dental

materials
Dental Materials 18 (2002) 12±19
www.elsevier.com/locate/dental

Fracture toughness of dental ceramics: comparison of bending and

indentation method
Horst Fischer*, Rudolf Marx
Department of Dental Prosthetics, Section of Dental Materials, University of Technology, Aachen, Germany
Received 18 April 2000; revised 9 October 2000; accepted 11 January 2001

Abstract
Objectives: The aim of this study was to compare two fracture toughness methods, the bending method on notched specimens as reference
and the indentation method as comparison. Potentialities and limitations of the indentation method were analyzed.
Methods: Fracture toughness values were determined for seven dental ceramic materials on `single-edge-v-notched beams' (bending
method) as our standard method. Additionally indentation tests were done on the identical samples from the bending tests. The results were
determined before and after annealing of the samples. With the reference fracture toughness results from the bending tests, the prefactor in
the indentation test formula was individually adapted for each tested material.
Results: The individual prefactors varied between 0.0122 and 0.0253 for the specimens before annealing and between 0.0150 and 0.0267
for the annealed specimens. Subsequently the differences between the KIc-values calculated by the ISO draft (TC 206) direction and
calculated by the modi®ed formulae with the material speci®c prefactors were up to 48% for unannealed and up to 33% for annealed
specimens, respectively.
Signi®cance: The indentation method is not an adequate tool to exactly determine the fracture toughness of an unknown ceramic material.
This method can only be used for a ®rst rough KIc estimation. q 2002 Academy of Dental Materials. Published by Elsevier Science Ltd. All
rights reserved.
Keywords: Dental ceramics; Fracture toughness; Bending method; Indentation method

1. Introduction the fracture toughness values. One method is the test on

As patients have become more and more demanding
regarding the esthetic and biocompatibility aspects of their
dental restorations, ceramic as a material for inlays, onlays,
crowns and bridges has become a main goal of scienti®c
interest especially from the material point of view. One
problematic aspect of ceramic materials in general and of
dental ceramics in particular is their low fracture toughness
(KIc) [1]. The lower the fracture toughness the lower is the
clinical reliability of the ceramic restoration because the KIc
value de®nes the critical stress intensity level at which cata-
strophic failure occurs due to a (critical) micro defect.
Therefore, every new dental ceramic material should be
tested not only with respect to its ¯exural strength but also
with respect to its fracture toughness before introducing into
the market.
Different test methods have been established to evaluate
* Corresponding author. Tel.: 149 241 8089896; fax: 149 241 8089896.
E-mail address: h.®scher@rwth-aachen.de (H. Fischer).
0109-5641/02/$22.00 + 0.00 q 2002 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved.
PII: S 0109- 564 1( 01) 00005- 7
`single-edge-v-notched beams' (SEVNB) [2±4], named
`bending test' further on in this paper. In this test bar speci-
mens with a v-notch of ®nite width introduced by a saw cut
and sharpened by a razor-blade and oil lubricated diamond
paste [5,6] are loaded in a four-point bending test arrange-
ment. The fracture toughness value can be evaluated from
the maximum load and the dimensions of the specimen and
the notch. One of the advantages of this test consists in the
small number of specimens which are needed for good
statistics. A fracture toughness round robin test showed
that the bending method having its simplicity in mind
gives very accurate and true toughness values which can
be highly reproduced [7].
The fracture toughness can also be determined by analyz-
ing indentation cracks, for example Vickers indentation
cracks [8]. During loading and unloading of a Vickers
indenter two perpendicular cracks are initiated starting at
the deepest location of the deformation zone which propa-
gate to the material's surface. From theoretical considera-
tions it follows that the KIc value can be calculated as a
 H. Fischer, R. Marx / Dental Materials 18 (2002) 12±19 13

Table 1
Annealing temperatures of batch 2, fracture toughnesses (KIc) evaluated by the bending test, Young's moduli evaluated by resonance frequency method and KIc
values calculated on the basis of the results of the indentation method using the constant prefactor j ˆ 0.018 given in ISO XXXXX draft (TC 206)

Ceramic material Annealing Fracture toughness Young's modulus Fracture toughness indentation method KIc (MPam 0.5) a
temperature bending method KIc E (GPa)
q (8C) (MPam 0.5)
Unannealed Batch 1 Annealed Batch 2

