In vitro chipping behaviour of all-ceramic crowns with a

zirconia framework and feldspathic veneering: comparison

of CAD/CAM-produced veneer with manually layered
veneer
M. SCHMI TTER, D. MUELLER & S. RUES Section for Biomaterial Research, Department of Prosthodontics,
University of Heidelberg, Heidelberg, Germany
SUMMARY The purpose of this in vitro study was to
assess the breaking load of zirconia-based crowns
veneered with either CAD/CAM-produced or
manually layered feldspathic ceramic. Thirty-two
identical zirconia frameworks (Sirona inCoris ZI,
mono L F1), 06 mm thick with an anatomically
shaped occlusal area, were constructed (Sirona
inLab 3.80). Sixteen of the crowns were then
veneered by the use of CAD/CAM-fabricated
feldspathic ceramic (CEREC Bloc, Sirona) and 16 by
the use of hand-layered ceramic. The CAD/CAM-
manufactured veneer was attached to the
frameworks by the use of Panavia 2.0 (Kuraray).
Half of the specimens were loaded until failure
without articial ageing; the other half of the
specimens underwent thermal cycling and cyclic
loading (12 million chewing cycles, force
magnitude F
max
= 108 N) before the assessment of
the ultimate load. To investigate the new technique
further, nite element (FE) computations were
conducted on the basis of the original geometry.
Statistical assessment was made by the use of non-
parametric tests. Initial breaking load was
signicantly higher in the hand-layered group than
in the CAD/CAM group (mean: 116586 N versus
39545 N). During chewing simulation, however,
875% (7/8) of the crowns in the hand-layered
group failed, whereas no crown in the CAD/CAM
group failed. The CAD/CAM-produced veneer was
signicantly less sensitive to ageing than the hand-
layered veneer.
KEYWORDS: CAD/CAM, ceramic, veneer
Accepted for publication 18 April 2013
Background
Several years ago, zirconia was introduced in
dentistry, enabling the construction of extended all-
ceramic restorations (1). The initial euphoria gener-
ated by this material has, however, decreased as a
result of the large number of veneering failures: both
adhesive (delamination) and cohesive (chipping) fail-
ures occur more often than for metalceramic restora-
tions (2, 3). When failure of the veneer affects
functional (e.g. failure in the area of the approximal
contacts) or aesthetic aspects, especially, repeated
manufacture of the restoration is mandatory. This is
both labour- and time-consuming and, consequently,
a worst-case scenario for both patient and dentist.
Although one reason for this mode of failure was
identied early (the design of the core (4) material
has to support the veneer material) (5), failure of the
veneer on zirconia-based restorations remained higher
than for veneered metallic restorations (3, 6). Later, it
was shown that a modied ring procedure with
additional cooling time resulted in greater fracture
resistance of the veneered zirconia restorations (7).
Although this approach improved the clinical perfor-
mance of veneered zirconia restorations, other
approaches have also been developed.
2013 John Wiley & Sons Ltd doi: 10.1111/joor.12061
Journal of Oral Rehabilitation 2013 40; 519--525
J o u r n a l o f
Oral Rehabilitation
One approach is the computer-aided design and
computer-aided manufacturing (CAD/CAM) of both
zirconia framework and veneer (8), resulting in
almost faultless components, because the ceramic
blanks are produced industrially. The parts of the res-
toration are joined together by the use of resin
cement. Although, in a recent study (9), the perfor-
mance of zirconia molar crowns with lithium disilicate
veneering was assessed, little information is available
about the behaviour of these restorations when CAD/
CAM-produced feldspathic veneer is used.
In this study, two groups of zirconia-based crowns
with feldspathic veneer were investigated by the use
of in vitro experiments. The hypothesis of the study
was that zirconia-based crowns veneered with CAD/
CAM-produced feldspathic ceramic would have a
higher breaking load than conventionally veneered
zirconia crowns.
Methods
Test series
With regard to chipping and/or delamination of the
veneer, the performance of all-ceramic molar crowns
fabricated with the new CAD/CAM technique was of
interest. A group of conventionally manufactured
crowns with manually layered veneer was tested as
reference, resulting in two test series:
1 Multilayer technology.*
2 Conventional layering technique (without modied
cooling phase).
Manufacture of the crowns
All specimens were produced by one trained techni-
cian (D.M.) to avoid any operator effect.
Manufacture of the zirconia framework. A gypsum cast
containing a prepared molar was fabricated and the
prepared die was duplicated 16 times in CoCrMo alloy
by selective laser melting.
For these dies, 32 identical zirconia frameworks
(mono L F1; Sirona inCoris ZI*), 06 mm thick with
an anatomically shaped occlusal area, were con-
structed (Sirona inLab 3.80*). The frames were then
milled (Sirona inLab MC XL*, Fig. 1a) and sintered.

