DENTAL-2455; No.

of Pages 11
ARTICLE IN PRESS
d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.intl.elsevierhealth.com/journals/dema

Slow cooling protocol improves fatigue life of

zirconia crowns

Vitor G. Paula a , Fabio C. Lorenzoni d , Estevam A. Bonfante a,∗ ,

Nelson R.F.A. Silva b , Van P. Thompson c , Gerson Bonfante a
a Department of Prosthodontics, University of São Paulo – Bauru College of Dentistry, Bauru, SP, Brazil
b Department of Restorative Dentistry, Federal University of Minas Gerais, UFMG, Belo Horizonte, MG, Brazil
c Department of Biomaterials, Biomimetics and Biophotonics, King’s College London, Guy’s Hospital, London, United

Kingdom
d Department of Prosthodontics, UNIGRAN University, Dourados, MS, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Objective. To compare the fatigue life and damage modes of zirconia crowns fabricated with
Received 2 September 2013 and without framework design modification when porcelain veneered using a fast or slow
Received in revised form cooling protocol.
22 December 2013 Methods. Composite resin replicas of a first molar full crown preparation were fabricated.
Accepted 27 October 2014 Zirconia copings were milled as conventional (0.5 mm even thickness, Zr-C, n = 20,) or mod-
Available online xxx ified (lingual margin of 1.0 mm thickness, 2.0 mm height connected to two proximal struts
of 3.5 mm height, Zr-M, n = 20). These groups were subdivided (n = 10 each) according to the
Keywords: veneer cooling protocol employed: fast cooling (Zr-CFast and Zr-MFast) and slow cooling
Ceramics (Zr-CSlow and Zr-MSlow). Crowns were cemented and fatigued for 106 cycles in water. The
Fatigue number of cycles to failure was recorded and used to determine the interval databased 2-
Weibayes parameter probability Weibull distribution parameter Beta (ˇ) and characteristic life value
Cooling rate Eta ().
Fractography Results. 2-parameter Weibull calculation presented ˇ = 5.53 and ˇ = 4.38 for Zr-MFast and Zr-
Zirconia CFast, respectively. Slow cooled crowns did not fail by completion of 106 cycles, thereby
Weibayes calculation was applied. Increased fatigue life was observed for slow cooled crowns
compared to fast cooled ones. Groups Zr-MFast and Zr-MSlow presented no statistical differ-
ence. Porcelain cohesive fractures were mainly observed in fast cooled groups. Slow cooled
crowns presented in some instances inner cone cracks not reaching the zirconia/veneer
interface.
Significance. Improved fatigue life in tandem with the absence of porcelain fractures were
observed in slow cooled crowns, regardless of framework design. Crowns fast cooled chiefly
failed by porcelain cohesive fractures.
© 2014 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

