Physiological aging of an all-ceramic restorative material J.L. Drummond D. Novickas
Department of Operative Dentistry University of Illinois at Chicago Chicago, IL 60612 J.W. Lenke
Health Foundation American Dental Association Chicago, IL 60611 Received April 16, 1990 Accepted January 2, 1991 This project was supported in part by the Wach Fund, College of Dentistry, University of Illinois at Chicago. Dent Mater 7:133-137, April, 1991 Abstract-Ceramic materials exposed to a liquid environment may be subject to stress corrosion and/or time-delayed failure. The intent of this project was to evaluate the susceptibility of a magnesia alumina spinel (Cerestore~) to stress corrosion and degradation. Bars 2.5 x 2.5 x 30.0 mm were prepared according to manufacturer's instructions. Specimens were aged in distilled water or air at 37°C. The modulus of rupture was evaluated at zero, six, and 12 months in four-point loading at loading rates of 0.05, 0.5, and 5.0 mm/min. The modulus of rupture of each specimen was tested in its respective aging medium. The data were analyzed by oneway analysis of variance with a multiplemeans comparison test, linear regression analysis, and Weibull statistics. The pooled data sets of specimens aged in water vs. those aged in air showed a significant difference in the respective modulus of rupture (air, 108.50 _ 16.11; water, 96.94 ± 15.04 MPa). The one-way analysis of variance showed no significant difference between the aging times zero, six, and 12 months in each respective aging medium. The Weibull analysis also showed no difference between the Weibull constants, 7.66 air vs. 7.64 water, but a significant difference between the characteristic strengths, 115.22 air vs. 103.02 water. This study indicates that distilled water has a significant degradative effect on a magnesia alumina spinel, more likely affecting the mode of fracture rather than the stress corrosion characteristics.
he increased utilization of all-ceramic restorations minimizes two clinical problems: the bonding of the ceramic to a metal substrate and the metal-tissue junction. A generalized problem for ceramic materials is a large discrepancy between theoretical strengths as calculated from atomic bonding forces (7000-70,000 MPa) and the strength values commonly observed (7-700 MPa). This discrepancy is a t t r i b u t e d to the presence of microdefects within the material (Kingery, 1960). These microdefects may be pre-existent within the material as a result of the method of preparation or may be generated during mechanical and/or environmental (chemical attack) loading. The interaction of the applied load and the microdefects results in a macroscopic crack which propagates in a slow m a n n e r until avalanche-like fracture occurs. Thus, the fracture behavior of materials requires two steps: the formation of a crack and then its propagation to fracture. Another consideration of ceramic
T
materials is the t i m e - d e p e n d e n t strength degradation in corrosive environments due to subcritical crack growth (Evans, 1974; Evans and Johnson, 1975; Wiederhorn, 1968; Ferber, 1981). Crack propagation is enhanced when ceramics (glass, alumina, porcelain) are placed in aqueous environments (Wiederhorn and Bolz, 1970; Wiederhorn, 1974). This time-dependence of strength or delayed failure occurs without warning weeks or even months after application of the load. Moisture can assist in failure of a material in several ways: (1) corroding of the surface, which leads to surface flaws; (2) condensing into a crack tip, exerting considerable capillary pressure, acting to open the flaw; (3) embrittling the material around the crack tip, (4) reducing the surface energy necessary to form the new surfaces, and (5) lowering the activation energy associated with slow crack growth (Bascom, 1974). The crack growth speed, V, is related empirically to the stress inten-
150
125 Q.
1 oo
,,,.
[] • • [] [] •
75
0 -.J .J
50.
a o
AIR 5MM/MIN AIR 0.SMM/MIN AIR 0,05MM/MIN WATER S MM/MIN WATER 0.5MM/MIN WATER 0.05MM/MIN
25.
0i 0
6
12
AGING TIME (MONTHS)
Fig. 1. Modulus of rupture for aged magnesia alumina spinel specimens.
Dental Materials/April 1991 133
150
In af = bzln # + bo where bo = a constant 1 bl ~ n+l
125,
A second relationship can be developed between the time of aging and the fracture stress, af:
L L~
~ s s s ~ s s s
75 [] []
0
50
=.
