journal of the mechanical behavior of biomedical materials 37 (2014) 100 108

Available online at www.sciencedirect.com

www.elsevier.com/locate/jmbbm

Research Paper

Strengthening of dental adhesives

via particle reinforcement

Renan Bellia, Stefan Kreppelb, Anselm Petschelta, Helga Hornbergerb,

Aldo R. Boccaccinib, Ulrich Lohbauera,n
a
Laboratory for Biomaterials Research, Dental Clinic 1 Operative Dentistry and Periodontology,
Friedrich-Alexander-University of Erlangen-Nuremberg, 91054 Erlangen, Germany
b
Institute of Biomaterials, Department of Materials Science and Engineering, Friedrich-Alexander-University
of Erlangen-Nuremberg, 91058 Erlangen, Germany

ar t ic l e in f o abs tra ct

Article history: The bond between methacrylic polymer adhesives and dental restoratives is not perfect and
Received 12 February 2014 may fail either in the short or in the long term. This study aims to evaluate the effects of
Received in revised form particle incorporation in a self-etch model adhesive on mechanical and physical properties that
24 April 2014 are relevant during application and service. Filled adhesives containing 5, 10, 15 or 25 wt% glass
Accepted 3 May 2014 llers were compared to their unlled counterpart in terms of water sorption and solubility;
Available online 11 May 2014 viscosity and dynamic viscosity during polymerization were recorded using rheological

Keywords: measurements and compared to FTIR analysis of the real-time degree of cure. Elastic modulus

Nanoparticles and ultimate tensile strength measurements were performed in uniaxial tension; the energy to

Bond strength fracture was used to calculate the fracture toughness of the adhesives. Finally, the experimental

Elasticity adhesives were applied on dentin substrate to test the bond strength using the microtensile

Adhesives test. Results showed that the incorporation of 510 wt% nanoller to self-etching dental

Dentin adhesives is efcient in accelerating the polymerization reaction and increasing the degree of
cure without compromising the lm viscosity for good wettability or water sorption and
solubility. Fillers increased the elastic modulus, tensile strength and fracture toughness to a
plateau between 5 and 15 wt% ller concentration, and despite the tendency to form
agglomerations, active crack pinning/deection toughening mechanisms have been observed.
The bond strength between resin composite and dentin was also improved when adhesives
with up to 10 wt% llers were used, with no additional improvements with further packing. The
use of llers to reinforce dental adhesives may therefore be of great practical benet by
improving curing and mechanical properties.
& 2014 Elsevier Ltd. All rights reserved.

n
Corresponding author. Dental Clinic 1 Operative Dentistry and Periodontology, Glueckstr. 11 D-91054 Erlangen, Germany.
Tel.: 49 9131 854 3740; fax: 49 9131 853 4207.
E-mail address: lohbauer@dent.uni-erlangen.de (U. Lohbauer).