Cerec 685 1.36 (0.05) 69 1.56 1.07

Duceram Opaker 440 1.45 (0.05) 73 1.76 1.57
Empress 1 525 1.18 (0.08) 67 1.74 1.42
Empress 2 435 2.48 (0.22) 96 1.76 1.67
Hydroxyapatite (HA) 600 0.90 (0.07) 100 b 0.93 0.88
In-Ceram Alumina Celay 600 5.00 (0.39) 251 6.00 4.35
Vita Omega Opaker 475 1.34 (0.07) 66 1.49 1.24
a
KIc values based on a constant prefactor j ˆ 0.018 as recommended in ISO XXXXX draft (TC 206).
b
The Young's modulus for hydroxyapatite (HA) was taken from the literature [20].

function of crack length at the surface, hardness and examination the potentialities and limitations of the inden-
Young's modulus of the tested material [9]. The advantage tation method were analyzed.
of this test, named `indentation test' further on in this paper,
is the simple test method itself. No standard test specimens
have to be prepared to evaluate the KIc-value. This can help
2. Materials and methods
to save time and costs. The Vickers diamond indenter theo-
retically can be set on every ceramic component no matter
The tests were done exemplarily on seven dental ceramic
what the geometry of the component is. Only a polished and
materials: Cerec (type Mark II, Vita, Bad SaÈckingen, FRG),
locally ¯at surface (less than 1 mm 2) is required. Therefore,
Duceram Opaker (Ducera, Rosbach, FRG), Empress 1
the indentation test seems to be very suitable, in particular
(OpazitaÈt OII) and Empress 2 (both Ivoclar, Schaan, FL),
for the dental ®eld, because fracture toughness measure-
HA (high density hydroxyapatite, Osprovit, CeramTec,
ments could be done directly on ®nished components, i.e.
Plochingen, FRG), In-Ceram Alumina Celay and Vita
on ceramic inlays, crowns and bridges for example.
Omega Opaker (both Vita, Bad SaÈckingen, FRG).
However, the problem of the indentation test is the calcu-
For the bending test n ˆ 8 bar specimens
lation of the fracture toughness value from the raw data.
(3.0 £ 6.0 £ 30.0 mm) of each tested material were manu-
Different authors have worked out different rules for
factured according to the company's recommendations. The
computing [10±12]. The formula to calculate KIc from the
surfaces of the specimens were ground on a rotation ground
characteristic pattern after indentation given in ISO
machine (Typ AW-10, Exakt Apparatebau, Norderstedt,
XXXXX draft (TC 206) [13] was stated for the ®rst time
FRG). Silicon nitride ground papers (grit 800, 1200, 2400
in the early 1980s [14]. The main problem of the rules for
and 4000, Struers, Willich, FRG) were used for the grinding
computing given in the ISO draft is a prefactor j inserted in
process. The ground bar specimens were notched. The ratio
the formula. A constant value of j ˆ 0.018 is given for that
between notch depth and specimen thickness was 0.5. The
prefactor. The hypothesis of this study is that this prefactor
notches were sharpened by the razor blade method as
is not a universal constant and depends on the speci®c char-
described above so that the tip of the notch had a width of
acteristics of the individual ceramic material. Consequently,
less than 30 mm. The specimens were loaded in a four-
individual, material depending prefactors, i.e. individual
point-bending test arrangement. The speed of the cross-
rules for computing have to be de®ned to calculate the
head of the testing machine (Z030, Zwick, Ulm, FRG)
true fracture toughness value KIc for each ceramic material.
was 1 mm/min. The fracture toughness was calculated
The width of range of these individual prefactors can be
using the following equation [15]:
taken as a criterion for the reliability of the indentation
method. The wider the range is, the more inaccurate is a r
d
calculated fracture toughness value using the universal 3´G M
Lmax lo 2 li h
prefactor given in the ISO XXXXX draft (TC 206). K Ic ˆ p ´ ´  3 …1†
For this study fracture toughness values of dental ceramic t h h d 2
2´ 1 2
materials were determined by the bending method as refer- h
ence. Using these reference fracture toughness results the
