Influence of the antagonist material on the wear of different composites using two different wear simulation methods

Dental Materials (2006) 22, 166–175

www.intl.elsevierhealth.com/journals/dema

Influence of the antagonist material on the wear of different composites using two different wear simulation methods S.D. Heintzea,*, G. Zellwegera, A. Cavalleria, J. Ferracaneb a

Research & Development, In vitro-Research, Ivoclar Vivadent, Bendererstrasse 2, FL-9494 Schaan, Liechtenstein b Division of Biomaterials and Biomechanics, Department of Restorative Dentistry, Oregon Health & Science University, Portland, Oregon, USA Received 6 January 2005; accepted 12 April 2005

KEYWORDS Wear; Composite; Ceramic; Stylus

Summary Objectives. The aim of the study was to evaluate two ceramic materials as possible substitutes for enamel using two wear simulation methods, and to compare both methods with regard to the wear results for different materials. Methods. Flat specimens (OHSU nZ6, Ivoclar nZ8) of one compomer and three composite materials (Dyract AP, Tetric Ceram, Z250, experimental composite) were fabricated and subjected to wear using two different wear testing methods and two pressable ceramic materials as stylus (Empress, experimental ceramic). For the OHSU method, enamel styli of the same dimensions as the ceramic stylus were fabricated additionally. Both wear testing methods differ with regard to loading force, lateral movement of stylus, stylus dimension, number of cycles, thermocycling and abrasive medium. In the OHSU method, the wear facets (mean vertical loss) were measured using a contact profilometer, while in the Ivoclar method (maximal vertical loss) a laser scanner was used for this purpose. Additionally, the vertical loss of the ceramic stylus was quantified for the Ivoclar method. The results obtained from each method were compared by ANOVA and Tukey’s test (p!0.05). To compare both wear methods, the log-transformed data were used to establish relative ranks between material/stylus combinations and assessed by applying the Pearson correlation coefficient. Results. The experimental ceramic material generated significantly less wear in Tetric Ceram and Z250 specimens compared to the Empress stylus in the Ivoclar method, whereas with the OHSU method, no difference between the two ceramic antagonists was found with regard to abrasion or attrition. The wear generated by the enamel stylus was not statistically different from that generated by the other two ceramic materials in the OHSU method. With the Ivoclar method, wear of the ceramic stylus was only statistically different when in contact with Tetric Ceram.

* Corresponding author. Tel.: C41 235 3570; fax: C41 235 1279. E-mail address: [email protected] (S.D. Heintze).

0109-5641/$ - see front matter Q 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dental.2005.04.012

Influence of the antagonist material

167

There was a close correlation between the attrition wear of the OHSU and the wear of the Ivoclar method (Pearson coefficient 0.83, pZ0.01). Significance: Pressable ceramic materials can be used as a substitute for enamel in wear testing machines. However, material ranking may be affected by the type of ceramic material chosen. The attrition wear of the OHSU method was comparable with the wear generated with the Ivoclar method. Q 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

Introduction The term ‘wear’ refers to the net loss of a material from its surface under operational conditions. The phenomenon is dependent on many different factors occurring almost simultaneously. In the oral cavity, many factors contribute to the wear of enamel and dentin, such as the nature of the occlusal contacts with antagonist teeth (attrition), chewing of food items, toothbrushing with toothpaste, inhalation of dust (abrasion), acidic attack due to the consumption of certain fruits and beverages, inhalation of industrial acids or vomiting and regurgitation of gastric juice as in the case of bulimia and anorexia nervosa (erosion) [1]. The annual wear rate of molar enamel in non-bruxers is reportedly very low (29 mm) [2], though a high biting force and parafunctional habits such as bruxism can accelerate tooth wear. As with enamel or dentin, restorative materials are subjected to wear, and the wear mode depends on the type of restorative material. Material loss occurs through microploughing, microcutting, microcracking and microfatigue [3]. The various dental materials may be grouped in four different categories: metal alloys, ceramics, composites and unfilled polymers. Of all these materials, the resin composites are somewhat unique in that many variables related to their composition directly influence their wear resistance. Composites consist of filler particles dispersed in a brittle polymer matrix. Optimally, the loading force applied to a composite restoration is completely transferred from the matrix to the stronger, harder filler particles. The size, shape and hardness of the fillers, the quality of the bonding between the fillers and the polymer matrix, and the extent of polymerization of the polymer matrix influence the wear characteristics of the composite. Furthermore, the composition of the material influences physical parameters such as flexural strength, fracture toughness, hardness, modulus of elasticity, and curing depth, all of which may influence the wear [4]. Different approaches have been taken to relate physical properties, such as fracture toughness, to