http://dx.doi.org/10.1016/j.jmbbm.2014.05.007
1751-6161/& 2014 Elsevier Ltd. All rights reserved.
 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108
1. Introduction
Tailoring of mechanical and physical properties of materials
through changes in formulations is a process based on com-
promises between desirable and undesirable effects. For dental
adhesives that join restoratives to tooth tissues, mechanical
stiffness and stability are the goals in order to endure shrinkage
stresses and cyclic mastication forces (Braga et al., 2005;
Frankenberger et al., 2005). The initial state of this bond has
shown to reach satisfactory levels for currently marketed
adhesives (Van Meerbeek et al., 2010). Its degradation, however,
is rather expedited by the susceptibility to water diffusion from
the underlying wet substrate and the environment, and is
mainly dictated by the hydrophilicity of the adhesive
composition (Tjderhane et al., 2013). Prevailing strategies to
preserve the bond integrity focus on minimizing the polymer
water uptake by using more hydrophobic compositions (Tay
et al., 2007; Tjderhane et al., 2013). This should be achievable
using monomer solutions that maintain a uidic character
for proper substrate wettability and still reach adequate
degrees of cure.
Improvements in bond strength can also be accomplished
by reinforcing the adhesive polymer with llers. In a previous
report (Lohbauer et al., 2010a) we showed that uncoated
spherical zirconia (ZrO2) nanoparticles can be added to either
components of a particular 3-step dental adhesive as an
attractive strategy to improve the dentin-composite bond.
That study corroborated previous attempts to reinforce the
bond between resin-based restoratives and tooth tissues
through ller incorporation into dental adhesive agents
(Miyazaki et al., 1995; Kim et al., 2005b; Sadat-Shojai et al.,
2010). Meanwhile nano- and micro-sized particles (mainly
colloidal silica or glass llers) compose many marketed
dental adhesives in concentrations between 5 and 48 wt%
(Van Landuyt et al., 2007).
Consequences of incorporating llers into dental adhesives
go beyond mechanical effects, and may further alter physical
aspects that have a potential clinical relevance during applica-
tion and service. One of these parameters is the lm viscosity,
which is increased by ller incorporation with the purpose of
thickening the adhesive lm and preventing full-thickness
oxygen inhibition (Frankenberger et al., 2002). Overdoing in ller
loading may increase the adhesive viscosity to a point that resin
inltration into the demineralized substrate is compromised.
That represents a bigger issue for etch-and rinse adhesives,
since the self-etching formulations promote resin inltration at
the pace of etching (Van Meerbeek et al., 2011). In increasing the
viscosity, llers may further affect the polymerization kinetics
of the resin by acting as rigid centers for chain initiation during
polymerization, with expected accompanying effects on the
degree of cure.
The present study aims to clarify these questions. Here we
address the effects of incorporating silanized sub-micrometer
glass llers into an experimental self-etch adhesive formulation
in terms of viscosity, dynamic viscosity during curing, degree of
conversion, water sorption/solubility, elastic modulus develop-
ment, ultimate tensile strength, fracture toughness and bond
strength to dentin. The null hypothesis tested was that ller
concentration has no effect in the aforementioned parameters.
101
2. Materials and methods
2.1. Experimental adhesives
Due to the uncertainty surrounding the precise chemical com-
position and monomer concentrations of marketed dental adhe-
sives, an experimental self-etching adhesive was produced
specically for this study (Voco, Cuxhaven, Germany). The
following monomers/additives and corresponding weight con-
centrations were stirred until homogeneously mixed and stored
in a PE beaker: Bis-2-methacryloyoxy)ethylphosphate (Bis-MEP;
33.3 wt%), Bisphenol A diglycidyl methacrylate (Bis-GMA; 33.3 wt
%), 2-hydroxyethyl methacrylate (HEMA; 32.5 wt%), Camphoro-
quinone (0.3 wt%), Ethyl-4-dimethyllaminobenzoate (DABE;
0.45 wt%) and Butylhydroxytoluene (BHT; 0.01 wt%). From this
composition, four lled versions (5, 10, 15 and 20 wt%, or 2.5, 5,
7.5 and 10 vol%, respectively) were produced by incorporation of
SiO2BaO glass llers (NanoFines NF180 of glass GM27884,
Schott, Landshut, Germany) with grain size of d50/d99 180/
500 nm and irregular shape. Silanization of llers was under-
taken by the company delivering the llers (Schott) by using
13 wt% relative to silica of the organosilane -methacryloxypro-
pyltrimethoxysilane as a coupling agent. Further specications
of the used particles can be found in the product information
sheet (Schott, 2010). Dispersion of ller in the resin was achieved
in three steps via kneading, rolling and ultrasonication.
2.2. Water sorption and solubility
Five disc-shaped specimens ( 15 mm; thickness 1 mm) were
produced according to ISO 4049 for each adhesive. The adhesives
were poured into a silicon mold, covered with a glass cover slip,
clamped and light-activated for 5 min using a light-curing oven
(Unilix AC, Heraeus Kulzer, Wehrheim, Germany). The speci-
mens were dried in a desiccator at 37 1C for 22 h and transferred
to another desiccator at 23 1C for 2 h. The specimens were
subsequently weighted using a high accuracy balance (YDK01
model, Sartorius, Goettingen, Germany). After a stable mass
reading was obtained, this value was taken as the initial mass
(minitial). The dimensions of the specimens were measured by
taking 2 points for the diameter and 5 points for the height to
calculate the volume V of the specimens. The specimens were
then put in containers with distilled water and stored for 7 days
at 37 1C. It was made ascertained that the minimum distance
between the specimens was Z3 mm and at least 10 ml water
was supplied per specimen. After the storage period the speci-
mens were rinsed, gently air-dried and weighted within 1 min.
The weight measured is referred as mwet. To determine the
solubility in water (representing the amount of diluted mono-
mer), the rst step needed to be repeated with the wet samples,
obtaining the weight mdry. For the calculation of water sorption
and solubility, the following equations were used:
mwet minitial
Wsorption 1
V
minitial mdry
Wsolubility 2
V
102 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108
2.3. Viscosity and polymerization kinetics
The viscosities of the unlled and lled adhesives were
measured using a rheometer (Physica MCR301, AntonPaar,
Ostldern, Germany) and processed with the software Rheo-
plus 3.40 (AntonPaar). The plate-on-plate testing set-up used
a oscillating cylinder ( 28 mm) at 2 Hz over an aluminum
plate at a distance of 1 mm. The lower plate was cooled to
23 1C by a Peltier water cooling system (HWE, Huber, Offen-
burg, Germany) and the tests were conducted in light-
protected environment.
To evaluate the dynamic viscosity of the adhesives during
polymerization, an integrated-controlled custom LED light-
source was built in the lower chamber of the rheometer to
provide polymerizable energy 1 mm beneath the glass plate
(1 mm-thick) over which the adhesives were placed. The light
energy provided by the light source was measured using a
Ulbricht integrating sphere (Gigahertz-Optik, Tuerkenfeld,
Germany) to 500 mW/cm2. The dynamic viscosity during
light-activation was measured using the same above descri-
bed set-up and measuring parameters.
2.4. Real-time degree of conversion
In order to evaluate the effect of ller concentration on the
conversion rate, the degrees of conversion (DC) of the unlled
and lled adhesives were measured after different light-
activation times by Fourier transform infrared reectance
spectroscopy (FTIR). A FTIR spectrometer (Nicolet 6700,
Thermo Scientic, Waltham, USA), operated under 0.5 cm  1
resolution and 64 scan coadditions, was used to capture the
ongoing curing process using the two frequency technique,
where the net peak absorbance areas of the aliphatic CQC
stretching vibrations at 1638 cm  1 and the reference aromatic
C    C stretching vibrations at 1608 cm  1 were recorded.
Continuous readings were made every 10 s over 50 s of light-
activation (Elipar TriLight, 3MESPE, output intensity: 850 mW/
cm2) on 1-mm-thick samples. The baseline method was
employed for the peak area ratios, where the quotient of areas
of the aromatic C    C peak with the methacrylic CQC peak
was plotted against curing time.
2.5. Elastic modulus, tensile strength and fracture
toughness measurements
Glass-hour-shaped specimens were produced by pouring the
adhesives into a custom built silicon mold. A thin plastic foil
covered the adhesive and glass plates were xed with clamps
on both sides to retain constant specimen dimensions
(3871 mm3). Curing was performed in a light-curing oven
(Unilux AC) for 5 min. Specimens were stored in distilled
water at 37 1C for 24 h. Specimens were tested in uniaxial
tension using a low-compliance universal testing machine
(Z2.5, Zwick, Ulm, Germany) until fracture. Stressstrain
curves were used to derive the Elastic modulus (E) from the
linear elastic range. Loads at fracture and specimens cross-
section areas were used to calculate the ultimate tensile
strength.
At the moment of fracture, the total external work applied
to the specimen can be retrieved from the area under the
stressstrain curve (Wf), to obtain the work-of-fracture f Wf/
2S, where S is the cross-sectional area of each respective
specimen. For materials with negligible dissipation of energy
in the form of plastic deformation (which can be assessed
from the stressstrain curve linearity), the work-of-fracture
equals the critical energy release rate Gc, which can be used
to calculate the fracture toughness, or critical stress intensity
factor KIc from the relationship proposed by Irwin (1964):
 
