- Email: [email protected]
Influence of cyclic fatigue in water on the load-bearing capacity of dental bridges made of zirconia
Available online at www.sciencedirect.com
Acta Biomaterialia 4 (2008) 1440–1447 www.elsevier.com/locate/actabiomat
Influence of cyclic fatigue in water on the load-bearing capacity of dental bridges made of zirconia Philipp Kohorst *, Marc Philipp Dittmer, Lothar Borchers, Meike Stiesch-Scholz Department of Prosthetic Dentistry and Biomedical Materials Science, Hannover Medical School, Carl-Neuberg-Straße 1, 30625 Hannover, Germany Received 26 October 2007; received in revised form 2 April 2008; accepted 14 April 2008 Available online 1 May 2008
Abstract The humid atmosphere and permanent occurrence of chewing forces in the oral environment lead to degradation of ceramics used for prosthetic restorations. The aim of this in vitro study was to evaluate the influence of artificial aging on the load-bearing capacity of fourunit bridges, with both undamaged and predamaged zirconia frameworks. Additionally, different parameters for chewing simulation have been investigated and a finite element analysis was made to predict the location of highest tensile stresses within the bridges. A total of 60 frameworks were milled from presintered zirconia and divided into six homogeneous groups. Prior to veneering, frameworks of two groups were ‘‘damaged” by a defined saw cut similar to an accidental flaw generated during shape cutting. After veneering, FPDs were subjected to thermal and mechanical cycling – with the exception of control groups. The load-bearing capacity of tested FPDs was significantly reduced by artificial aging. In comparison to unaged specimens, fracture resistance decreased by about 40%, whereas preliminary damage did not have a significant effect. Increasing number of cycles and increasing upper load limit failed to show any additional effect on fracture force. To predict the progression of degradation under the terms of in vitro simulation for even longer periods, further aging experiments are required. Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Dental bridges; Zirconia; Load-bearing capacity; Cyclic loading; FEA
1. Introduction All-ceramic dental restorations have become more and more important, in particular, due to their favorable esthetics and their outstanding biological compatibility. The recent incorporation of high-strength zirconia into dentistry has increased the spectrum of indications for ceramic materials, allowing fabrication of implants, allceramic crowns and fixed partial dentures (FPDs) for the highly loaded posterior region [1]. This type of zirconia owes its high-strength to so-called transformation reinforcement, a complex mechanism involving transformation from tetragonal-to-monoclinic structure and associated with a local 4% increase in volume [2,3]. For this process to be effective, the tetragonal structure must be stabilized *
Corresponding author. Tel.: +49 511 532 4773; fax: +49 511 532 4790. E-mail address: [email protected] (P. Kohorst).
at room temperature, for which reason almost all zirconium dioxide ceramics used for dental restorations contain yttrium oxide (Y2O3). Under functional loading in the oral environment, zirconia-based materials, despite their remarkably high-strength, undergo fatigue and subcritical crack growth that can significantly reduce their strength over time [4]. Subcritical crack growth occurs due to the stress-assisted reaction of water molecules with the ionic–covalent bonds at the crack tip [5]. Furthermore, water molecules can be incorporated into the zirconia lattice by filling oxygen vacancies [6]. This reduces the energy barrier for the tetragonal-to-monoclinic transformation and thus increases the rate of transformation which, due to the associated volume increase, results in microcrack formation within the lattice. Although this mechanism is usually very slow at oral temperatures, it may lead to a significant decrease in the strength of dental restorations over periods of wear of several decades [7].
1742-7061/$ - see front matter Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2008.04.012
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
Mechanical loading and a moist environment are the conditions encountered in the mouth during mastication. Thus, determining the influence of degradation effects to load-bearing capacity under cyclic stresses in water is prerequisite to predict the sustainable success of zirconia-based restorations in dentistry. There are only a few investigations that have dealt with the strength-degrading effect of cyclic loading in water on zirconia ceramics [8–10]. Besides cyclic loading in a moist environment, flaws within the ceramic structure could lead to subcritical crack growth and, eventually, catastrophic failure [11]. These defects may be conditional on the manufacturing of the zirconia blanks used or could be created by the technician on the material’s surface during production or finishing. In particular, the gingival embrasure of FPDs is the most sensitive site for any kind of flaw. By means of finite element analysis (FEA) and fractography, Kelly et al. determined that the highest tensile stress within the connectors of an all-ceramic three-unit FPD is at this site [12]. They evaluated the failure mode of FPDs made of glass-infiltrated alumina, taking into account the resilient embedding of abutment teeth. The aim of the present study was to test the hypothesis that the in vitro load-bearing capacity of a posterior dental bridge made of zirconia is reduced by aging in an artificial oral environment. Additionally, the hypotheses were tested that an increasing number of mechanical cycles or an increasing upper load limit and a defined mechanical damage to the framework at a sensitive location causes a decrease in load-bearing capacity. Furthermore, the site of the highest tensile stresses within the zirconia FPD was determined by means of FEA. 2. Materials and methods 2.1. Preparation
1441
many). The frames were made by repeated copying of a master frame in a computer-aided machinery (CAM) unit (Cercon brain, DeguDent, Hanau, Germany) with subsequent firing in the system oven (Cercon heat, DeguDent, Hanau, Germany). The dimensions of all frames were practically the same, connector width and height differing by less than 0.2 mm. Connector cross-sections were elliptically shaped, their areas being (from mesial to distal) 12.5, 15.6 and 11.6 mm2, respectively. Before veneering, 20 frames were selected at random and a U-shaped cut of width 180 lm and depth 60 lm was made at the gingival surface of the connector between teeth 25 and 26 (the presumed location of highest tensile stress during loading) (Fig. 1). This was done to simulate possible damage to the core during the manufacturing process and to test the influence on load-bearing capacity. Such cutting (subsequently referred to as ‘‘predamaging”) was performed with an annular diamond saw (Microslice 2, Metals Research Ltd., Royston, UK). After veneering, the undamaged specimens were also randomized and divided into homogeneous groups of 10 each, resulting in a total of six groups. 2.3. Aging The bridges were fixed onto the PUR abutments by means of glass-ionomer cement (Ketac-Cem, 3M ESPE, Seefeld, Germany). With the exception of two groups (undamaged/predamaged), the bridges were subjected to thermal and mechanical cycling (TMC) during 200 days storage in distilled water at 36 °C. During this period, 1 104 thermal cycles between 5 and 55 °C (30 s dwell time at each temperature) and a mechanical loading (load frequency 2.5 Hz) with varying number of cycles (1 106/ 2 106) and varying upper load limits (100 N/200 N) were applied successively (Table 1 and Fig. 2).
In an upper jaw plastic model (Frasaco OK 119, A-3 T, Franz Sachs & Co., Tettnang, Germany), teeth 24 and 27 were prepared to accommodate a four-unit all-ceramic FPD. Stone casts (Fuji Rock, GC, Leuven, Belgium) were produced by means of individual impressions of the prepared teeth, and were used as a basis for manufacturing zirconium dioxide bridge frames. During aging and static testing, the FPDs were cemented to and supported by duplicates of the prepared original abutment teeth. These duplicates were made of reinforced polyurethane (PUR) resin (Alpha-Die-Top, Schu¨tz Dental, Rosbach, Germany). Natural periodontal resilience was simulated by coating the roots of these stumps with an elastic latex material (Erkoskin, Erkodent, Pfalzgrafenweiler, Germany). The latexcoated roots were imbedded in a PUR base that reached as far as 3 mm below the preparation margin. 2.2. Manufacture of FPDs A total of 60 frameworks were fabricated in partially sintered zirconia (Cercon base, DeguDent, Hanau, Ger-
Fig. 1. SEM image of preliminary damage generated on the gingival embrasure of a zirconia framework.