wear [5–7]. Although properties such as fracture toughness and flexural strength have been identified as potential predictors of wear, both university and industry scientists rely more on in vitro devices that simulate wear than on physical properties alone to predict it. These devices are based on different approaches for both wear simulation and wear analysis, including chewing simulators, rotating disk machines, pinon-disk-machines, etc. [8]. Most often, the approaches are related to one or two wear mechanisms that occur in the mouth. The Ivoclar method that uses the Willytec simulator focusses on two-body wear (attrition) [7,9], while the OHSU wear simulator developed at the Oregon Health & Science University attempts to combine abrasion and attrition by including different forces, a transversal movement of the antagonist, as well as an abrasive medium [10]. There is no agreement in the literature on which material should be used as the antagonist material for wear simulation. While some methods, like those used at OHSU and ZURICH, have advocated enamel, others have suggested the use of steatite, a synthetic material mainly composed of magnesium silicate as a suitable substitute [11,12]. A substitute for enamel as antagonist may reduce the variability of test results due to the uniform shape and limited variation of material properties. Other advantages worth considering are the reduced time required for the fabrication of the antagonist. Furthermore, the limited availability of extracted teeth makes the use of a synthetic material for in vitro wear testing more desirable. It was recently reported that Empress ceramic antagonists produced similar wear as enamel on different composites [13]. The aim of the present study was to compare the wear of four different composite materials with two wear simulating methods using two ceramic materials as stylus as well as natural enamel. The overall goal was to determine the appropriateness of two different ceramic alternatives to enamel for use in future wear studies. Four hypotheses were tested:

168 Table 1

S.D. Heintze et al. Composition and physical properties of the ceramic materials.

Composition

Batch Linear thermal expansion coefficient A (100–400 8C) Chemical solubility (according ISO 6872) Flexural strength (according ISO 6872) Weibull modulus Elastic modulus Vickers hardness Fracture toughness (IS)

Empress TC1 (weight%)

Experimental ceramic—ExpC (weight%)

SiO2 (59–63), Al2O3 (17–21), K2O(10–14), Na2O(3.5–6.5), Pigments (!6): B2O3, BaO, CaO, CeO2, TiO2 E50546 17.0G0.5 [*10K6KK1]

SiO2 (57–80), Li2O (11–19), K2O (0–13.5), MgO (0.1–5), ZnO (0–6), Al2O3 (0–2.5), La2O3 (0.1–6), ZrO2 (0–8), P2O5 (0–11) VP2368 10.5G0.25 [*10K6KK1]

!200 mg/cm2

!100 mg/cm2

120–140 N/mm2

400–500 N/mm2

10.5 65.4 GPa 6.58 GPa 0.83 MPam

7.4 91.6 GPa 5.48 GPa 1.54 MPam

Data obtained at Ivoclar and by Dr. Michael V. Swain, University of Sydney.

1. Wear of the composite materials tested with the Empress material as antagonist will be comparable to the wear generated by human enamel. 2. Due to different material properties, there will be a difference between both ceramic materials with regard to material and antagonist wear. 3. The variability of test results with the ceramic materials will be lower than that achieved with human enamel. 4. There will be an agreement in the wear rankings for the materials tested in the two methods.

Material and methods The two ceramic materials that were used for the antagonists were fabricated by means of the lostTable 2

wax technique and differed with regard to their composition and physical properties (Table 1). One material was the leucite containing IPS Empress ceramic (Ivoclar Vivadent, Schaan; Liechtenstein), and the other a lithium disilicate material based on a new experimental pressable ceramic material featuring high translucency. The Empress antagonists were glazed two times at a temperature of 870 8C, whereas the antagonists made of the experimental ceramic material were double-glazed at 725 8C. The shape of the antagonists is described in the following sections. Three composites and one compomer were evaluated in order to provide a diverse cross-section of available materials to test the adequacy of the different antagonists on the two machines (Table 2). Because the work at the two sites was not conducted at the same time, the batches were not the same.

List of tested materials with batch numbers and composition.

Material

Manufacturer

Ivoclar Batch

OHSU Batch

Composition Matrix UDMA Polymethcrylate TCB BisGMA UDMA TEGDMA BisGMA UDMA Triethyleneglycol-DMA

Dyract AP

Dentsply

0309000790

na

Experimental Composite Tetric Ceram

BISCO

384-11A

384-11A

Ivoclar Vivadent

C16761

B42131

Z250

3M ESPE

6020

na

BisGMA Triethylenglyco-DMA

Filler Strontium fluorosilicate glass 0.8 mm Strontium glass 1–3 mm, SiO2 Barium glass 1 mm Ba-Al-FB-Silicate 1 mm, SiO2 40 nm, Spherical mixed oxide 0.2 mm, Ytterbiumtrifluoride SiO2, zirconium oxide

Influence of the antagonist material

Ivoclar Method All materials were placed onto an aluminum SEM specimen holder (: 8 mm, depth 2 mm) in one layer, light cured at 500 mW/cm2 for 120 s with an Astralis 5 polymerization unit (Ivoclar Vivadent) and then cured for an additional 10 min in a polymerization device (Spectramat, Ivoclar Vivadent, Liechtenstein). For each material, 8 specimens were fabricated. After processing and before testing, the flat specimens were kept dry at a temperature of 37 8C for 24 h. After storage, the specimens were polished with 600 grit SiC, 1200 grit SiC and 2500 grit SiC grit by means of a polishing device (Phoenix 4000, Buehler GmbH, Du ¨sseldorf, Germany). The specimens were mounted in a commercially available chewing simulator (Willytec, Gra ¨ felfing, Germany). The radius of the conical shaped antagonist was 1.18 mm at a height of 600 mm from the cuspal tip to the base. This geometry was chosen to mimic the curvatures of the palatal cusp of upper first molars of young adults (unpublished data). The antagonists were fixed on aluminum SEM holders with the resin cement (Dual Cement, Ivoclar Vivadent) that was light-cured for 40 s with the Astralis 5, and then cured for an additional 10 min in the Spectramat. For referencing baseline and follow-up models, three notches were drilled into the resin base of the ceramic antagonist. Before mounting the specimens in the simulator, impressions were made of the antagonists (see below). The load during the two-body wear simulation was fixed at 50 N, and the sliding movement at 0.7 mm. The frequency of the antagonist movement was 1.6 Hz. A total of 120,000 cycles of unidirectional antagonist movements were carried out. Thermocycling with a frequency of 320/120, 000 cycles was included during the wear testing with a temperature difference between 5 and 55 8C. The thermocycling process was conducted using magnetic valves in conjunction with a heating and cooling system controlled by programmable logic controllers (PLCs). After completing the wear generating procedure, impressions of both the material and the antagonist were made with a low viscosity vinyl polysiloxane material (President light, Colte `ne, Altsta ¨tten, Switzerland). After four hours, replicas were made from the impressions with white, super hard plaster (Fuji Superhard Rock, GC Corporation, Tokyo, Japan) by means of a vacuum, vibrator and pressure device (2 bar). The plaster replicas were analzed by means of a commercially available laser-scanner device (Laserscan 3D, Willytec) and the appropriate

169 match-3D software [14]. In principle, a light line with a width of 22 mm created by a laser diode is projected on the surface and onto a CCD-chip under a triangulation angle of 258, thus encoding the height of every surface point within the lateral displacement of the light line on the CCD-chip. The specimen is shifted with a stepper motor along the y direction. After each step, the CCD image of the light line is stored in a frame grabber. A digital signal processor allows the storage and measurement of 8.33 frames (580 lines) per second which results in a scanning of approx. 5000 surface points per second. The precision of 3-D data acquisition in flat specimens was calculated to be 2.9 mm (G0.5) [14], and in specimens with a wear facet a precision of 5–8 mm is assumed. Vertical resolution is 2–5 mm, and horizontal resolution in the y direction is 30 mm and less than 30 mm in the x direction [15]. For quantification of the material loss the area around the wear facet was used for reference, and for the 3D calculation of the material loss the procedures ‘fit plane’, ‘subtract plane’ and ‘statistics’ were used. For the quantification of the material loss of the ceramic stylus, baseline and follow-up pictures were superimposed by referencing the picture with the three notches and matching the objects with the match-3D procedure until a standard deviation of less than 10 mm was obtained (8000 iterations, minimum 1200 points per matching procedure). The software then calculates the maximum vertical loss (99% quantile). The rationale to use the 99% quantile is to eliminate extreme values due to fine dust particles and other discrepancies.

OHSU Method Six bar-shaped specimens (12.5 mm!5 mm! 2.5 mm) for each of the materials were prepared in a steel mold. The specimens were cured in a Triad II unit (Dentsply, York, USA) for 40 s from the top and bottom surfaces, and aged in water at 37 8C for 7 days. They were then cast in acrylic rings using metallographic epoxy (Buehler, Lake Bluff, USA), wet sanded with 600 grit SiC, and polished with a 1000 grit aluminum oxide slurry. Abrasion and attrition wear were evaluated in an in vitro wear simulator (OHSU Oral Wear Simulator). This machine is an updated version of the original OHSU oral wear simulator previously described by Condon and Ferracane [10]. The specimen was placed in an acrylic chamber filled with 5 mL of fresh slurry prepared by partially grinding 6 g of poppy seeds (McCormick & Co., Hunt Valley, USA) and then mixing with 3 g PMMA beads (average size: 50 mm, range: 5–125 mm; Dentsply Repair Material

170 Powder) and 30 mL of water. Enamel antagonists were sectioned from adult human molars and attached onto knobs on the heads of nylon screws with metallographic epoxy. Each enamel-tipped stylus was milled into a hemispherical shape (: w13 mm) with a diamond bur in a custom shaping device, finished with 600 grit SiC and polished with 1000-grit aluminum oxide slurry. The hemispherical ceramic styli for the OHSU device (: 11 mm) were produced at Ivoclar as described above. The stylus was pushed down into contact with the composite surface through a solenoid, exerting a force of 18– 20 N. The stylus was then slid across the composite surface to produce three-body abrasion and a wear trace of 6–7 mm in length. At the end of the slide, a static load of 80–90 N was applied to produce attrition. The term ‘abrasion’ defines the wear produced on a contact-free area by the sliding action of a tooth against its antagonist in the presence of a third body. ‘Attrition’ is the wear resulting from direct occlusal contact under a heightened load level [4]. Each specimen was subjected to 100,000 cycles at a frequency of 1.9 Hz at room temperature (25 8C). The wear facets were analyzed by making 10 equal spaced passes across the wear trace with an automated contact profilometer having a vertical resolution of approximately 1 mm (Prototech, Portland, USA). The profilometer employed a diamondtipped stylus (: 10 mm) whose vertical travel was sensed with a non-contact eddy-current sensor probe (Kaman Instruments, Colorado Springs, USA). The data was collected with an A/D board (DATAQ/CODAS, Akron, USA). The average depth of the wear path was measured relative to the unworn composite in the area not contacted by the enamel stylus. Traces 4–6 (region of sliding motion of the stylus under a constant load of 18–20 N) represented abrasion wear, and traces 8 and 9 (region of heightened wear under a static load of 80–90 N) represented attrition wear.

Statistical analysis Ivoclar Method The evaluation of normality of the data on scatter diagrams showed a similar pattern for all materials with one outlier in almost all materials. The Kolomogorov Smirnov test on normality revealed that a non-normal distribution was present in only 1 out of 16 material groups. Therefore, the more robust analysis on variance was applied. The Levene statistic indicated that some sort of heterogeneity of the data was present which would require a data transformation. As the value was rather small

S.D. Heintze et al. (2.2) and as transformation would have been necessary only for material wear, a data transformation was not performed. To detect differences between the total wear generated by the different stylus materials the unpaired t-test was used. For multiple comparisons of the materials, a one-way ANOVA with Tukey B correction was applied to the Ivoclar data. The significance level for all tests was set at p!0.05. OHSU Method For the OHSU data, individual 2-way ANOVAs were run for abrasion and attrition to compare the variables of materials and antagonist (3 styli) and to identify any interactions. There was no interaction, and only the variable of materials was significant. The attrition data passed the test for normality and homogeneity of variance, but the abrasion data did not pass the test for normality. However, since the abrasion data did pass the test for homogeneity of variance, all of the data was further analyzed with the Tukey test for pair-wise multiple comparisons. The significance level for all tests was set at p!0.05. Comparison Ivoclar Method/OHSU Method For the comparison of both wear methods a logarithmic transformation of the raw data was conducted as the variables had a different scale. To obtain more data points for the statistical analysis, the results for both antagonist materials were pooled. The mean values for each material/antagonist combination and method were ranked from the lowest log-transformed wear to the highest logtransformed wear. For each variable, the relative rank 1 was assigned to the material/antagonist combination showing the lowest wear, denoted by m, and the relative rank 8 to the material/antagonist combination with the highest wear, denoted by M. Then, relative ranks were defined in order to respect the relative differences between the materials. Thus, the relative rank of a material with mean x was set to 1C7(x-m)/(M-m). The Pearson correlation coefficient was calculated to assess the correlation between both test methods.

Results Ivoclar Method Mean maximal vertical material loss was statistically greater when using the Empress antagonist for Tetric Ceram and Z250, but there was no difference for Dyract or the experimental composite (Table 3).

Influence of the antagonist material Table 3

171

Ivoclar method. Material

Antagonist

Mean

SD

N

Exp. Composite Exp. Composite Dyract AP Dyract AP Z250 Z250 Tetric Ceram Tetric Ceram

Empress ExpC Empress ExpC Empress ExpC Empress ExpC

342.0 325.5 338.1 388.0 307.1 268.5 329.6 240.9

25.8 36.4 42.8 58.2 28.1 19.3 24.3 48.8

8 8 8 8 8 8 8 8

Exp. Composite Exp. Composite Dyract AP Dyract AP Z250 Z250 Tetric Ceram Tetric Ceram

Empress ExpC Empress ExpC Empress ExpC Empress ExpC

204.8 224.2 77.0 81.8 164.2 192.9 131.1 205.4

47.8 29.2 16.7 22.2 28.4 37.5 30.3 26.2

8 8 8 8 8 8 8 8

p

Wear of material pZ0.31 pZ0.071 pZ0.000 pZ0.01

Wear of antagonist pZ0.34 pZ0.631 pZ0.057 pZ0.01

Mean maximal vertical loss (mm) and standard deviation (SD) of 4 composite materials and the corresponding antagonist after 120,000 cycles of thermomechanical loading—results for Empress and experimental ceramic (ExpC) as antagonist. Statistics: unpaired t-test.

There was no significant difference in wear for any of the materials when tested with the Empress antagonist (Table 5). In conjunction with the experimental ceramic antagonist, Dyract showed significantly more wear than any of the composites, and the experimental composite had significantly more wear than Tetric Ceram and Z250. Wear of the ceramic antagonists was only significantly different when in contact with Tetric Ceram (Table 3). Antagonist wear was highest for the experimental composite and lowest for Dyract AP.

OHSU method The ANOVA revealed that there was no difference between the three antagonists for abrasion or attrition wear for any of the materials tested (Table 4). There was a difference between the materials in both abrasion and attrition. When the data was pooled in the ANOVA, Dyract experienced significantly more abrasion wear than the three composites, which were not different from one another. However, when evaluated individually by antagonist, Dyract had greater abrasion wear than only Z250 for enamel, Dyract had greater abrasion than experimental composite C and Z250 for Empress, and Dyract had greater wear than all of the composites for the experimental ceramic (Table 5). For the individual comparisons for attrition wear, Dyract only showed more attrition

than Z250 for enamel, there was no difference between any of the materials for Empress, and Dyract showed more attrition wear than all of the composites for the experimental ceramic. Use of the synthetic antagonists did not result in reduced variability in the wear data as compared with enamel.

Comparison of Ivoclar Method/OHSU Method The different materials showed a slightly different abrasion wear ranking in the two wear testing methods, especially for the Empress antagonist (Table 5). When calculating relative ranks for the different material/antagonist combinations one can see that the Ivoclar Method was closer to the OHSU attrition wear than to the OHSU abrasion wear, as confirmed by the Pearson correlation coefficients (Table 6). There was a tendency for better agreement between the two methods when the experimental ceramic was used as the stylus.

Discussion Enamel is an ideal antagonist due to its relevancy, but variations in the natural substrate and the extensive machining process required to shape it makes it somewhat less convenient and precise compared to synthetic materials. Therefore, it

172

S.D. Heintze et al.

Table 4

OHSU method.

Material

Antagonist

Abrasion Mean

Attrition SD

N

p

Mean

SD

N

p

Exp. composite Enamel Empress ExpC

23.2 25.8 23.2

7.6 5.7 9.2

6 5 6

ns

64.6 66.3 48.8

6.0 9.8 13.4

6 5 6

ns

Enamel Empress ExpC

34.8 40.6 37.2

5.6 19.6 8.4

6 6 6

ns

75.9 64.4 78.1

18.6 33.2 14.9

6 6 6

ns

Enamel Empress ExpC

19.2 21.6 17.4

1.6 5.5 9.9

6 6 6

ns

43.6 43.3 45.9

9.2 11.0 11.3

6 6 6

ns

Enamel Empress ExpC

23.6 27.3 19.7

5.1 11.4 6.5

6 6 6

ns

56.6 47.7 40.7

8.1 14.9 11.3

6 6 6

ns

Dyract

Z250

Tetric Ceram

Mean vertical loss (mm) and standard deviation (SD) of 4 composite materials after 100,000 cycles of abrasion/attrition loadingresults for Enamel, Empress and the experimental ceramic (ExpC) as antagonist. Statistics: 2-way ANOVA; nsZnot significantly different.