Gc E 1=2
KIc 3
1 2
where E is the elastic modulus measured for each adhesive
loading level, and the Poisson's ratio of the adhesives (all
assumed to be 0.35 in this study) (Greaves et al., 2011).
2.6. Microtensile bond strength to dentin
Bovine incisors were stored in 0.5% Chloramine-T solution at
4 1C and used within 3 months of extraction. From each tooth
the buccal enamel was grinded at until dentin exposure,
which was further wet grinded for 30 s with a 600-grit SiC
paper to create a clinically relevant smear layer.
For the application as a dentin adhesive system, the
formulated adhesive blends were diluted in a mixture of
water and acetone (50 vol% of adhesive, 25 vol% water and
25 vol% acetone) to be applied as a primer solution. The
primer was applied over the dry substrate in two consecutive
coats 20 s apart and dried with oil-free air-spray for 20 s. Non-
diluted solutions of the monomer formulations were applied
on top of the primer as an adhesive, following the so-called
two-step self-etch approach. After application, the adhesive
was polymerized for 40 s (Elipar Trilight) with an output
intensity of 850 mW/cm2. A resin composite (Grandio Flow,
Voco) was then placed incrementally in four 1-mm horizontal
increments; each individually cured for 40 s with the same
light-curing unit.
For the microtensile bond strength (TBS) testing, each
tooth was longitudinally sectioned in both x and y direc-
tions across the bonded interface with a low-speed diamond
saw under substantial water lubrication. This procedure
resulted in specimens with cross-sectional area of 0.957
0.05 mm2. After cutting, the specimens were stored for 7 days
in distilled water at 37 1C before testing. The specimens were
then xed to microtensile bond strength testing jigs with
cyanoacrylate glue and loaded in tension using a universal
testing machine (Z2.5, Zwick) with a 100 N load cell traveling
at a cross-head speed of 0.5 mm/min. After beam fracture,
the dimensions of the cross-sectional area were measured
using a digital caliper and used to calculate the TBS by
dividing the imposed force (in N) at the time of fracture by the
bonded area (in mm2). The failure modes were evaluated at
20  in a light microscope (SV-6, Zeiss, Oberkochen, Ger-
many) and classied as cohesive (failure entirely within
dentin substrate or resin composite), mixed (failure mainly
in the adhesive resin with excursion to the dentin/adhesive
interface) or adhesive (failure only at the dentin/adhesive
interface). Only specimens presenting adhesive or mixed
failure modes were included in the calculation of mTBS
means, and a minimum of n 30 was assured for the
calculation.
 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108 103

