Fracture and Fatigue Resistance of Cemented versus Fused

CAD-on Veneers over Customized Zirconia Implant

Abutments
Shereen Ahmed Nossair, BDS, PhD,1 Moustafa N. Aboushelib, DDS, PhD,1 & Tarek Salah Morsi, BDS,
PhD2
1
Department of Fixed Prosthodontics, Future University, Cairo, Egypt
2
Fixed Prosthodontics Department, Ain Shams University, Cairo, Egypt

Keywords
CAD-on veneers; fractography; fracture
mechanics; fatigue.
Correspondence
Moustafa N. Aboushelib, Dental Biomaterials
Department, Faculty of Dentistry, Alexandria
University, Egypt. E-mail: bluemarline_1@
yahoo.com
This study was supported by Science and
Technology Development Fund (STDF) grant
number 489.
The authors deny any conflicts of interest.
Accepted June 3, 2014
doi: 10.1111/jopr.12253
Abstract
Purpose: To evaluate the fracture mechanics of cemented versus fused CAD-on
veneers on customized zirconia implant abutments.
Materials and Methods: Forty-five identical customized CAD/CAM zirconia im-
plant abutments (0.5 mm thick) were prepared and seated on short titanium im-
plant abutments (Ti base). A second scan was made to fabricate 45 CAD-on veneers
(IPS Empress CAD, A2). Fifteen CAD-on veneers were cemented on the zirconia
abutments (Panavia F2.0). Another 15 were fused to the zirconia abutments using
low-fusing glass, while manually layered veneers served as control (n = 15). The
restorations were subjected to artificial aging (3.2 million cycles between 5 and 10 kg
in a water bath at 37°C) before being axially loaded to failure. Fractured specimens
were examined using scanning electron microscopy to detect fracture origin, location,
and size of critical crack. Stress at failure was calculated using fractography principles
(alpha = 0.05).
Results: Cemented CAD-on restorations demonstrated significantly higher (F = 72,
p < 0.001) fracture load compared to fused CAD-on and manually layered restora-
tions. Fractographic analysis of fractured specimens indicated that cemented CAD-on
veneers failed due to radial cracks originating from the veneer/resin interface. Branch-
ing of the critical crack was observed in the bulk of the veneer. Fused CAD-on veneers
demonstrated cohesive fracture originating at the thickest part of the veneer ceramic,
while manually layered veneers failed due to interfacial fracture at the zirconia/veneer
interface.
Conclusions: Within the limitations of this study, cemented CAD-on veneers on
customized zirconia implant abutments demonstrated higher fracture than fused and
manually layered veneers.

Computerized dentistry is gaining attention, allowing a quick
and thorough diagnosis of clinical problems, offering multiple
solutions for the patient, and allowing fabrication of accurate
and complex restorations with nothing more than a few
keyboard clicks. Early computer-assisted design/computer-
assisted manufacturing systems (CAD/CAM) focused on
fabrication of customized all-ceramic frameworks and copings
that were later veneered using different techniques.1 With
improvements in software and advances in milling potential,
CAD/CAM technology was deployed to fabricate implant
abutments,2 temporary restorations, metallic frameworks,
esthetic veneers, and even larger objects, such as surgical
guides and removable denture bases.3
Meanwhile, the fabrication of CAD/CAM zirconia or metal-
lic frameworks became a totally digital process. Application
of the veneer ceramic using either manual layering or press-on
veneers requires an investment of much time and effort from
the dental ceramist and remains liable to human errors, such
as introduction of structural defects in the veneer material,
occlusal misfits, or shade mismatch, making the final esthetic
outcome difficult to predict. Additionally, manually layered
veneers are associated with creation of internal stresses at the
framework/veneer interface related to mismatch in thermal
expansion coefficient and to cooling stresses. These internal
stresses were recently deemed a probable cause for chipping
failure of the veneer ceramic.4 Additional factors such as

