To develop a novel methodology to generate specimens for bi-axial flexure strength (BFS) determination from a ‘chair-side’ CAD–CAM feldspathic ceramic with surface defect integrals analogous to the clinical state. The hypotheses tested were: BFS and surface roughness ( R a ) are independent of machining variability introduced by the renewal or deterioration of form-grinding tools and that a post-machining annealing cycle would significantly modify BFS.
Nominally identical disc-shaped specimens (11 mm diameter, 1.1 mm thickness) were machined with identical design and operative parameters from Vita Mark II feldspathic ceramic. Six individual bur sets (Groups A–F) generated 14 specimens each. Three groups were annealed between glass transition and softening temperatures. 3D contact profilometry determined surface roughness before and following annealing and prior to BFS determination. Scanning electron microscopy was undertaken to examine machining tools and perform fractographic analyses of ceramic fracture fragments. Statistical analysis included independent and pairwise analyses of R a -values ( P < 0.05), Weibull analysis of BFS data and Pearsons correlations.
Mean R a differed significantly between groups ( P < 0.01) but was unaffected by annealing ( P = 0.42). Mean BFS was significantly altered by bur-set ( P < 0.01). Annealing resulted in no significant modification of the Weibull parameters as the 95% confidence intervals overlapped. No significant correlations between machining order and BFS or R a were observed ( P > 0.05).
Surface roughness and the nature of strength limiting defects appear to be probabilistic with flaw generation dependent on a random selection of a bur and a random machining sequence. The variability in BFS with machining could account for premature clinical failures.
The development of Computer Aided Design–Computer Aided Machining (CAD–CAM) technologies for dental applications has enabled clinicians to prepare and indirectly restore tooth tissue with an esthetic all-ceramic restoration, manufactured at the ‘chair-side’ in a single patient visit . Chair-side CAD–CAM techniques offer advantages to the patient including eliminating the laboratory processing costs and the requirement for intra-visit temporization of the prepared tooth structure . A range of dental ceramic substrates have been developed for ‘chair-side’ machining and are presented as pre-fabricated blocks, manufactured using processing routes identified to reproducibly control the resultant ceramic composition and microstructure . However, effective machining of the fully sintered ceramics utilized in ‘chair-side’ CAD–CAM processes must balance machining accuracy and efficiency with instrument durability and the inevitable damage accumulation that occurs at the surface and within the machined substrate .
Despite the development of novel techniques including electro-discharge machining , laser assisted machining , abrasive jet machining and chemically assisted machining , form-grinding remains the most economical and accessible machining process for dental application . As a result, form-grinding is almost universally employed in commercial chair-side dental CAD–CAM equipment . Form-grinding is an abrasive process in which the removal of the ceramic surface is achieved by the generation of chip fragments produced by the cutting edges of relatively hard particles located on the surface of the grinding tool . Analysis and modeling of the chip fragments and cutting forces in combination with surface metrology and qualitative assessment of the machined substrate has demonstrated that ceramic material removal is subject to fracture dominant mechanisms . Loading with the machining tool induces strain at the ceramic surface which precedes crack initiation, propagation, chip formation and results in the generation of a new surface texture . However, the formation of cracks on and beneath the newly formed surface, classified as surface and sub-surface damage, respectively, is associated with the material removal . The most common ‘chair-side’ CAD–CAM ceramic substrates, namely feldspathic or glass–ceramics, are susceptible to such damage accumulation as they possess relatively low fracture toughness and have low damage tolerance . A further consequence of form grinding is the coexistence of brittle fracture with plastic deformation of the ceramic surface and the development of surface irregularities consistent in pattern with the nature and incidence of the machining tool . As a result the strength of the machined restoration may differ significantly from the true strength of the machined substrate, since strength will be determined by the character and distribution of the surface and subsurface defects, stress concentrators and transient and residual stress states . When the first dental CAD–CAM system was introduced in 1985, form grinding was achieved using a single axis abrasive wheel but was further developed to employ a combination of a grinding wheel and a diamond impregnated bur . Most contemporary dental CAD–CAM systems now utilize paired diamond-impregnated burs which can possess matching or dissimilar geometries . However, dental form-grinding innovation to date has largely focused on the generation of more complex geometries to an improved dimensional accuracy and the role of the machining process on restoration performance has yet to be explored.
The overall objective of the current investigation was to develop a methodology to enable the generation of nominally identical ‘as-machined’ geometric test specimens from a chair-side CAD–CAM dental ceramic. The use of commercial dental CAD–CAM form-grinding equipment to fabricate the specimens would enable the determination of the biaxial flexure strength (BFS) of the ceramic in the equivalent ‘as-machined’ state encountered in dental restorations clinically. The specific hypotheses tested were that the BFS is independent of machining induced variability determined by the renewal or deterioration with aging of the form-grinding tools. A secondary aim was to explore the possibility of modifying the impact of the machining induced damage accumulation on the BFS of the ceramic substrate by subjecting the ceramic to a post-machining annealing treatment which could be performed in the dental office. The hypothesis tested was that the BFS of the ‘as-machined’ ceramic would be significantly altered following annealing using a regime previously demonstrated to be effective in reducing surface crack length in the ceramic substrate under investigation.