individual prefactors in the indentation test formula were Where Lmax is the maximum load, lo and li are the outer and
determined for the ceramic materials tested. Based on this inner roll spans, respectively, t and h are thickness and
14 H. Fischer, R. Marx / Dental Materials 18 (2002) 12±19
height of the specimens, d is the depth of the notch and
d ments were recorded within 1 min of indentation.
G M ˆ 1:9887 2 1:326
h Furthermore, Vickers indentations were done at ®ve
"  2 #   different loads with n ˆ 5 indentations per load and batch.
d d d d
3:49 2 0:68 1 1:35 12
h h h h
2  2 …
d ing. Loads from 1 to 5 kg (i.e. 1 £ 9.81 to 5 £ 9.81 N) in
11
h steps of 1 kg were used for Cerec, Duceram Opaker,
The two halves of the broken samples of the bending test
were subsequently utilized for the indentation test. The
surfaces of the (ground) halves which were used for the
indentation test were additionally polished before being
loaded by the Vickers indenter by a diamond paste (Diapast
883970, Girrbach, Pforzheim, FRG) in order to get sharp
imprints.
Two batches of specimens were tested by the indentation
test. The ®rst batch of polished broken halves from the
bending test were loaded by the Vickers indenter without
any other pretreatment. The polished halves of the second
batch were additionally annealed in a ceramic oven (Austro-
mat 3001, Dekema, Freilassing, FRG) in order to get the
surfaces free of residual stresses which were possibly intro-
duced by the grinding and polishing process. The annealing
time was 10 h. The annealing temperature for the ceramic
materials Cerec, Duceram Opaker, Empress 1, Empress 2
and Vita Omega Opaker was 100 K below the glass transi-
tion temperature of the materials, respectively. For In-
Ceram Alumina Celay the annealing temperature was
100 K below the glass transition temperature of the La-
glass. The values for the glass transition temperatures of
the ceramic materials came from the manufacturers, respec-
tively. The annealing temperature of hydroxyapatite (HA)
was 6008C. In a pilot study it was found that at higher
annealing temperatures alterations of the HA surface
occurred. The annealing temperatures applied for the
different ceramic materials are listed in the second column
of Table 1. The specimens were heated up slowly to the
annealing temperature within 2 h. After 10 h of the anneal-
ing process the specimens were slowly cooled down to room
temperature over night in the closed oven.
In a ®rst series of measurements the effect of time of
loading on the crack length and on the indentation diagonal
length was examined. These results were intended to
con®rm the loading time of 15 s recommended in ISO
XXXXX draft (TC 206). For the indentation tests Vickers
indentations were set with a hardness tester (Typ 3212001,
Zwick, Ulm, FRG) at a constant load of 3 kg (i.e.
3 £ 9.81 N) for 1, 5, 10, 15, 30 and 60 s (n ˆ 10 indentations
for each loading time) on a Duceram Opaker specimen of
batch 2 (annealed sample). The crack lengths and the diag-
onal lengths for this and the following series were measured
from the pattern of the impressions with a light-optical
microscope (magni®cation: 200 £ ). The microscope was
integrated into the hardness tester. The diagonal length 2a
and the crack length 2c were calculated from the arithmetic
means of 2a1, 2a2 and 2c1, 2c2, respectively. The measure-
The criterion for the minimum load for each material was
that a sharp, regular imprint could be detected after unload-
2†
Empress 1 and Empress 2 and for Vita Omega Opaker.
Loads from 0.5 to 1.5 kg in steps of 0.25 kg were used for
HA. Loads from 5 to 9 kg in steps of 1 kg were used for In-
Ceram Alumina Celay. The loading time for all of these
measurements was 15 s, respectively.
The formula to calculate the fracture toughness is given in
the ISO XXXXX draft (TC 206) as follows:
r
E P
KIc ˆ j´ ´ …3†
H c 1:5
where E is Young's modulus, P is the indenter load, c is the
crack length, j is the prefactor and H is the hardness. H can
be calculated as follows:
 