Values of water sorption, monomer solubility, elastic The FTIR spectrum of the unlled and the 20 wt% lled
modulus, tensile strength, fracture toughness and microten- adhesive are shown in Fig. 3 for the different curing times. The
sile strength were evaluated for normality and statistically decrease in resin content with increase in ller incorporation
treated using a one-way Analysis of Variance and LSD post- was also conrmed from the progressive decline of the peaks at
hoc test at pr0.05 (SPSS 21.0 software, IBM, Chicago, IL, USA). 1455 cm  1 and 1510 cm  1, related to CH2 and CQO, respec-
tively (not shown).
From the SEM images of the fractured specimens, a homo-
geneous ller distribution was observed for all lled adhesives
3. Results despite their tendency to form agglomerations. Maximum
cluster sizes were found to reach up to 10, 20, 50 and 80 mm
In Table 1 the values for viscosity, water sorption, monomer
solubility and for the mechanical testing are summarized. An
increasing trend of water uptake was observed for adhesives
with a ller concentration higher than 5 wt%, although no
signicant statistical differences between them and the
unlled adhesive could be observed. When the water sorption
is normalized by the monomer fraction in weight, the values
for the 15 and 20 wt% lled adhesives become statistically
higher than the unlled resin (Fig. 1). Monomer solubility
values showed a different trends, with the unlled adhesive
showing signicantly higher values in comparison to adhe-
sives with 10 and 15 wt% ller concentration. However when
the obtained values are again normalized by the monomer
fraction, no statistically signicant differences were found
between unlled and 5, 15 or 20 wt% lled adhesives.
The viscosity of the different adhesives increased with
ller concentration, which also signicantly inuenced their
dynamic viscosity during polymerization. Fig. 2a illustrates
the changes in viscosity of the unlled and lled adhesives
during polymerization. The artifacts in the graph are the
result of the spontaneous peeling of some samples after full
cure was achieved. The peak value represents the full state of
hardening, from which the distance to the time of light
activation may be taken as the curing time. While the unlled
resin took 10 s to cure, the adhesive containing 20 wt% llers
is fully cured after 2 s of light activation. The addition of 5 wt
% ller reduced the curing time to 4 s. The reduction in
polymerization time with increasing ller content could also
be observed from the values obtained from the real-time
degree of conversion (Fig. 2b). The unlled adhesive showed
to polymerize linearly with curing time (i.e., energy density
provided) until no signicant changes in conversion were
detected, whereas lled adhesives showed an exponential
behavior. After 20 s of light-activation, the unlled adhesive Fig. 1 Values for the sorption and solubility measured for
showed approximately 50% of its maximum conversion; at the unlled and lled adhesives. Absolute values and values
this stage all lled adhesives had achieved their nal normalized by the wt% of resin phase are plotted side
curing state. by side.

Table 1 Mean values (S.D.) for viscosity, water sorption, water solubility, elastic modulus (E), ultimate tensile strength
(UTS), fracture toughness (KIc) and microtensile bond strength (TBS) measurements for the tested adhesives.

Viscosity (Pa s) Water sorption Water solubility Emod (GPa) UTS (MPa) KIc (MPa m0.5) lTBS (MPa)
(lg/mm3) (lg/mm3)

Unlled 1 20.5 (5.1) ab 28.7 (3.7) a 0.57 (0.09) c 1.23 (0.21) c 0.74 (0.13) c 18.4 (8.5) c
5 wt% 1.4 15.5 (3.8) b 27.3 (7.1) ad 1.24 (0.27) b 2.66 (0.72) a 1.18 (0.26) a 26.9 (9.4) b
10 wt% 3.4 21.9 (5.1) ab 19.5 (1.4) bc 0.97 (0.20) b 1.93 (0.50) b 0.89 (0.22) bc 39.7 (10.5) a
15 wt% 7.3 24.4 (5.5) a 22.2 (1.8) bd 1.58 (0.42) a 2.33 (0.97) ab 1.03 (0.41) ab 39.1 (14.6) a
20 wt% 20.3 26.1 (6.3) a 23.8 (1.6) acd 1.58 (0.33) a 2.01 (0.68) ab 0.95 (0.24) ab 37.9 (16.2) a

Different letters in columns designate groups with different statistical subsets at 0.05.
104 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108

Fig. 2 (a) Viscosity curves during polymerization of the unlled and lled adhesives. (b) Conversion vs. curing time measured
by FTIR.

Fig. 3 FTIR spectra for every measuring time interval of (a) the unlled adhesive and (b) the 20 wt% lled adhesive.

Fig. 4 SEM images in backscattered mode of the fracture surfaces of the 10 wt% (a and b) and 20 wt% (c) lled adhesives. In
(a) unidirectional microcracks around one hemisphere of a particle agglomeration is pointed by the arrow. In (b) and (c) the
fracture surfaces show increased roughness due to crack deection around particles, which fractured partally or completely.
 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108 105

Fig. 5 Mean values (error bars are standard deviations) of the mechanical properties of the adhesives and microtensile bond
strength to dentin. The same letters designate the same statistical subsets.