Journal of Prosthodontics 24 (2015) 543–548 

C 2014 by the American College of Prosthodontists 543
Fracture of CAD-on Veneers Nossair et al

milling procedure, polishing, and glazing contribute to this

issue.
In an attempt to digitize the veneering process, CAD/CAM
technology has been used to fabricate CAD-on veneers milled
from a glass ceramic block of the desired shade, translucency,
and strength. The veneers are milled into the required shape and
dimensions to accurately fit the underlying zirconia frameworks
and to establish proper occlusion with the opposing dentition
and neighboring teeth. The software allows accurate control
of the emergence profile of the veneers and of the required
cement space. After milling, the veneers are manually polished
and placed in a ceramic furnace to receive final glaze or external
staining, if required.5
The dentist has the option either to cement these CAD-on ve-
neers on the frameworks in office or to use a low-fusing glass, Figure 1 Student working model demonstrating CAD/CAM zirconia im-
which will require an additional heating cycle. CAD-on veneer- plant abutment seated on short TiBase.
ing in combination with customized zirconia implant abutments
has gained popularity, as the entire restorative process could be
controlled using CAD/CAM technology; however, not much
can found in the dental literature about the long-term perfor-
mance of cemented versus fused CAD-on veneers. The aim of
this study was to evaluate the fatigue and fracture resistance of
cemented versus luted CAD-on veneers in combination with
customized zirconia implant abutments.

Materials and methods

Specimen preparation
A maxillary right first premolar was removed from a student
teaching model to create a space for an implant-supported
restoration. A titanium implant (TSV 4.1 × 12 mm; Zimmer
Dental Inc., Carlsbad, CA) was placed in place of the missing
bicuspid, 2 mm cervical to the gingival margin of the adjacent
teeth. A transfer coping was attached to the implant, and an elas-
tomeric impression was made of the prepared model, poured in
type IV dental stone, and mounted on a fully adjustable articu-
lator. A short Ti abutment (TiBase kit compatible with Zimmer
system) was fastened on the inserted implant. Each TiBase was
supplied with a scan body suitable for single use, transferring
the exact location and inclination of the implant to the design
software. The working model was mounted on the articulator Figure 2 Stereomicroscopic image of cross section of cemented CAD-
with the opposing dentition in centric occlusion. on veneer over customized zirconia implant abutment. Arrow indicates
The model was laser scanned (Cerec 3D 3.88 abutment soft- location of contact with loading indenter.
ware; Sirona Dental Systems, GmbH, Bensheim, Germany),
and 45 customized CAD/CAM zirconia abutments (inCoris Zi, abutments guaranteed a 0.5 to 0.7 mm minimum thickness at
meso block; Sirona Dental Systems, GmbH) were constructed. the screw channel region and 1 mm thick circumferential cervi-
The software starts by designing the abutment into full anatom- cal shoulder. Measurements were performed on cut sections of
ical form. Then the abutment is reduced with a virtual cutback the control specimens (Figs 2 and 3). Software ensured central-
process to accommodate the overlying veneer in a single de- ization of the screw access channel in the zirconia abutments.
signing step. The milling process (Inlab MCXL; Sirona Den- The overlying CAD-on veneers were checked for proper fit us-
tal Systems) is divided into two phases: milling the zirconia ing fit check powder spray and were glazed. For all specimens,
implant abutment (4° tapered angle, equi-gingival finish line, the abutment screw access channel was kept patent by drilling
and 1 mm shoulder) then the CAD-on veneers (IPS Empress through the veneer directly opposing the access channel (Figs
CAD, LT A2; Ivoclar Vivadent, Schaan, Liechtenstein). The ve- 4 and 5).
neers maintained a 1 mm shoulder, 1 mm radial thickness, and The veneers and the zirconia abutments were cleaned in an
1.5 mm minimum occlusal thickness. After milling, the zirco- ultrasonic bath containing 90% ethyl alcohol and randomly
nia abutments were sintered (Sirona inFire HTC; Sirona Den- divided by blinded operator into three groups (n = 15). Fif-
tal Systems) using a computer-programmed (Sirona inCoris teen veneers were etched with hydrofluoric acid (Vita Ceramic
ZI program) electric furnace (Fig 1). Final dimensions of the Etch; Vident, Brea, CA) and coated with a silane coupling agent

