The manufacture of dental crowns and bridges generates residual stresses within the veneering ceramic and framework during the cooling process. Residual stress is an important factor that control the mechanical behavior of restorations. Knowing the stress distribution within the veneering ceramic as a function of depth can help the understanding of failures, particularly chipping, a well-known problem with Yttria-tetragonal-zirconia-polycrystal based fixed partial dentures. The objective of this study is to investigate the cooling rate dependence of the stress profile in veneering ceramic layered on metal and zirconia frameworks.
The hole-drilling method, often used for engineering measurements, was adapted for use with veneering ceramic. The stress profile was measured in bilayered disc samples 20 mm in diameter, with a 0.7 mm thick metal or Yttria-tetragonal-zirconia-polycrystal framework and a 1.5 mm thick veneering ceramic. Three different cooling procedures were investigated.
The magnitude of the stresses in the surface of the veneering ceramic was found to increase with cooling rate, while the interior stresses decreased. At the surface, compressive stresses were observed in all samples. In the interior, compressive stresses were observed in metal samples and tensile in zirconia samples.
Cooling rate influences the magnitude of residual stresses. These can significantly influence the mechanical behavior of metal-and zirconia-based bilayered systems. The framework material influenced the nature of the interior stresses, with zirconia samples showing a less favorable stress profile than metal.
Mismatch in thermal expansion properties between core and veneering ceramic, and thermal gradients occurring during the cooling/solidification period of the veneer firing process induce residual stresses in crowns and fixed dental prostheses (FDPs) . These residual stresses greatly influence the strength and fracture characteristics of metal- and zirconia-based restorations. Particularly, the stress vs. depth profile is a key factor for understanding veneering ceramic fractures . Such fractures are a major cause of failure with Yttria-tetragonal-zirconia-polycrystal (Y-TZP) based FPDs . As the veneering ceramic cools from the surface to the center, thermal gradients induce non-uniform solidification, thereby causing contraction mismatch within the ceramic . The resulting residual stresses are compressive at the surface and tensile in the interior, as also occurs during the industrial manufacture of tempered glass. The size and distribution of the stresses generated by thermal gradients depends on material conductivity, thickness, and cooling rate at temperatures above the glass transition temperature T g . In the case of ceramic fused to metal frameworks, the thermal expansion coefficient (CTE) of the ceramic is generally slightly lower than the framework so that during cooling from T g to room temperature, compressive stresses are developed within the ceramic, with compensating tensile stresses developed within the framework . The thermal gradients and CTE mismatch effects combine to form the stress profile .
Until now, the cooling rate dependence of stress profiles in veneering ceramics layered on a core material has been studied through mathematical models . Swain studied the independent influence of the cooling rate, thickness and thermal expansion coefficient on residual stresses profile within a simple bilayer model composed of glass ceramic, alumina or zirconia substrates . The drawback of such models is the difficulty encountered when mimicking variations of thermal gradients and materials thermal properties during the firing process. Asaoka and Tesk developed an analytical model of a porcelain-fused-to-metal (PFM) beam that incorporates variations of several temperature-dependent factors, and using a constant cooling rate on both specimen surfaces and then a constant temperature distribution. They studied the influence of cooling rate and thermal expansion coefficient mismatch on residual stress profiles. They found compressive residual stresses at the surface of the ceramic and tensile residual stress at the opaque interface. Stresses were found to increase with cooling rate. DeHoff et al. developed a metal–ceramic disk model also based on the visco-elastic theory and reported that tempering by air blasting induces large compressive stresses in ceramic surface while cooling in air induces small tensile stresses. They also showed that length of cracks induced in the surface of the ceramic by microhardness indentation was smaller for tempered samples.
Tempering has often been reported to strengthen the veneering ceramic by developing compressive stresses in the surface . As with tempered glass, these stresses are intended to counteract external loads. In another hand some ceramic manufacturers have changed their recommendations relating to the veneering process of zirconia frameworks. A slow cooling procedure, from temperature of liquid phase sintering to T g , furnace closed, is now proposed for the last firing cycle (VM9 firing procedure, Vita Zahnfabrik, Bad Säckingen, Germany, www.vita-zahnfabrik.com/resourcesvita/shop/en/en_3055371.pdf ). Manufacturers contend that this procedure reduces tensile zones, and then cracks and chipping within the veneering thickness, but at this time there is no scientific evidence to support this contention.
