- Email: [email protected]
Effect of water exposure on the fracture toughness and flexure strength of a dental glass
dental materials Dental Materials 17 (2001) 367±371
www.elsevier.com/locate/dental
Effect of water exposure on the fracture toughness and ¯exure strength of a dental glass Susanne S. Scherrer a,*, Isabelle L. Denry b, H.W. Anselm Wiskott a, Urs C. Belser a a
Department of Prosthodontics, School of Dental Medicine, University of Geneva, Geneva, Switzerland b Department of Restorative and Prosthetic Dentistry, The Ohio State University, Columbus, OH, USA Received 9 August 2000; revised 6 December 2000; accepted 30 January 2001
Abstract Objectives: The low fusing dental glass (Duceram LFC) has been advertised as presenting a superior chemical resistance and augmented strength after 16 h exposure to water or 4% acetic acid. The purpose of this study was to evaluate the effect of prolonged exposure to water on two mechanical properties (fracture toughness and ¯exure strength) of LFC. Methods: Disks and bars were mirror polished and annealed prior to aging in: (1) air (control), (2) water for 24 h at 808C and (3) water for 8 weeks at 808C. Fracture toughness (KIc) was determined by indentation fracture (IF) and indentation strength (IS) using a 19.6 N Vickers indentation load. Flexure strength values were obtained from three-point bending at 0.1 mm/min. Statistical analysis was performed using the Weibull distribution, Tukey and Bartlett tests (P , 0.05). Results: Both techniques (IS and IF) showed a signi®cant improvement in the K of Duceram LFC after 8 weeks in water (0.88 and 1.14 MPa m 0.5) as opposed to the 24-h values both in water and air (0.77±0.78 MPa m 0.5). However, for ¯exure strength the Weibull characteristic (S0) and the m parameter did not change signi®cantly with water storage (S0 90±100 MPa, Weibull m 7±8). Signi®cance: The increase in toughness of Duceram LFC after aging in water is an interesting and favorable observation for a restorative material exposed to the oral environment. Nevertheless, in comparison with other contemporary ceramics, the toughness of this LFC remains in the range of soda-lime-glass or classic feldspar porcelains. q 2001 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Water exposure; Fracture toughness; Flexure strength; Dental glass
1. Introduction The low fusing dental glass Duceram LFC sinters at 6608C. This temperature is ca. 30% lower than that of conventional materials and extends the application range of this material to the repair of porcelain fractures (also after post-ceramic soldering) and to the ®nal adaptation of porcelain margins. Duceram LFC was initially marketed as a restorative material for the fabrication of inlays, onlays and veneers. The manufacturer (Ducera Dental, Germany) especially emphasized the material's superior resistance to chemical degradation in aqueous and acidic media. It was also determined that the ¯exure strength (S) of the material increased by 24% (S 123 MPa) after 16 h exposure to 4% acetic acid [1]. Such a behavior suggests an ion exchange mechanism which modi®es the surface's structural arrangement after exposure to speci®c environments. In a recently * Corresponding author. Tel.: 141-22-382-9129; fax: 141-22-781-1297. E-mail address: [email protected] (S.S. Scherrer).
reported experiment, Jestel et al. [2] used depth-resolved Raman microprobe analyses to examine the surface of Duceram LFC samples after exposure to water and 4% acetic acid. They demonstrated that structured zones containing metaborate, pyroborate, boroxol rings and a mixture of borate and silicate tetrahedra developed on the outer surface of the glass. Pending further investigation, they hypothesized that the increase in strength after exposure to water or acetic acid was due to this newly developed 3±10 mm thick boron-containing surface layer. Yet a strengthening mechanism linked to the exposure to water would be unique among ceramic restorative materials. Indeed most feldspar-based ceramics show strong tendencies towards stress corrosion [3±7]Ða phenomenon which originates in an increasing hydrolysis of the ceramic when subjected to stress application and results in a time-dependent reduction in ¯exural strength [8±13]. The limitations of ¯exural strength as an indicator of the structural performance of brittle materials have been pointed out by Kelly [14]. By contrast, an appropriate parameter would
0109-5641/01/$20.00 + 0.00 q 2001 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S 0109-564 1(01)00002-1
368
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
be the ceramics' fracture toughness (KIc), that is the material's intrinsic resistance to crack propagation [15]. Ceramic materials are susceptible to stress corrosion, dynamic fatigue and surface degradation which all affect their ¯exure strength. It is unclear, however, whether their toughness also is affected by exposure to an accelerated aging environment. Therefore, the aim of the present study was to determine the fracture toughness and the Weibull strength parameters (S0 and m) of a lowfusing dental glass (Duceram LFC) which had been previously aged up to 2 months in 808C water. 2. Materials and methods 2.1. Specimen preparation Disks (13 £ 2 mm 2) and bars (25 £ 4.5 £ 2 mm 3) were prepared from a low fusing dental glass (Duceram LFC dentin, Ducera Dental GmbH, Rosbach, Germany, batch no. CE 0483). After ®ring, the specimens were ground ¯at and parallel with a diamond abrasive disk (600 grit). The test side was further polished using a series of SiC abrasive disks (P#1200, P#2400 and P#4000) and diamond pastes (6, 3 and 1 mm). To relieve residual surface stresses due to grinding and polishing, all samples were annealed. The annealing temperatures for the low fusing dental glass and the feldspathic porcelain were determined from dilatometric analyses (Orton dilatometer, Model 1600D) and reported in a previous study (Scherrer et al. [16]). Annealed specimens were then distributed in three groups: (1) aging in ambient air (room temperature, 1 week), (2) aging in deionized water at 808C for 24 h and (3) aging in deionized water at 808C for 8 weeks. After aging, all specimens were ultrasonically cleaned in ethanol for 5 min and air dried prior to testing. 