(2)
NR
WATER
# s s s ~ s s s ~ s s s ¢ s s s ¢%s%S%s ~
~ s s s ~s~s ~s~s
0.5
0.05
LOADING RATES (MM/MIN)
F/g. 2. Modulus of rupture for pooled magnesia alumina spinel specimens.
sity factor, KI, by the following relationship: v = AK~
(1)
where A and n are constants. For the case of a normalized K1 (divided by K~c, the critical fracture stress intensity factor), the higher n the more resistant the material to stress
corrosion and the less susceptible to degradation in a corrosive environment (Palmer et al., 1979). Subcritical crack growth can be related to the fracture stress, af, and the stress rate, ~, such that an increase in the applied stressing rate results in a higher fracture stress. A relationship can be shown to exist that
In t = bsln af + b2 where b= = a constant b8 = - n
(3)
By plotting In a t vs. In d and In t vs. In at, the slope can be used to calculate n (Evans, 1974; Ferber, 1981). The purpose of this experiment was to utilize two different approaches to obtain a value for the stress corrosion coefficient and determine the effect, if any, on the modulus of rupture on an all-ceramic restorative material after being aged in air or distilled water. MATERIALS AND METHODS Sample bars 2.5 mm x 2.5 mm x 30 mm were waxed, invested, injected, and heat-treated according to the manufacturer's instructions for the processing of the magnesia alumina spinel (Cerestore~, Johnson and Johnson Dental Products, East Windsor, N J). After being fired, the specimens were ground to a surface
TABLE1 AGING AND WEIBULL PARAMETERSOF MAGNESIAALUMINA SPINEL SPECIMENS Group N Pooled Loading Rate Data Air Control 57 Water Control 62 Air 6 Months 54 Water 6 Months 52 Air 12 Months 51 Water 12 Months 52 Pooled Aging Time Data Air 5 mm/min 51 Water 5 mm/min 67 Air 0.5 mm/min 60 Water 0.5 mm/min 50 Air 0.05 mm/min 51 Water 0.05 mm/min 49 Pooled Loading Rate and Aging Time Data Air 162 Water 166 N, number of specimens. MOR, modulus of rupture. r2, correlation coefficient from linear regression analysis. m, Weibull constant. so, characteristic strength from Weibull analysis.
r2
MOR _ SO (mPa)
m
So
0.19 0.28 0.00 0.06 0.01 0.26
98.63 92.62 111.43 101.43 115.43 97.61
- 15.73 _ 13.40 _ 14.13 _ 17.10 _+ 13.56 __ 13.45
6.49 7.89 8.99 6.21 9.58 7.64
105.91 98.28 117.51 109.07 121.38 103.80
0.02 0.05 0.29 0.00 0.50 0.04
109.45 88.28 113.00 102.06 102.26 103.57
__ 13.66 _ 13.69 _ 17.76 _+ 15.39 +__14.54 -+ 10.07
9.14 7.34 6.27 6.91 7.65 11.82
115.33 94.03 118.86 109.72 108.30 108.05
0.00 0.36
108.50 __ 18.11 96.94 _+ 15.04
7.66 7.64
115.22 103.02
134 DRUMMOND et aL/AGING OF A CERAMIC RESTORATIVE MATERIAL
finish of 600-grit and the edges beveled. The study was conducted by two different a p p r o a c h e s - a dynamic fatigue study and an aging s t u d y - t o evaluate any possible changes in the modulus of rupture (MOR). The specimens were aged in sealed polyethylene containers in air at 37°C or with 500 mL of distilled water for six and 12 months at 37°C to determine the effect of aging on the MOR. The specimens were tested in their respective aging media in four-point loading. The MOR was tested at loading rates of 0.05, 0.5, and 5.0 mrn/min in an Instron 1125 testing system (Instron Corporation, Inc., Canton, MA). The MOR was calculated by the following equation: 4P1 (4) 3wh 2 where 1 is the span between the outside supports (the inside supports span is 1/21), P the breaking load, w the sample width, and h the sample height. The aging results were then compared by one-way analysis of variance (ANOVA) of the means, followed by a S t u d e n t - N e w m a n Keuls (SNK) test to determine which aging times and/or solutions were significantly d i f f e r e n t . L i n e a r regression analysis was employed to obtain a best fit for the three stressing rates for each respective aging medium and time of aging. If the distribution of the fracture data was not applicable to a normal distribution, Weibull analysis was applied to the data to obtain the Weibull constant and the characteristic s t r e n g t h (Drummond and Mieske, 1991). The chemical analysis was determined by energy-dispersive x-ray analysis (EDS, P r i n c e t o n - G a m m a - T e c h , Princeton, N J).