Materials and methods
Generation of a CAD template
Nominally identical disc-shaped fine grain feldspathic ‘as-machined’ ceramic specimens were manufactured using the CEREC ‘chair-side’ dental CAD–CAM system (Sirona Dental Systems GmbH, Bensheim, Germany). A disc-shaped template (11 mm diameter and 1.1 mm thickness) was formed by condensing a commercial dental resin-based composite (RBC) (Filtek Z250, Shade A3, Lot 8TL, 3M ESPE, St. Paul, MN, USA) into a Teflon cylindrical split-mold under acetate and glass coverage. The RBC was light irradiated from both exposed aspects for 20 s with an Optilux 501 (SDS Kerr, Danbury, CT, USA) light-curing unit at an output intensity of 740 mW cm −2 delivered through an 11 mm light curing tip diameter. The irradiated RBC disc-shaped template was embedded into a polyvinylsiloxane impression material (President Putty, Coltene Whaledent AG, Altstätten, Switzerland) to create a disc-shaped relief of equal geometry to the template following its removal from the hardened silicone mass. The CEREC 3D image acquisition unit was employed to obtain an optical impression of the disc-shaped relief cavity, the CEREC 3D software was used to generate a disc-shaped pattern and subsequently refine and level the upper surface of the CAD template.
Fabrication of the geometric test specimens
A CEREC MC XL milling unit was used to reduce Vita Mark II fine grain feldspathic ceramic blocks (10 mm × 12 mm × 15 mm) (2M3C-I12, Vita, Bad Säckingen, Germany) to the prescribed CAD template using diamond impregnated burs and water coolant ( Fig. 1 ). Six new paired bur sets comprising each of one cylindrical (Cylinder pointed bur 12, Sirona Dental Systems GmbH, Bensheim, Germany) and one stepped pattern (Cylinder step bur 12, Sirona Dental Systems GmbH, Bensheim, Germany) were used in the current investigation to generate six groups of ‘as-machined’ discs (Groups A–F). Bur sets were replaced as a pair following bur failure either due to fracture of the bur shank or when indicated by the milling unit. The earliest bur failure determined the maximum sample size for all individual groups thereby ensuring even group sizes. Specimen machining sequence and machining duration was recorded for each individual specimen. The sprue attachment created by machining was located at a single point on the circumference of the upper surface of the disc-shaped specimens and was removed manually using P500 and P1000 grades of silicon carbide abrasive paper (Struers, Glasgow, UK) sequentially. The position of the sprue attachment allowed accurate re-orientation of specimens throughout the experiment.
Machined surface roughness characterization
A contact stylus profilometer (Talysurf CLI 2000 Taylor-Hobson Precision, Leicester, UK) was used to measure the surface roughness ( R a -value) of the lower surface of the ‘as-machined’ ceramic discs. A central area 2.5 mm width and 5.5 mm length was scanned using an inductive gauge (500 μm range) with a 2 μm radius diamond stylus 90° conisphere tip. Scans were performed using a 10 μm spacing in the x -axis and a 4 μm spacing in the y -axis at a scanning speed of 1 mm s −1 providing a 10 nm resolution in the z -axis. The R a -value generated was a measure of the arithmetic mean of the absolute departures of the roughness profile from a mean (horizontal) line and was calculated from the mean of the 626 profiles produced from the area examined. A Gaussian filter conforming to ISO3274:1996 and a 0.25 mm cut-off enabled the removal of waves above certain amplitudes allowing visualization of the introduced resident surface texture as a three-dimensional (3D) image generated from the multiple traces recorded. The mean and standard deviations of the R a -values were calculated and representative profiles examined.
Specimen storage and annealing regime
Following surface profilometry, Groups A–C specimens were stored in a dessicator maintained at 23 ± 1 °C and a relative humidity of 50 ± 5% until required for testing whereas Groups D and E specimens were subjected to an annealing regime. The disc-shaped specimens were positioned with the lower surfaces in contact with a flat silicon nitride refractory tray. Specimens were annealed according to a regime identified by Denry et al. which was reported to result in surface crack length reduction for the feldspathic substrate investigated. The Vita Mark II specimens were heated to at 900 °C for 1 h prior to slow cooling to room temperature over 2.5 h. The annealing temperature was above the glass transition (796 ± 5 °C) and below the softening temperature (914 ± 8 °C) of Vita Mark II . Specimens were orientated to allow re-measurement of the surface roughness in the identical region that was characterized before annealing and using the methods outlined in the previous section.
Bi-axial flexure strength determination
The disc-shaped specimens were supported on a 10 mm diameter knife-edge ring and the upper surface was centrally loaded with a spherical ball indenter (4 mm diameter) at a cross-head speed of 1 mm min −1 . The BFS was then calculated according to the Timoshenko and Woinisky Kriegers analysis (Eq. (1) ):
σ max = P h 2 ( 1 + ν ) 0.485 × ln a h + 0.52 + 0.48