1368 P P
H ˆ 2´sin ´ 2
ˆ 1:8543677´ …4†
2 …2a† …2a†2
where 2a is the length of the diagonal of the imprint.
To calculate the prefactor j for each tested material, Eq.
(3) was transformed to:
K
j ˆ r Ic …5†
E P
´
H c1:5
H, P and c were measured and calculated from the imprints
of the indentation examinations and the values for KIc
were taken from the results of the (standard) bending test,
respectively.
The Young's moduli of the tested ceramic materials
except for hydroxyapatite were evaluated on rectangular
specimens (3.0 £ 9.0 £ 40.0 mm) by the resonance
frequency method [16] (apparatus type MK4-1, Grindo-
sonic, KoÈln, FRG) according to DIN V ENV 843-2 [17]
as this is a test method with high accuracy for brittle materi-
als [18,19]. The Young's modulus value of hydroxyapatite
was taken from the literature [20].
3. Results
In the third column of Table 1 the fracture toughness
values (KIc) evaluated by the bending test are listed. The
mean values and the standard deviations were calculated
from n ˆ 8 specimens for each material, respectively. The
standard deviations were less than 9% of the mean values in
all cases. HA shows the lowest fracture toughness
(0.90 MPam 0.5). The highest KIc values were measured for
In-Ceram (5.0 MPam 0.5) and Empress 2 (2.48 MPam 0.5).
 H. Fischer, R. Marx / Dental Materials 18 (2002) 12±19
The results from the ®rst series of the indentation test are
illustrated in Fig. 1. In this test the effect of loading time on the
crack lengths and indentation diagonal lengths was examined.
This effect was examined for Duceram Opaker after annealing
at 4408C for 10 h. Fig. 1 shows the mean values and the stan-
dard deviations calculated from n ˆ 10 indentations for each
loading time, respectively. The straight line parallel to the
abscissa re¯ects the mean value of all 60 crack lengths and
indentation diagonal lengths, respectively. For a loading time
of 10 s and more no effect on the crack lengths and on the
indentation diagonal lengths can be observed.
The results of the resonance frequency test are documen-
ted in the fourth column of Table 1. The common dental
ceramic materials (Cerec, Duceram Opaker, Empress 1 and
Vita Omega Opaker) have a Young's modulus between 65
and 75 GPa. Empress 2 has a 30±40% higher modulus
because of its lithium-disilica particles. The material with
the lowest compliance, i.e. the highest Young's modulus in
this test was In-Ceram Alumina Celay.
To evaluate, whether in Eq. (3) (refer also to ISO
XXXXX draft/TC 206) the term P/c 1.5 is conclusive, P/c 1.5
vs. P was calculated and plotted in Fig. 2 for batch 1 and in
Fig. 3 for batch 2. The error bars represent standard devia-
tions for a minimum of ®ve indentations at each load. The
®tted lines are means over all indentations (n ˆ 50) for each
material. The results of the tests of batch 1 (before anneal-
ing) show that P/c 1.5 is almost constant over working range
of P for all but one (Cerec) of the ceramic materials tested.
After annealing (batch 2) P/c 1.5 is effectively invariant with
respect to the load as predicted in Eq. (3).
Figs. 4 and 5 show the results for the calculated prefactors
(see (Eq. (5)) for both batches, before (batch 1) and after the
annealing process (batch 2). Additionally the mean value of
all prefactors calculated from the seven individual values is
listed for both batches. These mean values were j ˆ 0.0166
for the specimens which were not annealed (batch 1) and