2
I
0
"J
-2
/
~
AIR
-3
MOR =
-4
"5
•
3.8
4.0
4.2
4.4
•
|
I
4.8
4.6
|
5.0
LOG(S) S=STRESS AT FAILURE (MPA) F/g. 3. Weibullplotsfor pooledagedmagnesiaaluminaspinelspecimenstestedat 5.0 mm/min.
2 1
WATER o
~" | ~ .,-. 1~ .j
RESULTS AND DISCUSSION
~ 0
The ANOVA analysis of the specimens MOR aged in water or air indicated no significant difference between the controls or the specimens aged six or 12 months at any of the three testing rates (Fig. 1). The observed flexure strength of magnesia alumina spinel was in the same range as that observed in previous studies (Campbell, 1989; McLean and Kedge, 1987). ANOVA analysis of the loading
"J
-1
-2
AIR
-3 [] B [] []
-4'
[]
-5 ' 3.8
~ 4.0
4.2
4.4
4.6
4.8
5.0
LOG(S) S=STRESS AT FAILURE (MPA) Fig. 4. Weibull plots for all pooled air and water magnesia alumina spinel specimens.
Dental Materials~April 1991 135
TABLE 2
ENERGY-DISPERSIVEANALYSISOF MAGNESIAALUMINASPINEL VALUES ARE WEIGHTPERCENTOF OXIDE(± 5.0%) Area AI203 MgO P1 area 63.83 10.32 P2 spot 16.05 0.00 P3 spot 14.99 0.00 P4 spot 35.43 14.08 P5 spot 5.65 92.21 P6 spot 96.67 0.39 P7 spot 0.53 0.00 P1 is the matrix of the spinel. P2-P7 are selected particles in the matrix.
Si02 20.31 57.65 56.47 50.30 2.05 2.71 3.33
CaO 0.35 0.93 0.83 0.08 0.08 0.12 0.00
Ba203 5.18 25.38 27.72 0.10 0.01 0.12 0.09
Zr02 0.00 0.00 0.00 0.00 0.00 0.00 96.04
Fig. 5. Secondary electron image of energy-dispersive x-ray analysis area of a magnesia alumina spinel.