1442
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
Table 1 Variations in fatigue parameters and preliminary damage Group Cer Cer Cer Cer Cer Cer
I II III IV V VI
Predamage
Thermal cycles
Mechanical cycles
Upper load limit
1 104 1 104 1 104 1 104
1 106 1 106 2 106 1 106
100 N 100 N 100 N 200 N
+ +
not influenced by artificial aging and predamaging. Weibull parameters F0 and m were determined for each test group by fitting a Weibull distribution to the respective data set. The parameter F0 (characteristic force at failure) is associated with 63.2% probability of failure, whereas the Weibull modulus, m, is a measure of the scatter in the force at failure and of the reliability of the material investigated. The greater the value for m, the steeper the transition from survival to failure for the probability distribution against force at failure. 2.6. Finite element analysis One of the investigated FPDs was scanned optically before and after veneering by the principle of triangulation (Atos II SO, GOM, Braunschweig, Germany). Based on the resulting polygon meshes, two three-dimensional bulk models were created by reverse engineering. Afterwards models were imported into a CAD program (DesignModeler, Ansys Inc., Canonsburg, PA). In a Boolean operation, the zirconia framework was virtually subtracted from the veneered bridge, whereby a separate model of the veneering material was generated. According to the in vitro study model, abutment teeth, including the cement layer and resilient support in the polyurethane block, were designed virtually by means of the CAD software. Then the model was transferred to an FEA program (DesignSpace, Ansys Inc.) for finite element stress analysis. 3. Results
Fig. 2. Schematic of cyclic loading set-up.
2.4. Determination of load-bearing capacity Following the aging procedure, the specimens were loaded until fracture. This was carried out in a universal testing machine (Type 20K, UTS Testsysteme, Ulm-Einsingen, Germany) at a crosshead speed of 1 mm min 1 with the force transferred to the occlusal connector area between teeth 25 and 26 via a tungsten carbide ball (diameter 6.0 mm) on an interposed tin foil (thickness 0.2 mm). A sudden decrease in force of more than 15 N was regarded as indication of failure and the maximum force up to this point recorded as force at fracture. 2.5. Statistics The statistical analysis of force at fracture data was performed using one-way analysis of variance (ANOVA) for detecting the influence of artificial aging and two-factor ANOVA for detecting the influence of predamage, with the level of significance chosen at 0.05 (SPSS 15.0, SPSS Software Corp., Chicago, IL). A direct comparison of group means was carried out with post hoc Scheffe´ test. The null hypotheses were that load-bearing capacity is
All FPDs tested survived thermal and mechanical cycling in the artificial oral environment. Table 2 shows the results of load-bearing capacity testing. After the failure criterion (15 N load drop) had been fulfilled, visual inspection revealed that 57 of 60 FPDs (95%) had failed with sudden bulk fracture of the zirconia framework. Two bridges showed Hertzian cracking beneath the indenter and one bridge extensive delamination of the veneering ceramic before bulk fracture occurred at higher force levels. The artificial aging by thermal and mechanical cycling proved to have a significant influence on the load-bearing capacity of both the undamaged and the predamaged zirconia FPDs (p < 0.001). In comparison to the control groups, Cer I (1525 N) and Cer II (1335 N), the aged specimens exhibited a broad decrease in average force at fracture (Fig. 3). However, variation of mechanical cycling parameters (number of cycles/upper load limit) failed to show a significant influence on load-bearing capacity (Table 3). Means were 904 N (Cer III), 924 N (Cer V) and 952 N (Cer VI) (Table 2 and Fig. 4). Furthermore, the effect of mechanical predamage did not prove to be statistically significant (p = 0.213) for either the unaged (Cer II/1335 N) or the aged (Cer IV/921 N) specimens (Table 2 and Fig. 3). Therefore, the first hypothesis formulated in the introduction could be accepted, whereas the others had to be rejected.
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
1443
Table 2 Mean force at fracture (SD) and Weibull parameters F0 and m, together with their 95% confidence intervals Group
Load at fracture (N)
F0 (N)
95% confidence interval Lower limit
Upper limit
Cer Cer Cer Cer Cer Cer
1525.0 1334.7 903.7 921.1 923.5 952.4
1625.8 1445.9 958.6 990.6 976.3 1015.8
1580.3 1391.4 927.8 955.5 948.5 957.4
1671.4 1500.3 989.3 1025.8 1004.1 1074.1
I II III IV V VI
(76.5) (89.4) (40.8) (55.6) (40.3) (51.4)
Fig. 3. Box chart representing forces at fracture vs. preliminary damage, with dependency on cyclic loading (1 106 cycles/100 N upper load limit). Medians, quartiles and extremes are given.
Table 3 Results of statistical post hoc comparison of testing groups subjected to different fatigue regimes (p < 0.05) Cer I Cer Cer Cer Cer
I III V VI
<0.001 <0.001 <0.001
Cer III
Cer V
Cer VI
<0.001
<0.001 0.995
<0.001 0.942 0.987
0.995 0.942
0.987
As follows from the peculiarities of Weibull statistics, Weibull characteristic forces were around 6–8% higher than the corresponding mean forces at fracture and exhibited the same tendencies (Table 2). A higher number of mechanical cycles was associated with the statistically non-significant increase in the Weibull modulus from 7.2 (Cer I) to 8.0 (Cer III) to 8.6 (Cer V). However, raising the upper load limit did not have the same effect, with a Weibull modulus of 7.2 (Cer VI), as in the control group (Cer I) (Table 2). Mechanical predamage resulted in a decrease in the Weibull modulus m. This could be observed for both the unaged (decrease from 7.2 to 5.4) and aged (decrease from 8.0 to 6.2) specimens.
m
95% confidence interval Lower limit
Upper limit
7.2 5.4 8.0 6.2 8.6 7.2
6.1 4.6 6.4 5.1 6.9 4.7
8.4 6.3 9.7 7.2 10.3 9.8
Fig. 4. Box chart representing forces at fracture vs. different cyclic loading parameters. Medians, quartiles and extremes are given.
FE analysis revealed the location of highest tensile stresses to be in the zirconia framework close to its surface at the gingival embrasure of the connector between premolar and molar (Figs. 5 and 6). 4. Discussion In most studies investigating load-bearing capacity of all-ceramic restorations, alloys are used to create abutment teeth for in vitro testing. In the present study, model materials and testing conditions were chosen carefully to imitate clinical reality as faithfully as possible. The abutment teeth and their bases were made of reinforced PUR, which has an elastic modulus somewhat lower than dentin and bone but represents a better approximation of natural conditions than alloys, for example. Models made of alloys exhibit very high rigidity, so dental bridges are better supported during static and cyclic loading compared to conditions in the oral cavity. This results in an artificially high loadbearing capacity [13]. The latex layer around the abutment roots simulated a periodontal resilience of 30–95 lm against forces of between 50 and 100 N in the axial direction [14]. Extrapolating measurements of natural periodontal resilience men-
1444
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
Fig. 5. Stress distribution on framework surface during occlusal loading according to FEA (veneering layer removed for clarity). Highest tensile stress occurred within the framework at the gingival area of the connector.