544 Journal of Prosthodontics 24 (2015) 543–548 

C 2014 by the American College of Prosthodontists
Nossair et al
Figure 3 Stereomicroscopic image focusing on the cervical region of the
restoration demonstrating precision contract between cemented CAD-
on veneer, the TiBase, and the implant body.
Figure 4 CAD-on veneer prepared to be cemented or fused on
CAD/CAM zirconia implant abutment.
(Monobond S; Ivoclar Vivadent), then directly cemented on the
zirconia abutments using resin cement (Panavia F2.0; Kuraray,
Osaka, Japan) under a constant 500 g load. Excess cement
was wiped off, and an air barrier was applied on the margins.
Low-fusing glass (Fusion Glass; Ivoclar Vivadent) was coated
on the fitting surface of 15 veneers seated on the zirconia abut-
ments and heated in a porcelain furnace (Programat CS; Ivoclar
Vivadent) according to manufacturer recommendations. The
remaining 15 zirconia abutments were manually layered us-
ing porcelain slurry (IPS e.max Ceram, A2; Ivoclar Vivadent).
The CAD-on veneers, seated on the zirconia abutments, were
screwed on the implant (master model) using a torque control
device (35 N/cm).
Journal of Prosthodontics 24 (2015) 543–548 
C 2014 by the American College of Prosthodontists
Fracture of CAD-on Veneers
Figure 5 Access hole used to tighten the fixation screw of the CAD-on
veneered restoration.
Artificial aging
The specimens were mounted in a custom-made pneumatic
cyclic loading machine calibrated to deliver 3.2 million cycles
between 5 and 10 kg under a water bath set at 37°C. A tough
rubber sheet was placed between the specimens and the loading
indenter (4 mm stainless steel) to prevent cone crack generation.
The device maintained a constant load with the restorations
for 2 seconds. Further data about the loading protocol and
precautions against generation of cone cracks are mentioned
elsewhere.6 All specimens were immersed in 1% methylene
blue dye every 10,000 cycles to detect any cracks in the veneer
ceramic.
Fracture strength test
After completion of artificial aging, the specimens were axially
loaded (0.5 mm/min) to fracture in a universal testing machine
(Instron 6022; Instron Limited, High Wycombe, UK). A tough
rubber sheet (1 mm thick) was placed between the loading
indenter (5 mm round stainless steel) and the occlusal surface of
the restorations to prevent stress concentration.6 The load was
delivered vertically at the center of the occlusal surface with
a contact area larger than the access hole made in the center
of the restoration. The test was stopped at the first incidence
of failure as detected by a sudden drop in applied load or if
a cracking sound was observed. Crosshead speed was set to
0.2 mm/min, and a 5000 N load cell was selected. One-way
ANOVA and Bonferroni post hoc tests were used to analyze
the data (alpha = 0.05) using computer software (SPSS 14.0;
SPSS Inc, Chicago, IL).
Fractographic analysis
All fractured parts were collected and heated at 500°C to burn
all traces of resin cement, cleaned in an ultrasonic bath con-
taining 70% ethyl alcohol, dried at 80°C for 60 minutes, gold
sputter coated, and examined under scanning electron micro-
scope (SEM; XL30; Philips, Eindhoven, The Netherlands).7
Fracture origin, critical crack, and fracture landmarks were ob-
served on fracture segments. Critical crack dimensions (crack
depth and width) were measured and were used to calculate
stress at failure using fractography principles.8,9
545
Fracture of CAD-on Veneers
Figure 6 SEM image (75×) demonstrating radial crack originating at
CAD-on veneer/resin interface and propagating in occlusal direction. The
crack front starts to branch into four or five accessory cracks as the main
crack proceeds deeper in the material.
Figure 7 SEM image of the previous specimen (200×) demonstrating
origin of radial crack originating at CAD-on veneer/resin interface.
Results