for the 5, 10, 15 and 20 wt% ller concentration, respectively. At
preferential hemispheres of agglomerations of diameters larger
than 10 mm, unidirectional microcracks were observed occa-
sionally (Fig. 4a). In general, fracture surfaces became rougher
with the increase in ller content, and crack deection fracto-
graphic markings were visible around particles and agglomera-
tions, revealing the crack trajectory (Fig. 5b and c). Agglome-
rations involved in the crack path were observed to be partially
or completely fractured.
A summary of the results from the mechanical tests are
illustrated in Fig. 5. The ultimate tensile strength of the
adhesives markedly increased with the addition of llers,
but not further beyond a 5 wt% ller level, which in fact lead
to a slight decrease in strength. A 5 wt% incorporation of
llers signicantly increased the elastic modulus of the
adhesive to approximately 2 times the values of the neat
resin, reaching a 3-fold increase when 15 wt% llers were
added. Elastic modulus increase found a plateau at 15 wt%
ller level, with a signicant decrease when the ller con-
centration was further increased to 20 wt% (Table 1). Fracture
toughness results followed the trend observed for the elastic
modulus and tensile strength, once these properties shape the
fracture behavior of brittle materials. The mTBS to dentin showed
a signicant increase for ller levels Z5 wt% (Table 1). mTBS
values for the adhesives containing 10, 15 and 20 wt% were not
statistically different (Fig. 5a). A trend of increasing cohesive
failure of either substrate was observed with the increase in ller
incorporation (not shown).
4. Discussion
In the present study the incorporation of glass llers to a self-
etch adhesive has shown to produce signicant effects in all
tested properties, leading us to reject the null hypothesis.
The testing of dental adhesives under pure tension is
perhaps of higher practical relevance given the high shrinkage
stress undergone by the adhesive interface during polymeriza-
tion. To that adds the higher susceptibility of materials show-
ing brittle-like fracture response to tensile stresses. Packing of
a more compliant material, such as methacrylate-based poly-
mers for dental use, with stiffer llers leads clearly to an
increase in the elastic modulus of the composite. Locally, a
well-bonded particle constrains the surrounding matrix
against deformation. Under uniaxial loading, closely packed
particles act by hindering the transverse strain of the matrix in
the interstitial spaces. The global effect is the decrease of the
Poisson0 s ratio of the composite as the ller loading increases
(Chabrier et al., 1999; Chung et al., 2004; Masouras et al., 2008).
Following the trend of the elastic modulus, further packing
106 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108
also increases the bulk modulus, rendering dental composites
also more resistant to shear stresses (Masouras et al., 2008).
Bond strength increases as a consequence of ller addition
was expected from past results. Particles in adhesives may
increase the debonding energy in different ways. In a favor-
able scenario where the dentinadhesive interface is strong,
aws responsible for failure may rather origin in the adhesive
resin, with the crack propagating therein. Consequently, as it
occurs in resin composite restoratives, particles provide
obstacles at the crack front, consuming energy during crack
extension (Lohbauer et al., 2013). In a low stiffness Bis-GMA/
TEGDMA-based polymer, the high modulus of a particle (e.g.,
glass) has the potential to attract an offset traveling crack,
causing a temporary reduction in the energy release rate due
to changes in crack trajectory (Wang et al., 2008). In compo-
sites composed of 40120 nm spherical silica particles or
0.4 mm irregular glass particles, crack deection was found
to be the main mechanisms of polymer toughening, with the
particlematrix interface playing an important role (Chan
et al., 2007). In (Lohbauer et al., 2010a), complementary
particle-induced toughening mechanisms have been detected
contributing to the increase in debonding energy.
Particle clustering is an important obstacle when adding
llers to adhesives, since particles aggregate before incor-
poration or in suspension. By increasing particle concentra-
tion in a medium the distances between particles are
reduced, and inter-particle interactions become more likely.
Secondary cluster formation as response to increases in ller
loading was related to bond strength deterioration in several
studies of the kind (Kim et al., 2005b; Sadat-Shojai et al., 2010;
Chen et al., 2011; Solhi et al., 2011). Spherical particles are
advantageous in this respect since few contacts can be
established to neighboring particles, an less energy is needed
to break those interactions. Geometries that increase the
potential area for contact with other particles will be there-
fore more prone to form agglomerations. When using rod-
shaped hydroxyapatite nanollers we also observed severe
particle agglomeration already at low ller concentrations,
which prevented meaningful improvements in bond strength
(Wagner et al., 2013). In that study, parabolic markings (Kies
gures) (Kies et al., 1950) associated with micrometer-size
