Académique Documents
Professionnel Documents
Culture Documents
Dental Research
http://jdr.sagepub.com/
Published by:
http://www.sagepublications.com
On behalf of:
International and American Associations for Dental Research
Additional services and information for Journal of Dental Research can be found at:
Email Alerts: http://jdr.sagepub.com/cgi/alerts
Subscriptions: http://jdr.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://jdr.sagepub.com/content/69/12/1791.refs.html
Downloaded from jdr.sagepub.com at MCGILL UNIVERSITY LIBRARY on September 26, 2013 For personal use only. No other uses without permission.
A bi-axial flexure test (piston-on-three-balls), a four-point flexure test, and a diametral tensile test were used to measure the
failure stress of four brittle dental materials: zinc phosphate
cement, body porcelain, opaque porcelain, and visible lightcured resin composite. Furthermore, the fracture probability
of the bi-axial test specimens was predicted from the results
of the four-point flexure test, with use of statistical fracture
theory. Bi-axial failure stresses calculated from an equation
developed by Marshall (1980) exhibited no significant difference for zinc phosphate cement as a function of piston size,
specimen thickness, presence or absence of a stress-distributing film, and loading rate. The four-point flexure strength values of zinc phosphate cement and opaque porcelain were
significantly lower (p<0.05) than the corresponding mean biaxial strength values, while the mean four-point flexure strength
values of body porcelain and resin composite were not significantly lower (p>0.05) than the corresponding mean bi-axial
strength values. The diametral tensile strength of all materials
was significantly lower than the bi-axial flexure strength. The
mean bi-axial flexure strengths of zinc phosphate cement and
opaque porcelain were much higher than the theoretical values
predicted from surface flaw theory, while the strength values
for body porcelain and resin composite were comparable with
those determined from the four-point flexure test. These results
demonstrate that the strength of zinc phosphate cement depends not only upon the geometric factors, but also upon sample preparation conditions.
J Dent Res 69(12):1791-1799, December, 1990
Introduction.
Mechanical strength is an important factor that controls the
clinical success of dental restorations. Usually, complex stress
distributions that are induced by compressive, tensile, and shear
stresses are present in most specimens under practical conditions. It is extremely difficult to induce a pure stress of a single
type in a body. In general, tensile strength is easily determined
for ductile materials such as metals. For convenience, compressive strength is often measured for brittle materials such
as porcelains, cements, amalgams, and resin composites.
However, brittle materials are much weaker in tension than in
compression, and tensile stresses in some materials are (in
certain cases) larger than compressive stress (Anusavice and
Hojjatie, 1987). Therefore, tensile strength is generally considered as the more meaningful property for these brittle materials (compared with compressive strength) for assessment
of the failure potential of dental restorations, especially in the
presence of critical surface flaws.
To determine the tensile strength for brittle dental materials,
the diametral tensile test has been used frequently (Earnshaw
and Smith, 1966; Williams and Smith, 1971; Powers et al.,
Received for publication February 27, 1990
Accepted for publication August 2, 1990
This study was supported by NIDR Grant DE 06672.
1976). The traditional tensile test has rarely been used for
brittle materials (Bowen and Rodriguez, 1962; Zidan et al.,
1980) because of the difficulty associated with gripping and
aligning the specimens. The diametral tensile test provides a
simple experimental method for measurement of the tensile
strength of brittle materials. However, the complex stress distribution developed in the specimen can lead to various modes
of fracture. If the specimen deforms significantly before failure, the data may not be valid. Zidan et al. (1980) suggested
that the diametral tensile test cannot be considered reliable for
dental resinous materials. Chiang and Tesk (1989) demonstrated that a correction of the stress calculation equation for
diametral tension is needed with double cleft fracture.
The main advantage of the flexure test is that a state of pure
tension can be established on one side of the specimen (Berenbaum and Brodie, 1959). Three-point and four-point flexure
tests have also been used for strength evaluation of singlecomponent brittle materials (Bryant and Mahler, 1986; Soderholm, 1986) and metal-ceramic structures (Coffey et al.,
1988). For these uni-axial flexure tests, the principal stress on
the lower surfaces of the specimens is tensile, and it is usually
responsible for crack initiation in brittle materials. However,
undesirable edge fracture (which can increase the variance of
the failure stress value) can occur. Furthermore, these methods
were designed for engineering materials that are usually associated with relatively large specimens. For brittle dental materials, construction of such specimens is not usually convenient
because suitable quantities of dental restorative materials are
not often available to prepare a sufficient number of specimens
for assessment of statistically significant differences. Furthermore, the residual stress states due to polymerization shrinkage
or thermal contraction difference and the flaw characteristics
induced in large specimens may not be representative of those
that are present in smaller clinical restorations.