would be beneficial to have a synthetic material for use as a relevant antagonist for in vitro composite wear studies. The results from the OHSU simulator suggest that both glass ceramics may serve as substitutes for natural tooth enamel to study the wear of composite materials. For the Empress material, this result is in agreement with that of Shortall et al. [13] who analyzed Vickers microhardness, surface roughness, and the coefficient of friction of Empress, stainless steel and steatite compared to human enamel, and concluded that Empress had the best ability to simulate natural human enamel. In that study, wear testing of the three selected counterspecimen materials was done using a test

apparatus in which an enamel specimen was placed under a vertical load of 10 N against the rotating cylindrical counterspecimen for 3 h. The Ivoclar wear testing method did not show a close similarity for the wear testing results for both ceramic materials. Wear was significantly lower for Z250 and Tetric Ceram when tested with the experimental ceramic antagonist compared to the Empress material. An explanation for these findings could be the lower hardness of the experimental ceramic compared to Empress (difference of approximately 1 GPa). When the glass particles of the composite come into contact with the ceramic antagonist, their physical properties, shape and size determine the loss of substance. Thus,

Table 5 Ranking (least to most wear) of the materials in the two wear methods using two different antagonist materials. OHSU

Ivoclar

Enamel Abrasion a

Z250 Exp Coab Tetric Cerab Dyractb

Empress Attrition a

Z250 Tetric Cerab Exp Coab Dyractab

Abrasion a

Z250 Exp Coa Tetric Cerab Dyractb

ExpC Attrition a

Z250 Tetric Cera Dyracta Exp Coa

Abrasion a

Z250 Tetric Cera Exp Coa Dyractb

Empress

ExpC

Z250a Tetric Cera Dyracta Exp Coa

Tetric Cera Z250a Exp Cob Dyractc

Attrition Tet Cera Z250a Exp Coa Dyractb

ExpC, experimental ceramic; Exp Co, experimental composite; Tetric Cer, Tetric Ceram. Superscript letter indicates that the material is in the same statistical subgroup (OHSU and Ivoclar ANOVA post hoc Tukey B, p!0.05).

Influence of the antagonist material

173

Table 6 Relative ranks of wear material/antagonist combinations in Ivoclar and OHSU wear method.

Exp.Co./ Empress Dyract/ Empress Z250/ Empress Tetric Ceram/ Empress Exp.Co./ ExpC Dyract/ ExpC Z250/ExpC Tetric Ceram/ ExpC

Ivoclar

OHSU attrition

OHSU abrasion

6.14

6.23

3.64

5.97

5.92

8.00

4.56

1.66

1.00

5.60

2.69

2.13

5.41

2.94

3.64

8.00

8.00

8.00

2.59 1.00

2.28 1.00

1.00 2.13

Pearson correlation coefficient: Ivoclar-OHSU attrition: 0.83 (pZ0.01) Ivoclar-OHSU abrasion: 0.69 (pZ0.06).

composites may wear less against a ceramic antagonist with lower hardness. Additionally, the higher fracture toughness of the experimental ceramic compared to Empress may cause less chipping of the ceramic material, which eventually may induce less wear of the composite. The generally lower wear of the Empress antagonist material using the Ivoclar Method, which was only significant for Tetric Ceram, may also be explained by its lower fracture toughness. Cracks and chipping of the ceramic may induce more wear of the opposing material, but may also limit the wear of the antagonist. In materials with higher fracture toughness and lower hardness, wear of the composite may be lower but wear of the antagonist may be higher. Therefore, the first hypothesis, that the ceramics would produce similar results as enamel, was shown to be true with the OHSU device. The second hypothesis, that the wear of materials against the two different ceramic antagonists would be different due to their different properties, was partially shown for the Ivoclar Method, but was not substantiated for the OHSU Method. This may be due to the greater overall wear using the Ivoclar Method, which made it more discriminating for this effect. The third hypothesis, which predicted a lower variability of the wear results when using ceramic compared to enamel as the antagonist, was not confirmed. The average coefficient of variation