rates indicated that the MOR was significantly lower for the specimens tested in air at 0.5 mm/min vs. 5.0 and 0.05 mm/min. For the specimens tested in water, the MOR tested at 5 mm/min was significantly lower than at 0.5 and 0.05 mm/min. Since there was no significant difference in the MOR dependent on aging time, the data were pooled. There is a clear effect of water to decrease the modulus of rupture. For the pooled data at 5.0 mm/min and 0.5 mm/min, the MOR in water is significantly less than in air (Fig. 2). At 0.05 mm/min, the MOR was essentially the same for air or water. Linear regression analysis of the data to determine n resulted in a correlation coefficient in the range of 0.00 to 0.50, thereby showing very
little correlation to determine n by either aging or dynamic fatigue (Table 1). Due to the lack of correlation as determined by linear regression analysis, Weibull analysis was performed. This analytical approach to Weibull is explained in greater detail in Drummond and Mieske (1991). The Weibull plots for all pooled aged data in air or water tested at 5.0 ram/ min and all pooled data in air or water are shown in Figs. 3 and 4, respectively. The 5.0 mm/min plots give an excellent fit to the data. The pooled plots (Fig. 4) indicate one mode of fracture in water and two modes of fracture in air, one above 4.5 MPa and a second below 4.5 MPa. In use of three loading rates, the theory holds that the slower the loading rate, the lower the MOR. This decrease is
136 DRUMMOND et aL/AGING OF A CERAMIC RESTORATIVE MATERIAL
attributed to the more time available for the microcracks present on the surface to grow and lead to catastrophic failure. The EDS analysis is presented in Table 2. The data are from a specimen aged for 458 days in air at 37°C, but are representative of all the samples analyzed. The actual values of the oxides will vary depending on the exact area analyzed. P1 is the entire area analysis of the material, and P2-P7 are spot analyses of the individual particles of the magnesia alumina spinel. As these spot analyses indicate, this is not a homogeneous m a t e r i a l b u t a m a t r i x of alumina and silica with particles of high concentrations of alumina, silica, baria, magnesia, and zirconia. The scanning electron micrographs of the EDS-analyzed area and spots are presented in Fig. 5 (electron beam activation) and Fig. 6 (back-scatter activation). The back-scatter image excitation is dependent on the atomic m a s s - t h a t is, the higher the atomic mass, the more intense the image, as is evident for P7, zirconia. The observation of incomplete or weak bonding between the particles and the matrix may be one reason this material is susceptible to stress corrosion. The magnesia alumina spinel did not show any significant difference in the MOR during aging in air or water at 37°C, but did show a significant difference between the MOR for the specimens tested in air vs. water. This result was also confirmed by Weibull analysis for the characteristic strengths. This decrease in strength in water is attributed to the effect of water in aiding the crack propagation to failure, assisted by the incomplete bonding between the particles and matrix of the spinel. REFERENCES
BASCOM, W.D. (1974): The Surface Chemistry of Moisture-induced Composite Failure. In: Interfaces in Polymer Matrix Composites, Vol. 6, E.P.
Plueddemann, Ed., New York: Academic Press, pp. 79-108. CAMPBELL,S.D. (1989): A Comparative Strength Study of Metal Ceramic and All-Ceramic Esthetic Materials, J Prosthet Dent 62: 476-479. DRUMMOND,J.L. and MIESKE, K. (1991): Weibull Statistical Analysis of Dental
Composite Data: Aged in Physiological Media and Cyclic Fatigued, Dent Mater 7: 25-29. EVANS, A.G. (1974): Slow Crack Growth in Brittle Materials Under Dynamic Loading Conditions, Int J Fract Mechan 10: 251-259. EVANS, A.G. and JOHNSON, H. (1975): The Fracture Stress and Its Dependence on Slow Crack Growth, J Mater Sci 10: 214-222. FERBER, M.K. (1981): Delayed Failure Characteristics of Plasma-Sprayed Alumina Applied to 316L Stainless Steel and Ti-6AI-4V E L I Substrates in Various Physiological Media, Ph.D. Thesis, University of Illinois, pp. 1416. KINGERY,W.D. (1960): Introduction to Ceramics. New York: John Wiley and Sons, pp. 591-627. MCLEAN, J.W. and KEDGE, M.I. (1987): High-Strength Ceramics, Quint Int 18: 97-106. PALMER, R.A.; LINDBERG, W.R.; and HENCH, L.L. (1979): Fatigue Properties of Li20"2Si02 Glass and Glass-Ceramic, J A m Ceram Soc 62: 319-320. WIEDERHORN, S.M. (1968): Moisture Assisted Crack Growth in Ceramics, Int J Fract Mechan 4: 171-177. WIEDERHORN, S.M. (1974): Subcritical Crack Growth in Ceramic. In: Frac-
Fig. 6. Back-scattered electron image of energy-dispersive x-ray analysis area of a magnesia alumina spinel.
ture Mechanics of Ceramics, Vol. 2, R. Bradt, D.P. Hasselman, and F.F. Lange, Eds. New York: Plenum Press, pp. 613-645.
WIEDERHORN,S.M. and BOLZ, L.H. (1970): Stress Corrosion and Static Fatigue of Glass, J A m Ceram Soc 53: 543-548.