∗
Corresponding author at: Alameda Octávio Pinheiro Brisola 9-75, 17.012-901 Bauru, SP, Brazil. Tel.: +55 14 98153 0860;
fax: +55 14 3234 5805.
E-mail address: estevamab@gmail.com (E.A. Bonfante).
http://dx.doi.org/10.1016/j.dental.2014.10.005
0109-5641/© 2014 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
2 d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx
1. Introduction
Over the past ten years several in vitro [1–5] and clinical stud-
ies [6,7] have attempted to document and understand the
nature of porcelain veneer cohesive fractures in all-ceramic
restorations, specially yttria-tetragonal zirconia polycrystals
(Y-TZP) [6,7]. Besides issues related to manufacturing process
(i.e., porosities and micro-structural defects) [8], chipping has
been suggested to be strongly related with; residual thermal
stress [9–12], coefficient of thermal expansion mismatches [1],
poor framework design [3,4,13–15], and others [12].
Changes in the porcelain physical properties (during the
transition from the visco-elastic to solid state) when it is
cooled down from the sintering temperature to room tem-
perature take place in every firing cycle [16,17]. As a result,
non-uniform residual thermal stresses may be produced
throughout the porcelain veneer as a function of cooling
rates [18,19]. Such residual stress may be induced within the
porcelain veneer either by thermal expansion mismatches
or by the temperature gradients produced during cooling or
both [12,17]. Similarly to tempered glass, during a fast cool-
ing process the outer crown surface becomes solid first and
starts to contract inducing a compressive stress. However,
while the inner surface is still in a visco-elastic state and
at high temperature, leading to high tensile stress when the
system reaches room temperature [12,17,19]. This process is
expected given that the porcelain veneer inner bulk remains
in a temperature above its glass-transition temperature (Tg ),
leading to the gradient temperature (difference between the
inner and outer temperature) [12]. As a consequence, resid-
ual thermal stresses are held within the porcelain veneer
[9,12,20], also known as tempering. In addition, Y-TZP shows a
low thermal diffusivity (thermal conductivity of 2 W m−1 K−1
whereas metals vary from 100 to 300 W m−1 K−1 ) [21,22],
which may increase the thermal gradient intensity within
veneer porcelain [9,12,20,23]. Therefore, the multifaceted role
of the above factors, has been speculated as potential sources
related to high Y-TZP chipping rates reported in the literature
[5,8,10,12,23].
Recently, efforts have been devoted to the reduction of
residual thermal stresses in Y-TZP restorations by managing
the rates of cooling [9,10,24,25]. It has been reported that over-
all thermal gradients within the porcelain veneered to Y-TZP
decreased when a slow cooling protocol was used [23,26]. Also,
thermodynamic analysis performed in molar crowns showed
that different results may be achieved as a function of cooling
protocols [23]. It has been shown that slow cooling (the furnace
door opened when the temperature was below the porcelain
veneer Tg ) produced a small temperature variation between
the inner and outer surfaces of crowns and, more important,
it decreased the residual thermal stresses within the veneer
layer [23]. However, to date few attempts regarding the influ-
ence of cooling rate on the fatigue life of Y-TZP crowns are
available.
Most manufacturers of porcelain veneers for Y-TZP restora-
tions have modified their firing guidelines, advocating that
technicians adopt a slow cooling rate protocol. In 2009 for
example, one manufacturer recommended that the furnace
was opened soon after the end of the firing cycle [27], and
Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
a year later it was modified and crowns should be kept into
the closed furnace up to temperature reaching 450 ◦ C, slightly
below the Tg (490 ◦ C) of the veneer porcelain [28].
Another topic that has frequently been revisited is frame-
work design modifications in Y-TZP restorations [4,14,29].
Clinical evaluations have suggested that porcelain chipping
is frequently associated with areas without appropriate porce-
lain support (proximal areas) [14,15,30]. This empirical concept
relies on the fact that the porcelain veneer should have an
even thickness, of no more than 2 mm, and be supported by
the core [31]. Although framework design modifications have
shown to decrease chipping rates in clinical trials [15,32,33]
no control groups (i.e., even thickness framework) have been
included, hindering a more comprehensive understanding
between framework design modification and chipping [15,32].
In light of the aforementioned information, this study
sought to evaluate the fatigue life and failure modes of Y-TZP
molar crowns fabricated with two different porcelain veneer
cooling protocols with and without framework design modifi-
cation. The following null hypotheses were tested: (1) cooling
protocols during glazing will not improve the fatigue life of Y-
TZP systems and (2) framework design modification will not
improve the fatigue life of Y-TZP crown systems, irrespective
of cooling rate.
2. Materials and methods
2.1. Sample preparation
A plastic maxillary first molar positioned in a mannequin
(Plastic mannequin tooth – MOM, Marilia, SP, Brazil) was pre-
pared to receive a full crown by reducing the axial walls by
1.5 mm, the occlusal surface by 2.0 mm and shoulder margin
(1.2 mm of thickness) with rounded internal angles. After-
wards, 40 composite resin replicas were obtained from a vinyl
polysiloxane impression (Express – 3M ESPE, St. Paul, MN,
USA) of the prepared tooth, which was incrementally (2.0 mm
of thickness) packed with composite resin (Z100 – 3M ESPE,
St. Paul, MN, USA) and light-cured (Ultralux, Dabi Atlante,
Ribeirão Preto, SP, Brazil) according to manufacturer’s instruc-
tions. All resin tooth replicas were stored in distilled water
for at least 30 days to minimize dimensional changes during
testing [34]. Each replica was vertically positioned in 25 mm
diameter PVC tubes and potted with acrylic resin (Jet, Clássico
Artigos Odontológicos, São Paulo, SP, Brazil). The preparation
finish line was maintained 2 mm above the potting sur-
face. Stone dies were acquired from polyether impressions
(Impregum F – 3M-ESPE, St. Paul, MN, USA) of each resin-tooth
replica and sent to a dental laboratory.
Stone dies were randomly divided in 2 groups (n = 20, each)
according to framework design, as: conventional (Zr-C) or
modified design (Zr-M). Subsequently, each group was subdi-
vided in two subgroups (n = 10) according to thermal cooling
protocol as: fast cooling protocol (Zr-CFast and Zr-MFast) and
slow cooling protocol (Zr-CSlow and Zr-MSlow).
Frameworks were milled from pre-sintered Y-TZP blocks
(e.max ZirCAD, shade MO 0, block size 15C – Ivoclar Vivadent,
Schaan, Liechtenstein) and sintered in the furnace (Sintra-
mat, Ivoclar Vivadent AG, Schaan, Liechtenstein) at 1500 ◦ C
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx 3

Fig. 1 – Proximal view of (a) standard and (b) modified core designs before veneering with porcelain.

Table 1 – Programat® EP 3000 – Firing program table

(Ivoclar Vivadent) – March 2010 [31].
Program ZirLiner e.Max Ceram
◦
Stand-by temperature 403 C 403 ◦ C
Stand-by time 4 min 4 min
Temperature increase 40 ◦ C/min 40 ◦ C/min
Holding temperature 960 ◦ C 750 ◦ C
Holding time 1 min 1 min
Vacuum on 450 ◦ C 450 ◦ C
Vacuum off 959 ◦ C 749 ◦ C
Opening furnace 960 ◦ C 750 ◦ C
Cooling rate – –

