The objective was to explore how clinically relevant machining process and heat treatment influence damage accumulation and strength degradation in lithium silicate-based glass ceramics machined in the fully crystallized state.
A commercial zirconia-reinforced lithium silicate (ZLS) glass ceramic with a fully developed microstructure (Celtra® Duo) was studied. Disk-shaped specimens (nominal 10 mm diameter and 1 mm thickness) were fabricated either using a CAD-CAM process, creating a clinically relevant dental restoration surface, or were sectioned from water-jet cut cylindrical blocks with their critical surfaces consistently polished. Bi-axial flexure strength (BFS) was determined in a ball-on-ring configuration, and fractographic analysis was performed on failed specimens. XRD, AFM and SEM measurements were conducted before and after heat treatment. For each sample group, BFS was correlated with surface roughness. A two-way ANOVA and post-hoc Tukey tests were used to determine differences in BFS between machining and heat treatment groups (ɑ = 0.05).
A two-way ANOVA demonstrated that BFS was influenced by fabrication route (p < 0.01) with CAD-CAM specimens exhibiting significantly lower mean BFS. A factorial interaction was observed between heat treatment and machining route (p < 0.01), where a significant strengthening effect of post-manufacture heat treatment was noted for CAD-CAM specimens but not sectioned and polished samples. CAD-CAM specimens exhibited sub-surface lateral cracks alongside radial cracks near fracture origin which were not observed for polished specimens. BFS did not correlate with surface roughness for polished specimens, and no change in microstructure was detectable by XRD following heat treatment.
The mechanical properties of the ZLS ceramic material studied were highly sensitive to the initial surface defect integral associated with manufacturing route and order of operations. CAD-CAM manufacturing procedures result in significant strength-limiting damage which is likely to influence restoration performance; however, this can be partially mitigated by post-machining heat treatment.
Developments in intra-oral three-dimensional (3D) imaging and ‘chair-side’ computer aided design – computer aided manufacturing (CAD-CAM) have resulted in efficient dental digital workflows for prosthesis fabrication, which reduce patient treatment times compared with traditional methodologies. However, a rate-limiting step for the manufacturing of all-ceramic restorations is the necessity to perform post-machining treatments to improve the mechanical properties including toughness and flexural strength prior to clinical use [ ]. Machining of the ceramic substrate in a precursor form – which lacks the final microstructure, is typically required to balance machining efficiency and machining tool longevity with the accumulation of strength-limiting damage in the machined prosthesis [ ]. Heat treatments are subsequently used to develop the microstructure by sintering of zirconia ceramics and crystallization of lithium silicate glass-ceramics. However, heat treatments are often undesirable, increasing patient treatment times between tooth preparation and restoration placement. As a consequence, material developments have been targeted at eliminating or reducing lengthy post-machining heat treatment processes.
Currently, the most common substrates used for the ‘chair-side’ manufacture of all-ceramic restorations are lithium silicate based glass-ceramics, which can accurately mimic the shade, translucency and lustre of the natural dentition [ ]. Much of the evidence demonstrating the long-term clinical success of CAD-CAM lithium silicate glass-ceramics has been generated for one commercial material, namely, IPS e.max® CAD (Ivoclar Vivadent Inc) [ , ]. IPS e.max® CAD contains a lithium metasilicate crystalline phase and requires a post-machining heat treatment to crystallise the lithium disilicate which largely replaces the lithium metasilicate [ ]. A recent development of lithium silicate glass-ceramics is the incorporation of zirconia dissolved in the parent glass. The zirconia is not present in a crystalline form, but rather acts alongside phosphorus pentoxide as a nucleating agent and results in a final ceramic microstructure comprised of lithium metasilicate crystals surrounded by nanometric lithium orthophosphate crystals and a minor lithium disilicate crystalline fraction within the residual glass [ ]. Zirconia-reinforced lithium silicate (ZLS) glass-ceramics are advocated to be machined in either a partially crystallized form (Vita Suprinity®, Vita Zahnfabrik) or alternatively with a ‘fully developed’ microstructure (Celtra® Duo, Dentsply Sirona). The manufacturer of Celtra® Duo advocates that considerable clinical efficiencies can be gained by machining ZLS with a ‘fully developed’ microstructure [ ], such that dental prostheses can be milled and manually polished prior to insertion in as little as ∼15 min., or with the addition of a glazing firing cycle – which reportedly acts to strengthen the ceramic, in ∼30 min [ ].
The multipoint contact grinding processes used in dentistry to shape the prosthesis in chair-side CAD-CAM are known to introduce complex patterns of surface and subsurface damage which limit the strength of the prosthesis [ , ]. Material removal during machining occurs when diamond particles embedded in the surfaces of rotating burs contact the ceramic thereby generating high surface strains which initiate microcracking and chip fragment formation [ , ]. The contact between the abrasive and the surface also results in residual cracks, both normal to and parallel with the cut surface, which interact with a subsurface residual stress field created through both plastic deformation and heat generated by friction [ , ]. Machining damage has been shown to extend to at least 50 μm from the surface and is microstructure dependent, with the hardness and toughness of the substrate determining the magnitude of the machining contact forces [ , ]. Therefore, the resultant surface defect population and the residual stresses dictate the effective strength of the restoration, but both may be modified by post-machining heat treatments [ ].