After sintering, all frameworks were checked for

aws, etc., by light microscopy.

Also, for each frame,

the thickness was measured with a calliper at ve
previously dened points.
Construction of the veneer. After construction of the
veneer by the use of inLab 3.80 (Sirona), the mesio-
lingual cusp was modied with Rapidform Dental
software

as follows: (i) constant 30 angle between

the occlusal surface of the cusp and the tooth axis
and (ii) construction of a mould (radius: 30 mm,
depth: 01 mm) dening the loading site for mechani-
cal testing. The thickness of the veneer at the mesio-
lingual cusp was approximately 20 mm.
Multilayer technology. After construction and modica-
tion of the veneer, it was milled from feldspathic cera-
mic (CEREC Bloc). All veneer was wetted with glaze
and then red (Ivoclar Programat 700; details are
given in Table 1). After sintering, the thickness of the
veneer was checked, and its inner surface and the
framework were sand-blasted (50 lm/2bar). Finally,
the CAM-manufactured veneer was attached to the
framework by the use of Panavia F 2.0 (Fig. 1b).
Conventional layering technique. In this group, the
framework was veneered manually. The powder-to-
liquid ratio was 028 g to 012 mL. Two bonder r-
ings, two dentine rings and one glaze ring were
performed (details are given in Table 1). To produce
crowns comparable with those described above, sec-
tions of a silicone impression taken from completely
CAD/CAM-produced crowns (Section Multilayer tech-
nology) were used to monitor the shape of the hand-
made veneer.
In vitro tests
The purpose of this study was to assess adhesive fail-
ure at the frame/veneer interface and/or cohesive fail-
ure within the veneer ceramics of molar crowns.
Thus, a suitable load had to be applied to achieve this
mode of failure. During all mechanical tests, loads
*CEREC Bloc; Sirona, Bensheim, Germany.

Vita ZYrcomat; Vita Zahnfabrik, Bad Saeckingen, Germany.

STEMI SR; Carl Zeiss, Oberkochen, Germany.

INUS Technology, Seoul, Korea.

2013 John Wiley & Sons Ltd
M. S C HMI T T E R et al. 520
were applied to the mesio-lingual cusp at an angle of
30 to the direction of insertion (Fig. 2a). The speci-
mens were embedded in the steel mould at an angle
of 30 to the direction of insertion, because vertical
loading during chewing simulation and fracture tests
was mandatory. This placement could be exactly
reproduced by the use of a special metal key orienting
the metal dies in the desired direction during the use
of a parallelometer table. The metal dies were embed-
ded in steel moulds with cold-curing resin.

The
crowns were luted to the metal dies with Panavia F
2.0.** Some important properties of this material are
given in Table 2. In vitro testing started after at least
1 day of hardening of the resin cement.
For each of the two techniques, two series of tests
(sample size n = 8 for each series) were conducted.
First, initial fracture resistance

directly after manu-

facture was tested. Second, performance during a sim-
ulated ageing process consisting of 10 000 thermal
cycles (65 C/60 C, dwell time 45 s

; Willytec THE-
1100) and 12 million chewing cycles (force magni-
tude F
max
= 108 N, water storage; Willytec CS3