Recently a residual stress measurement method was introduced for dental applications . The hole-drilling method, a standardized method developed for industrial applications , was adapted for measuring residual stresses in the veneering ceramic. This method is based on the removal of some stressed material and the measurement of the resulting deformations in the adjacent material . The deformations are measured on the surface, typically using strain gages, from which the residual stresses can be calculated. Stresses are calculated from surface to depth, typically with 0.1 mm steps, and giving a stress profile within a 1.2 mm depth.
The first objective of this study is to investigate the cooling rate dependence of stress profiles in PFM and Y-TZP structures using the hole-drilling method. A second objective is to investigate how modified firing procedures can influence mechanical behavior of zirconia-based restorations, comparing stress profile in metal- and zirconia-based structures.
Materials and methods
Bilayered disc samples composed of veneering ceramic sintered either on Y-TZP framework (VZr, 6 samples), or on dental CoCr alloy framework (VM, 6 samples) were manufactured following standard dental laboratory procedures and manufacturer’s recommendations. CoCr core discs (Duceralloy C, DeguDent GmbH, Hanau, Germany), 20 mm diameter, were cast and sequentially ground with 80-grit, 180-grit and 500-grit silicon carbide discs (Struers LabPol polishing machine, Copenhagen, Denmark) to a thickness of 0.70 ± 0.02 mm. The surface to be veneered was sandblasted at 4 bars with 125 μ alumina particles.
Y-TZP core discs were cut out of a pre-sintered Vita In-Ceram YZ blocks (Vita Zahnfabrik, Bad Säckingen, Germany), were rounded by polishing, and densely sintered at 1530 °C for 120 min with heating rate 10 °C/min, and heating time 149 min (Zircomat furnace, Vita Zahnfabrik, Bad Säckingen, Germany). The sintered Y-TZP discs were ground and dimensioned in the same way as CoCr, but not sandblasted.
CoCr and Y-TZP discs were veneered respectively with Vita VM 13 and Vita VM9 feldspar veneering ceramic (shade 3M2) (Vita Zahnfabrik, Bad Säckingen, Germany). A Vita Vacumat 4000 Premium furnace (Vita Zahnfabrik, Bad Säckingen, Germany) was used for all firing procedures, as summarized in Table 1 . All samples were baked on the same ceramic mesh-tray.
|Starting T (°C)||Pre-drying t (min), closing t||Heating rate (°C/min)||Heating t (min)||Firing T (°C)||Holding t (min)||Vacuum holding t (min)||Slow cooling ending T (°C)||Slow cooling rate (°C/min)|
|Alloy core oxidation||600||3||75||4||900||2||4|
|Vita VM13 opaque||600||2||75||4||900||1||4|
|Vita VM13 dentin||600||8||50||6||900||6||6|
|Y-TZP core regeneration firing||500||–||100||5||1000||15||–|
|Vita VM9 effect bonder||500||6||75||6||950||1||6|
|Vita VM9 dentin||500||6||55||7.27||910||4||7.27|
|Last firing cycle|
|Classic Cooling (CC)||600||8||50||6||900||6||6|
|Slow cooling (SC)||Room t °||10||900||6||Room t °||2|
|Modified Cooling (MC)||600||8||50||6||900||6||6||600||33|
The sandblasted surfaces of the VM samples were oxidized before ceramic layering according to the manufacturer’s guidelines. Vita VM 13 Opaque ceramic powder mixed with Vita VM opaque fluid was applied to the substructure with a brush, and fired to enhance the bond to the alloy surface. Three layers of dentin ceramic were successively fired. This layering technique promotes adhesion between opaque and dentin ceramic and reproduces the dental laboratory procedure. Samples were sequentially ground with 80-grit, 180-grit and 500-grit silicon carbide discs to the thickness of 2.2 ± 0.02 mm to create a 1.5 mm thick ceramic layer on a 0.7 mm thick framework.
For the VZr samples preparation, the Y-TZP discs were exposed to a “regeneration firing”, which is a final thermal treatment of the core to reverse any phase transitions in the zirconia due to the grinding procedures. A thin coat of Vita VM 9 Effect Bonder was applied and fired on the surface to be veneered. Then, Vita VM 9 Base Dentin was progressively layered on the effect bonder and samples were dimensioned in the same way than VM samples.
After final polishing, all specimens were exposed one by one to a last firing cycle. This last firing cycle restores the residual stress profile through the veneering ceramic thickness. All samples were placed in the same position, on the center of the mesh-tray and of the furnace. