within both intertubular and peritubular dentin
and/or may interfere with conversion of adhesive
monomers to polymeric networks.
Although the adhesive resin may have followed the
primer, it may have not copolymerized well with the
primer. Thus, the adhesion of resins to caries-affected dentin
may be inferior to that of normal dentin, due to weaker
collagen and/or weaker resin. Further studies of the sealing
qualities of the hybrid layer of caries-affected dentin by All
Bond 2 are indicated.
For the Scotchbond Multi-Purpose adhesive system, the
remaining dentin thickness (RDT) was higher for both
normal and caries-affected dentin (Table 2) than for the
other two materials. This means that the Scotchbond Multi-
Purpose bonded specimens were bonded to dentin that was
further from the pulp chamber than that in the other two
groups. Generally, this should have produced higher bond
strengths. While this was true for the caries-affected dentin,
it was not true for normal dentin, which gave somewhat
lower bond strengths with Scotchbond Multi-Purpose
relative to the other two products. If the remaining dentin
thickness does have a physical relationship to dentin
quality, the interaction of material and dentin substrate may
be caused by differences in remaining dentin thickn-ess
(Table 3) and not by materials differences. This problem
deserves further study.
There were no significant differences in the
morphological appearance of the adhesive interface (Figs.
3a,b,c) bonded to either caries-affected or normal dentin.
The top half of the hybrid layer in both types of dentin
showed acid and base resistance in both substrates. The
susceptibility of the bottom half of the hybrid layer to
acid/base treatment (Figs. 3b,c) correlates well with a recent
study by Sano et al. (1995b). According to their report, silver
nitrate tracer accumulated mainly at the bottom half of the
hybrid layer, which might not have contained a sufficient
amount of adhesive resin. Presumably, the collagen fibers of
the bottom half of the hybrid layer of this system may not be
fully enveloped by primer or adhesive resin, leaving spaces
accessible to the acid/base attack. The similarity of tensile
bond strengths of both the caries-affected and the normal
dentin groups (Tables 2,4) indicates that the quality of resin
infiltration into the top of the hybrid layer may be more
important than the depth of resin penetration. It may be
possible that the photo-initiator system that has been
included in the Scotchbond Multi-Purpose system produced
slightly higher degrees of polymerization and cross-linking
I Dent Res 7400) 1995 Bonding to Caries-affected Dentin
of resins in caries-affected dentin compared with the other
bonding systems. Further studies are needed to evaluate the
extent of polymerization and cross-linking of adhesive
resins in caries-affected dentin.
Smear layer removal, or its modification, by Clearfil
Liner Bond II is performed by use of a mixture of Phenyl-P
and HEMA. This mixture has been shown to be effective in
obtaining good bonding to normal ground dentin (Chigira et
al., 1994; Watanabe et al., 1994). The thickness of the hybrid
layer was approximately 2.5 pm in normal dentin (Fig. 4c)
and slightly thicker in caries-affected dentin (Fig. 4b). The
hybrid layer of normal dentin created by Clearfil Liner Bond
II was largely resistant to the effects of acid/base treatment.
The Clearfil Liner Bond II system bonded to normal dentin
showed the greatest tensile bond strength among the groups
tested in the present study (Table 2), although, due to the
large standard deviations, the difference was not significant.
Burrow et al. (1993, 1994), using conventional tensile
bond testing, showed that the failure mode was material-
dependent. During testing for bond strength, shear bond
strength measurements have been reported to show greater
abnormalities in stress distribution at the adhesive interface
than tensile bond strength measurement (Van Noort et al.,
1989, 1991). The testing method that we used in this study
may create a more uniform stress distribution compared
with the conventional methods of either shear or tensile
testing, which utilize much greater surface areas for testing.
Since the bonded surface areas that were tested were small
(ca. 1 mm2), they may contain fewer defects that can serve as
focal points of stress. Only one case of complete cohesive
failure was observed within the resin composite. There were
no instances of cohesive failure in dentin, even when the
relatively soft caries-affected dentin was used as a substrate,
indicating that this test method may produce better stress
distribution during testing. Currey and Brear (1990)
developed an empirical equation that related the hardness of
mineralized tissues to their yield stress. Using the lowest
Knoop hardness value in Table 2 (caries-affected dentin
bonded with Scotchbond Multi- Purpose = 22.17 kg mm-2),
their equation would predict a yield strength of 44.2 MPa.
The measured bond strength to that same substrate was 19
MPa, a value much lower than its predicted intrinsic
strength. Thus, it is not surprising that even the softest
dentin did not fail cohesively. It is thought that the reason
for the frequent occurrence of mixed failures was the wide
variations in elastic modulus at the adhesive interface,
which might cause stress concentrations at the top of the
adhesive layer. There is no information available for the
elastic modulus of caries-affected dentin or that of hybrid
layers formed in caries-affected dentin. Further study is
needed to clarify the mechanical properties of caries-affected
dentin and its hybrid layers.