) was
assessed. Finally, the remaining fracture resistance
was evaluated. For both chewing simulation and frac-
ture tests, steel spheres 6 mm in diameter were used
as antagonists. The arrangement for the fracture tests
is illustrated in Fig. 2a.
After each step, that is, manufacture, thermal
cycling and chewing simulation, the crowns were
inspected under a microscope to detect the formation
of aws or cracks which might have occurred.
Statistics
Descriptive statistics were used to analyse the means,
standard deviations and medians.
To assess the effect of the different veneer on
fracture load, non-parametric tests were used
(KruskalWallis test and MannWhitney U-test with
a < 005 as signicance level). A chi-squared test was
used to assess differences between survivals during
articial ageing. The results were depicted by the use
of whiskerboxplots.
Finite element computations
Because the stiffness of the interface layer between
frame and veneer will affect the distribution of stress,
(a) (b)
Fig. 1. (a) frame milled from
zirconia. (b) complete crown after
fusion by the use of Panavia F2.0.
Table 1. Firing procedures used for manufacture of the crowns
Firing
Pre-heating
temp. [C]
Dry phase
[min]
Temp.
increase
[C min
1
]
Final
temp. [C]
Hold time
[min]
Vac. on
[C]
Vac. off
[C]
Conventional veneering porcelain
Bonder 500 6 80 980 2 500 980
Dentin 1 500 6 55 910 1 500 910
Dentin 2 500 6 55 900 1 500 900
Glaze 500 4 80 900 1

Technovit 4071; Heraeus Kulzer GmbH, Hanau, Germany.

**Kuraray Medical Inc., Osaka, Japan.

Zwick/Roell Z005; Zwick/Roell AG, Ulm, Germany.

SD Mechatronik, Feldkirchen, Germany.

2013 John Wiley & Sons Ltd
C A D / C A M - P R O D U C E D V E R S U S M A N U A L L Y L A Y E R E D V E N E E R S 521
nite element computations

were performed with

different interface materials based on the original
geometry (Fig. 2b). The Youngs modulus of the inter-
face layer was varied, that is, E = 35 GPa, E = 186
GPa (range for resin cements) and E = 70 GPa (no
interface layer properties of conventional veneering
porcelain). The properties of all the linearelastic
materials are listed in Table 2. The FE model consisted
of approximately 65 000 20-node hexahedral ele-
ments and 240 000 nodes. The CoCr stump was
assumed to be rigid, that is, all displacements of nodes
located on the inner surface (stump/cement interface)
were restricted. As a consequence, the stump was not
included in the computations. Loading took place by
contact (area-to-area contact formulation) with a steel
sphere (6 mm diameter) analogous to loading in the
in vitro tests (Fig. 2b). The nal load (F = 600 N) was
chosen in accordance with the highest ultimate loads
recorded for crowns veneered with CEREC Bloc shells
(Fig. 3). Thermal stress because of the manufacturing
process was not of interest in this analysis and was
therefore not included.
Results
Tests without articial ageing initial fracture resistance
The ultimate load values recorded (mean value and
standard deviation) corresponding to fracture were as
follows:
crowns veneered with CEREC Bloc shells (n = 8):
F = 395 96 N
conventionally veneered crowns (n = 7):
F = 1166 189 N
Fig. 3. Ultimate loads for crowns without ageing and for those
surviving thermal cycling and chewing simulation.
(a)
(b)
Fig. 2. (a) Schematic diagram of sample orientation and loading
during chewing simulation and fracture tests. (b) Section
through the nite element model with formulation of the con-
tact between sphere and crown.
Table 2. Properties of the materials used in the FE analysis
Material
Youngs
modulus [GPa]
Poissons
ratio []
Cerec Bloc* 70 021
Panavia F2.0 186 028
Multilink Implant* 35 035
Zirconia* 210 026
*According to information provided by the manufacturer.

ANSYS 13.0 ; ANSYS Inc., Canonsburg, PA, USA.