j ˆ 0.0200 for the annealed specimens (batch 2). These
mean values are in good agreement with the prefactor
given in ISO XXXXX draft (TC 206) (j ˆ 0.018). The
difference between the lowest and the highest calculated
prefactor is about 200% for batch 1 and about 180% for
batch 2, respectively. The lowest prefactor of batch 2
(j ˆ 0.0122 for Empress 1) is 73% and the highest prefactor
of batch 2 (j ˆ 0.0267 for Empress 2) is 134% of the calcu-
lated mean prefactor.
4. Discussion
Fracture toughness (KIc) is an important material charac-
terization value. Its value characterizes the resistance of a
material against a propagating crack. The higher the KIc
value is, the better is the mechanical behavior of a compo-
nent fabricated out of this material. Unfortunately, ceramic
materials have low fracture toughnesses. This re¯ects one
main problematic aspect of ceramics in general. Dental
15
Fig. 1. Effect of time of loading (F ˆ 30 N, constant) on crack length 2c and
on indentation diagonal length 2a. Sample: Duceram Opaker after anneal-
ing at 4408C for 10 h (batch 2). Number of imprints: n ˆ 10 for each
indentation time. The straight line parallel to abscissa illustrates the mean
value of all 60 values for the crack lengths and indentation diagonal lengths,
respectively.
ceramics have even lower KIc values than (high perfor-
mance) polychristalline ceramics because they mostly
consist of crystalline particles in a glassy matrix. These
glassy matrices are very brittle (KIc < 1 MPam 0.5). Never-
theless those ceramic materials are frequently used for
dental applications because of their esthetic advantages.
Typical dental ceramics in this test are the feldspar or
leucite containing materials Cerec, Duceram Opaker,
Empress 1 and Vita Omega. For these ceramics fracture
toughnesses between 1.2 and 1.5 MPam 0.5 were determined
by the bending test. Hydroxyapatite (HA) is a very brittle
ceramic material although it has no glassy matrix. The
measured KIc value is 0.9 MPam 0.5. This low value implies
that HA is not suitable for load-carrying components.
Therefore, it is used only to restore bone defects (ability
of bony ingrowth). Empress 2 shows a much better mechan-
ical behavior because of its high fracture toughness value
which is double the value (KIc ˆ 2.5 MPam 0.5) compared for
example to Empress 1. The higher fracture toughness of
Empress 2 can be explained with the chemical composition
and the outstanding structure of this material. Note that
Empress 1 and 2 are very different ceramic materials, shar-
ing only a trade name. Empress 2 consists of 60 wt.% of
crystalline phase (main crystalline phase: Li2Si2O5). The
Li2Si2O5 crystals form an interlocking structure which
causes the favorable mechanical properties. In-Ceram
Alumina Celay has the highest fracture toughness of the
dental ceramics tested in this study (KIc ˆ 5.0 MPam 0.5).
In-Ceram has a completely different structure. It consists
of an alumina matrix which is in®ltrated with a Lanthanum
borisilicate glass. The high fracture toughness of the
alumina-glass composite is mainly due to alumina particle
diffuse toughening and the mismatch of thermal expansion
coef®cient between alumina and Lanthanum borisilicate
glass [21]. Note that dense Alumina has also a fracture
toughness of KIc < 5 MPam 0.5 [1].
In another study KIc values for Cerec, Duceram Opaker
and Empress 1 were evaluated by the bending test as well
16 H. Fischer, R. Marx / Dental Materials 18 (2002) 12±19