Fig. 6. Cross-sectional area of the FE model with mesial view of the median connector. Highest tensile stress within the zirconia framework at transition to veneering ceramic.
tioned in the literature [15,16] as the chosen conditions is a sufficient approximation. A simple rigid support would not consider additional compressive, tensile and shear stresses caused by occlusal loading and lateral movements. Aqueous environments like the oral cavity generally facilitate subcritical crack growth in ceramics [17] and may cause uncontrolled transitions from the tetragonal to the monoclinic structure of Y-TZP [7]. Both produce a
decrease in strength of the ceramic. Other investigations have shown that the spontaneous transformation of zirconia in water initially proceeds rapidly but remains static after some point. For example, Drummond reported only a minor further decrease in strength of zirconia after approximately 300 days of water storage [18]. Consequently the bulk of degradation could be expected to have appeared during the 200 days of water storage applied in
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
the present study. However, it has to be taken into account that the zirconia specimens evaluated in the present study were veneered with a feldspathic ceramic and hence were not in direct contact with the humid environment. It remains to be clarified if the veneering layer permitted sufficient water diffusion to promote zirconia degradation or if it rather acted as a protective layer, as was reported for a combination of silica and zirconia [19]. In the course of water storage, specimens were also subjected to thermal cycling. Material strength is degraded by the repeated thermal stressing of prosthetic restorations, which results from temperature changes caused by the consumption of hot and cold food and drink, and breathing. This thermal stressing leads to tensions within the specimens and become manifest in slow subcritical crack growth and catastrophic failure eventually. Several investigators found a decrease in strength [20] and the development of flaws [21] during thermal cycling for specimens made of feldspathic ceramics. Finally, FPDs were subjected to mechanical cycling in an aqueous environment. The repeated application of chewing forces contributes to the decrease in ceramic strength during service [13]. This was simulated by mechanical cycling with upper limits of 100 and 200 N. Since a recent survey has revealed the projected number of chewing cycles to be about 800,000 per year [13], the 1 106 and 2 106 cycles applied in this study corresponded to an in vivo service period of approximately 15 and 30 months, respectively. The cyclic loads applied in the present study represent high levels of physiological chewing forces, which have been found to range between 20 and 120 N, depending on food consistency [22]. Rosentritt et al. found a decrease in load-bearing capacity of all-ceramic restorations due to increased simulated chewing forces [13]. Based on these results, it appears reasonable to raise the upper load limit during mechanical cycling beyond physiological forces, thus improving the estimation of the restorations’ performance under more severe conditions like bruxism. In all groups tested, artificial aging resulted in a significant decrease in load-bearing capacity of approximately 40%, whereas variation in aging parameters did not show significant influences. Other authors investigating dental restorations made of zirconia also recorded reduced strength by about 20% due to aging [9,23]. Rosentritt et al. reported a reduction in strength in an artificial oral environment for FPDs made of a pressable glass–ceramic. But, contrary to the present study, they described a significant decrease in load-bearing capacity due to an increase in the upper load limit from 50 to 150 N [13]. These findings are most likely associated with the special properties of the glass–ceramic tested, which is of lower strength and also greater sensitivity to subcritical crack growth than zirconia [10]. Contrary to the above investigations, Curtis et al. did not observe a decrease in load-bearing capacity of standardized zirconia specimens after mechanical cycling, either in dry or in humid atmospheres [8]. They even
1445
increased the upper load limit to a range from 500 to 800 N, whereas only 2000 cycles were performed. Sundh et al. also did not detect any reduction in the strength of zirconia bridges subjected to dynamic loading in water for 1 105 cycles with an upper load limit of 50 N [24]. An analogous observation was made by Studart et al. for zircona samples [4]. Specimens’ strength remained unaltered for a long period during mechanical cycling, but then abruptly dropped and settled at a lower level. This material behavior shows certain parallels to the results of the present study where neither doubling the load nor doubling the number of cycles induced a change in load-bearing capacity. For zirconia specimens stored in a humid atmosphere, Drummond observed an analogous phenomenon [18]. Though mechanical cycling did not greatly reduce the strength of zirconia specimens in aforementioned investigations, numerous studies have showed a significant influence of mechanical cycling on the load-bearing capacity of dental restorations [9,13,23]. Most of these investigations applied additional thermal cycling which, as suggested by Rosentritt et al., is accountable for a large decrease in strength [13]. Besides the effects of thermal and mechanical cycling, the present study also investigated the influence of preliminary damage to the load-bearing capacity of zirconia FPDs. The nature of the damage was deliberately chosen to resemble a flaw unintentionally generated by a technician during fabrication of a bridge. No significant difference in load-bearing capacity from preliminary damage was found for unaged or for aged specimens. This observation was also reported by Schneemann et al. for restorations made from zirconia [9]. On the other hand, all other studies dealing with preliminary damage of zirconia specimens have described a significant decrease in strength as a result of machining procedure [25–28]. In principle, grinding of ceramics can act in two different directions. First, it may cause residual surface compressive stress which can considerably increase strength of Y-TZP ceramics [29]. Second, it may induce surface flaws which act as stress concentrators and may become strength-determining if they exceed the depth of the grinding-induced surface compressive layer [30]. The minor influence of preliminary damage in the present survey may be explained by the flaw being only 180 lm in width and 60 lm in depth. Moreover, scratches were induced during water cooling and showed a rounded inner edge design caused by the geometry of the diamond saw. However, this design seems to be advantageous with regard to tensile stresses during functional loading. Though the likely site of highest tensile stresses during occlusal loading is located in the middle of the framework’s central gingival embrasure (Figs. 5 and 6), preliminary damage inflicted in the vicinity of this area did not create a stress concentration sufficient to promote a crack origin. Furthermore, the cuts were covered by veneering material and therefore possibly shielded against the degrading effect of water during TMC and storage in water. However, it is
1446
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
difficult to assess how far this protecting effect could be achieved by the feldspathic ceramic used. Preliminary damage did not induce a significant decrease in load-bearing capacity. In contrast, Weibull analysis shows a decrease in the Weibull modulus m due to preliminary damage for both aged specimens (from 8.0 to 6.2) and unaged specimens (from 7.2 to 5.4). This means that preliminary damage increases scatter in the force at failure and the spectrum of fracture-initiating structural defects. Therefore the probability of failure rises at lower forces, correlating with the decreased reliability of restorations during clinical use. A comparable decrease in the Weibull modulus m has also been reported for zirconia from surface processing with coarse milling tools [26,31,32]. In contrast to preliminary damage, an increased number of mechanical cycles did not effect a significant change in the Weibull modulus m, whereas the load-bearing capacity of the restorations significantly decreased in comparison to the control. Over all groups, the values of the Weibull modulus were similar to those found in other studies dealing with restorations made of zirconia [33]. Kohorst et al. investigated the fracture origin of failed four-unit zirconia FPDs by means of SEM analysis [1]. The results show a close relationship with the location of highest tensile stresses evaluated by FEA in the present study (Figs. 5 and 6). Ceramics are particularly susceptible to tensile stresses which promote subcritical crack growth in brittle materials. Promoted by flaws and defects within the lattice, crack growth and eventually catastrophic failure of ceramic components may occur under critical fracture stresses [34]. Up to now there have been no studies using 3D FEA to determine the stress distribution in all-ceramic four-unit bridges, but research on three-unit restorations has shown that the highest tensile stresses occur within the connectors, where the fracture origin would be expected [35]. FEA suggests the possibility of detecting the weakest points of dental restorations by virtual simulation of different loading conditions. Based on collected data, the design of dental restorations could be optimized to achieve a smoother distribution of stresses and, at the same time, a higher load-bearing capacity [36,37].
References [1] Kohorst P, Herzog TJ, Borchers L, Stiesch-Scholz M. Load-bearing capacity of all-ceramic posterior four-unit fixed partial dentures with different zirconia frameworks. Eur J Oral Sci 2007;115:161–6. [2] Hannink RHJ, Kelly PM, Muddle BC. Transformation toughening in zirconia-containing ceramics. J Am Ceram Soc 2000;83:461–87. [3] Luthardt RG, Sandkuhl O, Reitz B. Zirconia-TZP and alumina – advanced technologies for the manufacturing of single crowns. Eur J Prosthodont Restor Dent 1999;7:113–9. [4] Studart AR, Filser F, Kocher P, Gauckler LJ. Fatigue of zirconia under cyclic loading in water and its implications for the design of dental bridges. Dent Mater 2007;23:106–14. [5] Chevalier J, Olagnon C, Fantozzi G. Subcritical crack propagation in 3Y-TZP ceramics: static and cyclic fatigue. J Am Ceram Soc 1999;82:3129–38.
[6] Guo X. Property degradation of tetragonal zirconia induced by lowtemperature defect reaction with water molecules. Chem Mater 2004;16:3988–94. [7] Thompson I, Rawlings RD. Mechanical behaviour of zirconia and zirconia-toughened alumina in a simulated body environment. Biomaterials 1990;11:505–8. [8] Curtis AR, Wright AJ, Fleming GJ. The influence of simulated masticatory loading regimes on the bi-axial flexure strength and reliability of a Y-TZP dental ceramic. J Dent 2005;34:317–25. [9] Schneemann P, Borchers L, Stiesch-Scholz M. Belastbarkeit 4gliedriger Seitenzahnbru¨cken aus Vollkeramik [Fracture load of 4unit all-ceramic fixed partial dentures]. ZWR 2005;114:28–36. [10] Studart AR, Filser F, Kocher P, Gauckler LJ. In vitro lifetime of dental ceramics under cyclic loading in water. Biomaterials 2007;28:2695–705. [11] Zhang Y, Lawn B. Long-term strength of ceramics for biomedical applications. J Biomed Mater Res 2004;69B:166–72. [12] Kelly JR, Tesk JA, Sorensen JA. Failure of all-ceramic fixed partial dentures in vitro and in vivo: analysis and modeling. J Dent Res 1995;74:1253–8. [13] Rosentritt M, Behr M, Gebhard R, Handel G. Influence of stress simulation parameters on the fracture strength of all-ceramic fixedpartial dentures. Dent Mater 2006;22:176–82. [14] Stiesch-Scholz M, Schulz K, Borchers L. In vitro fracture resistance of four-unit fiber-reinforced composite fixed partial dentures. Dent Mater 2006;22:374–81. [15] Parfitt GJ. Measurement of the physiological mobility of individual teeth in an axial direction. J Dent Res 1960;39:608–18. [16] Heners M. Elektronische Untersuchung zur Reproduzierbarkeit des okklusalen Traumas [Electronic studies on the reproducibility of occlusal trauma]. Dtsch Zahna¨rztl Z 1977;32:433–9. [17] Wiederhorn SM. Moisture assisted crack growth in ceramics. Int J Fracture Mech 1968;4:171–7. [18] Drummond JL. In vitro aging of yttria stabilized zirconia. J Am Ceram Soc 1989;72:675–6. [19] Koh Y-H, Kong Y-M, Kim S, Kim H-E. Improved low-temperature environmental degradation of yttria-stabilized tetragonal zirconia polycrystals by surface encapsulation. J Am Ceram Soc 1999;82:1456–8. [20] Addison O, Fleming GJ, Marquis PM. The effect of thermocycling on the strength of porcelain laminate veneer (PLV) materials. Dent Mater 2003;19:291–7. [21] Magne P, Kwon KR, Besler C, Hodges JS, Douglas WH. Crack propensity of porcelain laminate veneers – a simulated operatory evaluation. J Prosthet Dent 1999;81:327–34. [22] Schindler HJ, Stengel E, Spiess WE. Feedback control during mastication of solid food textures – a clinical–experimental study. J Prosthet Dent 1998;80:330–6. [23] Rountree P, Nothdurft F, Pospiech P. In vitro-investigations on the fracture strength of all-ceramic posterior bridges of ZrO2-ceramic. J Dent Res 2001;80(Spec. Iss. A):57 (Abstract 173). [24] Sundh A, Molin M, Sjogren G. Fracture resistance of yttrium oxide partially stabilized zirconia all-ceramic bridges after veneering and mechanical fatigue testing. Dent Mater 2005;21:476–82. [25] Lee SK, Tandon R, Readey MJ, Lawn B. Scratch damage in zirconia ceramics. J Am Ceram Soc 2000;83:1428–32. [26] Luthardt RG, Holzhu¨ter M, Sandkuhl O, Herold V, Schnapp JD, Kuhlisch E, Walter M. Reliability and properties of ground Y-TZPzirconia ceramics. J Dent Res 2002;81:487–91. [27] Fischer H, Weinzierl P, Weber M, Marx R. Bearbeitungsinduzierte Scha¨digung von Dentalkeramik [Processing related damages of dental ceramic]. Dtsch Zahna¨rztl Z 1999;54:484–8. [28] Zhang Y, Pajares A, Lawn BR. Fatigue and damage tolerance of YTZP ceramics in layered biomechanical systems. J Biomed Mater Res 2004;71B:166–71. [29] Kosmac T, Oblak C, Jevnikar P, Funduk N, Marion L. The effect of surface grinding and sandblasting on flexural strength and reliability of Y-TZP zirconia ceramic. Dent Mater 1999;15:426–33.
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447 [30] Sindel J, Petschelt A, Grellner F, Dierken C, Greil P. Evaluation of subsurface damage in CAD/CAM machined dental ceramics. J Mater Sci-Mater Med 1998;9:291–5. [31] Curtis AR, Wright AJ, Fleming GJ. The influence of surface modification techniques on the performance of a Y-TZP dental ceramic. J Dent 2006;34:195–206. [32] Kosmac T, Oblak C, Jevnikar P, Funduk N, Marion L. Strength and reliability of surface treated Y-TZP dental ceramics. J Biomed Mater Res 2000;53:304–13. [33] Lu¨thy H, Filser F, Loeffel O, Schumacher M, Gauckler LJ, Hammerle CH. Strength and reliability of four-unit all-ceramic posterior bridges. Dent Mater 2005;21:930–7. [34] Marx R, Fischer H, Weber M, Jungwirth F. Rissparameter und Weibullmodule: unterkritisches Risswachstum und Langzeitfestigkeit
1447
vollkeramischer Materialien [Crack parameters and Weibull moduli: subcritical crack growth and long time durability of all-ceramic materials]. Dtsch Zahna¨rztl Z 2001;56:90–8. [35] Oh W, Gotzen N, Anusavice KJ. Influence of connector design on fracture probability of ceramic fixed-partial dentures. J Dent Res 2002;81:623–7. [36] Augereau D, Pierrisnard L, Barquins M. Relevance of the finite element method to optimize fixed partial denture design. Part I. Influence of the size of the connector on the magnitude of strain. Clin Oral Investig 1998;2:36–9. [37] Couegnat G, Fok SL, Cooper JE, Qualtrough AJE. Structural optimization of dental restorations using the principle of adaptive growth. Dent Mater 2006;22:3–12.