ranged between 18–30% for the three antagonists in the OHSU Method, with the variation being similar for all three antagonists in abrasion but actually lowest for enamel in attrition. The lack of improvement in the variability by using a standardized cusp in this study simply emphasizes the inherent variability of wear testing and the wear phenomenon in general. The reason a reduced scatter with ceramic cusps was predicted was based on the results of a recent round robin test involving 10 restorative materials, which produced results with a large variability [16]. In this study, eight composites, one amalgam, and one ceramic were evaluated with five different wear generating methods at five different sites, with every site receiving specimens made at Ivoclar by the same employee from the same batch and with the test site unaware of the brand. The five different wear simulators included the Ivoclar Method and OHSU Method (different lab than in the present study). The comprehensive statistical analysis revealed a low discrimination power and high coefficient of variation for the Zurich Method; with the best methods in this regard being those from ACTA and Ivoclar, with the OHSU Method being close behind. As both the Zurich and the OHSU Method typically use enamel as antagonists, the high scatter in that study may be explained by the different configuration and processing methods. While for OHSU the enamel of the antagonist is prepared to standardized shape and polished, the enamel cusps are only standardized by subjective molar selection and cleaned with a rotating nylon brush for the Zurich Method. Thus, the cusps may differ widely with regard to anatomical form, fluoride content in the outer surface and the amount of aprismatic enamel, with the latter two parameters having significant influence on the hardness of the enamel stylus. However, the similar amount of variability between synthetic and natural ceramic cusps in the present study suggest that the antagonist is not likely to be the most important factor determining the variability in wear results among different test methods. The fourth hypothesis was that there would be a good correlation between the two wear testing methods, and this was essentially shown to be true, especially when comparing the Ivoclar Method to the attrition wear region for the OHSU Method. It can be hypothesized that the higher load in the attrition region in the OHSU device more closely matches the wear conditions in the Ivoclar Method than does the low load abrasion wear. The Ivoclar Method showed higher wear overall, possibly due to the higher load in the region where movement of the antagonist on the surface occurs. However, it is

174 emphasized that the Ivoclar results represent maximum vertical loss, whereas the wear values reported from the OHSU Method represent mean wear depths. Another difference between the two devices is that the OHSU Method uses an abrasive medium and the Ivoclar Method does not. Therefore, the wear in the Ivoclar Method can be regarded as a consequence of a direct contact between material and antagonist, while the third body is primarily responsible for wear in the OHSU Method. However, it is likely that in both methods, wear can be described as a mixed mode process, involving abrasion, adhesion, attrition and fatigue. The round robin test comparing ten dental restorative materials with five different wear methods revealed a substantial correlation between the Ivoclar Method and OHSU Method [16]. However, 50 N was chosen for the abrasion load in the OHSU machine in that study, whereas 20 N was used in the present study. For composite materials the discriminating characteristics that control the wear rate are friction coefficient, hardness, elastic modulus, shear strength, dimension and volume of filler particles, and degree of cure of the matrix. A low elastic modulus, for example, leads to a higher contact area, thus to a lower pressure (stress). Large filler particles cause high friction that leads to high internal shear stresses in the polymer matrix. A force of 50 N is used with the Ivoclar Method. This value is regarded to be a mean value of physiological biting forces of non-bruxing patients [17]. The OHSU method in the present study, however, used a lower force for the abrasion and a higher force for the attrition. Higher forces during in vitro simulation lead to higher wear rates [18]. But the wear rate does not solely depend on the actual force that is created on the material, but also on the contact area, which may vary due to differences in the configuration of the antagonist and flattening of the antagonist during the wear test. The sharper the antagonist, the higher the wear rate [19,20]. In a study on 20 extracted first upper molars, a ball with a radius of 0.6 mm was assumed to be most suitable to the anatomical variations in human molar cusps [21]. No other measurements of this kind were found in the literature. However, analyses of plaster models of ten 20–25 year-old patients analyzed using 3-D laser equipment revealed a mean radius of 1.04 mm in the frontal segment of the cusp and 1.79 mm in the sagital segment (unpublished data by Ivoclar). The latter finding was the basis for choosing a radius of 1.18 mm for the antagonist used in the Ivoclar Method. With the Ivoclar Method the sliding movements simulate the sliding of teeth in contact [22]. Therefore, particles worn during the simulation

S.D. Heintze et al. may act as abrasive medium during the simulation. A sliding movement of 6–7 mm is a component in the OHSU method. An early publication indicated that the sliding distance should be at least twice the diameter of the stylus to ensure that particles of the materials that are abraded are easily washed away and to reduce creep effects [23]. It is likely that despite the strong mixing effect, the wear particles created in the OHSU Method become part of the abrasive, and contribute to greater wear, as has previously been demonstrated [24]. Therefore, due to the many differences between the Ivoclar and OHSU Method, it is perhaps surprising that there is a reasonable correlation between the results obtained from the two test methods. A systematic study of experimental materials with controlled differences in formulation would be necessary to further elucidate the strength of this correlation.