agglomerations pointed to independent crack initiation sites
in front of the running crack, indicating that particle agglom-
erations may potentially act as stress concentrating defect
sites. There, agglomerations reached a critical size (450 mm)
beyond 5 wt% of loading level, which led to a signicant
decrease in bond strength. Particle cluster dimensions up to
80 mm were found in the present study for the 20 wt% lled
adhesive. In the availability of KIc values, the critical aw size
(ac) for unstable fracture initiation can be calculated using the
basic fracture mechanics relation of Irwin (Irwin, 1964)
KIc Ysfac, where Y is the crack geometric factor ( 1.3 for
elongated agglomerations) and sf the ultimate tensile
strength. For the 20 wt% lled adhesive aco10 mm, showing
that large agglomerations although not readily benecial,
apparently did not act as weakening sites. Flaws matching
the theoretical ac dimensions were occasionally found around
particle clusters 415 mm, as shown in Fig. 4a. Their morphol-
ogy, namely the nonorthogonal orientation of the cracks in
relation to the surface of the agglomeration, but rather a
collective unidirectional elongation, is suggestive of cohesive
cracking of the polymer matrix resulting from a tensile stress
zone created during copolymerization of resin monomers
with the silanized llers. These cracks may be a repercussion
of inhomogeneities in the local onset of the curing reaction,
leading to stress competition zones around an agglomeration
axis. The increase in size of agglomerations may aggravate
this effect in more highly loaded systems. At an interface
under stress this is critical, and may explain the increase in
bond strength scatter for the adhesives having ller loading
levels beyond 10 wt%.
Comparatively, KIc values of pure Bis-GMA and TEGDMA
polymers were measured to 1.45 and 1.97 MPam (Cook and
Moopnar, 1990), respectively, and 0.22 MPam for a blend of
37.5 wt% Bis-GMA, 37.5 wt% Bis-EMA and 25 wt% TEGDMA
(Chan et al., 2007). Considering the low cross-linking potential
of our monomer mixture, a nearly 60% increase in fracture
toughness values took place from 0 to 5 wt% ller loading.
Higher mTBS means for lled adhesives may reect the
observed enhancements in UTS and KIc, but the relation
ship between these properties is not at all straightforward.
Differences in specimen dimensions create clear volume
dependent aw population size and distribution effects.
Additionally, lm thickness, geometry and elastic moduli
ratio at the interface, all affect the stress distribution during
microtensile testing, as highlighted by Soderholm (2010). The
contribution of llers to higher bond strengths to dentin may
be rather associated with the increase in interface stiffness,
as found by Tam and Pilliar (2000). From a subsequent work it
could also be shown the important role of adhesive packing
in increasing the fracture toughness of the adhesivedentin
interface (Tam et al., 2001). The change in fracture mode
patterns observed herein also point to a crack trajectory
moving away from the adhesive joint constituents toward
the adhesivedentin interface as a result of increased resis-
tance to fracture propagation.
The tendency to form agglomerations is mainly regulated
by the zeta () potential of the particles in suspension, which
is in turn dependent on the particle surface charge and pH of
the solution. These properties in an adhesive solution are of
high relevance when dealing with particles that may be sur-
face modied and embedded in solutions that undergo wide
changes in pH during application, as is the case for self-etching
systems. For the application as primer, the monomer Bis-MEP
in our self-etch adhesive is ionized by water and dissociates
into phosphoric acid and two HEMA molecules. During this
period, before the phosphoric acid is completely neutralized by
calcium from the dentin surface, the pH of the solution drops
considerably and particles become more mobile due to a
decrease in viscosity. Below a pH of  3, silica particles become
highly unstable (Kim and Lawler, 2005a). Silanization of silica
particles helps to increase the stability of the dispersion at
lower pH values by shifting the isoelectric point toward higher
pH (Jesionowski, 2003). During neutralization, an incre-
ase in pH beyond 5 may destabilize the suspension, as the -
potential of silanized particles enter a range of higher instabil-
ity (Jesionowski, 2003). Our adhesive solution was, however,
only ionized for application onto dentin for the mTBS test.
Nonetheless, deeper knowledge of this dynamic in self-etching
systems is of high interest and warrants further research.
 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108
Monomer composition also regulates the overall hydro-
philicity of the resin mixture and consequently the quantity
of water that is absorbed by the cured polymer (Ito et al.,
2005). In the monomer blend used in the present study,
carbonyl groups in HEMA/Bis-MEP along with the two hydro-
xyl groups in Bis-GMA and in the phosphate functional group
of Bis-MEP form polar sites to which water molecules are able
to form hydrogen bonds. With the incorporation of silanized
particles to the monomer solution, additional OH that