for 8 h. Zr-C cores were manufactured with an even thick-

ness of 0.5 mm. For Zr-M cores, a lingual margin (thickness of
1.0 mm with 2.0 mm height) connected to two proximal struts
of 3.5 mm height were added to an even thickness of 0.5 mm,
as described previously [3] (Fig. 1).
All Y-TZP cores were veneered with porcelain (IPS e.max
Ceram, shade Transpa Clear – Ivoclar Vivadent AG, Schaan,
Liechtenstein) and sintered in a ceramic furnace (Progra-
mat EP3000 – Ivoclar Vivadent, Schaan, Liechtenstein). The
hand-layering process was performed in accordance with the
manufacturer’s recommendation, starting with one liner firing
(ZirLiner – Ivoclar Vivadent AG, Schaan, Liechtenstein) fol-
lowed by firing two layers of porcelain. Finally, the crowns
were manually finished and polished before the glaze fir-
ing. The firing schedule of liner and veneering porcelain are
shown in Table 1 and Fig. 2A. A silicone matrix (Zetalabor
– Zhermack, Badia Polesine, Rovigo, Italy) of the unprepared
plastic mannequin tooth was used to guide and standardize
the porcelain veneer contour. The occlusal and axial surfaces
showed approximately 1.5 and 1.0 mm of thickness, respec-
tively. The glaze layer (Glaze Paste and Stain Liquid Long Life,
IPS e.Max Ceram – Ivoclar Vivadent AG, Schaan, Liechtenstein)
was applied onto the porcelain surface and treated according
to one of the following cooling protocols (Fig. 2B and Table 2):

Fig. 2 – (A) Programat® EP 3000 – Firing program (Ivoclar

• Fast cooling: after the end of the glaze firing cycle, the fur-
Vivadent) – March 2010 [28]. (B) Glaze firing program.
nace head was opened. Then, crowns were immediately
taken out from the mesh-tray to the bench, leading to cool

Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
4 d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx
Table 2 – Glaze firing program table.
Program Fast cooling Slow cooling
Stand-by temperature 403 ◦ C 403 ◦ C
Stand-by time 6 min 6 min
Temperature increase 60 ◦ C/min 60 ◦ C/min
Holding temperature 725 ◦ C 725 ◦ C
Holding time 1 min 1 min
Vacuum on 450 ◦ C 450 ◦ C
Vacuum off 724 ◦ C 724 ◦ C
Opening furnace 725 ◦ C 50 ◦ C
Cooling rate – 10 ◦ C/min
at room temperature. This thermal cycle followed manufac-
turer’s recommendation published in 2009 [27].
• Slow cooling protocol: the furnace had the cooling rate set to
10 ◦ C/min and crowns were kept into the furnace until the
temperature reached 50 ◦ C, after which the furnace head
was opened.
The glazed crowns were cemented to the aged resin-
composite tooth replicas using a dual cure resin cement (Rely
X ARC – 3M-ESPE, St. Paul, MN, USA) following the manu-
facturer’s instructions. No treatment was made in the Y-TZP
cementation surface. Prior to mechanical testing, samples
were incubated in distilled water at 37 ◦ C for at least 24 h.
2.2. Fatigue test
Fatigue test was carried out in a fatigue machine (Model MSFM
– Elquip – São Carlos, SP, Brazil) in r-ratio mode at a 30–300 N
load range for 106 cycles or until failure under distilled water
(37 ◦ C) [3]. As the indenter was in contact with the specimen,
the initial contact load of 30 N was applied without abrasion,
attrition or impact, steadily increasing to 300 N and decreasing
back to 30 N. The load was applied through a spherical steel
indenter (3.18 mm radius) on the center of the occlusal sur-
face, at 2 Hz. Our attempt was to simulate the most typical
molar relationship in natural dentition (Angle class I) where
the mesiolingual cusp of the maxillary first molar is situated
in the central fossa area of the mandibular first molar [35].
After a predetermined number of cycles (125,000) the test was
interrupted for crown surface damage accumulation inspec-
tion under stereomicroscopy (Leica Zeiss MZE, Mannheim,
Germany). Specimens were then repositioned in the machine
for an additional 125,000 cycles and inspections until comple-
tion of 106 cycles or failure.
2.3. Data analysis
The failure during fatigue was recorded as a function of the
number of cycles. This information was used to determine
the interval databased 2-parameter probability Weibull dis-
tribution for Zr-CFast and Zr-MFast (Weibull 7++; Reliasoft,
Tucson, AZ, USA). Weibull shape parameter Beta (ˇ) and char-
acteristic value Eta () (here the number of stress cycles (n) at
which 62.3% of the specimens would fail) were determined
employing 90% confidence levels for upper and lower lim-
its where applicable. If either one of groups displayed no
failures, a Weibayes [36] calculation was employed based on
Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
the ˇ of the counter group’s distribution. Weibayes is a one-
parameter Weibull analysis technique developed to “solve
problems when traditional Weibull analysis has large uncer-
tainties or cannot be used because there are no failures” [36].
“A basic assumption governing the accuracy of Weibayes anal-
ysis and its associated test regimens is that the value of the
Weibull shape parameter (ˇ) is known or can be reasonably
estimated”. “Knowledge of ˇ can be derived from historical
failure data, prior experience, or from engineering knowledge
of the physics of the failure”. Given ˇ, an equation may be
derived using the method of maximum likelihood to deter-
mine the characteristic life, , as follows [36,37]:
  ˇ1