2.2. Fracture toughness and ¯exure strength determination Fracture toughness (KIc) was determined using the indentation fracture (IF) [17] and the indentation strength (IS) [18] technique. For the IF technique, disks were gold-coated prior to indentation to facilitate the crack length readings. Vickers indentations were produced in ambient air at 19.6 N until 10 acceptable crack patterns were obtained [16]. No drop of oil was used to prevent slow crack growth. However, crack length readings were performed using the indenters microscope within 30 s after indentation to minimize slow crack growth due to moisture presence in ambient air. Fracture toughness was calculated according to the equation developed by Anstis et al. [17] which relates the indentation load (P), the size of the median cracks (c), the modulus of elasticity (E) and the material' hardness (H): KIc 0:016 E=H0:5 P=c1:5
1
Hardness (H) was de®ned as the load divided by the projected area on the surface and calculated as: H 0.5P/ a 2 [19] where a is the average indentation half-diagonal. It was determined from 10 indentations using a load of 1.96 N,
a magnitude that prevented the formation of radial cracks. The material's density, dynamic E and Poisson's ratio, were previously measured using the ultrasonic velocity method [16]. Only E values for air and 24-h water specimens were measured. As the modulus of elasticity did not change, the value obtained for the 24-h water group was also used for the 8-week water specimens for toughness calculation. The second procedure used for determining toughness was the indentation strength (IS) which consisted of producing a Vickers indentation of 19.6 N on the center of the bar's mirror polished side after aging and shortly before testing. Ten indented specimens per aging group were subjected to a three-point-bending test under the following conditions: cross-head speed: 0.1 mm/min, span: 21 mm, ambient air. The stress at fracture (s f) was calculated as:
s f 3WL=2BD2
2
where W is the breaking load, L the span, B the specimen's width and D its thickness. Fracture toughness was calculated from a uni®ed indentation-fracture/tensile-failure equation [18] that relates the modulus of elasticity (E), indentation load (P), hardness (H) and stress at fracture (s f): K Ic 0:59 E=H1=8 s f P1=3 3=4
3
Hardness values were obtained from the disk specimens as previously described. A test was considered successful after inspection of the specimen's matching halves under magni®cation when it was veri®ed that the fracture process did indeed start from the indentation ¯aw. Again, the modulus of elasticity value for the 24-h group was used for the 8week water group. LFC ¯exure strength was determined from 20 bars per aging group. Specimens were polished and beveled on their edges, annealed and aged as described above. The ¯exure strength was determined from three-point-bending at 0.1 mm/min and 21 mm span. 2.3. Statistical analysis The mean fracture toughness of each group was computed and differences among groups tested for statistical signi®cance with Bartlett's and Tukey's multiple range test. A P , 0.05 was considered signi®cant. The Weibull distribution was used to calculate the fracture probability as a function of applied stress [20]. The Weibull formula (for the two parameter distribution) which yields the cumulative probability of failure F(S) as a function of applied stress (S) is given by: m F S 1 2 e
2
S S0
4
where S is the stress at a given point, S0 is the characteristic strength i.e. the stress level at which 63% of the specimens have failed and m (the Weibull modulus) a constant which characterizes the spread of the failure data. The distribution function can also be expressed in (natural) logarithmic
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
369
Table 1 Fracture toughness (KIc) determined by indentation fracture (IF) and indentation strength (IS) techniques after aging Duceram LFC in deionized water at 808C for 24 h and 8 weeks. Number of tests (n), Hardness (H), modulus of elasticity (E). The modulus of elasticity (E) were determined previously [16]. Values with the same superscript were not signi®cantly different (P . 0.05) Duceram LFC
n
E (GPa)
H (GPa) 1.96 N load
IF KIc (MPa m 0.5) 19.6 N load
Crack length IF (mm)
IS KIc (MPa m 0.5) 19.6 N load
Air H2O/24 h H2O/8 weeks
10 10 10
69.6 (0.6) 68.2 (2.5) ±
6.4 (0.5) 5.5 (0.4) 4.6 (0.2)
0.77 (0.09) a 0.77 (0.06) a 1.14 (0.09) b
122.4 (10.1) 127.6 (7.6) 104.4 (5.8)
0.78 (0.08) a 0.78 (0.08) a 0.88 (0.10) b
Table 2 Weibull characteristic strength (S0) and m parameter with corresponding con®dence intervals (CI) of LFC aged specimens tested in three-point bending. Number of specimens (n). Values with the same superscript are not signi®cantly different using con®dence intervals at 95% LFC aging groups
n
Weibull parameter m
Con®dence interval of m
Characteristic strength S0 (MPa)
Con®dence interval of S0 (MPa)
Correlation coef®cient r
Air 24 h H2O 8 weeks H2O
20 20 20
8.0 a 6.9 a 7.2 a
5.0±10.7 4.3±9.2 4.6±9.7
92.1 b 100.2 b 89.6 b
86.1±98.7 92.6±108.6 83.1±96.6
0.940 0.994 0.982
form: lnln
1 1 2 F S
nearly overlaps the air-aging group. The lnln transforms represented in Fig. 2 show similar slopes.
ln S 2 ln S0
5
The linear correlation coef®cient r of the data obtained from the lnln plots provides a measure of the applicability of the distribution. A r value above 0.80 is considered acceptable. The maximum likelihood estimates for the characteristic strength S0 and m value were determined for each group using iterative procedures (Newton±Raphson) as described by Mann et al. [21]. The 95% con®dence intervals were calculated for S0 and m using the tables established by Thoman et al. [22].