Acta Biomaterialia 4 (2008) 1440–1447 www.elsevier.com/locate/actabiomat
Influence of cyclic fatigue in water on the load-bearing capacity of dental bridges made of zirconia Philipp Kohorst *, Marc Philipp Dittmer, Lothar Borchers, Meike Stiesch-Scholz Department of Prosthetic Dentistry and Biomedical Materials Science, Hannover Medical School, Carl-Neuberg-Straße 1, 30625 Hannover, Germany Received 26 October 2007; received in revised form 2 April 2008; accepted 14 April 2008 Available online 1 May 2008
Abstract The humid atmosphere and permanent occurrence of chewing forces in the oral environment lead to degradation of ceramics used for prosthetic restorations. The aim of this in vitro study was to evaluate the influence of artificial aging on the load-bearing capacity of fourunit bridges, with both undamaged and predamaged zirconia frameworks. Additionally, different parameters for chewing simulation have been investigated and a finite element analysis was made to predict the location of highest tensile stresses within the bridges. A total of 60 frameworks were milled from presintered zirconia and divided into six homogeneous groups. Prior to veneering, frameworks of two groups were ‘‘damaged” by a defined saw cut similar to an accidental flaw generated during shape cutting. After veneering, FPDs were subjected to thermal and mechanical cycling – with the exception of control groups. The load-bearing capacity of tested FPDs was significantly reduced by artificial aging. In comparison to unaged specimens, fracture resistance decreased by about 40%, whereas preliminary damage did not have a significant effect. Increasing number of cycles and increasing upper load limit failed to show any additional effect on fracture force. To predict the progression of degradation under the terms of in vitro simulation for even longer periods, further aging experiments are required. Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Dental bridges; Zirconia; Load-bearing capacity; Cyclic loading; FEA
1. Introduction All-ceramic dental restorations have become more and more important, in particular, due to their favorable esthetics and their outstanding biological compatibility. The recent incorporation of high-strength zirconia into dentistry has increased the spectrum of indications for ceramic materials, allowing fabrication of implants, allceramic crowns and fixed partial dentures (FPDs) for the highly loaded posterior region [1]. This type of zirconia owes its high-strength to so-called transformation reinforcement, a complex mechanism involving transformation from tetragonal-to-monoclinic structure and associated with a local 4% increase in volume [2,3]. For this process to be effective, the tetragonal structure must be stabilized *
Corresponding author. Tel.: +49 511 532 4773; fax: +49 511 532 4790. E-mail address: [email protected] (P. Kohorst).
at room temperature, for which reason almost all zirconium dioxide ceramics used for dental restorations contain yttrium oxide (Y2O3). Under functional loading in the oral environment, zirconia-based materials, despite their remarkably high-strength, undergo fatigue and subcritical crack growth that can significantly reduce their strength over time [4]. Subcritical crack growth occurs due to the stress-assisted reaction of water molecules with the ionic–covalent bonds at the crack tip [5]. Furthermore, water molecules can be incorporated into the zirconia lattice by filling oxygen vacancies [6]. This reduces the energy barrier for the tetragonal-to-monoclinic transformation and thus increases the rate of transformation which, due to the associated volume increase, results in microcrack formation within the lattice. Although this mechanism is usually very slow at oral temperatures, it may lead to a significant decrease in the strength of dental restorations over periods of wear of several decades [7].
1742-7061/$ - see front matter Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2008.04.012
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
Mechanical loading and a moist environment are the conditions encountered in the mouth during mastication. Thus, determining the influence of degradation effects to load-bearing capacity under cyclic stresses in water is prerequisite to predict the sustainable success of zirconia-based restorations in dentistry. There are only a few investigations that have dealt with the strength-degrading effect of cyclic loading in water on zirconia ceramics [8–10]. Besides cyclic loading in a moist environment, flaws within the ceramic structure could lead to subcritical crack growth and, eventually, catastrophic failure [11]. These defects may be conditional on the manufacturing of the zirconia blanks used or could be created by the technician on the material’s surface during production or finishing. In particular, the gingival embrasure of FPDs is the most sensitive site for any kind of flaw. By means of finite element analysis (FEA) and fractography, Kelly et al. determined that the highest tensile stress within the connectors of an all-ceramic three-unit FPD is at this site [12]. They evaluated the failure mode of FPDs made of glass-infiltrated alumina, taking into account the resilient embedding of abutment teeth. The aim of the present study was to test the hypothesis that the in vitro load-bearing capacity of a posterior dental bridge made of zirconia is reduced by aging in an artificial oral environment. Additionally, the hypotheses were tested that an increasing number of mechanical cycles or an increasing upper load limit and a defined mechanical damage to the framework at a sensitive location causes a decrease in load-bearing capacity. Furthermore, the site of the highest tensile stresses within the zirconia FPD was determined by means of FEA. 2. Materials and methods 2.1. Preparation
1441
many). The frames were made by repeated copying of a master frame in a computer-aided machinery (CAM) unit (Cercon brain, DeguDent, Hanau, Germany) with subsequent firing in the system oven (Cercon heat, DeguDent, Hanau, Germany). The dimensions of all frames were practically the same, connector width and height differing by less than 0.2 mm. Connector cross-sections were elliptically shaped, their areas being (from mesial to distal) 12.5, 15.6 and 11.6 mm2, respectively. Before veneering, 20 frames were selected at random and a U-shaped cut of width 180 lm and depth 60 lm was made at the gingival surface of the connector between teeth 25 and 26 (the presumed location of highest tensile stress during loading) (Fig. 1). This was done to simulate possible damage to the core during the manufacturing process and to test the influence on load-bearing capacity. Such cutting (subsequently referred to as ‘‘predamaging”) was performed with an annular diamond saw (Microslice 2, Metals Research Ltd., Royston, UK). After veneering, the undamaged specimens were also randomized and divided into homogeneous groups of 10 each, resulting in a total of six groups. 2.3. Aging The bridges were fixed onto the PUR abutments by means of glass-ionomer cement (Ketac-Cem, 3M ESPE, Seefeld, Germany). With the exception of two groups (undamaged/predamaged), the bridges were subjected to thermal and mechanical cycling (TMC) during 200 days storage in distilled water at 36 °C. During this period, 1 104 thermal cycles between 5 and 55 °C (30 s dwell time at each temperature) and a mechanical loading (load frequency 2.5 Hz) with varying number of cycles (1 106/ 2 106) and varying upper load limits (100 N/200 N) were applied successively (Table 1 and Fig. 2).
In an upper jaw plastic model (Frasaco OK 119, A-3 T, Franz Sachs & Co., Tettnang, Germany), teeth 24 and 27 were prepared to accommodate a four-unit all-ceramic FPD. Stone casts (Fuji Rock, GC, Leuven, Belgium) were produced by means of individual impressions of the prepared teeth, and were used as a basis for manufacturing zirconium dioxide bridge frames. During aging and static testing, the FPDs were cemented to and supported by duplicates of the prepared original abutment teeth. These duplicates were made of reinforced polyurethane (PUR) resin (Alpha-Die-Top, Schu¨tz Dental, Rosbach, Germany). Natural periodontal resilience was simulated by coating the roots of these stumps with an elastic latex material (Erkoskin, Erkodent, Pfalzgrafenweiler, Germany). The latexcoated roots were imbedded in a PUR base that reached as far as 3 mm below the preparation margin. 2.2. Manufacture of FPDs A total of 60 frameworks were fabricated in partially sintered zirconia (Cercon base, DeguDent, Hanau, Ger-
Fig. 1. SEM image of preliminary damage generated on the gingival embrasure of a zirconia framework.