N
Tˇi
= (1)
r
i=1
where Ti , test time for each sample; r, number of failures; N,
sample size; , characteristic life of the cumulative density
function (CDF) = 63.2%; ˇ, Weibull shape parameter.
A Weibull distribution is derived from  (determined by Eq.
(1)) using the assumed ˇ. The 2-parameter Weibull probability
result was plotted using the counter group Weibayes value for
pairwise comparisons. If two or more groups showed no fail-
ures, the Weibayes lines were plotted and a probability Weibull
graphic generated.
2.4. Failure mode characterization
The criteria for failure were inner cone cracks reaching
the Y-TZP framework, porcelain veneer cohesive fracture
or Y-TZP core fracture. Representative fractured surfaces
were characterized under polarized-light microscopy (MZ-
APO stereomicroscope, Carl Zeiss MicroImaging, Thornwood,
NY, USA) followed by Scanning Electron Microscopy (SEM)
(Model S-3500N; Hitachi, Japan). If no fractures were noted
after completion of final fatigue interval, the specimen was
considered as suspended. Subsequently, to characterize the
subsurface damage, suspended crowns were embedded in
epoxy resin (Resina Epoxi RD6921, Redelease, São Paulo,
Brazil), sectioned and then serially polished along the plane of
crack extension with silicon carbide papers (320, 400, 500, 600,
1200, 2000 and 2500 grits), under copious water irrigation. The
damage accumulation was inspected and registered under
stereomicroscopy (Leica Zeiss MZE, Mannheim, Germany).
3. Results
During fatigue testing, all crowns fast cooled, from Zr-CFast
and Zr-MFast groups, failed before completion of 1,000,000
cycles chiefly by porcelain cohesive fracture. Three crowns
with conventional framework and five from the modified fast
cooled groups did not present chippings. However, the sub-
surface damage inspection of these crowns revealed inner
cone cracks reaching the core/veneer interface. While inner
cone cracks reaching the interface reflect the porcelain’s fail-
ure to resist crack growth, they do not represent a clinical
fracture per se, only the porcelain fractured crowns were con-
sidered for further analysis (n = 7 from the Zr-CFast and n = 5
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx 5

Fig. 3 – (a) Probability Weibull 2P plot showing probability line for Zr-CFast and Weibayes line projection of group Zr-CSlow.
Note that 90%CB calculation for Zr-CFast did not allow confidence intervals. (b) Probability Weibull 2P plot showing
probability line calculation of Zr-MFast and Weibayes probability line projection of group Zr-MSlow. (c) Probability Weibull
plot showing the Zr-CSlow and Zr-MSlow Weibayes projection lines. Note the different inclination of the probability lines as
a result of the different Beta values: ˇ = 4.38 for Zr-MSlow and ˇ = 5.53 for Zr-CSlow.

from the Zr-MFast). Conversely, crowns subjected to the slow are shown in Fig. 3. For the results distribution, the groups
cooling protocol, regardless of framework design, (Zr-CSlow were plotted based on the core design varying the cooling rate
and Zr-MSlow) were suspended since they did not fail at the process as described in materials and methods. 2-Parameter
completion of 106 cycles. Results of the Weibull calculation Weibull calculation was performed for Zr-CFast and Zr-MFast.

Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
6 d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx

Fig. 4 – (a) Light-polarized microscope and SEM images of a representative failed Zr-MFast crown. Palatal view of cohesive
porcelain fracture shows that the crack involved the lingual cusp. (b) SEM lingual view displaying the indentation area
(asterisk) and the dotted arrows shows the crack propagation toward the cervical and proximal areas. (c) and (d) The
concave portion of the arrest lines (pointer) points to the origin (asterisk). Continued propagation toward the margins is
suggested by the series of hackles (bowshot). (e) and (f) Magnification of wake-hackles (pointers) corroborating the origin
and direction of crack propagation along with twist-hackles (white arrows).

Groups Zr-CSlow and Zr-MSlow did not present any failure,
and therefore Weibayes calculation was performed [36]. In our
study we used the ˇ value from the respective counter group’s
design and ran Weibayes calculation. 2-parameter Weibull
calculation presented ˇ = 5.53 and ˇ = 4.38 for Zr-MFast and Zr-
CFast respectively. The Weibayes calculation using the same
ˇ values for the respective counter presented a shift to the
right of the Weibull probability lines (longer fatigue life). The 2-
parameter Weibull calculation for group Zr-CFast did not allow
90% confidence intervals calculation as a result of the distri-
bution of the failed specimens. However there is a significant
shift to the right of the Weibayes probability line for its counter
part group. Groups Zr-MFast and Zr-MSlow presented no sta-
tistical difference as a result of overlap of the upper limit of
the confidence bounds. Fig. 3 also presents the Weibayes prob-
ability line calculations for groups Zr-MSlow and Zr-CSlow.
3.1. Failure modes
The chief failure mode observed for crowns subjected to the
fast cooling protocol was porcelain veneer cohesive fracture,
regardless of framework design (Zr-CFast, Zr-MFast), with-
out exposing the Y-TZP surface. Of special interest was the
fact that smaller chippings were consistently observed for
Zr-MFast relative to Zr-CFast group. In the Zr-MFast group,
chippings were generally confined to the lingual cusp (Fig. 4),
whereas chip-off fractures extending to the proximal sur-
faces were observed in Zr-CFast crows (Fig. 5). Qualitative

Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx 7

Fig. 5 – (a) Light-polarized microscope image of a Zr-CFast crown shows the extension of the fractured porcelain veneer.
Note that a layer of porcelain remained onto the Y-TZPs framework surface. (b) SEM lingual overview of the fractured crown
and indentation area (asterisk) where the fracture initiated and propagated toward margins and proximal areas (dotted
arrows). (c) Magnification of (b) showing the presence of a series of wake-hackles (pointers) confirming the fracture origin
and direction of the crack propagation (dotted arrow). (d) A series of twist-hackles (white arrows) that strongly suggests the
presence of residual thermal stress and further confirms the direction of crack propagation toward the margin of the fracture.

fractography performed in polarized-light microscopy and
SEM showed the presence of quasiplastic deformation at the
indentation area. Telltale fractographic marks including hack-
les, wake hackles, and arrest lines suggesting the direction of
crack propagation from the indentation area toward the mar-
gins of the fractured porcelain veneer were observed. Twist
hackles were only identified in those groups subjected to fast
cooling.
Crowns subjected to the slow cooling protocol of both
framework designs (Zr-CSlow and Zr-MSlow) were consid-
ered suspended since they did not present failures at the
completion of the 106 cycles. The subsurface damage inspec-
tion of these specimens revealed inner cone cracks as the
dominant crack system mechanism in all samples from both
groups, although they did not reach the core/porcelain inter-
face (Fig. 6). No radial crack and bulk fractures were observed.
4. Discussion
The present study evaluated the effect of cooling processes
and framework designs on Y-TZP molar crowns fatigue
life. The absence of statistical difference in fatigue life in
crowns submitted to the slow cooling protocol, regardless
of framework design, lead us to reject the first postulated
null hypothesis. The second null hypothesis was accepted
since framework design modification did not improve fatigue
life within either the fast or slow cooled crowns. However
when intergroup comparisons were made, the relevance of a
modified framework should not be overlooked since the Zr-
MFast crowns were not statistically different from slow cooled
crowns of any framework design. Nevertheless, it appears that
cooling protocols played a major role in fatigue life in the
current study. As expected from previous studies [6,9,38], no
Y-TZP core fractures were detected. When porcelain veneer
fractures occurred, they were always cohesive within the
porcelain veneer, without exposing the Y-TZP core [3,5,39–42].
Crowns cooled slowly did not present porcelain chippings
within the established limit of one million fatigue cycles and
the only observed failure mode was near-field damage (inner
cone crack), in both framework designs. It is likely that such
fatigue performance improvement may be explained by a
reduction in residual stress entrapment within the veneer
layer [16,20,23]. A recent investigation evaluated the influ-
ence of cooling protocols on thermal stress formation in Y-TZP
molar crowns showing that when samples were cooled slowly
the outer porcelain veneer surface did not become solid first
(due to the absence of thermal gradient) allowing residual
thermal stress relaxation by viscosity flow [23]. As fatigue
testing was suspended at one million cycles to allow for dam-
age inspection of groups within the same range of cycles,
it must be acknowledged that extended fatigue could have
provided a more clinically realistic failure scenario, although
increasing testing time to an unknown extent. In addition, as

Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
8 d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx

Fig. 6 – (a) Occlusal light-polarized micrograph shows the indenter contact area on the porcelain surface and the resulting
quasiplastic deformation on a Zr-MSlow crown. (b) Magnification of the square area showed in (a), reveals the presence of
quasiplatic deformation, which is a coalescing microcrak involving the indentation area (bowshot) after 106 cycles. (c)
Bucco-lingual section of the same crown, polished and examined throughout the damaged area, showing inner cone crack
formation immediately below the indentation area, not reaching the zirconia/veneer interface (d) Magnification of the inner
cone crack confirming its extension. (e) Bucco-lingual section micrograph of a polished Zr-CSlow crown showing the
indentation area in the blue square and in (f) its magnification revealing inner cone cracks (pointer) formed immediately
bellow the indentation area, also not reaching the interface. R = resin composite; Y-TZP = zirconia framework; V = veneering
porcelain.

sliding-fatigue was omitted, porcelain fractures revealed ori-
gin from the occlusal surface, but not related to surface
roughness due to wear, as reported in clinical fractogra-
phy [43] and laboratory studies involving such methodology
[2,44,45,47].
Regarding the slow cooling protocol used in this study,
it must be emphasized that it is experimental for having
extended the slow cooling process below the Tg , in contrast to
what is currently recommended by most manufacturers (slow
cooling in the last stain/glaze firing cycle until Tg ). Whereas
the maintenance of the crowns into the furnace until 50 ◦ C
demanded additional laboratory time, it may be important to
reduce thermal gradients [12,23,46]. A comprehensive eval-
uation of stress distribution within the veneering ceramic
as a function of depth toward the substructure material has
shown that zirconia veneered samples present tensile stresses
around 0.5 mm from the surface with magnitude varying
inversely to the cooling rate [26].
Y-TZP crowns cooled fast resulted in similar fatigue life,
irrespective of framework design, but remarkable differences
were observed in porcelain cohesive fracture sizes. Whereas
fractures in the modified Y-TZP framework were confined to
lingual cusp, those in the conventional designed were more
extensive often involving both approximals in addition to the

Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx
buccal cusp. Despite such differences, from a clinical stand-
point only crowns with modified framework would potentially
be repaired, although ideally both would require replacement.
Such an observation questions the suitability of this partic-
ular framework design modification since it results in an
even porcelain thickness only on the lingual side. Perhaps, an
anatomic core design that provides an even thickness in all
cusps and proximal areas could be more relevant as recom-
mended by several manufacturers. A recent study shows that
an anatomical support of the veneering ceramic resulted in
reduced veneer chip size fractures and significantly increased
reliability with press and hand-layer veneering ceramics and
appears to be a promising tool to optimize long-term perfor-
mance of veneered Y-TZP crowns in clinical application [29].
One important fractographic finding was the consistent
presence of twist hackles in the groups cooled by the fast
protocol. Although it was not our aim to perform residual
thermal stress measurements, the presence of twist hackles,
strongly suggest the presence of significant thermal stress gra-
dients [47], which were induced in our study by the fast cooling
protocol [16,23,46].
Our Weibayes use of an assumed beta value means that
always there is a “risk” of using a beta higher than the “true”
value [37]. In this context, the impact of a higher beta is being
overly optimistic in the interpretation of the demonstrate reli-
ability and the possible exclusion of early failures. In our
study the beta value employed is based upon the ˇ value from
the respective counter group (fast cooled), which displayed a
higher number of failures than the slow cooled groups. More-
over, a way to investigate the effect of changes in an assumed
beta is by assuming two beta values (two beta lines), where
the higher the cumulative density function (CDF) level of the
beta line intersection point, the greater the risk of reaching
an overly optimistic conclusion for a fixed By life below that
intersection point [37].
Also, if two beta values were assumed, one as calculated
from results of the counter group and the other reduced by
two units, one graphic including both beta lines would be
generated. For a given N = 10 (based on our sample size), the
intersection of two beta lines is known as beta intersection
point and it most likely occurs at approximately CDF of 10%
(here the test time on each of the “N” samples is equal). For a
larger “N”, the beta intersection point occurs at less than 10%.
Such assumption is based on a general rule for beta intersec-
tion point where the smaller the Weibayes test sample size,
the higher the CDF value at which the beta intersection point
occurs [37]. Thus, our limited number of samples would allow
a lower beta intersection point close to the B10 life where 10%
of samples would fail.
It is acknowledged that several damage modes can com-
pete during the fatigue test such as radial cracks (far-field
damage) and inner, outer, and partial cone cracks (near-field
damage) [48–50]. The failure characterization by polishing
showed that some of the suspended samples (three ZR-CFast
and four Zr-MFast) presented inner cone cracks reaching the
core/veneer interface. Because from a clinical perspective the
absence of external porcelain fracture, as in the presence of
inner cone cracks reaching the interface, leads to a diagnosis
of crown survival, they were not deemed as failures for data
analysis. With continued load cycling it would likely evolve to
Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
9
a cohesive fracture from the extension of the inner cone crack-
ing [49–52]. Suspended crowns subjected to the slow cooling
protocol showed cone cracks formation virtually confined to
the quasiplastic zone beneath the indenter and therefore not
reaching the Y-TZP framework and are much less likely to lead
to cohesive failure with continued loading.
As observed in clinical trials, the proximal region has been
regarded as an area specially affected by chipping [6,14,30,53].
In a natural occlusal scheme the mesio bucal cusp of the first
mandibular molar may contact the marginal ridges of both
second premolar and first maxillary molar [54,55]. This sug-
gests that future fatigue testing in the proximal areas may be
of clinical relevance.
5. Conclusions
The first postulated null hypothesis which stated that cool-
ing protocols during glazing would not improve the fatigue
life of Y-TZP crowns was partially rejected. Decreased fatigue
life was observed only for crowns fast cooled of conventional
framework compared to any slow cooled crowns, whereas
modified framework fast cooled crowns were not signifi-
cantly different from slow cooled ones. The second null
hypothesis was accepted since framework design modifica-
tion did not improve fatigue life within either the fast or
slow cooled crowns. Slow cooling protocol is recommended for
more predictable mechanical performance of layered-zirconia
restorations.
Acknowledgements
We thank Dr. Raphael Piazzarollo Loureiro for his input on
ceramic processing engineering and the dental technician José
Carlos Romanini for crown fabrication. This study was sup-
ported, in part, by CNPq, and FAPESP Grant 2010/06152-9. The
authors acknowledge Hitachi S3500N SEM imaging that was
made possible by the New York University College of Den-
tistry’s cooperative agreement with the NIH/NIDCR.
references
[1] Bonfante EA, Rafferty B, Zavanelli RA, Silva NR, Rekow ED,
Thompson VP, et al. Thermal/mechanical simulation and
laboratory fatigue testing of an alternative yttria tetragonal
zirconia polycrystal core-veneer all-ceramic layered crown
design. Eur J Oral Sci 2010;118:202–9.
[2] Coelho PG, Silva NR, Bonfante EA, Guess PC, Rekow ED,
Thompson VP. Fatigue testing of two porcelain-zirconia
all-ceramic crown systems. Dent Mater 2009;25:1122–7.
[3] Lorenzoni FC, Martins LM, Silva NR, Coelho PG, Guess PC,
Bonfante EA, et al. Fatigue life and failure modes of crowns
systems with a modified framework design. J Dent
2010;38:626–34.
[4] Silva NR, Bonfante EA, Rafferty BT, Zavanelli RA, Rekow ED,
Thompson VP, et al. Modified Y-TZP core design improves
all-ceramic crown reliability. J Dent Res 2011;90:104–8.
[5] Silva NR, Bonfante EA, Zavanelli RA, Thompson VP, Ferencz
JL, Coelho PG. Reliability of metalloceramic and
zirconia-based ceramic crowns. J Dent Res 2010;89:1051–6.
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
10 d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx
[6] Al-Amleh B, Lyons K, Swain M. Clinical trials in zirconia: a
systematic review. J Oral Rehabil 2010;37:641–52.
[7] Sailer I, Feher A, Filser F, Gauckler LJ, Luthy H, Hammerle CH.
Five-year clinical results of zirconia frameworks for posterior