remain available from free silanol groups at the glass particle
surface, as well as carbonyl groups from the silane backbone
molecule add to the hydrophilicity of the resin (Karabela and
Sideridou, 2008; Sideridou and Karabela, 2009). Water sorp-
tion then increases for the lled adhesives, even though they
show higher monomer conversion (Figs. 1 and 2). Since a small
amount of resin cannot absorb more water than a large amount
of the same resin, the elevated water sorption values were
normalized by the wt% of resin phase to reveal the contribution
of the ller and silane polar groups to the water uptake in the
lled adhesives (Fig. 1). The large solubility values for the lled
adhesives also reect the higher amount of water that they
absorbed, which wash out the unreacted monomer fraction
during desorption. Due to the limited observation period
(7 days), the solubility values for the lled adhesives may also
have been underrepresented, as more hydrophilic blends loose
water at a slower rate due to the higher density of dipoledipole
interactions (Dickens et al., 2010).
An improvement in degree of conversion, shown by the
progressive decrease in the methacrylic/aromatic carbon
double bond ratio, occurred after addition of llers. The active
role of llers during polymerization is supported by the accel-
eration of the reaction shown during viscosity measurement
(Fig. 2). The viscosity curves during polymerization show that
the polymeric network reaches higher levels of rigidity earlier
for lled adhesives. Harsch et al. (2007) demonstrated this effect
for epoxy resins lled with silica particles, which showed to
induce a reduction in the time the network needs to reach its
gelation state. Their results highlighted the role of surface
functionalization of particles as essential in this mechanism,
acting by absorbing the reacting monomers at the surface.
Fillers then serve as xed sites for chain growth initiation. By
increasing ller concentration, neighboring particles become
closer together and narrow the interstice spacings, where the
copolymerization of opposite growing polymeric chains occurs.
This change in polymerization kinetics expedited by llers
seems also to potentialize monomer conversion, probably by
spatially homogenizing the sites for monomer reaction. A faster
growing polymeric chain reaction may bring important clinical
benets for adhesive procedures. In scenarios of light energy
density attenuation (e.g., luting of indirect restorations)
(Lohbauer et al., 2010b), higher degrees of cure may be assured
for more efcient curing reactions.
Finally, the compromises that are to be reached when
incorporating llers to dental adhesives must balance
the improvements in polymerization quality and speed,
increased mechanical properties and bond strength against
increased water sorption and solubility. These effects were
studied here using only one monomer composition of high
hydrophilicity, which may be further tailored using different
monomer mixtures.
107
5. Conclusions
Considering the limitations of the methods used in the
present study, we can conclude that the incorporation of
nanoparticle llers in a particular self-etching monomer
mixture has the potential to promote the strengthening of
the adhesive resin and the adhesive interface, with improve-
ments to the degree of conversion and polymerization ef-
cacy. As drawbacks, llers may substantially increase the
viscosity of the resin and expedite water sorption into the
adhesive layer and the elution of solubilized components.
Acknowledgments
Experimental adhesives were supplied by Voco, Germany.
r e f e r e nc e s
Braga, R.R., Ballester, R.Y., Ferracane, J.L., 2005. Factors involved in
the development of polymerization shrinkage stress in resin-
composites: a systematic review. Dent. Mater. 21, 962970.
Chabrier, F., Lloyd, C.H., Scrimgeour, S.N., 1999. Measurement at
low strain rates of the elastic properties of dental polymeric
materials. Dent. Mater. 15, 3338.
Chan, K.S., Lee, Y.-D., Nicolella, D.-P., Furman, B.R., Wellinghoff, S.,
Rawls, H.R., 2007. Improving fracture toughness of dental
nanocomposites by interface engineering and micromechanics.
Eng. Fract. Mech. 74, 18571871.
Chen, L., Yu, Q., Wang, Y., Li, H., 2011. BisGMA/TEGDMA dental
composite containing high aspect-ratio hydroxyapatite
nanofibers. Dent. Mater. 27, 11871195.
Chung, S.M., Yap, A.U.J., Koh, W.K., Tsai, K.T., Lim, C.T., 2004.
Measurement of Poissons ratio of dental composite
restorative materials. Biomaterials 25, 24552460.
Cook, W.D., Moopnar, M., 1990. Influence of chemical structure on
the fracture behaviour of dimethacrylate composite resins.
Biomaterials 11, 272276.
Dickens, S.H., Flaim, G.M., Floyd, C.J.E., 2010. Effects of adhesive,
base and diluent monomers on water sorption and conversion
of experimental resins. Dent. Mater. 26, 675681.
Frankenberger, R., Lopes, M., Perdigao, J., Ambrose, W.W., Rosa, B.T.,
2002. The use of flowable composites as filled adhesives. Dent.
Mater. 18, 227238.
Frankenberger, R., Pashley, D.H., Reich, S.M., Lohbauer, U.,
Petschelt, A., Tay, F.R., 2005. Characterization of resindentine
interfaces by compressive cyclic loading. Biomaterials 26,
20432052.
Greaves, G.N., Greer, A.L., Lakes, R.S., Rouxel, T., 2011. Poissons