Great care was taken in this study to characterize the
bonded substrate. In addition to the use of caries detector
dye and an explorer to estimate the hardness of excavated
carious dentin, Knoop hardness measurements were made,
and the dentin was examined by SEM. The hardness values
that were obtained in our normal dentin (from 60 to 63 kg
mm 2, Table 2) were similar to those reported by others
(Craig et al., 1959; Fusayama et al., 1966; Pashley et al., 1985).
1687
The low hardness values for caries-affected dentin (from 22
to 34 KHN, Table 2) are similar to that reported by
Fusayama et al. (1966) for transparent dentin which exists
beneath carious lesions. This softness is due to the loss of
apatite mineral from intertubular dentin (Shimizu et al.,
1981), even though most of the tubules in such dentin are
filled with mineral deposits. These intratubular crystals are
not well-packed and are softer than well-packed apatite,
even though they are more acid-resistant (Ogawa et al.,
1983). Thus, the lower bond strengths of adhesive resins to
caries-affected dentin could be due to either weakened resin
polymers or weakened collagen fibers. Future research
should be directed at increasing both the strength of
collagen and the strength of the bonding resins in these
softer substrates.
The only report in the literature that we could find on the
bond strength of adhesive resins to caries-affected dentin
was that of Fusayama et al (1979). Using Clearfil Bond
System-F (a Phenyl-P based system), they obtained bond
strengths of 1.7 MPa to caries-affected dentin. Since that
time, the formulations of adhesive systems have
dramatically improved, as have bond strengths. The recent
attempt to simulate caries-affected dentin for bonding
studies reported very low (ca. 1 to 3 MPa) shear bond
strengths of All Bond 2 and Scotchbond Multi Purpose
(Perdigao et al., 1994). These are far below the bond
strengths that were obtained (ca. 13 to 18 MPa) in the present
study, indicating that their procedure may not provide
appropriate simulations of caries-affected dentin. The
Knoop hardness of caries-affected dentin has been reported
to be lower than normal (Fusayama et al., 1966; Ogawa et al.,
1983). Thus, even though the dentinal tubules of the
substrates are occluded with mineral, the intertubular
dentin is hypomineralized (Ogawa et al., 1983) and exhibits
lower than normal Knoop hardness values (Table 2).
Apparently, the acidic conditioners in the bonding
systems that were tested were able to demineralize caries-
affected dentin at least as much as, if not more than, normal
dentin. Since the adhesive resins infiltrated the
demineralized caries-affected dentin more fully, they
produced thicker hybrid layers. Even though the hybrid
layers of caries-affected dentin were thicker than those of
normal dentin, their tensile bond strengths were lower than
those of normal dentin, suggesting that there is little
correlation between hybrid layer thickness and bond
strength of adhesive resins. Finger et al. (1994) found no
correlation between hybrid layer thickness and bond
strength. The results indicate that there may be a
relationship between the quality of the hybrid layer and
bond strength, although there are no objective measures of
hybrid layer quality. Little is known about the surface
energy of normal demineralized dentin collagen or of the
collagen of caries-affected dentin. It is theoretically possible
for adhesive resin monomers to penetrate further into less
mineralized, more porous demineralized dentin but not wet
the collagen fibers very well. Thus, even though the hybrid
layers of caries-affected dentin appear by SEM to be thicker,
their quality must be evaluated by TEM studies. Another
variable is the degree of polymerization of the primers and
adhesive resins that infiltrate the demineralized caries-
1688 Nakajima et al.
affected dentin.
The results of this study support the hypothesis that
bonding to caries-affected dentin with All Bond 2 and
Clearfil Liner Bond II produces bond strengths lower than
these systems achieve to normal dentin. An exception to that
general finding was that bonds made to caries-affected
dentin with Scotchbond Multi-Purpose were not
significantly different from those made to normal dentin,
although the latter were somewhat lower (but not
significantly) than bonds made to normal dentin with All
Bond 2 or Clearfil Liner Bond II.
The development of new adhesive systems must aim to
create strong adhesion to all types of tooth substrates,
particularly dentin affected by caries. In addition, these
systems must also aim to create a perfect seal at the adhesive
interface and within the hybrid layer to prevent leakage,
which may lead to degradation of the bond, secondary
caries, and pulpal irritation.
Acknowledgments
The authors are grateful to Shirley Johnston for outstanding
secretarial support. This work was supported, in part, by
grant DE 06427 from the National Institute of Dental
Research and by the Medical College of Georgia
Biocompatibility Program.
References
Burrow MF, Tagami J, Hosoda H (1993). The long term
durability of bond strengths to dentin. Bull Tokyo Med Dent
Univ 40:173-191.
Burrow MF, Tagami J, Negishi T, Nikaido T, Hosoda H (1994).
Early tensile bond strengths of several enamel and dentin
bonding systems. J Dent Res 73:522-528.
Chigira H, Yukitani W, Hasegawa T, Manabe A, Itoh K,
Hayakawa T, et al. (1994). Self-etching dentin primers
containing Phenyl-P. J Dent Res 73:1088-1095.
Craig RG, Gehring PE, Peyton FA (1959). Relation of structure to
the microhardness of human dentin. J Dent Res 38:624-630.