Dental Materials~April 1991 137
Department of Operative Dentistry University of Illinois at Chicago Chicago, IL 60612 J.W. Lenke
Health Foundation American Dental Association Chicago, IL 60611 Received April 16, 1990 Accepted January 2, 1991 This project was supported in part by the Wach Fund, College of Dentistry, University of Illinois at Chicago. Dent Mater 7:133-137, April, 1991 Abstract-Ceramic materials exposed to a liquid environment may be subject to stress corrosion and/or time-delayed failure. The intent of this project was to evaluate the susceptibility of a magnesia alumina spinel (Cerestore~) to stress corrosion and degradation. Bars 2.5 x 2.5 x 30.0 mm were prepared according to manufacturer's instructions. Specimens were aged in distilled water or air at 37°C. The modulus of rupture was evaluated at zero, six, and 12 months in four-point loading at loading rates of 0.05, 0.5, and 5.0 mm/min. The modulus of rupture of each specimen was tested in its respective aging medium. The data were analyzed by oneway analysis of variance with a multiplemeans comparison test, linear regression analysis, and Weibull statistics. The pooled data sets of specimens aged in water vs. those aged in air showed a significant difference in the respective modulus of rupture (air, 108.50 _ 16.11; water, 96.94 ± 15.04 MPa). The one-way analysis of variance showed no significant difference between the aging times zero, six, and 12 months in each respective aging medium. The Weibull analysis also showed no difference between the Weibull constants, 7.66 air vs. 7.64 water, but a significant difference between the characteristic strengths, 115.22 air vs. 103.02 water. This study indicates that distilled water has a significant degradative effect on a magnesia alumina spinel, more likely affecting the mode of fracture rather than the stress corrosion characteristics.
he increased utilization of all-ceramic restorations minimizes two clinical problems: the bonding of the ceramic to a metal substrate and the metal-tissue junction. A generalized problem for ceramic materials is a large discrepancy between theoretical strengths as calculated from atomic bonding forces (7000-70,000 MPa) and the strength values commonly observed (7-700 MPa). This discrepancy is a t t r i b u t e d to the presence of microdefects within the material (Kingery, 1960). These microdefects may be pre-existent within the material as a result of the method of preparation or may be generated during mechanical and/or environmental (chemical attack) loading. The interaction of the applied load and the microdefects results in a macroscopic crack which propagates in a slow m a n n e r until avalanche-like fracture occurs. Thus, the fracture behavior of materials requires two steps: the formation of a crack and then its propagation to fracture. Another consideration of ceramic
T
materials is the t i m e - d e p e n d e n t strength degradation in corrosive environments due to subcritical crack growth (Evans, 1974; Evans and Johnson, 1975; Wiederhorn, 1968; Ferber, 1981). Crack propagation is enhanced when ceramics (glass, alumina, porcelain) are placed in aqueous environments (Wiederhorn and Bolz, 1970; Wiederhorn, 1974). This time-dependence of strength or delayed failure occurs without warning weeks or even months after application of the load. Moisture can assist in failure of a material in several ways: (1) corroding of the surface, which leads to surface flaws; (2) condensing into a crack tip, exerting considerable capillary pressure, acting to open the flaw; (3) embrittling the material around the crack tip, (4) reducing the surface energy necessary to form the new surfaces, and (5) lowering the activation energy associated with slow crack growth (Bascom, 1974). The crack growth speed, V, is related empirically to the stress inten-
150
125 Q.
1 oo
,,,.
[] • • [] [] •
75
0 -.J .J
50.
a o
AIR 5MM/MIN AIR 0.SMM/MIN AIR 0,05MM/MIN WATER S MM/MIN WATER 0.5MM/MIN WATER 0.05MM/MIN
25.
0i 0
6
12
AGING TIME (MONTHS)
Fig. 1. Modulus of rupture for aged magnesia alumina spinel specimens.
Dental Materials/April 1991 133
150
In af = bzln # + bo where bo = a constant 1 bl ~ n+l
125,
A second relationship can be developed between the time of aging and the fracture stress, af:
L L~
~ s s s ~ s s s
75 [] []
0
50
=.