1442
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
Table 1 Variations in fatigue parameters and preliminary damage Group Cer Cer Cer Cer Cer Cer
I II III IV V VI
Predamage
Thermal cycles
Mechanical cycles
Upper load limit
1 104 1 104 1 104 1 104
1 106 1 106 2 106 1 106
100 N 100 N 100 N 200 N
+ +
not influenced by artificial aging and predamaging. Weibull parameters F0 and m were determined for each test group by fitting a Weibull distribution to the respective data set. The parameter F0 (characteristic force at failure) is associated with 63.2% probability of failure, whereas the Weibull modulus, m, is a measure of the scatter in the force at failure and of the reliability of the material investigated. The greater the value for m, the steeper the transition from survival to failure for the probability distribution against force at failure. 2.6. Finite element analysis One of the investigated FPDs was scanned optically before and after veneering by the principle of triangulation (Atos II SO, GOM, Braunschweig, Germany). Based on the resulting polygon meshes, two three-dimensional bulk models were created by reverse engineering. Afterwards models were imported into a CAD program (DesignModeler, Ansys Inc., Canonsburg, PA). In a Boolean operation, the zirconia framework was virtually subtracted from the veneered bridge, whereby a separate model of the veneering material was generated. According to the in vitro study model, abutment teeth, including the cement layer and resilient support in the polyurethane block, were designed virtually by means of the CAD software. Then the model was transferred to an FEA program (DesignSpace, Ansys Inc.) for finite element stress analysis. 3. Results
Fig. 2. Schematic of cyclic loading set-up.
2.4. Determination of load-bearing capacity Following the aging procedure, the specimens were loaded until fracture. This was carried out in a universal testing machine (Type 20K, UTS Testsysteme, Ulm-Einsingen, Germany) at a crosshead speed of 1 mm min 1 with the force transferred to the occlusal connector area between teeth 25 and 26 via a tungsten carbide ball (diameter 6.0 mm) on an interposed tin foil (thickness 0.2 mm). A sudden decrease in force of more than 15 N was regarded as indication of failure and the maximum force up to this point recorded as force at fracture. 2.5. Statistics The statistical analysis of force at fracture data was performed using one-way analysis of variance (ANOVA) for detecting the influence of artificial aging and two-factor ANOVA for detecting the influence of predamage, with the level of significance chosen at 0.05 (SPSS 15.0, SPSS Software Corp., Chicago, IL). A direct comparison of group means was carried out with post hoc Scheffe´ test. The null hypotheses were that load-bearing capacity is
All FPDs tested survived thermal and mechanical cycling in the artificial oral environment. Table 2 shows the results of load-bearing capacity testing. After the failure criterion (15 N load drop) had been fulfilled, visual inspection revealed that 57 of 60 FPDs (95%) had failed with sudden bulk fracture of the zirconia framework. Two bridges showed Hertzian cracking beneath the indenter and one bridge extensive delamination of the veneering ceramic before bulk fracture occurred at higher force levels. The artificial aging by thermal and mechanical cycling proved to have a significant influence on the load-bearing capacity of both the undamaged and the predamaged zirconia FPDs (p < 0.001). In comparison to the control groups, Cer I (1525 N) and Cer II (1335 N), the aged specimens exhibited a broad decrease in average force at fracture (Fig. 3). However, variation of mechanical cycling parameters (number of cycles/upper load limit) failed to show a significant influence on load-bearing capacity (Table 3). Means were 904 N (Cer III), 924 N (Cer V) and 952 N (Cer VI) (Table 2 and Fig. 4). Furthermore, the effect of mechanical predamage did not prove to be statistically significant (p = 0.213) for either the unaged (Cer II/1335 N) or the aged (Cer IV/921 N) specimens (Table 2 and Fig. 3). Therefore, the first hypothesis formulated in the introduction could be accepted, whereas the others had to be rejected.
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
1443
Table 2 Mean force at fracture (SD) and Weibull parameters F0 and m, together with their 95% confidence intervals Group
Load at fracture (N)
F0 (N)
95% confidence interval Lower limit
Upper limit
Cer Cer Cer Cer Cer Cer
1525.0 1334.7 903.7 921.1 923.5 952.4
1625.8 1445.9 958.6 990.6 976.3 1015.8
1580.3 1391.4 927.8 955.5 948.5 957.4
1671.4 1500.3 989.3 1025.8 1004.1 1074.1
I II III IV V VI
(76.5) (89.4) (40.8) (55.6) (40.3) (51.4)
Fig. 3. Box chart representing forces at fracture vs. preliminary damage, with dependency on cyclic loading (1 106 cycles/100 N upper load limit). Medians, quartiles and extremes are given.
Table 3 Results of statistical post hoc comparison of testing groups subjected to different fatigue regimes (p < 0.05) Cer I Cer Cer Cer Cer
I III V VI
<0.001 <0.001 <0.001
Cer III
Cer V
Cer VI
<0.001
<0.001 0.995
<0.001 0.942 0.987
0.995 0.942
0.987
As follows from the peculiarities of Weibull statistics, Weibull characteristic forces were around 6–8% higher than the corresponding mean forces at fracture and exhibited the same tendencies (Table 2). A higher number of mechanical cycles was associated with the statistically non-significant increase in the Weibull modulus from 7.2 (Cer I) to 8.0 (Cer III) to 8.6 (Cer V). However, raising the upper load limit did not have the same effect, with a Weibull modulus of 7.2 (Cer VI), as in the control group (Cer I) (Table 2). Mechanical predamage resulted in a decrease in the Weibull modulus m. This could be observed for both the unaged (decrease from 7.2 to 5.4) and aged (decrease from 8.0 to 6.2) specimens.
m
95% confidence interval Lower limit
Upper limit
7.2 5.4 8.0 6.2 8.6 7.2
6.1 4.6 6.4 5.1 6.9 4.7
8.4 6.3 9.7 7.2 10.3 9.8
Fig. 4. Box chart representing forces at fracture vs. different cyclic loading parameters. Medians, quartiles and extremes are given.
FE analysis revealed the location of highest tensile stresses to be in the zirconia framework close to its surface at the gingival embrasure of the connector between premolar and molar (Figs. 5 and 6). 4. Discussion In most studies investigating load-bearing capacity of all-ceramic restorations, alloys are used to create abutment teeth for in vitro testing. In the present study, model materials and testing conditions were chosen carefully to imitate clinical reality as faithfully as possible. The abutment teeth and their bases were made of reinforced PUR, which has an elastic modulus somewhat lower than dentin and bone but represents a better approximation of natural conditions than alloys, for example. Models made of alloys exhibit very high rigidity, so dental bridges are better supported during static and cyclic loading compared to conditions in the oral cavity. This results in an artificially high loadbearing capacity [13]. The latex layer around the abutment roots simulated a periodontal resilience of 30–95 lm against forces of between 50 and 100 N in the axial direction [14]. Extrapolating measurements of natural periodontal resilience men-
1444
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
Fig. 5. Stress distribution on framework surface during occlusal loading according to FEA (veneering layer removed for clarity). Highest tensile stress occurred within the framework at the gingival area of the connector.