Fig. 2. 2 Plot of P/c 1.5 over working range of load P for each tested material before annealing (batch 1). Number of values for crack length c used to determine
mean curves is indicated in each case.
 H. Fischer, R. Marx / Dental Materials 18 (2002) 12±19 17

Fig. 3. Plot of P/c 1.5 over working range of load P for each tested material after annealing (batch 2). Number of values for crack length c used to determine
mean curves is indicated in each case.
18 H. Fischer, R. Marx / Dental Materials 18 (2002) 12±19
Fig. 4. Calculated prefactors j of specimens before annealing process
(batch 1).
[22]. The fracture toughnesses evaluated in that study were
up to 39% higher. This re¯ects that the notch radius width of
the specimens has a big effect on the results. The notches in
the cited study had widths of w $ 100 mm as documented
by the authors. The notch widths realized in this study was
w < 30 mm. On specimens with too wide notches
(w $ 60 mm) apparently higher fracture toughnesses will
be measured [7].
Fig. 1 documents that the time of loading of the indenta-
tion test is not critical with respect to the crack lengths and
the indentation diagonal lengths. From these results it can be
stated that the crack lengths are reproducible for t $ 10 s.
The value for the indentation diagonal length from which
the hardness can be calculated (see Eq. (4)) is almost inde-
pendent of the time of indentation. The value is even stable
for very short indentation times (t ˆ 1 s).
Precaution must be taken in selecting a working range of
indentation loads if the test surface contains residual stresses
prior to indentation (batch 1, see Fig. 2). The presence of
residual stresses leads to a systematic variation in P/c 1.5.
This phenomenon was also observed by other scientists
[14]. Especially for Cerec P/c 1.5 vs. P is not constant for
the unannealed specimens. This variation can be minimized
by annealing the tested surfaces (see Fig. 3). Therefore Eq.
(3) is conclusive over a wide working range of indentation
loads only for annealed, i.e. residual stress free specimens.
The correlation of the KIc-values, determined by the (stan-
dard) bending test with the results of the indentation test
show that the prefactor in Eq. (3) to calculate KIc from the
indentation pattern is not a constant value but a function of
the ceramic material. The results con®rm the hypothesis that
the prefactor stated in the ISO XXXXX draft (TC 206)
(j ˆ 0.018) represents only a mean value and is not suitable
to calculate true fracture toughnesses. From Fig. 5 it can be
taken that the mean prefactor j of the seven (annealed)
ceramics (batch 2) tested in this study was 0.02. The corre-
sponding mean factor j of batch 1 (measurements before
annealing, see Fig. 4) was 0.0166, i.e. a 17% lower value.
Correspondingly all of the individual prefactors j of all
Fig. 5. Calculated prefactors j of specimens after annealing process
(batch 2).
unannealed tested ceramic materials showed lower values.
A lower prefactor implies shorter crack lengths c. The crack
cannot propagate in its natural kind if compression stresses
are induced into the surface of the ceramic material. The
lower prefactors therefore indicate that residual compres-
sion stresses have been induced by the grinding and polish-
ing process. After annealing for 10 h these residual stresses
had relaxed. Therefore, components on which fracture
toughness measurements with the indentation method will
be done should be annealed ®rst in order to remove possible
residual stresses.
For Empress 1 the lowest prefactor (75% of the mean
value of the annealed specimens) and for Empress 2 the
highest prefactor (134% of the mean value of the annealed
specimens) were calculated. The last two colums of Table 1
list the fracture toughness values tested by the indentation
method under the assumption that the constant prefactor
j ˆ 0.018 as given in ISO XXXXX draft (TC 206) is
used. The constant prefactor leads to errors up to 48% for
the unannealed specimens (Empress 1, batch 1) and up to
33% for the annealed specimens (Empress 2, batch 2).
The physical meaning of the prefactor was worked out by
Lawn et al. [10]. It was shown that the prefactor consists of
an elastic and of a plastic component in the median/radial
crack system. The elastic component is independent of the
median/radial crack system, but the plastic component is
material sensitive and depends on the crack system, i.e.
whether the median/radial crack system is well developed
(half-penny shaped cracks) or not [11,23]. Because the
P/c 1.5 ratio was constant over a wide working range of
indentation loads for the annealed specimens (batch 2),
half-penny shaped cracks can be assumed at least for the
specimens of this batch [14]. Despite these limitations it