3. Results Table 1 lists the hardness (H), the modulus of elasticity (E) and the toughness obtained using the indentation fracture and indentation strength techniques. The toughness increased signi®cantly for both the IF and the IS techniques after 2 months of aging in water. The difference was slight (13%) for the IS technique and marked for IF (48%). No differences were found between 24-h water exposure and aging in air using either IF or IS. Table 2 lists the Weibull parameters S0 and m as well as their respective con®dence intervals. The correlation coef®cient r as determined from the lnln transformed ¯exure strength data is also included. The S0 and m parameters were not signi®cantly different for all three aging groups. The cumulative failure plots and lnln transforms are shown in Figs. 1 and 2. A slight shift to the right (i.e. a minor improvement) was observable in Fig. 1 for the 24-h water specimens. However, this tendency was not con®rmed after 2 months in water and the cumulative probability of failure
4. Discussion The purpose of this study was to determine the ¯exure strength and the toughness of a dental glass ceramic (Duceram LFC) after prolonged exposure to water (8 weeks). The manufacturer promoted LFC as a material whose ¯exure strength was augmented by exposure to waterÐa claim that was indeed con®rmed in one study [1]. The shortcomings of ¯exure tests for ceramic materials have been reviewed by Quinn and Morrell [23] especially with respect to their sensitivity to surface ®nishing procedures. Brittle materials such as ceramics will fail due to the progression of existing ¯aws when
Fig. 1. Plot of the cumulative probability of failure vs. the applied stress of Duceram LFC after aging in air, 24 h in water at 808C and 8 weeks in water at 808C.
370
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
Fig. 2. Plot of the lnln transform of the cumulative probability of failure vs. the ln of the applied stress for the LFC after aging in air, 24 h in water and 8 weeks in water.
subjected to stresses above a critical level. Hence, using the Weibull distribution for the assessment of the material's characteristic ¯exure strength (S0) and Weibull modulus (m) is consistent with the `weakest link theory' [20], which states that a brittle specimen's breakage process will originate from its largest ¯aw. The size of the worst defect depends on the sample and thus induces a scatter in the recorded material strength values. This variability can be analyzed statistically using the Weibull distributionÐa model which assumes that a ®nite strength value is associated with each volume or surface element in the material. In a bend test, the stress distribution is nonuniform varying from zero at the neutral axis to a maximum at the surface under tension. This markedly increases the effect of the surface state on an experimentally determined ¯exure strength. Therefore, care was taken in the present study to mirror polish (^1 mm) and anneal all LFC specimens in order to eliminate residual compressive stresses left after polishing. When these ®nishing procedures were applied, no changes in the Weibull parameters were detected among the air-, 24-h water- and 8-week water specimens. A slight shift to the right of the 24-h cumulative failure curve (Fig. 1) was observed but this trend was not con®rmed after 2 months of exposure to water. All Weibull m parameters were extremely close (Table 2, Fig. 2), indicating that no changes occurred in the material's ¯aw population subsequently to water exposure. The second material property evaluated in the present study was fracture toughness as a function of time and water exposure. As shown in Table 1, a signi®cant increase was observed in the 8-week water specimens regardless of which testing technique was used. After 2 months of exposure to water, the IF technique yielded an increase of 48% in toughness and the IS procedure an increase of 13%. The origin of the observed improvement is unclear. Several possible explanations are given below. First, hardness (H) determined by indentation gradually decreased with time and water exposure (Table 1), thereby
strongly suggesting changes in surface structure. This observation has a bearing on the increase in toughness. Indeed, if we assigned a constant value to hardness (that of the air group) and used Eq. (1) (IF), the recomputed KIc for the 8week water group would be 0.96, that is an increase of 25% (vs. 48%). This increase of 25% is still signi®cantly different from the air- and 24-h water groups and demonstrates the relation between hardness and toughness when applying Eq. (1). It should also be mentioned that the indentation load for hardness measurements was small (1.96 N) to avoid the formation of radial cracks. Under this load, the Vickers indenter's half diagonal is approx. 15 mm and the depth of the indent about 6 mm. With reference to the data by Jestel et al. [2], our Vickers indentation for the H measurements was thus mainly determined by the boron-containing surface layer. Second, one should remember the intricacies of indentation fracture techniques [16,17] with respect to crack lengths measurements. The IF technique is based on a Vickers induced crack system of the radial/median type. This penny-shape crack con®guration may be modi®ed by the presence crack neighbours (i.e. lateral cracks) and the immediate exposure of the indented zone to a reactive environment (moisture) will contribute to slow crack growth. As oil was not used to prevent slow crack growth during indentation it could well be that crack length readings were affected by moisture in air even though the time between indentation and reading was kept to a minimum. The toughness results obtained with the IS technique seemed to be more reliable that those of the IF due to the absence of possible error during crack length readings. However, the same critique about moisture interaction would apply to the IS technique as the time between indentation and testing was approximately 30 min to 1 h. Even though the testing timeframe was similar for all groups, the 8-week water specimens may perhaps have had more slow crack growth due to the presence of water molecules within the LFC's surface. Thus, due to the presence of the residual indentation stress ®eld, both median/radial and lateral cracks (present in all indented specimens) can grow sub-critically in non-inert environments [24]. The consequence of this process is that the residual stress ®eld is partially relieved which in turn increases the strength. That could also have accounted for the 12% increase. The decrease in hardness from 6.4 to 4.6 GPa would also automatically lead to an increase in toughness of about 4%. An additional validation of the KIc values obtained with the IS technique using fractographic analyses [25] was not feasible. It was indeed impossible to assess properly the dimensions of the pertinent semi-elliptical cracks due to the slow crack growth that occurred during testing. Slow crack growth (SCG) is a primary origin in the failure of structural materials including ceramics and glasses [8±10]. It is often caused by the presence of reactants at low concentrations which are innocuous in the absence of a stress ®eld but may cause signi®cant damage when the structure is placed under load. SCG is described by
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
a fatigue parameter, the crack growth exponent n, which is a material constant for a given environment. For water, crack growth exponents have been determined for a feldspathic porcelain (Ceramco): n 14 [4], for a model feldspathic porcelain: n 28 [5], for an aluminous ceramic (VitadurN): n 29 [4], for a leucite reinforced ceramic (Optec-hsp): n 26 [6] and a heat-pressed ceramic (IPS-Empress): n 31) [7]. The larger the n value, the greater the resistance to crack growth in a corrosive environment. In that respect, determining the stress-corrosion susceptibility (n) of LFC would add to our understanding of the behavior of this material. In conclusion, the low fusing glass Duceram LFC exhibited mechanical properties (strength and toughness) that were comparable to those of conventional feldspathic porcelains. Although a signi®cant increase in toughness after water aging was observed, its importance should not be overestimated since the LFC's toughness still remained in the lower range of currently available ceramic materials. Acknowledgements
[5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]
The authors greatfully acknowledge the help of the following individuals: Mr A. Lavrentyev (Ohio State University), Dr L. Haenny and Mr D. Werner (Engineering School, Geneva), Dr C. Susz (Qualident), Mr R. Renevey CDT, Mrs M. C. Bijon as well as the Ducera company for providing the LFC material. This work has been supported in part by a research grant from the Swiss Prosthodontic Society.