Fig. 6. Cross-sectional area of the FE model with mesial view of the median connector. Highest tensile stress within the zirconia framework at transition to veneering ceramic.
tioned in the literature [15,16] as the chosen conditions is a sufficient approximation. A simple rigid support would not consider additional compressive, tensile and shear stresses caused by occlusal loading and lateral movements. Aqueous environments like the oral cavity generally facilitate subcritical crack growth in ceramics [17] and may cause uncontrolled transitions from the tetragonal to the monoclinic structure of Y-TZP [7]. Both produce a
decrease in strength of the ceramic. Other investigations have shown that the spontaneous transformation of zirconia in water initially proceeds rapidly but remains static after some point. For example, Drummond reported only a minor further decrease in strength of zirconia after approximately 300 days of water storage [18]. Consequently the bulk of degradation could be expected to have appeared during the 200 days of water storage applied in
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
the present study. However, it has to be taken into account that the zirconia specimens evaluated in the present study were veneered with a feldspathic ceramic and hence were not in direct contact with the humid environment. It remains to be clarified if the veneering layer permitted sufficient water diffusion to promote zirconia degradation or if it rather acted as a protective layer, as was reported for a combination of silica and zirconia [19]. In the course of water storage, specimens were also subjected to thermal cycling. Material strength is degraded by the repeated thermal stressing of prosthetic restorations, which results from temperature changes caused by the consumption of hot and cold food and drink, and breathing. This thermal stressing leads to tensions within the specimens and become manifest in slow subcritical crack growth and catastrophic failure eventually. Several investigators found a decrease in strength [20] and the development of flaws [21] during thermal cycling for specimens made of feldspathic ceramics. Finally, FPDs were subjected to mechanical cycling in an aqueous environment. The repeated application of chewing forces contributes to the decrease in ceramic strength during service [13]. This was simulated by mechanical cycling with upper limits of 100 and 200 N. Since a recent survey has revealed the projected number of chewing cycles to be about 800,000 per year [13], the 1 106 and 2 106 cycles applied in this study corresponded to an in vivo service period of approximately 15 and 30 months, respectively. The cyclic loads applied in the present study represent high levels of physiological chewing forces, which have been found to range between 20 and 120 N, depending on food consistency [22]. Rosentritt et al. found a decrease in load-bearing capacity of all-ceramic restorations due to increased simulated chewing forces [13]. Based on these results, it appears reasonable to raise the upper load limit during mechanical cycling beyond physiological forces, thus improving the estimation of the restorations’ performance under more severe conditions like bruxism. In all groups tested, artificial aging resulted in a significant decrease in load-bearing capacity of approximately 40%, whereas variation in aging parameters did not show significant influences. Other authors investigating dental restorations made of zirconia also recorded reduced strength by about 20% due to aging [9,23]. Rosentritt et al. reported a reduction in strength in an artificial oral environment for FPDs made of a pressable glass–ceramic. But, contrary to the present study, they described a significant decrease in load-bearing capacity due to an increase in the upper load limit from 50 to 150 N [13]. These findings are most likely associated with the special properties of the glass–ceramic tested, which is of lower strength and also greater sensitivity to subcritical crack growth than zirconia [10]. Contrary to the above investigations, Curtis et al. did not observe a decrease in load-bearing capacity of standardized zirconia specimens after mechanical cycling, either in dry or in humid atmospheres [8]. They even
1445
increased the upper load limit to a range from 500 to 800 N, whereas only 2000 cycles were performed. Sundh et al. also did not detect any reduction in the strength of zirconia bridges subjected to dynamic loading in water for 1 105 cycles with an upper load limit of 50 N [24]. An analogous observation was made by Studart et al. for zircona samples [4]. Specimens’ strength remained unaltered for a long period during mechanical cycling, but then abruptly dropped and settled at a lower level. This material behavior shows certain parallels to the results of the present study where neither doubling the load nor doubling the number of cycles induced a change in load-bearing capacity. For zirconia specimens stored in a humid atmosphere, Drummond observed an analogous phenomenon [18]. Though mechanical cycling did not greatly reduce the strength of zirconia specimens in aforementioned investigations, numerous studies have showed a significant influence of mechanical cycling on the load-bearing capacity of dental restorations [9,13,23]. Most of these investigations applied additional thermal cycling which, as suggested by Rosentritt et al., is accountable for a large decrease in strength [13]. Besides the effects of thermal and mechanical cycling, the present study also investigated the influence of preliminary damage to the load-bearing capacity of zirconia FPDs. The nature of the damage was deliberately chosen to resemble a flaw unintentionally generated by a technician during fabrication of a bridge. No significant difference in load-bearing capacity from preliminary damage was found for unaged or for aged specimens. This observation was also reported by Schneemann et al. for restorations made from zirconia [9]. On the other hand, all other studies dealing with preliminary damage of zirconia specimens have described a significant decrease in strength as a result of machining procedure [25–28]. In principle, grinding of ceramics can act in two different directions. First, it may cause residual surface compressive stress which can considerably increase strength of Y-TZP ceramics [29]. Second, it may induce surface flaws which act as stress concentrators and may become strength-determining if they exceed the depth of the grinding-induced surface compressive layer [30]. The minor influence of preliminary damage in the present survey may be explained by the flaw being only 180 lm in width and 60 lm in depth. Moreover, scratches were induced during water cooling and showed a rounded inner edge design caused by the geometry of the diamond saw. However, this design seems to be advantageous with regard to tensile stresses during functional loading. Though the likely site of highest tensile stresses during occlusal loading is located in the middle of the framework’s central gingival embrasure (Figs. 5 and 6), preliminary damage inflicted in the vicinity of this area did not create a stress concentration sufficient to promote a crack origin. Furthermore, the cuts were covered by veneering material and therefore possibly shielded against the degrading effect of water during TMC and storage in water. However, it is
1446
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447
difficult to assess how far this protecting effect could be achieved by the feldspathic ceramic used. Preliminary damage did not induce a significant decrease in load-bearing capacity. In contrast, Weibull analysis shows a decrease in the Weibull modulus m due to preliminary damage for both aged specimens (from 8.0 to 6.2) and unaged specimens (from 7.2 to 5.4). This means that preliminary damage increases scatter in the force at failure and the spectrum of fracture-initiating structural defects. Therefore the probability of failure rises at lower forces, correlating with the decreased reliability of restorations during clinical use. A comparable decrease in the Weibull modulus m has also been reported for zirconia from surface processing with coarse milling tools [26,31,32]. In contrast to preliminary damage, an increased number of mechanical cycles did not effect a significant change in the Weibull modulus m, whereas the load-bearing capacity of the restorations significantly decreased in comparison to the control. Over all groups, the values of the Weibull modulus were similar to those found in other studies dealing with restorations made of zirconia [33]. Kohorst et al. investigated the fracture origin of failed four-unit zirconia FPDs by means of SEM analysis [1]. The results show a close relationship with the location of highest tensile stresses evaluated by FEA in the present study (Figs. 5 and 6). Ceramics are particularly susceptible to tensile stresses which promote subcritical crack growth in brittle materials. Promoted by flaws and defects within the lattice, crack growth and eventually catastrophic failure of ceramic components may occur under critical fracture stresses [34]. Up to now there have been no studies using 3D FEA to determine the stress distribution in all-ceramic four-unit bridges, but research on three-unit restorations has shown that the highest tensile stresses occur within the connectors, where the fracture origin would be expected [35]. FEA suggests the possibility of detecting the weakest points of dental restorations by virtual simulation of different loading conditions. Based on collected data, the design of dental restorations could be optimized to achieve a smoother distribution of stresses and, at the same time, a higher load-bearing capacity [36,37].