Recently, the bi-axial flexure test has been used frequently
for the determination of fracture characteristics of brittle materials. The measurement of the strength of brittle materials
under bi-axial flexure conditions rather than uni-axial flexure
is often considered more reliable, because the maximum tensile
stresses occur within the central loading area and spurious edge
failures are eliminated. This allows slightly warped specimens
to be tested and produces results unaffected by the edge condition of the specimen. This feature makes the method suitable
for assessment of the effects of surface conditions on strength.
A wide variety of loading arrangements has been developed
for bi-axial flexure tests: (1) ring-on-ring (Kao et al., 1971),
(2) piston-on-ring (Wilshaw, 1968), (3) ball-on-ring (McKinney and Herbert, 1970), (4) ring-on-ball (Shetty et al.,
1983), (5) piston-on-three-ball (Kirstein and Woolley, 1967),
and (6) ring-on-spring (Marshall, 1980). For this study, the
fifth option was used, since it is suitable for slightly warped
and small specimens such as brittle dental materials, and excellent results by this method have been reported previously
for some glasses and ceramics. The bi-axial flexure strength
is determined by support of a disc specimen on three metal
spheres positioned at equal distances from each other and from
the center of the disc. The load is applied to the center of the
1791
1792
Code
Type
ZP
BP
OP
RC
Body Porcelain
Opaque Porcelain
Resin Composite
Manufacturer
Ormco, Glendora, CA
J.F. Jelenko & Co., Armonk, NY
J.F. Jelenko & Co.
Kerr Manufacturing Co., Romulus, MI
TABLE 2
ZP
BP
OP
RC
D,
t
6.7 0.2
4.3 0.2
7.7 0.1
4.3 0.4
7.8 0.1
3.1 + 0.1
6.2 + 0.0
small; M, medium; and L, large.
3.2 0.1
3.5 + 0.5
3.3 0.3
3.2 0.1
minutes after the mix was started, the glass plates and cements
were placed in a humidor at 37C for one h. After removal
from the humidor, the specimens were separated from the glass
and stored in water at 37C for 24 h.
Porcelain specimens were prepared by normal fabrication
procedures. A slurry of porcelain powder was vibrated and
condensed into a mold 16 mm in diameter and 2 mm in depth.
The discs were fired in a dental oven (Mark IV, J.M. Ney
Co., Bloomfield, CT) at a heating rate of 55C/min under
vacuum to 982C followed by a 90-second holding time in air.
The specimens were removed from the furnace and rapidly
cooled in ambient air by natural convection. The porcelain
discs were ground from 120-grit to 600-grit papers and polished with 1->Lm and 0.3-gm A1203 powder on a metallographic polishing wheel.
Visible-light-cured resin composite specimens were made as
follows: Approximately 0.5 mL of resin paste was placed on
a flat glass plate, 1.0 mm in thickness. A flexible ring, approximately 14 mm in inner diameter and 1.2 mm in thickness,
was placed in the paste, and another glass plate was used to
press the resin paste into a disc. Light emitted from a fiber
optic handpiece (Translux, Kulzer & Co., Bad Homburg, Germany) passed through the glass plates for a total of 200 s so
that adequate polymerization of each side would be ensured.
After light irradiation, specimens were separated from the glass
plates and stored in water at 37C for 24 h before being tested.
Specimens for the four-point flexure and the diametral tensile tests were also prepared in a similar manner. The final
dimensions of the specimens are listed in Table 2.
Determination of fracture strength. -The bi-axial flexure
test apparatus is described in the ASTM Standard F394 for biaxial flexure testing of ceramic substrates. In this study, the
dimensions of the apparatus were smaller than those described
in the ASTM standard so that the small specimen size of brittle
dental materials typically used in dental restorations would be
accommodated.