2013 John Wiley & Sons Ltd
M. S C HMI T T E R et al. 522
When not aged, conventionally veneered crowns
had signicantly higher fracture resistance (Mann
Whitney U-test, P < 0001) than crowns veneered
with CEREC Bloc shells (Fig. 3).
Tests with articial ageing
Chewing simulation. During chewing simulation, no
specimen in the CEREC Bloc group failed; 875% (i.e.
7 of 8) conventionally veneered crowns failed, how-
ever (Fig. 4). Because all failures occurred after a
rather short period (less than one-tenth of the com-
plete chewing simulation), the difference between the
two groups was signicant (chi-squared test;
P < 0001).
Remaining fracture resistance. For crowns surviving
thermal cycling and chewing simulation, force values
recorded in the fracture tests were as follows:
crowns veneered with CEREC Bloc shells (n = 8):
F = 437 N 35 N
conventional veneered crowns (n = 1): F = 934 N
Because one of the groups contained one specimen
only, no statistical assessment was possible (Fig. 3).
Effect of articial ageing. On the basis of the results
presented above, crowns with CAD/CAM-produced
veneer do not seem to be affected by articial ageing,
whereas almost all manually layered crowns failed
during chewing simulation (Figs 3 and 4).
Effect of the stiffness of the interface layer
Tensile stress is of most interest when brittle materials
fracture. As is apparent from Fig. 5a, tensile stress
within the veneer was highly affected by the stiffness
of the interface layer. Along the inner surface of the
veneer, maximum tensile stress within the veneer
material decreased with increasing Youngs modulus,
E
int
, of the interface layer: Whereas a load of
F = 600 N caused high tensile stress for E
int
= 35 GPa
(r
I,max
= 155 MPa), moderate stress (r
I,max
= 83 MPa)
was recorded for E = 186 GPa, and low stress
(r
I,max
= 30 MPa) correlated with conventional
veneer, that is, E = 70 GPa.
Discussion
This study revealed that zirconia frameworks veneered
with CAD/CAM-produced feldspathic ceramic are less
sensitive to ageing than zirconia crowns with layered
feldspathic veneer.
In recent years, it has become obvious that cohesive
and adhesive failures occur more often for zirconia-
based, veneered restorations (3, 10). Several
approaches have therefore been assessed as possible
means of reducing this higher occurrence of failure,
including optimisation of the ring procedure (7),
framework design (4), the fabrication process (11)
and the use of pressed veneer ceramic (12). Although
these approaches have reduced the incidence of chip-
ping, additional effort is necessary to improve the per-
formance of these restorations further. One possible
approach could be use of CAD/CAM technology for
production of both the veneer and the framework.
Furthermore, CAD/CAM-fabricated veneer could
complete the digital production of dental restorations.
It might, thus, be the optimum approach for solution
of the aforementioned problems.
Fabrication of veneer from industrially produced
ceramic blanks with fewer faults than manually pro-
duced veneer (12) is believed to optimise fatigue
behaviour and to increase fracture strength compared
with hand-layered ceramics (2). If composite resin is
used to lute the framework to the veneer, however,
intra-oral conditions (saliva, mechanical loading and
Fig. 4. Survival of veneered crowns [%] during mechanical
articial ageing.
2013 John Wiley & Sons Ltd
C A D / C A M - P R O D U C E D V E R S U S M A N U A L L Y L A Y E R E D V E N E E R S 523
thermal stress) (13) might result in clinical failure.
Furthermore, discoloration of the composite resin that
lls the gap between the framework and the veneer
could occur with time (14).
All specimens in this study were produced without
the use of a modied ring procedure. Although use
of an optimised ring procedure has a positive effect
on breaking load (7), this approach was chosen for
two reasons: rst, this procedure was commonly used
for all-ceramic restorations until a few years ago and
problems originating from this manufacturing tech-