fixed partial dentures. Int J Prosthodont 2007;20:383–8.
[8] Zarone F, Russo S, Sorrentino R. From
porcelain-fused-to-metal to zirconia: clinical and
experimental considerations. Dent Mater 2011;27:83–96.
[9] Guazzato M, Walton TR, Franklin W, Davis G, Bohl C,
Klineberg I. Influence of thickness and cooling rate on
development of spontaneous cracks in porcelain/zirconia
structures. Aust Dent J 2010;55:306–10.
[10] Rues S, Kroger E, Muller D, Schmitter M. Effect of firing
protocols on cohesive failure of all-ceramic crowns. J Dent
2010;38:987–94.
[11] Asaoka K, Tesk JA. Transient and residual stress in a
porcelain-metal strip. J Dent Res 1990;69:463–9.
[12] Swain MV. Unstable cracking (chipping) of veneering
porcelain on all-ceramic dental crowns and fixed partial
dentures. Acta Biomater 2009;5:1668–77.
[13] Bonfante EA, da Silva NR, Coelho PG, Bayardo-Gonzalez DE,
Thompson VP, Bonfante G. Effect of framework design on
crown failure. Eur J Oral Sci 2009;117:194–9.
[14] Marchack BW, Futatsuki Y, Marchack CB, White SN.
Customization of milled zirconia copings for all-ceramic
crowns: a clinical report. J Prosthet Dent 2008;99:169–73.
[15] Tinschert J, Schulze KA, Natt G, Latzke P, Heussen N,
Spiekermann H. Clinical behavior of zirconia-based fixed
partial dentures made of DC-Zirkon: 3-year results. Int J
Prosthodont 2008;21:217–22.
[16] Choi JE, Waddell JN, Swain MV. Pressed ceramics onto
zirconia, Part 2: Indentation fracture and influence of
cooling rate on residual stresses. Dent Mater 2011;27:1111–8.
[17] Gostemeyer G, Jendras M, Dittmer MP, Bach FW, Stiesch M,
Kohorst P. Influence of cooling rate on zirconia/veneer
interfacial adhesion. Acta Biomater 2010;6:4532–8.
[18] Fischer J, Stawarczyk B, Tomic M, Strub JR, Hammerle CH.
Effect of thermal misfit between different veneering
ceramics and zirconia frameworks on in vitro fracture load
of single crowns. Dent Mater J 2007;26:766–72.
[19] Fairhurst CW, Hashinger DT, Twiggs SW. Glass transition
temperatures of dental porcelain. J Dent Res 1981;60:995–8.
[20] Baldassarri M, Stappert CF, Wolff MS, Thompson VP, Zhang
Y. Residual stresses in porcelain-veneered zirconia
prostheses. Dent Mater 2012;28:873–9.
[21] O’Brien WJ. Dental materials and their selection. Hanover
Park, IL: Quintessence Publishing Co, Inc.; 2008.
[22] Guazzato M, Albakry M, Ringer SP, Swain MV. Strength,
fracture toughness and microstructure of a selection of
all-ceramic materials, Part II. Zirconia-based dental
ceramics. Dent Mater 2004;20:449–56.
[23] Tholey MJ, Swain MV, Thiel N. Thermal gradients and
residual stresses in veneered Y-TZP frameworks. Dent Mater
2011;27:1102–10.
[24] Komine F, Saito A, Kobayashi K, Koizuka M, Koizumi H,
Matsumura H. Effect of cooling rate on shear bond strength
of veneering porcelain to a zirconia ceramic material. J Oral
Sci 2010;52:647–52.
[25] Tan JP, Sederstrom D, Polansky JR, McLaren EA, White SN.
The use of slow heating and slow cooling regimens to
strengthen porcelain fused to zirconia. J Prosthet Dent
2012;107:163–9.
[26] Mainjot AK, Schajer GS, Vanheusden AJ, Sadoun MJ.
Influence of cooling rate on residual stress profile in
veneering ceramic: measurement by hole-drilling. Dent
Mater 2011;27:906–14.
[27] Ivoclar. Programat® EP 3000 – program table; 2009 http://
icubusivoclarvivadentcom/Secure30/media area/ivoclar/C/
Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005
DATA%5Cclient%5Civoclar%5Cmedia%5Cdownloads%5C/
Media 611442002pdf/true/application%5Cpdf/EP3000
progtab cf enpdf
[28] Ivoclar. Programat® EP 3000 – program table; 2010
http://wwwivoclarvivadentcom/zoolu-website/media/
document/4611/Programat+-+Firing+program+tables
[29] Guess PC, Bonfante EA, Silva NR, Coelho PG, Thompson VP.
Effect of core design and veneering technique on damage
and reliability of Y-TZP-supported crowns. Dent Mater
2013;29:307–16.
[30] Scherrer SS, Quinn JB, Quinn GD, Wiskott HW. Fractographic
ceramic failure analysis using the replica technique. Dent
Mater 2007;23:1397–404.
[31] McLean JW. Preparation an the design of metal
substructures. In: Quintessence, editor. Dental ceramics.
Hanover Park: Quintessence; 1983.
[32] Molin MK, Karlsson SL. Five-year clinical prospective
evaluation of zirconia-based Denzir 3-unit FPDs. Int J
Prosthodont 2008;21:223–7.
[33] Schmitter M, Mussotter K, Rammelsberg P, Stober T,
Ohlmann B, Gabbert O. Clinical performance of extended
zirconia frameworks for fixed dental prostheses: two-year
results. J Oral Rehabil 2009;36:610–5.
[34] Huang M, Thompson VP, Rekow ED, Soboyejo WO. Modeling
of water absorption induced cracks in resin-based
composite supported ceramic layer structures. J Biomed
Mater Res B: Appl Biomater 2008;84:124–30.
[35] Okeson J. Alignment and occlusion of the dentition.
Management of temporomandibular disorders and
occlusion. 7th ed. Mosby; 2012. p. 58–80.
[36] Abernethy RB. In: Abernethy RB, editor. The new Weibull
handbook fifth edition, reliability and statistical analysis for
predicting life, safety, supportability, risk, cost and warranty
claims. fifth ed. 2007.
[37] Nicholls D, Lein P. Weibayes testing: what is the impact if
assumed beta is incorrect? Reliability and maintainability
symposium, 2009 RAMS 2009 Annual 2009. 2009. p. 37–42.
[38] Denry I, Kelly JR. State of the art of zirconia for dental
applications. Dent Mater 2008;24:299–307.
[39] Guess PC, Zhang Y, Thompson VP. Effect of veneering
techniques on damage and reliability of Y-TZP trilayers. Eur J
Esthet Dent 2009;4:262–76.
[40] Rekow D, Zhang Y, Thompson V. Can material properties
predict survival of all-ceramic posterior crowns? Compend
Contin Educ Dent 2007;28:362–8, quiz 69, 86.
[41] Tholey MJ, Berthold C, Swain MV, Thiel N. XRD2
micro-diffraction analysis of the interface between Y-TZP
and veneering porcelain: role of application methods. Dent
Mater 2010;26:545–52.
[42] Tholey MJ, Swain MV, Thiel N. SEM observations of porcelain
Y-TZP interface. Dent Mater 2009;25:857–62.
[43] Koenig V, Vanheusden AJ, Le Goff SO, Mainjot AK. Clinical
risk factors related to failures with zirconia-based
restorations: an up to 9-year retrospective study. J Dent
2013;41:1164–74.
[44] Belli R, Frankenberger R, Appelt A, Schmitt J, Baratieri LN,
Greil P, et al. Thermal-induced residual stresses affect the
lifetime of zirconia-veneer crowns. Dent Mater
2013;29:181–90.
[45] Coelho PG, Bonfante EA, Silva NR, Rekow ED, Thompson VP.
Laboratory simulation of Y-TZP all-ceramic crown clinical
failures. J Dent Res 2009;88:382–6.
[46] Mainjot AK, Schajer GS, Vanheusden AJ, Sadoun MJ. Residual
stress measurement in veneering ceramic by hole-drilling.
Dent Mater 2011;27:439–44.
[47] Quinn GD. Fractography of ceramics and glasses.
Washington, DC: National Institute of Standards and
Technology; 2007.
DENTAL-2455; No. of Pages 11
ARTICLE IN PRESS
d e n t a l m a t e r i a l s x x x ( 2 0 1 4 ) xxx–xxx 11