As shown in Fig. 1, specimens were supported on three steel
spheres (3.2 mm in diameter) equally spaced along a diameter
of 10 mm. For zinc phosphate cement specimens, loading was
applied by a steel piston (with flat areas of 1.2 mm and 1.6
mm in diameter ground along the surface of contact) until
fracture occurred. For the porcelain and resin composite specimens, the piston with a diameter of 1.2 mm was used.
The failure stress, cr, at the center of the lower surface was
calculated by equations developed by Marshall (1980). These
Vol. 69 No. 12
1 793
Load
Piston
2ro
jXTjj~~~ _Specimen
Support
Ball Bearing
Specimen
Holder
1794
TABLE 3
BI-AXIAL FLEXURE TEST RESULTS FOR ZINC PHOSPHATE CEMENT
Condition
Specimen
Diameter
LoadLoading
Meant
of Piston
distributing
Speed
No. of
Strength
Size*
Film
Tests
Code
(mm)
(mm/min)
(MPa)
Yes
0.1
M
1.2
10
18.1
ZP
10
Yes
0.1
19.0
1.2
ZP1
S
16.9
Yes
0.1
10
L
1.2
ZP2
0.1
10
19.2
Yes
M
1.6
ZP3
17.3
1.2
No
0.1
10
ZP4
M
17.5
Yes
1.0
10
M
1.2
ZP5
60
18.0
Total
*Dimensions of S, M, and L specimens are indicated in Table 2.
tCommon vertical lines indicate no significant difference at ox = 0.05.
Result
Standard
Coefficient
Deviation
of
Variation
0.073
0.098
(MPa)
1.3
1.9
2.6
2.1
2.5
2.0
2.2
0.155
0.108
0.142
0.113
0.121
L,
(12)
and
a
3P(3+v)a2/8
t2oal
(13)
(14)
Pf
exp [-B]
(6)
Results.
The mean bi-axial flexure strength values of zinc phosphate
cement as a function of six different test conditions are listed
Vol. 69 No. 12
1 795
TABLE 4
Specimen
Bi-axial flexure test
ZP*
BP
OP
RC
Four-point flexure test
ZP
BP
OP
RC
Diametral tensile test
ZP
BP
OP
RC
*Same data as in Table 3
Ratio of
Strength to
Bi-axial Flex-
No. of
Tests
Meant
Strength
(MPa)
Standard
Deviation
(MPa)
Coefficient
of
Variation
10
10
10
10
18.1
52.4
75.6
103.9
1.3
5.6
8.6
15.6
0.073
0.106
0.114
0.150
1
1
1
1
10
10
10
10
6.8
48.4
52.4
98.4
0.6
5.8
10.2
11.4
0.088
0.125
0.195
0.116
0.375
0.923
0.693
0.947
4.5
22.4
23.8
24.6
0.5
8.7
6.0
5.8
0.088
0.391
0.253
0.234
0.249
0.427
0.315
0.236
10
10
10
10
for group ZP.
ure
Strength
0.427 for the four materials. There were no significant differences (p>0.05) among the diametral tensile strengths of body
porcelain, opaque porcelain, and resin composite. The diametral tensile test data revealed that zinc phosphate cement
exhibited the lowest coefficient of variation of 0.088, while
other materials exhibited relatively larger values of 0.234 to
0.391.
Shown in Fig. 2 are Weibull plots of fracture stresses, In In
[1/(1 -Ff)] vs. ln u, for the zinc phosphate cement tested according to the three methods. The data points were described
by a straight line produced by least-squares fit of the fracture
data by use of a computer, and the Weibull modulus, characteristic strength, and strength at a predicted failure level of
5% were also calculated by computer. These results are listed
in Table 5. The Weibull moduli for both bi-axial and fourpoint flexure tests were larger than those for the diametral
tensile test, except for zinc phosphate cement. The Weibull
modulus of zinc phosphate cement exhibited the largest value
for each test method.
Shown in Table 6 are the failure probability parameters for
the bi-axial flexure test of four brittle dental materials derived
from the analytical solutions of Eqs. (11) to (20) by use of
Weibull analysis results for the four-point flexure test in Table
5. Predicted plots (Bs and B,) for the Weibull analysis of the
zinc phosphate cement are presented as dashed lines in Fig. 2.