Currey JD, Brear K (1990). Young's modulus and yield strength
in mammalian mineralized tissues. J Mater Sci Mater Med
1:14-20.
Ehudin DZ, Thompson VP (1994). Tensile bond strength of
dental adhesives bonded to simulated caries-exposed
dentin. J Prosthet Dent 71:165-173.
Finger WJ, Inoue M, Asmussen E (1994). Effect of wetability of
adhesive resins on bonding to dentin. Am J Dent 7:35-38.
Fusayama T, Okuse K, Hosoda H (1966). Relationship between
hardness, discoloration, and microbial invasion in carious
dentin. J Dent Res 45:1033-1046.
Fusayama T, Nakamura M, Kurosaki N, Iwaku M (1979). Non-
pressure adhesion of a new adhesive restorative resin. J
Dent Res 58:1364-1370.
Garberoglio R, Brannstrom M (1976). Scanning electron
microscopic investigation of human dentinal tubules. Arch
Oral Biol 21:355-362.
Gwinnett AJ (1992). Moist versus dry dentin: Its effect on shear
bond strength. Am J Dent 5:127-129.
Gwinnett AJ, Kanca J (1992). Interfacial morphology of resin
composite and shiny erosion lesions. Am J Dent 5:315-317.
Harnirattisai C, Inokoshi S, Hosoda H, Shimada Y (1992).
j Dent Res 74(10) 1995
Interfacial morphology of an adhesive composite resin and
etched caries-affected dentin. Oper Dent 17:222-228.
Harnirattisai C, Inokoshi S, Shimada Y, Hosoda H (1993).
Adhesive interface between resin and dentin of cervical
erosion/abrasion lesions. Oper Dent 18:138-143.
Inokoshi S, Hosoda H, Harnirattisai C, Shimada Y (1993).
Interfacial structure between dentin and dentin bonding
systems revealed using argon ion beam etching. Oper Dent
18:8-16.
Kanca J (1992). Resin bonding to wet substrate. I. Bonding to
dentin. Quintessence Int 23:39-41.
Nakabayashi N (1985). Bonding of restorative materials to
dentin: The present status in Japan. Int Dent J 35:145-154.
Nakajima M (1991). Effects of CO2 laser irradiation on
microstructure of dentin, dentin adhesion of resin
composites and dental pulp. J Stomatol Soc Jpn 58:263-293.
Ogawa K, Yamashita Y, Ichijo T, Fusayama T (1983). The
ultrastructure and hardness of the transparent layer of
human carious dentin. J Dent Res 62:7-10.
Pashley DH, Okabe A, Parham P (1985). The relationship
between dentin microhardness and tubule density.
Endodont Dent Traumatol 1:176-179.
Perdigao J, Swift EJ, Denehy GE, Wefel JS, Donly KJ (1994). In
vitro bond strengths and SEM evaluation of dentin bonding
systems to different dentin substrates. J Dent Res 73:44-55.
Perinka L, Sano H, Hosoda H (1992). Dentin thickness, hardness
and Ca-concentration versus bond strength of dentin
adhesives. Dent Mater 8:229-233.
Sano H (1987). Relationship between caries detector staining
and structural characteristics of carious dentin. J Stomatol
Soc Jpn 54:241-270.
Sano H, Sonoda H, Shono T, Takatsu T, Ciucchi B, Carvalho R,
et al. (1994). Relationship between surface area for adhesion
and tensile bond strength. Evaluation of a micro tensile
bond test. Dent Mater 10:236-240.
Sano H, Takatsu T, Ciucchi B, Horner JA, Matthews WG,
Pashley DH (1995). Nano-leakage: Leakage within the
hybrid layer. Oper Dent 20:18-25.
Shimizu C, Yamashita Y, Ichijo T, Fusayama T (1981). Carious
change of dentin observed on long span ultrathin sections.
J Dent Res 60:1826-1831.
Suzuki T, Finger WJ (1986). Dentin adhesives: site of dentin vs
bonding of composite resins. Dent Mater 4:379-383.
Tay FR, Gwinnett AJ, Pang KM, Wei SHY (1994). Structural
evidence of a sealed interface with a total etch wet-bonding
technique in vivo. J Dent Res 73:629-636.
Van Meerbeek B, Braem M, Lambrechts P, Vanherle G (1994).
Morphological characterization of the interface between
resin and sclerotic dentine. J Dent 22:141-146.
Van Noort R, Noroozi S, Howard IC, Cardew GE (1989). A
critique of bond strength measurements. J Dent 17:61-67.
Van Noort R, Cardew GE, Howard IC, Noroozi S (1991). The
effect of local interfacial geometry on the measurement of
the tensile bond strength to dentin. J Dent Res 70:889-893.
Wang T, Nakabayashi N (1991). Effect of 2-(methacryloxy)ethyl
phenyl hydrogen phosphate on adhesion to dentin. J Dent
Res 70:59-66.
Watanabe I, Nakabayashi N, Pashley DH (1994). Bonding to
ground dentin by a Phenyl-P self-etching primer. J Dent Res
73:1212-1220.