[17] [18] [19] [20]
References [1] Risito C, LuÈthy H, Loeffel O, SchaÈrer P. Chemische LoÈslichkeit und Festigkeit von niedrigschmelzenden Dentalporzellanen. Schweiz Monatsschr Zahnmed 1995;105:611±6. [2] Jestel NL, Morris MD, O'Brien WJ. Depth-resolved Raman microprobe examination of a commercial dental porcelain exposed to a simulated oral environment. Dent Mater 1998;14:375±82. [3] Anusavice KJ, Lee RB. Effect of ®ring temperature and water exposure on crack propagation in unglazed porcelain. J Dent Res 1989;68:1075±81. [4] Morena R, Beaudreau GM, Lockwood PE, Evans AL, Fairhurst CW.
[21] [22] [23] [24] [25]
371
Fatigue of dental ceramics in simulated oral environment. J Dent Res 1986;65:993±7. Fairhurst CW, Lockwood PE, Ringle RD, Twiggs SW. Dynamic fatigue of feldspathic porcelains. Dent Mater 1993;9:269±73. Myers ML, Ergle JW, Fairhurst CW, Ringle RD. Fatigue characteristics of a high-strength porcelain. Int J Prosthodont 1994;7:253±7. Myers ML, Ergle JW, Fairhurst CW, Ringle RD. Fatigue failure parameters of IPS-Empress porcelain. Int J Prosthodont 1994;7:549±53. Wiederhorn SM. Moisture assisted crack growth in ceramics. Int J Fract Mechan 1968;4:171±7. Wiederhorn SM. Subcritical crack growth in ceramics. In: Bradt RC, Hasselman DPH, Lange FF, editors. Fracture mechanics of ceramics, vol. 2. New York: Plenum Press, 1974. p. 613±45. Evans AG. Slow crack growth in brittle materials under dynamic loading conditions. Int J Fract Mechan 1974;10:251±9. White SN, Li ZC, Yu Z, Kipnis V. Relationship between static chemical and cyclic mechanical fatigue in a feldspathic porcelain. Dent Mater 1997;13:103±10. Sherrill CA, O'Brien WJ. Transverse strength of aluminous and feldspathic porcelain. J Dent Res 1974;53:683±90. Drummond JL, Novickas D, Lenke JW. Physiological aging of an allceramic restorative material. Dent Mater 1991;7:133±7. Kelly JR. Perspectives on strength. Dent Mater 1995;11:103±10. Mecholsky Jr JJ. Fracture mechanics principles. Dent Mater 1995;11:111±2. Scherrer SS, Denry IL, Wiskott HWA. Comparison of three fracture toughness testing techniques using a dental glass and a dental ceramic. Dent Mater 1998;14:246±55. Anstis GR, Chantikul P, Lawn BR, Marshall DB. A critical evaluation of indentation techniques for measuring fracture toughness: I. Direct crack measurements. J Am Ceram Soc 1981;64:533±8. Chantikul P, Anstis GR, Lawn BR, Marshall DB. A critical evaluation of indentation techniques for measuring toughness: II. Strength method. J Am Ceram Soc 1981;64:539±43. Lawn BR, Marshall DB. Hardness, toughness, and brittleness: an indentation analysis. J Am Ceram Soc 1979;62:347±50. Weibull W. A statistical theory of the strength of materials. Ing Vetensk Akad Proc 1939;151:1±45. Mann NR, Schafer RE. Methods for the analysis of reliability and life data, New York: Wiley, 1974. p. 185±91. Thoman DR, Bain LJ, Antle CE. Inferences on the parameters of the Weibull distribution. Technometrics 1969;11:445±60. Quinn GD, Morrell R. Design data for engineering ceramics: a review of the ¯exure test. J Am Ceram Soc 1991;74:2037±66. Sglavo VM, Green DJ. In¯uence of indentation crack con®guration on strength and fatigue behaviour of sodal-lime silicate glass. Acta Metall Mater 1995;43:965±72. Scherrer SS, Kelly JR, Quinn GD, Xu K. Fracture toughness (KIc) of a dental porcelain determined by fractographic analysis. Dent Mater 1999;15:342±8.