should be noted that the indentation method can exhibit
exceptional crack development behavior especially on
coarse-grained materials, anomalous glasses, softer cera-
mics, and monocrystals [14]. That means, the deviations
 H. Fischer, R. Marx / Dental Materials 18 (2002) 12±19
of the individual prefactors evaluated in this study from a
constant mean prefactor may be caused by multiple effects
with respect to the individual material characteristics, stress
state and crack systems.
5. Conclusion
The indentation method is not an adequate tool to exactly
determine the fracture toughness of an unknown dental cera-
mic material because the material speci®c, individual
prefactor in the formula to calculate KIc is unknown as
long as no bending tests have been done. Using the constant
prefactor (j ˆ 0.018) given in the formula of the ISO
XXXXX draft (TC 206) and con®rmed as a mean value in
this study, only a ®rst rough estimation of the fracture
toughness can be calculated.
Acknowledgements
The authors gratefully acknowledge the assistance of Mr
Michael Weber for preparing the bending bar specimens and
for the bending bar measurements.
References
[1] Munz D, Fett T. Ceramics: mechanical properties, failure behavior,
materials selection. 1st ed. Berlin Heidelberg New York: Springer,
1999.
[2] EN NMP 291 AK 3. Hochleistungskeramik-Monolithische Keramik,
Ermittlung der RisszaÈhigkeit KIc fuÈr Biegeproben mit ebener Kerbe
oder ebenem Riss bei Raumtemperatur. European Committee for
Standardization, BruÈssel, 1995.
[3] DIN 51109. PruÈfung von keramischen Hochleistungswerkstoffen,
Ermittlung der RiûzaÈhigkeit KIc. Beuth-Verlag, Berlin, 1991.
[4] Pabst RF. Determination of KIc-factors with diamond-saw cuts in
ceramic materials. In: Bradt RC, Hasselman DPH, editors. Fract
Mech Ceram, vol. 2. New York: Plenum Press, 1974. p. 555±65.
[5] Nishida T, Pezzotti G, Mangialardi T, Paolini AE. Fracture mechanics
evaluation of ceramics by stable crack propagation in bend bar speci-
mens. Fract Mech Ceram 1996;11:107±14.
[6] Nishida T, Hanaki Y, Pezzotti G. Effect of notch-root radius on the
19
fracture toughness of a ®ne grained alumina. J Am Ceram Soc
1994;77:606±8.
[7] KuÈbler J. Fracture toughness of ceramics using the SEVNB method;
round robin, VAMAS-Report No. 37. EMPA, DuÈbendorf, 1999.
[8] Evans AG, Charles EA. Fracture toughness determinations by inden-
tation. J Am Ceram Soc 1976;59:371±2.
[9] Binner JGP, Stevens R. The measurement of toughness by indenta-
tion, review paper. British Ceramic 1984;83:168±72.
[10] Lawn BR, Evans AG, Marshall DB. Elastic/plastic indentation
damage in ceramics: the median/radial crack system. J Am Ceram
Soc 1980;63:574±81.
[11] Niihara K, Morena R, Hasselman DPH. Evaluation of KIc of brittle
solids by the indentation method with low crack-to-indent ratios.
Mater Sci Let 1982;1:13±6.
[12] Ostojic P, McPherson R. A review of indentation fracture theory: its
development, principles and limitations. Int J Fracture 1987;33:297±
312.
[13] ISO XXXXX draft. The test methods for fracture toughness of mono-
lithic ceramics at room temperature. Proposed by JISC, Secretariat of
ISO Technical Committee 206: Fine Ceramics, 1995.
[14] Anstis GR, Chantikul P, Lawn BR, Marshall DB. A critical evaluation
of indentation technique for measuring fracture toughness: I. Direct
crack measurements. J Am Ceram Soc 1981;64:533±8.
[15] Munz D, Fett T. Mechanisches Verhalten keramischer Werkstoffe.
Berlin Heidelberg New York: Springer, 1989.
[16] Spinner S, Tefft WE. A method for determining mechanical reso-
nance frequencies and for calculating elastic moduli from these
frequencies. Proceedings ASTM 1961;61:1221±38.
[17] DIN V ENV 843-2. Hochleistungskeramik, mechanische Eigenschaf-
ten bei Raumtemperatur, Teil 2: Bestimmung der elastischen
Konstanten. Beuth, Berlin, 1998.
[18] ASTM C 623-69 T. Tentative method of test for Young's modulus,
shear modulus and poisson's ratio for glass and glass-ceramics by
resonance. Book of ASTM Standards, 1971. p. 646±54.
[19] Meredith N. Determination of the elastic modulus of resin based
materials as a function of resonance frequency during polymerisation.
Dent Mater 1999;15:98±104.
[20] Willmann G. Medical grade hydroxyapatite: state of the art. British
Ceram Trans 1996;95:212±6.
[21] Hornberger H, Marquis PM. Mechanical properties and microstruc-
ture of In-Ceram, a ceramic-glass composite for dental crowns. Glas-
tech Ber Glass Sci Technol 1995;68:188±344.
[22] Bieniek KW, Marx R. Die mechanische Belastbarkeit neuer vollker-
amischer Kronen- und BruÈckenmaterialien. Schweiz Monatsschr
Zahnmed 1994;104:284±9.
[23] Quinn GD, Salem J, Bar-On I, Cho K, Foley M, Fang H. Fracture
toughness of advanced ceramics at room temperature. J Res Natl Inst
Stand Technol 1992;97:579±607.