nique resulted in the development of the new
approaches presented. Second, many technicians do
not use optimised ring programmes, because many
furnaces do not support this option. Pre-treatment of
the specimens (sand-blasting) before attaching the
feldspathic ceramic veneer to the core is another
aspect of this study that should be critically discussed.
This approach was chosen because the use of hydro-
uoric acid in dentistry is controversial (15) it is a
contact poison with the potential for deep, initially
painless, burns, with later tissue death (16). Sand-
blasting of the surface also results in acceptable bond-
ing to composite resin (17).
Clinically, chipping is often observed after the resto-
ration has been in situ for some time, that is, ageing
has occurred. Fracture of the veneer immediately after
the insertion is rare. Thus, lack of sensitivity to ageing
seems to be an essential issue in this context. The ini-
tial fracture load should, however, exceed expected
chewing forces; in this study, this was achieved for
both techniques (the lowest value (285 N) was mea-
sured for CAD/CAM veneer). The bite force is nor-
mally transferred to several cusps; consequently, the
magnitude of our breaking test force corresponds to
even higher bite forces (e.g. four cusps with 30 incli-
nation: 4 9 285N 9 cos 30 = 987 N).
In this study, fracture load for some specimens was
higher after ageing than before. There might be two
explanations of this, at rst glance, contradictory
result. First, the industrially manufactured veneer is
insensitive to ageing and the higher fracture load is
an effect of the small sample size. Second, wear
occurred during ageing at the loading point optimising
t of the sphere used during subsequent ultimate load
testing.
The results of this study also showed that hand-
layered zirconia crowns have higher initial fracture
resistance than CAD/CAM-produced feldspathic
veneer. This, at rst glance surprising, result might
become explicable when the mechanical properties of
the interface material are taken into consideration. As
the FE analysis revealed, tensile stress equalling the
materials strength (150 MPa for CEREC Bloc) was
reached at the inner surface of the veneer for a soft
resin cement when the crown was loaded with
F = 600 N, which would correlate perfectly with the
in vitro test results. With use of Panavia, however,
and the values given in the literature for its Youngs
modulus, fracture resistance approximately twice as
high would be expected (if the fracture still starts at
the inner veneer surface). The authors know of no
composites with E > 18 GPa, and E = 18 GPa can be
achieved only by the use of a high percentage of inor-
ganic ller, macroparticles and dual curing; the value
E = 186 GPa might be too high for Panavia.
To be able to show differences between the rather
small groups in the in vitro tests, a relatively high
Fig. 5. Maximum tensile stress r
I
for interface layers of different Youngs moduli between frame and veneer.
2013 John Wiley & Sons Ltd
M. S C HMI T T E R et al. 524
force magnitude, F
max
= 108 N, was chosen for the
chewing simulation. Because common bite forces are
most often transferred to several cusps, our cyclic test
force magnitude corresponds to even greater bite
forces. If all four cusps (assumption: 30 angle for all
cusps) are loaded, our test set-up simulated a vertical
bite force of 4 9 108 N 9 cos30 = 374 N. Hence,
depending on the number of loaded cusps, corre-
sponding bite forces in the range 108 to 374 N were
simulated. In a recent study, it was shown by nite
element analysis that the stumps stiffness had almost
no effect on breaking load in the study design investi-
gated due to the shielding effect of the zirconia
frame (9). The same design was used in our study
and, consequently, use of metal dies seems to be ade-
quate and might not bias the results.
To summarise, the hypothesis of this study, that
CAD/CAM-produced veneer has a higher breaking load
than manually layered veneer, had to be rejected.