www.elsevier.com/locate/dental
Effect of water exposure on the fracture toughness and ¯exure strength of a dental glass Susanne S. Scherrer a,*, Isabelle L. Denry b, H.W. Anselm Wiskott a, Urs C. Belser a a
Department of Prosthodontics, School of Dental Medicine, University of Geneva, Geneva, Switzerland b Department of Restorative and Prosthetic Dentistry, The Ohio State University, Columbus, OH, USA Received 9 August 2000; revised 6 December 2000; accepted 30 January 2001
Abstract Objectives: The low fusing dental glass (Duceram LFC) has been advertised as presenting a superior chemical resistance and augmented strength after 16 h exposure to water or 4% acetic acid. The purpose of this study was to evaluate the effect of prolonged exposure to water on two mechanical properties (fracture toughness and ¯exure strength) of LFC. Methods: Disks and bars were mirror polished and annealed prior to aging in: (1) air (control), (2) water for 24 h at 808C and (3) water for 8 weeks at 808C. Fracture toughness (KIc) was determined by indentation fracture (IF) and indentation strength (IS) using a 19.6 N Vickers indentation load. Flexure strength values were obtained from three-point bending at 0.1 mm/min. Statistical analysis was performed using the Weibull distribution, Tukey and Bartlett tests (P , 0.05). Results: Both techniques (IS and IF) showed a signi®cant improvement in the K of Duceram LFC after 8 weeks in water (0.88 and 1.14 MPa m 0.5) as opposed to the 24-h values both in water and air (0.77±0.78 MPa m 0.5). However, for ¯exure strength the Weibull characteristic (S0) and the m parameter did not change signi®cantly with water storage (S0 90±100 MPa, Weibull m 7±8). Signi®cance: The increase in toughness of Duceram LFC after aging in water is an interesting and favorable observation for a restorative material exposed to the oral environment. Nevertheless, in comparison with other contemporary ceramics, the toughness of this LFC remains in the range of soda-lime-glass or classic feldspar porcelains. q 2001 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Water exposure; Fracture toughness; Flexure strength; Dental glass
1. Introduction The low fusing dental glass Duceram LFC sinters at 6608C. This temperature is ca. 30% lower than that of conventional materials and extends the application range of this material to the repair of porcelain fractures (also after post-ceramic soldering) and to the ®nal adaptation of porcelain margins. Duceram LFC was initially marketed as a restorative material for the fabrication of inlays, onlays and veneers. The manufacturer (Ducera Dental, Germany) especially emphasized the material's superior resistance to chemical degradation in aqueous and acidic media. It was also determined that the ¯exure strength (S) of the material increased by 24% (S 123 MPa) after 16 h exposure to 4% acetic acid [1]. Such a behavior suggests an ion exchange mechanism which modi®es the surface's structural arrangement after exposure to speci®c environments. In a recently * Corresponding author. Tel.: 141-22-382-9129; fax: 141-22-781-1297. E-mail address: [email protected] (S.S. Scherrer).
reported experiment, Jestel et al. [2] used depth-resolved Raman microprobe analyses to examine the surface of Duceram LFC samples after exposure to water and 4% acetic acid. They demonstrated that structured zones containing metaborate, pyroborate, boroxol rings and a mixture of borate and silicate tetrahedra developed on the outer surface of the glass. Pending further investigation, they hypothesized that the increase in strength after exposure to water or acetic acid was due to this newly developed 3±10 mm thick boron-containing surface layer. Yet a strengthening mechanism linked to the exposure to water would be unique among ceramic restorative materials. Indeed most feldspar-based ceramics show strong tendencies towards stress corrosion [3±7]Ða phenomenon which originates in an increasing hydrolysis of the ceramic when subjected to stress application and results in a time-dependent reduction in ¯exural strength [8±13]. The limitations of ¯exural strength as an indicator of the structural performance of brittle materials have been pointed out by Kelly [14]. By contrast, an appropriate parameter would
0109-5641/01/$20.00 + 0.00 q 2001 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S 0109-564 1(01)00002-1
368
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
be the ceramics' fracture toughness (KIc), that is the material's intrinsic resistance to crack propagation [15]. Ceramic materials are susceptible to stress corrosion, dynamic fatigue and surface degradation which all affect their ¯exure strength. It is unclear, however, whether their toughness also is affected by exposure to an accelerated aging environment. Therefore, the aim of the present study was to determine the fracture toughness and the Weibull strength parameters (S0 and m) of a lowfusing dental glass (Duceram LFC) which had been previously aged up to 2 months in 808C water. 2. Materials and methods 2.1. Specimen preparation Disks (13 £ 2 mm 2) and bars (25 £ 4.5 £ 2 mm 3) were prepared from a low fusing dental glass (Duceram LFC dentin, Ducera Dental GmbH, Rosbach, Germany, batch no. CE 0483). After ®ring, the specimens were ground ¯at and parallel with a diamond abrasive disk (600 grit). The test side was further polished using a series of SiC abrasive disks (P#1200, P#2400 and P#4000) and diamond pastes (6, 3 and 1 mm). To relieve residual surface stresses due to grinding and polishing, all samples were annealed. The annealing temperatures for the low fusing dental glass and the feldspathic porcelain were determined from dilatometric analyses (Orton dilatometer, Model 1600D) and reported in a previous study (Scherrer et al. [16]). Annealed specimens were then distributed in three groups: (1) aging in ambient air (room temperature, 1 week), (2) aging in deionized water at 808C for 24 h and (3) aging in deionized water at 808C for 8 weeks. After aging, all specimens were ultrasonically cleaned in ethanol for 5 min and air dried prior to testing. 