(2)
NR
WATER
# s s s ~ s s s ~ s s s ¢ s s s ¢%s%S%s ~
~ s s s ~s~s ~s~s
0.5
0.05
LOADING RATES (MM/MIN)
F/g. 2. Modulus of rupture for pooled magnesia alumina spinel specimens.
sity factor, KI, by the following relationship: v = AK~
(1)
where A and n are constants. For the case of a normalized K1 (divided by K~c, the critical fracture stress intensity factor), the higher n the more resistant the material to stress
corrosion and the less susceptible to degradation in a corrosive environment (Palmer et al., 1979). Subcritical crack growth can be related to the fracture stress, af, and the stress rate, ~, such that an increase in the applied stressing rate results in a higher fracture stress. A relationship can be shown to exist that
In t = bsln af + b2 where b= = a constant b8 = - n
(3)
By plotting In a t vs. In d and In t vs. In at, the slope can be used to calculate n (Evans, 1974; Ferber, 1981). The purpose of this experiment was to utilize two different approaches to obtain a value for the stress corrosion coefficient and determine the effect, if any, on the modulus of rupture on an all-ceramic restorative material after being aged in air or distilled water. MATERIALS AND METHODS Sample bars 2.5 mm x 2.5 mm x 30 mm were waxed, invested, injected, and heat-treated according to the manufacturer's instructions for the processing of the magnesia alumina spinel (Cerestore~, Johnson and Johnson Dental Products, East Windsor, N J). After being fired, the specimens were ground to a surface
TABLE1 AGING AND WEIBULL PARAMETERSOF MAGNESIAALUMINA SPINEL SPECIMENS Group N Pooled Loading Rate Data Air Control 57 Water Control 62 Air 6 Months 54 Water 6 Months 52 Air 12 Months 51 Water 12 Months 52 Pooled Aging Time Data Air 5 mm/min 51 Water 5 mm/min 67 Air 0.5 mm/min 60 Water 0.5 mm/min 50 Air 0.05 mm/min 51 Water 0.05 mm/min 49 Pooled Loading Rate and Aging Time Data Air 162 Water 166 N, number of specimens. MOR, modulus of rupture. r2, correlation coefficient from linear regression analysis. m, Weibull constant. so, characteristic strength from Weibull analysis.
r2
MOR _ SO (mPa)
m
So
0.19 0.28 0.00 0.06 0.01 0.26
98.63 92.62 111.43 101.43 115.43 97.61
- 15.73 _ 13.40 _ 14.13 _ 17.10 _+ 13.56 __ 13.45
6.49 7.89 8.99 6.21 9.58 7.64
105.91 98.28 117.51 109.07 121.38 103.80
0.02 0.05 0.29 0.00 0.50 0.04
109.45 88.28 113.00 102.06 102.26 103.57
__ 13.66 _ 13.69 _ 17.76 _+ 15.39 +__14.54 -+ 10.07
9.14 7.34 6.27 6.91 7.65 11.82
115.33 94.03 118.86 109.72 108.30 108.05
0.00 0.36
108.50 __ 18.11 96.94 _+ 15.04
7.66 7.64
115.22 103.02
134 DRUMMOND et aL/AGING OF A CERAMIC RESTORATIVE MATERIAL
finish of 600-grit and the edges beveled. The study was conducted by two different a p p r o a c h e s - a dynamic fatigue study and an aging s t u d y - t o evaluate any possible changes in the modulus of rupture (MOR). The specimens were aged in sealed polyethylene containers in air at 37°C or with 500 mL of distilled water for six and 12 months at 37°C to determine the effect of aging on the MOR. The specimens were tested in their respective aging media in four-point loading. The MOR was tested at loading rates of 0.05, 0.5, and 5.0 mrn/min in an Instron 1125 testing system (Instron Corporation, Inc., Canton, MA). The MOR was calculated by the following equation: 4P1 (4) 3wh 2 where 1 is the span between the outside supports (the inside supports span is 1/21), P the breaking load, w the sample width, and h the sample height. The aging results were then compared by one-way analysis of variance (ANOVA) of the means, followed by a S t u d e n t - N e w m a n Keuls (SNK) test to determine which aging times and/or solutions were significantly d i f f e r e n t . L i n e a r regression analysis was employed to obtain a best fit for the three stressing rates for each respective aging medium and time of aging. If the distribution of the fracture data was not applicable to a normal distribution, Weibull analysis was applied to the data to obtain the Weibull constant and the characteristic s t r e n g t h (Drummond and Mieske, 1991). The chemical analysis was determined by energy-dispersive x-ray analysis (EDS, P r i n c e t o n - G a m m a - T e c h , Princeton, N J).