Experimental results for the bi-axial test of zinc phosphate
cement and opaque porcelain showed much higher strength
than that predicted by surface flaw analysis. The failure probability of both body porcelain and resin composite exhibited
good agreement with the values predicted for the surface flaw
condition.
Shown in Fig. 3 are SEM images of fracture surfaces produced by bi-axial and four-point flexure stress for zinc phosphate cement, body porcelain, opaque porcelain, and resin
composite. The arrows indicate the most likely sites of crack
initiation. The fracture surfaces for zinc phosphate cement
specimens exhibited a porous structure, especially in the specimens that were prepared for the four-point flexure test, which
had much larger pores than those for the bi-axial test. For the
specimens tested by four-point flexure, the fracture origin appeared to be located around pores located at the corner of the
1796
TABLE 5
SUMMARY OF WEIBULL ANALYSIS
2
ZP
Specimen
Bi-axial flexure test
ZP
BP
OP
RC
Four-point flexure test
ZP
BP
Cc
o.a05
(MPa)
(MPa)
10.45
7.54
7.47
6.27
18.6
54.9
80.2
111.6
14.0
37.0
53.9
69.5
Diametral
Four-point
Biaxial
A
CL
-
8.95
8.85
4.37
6.88
7.1
50.6
57.1
102.9
5.1
36.2
29.0
66.8
ZP
11.68
BP
2.10
3.25
OP
RC
3.39
m, Weibull modulus.
or, characteristic strength.
4.7
3.7
6.9
10.5
11.0
OP
RC
Diametral tensile test
1-
-1
-2
25.9
26.2
26.3
Discussion.
Effects of test conditions on bi-axialflexure strength. -Maximum tensile stresses produced by bi-axial flexure occur below
the central loading area on the bottom surfaces of disc specimens. However, because of the typically high elastic modulus
and hardness of brittle materials, any imperfect contact between the rigid loading tool and the test specimen can lead to
a substantial deviation from radial symmetry in the stress field
and, consequently, to errors in strength measurements. Wachtman et al. (1972) suggested that a layer of polyethylene between piston and test surface would assist in the obtaining of
uniform loading at the end of the piston. Based on strain gauge
measurements, Marshall (1980) found that the piston applies
the load uniformly over its contact area when a film is used.
However, in the present study, there were no significant differences (p>0.05) between the strengths of zinc phosphate
cement with and without a film. It seems that the specimen
surface of zinc phosphate cement, which is covered with a
precipitate by reaction between the cement surface and water
in the storage chamber, was flattened by the initial contact of
the piston and, consequently, developed a uniform loading
condition without an intermediate film. Furthermore, if the flat
surface of the piston is sufficiently parallel to the test surface,
one does not always need to place the film or cushion over the
specimen center.
The thickness of the specimens is one of the most important
factors in the determination of the bi-axial flexure strength,
since the calculated stress is inversely proportional to the second power of its thickness, as derived in Eq. (1). Furthermore,
the stress equation is valid only if the deflection does not exceed about one-half of the plate thickness (Wachtman et al.,
1972). Bending is directly proportional to the sustained load
and inversely proportional to Young's modulus, whereas it is
only slightly dependent on Poisson's ratio. The estimated min-
-3 - _
0.5
B;
1.0
1.5
Bv
2.0
In
2.5
3.0
3.5
(MPa)
orf
C 6rl/(l + n)
a
n
(21)
Vol. 69 No. 12
1797
TABLE 6
PREDICTED FAILURE PROBABILITY PARAMETERS FOR BI-AXIAL FLEXURE STRENGTH BASED ON FOUR-POINT FLEXURE DATA
Surface Flaw
Volume Flaw
In ('rr a2L9)/m
L
ln [(QT a2tl2)L,]/m
ln oC
Code
L,
cv
- 0.288
0.002
2.371
0.015
0.020
2.063
ZP
2.083
4.274
0.002
- 0.258
0.045
3.971
BP
4.016
0.017
0.161
0.026
3.984
0.005
-0.221
4.366
OP
4.145
5.062
0.020
4.691
0.003
- 0.304
0.067
RC
4.758
sistencies in mixing.
Comparison with four-point flexure strength.-Shetty et al.
during setting.
The second reason for the discrepancy is the specimen size
effect. The effect of specimen size on four-point flexure strength
has been reported by many investigators (Berenbaum and Brodie, 1959; Baratta, 1984; Ikeda et al., 1986; Lamon, 1988).