2.2. Fracture toughness and ¯exure strength determination Fracture toughness (KIc) was determined using the indentation fracture (IF) [17] and the indentation strength (IS) [18] technique. For the IF technique, disks were gold-coated prior to indentation to facilitate the crack length readings. Vickers indentations were produced in ambient air at 19.6 N until 10 acceptable crack patterns were obtained [16]. No drop of oil was used to prevent slow crack growth. However, crack length readings were performed using the indenters microscope within 30 s after indentation to minimize slow crack growth due to moisture presence in ambient air. Fracture toughness was calculated according to the equation developed by Anstis et al. [17] which relates the indentation load (P), the size of the median cracks (c), the modulus of elasticity (E) and the material' hardness (H): KIc 0:016 E=H0:5 P=c1:5
1
Hardness (H) was de®ned as the load divided by the projected area on the surface and calculated as: H 0.5P/ a 2 [19] where a is the average indentation half-diagonal. It was determined from 10 indentations using a load of 1.96 N,
a magnitude that prevented the formation of radial cracks. The material's density, dynamic E and Poisson's ratio, were previously measured using the ultrasonic velocity method [16]. Only E values for air and 24-h water specimens were measured. As the modulus of elasticity did not change, the value obtained for the 24-h water group was also used for the 8-week water specimens for toughness calculation. The second procedure used for determining toughness was the indentation strength (IS) which consisted of producing a Vickers indentation of 19.6 N on the center of the bar's mirror polished side after aging and shortly before testing. Ten indented specimens per aging group were subjected to a three-point-bending test under the following conditions: cross-head speed: 0.1 mm/min, span: 21 mm, ambient air. The stress at fracture (s f) was calculated as:
s f 3WL=2BD2
2
where W is the breaking load, L the span, B the specimen's width and D its thickness. Fracture toughness was calculated from a uni®ed indentation-fracture/tensile-failure equation [18] that relates the modulus of elasticity (E), indentation load (P), hardness (H) and stress at fracture (s f): K Ic 0:59 E=H1=8 s f P1=3 3=4
3
Hardness values were obtained from the disk specimens as previously described. A test was considered successful after inspection of the specimen's matching halves under magni®cation when it was veri®ed that the fracture process did indeed start from the indentation ¯aw. Again, the modulus of elasticity value for the 24-h group was used for the 8week water group. LFC ¯exure strength was determined from 20 bars per aging group. Specimens were polished and beveled on their edges, annealed and aged as described above. The ¯exure strength was determined from three-point-bending at 0.1 mm/min and 21 mm span. 2.3. Statistical analysis The mean fracture toughness of each group was computed and differences among groups tested for statistical signi®cance with Bartlett's and Tukey's multiple range test. A P , 0.05 was considered signi®cant. The Weibull distribution was used to calculate the fracture probability as a function of applied stress [20]. The Weibull formula (for the two parameter distribution) which yields the cumulative probability of failure F(S) as a function of applied stress (S) is given by: m F S 1 2 e
2
S S0
4
where S is the stress at a given point, S0 is the characteristic strength i.e. the stress level at which 63% of the specimens have failed and m (the Weibull modulus) a constant which characterizes the spread of the failure data. The distribution function can also be expressed in (natural) logarithmic
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
369
Table 1 Fracture toughness (KIc) determined by indentation fracture (IF) and indentation strength (IS) techniques after aging Duceram LFC in deionized water at 808C for 24 h and 8 weeks. Number of tests (n), Hardness (H), modulus of elasticity (E). The modulus of elasticity (E) were determined previously [16]. Values with the same superscript were not signi®cantly different (P . 0.05) Duceram LFC
n
E (GPa)
H (GPa) 1.96 N load
IF KIc (MPa m 0.5) 19.6 N load
Crack length IF (mm)
IS KIc (MPa m 0.5) 19.6 N load
Air H2O/24 h H2O/8 weeks
10 10 10
69.6 (0.6) 68.2 (2.5) ±
6.4 (0.5) 5.5 (0.4) 4.6 (0.2)
0.77 (0.09) a 0.77 (0.06) a 1.14 (0.09) b
122.4 (10.1) 127.6 (7.6) 104.4 (5.8)
0.78 (0.08) a 0.78 (0.08) a 0.88 (0.10) b
Table 2 Weibull characteristic strength (S0) and m parameter with corresponding con®dence intervals (CI) of LFC aged specimens tested in three-point bending. Number of specimens (n). Values with the same superscript are not signi®cantly different using con®dence intervals at 95% LFC aging groups
n
Weibull parameter m
Con®dence interval of m
Characteristic strength S0 (MPa)
Con®dence interval of S0 (MPa)
Correlation coef®cient r
Air 24 h H2O 8 weeks H2O
20 20 20
8.0 a 6.9 a 7.2 a
5.0±10.7 4.3±9.2 4.6±9.7
92.1 b 100.2 b 89.6 b
86.1±98.7 92.6±108.6 83.1±96.6
0.940 0.994 0.982
form: lnln
1 1 2 F S
nearly overlaps the air-aging group. The lnln transforms represented in Fig. 2 show similar slopes.
ln S 2 ln S0
5
The linear correlation coef®cient r of the data obtained from the lnln plots provides a measure of the applicability of the distribution. A r value above 0.80 is considered acceptable. The maximum likelihood estimates for the characteristic strength S0 and m value were determined for each group using iterative procedures (Newton±Raphson) as described by Mann et al. [21]. The 95% con®dence intervals were calculated for S0 and m using the tables established by Thoman et al. [22].