2
I
0
"J
-2
/
~
AIR
-3
MOR =
-4
"5
•
3.8
4.0
4.2
4.4
•
|
I
4.8
4.6
|
5.0
LOG(S) S=STRESS AT FAILURE (MPA) F/g. 3. Weibullplotsfor pooledagedmagnesiaaluminaspinelspecimenstestedat 5.0 mm/min.
2 1
WATER o
~" | ~ .,-. 1~ .j
RESULTS AND DISCUSSION
~ 0
The ANOVA analysis of the specimens MOR aged in water or air indicated no significant difference between the controls or the specimens aged six or 12 months at any of the three testing rates (Fig. 1). The observed flexure strength of magnesia alumina spinel was in the same range as that observed in previous studies (Campbell, 1989; McLean and Kedge, 1987). ANOVA analysis of the loading
"J
-1
-2
AIR
-3 [] B [] []
-4'
[]
-5 ' 3.8
~ 4.0
4.2
4.4
4.6
4.8
5.0
LOG(S) S=STRESS AT FAILURE (MPA) Fig. 4. Weibull plots for all pooled air and water magnesia alumina spinel specimens.
Dental Materials~April 1991 135
TABLE 2
ENERGY-DISPERSIVEANALYSISOF MAGNESIAALUMINASPINEL VALUES ARE WEIGHTPERCENTOF OXIDE(± 5.0%) Area AI203 MgO P1 area 63.83 10.32 P2 spot 16.05 0.00 P3 spot 14.99 0.00 P4 spot 35.43 14.08 P5 spot 5.65 92.21 P6 spot 96.67 0.39 P7 spot 0.53 0.00 P1 is the matrix of the spinel. P2-P7 are selected particles in the matrix.
Si02 20.31 57.65 56.47 50.30 2.05 2.71 3.33
CaO 0.35 0.93 0.83 0.08 0.08 0.12 0.00
Ba203 5.18 25.38 27.72 0.10 0.01 0.12 0.09
Zr02 0.00 0.00 0.00 0.00 0.00 0.00 96.04
Fig. 5. Secondary electron image of energy-dispersive x-ray analysis area of a magnesia alumina spinel.
rates indicated that the MOR was significantly lower for the specimens tested in air at 0.5 mm/min vs. 5.0 and 0.05 mm/min. For the specimens tested in water, the MOR tested at 5 mm/min was significantly lower than at 0.5 and 0.05 mm/min. Since there was no significant difference in the MOR dependent on aging time, the data were pooled. There is a clear effect of water to decrease the modulus of rupture. For the pooled data at 5.0 mm/min and 0.5 mm/min, the MOR in water is significantly less than in air (Fig. 2). At 0.05 mm/min, the MOR was essentially the same for air or water. Linear regression analysis of the data to determine n resulted in a correlation coefficient in the range of 0.00 to 0.50, thereby showing very
little correlation to determine n by either aging or dynamic fatigue (Table 1). Due to the lack of correlation as determined by linear regression analysis, Weibull analysis was performed. This analytical approach to Weibull is explained in greater detail in Drummond and Mieske (1991). The Weibull plots for all pooled aged data in air or water tested at 5.0 ram/ min and all pooled data in air or water are shown in Figs. 3 and 4, respectively. The 5.0 mm/min plots give an excellent fit to the data. The pooled plots (Fig. 4) indicate one mode of fracture in water and two modes of fracture in air, one above 4.5 MPa and a second below 4.5 MPa. In use of three loading rates, the theory holds that the slower the loading rate, the lower the MOR. This decrease is
136 DRUMMOND et aL/AGING OF A CERAMIC RESTORATIVE MATERIAL