Berenbaum and Brodie (1959) showed that the four-point flexure strength of pure plaster of Paris increased with a decrease
in specimen thickness. They suggested that the four-point flexure strength of porous, weak, and brittle materials such as
plaster is strongly influenced by specimen size. The surface
layer of these chemically setting materials is tougher than their
internal structure, since the surface tends to form a dense structure having few pores during setting under pressure with the
mold, as shown in Fig. 3. Furthermore, the properties of mixing materials such as plaster and cement are significantly affected by the total volume of mixing. For preparation of zinc
phosphate cement specimens for the four-point flexure test, the
mixture of 6 g of powder and 2 mL of liquid was used, while
a standard amount of powder and liquid (1.5 g and 0.5 mL)
was used for preparation of specimens for the bi-axial flexure
test. Larger pores were formed and remained in the specimens
for the four-point flexure test, compared with the bi-axial flexure test specimens, because of insufficient mixing and pressure. Thus, the bi-axial testing of zinc phosphate cement yielded
much higher strength values than those predicted for both surface-flaw failure and volume-flaw failure, possibly because of
differences in the homogeneity of specimens. Opaque porcelain specimens were also relatively inhomogeneous. Compared
with body porcelain, opaque porcelain is relatively difficult to
condense for large specimen volumes such as that used for the
four-point flexure test, since opaque porcelain contains a higher
fraction of opacifiers such as zirconium or titanium oxide,
which reduce the fluidity of the slurry. The results for body
porcelain and resin composite were comparable with those determined from the four-point flexure test, since test specimens
for both materials showed a similar structure for both bi-axial
and four-point flexure test specimens. These results suggest
that this statistical approach demonstrates reasonable agreement between the bi-axial and four-point flexure strengths of
brittle specimens with similar flaw characteristics.
Comparison with diametral tensile strength.-For the diametral tensile test, it is difficult for ideal loading to be produced
along a line when cylindrical specimens are used. A proper
load distribution is generally accomplished by placement of a
narrow pad of suitable materials between the specimen and the
loading platens. For example, in the diametral tensile test procedure for ADA Specification No. 27 for direct filling resins
(Council on Dental Materials and Devices, 1977), a thin piece
of paper (approximately 0.5 mm thick) wet with water must
be inserted between the platens of the testing machine along
each side of the specimen. In ADA Specification No. 1 for
amalgam, the specimen should be padded with two thicknesses
of 0.038-mm aluminum foil on each side. However, it should
be noted that the apparent strength changes with the type of
padding material and its thickness, because the uniformity of
the tensile stress distribution also changes (Rudnick et al.,
1963). Therefore, no pad was used in the present study.
Relative to the diametral tensile test, zinc phosphate cement
exhibited a small coefficient of variation and a large Weibull
modulus, as shown in Tables 4 and 5, whereas other materials
exhibited large coefficients of variation in diametral tensile
strength and small Weibull moduli, compared with those determined from both flexure tests. Although an advantage of
the diametral tensile test is that the maximum tensile stress was
not restricted to the surface, the surface effect on the fracture
value is large (Rudnick et al., 1963). However, it is difficult
to control the surface roughness on curved surfaces. It seems
1798
Fig. 3-SEM images of fractured surfaces that were subjected to bi-axial and four-point flexure for zinc phosphate cement (ZP), body porcelain (BP),
opaque porcelain (OP), and resin composite (RC) specimens.
Vol. 69 No. 12
Acknowledgments.
The authors gratefully acknowledge Professor Jiro Hasegawa, Aichi-Gakuin University, for his helpful suggestions
and Mr. Robert B. Lee, University of Florida, for his technical
support in conducting the mechanical tests.
REFERENCES
ANUSAVICE, K.J. and HOJJATIE, B. (1987): Stress Distribution in Metalceramic Crowns with a Facial Porcelain Margin, JDent Res 66:1493-1498.
BARATTA, F.I. (1984): Requirements for Flexure Testing of Brittle Materials.
In: Methods for Assessing the Structural Reliability of Brittle Materials, ASTM STP 844, S.W. Freiman and C.M. Hudson, Eds., Philadelphia: American Society for Testing and Materials, pp. 194-222.
1799