The statistical test of choice had adequate power (1.0) and ad-
justed beta-squared value (0.99), indicating precision of the
obtained data. All specimens survived the cyclic loading pro-
gram without fracture. There was a significant statistical dif-
ference (F = 73, p < 0.001) in fracture strength between the
three tested groups: cemented CAD-on veneers (1565 ± 105
N), fused CAD-on veneers (1289 ± 103 N), and manually
layered veneers (1152 ± 17 N). Fractographic analysis of frac-
tured specimens indicated that all cemented CAD-on veneers
failed due to radial cracking resulting in splitting of the CAD-
on veneer in two or three fragments without fracture of the
underlying zirconia abutment. The radial crack propagated ver-
tically to cause cohesive fracture of the veneer ceramic (Figs
6 to 8). Fused CAD-on specimens were cohesively fractured
without exposure of the underlying zirconia implant abutment.
The crack front branched as it travelled in an occlusal direction
(Fig 9). One specimen exhibited a minor cohesive fracture of
the underlying zirconia abutment. Manually layered specimens
demonstrated delamination interfacial failure that exposed the
546 Journal of Prosthodontics 24 (2015) 543–548 
Nossair et al
Figure 8 SEM image (150×) demonstrating critical crack dimensions of
cemented CAD-on veneer.
Figure 9 SEM image (100×) demonstrating progress of crack front at fu-
sion glass layer between CAD-on veneer and underlying zirconia coping.
Notice multi-stepping pattern indicating difficulty of fast crack propaga-
tion in this layer.
underlying zirconia implant abutment. Accurate crack mea-
surement was not possible in four specimens due to excessive
damage of the veneer ceramic. The estimated stress at failure
was 123 MPa for cemented CAD-on veneers, 110 MPa for fused
CAD-on veneers, and 98 MPa for manually layered specimens.
Discussion
The interface between the brittle veneer and the much tougher
zirconia is the site where active transmission of occlusal loads
occurs. This interface is also the site for entrapment of inter-
nal stresses generated due to mismatch in thermal expansion
coefficient, uneven cooling of the restoration, and entrapment
of structural defects.10 One of the advantages of CAD-on ve-
neering is elimination of these internal stresses, as the veneer is
manufactured in a stress-free condition and brought as one mass
onto the underlying zirconia coping.5 However, there was a sig-
nificant difference in the fracture strength between cemented
and fused CAD-on veneers.
C 2014 by the American College of Prosthodontists
Nossair et al
For cemented CAD-on veneers, the adhesive layer acted as a
resilient cushion under the brittle veneer, allowing absorption
of more stresses before fracture. Moreover, the cementation
process does not require an additional heating cycle compared
to fused CAD-on veneer where the low temperature fusing glass
is melted between the veneer and the zirconia, a process known
to generate internal stresses as the glass tends to cool down
and shrink in the tight space. The formation of a ceramic resin
hybrid layer was attributed to strengthening surface-damaged
specimens.11 These factors resulted in the significantly higher
fracture load observed for cemented CAD-on veneers, which
were able to withstand higher loads before giving up under
the influence of the vertically advancing radial crack, causing
splitting of the veneer and exposing the zirconia surface. This
fracture pattern protects the underlying implant abutment from
excessive damage and allows simple repair of the fractured
veneer. The estimated stress at failure (123 MPa) matches well
with the higher fracture loads observed for this group.
Fused CAD-on veneers failed due to cohesive fracture origi-
nating within the bulk of the veneer ceramic. A similar finding
was previously reported for the same materials.12 This could