attributed to the more time available for the microcracks present on the surface to grow and lead to catastrophic failure. The EDS analysis is presented in Table 2. The data are from a specimen aged for 458 days in air at 37°C, but are representative of all the samples analyzed. The actual values of the oxides will vary depending on the exact area analyzed. P1 is the entire area analysis of the material, and P2-P7 are spot analyses of the individual particles of the magnesia alumina spinel. As these spot analyses indicate, this is not a homogeneous m a t e r i a l b u t a m a t r i x of alumina and silica with particles of high concentrations of alumina, silica, baria, magnesia, and zirconia. The scanning electron micrographs of the EDS-analyzed area and spots are presented in Fig. 5 (electron beam activation) and Fig. 6 (back-scatter activation). The back-scatter image excitation is dependent on the atomic m a s s - t h a t is, the higher the atomic mass, the more intense the image, as is evident for P7, zirconia. The observation of incomplete or weak bonding between the particles and the matrix may be one reason this material is susceptible to stress corrosion. The magnesia alumina spinel did not show any significant difference in the MOR during aging in air or water at 37°C, but did show a significant difference between the MOR for the specimens tested in air vs. water. This result was also confirmed by Weibull analysis for the characteristic strengths. This decrease in strength in water is attributed to the effect of water in aiding the crack propagation to failure, assisted by the incomplete bonding between the particles and matrix of the spinel. REFERENCES
BASCOM, W.D. (1974): The Surface Chemistry of Moisture-induced Composite Failure. In: Interfaces in Polymer Matrix Composites, Vol. 6, E.P.
Plueddemann, Ed., New York: Academic Press, pp. 79-108. CAMPBELL,S.D. (1989): A Comparative Strength Study of Metal Ceramic and All-Ceramic Esthetic Materials, J Prosthet Dent 62: 476-479. DRUMMOND,J.L. and MIESKE, K. (1991): Weibull Statistical Analysis of Dental
Composite Data: Aged in Physiological Media and Cyclic Fatigued, Dent Mater 7: 25-29. EVANS, A.G. (1974): Slow Crack Growth in Brittle Materials Under Dynamic Loading Conditions, Int J Fract Mechan 10: 251-259. EVANS, A.G. and JOHNSON, H. (1975): The Fracture Stress and Its Dependence on Slow Crack Growth, J Mater Sci 10: 214-222. FERBER, M.K. (1981): Delayed Failure Characteristics of Plasma-Sprayed Alumina Applied to 316L Stainless Steel and Ti-6AI-4V E L I Substrates in Various Physiological Media, Ph.D. Thesis, University of Illinois, pp. 1416. KINGERY,W.D. (1960): Introduction to Ceramics. New York: John Wiley and Sons, pp. 591-627. MCLEAN, J.W. and KEDGE, M.I. (1987): High-Strength Ceramics, Quint Int 18: 97-106. PALMER, R.A.; LINDBERG, W.R.; and HENCH, L.L. (1979): Fatigue Properties of Li20"2Si02 Glass and Glass-Ceramic, J A m Ceram Soc 62: 319-320. WIEDERHORN, S.M. (1968): Moisture Assisted Crack Growth in Ceramics, Int J Fract Mechan 4: 171-177. WIEDERHORN, S.M. (1974): Subcritical Crack Growth in Ceramic. In: Frac-
Fig. 6. Back-scattered electron image of energy-dispersive x-ray analysis area of a magnesia alumina spinel.
ture Mechanics of Ceramics, Vol. 2, R. Bradt, D.P. Hasselman, and F.F. Lange, Eds. New York: Plenum Press, pp. 613-645.
WIEDERHORN,S.M. and BOLZ, L.H. (1970): Stress Corrosion and Static Fatigue of Glass, J A m Ceram Soc 53: 543-548.
Dental Materials~April 1991 137