References [1] Kohorst P, Herzog TJ, Borchers L, Stiesch-Scholz M. Load-bearing capacity of all-ceramic posterior four-unit fixed partial dentures with different zirconia frameworks. Eur J Oral Sci 2007;115:161–6. [2] Hannink RHJ, Kelly PM, Muddle BC. Transformation toughening in zirconia-containing ceramics. J Am Ceram Soc 2000;83:461–87. [3] Luthardt RG, Sandkuhl O, Reitz B. Zirconia-TZP and alumina – advanced technologies for the manufacturing of single crowns. Eur J Prosthodont Restor Dent 1999;7:113–9. [4] Studart AR, Filser F, Kocher P, Gauckler LJ. Fatigue of zirconia under cyclic loading in water and its implications for the design of dental bridges. Dent Mater 2007;23:106–14. [5] Chevalier J, Olagnon C, Fantozzi G. Subcritical crack propagation in 3Y-TZP ceramics: static and cyclic fatigue. J Am Ceram Soc 1999;82:3129–38.
[6] Guo X. Property degradation of tetragonal zirconia induced by lowtemperature defect reaction with water molecules. Chem Mater 2004;16:3988–94. [7] Thompson I, Rawlings RD. Mechanical behaviour of zirconia and zirconia-toughened alumina in a simulated body environment. Biomaterials 1990;11:505–8. [8] Curtis AR, Wright AJ, Fleming GJ. The influence of simulated masticatory loading regimes on the bi-axial flexure strength and reliability of a Y-TZP dental ceramic. J Dent 2005;34:317–25. [9] Schneemann P, Borchers L, Stiesch-Scholz M. Belastbarkeit 4gliedriger Seitenzahnbru¨cken aus Vollkeramik [Fracture load of 4unit all-ceramic fixed partial dentures]. ZWR 2005;114:28–36. [10] Studart AR, Filser F, Kocher P, Gauckler LJ. In vitro lifetime of dental ceramics under cyclic loading in water. Biomaterials 2007;28:2695–705. [11] Zhang Y, Lawn B. Long-term strength of ceramics for biomedical applications. J Biomed Mater Res 2004;69B:166–72. [12] Kelly JR, Tesk JA, Sorensen JA. Failure of all-ceramic fixed partial dentures in vitro and in vivo: analysis and modeling. J Dent Res 1995;74:1253–8. [13] Rosentritt M, Behr M, Gebhard R, Handel G. Influence of stress simulation parameters on the fracture strength of all-ceramic fixedpartial dentures. Dent Mater 2006;22:176–82. [14] Stiesch-Scholz M, Schulz K, Borchers L. In vitro fracture resistance of four-unit fiber-reinforced composite fixed partial dentures. Dent Mater 2006;22:374–81. [15] Parfitt GJ. Measurement of the physiological mobility of individual teeth in an axial direction. J Dent Res 1960;39:608–18. [16] Heners M. Elektronische Untersuchung zur Reproduzierbarkeit des okklusalen Traumas [Electronic studies on the reproducibility of occlusal trauma]. Dtsch Zahna¨rztl Z 1977;32:433–9. [17] Wiederhorn SM. Moisture assisted crack growth in ceramics. Int J Fracture Mech 1968;4:171–7. [18] Drummond JL. In vitro aging of yttria stabilized zirconia. J Am Ceram Soc 1989;72:675–6. [19] Koh Y-H, Kong Y-M, Kim S, Kim H-E. Improved low-temperature environmental degradation of yttria-stabilized tetragonal zirconia polycrystals by surface encapsulation. J Am Ceram Soc 1999;82:1456–8. [20] Addison O, Fleming GJ, Marquis PM. The effect of thermocycling on the strength of porcelain laminate veneer (PLV) materials. Dent Mater 2003;19:291–7. [21] Magne P, Kwon KR, Besler C, Hodges JS, Douglas WH. Crack propensity of porcelain laminate veneers – a simulated operatory evaluation. J Prosthet Dent 1999;81:327–34. [22] Schindler HJ, Stengel E, Spiess WE. Feedback control during mastication of solid food textures – a clinical–experimental study. J Prosthet Dent 1998;80:330–6. [23] Rountree P, Nothdurft F, Pospiech P. In vitro-investigations on the fracture strength of all-ceramic posterior bridges of ZrO2-ceramic. J Dent Res 2001;80(Spec. Iss. A):57 (Abstract 173). [24] Sundh A, Molin M, Sjogren G. Fracture resistance of yttrium oxide partially stabilized zirconia all-ceramic bridges after veneering and mechanical fatigue testing. Dent Mater 2005;21:476–82. [25] Lee SK, Tandon R, Readey MJ, Lawn B. Scratch damage in zirconia ceramics. J Am Ceram Soc 2000;83:1428–32. [26] Luthardt RG, Holzhu¨ter M, Sandkuhl O, Herold V, Schnapp JD, Kuhlisch E, Walter M. Reliability and properties of ground Y-TZPzirconia ceramics. J Dent Res 2002;81:487–91. [27] Fischer H, Weinzierl P, Weber M, Marx R. Bearbeitungsinduzierte Scha¨digung von Dentalkeramik [Processing related damages of dental ceramic]. Dtsch Zahna¨rztl Z 1999;54:484–8. [28] Zhang Y, Pajares A, Lawn BR. Fatigue and damage tolerance of YTZP ceramics in layered biomechanical systems. J Biomed Mater Res 2004;71B:166–71. [29] Kosmac T, Oblak C, Jevnikar P, Funduk N, Marion L. The effect of surface grinding and sandblasting on flexural strength and reliability of Y-TZP zirconia ceramic. Dent Mater 1999;15:426–33.
P. Kohorst et al. / Acta Biomaterialia 4 (2008) 1440–1447 [30] Sindel J, Petschelt A, Grellner F, Dierken C, Greil P. Evaluation of subsurface damage in CAD/CAM machined dental ceramics. J Mater Sci-Mater Med 1998;9:291–5. [31] Curtis AR, Wright AJ, Fleming GJ. The influence of surface modification techniques on the performance of a Y-TZP dental ceramic. J Dent 2006;34:195–206. [32] Kosmac T, Oblak C, Jevnikar P, Funduk N, Marion L. Strength and reliability of surface treated Y-TZP dental ceramics. J Biomed Mater Res 2000;53:304–13. [33] Lu¨thy H, Filser F, Loeffel O, Schumacher M, Gauckler LJ, Hammerle CH. Strength and reliability of four-unit all-ceramic posterior bridges. Dent Mater 2005;21:930–7. [34] Marx R, Fischer H, Weber M, Jungwirth F. Rissparameter und Weibullmodule: unterkritisches Risswachstum und Langzeitfestigkeit
1447
vollkeramischer Materialien [Crack parameters and Weibull moduli: subcritical crack growth and long time durability of all-ceramic materials]. Dtsch Zahna¨rztl Z 2001;56:90–8. [35] Oh W, Gotzen N, Anusavice KJ. Influence of connector design on fracture probability of ceramic fixed-partial dentures. J Dent Res 2002;81:623–7. [36] Augereau D, Pierrisnard L, Barquins M. Relevance of the finite element method to optimize fixed partial denture design. Part I. Influence of the size of the connector on the magnitude of strain. Clin Oral Investig 1998;2:36–9. [37] Couegnat G, Fok SL, Cooper JE, Qualtrough AJE. Structural optimization of dental restorations using the principle of adaptive growth. Dent Mater 2006;22:3–12.