CAD/CAM-fabricated veneer was, however, far less
sensitive to ageing than manually layered veneer.
Conclusions
The main ndings of this study were:
1 Manually veneered crowns (fast cooling) have high
initial fracture load but are very sensitive to ageing.
2 Attachment of CAD/CAM-fabricated feldspathic
ceramic veneer to a zirconia frame by the use of
resin cement might be associated with lower break-
ing load but with less sensitivity against ageing.
Conict of interest
The authors declare that they have no conict of
interest.
References
1. Schmitter M, Mussotter K, Rammelsberg P, Stober T,
Ohlmann B, Gabbert O. Clinical performance of extended
zirconia frameworks for xed dental prostheses: two-year
results. J Oral Rehabil. 2009;36:610615.
2. Guess PC, Zavanelli RA, Silva NR, Bonfante EA, Coelho PG,
Thompson VP. Monolithic CAD/CAM lithium disilicate versus
veneered Y-TZP crowns: comparison of failure modes and
reliability after fatigue. Int J Prosthodont. 2010;23:434442.
3. Silva NR, Bonfante EA, Zavanelli RA, Thompson VP, Ferencz
JL, Coelho PG. Reliability of metalloceramic and zirconia-
based ceramic crowns. J Dent Res. 2010;89:10511056.
4. Rosentritt M, Steiger D, Behr M, Handel G, Kolbeck C.
Inuence of substructure design and spacer settings on the
in vitro performance of molar zirconia crowns. J Dent.
2009;37:978983.
5. Lorenzoni FC, Martins LM, Silva NR, Coelho PG, Guess PC,
Bonfante EA et al. Fatigue life and failure modes of crowns
systems with a modied framework design. J Dent.
2010;38:626634.
6. Heintze SD, Rousson V. Survival of zirconia- and metal-
supported xed dental prostheses: a systematic review. Int J
Prosthodont. 2010;23:493502.
7. Rues S, Kroger E, Muller D, Schmitter M. Effect of ring
protocols on cohesive failure of all-ceramic crowns. J Dent.
2010;38:987994.
8. Beuer F, Schweiger J, Eichberger M, Kappert HF, Gernet W,
Edelhoff D. High-strength CAD/CAM-fabricated veneering
material sintered to zirconia copingsa new fabrication mode
for all-ceramic restorations. Dent Mater. 2009;25:121128.
9. Schmitter M, Mueller D, Rues S. Chipping behaviour of all-
ceramic crowns with zirconia framework and CAD/CAM
manufactured veneer. J Dent. 2012;40:154162.
10. Sailer I, Pjetursson BE, Zwahlen M, Hammerle CH. A
systematic review of the survival and complication rates of
all-ceramic and metal-ceramic reconstructions after an
observation period of at least 3 years. Part II: xed dental
prostheses. Clin Oral Implants Res. 2007;18(Suppl 3):8696.
11. Rosentritt M, Kolbeck C, Handel G, Schneider-Feyrer S,
Behr M. Inuence of the fabrication process on the in vitro
performance of xed dental prostheses with zirconia sub-
structures. Clin Oral Investig. 2011;15:10071112.
12. Christensen RP, Ploeger BJ. A clinical comparison of
zirconia, metal and alumina xed-prosthesis frameworks
veneered with layered or pressed ceramic: a three-year
report. J Am Dent Assoc. 2010;141:13171329.
13. Cavalcanti AN, Foxton RM, Watson TF, Oliveira MT,
Giannini M, Marchi GM. Bond strength of resin cements to
a zirconia ceramic with different surface treatments. Oper
Dent. 2009;34:280287.
14. Molin MK, Karlsson SL. A randomized 5-year clinical evalu-
ation of 3 ceramic inlay systems. Int J Prosthodont.
2000;13:194200.
15. Loomans BA, Mine A, Roeters FJ, Opdam NJ, De Munck J,
Huysmans MC et al. Hydrouoric acid on dentin should be
avoided. Dent Mater. 2010;26:643649.
16. Wu ML, Deng JF, Fan JS. Survival after hypocalcemia,
hypomagnesemia, hypokalemia and cardiac arrest following
mild hydrouoric acid burn. Clin Toxicol (Phila).
2010;48:953955.
17. Kato H, Matsumura H, Atsuta M. Effect of etching and
sandblasting on bond strength to sintered porcelain of
unlled resin. J Oral Rehabil. 2000;27:103110.
Correspondence: Marc Schmitter, Section for Biomaterial Research ,
Department of Prosthodontics, University of Heidelberg, 69120
Heidelberg, Germany.
E-mail: Marc_Schmitter@med.uni-heidelberg.de
2013 John Wiley & Sons Ltd
C A D / C A M - P R O D U C E D V E R S U S M A N U A L L Y L A Y E R E D V E N E E R S 525