in part be due to entrapment of internal stresses or due to the
pattern of stress distribution within the restoration. In the case
of a fixed partial denture (FPD) restoration, this could lead to a
fracture of the framework.13 However, the estimated stress (110
MPa) at failure is higher than the internal strength of the ma-
terial, suggesting the presence of reinforcing internal stresses,
which could be associated with shrinkage of fusing glass upon
cooling to room temperature. Similar observations and fracture
loads were reported in a previous publication.14 The higher
fracture strength of fused CAD-on veneers compared to man-
ually layered specimens could be related to shrinkage of the
fusion glass, which will put the veneer ceramic in a compres-
sive state. A previous study reported increase in the flexure
strength of glass ceramic materials when coated with a thin
layer of glaze.15 Such a finding explains the strengthening of
the low-fusion glass material used.
Manually layered veneers demonstrated cohesive fracture
within the bulk of the restoration and at the zirconia veneer
interface. Structural defects in the form of air bubbles and in-
ternal flaws were also associated with the fracture site.16 The
estimated stress at failure was close to the internal strength
of the material, which excludes the possible role of internal
stresses as a probable cause for crack initiation.17
Six cases of screw loosening were observed after approx-
imately one million cycles. The screws were properly tight-
ened, and the restorations continued the cyclic loading program.
None of the fabricated restorations failed during cyclic load-
ing, indicating the high reliability of zirconia implant-supported
restorations.18 This in particular was related to the selected
loading protocol, which prevented surface damage of the brit-
tle veneer and maintained the applied load at almost 70% of
the expected failure load of the veneer.19 Several studies that
incorporated fatigue testing of veneered zirconia restorations
did not report significant differences in the performance of the
tested materials.12 Fatigue requires a higher number of loading
cycles to cause sufficient weakening of the tested materials.
A point of concern was axial loading of the specimens af-
ter completion of the cyclic loading program. While many
Journal of Prosthodontics 24 (2015) 543–548 
C 2014 by the American College of Prosthodontists
Fracture of CAD-on Veneers
studies advise against the use of axial loading of brittle ma-
terials because it does not produce clinically relevant fracture
patterns,20,21 this procedure was made to assess the fracture
strength of the tested specimens after simulation of almost 7
years in function considering the mode of force delivery and
distribution in the specimens. It is well established that clini-
cal failure is a parameter of concentration of stress at structural
flaws and not a function of load concentration under the loading
indenter.22,23 However, the failure pattern of the fractured spec-
imens simulated the clinical pattern, as the fracture origin of the
specimen was located at the tensile surface of the specimens.
In this study, all restorations were subjected to artificial aging
then axially loaded to failure. Generation of cone cracks was
prevented by protection of the loading area with a tough plastic
sheet.6 However, under clinical conditions, restorations could
be subjected to higher functional loads and at faster loading
speeds, which could alter stress distribution in the restoration.
Moreover, CAD-on veneering would require a modified con-
nector design for FPD restorations.
Conclusions
Within the limitations of this study, cemented CAD-on veneers
on customized zirconia implant abutments demonstrated higher
fracture resistance compared to fused and manually layered
veneers.
References
1. Rosentritt M, Behr M, Kolbeck C, et al: Marginal integrity of