3. Results Table 1 lists the hardness (H), the modulus of elasticity (E) and the toughness obtained using the indentation fracture and indentation strength techniques. The toughness increased signi®cantly for both the IF and the IS techniques after 2 months of aging in water. The difference was slight (13%) for the IS technique and marked for IF (48%). No differences were found between 24-h water exposure and aging in air using either IF or IS. Table 2 lists the Weibull parameters S0 and m as well as their respective con®dence intervals. The correlation coef®cient r as determined from the lnln transformed ¯exure strength data is also included. The S0 and m parameters were not signi®cantly different for all three aging groups. The cumulative failure plots and lnln transforms are shown in Figs. 1 and 2. A slight shift to the right (i.e. a minor improvement) was observable in Fig. 1 for the 24-h water specimens. However, this tendency was not con®rmed after 2 months in water and the cumulative probability of failure
4. Discussion The purpose of this study was to determine the ¯exure strength and the toughness of a dental glass ceramic (Duceram LFC) after prolonged exposure to water (8 weeks). The manufacturer promoted LFC as a material whose ¯exure strength was augmented by exposure to waterÐa claim that was indeed con®rmed in one study [1]. The shortcomings of ¯exure tests for ceramic materials have been reviewed by Quinn and Morrell [23] especially with respect to their sensitivity to surface ®nishing procedures. Brittle materials such as ceramics will fail due to the progression of existing ¯aws when
Fig. 1. Plot of the cumulative probability of failure vs. the applied stress of Duceram LFC after aging in air, 24 h in water at 808C and 8 weeks in water at 808C.
370
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
Fig. 2. Plot of the lnln transform of the cumulative probability of failure vs. the ln of the applied stress for the LFC after aging in air, 24 h in water and 8 weeks in water.
subjected to stresses above a critical level. Hence, using the Weibull distribution for the assessment of the material's characteristic ¯exure strength (S0) and Weibull modulus (m) is consistent with the `weakest link theory' [20], which states that a brittle specimen's breakage process will originate from its largest ¯aw. The size of the worst defect depends on the sample and thus induces a scatter in the recorded material strength values. This variability can be analyzed statistically using the Weibull distributionÐa model which assumes that a ®nite strength value is associated with each volume or surface element in the material. In a bend test, the stress distribution is nonuniform varying from zero at the neutral axis to a maximum at the surface under tension. This markedly increases the effect of the surface state on an experimentally determined ¯exure strength. Therefore, care was taken in the present study to mirror polish (^1 mm) and anneal all LFC specimens in order to eliminate residual compressive stresses left after polishing. When these ®nishing procedures were applied, no changes in the Weibull parameters were detected among the air-, 24-h water- and 8-week water specimens. A slight shift to the right of the 24-h cumulative failure curve (Fig. 1) was observed but this trend was not con®rmed after 2 months of exposure to water. All Weibull m parameters were extremely close (Table 2, Fig. 2), indicating that no changes occurred in the material's ¯aw population subsequently to water exposure. The second material property evaluated in the present study was fracture toughness as a function of time and water exposure. As shown in Table 1, a signi®cant increase was observed in the 8-week water specimens regardless of which testing technique was used. After 2 months of exposure to water, the IF technique yielded an increase of 48% in toughness and the IS procedure an increase of 13%. The origin of the observed improvement is unclear. Several possible explanations are given below. First, hardness (H) determined by indentation gradually decreased with time and water exposure (Table 1), thereby
strongly suggesting changes in surface structure. This observation has a bearing on the increase in toughness. Indeed, if we assigned a constant value to hardness (that of the air group) and used Eq. (1) (IF), the recomputed KIc for the 8week water group would be 0.96, that is an increase of 25% (vs. 48%). This increase of 25% is still signi®cantly different from the air- and 24-h water groups and demonstrates the relation between hardness and toughness when applying Eq. (1). It should also be mentioned that the indentation load for hardness measurements was small (1.96 N) to avoid the formation of radial cracks. Under this load, the Vickers indenter's half diagonal is approx. 15 mm and the depth of the indent about 6 mm. With reference to the data by Jestel et al. [2], our Vickers indentation for the H measurements was thus mainly determined by the boron-containing surface layer. Second, one should remember the intricacies of indentation fracture techniques [16,17] with respect to crack lengths measurements. The IF technique is based on a Vickers induced crack system of the radial/median type. This penny-shape crack con®guration may be modi®ed by the presence crack neighbours (i.e. lateral cracks) and the immediate exposure of the indented zone to a reactive environment (moisture) will contribute to slow crack growth. As oil was not used to prevent slow crack growth during indentation it could well be that crack length readings were affected by moisture in air even though the time between indentation and reading was kept to a minimum. The toughness results obtained with the IS technique seemed to be more reliable that those of the IF due to the absence of possible error during crack length readings. However, the same critique about moisture interaction would apply to the IS technique as the time between indentation and testing was approximately 30 min to 1 h. Even though the testing timeframe was similar for all groups, the 8-week water specimens may perhaps have had more slow crack growth due to the presence of water molecules within the LFC's surface. Thus, due to the presence of the residual indentation stress ®eld, both median/radial and lateral cracks (present in all indented specimens) can grow sub-critically in non-inert environments [24]. The consequence of this process is that the residual stress ®eld is partially relieved which in turn increases the strength. That could also have accounted for the 12% increase. The decrease in hardness from 6.4 to 4.6 GPa would also automatically lead to an increase in toughness of about 4%. An additional validation of the KIc values obtained with the IS technique using fractographic analyses [25] was not feasible. It was indeed impossible to assess properly the dimensions of the pertinent semi-elliptical cracks due to the slow crack growth that occurred during testing. Slow crack growth (SCG) is a primary origin in the failure of structural materials including ceramics and glasses [8±10]. It is often caused by the presence of reactants at low concentrations which are innocuous in the absence of a stress ®eld but may cause signi®cant damage when the structure is placed under load. SCG is described by
S.S. Scherrer et al. / Dental Materials 17 (2001) 367±371
a fatigue parameter, the crack growth exponent n, which is a material constant for a given environment. For water, crack growth exponents have been determined for a feldspathic porcelain (Ceramco): n 14 [4], for a model feldspathic porcelain: n 28 [5], for an aluminous ceramic (VitadurN): n 29 [4], for a leucite reinforced ceramic (Optec-hsp): n 26 [6] and a heat-pressed ceramic (IPS-Empress): n 31) [7]. The larger the n value, the greater the resistance to crack growth in a corrosive environment. In that respect, determining the stress-corrosion susceptibility (n) of LFC would add to our understanding of the behavior of this material. In conclusion, the low fusing glass Duceram LFC exhibited mechanical properties (strength and toughness) that were comparable to those of conventional feldspathic porcelains. Although a signi®cant increase in toughness after water aging was observed, its importance should not be overestimated since the LFC's toughness still remained in the lower range of currently available ceramic materials. Acknowledgements
[5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]
The authors greatfully acknowledge the help of the following individuals: Mr A. Lavrentyev (Ohio State University), Dr L. Haenny and Mr D. Werner (Engineering School, Geneva), Dr C. Susz (Qualident), Mr R. Renevey CDT, Mrs M. C. Bijon as well as the Ducera company for providing the LFC material. This work has been supported in part by a research grant from the Swiss Prosthodontic Society.