The primary aim of the study was to determine the impact of surface damage generated by a clinically relevant CAD-CAM process on the strength of a ZLS glass-ceramic machined in a fully crystallized form and to relate the findings to the nature of surface damage accumulation during machining. The secondary aim was to assess how the order of post-machining procedures that result in a modification of surface defects and residual stress states, including heat treatment, influence the mitigation of machining associated surface damage.
Materials and methods
CAD-CAM test specimen preparation
A zirconia reinforced (10 wt% zirconium dioxide) lithium silicate ceramic, Celtra® Duo HT (Lot# 18028463, Dentsply Sirona, USA) was acquired in the form of prefabricated blocks (12 × 14 × 18 mm) for use in a ‘chair-side’ dental CAD-CAM unit. Two groups of nominally identical disk-shaped specimens (n = 15) were fabricated using a CAD-CAM workflow. In accordance with a previous study, 10 mm diameter and 1.0 ± 0.1 mm thickness disk-shaped specimens were machined using a Sirona CEREC MC X ‘chair-side’ milling unit (Dentsply Sirona, USA) from the prefabricated blocks [ ]. A CAD template was acquired by scanning an equivalent geometry disk-shaped mould with a CEREC Omnicam (Dentsply Sirona, USA) and the design was completed in CEREC v4.5 software. New bur sets were installed before the machining of each disk-group investigated, and each disk was labelled on the upper surface following machining to ensure a consistent orientation for both subsequent heat treatment and mechanical testing protocols. The tab left by machining (close to the upper surface) was manually removed using SiC abrasive paper under water lubrication prior to mechanical testing. One group of CAD-CAM specimens received no further processing (CAD) while the second group were heat treated (CAD-HT). The heat treated specimens were orientated with the marked upper surfaces not in contact with the firing slab and subjected to the manufacturer’s recommended ‘polish and fire’ cycle in an Ivoclar Vivadent Programat EP 5000 furnace. The ‘polish and fire’ cycle involved the disk-shaped specimens being pre-heated at 500 °C for 2 min., the temperature increased at 60 °C/min to 820 °C, held for 1 min prior to long term cooling from 750 °C.
Control surface test specimen preparation
Four additional groups of specimens (n = 15) were fabricated by first reducing the rectangular prefabricated Celtra® Duo blocks to 10 mm diameter cylinders (nominal) using water-jet cutting. Disks of 10 mm diameter and 1.0 ± 0.1 mm thickness were subsequently sectioned from the prepared cylinders using a diamond blade in an IsoMet 5000 precision saw (Buehler, USA) at a blade speed of 300 rpm and a feed rate of 4.0 mm/min, using water coolant. One group of disk-shaped specimens received no further treatment (CUT). In the second group, specimens were manually polished on one surface sequentially using SiC abrasive papers (P800, P1000, P1500 and P2000 grits) for fixed time intervals of 90 s using water as a lubricant (POL). A further group of disk-shaped specimens were manually polished in accordance with procedure above (POL) and then heat-treated according to ‘polish and fire’ cycle previously outlined (POL-HT). In the final group, specimens were heat-treated according to ‘polish and fire’ cycle prior to manual polishing (HT-POL). To account for any variability induced by the ceramic blocks, cutting sequence or different individual firings, ten additional disk-shaped specimens were sequentially cut from a single prefabricated block and randomised. Subsequently, five specimens were manually polished as previously described (POL-HT SUPP ) before all ten disk-shaped specimens were heat treated according to ‘polish and fire’ cycle in a single furnace firing cycle. The remaining five samples were then polished (HT-POL SUPP ) after heat treatment.
Surface metrology was conducted on each disk-shaped specimen to discriminate between differences in the surface roughness introduced by specimen manufacture and post-machining processes. A Nexview 5000 optical profilometer (Zygo, USA) with a 10× Mirau objective lens and 2.0 x magnification provided a resolution of 0.1 nm in the z-axis and was used to determine the mean Ra-value from a 0.36 × 0.27 mm measurement area arbitrarily chosen on each sample close to the specimen centre.
Bi-axial flexure strength (BFS) determination
An ElectroPuls E3000 universal testing apparatus (Instron, USA) was used for bi-axial flexure strength (BFS) testing of the disk-shaped specimens using a ball-on-ring configuration. Disks were coaxially positioned on an 8.8 mm diameter knife-edge ring support. For CAD and CAD-HT groups, the unmarked specimen surfaces were placed in contact with the ring support. For the POL, POL-HT and HT-POL groups, the polished specimen surfaces were placed in contact with the ring support. The disks were centrally loaded with a 7.9 mm diameter stainless steel ball indenter at a crosshead speed of 1 mm/min. BFS (MPa) was calculated using a Timoshenko and Woinowsky Krieger formula [ ] (Eq. 1):