CAD/CAM fixed partial dentures. Eur J Dent 2007;1:25-30
2. Bergler M, Holst S, Blatz MB, et al: CAD/CAM and telescopic
technology: design options for implant-supported overdentures.
Eur J Esthet Dent 2008;3:66-88
3. Sohmura T, Kusumoto N, Otani T, et al: CAD/CAM fabrication
and clinical application of surgical template and bone model in
oral implant surgery. Clin Oral Implan Res 2009;20:87-93
4. Aboushelib MN, Feilzer AJ, de Jager N, et al: Prestresses in
bilayered all-ceramic restorations. J Biomed Mater Res B Appl
Biomater 2008;87:139-145
5. Wiedhahn K: The impression-free Cerec multilayer bridge with
the CAD-on method. Int J Comput Dent 2011;14:33-45
6. Aboushelib MN: Simulation of cumulative damage associated
with long term cyclic loading using a multi-level strain
accommodating loading protocol. Dent Mater 2013;29:252-258
7. Scherrer SS, Quinn JB, Quinn GD, et al: Failure analysis of
ceramic clinical cases using qualitative fractography. Int J
Prosthodont 2006;19:185-192
8. Kelly JR, Giordano R, Pober R, et al: Fracture surface analysis of
dental ceramics: clinically failed restorations. Int J Prosthodont
1990;3:430-440
9. Mecholsky JJ, Jr: Fractography: determining the sites of fracture
initiation. Dent Mater 1995;11:113-116
10. Nakamura T, Wakabayashi K, Kawamura Y, et al: Analysis of
internal defects in all-ceramic crowns using micro-focus X-ray
computed tomography. Dent Mater J 2007;26:598-601
11. Addison O, Marquis PM, Fleming GJ: Quantifying the strength
of a resin-coated dental ceramic. J Dent Res 2008;87:542-547
12. Tsalouchou E, Cattell MJ, Knowles JC, et al: Fatigue and
fracture properties of yttria partially stabilized zirconia crown
systems. Dent Mater 2008;24:308-318
547
Fracture of CAD-on Veneers
13. Lohbauer U, Amberger G, Quinn GD, et al: Fractographic
analysis of a dental zirconia framework: a case study on design
issues. J Mech Behav Biomed Mater 2011;3:623-629
14. Beuer F, Schweiger J, Eichberger M, et al: High-strength
CAD/CAM-fabricated veneering material sintered to zirconia
copings-a new fabrication mode for all-ceramic restorations.
Dent Mater 2009;25:121-128
15. Cattell MJ, Chadwick TC, Knowles JC, et al: The development
and testing of glaze materials for application to the fit surface of
dental ceramic restorations. Dent Mater 2009;25:431-441
16. Zahran M, El-Mowafy O, Tam L, et al: Fracture strength and
fatigue resistance of all-ceramic molar crowns manufactured
with CAD/CAM technology. J Prosthodont 2008;17:370-377
17. Guess PC, Zhang Y, Thompson VP: Effect of veneering
techniques on damage and reliability of Y-TZP trilayers. Eur J
Esthet Dent 2009;4:262-276
18. Gabbert O, Karatzogiannis E, Ohlmann B, et al: Fracture load of
tooth-implant-retained zirconia ceramic fixed dental prostheses:
548 Journal of Prosthodontics 24 (2015) 543–548 
Nossair et al
effect of span length and preparation design. Clin Oral Implants
Res 2012;23:719-725
19. Wang R, Lu C, Arola D, et al: Plastic damage induced fracture
behaviors of dental ceramic layer structures subjected to
monotonic load. J Prosthodont 2013;22:456-
464
20. Kelly JR, Benetti P, Rungruanganunt P, et al: The slippery slope:
critical perspectives on in vitro research methodologies. Dent
Mater 2012;28:41-45
21. Scherrer SS, de Rijk WG: The fracture resistance of all-ceramic
crowns on supporting structures with different elastic moduli. Int
J Prosthodont 1993;6:462-467
22. Scherrer SS, Kelly JR, Quinn GD, et al: Fracture toughness (KIc)
of a dental porcelain determined by fractographic analysis. Dent
Mater 1999;15:342-348
23. Quinn GD, Studart AR, Hebert C, et al: Fatigue of zirconia and
dental bridge geometry: design implications. Dent Mater
2010;26:1133-1136
C 2014 by the American College of Prosthodontists
Copyright of Journal of Prosthodontics is the property of Wiley-Blackwell and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for
individual use.