[17] [18] [19] [20]
References [1] Risito C, LuÈthy H, Loeffel O, SchaÈrer P. Chemische LoÈslichkeit und Festigkeit von niedrigschmelzenden Dentalporzellanen. Schweiz Monatsschr Zahnmed 1995;105:611±6. [2] Jestel NL, Morris MD, O'Brien WJ. Depth-resolved Raman microprobe examination of a commercial dental porcelain exposed to a simulated oral environment. Dent Mater 1998;14:375±82. [3] Anusavice KJ, Lee RB. Effect of ®ring temperature and water exposure on crack propagation in unglazed porcelain. J Dent Res 1989;68:1075±81. [4] Morena R, Beaudreau GM, Lockwood PE, Evans AL, Fairhurst CW.
[21] [22] [23] [24] [25]
371
Fatigue of dental ceramics in simulated oral environment. J Dent Res 1986;65:993±7. Fairhurst CW, Lockwood PE, Ringle RD, Twiggs SW. Dynamic fatigue of feldspathic porcelains. Dent Mater 1993;9:269±73. Myers ML, Ergle JW, Fairhurst CW, Ringle RD. Fatigue characteristics of a high-strength porcelain. Int J Prosthodont 1994;7:253±7. Myers ML, Ergle JW, Fairhurst CW, Ringle RD. Fatigue failure parameters of IPS-Empress porcelain. Int J Prosthodont 1994;7:549±53. Wiederhorn SM. Moisture assisted crack growth in ceramics. Int J Fract Mechan 1968;4:171±7. Wiederhorn SM. Subcritical crack growth in ceramics. In: Bradt RC, Hasselman DPH, Lange FF, editors. Fracture mechanics of ceramics, vol. 2. New York: Plenum Press, 1974. p. 613±45. Evans AG. Slow crack growth in brittle materials under dynamic loading conditions. Int J Fract Mechan 1974;10:251±9. White SN, Li ZC, Yu Z, Kipnis V. Relationship between static chemical and cyclic mechanical fatigue in a feldspathic porcelain. Dent Mater 1997;13:103±10. Sherrill CA, O'Brien WJ. Transverse strength of aluminous and feldspathic porcelain. J Dent Res 1974;53:683±90. Drummond JL, Novickas D, Lenke JW. Physiological aging of an allceramic restorative material. Dent Mater 1991;7:133±7. Kelly JR. Perspectives on strength. Dent Mater 1995;11:103±10. Mecholsky Jr JJ. Fracture mechanics principles. Dent Mater 1995;11:111±2. Scherrer SS, Denry IL, Wiskott HWA. Comparison of three fracture toughness testing techniques using a dental glass and a dental ceramic. Dent Mater 1998;14:246±55. Anstis GR, Chantikul P, Lawn BR, Marshall DB. A critical evaluation of indentation techniques for measuring fracture toughness: I. Direct crack measurements. J Am Ceram Soc 1981;64:533±8. Chantikul P, Anstis GR, Lawn BR, Marshall DB. A critical evaluation of indentation techniques for measuring toughness: II. Strength method. J Am Ceram Soc 1981;64:539±43. Lawn BR, Marshall DB. Hardness, toughness, and brittleness: an indentation analysis. J Am Ceram Soc 1979;62:347±50. Weibull W. A statistical theory of the strength of materials. Ing Vetensk Akad Proc 1939;151:1±45. Mann NR, Schafer RE. Methods for the analysis of reliability and life data, New York: Wiley, 1974. p. 185±91. Thoman DR, Bain LJ, Antle CE. Inferences on the parameters of the Weibull distribution. Technometrics 1969;11:445±60. Quinn GD, Morrell R. Design data for engineering ceramics: a review of the ¯exure test. J Am Ceram Soc 1991;74:2037±66. Sglavo VM, Green DJ. In¯uence of indentation crack con®guration on strength and fatigue behaviour of sodal-lime silicate glass. Acta Metall Mater 1995;43:965±72. Scherrer SS, Kelly JR, Quinn GD, Xu K. Fracture toughness (KIc) of a dental porcelain determined by fractographic analysis. Dent Mater 1999;15:342±8.