ratio and modern materials. Nat. Mater. 10, 823837.
Harsch, M., Karger-Kocsis, J., Holst, M., 2007. Influence of fillers
and additives on the cure kinetics of an epoxy/anhydride
resin. Eur. Polym. J. 43, 11681178.
Irwin, G.R., 1964. Structural aspects of brittle fracture. Appl.
Mater. Res. 3, 6581.
Ito, S., Hashimoto, M., Wadgaonkar, B., Svizero, N., Carvalho, R.M.,
Yiu, C., Rueggeberg, F.A., Foulger, S., Saito, T., Nishitani, Y.,
Yoshiyama, M., Tay, F.R., Pashley, D.H., 2005. Effects of resin
hydrophilicity on water sorption and changes in modulus of
elasticity. Biomaterials 26, 64496459.
Jesionowski, T., 2003. Influence of aminosilane surface
modification and dyes absorption on zeta potential of
spherical silica particles formed in emulsion systems. Colloids
Surf. A 222, 8794.
108 journal of the mechanical behavior of biomedical materials 37 (2014) 100 108
Karabela, M.M., Sideridou, I., 2008. Effect of the structure of silane
coupling agent on sorption characteristics of solvents by
dental resin nanocomposites. Dent. Mater. 24, 16311639.
Kies, A.J., Sullivan, A.M., Irwin, G.R., 1950. Interpretation of
fracture markings. J. Appl. Phys. 21, 716.
Kim, J., Lawler, D.F., 2005a. Charateristics of zeta potential
distribution in silica particles. Bull. Korean Chem. Soc. 26,
10831089.
Kim, J.S., Cho, B.H., Lee, I.B., Um, C.M., Lim, B.S., Oh, M.H., Chang,
C.G., Son, H.H., 2005b. Effect of the hydrophilic nanofiller
loading on the mechanical properties and the microtensile
bond strength of an ethanol-based one-bottle dentin
adhesive. J. Biomed. Mater. Res. B: Appl. Biomater. 72, 284291.
Lohbauer, U., Belli, R., Ferracane, J.L., 2013. Factors involved in the
mechanical fatigue degradation of dental resin composites.
J. Dent. Res. 92, 584591.
Lohbauer, U., Wagner, A., Belli, R., Stoetzel, C., Hilpert, A.,
Kurland, H.-D., Grabow, J., Muller, F.A., 2010a. Zirconia
nanoparticles prepared by laser vaporization as fillers for
dental adhesives. Acta Biomater. 6, 45394546.
Lohbauer, U., Pelka, M., Belli, R., Schmitt, J., Mocker, E., Jandt, K.D.,
Muller, F.A., 2010b. Degree of conversion of luting resins
around ceramic inlays in natural deep cavities: a micro-
Raman spectroscopy analysis. Oper. Dent. 35, 579586.
Masouras, K., Silikas, N., Watts, D.C., 2008. Correlation of filler
content and elastic properties of resin-composites. Dent.
Mater. 24, 932939.
Miyazaki, M., Ando, S., Hinoura, K., Onose, H., Moore, B.K., 1995.
Influence of filler addition to bonding agents on shear bond
strength to bovine dentin. Dent. Mater. 11, 234238.
Sadat-Shojai, M., Atai, M., Nodehi, A., Khanlar, L.N., 2010.
Hydroxyapatite nanorods as novel fillers for improving the
properties of dental adhesives: synthesis and application.
Dent. Mater. 26, 471482.
SCHOTT, 2010. NanoFine NF180 Product Information Sheet.
Available at: http://ww.schott.com/epackiaging/russian/
download/schott_dental_glass_nanofine180_datasheet.pdf.
Sideridou, I., Karabela, M.M., 2009. Effect of the amount of 3-
methacyloxypropyltrimethoxysilane coupling agent on
physical properties of dental resin nanocomposites. Dent.
Mater. 25, 13151324.
Soderholm, K.-J., 2010. Review of the fracture toughness
approach. Dent. Mater. 26, e63e77.
Solhi, L., Atai, M., Nodehi, A., Imani, M., Ghaemi, A., Khosravi, K.,
2011. Poly(acrylic acid) grafted montmorillonite as novel fillers
for dental adhesives: synthesis, characterization and
properties of the adhesive. Dent. Mater. 28, 369377.
Tam, L.E., Khoshand, S., Pilliar, R.M., 2001. Fracture resistance of
dentin-composite interfaces using different adhesive resin
layers. J. Dent. 29, 217225.
Tam, L.E., Pilliar, R.M., 2000. The effect of interface stiffness on
dentin-composite interfacial fracture resistance. J. Dent. 28,
487493.
Tay, F.R., Pashley, D.H., Kapur, R.R., Carrilho, M.R., Hur, Y.B.,
Garret, L.V., Tay, F.R., 2007. Bonding BisGMA to dentin: a proof
of concept for hydrophobic dentin bonding. J. Dent. Res. 86,
10341039.
Tjaderhane, L., Nascimento, F.D., Breschi, L., Mazzoni, A.,
Tersariol, I.L., Geraldeli, S., Tezvergil-Mutluay, A., Carrilho, M.,
Carvalho, R.M., Tay, F.R., Pashley, D.H., 2013. Strategies to
prevent hydrolitic degradation of the hybrid layer: a review.
Dent. Mater. 29, 9991011.
Van Landuyt, K.L., Snauwaert, J., De Munck, J., Peumans, M.,
Yoshida, Y., Poitevin, A., Coutinho, E., Suzuki, K., Lambrechts,
P., Van Meerbeek, B., 2007. Systematic review of the chemical
composition of contemporary dental adhesives. Biomaterials
28, 37573785.
Van Meerbeek, B., Peumans, M., Poitevin, A., Mine, A., Van Ende,
A., Neves, A., De Munck, J., 2010. Relationship between bond-
strength tests and clinical outcomes. Dent. Mater. 26,
e100e121.
Van Meerbeek, B., Yoshihara, K., Yoshida, Y., Mine, A., De Munck,
J., Van Landuyt, K.L., 2011. State of the art of self-etch
adhesives. Dent. Mater. 27, 1728.
Wagner, A., Belli, R., Stoetzel, C., Hilpert, A., Muller, F.A.,
Lohbauer, U., 2013. Biomimetically- and hydrotermally-grown
Hap nanoparticles as reinforcing fillers for dental adhesives. J.
Adhes. Dent. 15, 413422.
Wang, W., Sadeghipour, K., Baran, G., 2008. Finite element
analysis of the effect of an interphase on toughening of a
particle reinforced composite. Composites Part A: Appl. Sci.
Manuf. 39, 956964.