Conclusions This study verified a good correlation between the wear results obtained for similar composites tested by the Ivoclar wear method as shown for the attrition wear in the OHSU oral wear simulator. While the Ivoclar Method showed significantly less wear with Tetric Ceram and Z250 and the experimental ceramic stylus than with the Empress stylus, both ceramic antagonist materials showed similar results in the OHSU Method for the four test composite materials, and were also comparable to natural enamel. Despite the more uniform preparation method, the variability in the data generated with the synthetic antagonists was not reduced compared with prepared human enamel, further emphasizing the general variability of wear testing both in vitro and in vivo.

Acknowledgements This study was supported in part by NIH/NIDCR grant DE07079.

References [1] Smith BG, Bartlett DW, Robb ND. The prevalence, etiology and management of tooth wear in the United Kingdom. J Prosthet Dent 1997;78:367–72. [2] Lambrechts P, Braem M, Vuylsteke-Wauters M, Vanherle G. Quantitative in vivo wear of human enamel. J Dent Res 1989;68:1752–4. [3] Suh NP. Tribophysics. New Jersey: Prentice-Hall; 1986.

Influence of the antagonist material [4] Roulet J-F. Degradation of dental polymers. Basel: Karger; 1987. [5] Lewis G. Predictors of clinical wear of restorative dental composite materials. Biomed Mater Eng 1993;3:167–74. [6] Tyas MJ. Correlation between fracture properties and clinical performance of composite resins in Class IV cavities. Aust Dent J 1990;35:46–9. [7] Manhart J, Kunzelmann KH, Chen HY, Hickel R. Dent Mater 2000;16:33–40. [8] Dental materials - Guidance on testing of wear, Part, 2 (2001) 14569–14562 [9] Krejci I, Reich T, Lutz F, Albertoni M. In-vitro-Testverfahren zur Evaluation dentaler Restaurationssysteme. Schweiz Monatsschr Zahnmed 1990;100:953–60. [10] Condon JR, Ferracane JL. Evaluation of composite wear with a new multi-mode oral wear simulator. Dent Mater 1996;12:218–26. [11] Wassell RW, McCabe JF, Walls AW. A two-body frictional wear test. J Dent Res 1994;73:1546–53. [12] Wassell RW, McCabe JF, Walls AW. Wear characteristics in a two-body wear test. Dent Mater 1994;10:269–74. [13] Shortall AC, Hu XQ, Marquis PM. Potential counterspecimen materials for in vitro simulation wear testing. Dent Mater 2002;18:246–54. [14] Mehl A, Gloger W, Kunzelmann KH, Hickel R. A new optical 3-D device for the detection of wear. J Dent Res 1997;76: 1799–807.

175 [15] Instructions of use—Match 3D. Gra ¨felfing: Willytec GmbH; 2001. [16] Heintze SD, Zappini G, Rousson V. Wear of ten dental restorative materials in five wear simulators - Results of a round robin test. Dent Mater 2005;21:304–17. [17] Gibbs CH, Mahan PE, Lundeen HC, Brehnan K, Walsh EK, Holbrook WB. Occlusal forces during chewing and swallowing as measured by sound transmission. J Prosthet Dent 1981;46:443–9. [18] Lutz F, Krejci I, Barbakow F. Chewing pressure vs. wear of composites and opposing enamel cusps. J Dent Res 1992;71: 1525–9. [19] Krejci I, Lutz F, Zedler C. Effect of contact area size on enamel and composite wear. J Dent Res 1992;71:1413–6. [20] Kunzelmann K-H. Verschleissanalyse und -quantifizierung ¨llungsmaterialien in vivo und in vitro. Aachen: von Fu Shaker Verlag; 1998. [21] Krejci I, Albert P, Lutz F. The influence of antagonist standardization on wear. J Dent Res 1999;78:713–9. [22] Gibbs CH, Lundeen HC, Mahan PE, Fujimoto J. Chewing movements in relation to border movements at the first molar. J Prosthet Dent 1981;46:308–22. [23] Powell JM, Phillips RW, Norman RD. In vitro wear response of composite resin, amalgam, and enamel. J Dent Res 1975; 54:1183–95. [24] Condon JR, Ferracane JL. Factors effecting dental composite wear in vitro. J Biomed Mater Res 1997;38:303–13.