[48] Kim JH, Kim JW, Myoung SW, Pines M, Zhang Y. Damage
maps for layered ceramics under simulated mastication. J
Dent Res 2008;87:671–5.
[49] Rekow D, Thompson V. Near-surface damage-a persistent
problem in crowns obtained by computer-aided design and
manufacturing. Proc Inst Mech Eng H: J Eng Med
2005;219:233–43.
[50] Rekow D, Thompson VP. Engineering long term clinical
success of advanced ceramic prostheses. J Mater Sci Mater
Med 2007;18:47–56.
[51] Wiskott HW, Nicholls JI, Belser UC. Stress fatigue: basic
principles and prosthodontic implications. Int J Prosthodont
1995;8:105–16.
[52] Bhowmick S, Melendez-Martinez JJ, Hermann I, Zhang Y,
Lawn BR. Role of indenter material and size in veneer failure
of brittle layer structures. J Biomed Mater Res B: Appl
Biomater 2007;82:253–9.
[53] Zhang Y, Chai H, Lee JJ, Lawn BR. Chipping resistance of
graded zirconia ceramics for dental crowns. J Dent Res
2012;91:311–5.
[54] Wiskott HW, Belser UC. A rationale for a simplified occlusal
design in restorative dentistry: historical review and clinical
guidelines. J Prosthet Dent 1995;73:169–83.
[55] Myers GE, Anderson Jr JR. Nature of contacts in centric
occlusion in 32 adults. J Dent Res 1971;50:7–13.

Please cite this article in press as: Paula VG, et al. Slow cooling protocol improves fatigue life of zirconia crowns. Dent Mater (2014),
http://dx.doi.org/10.1016/j.dental.2014.10.005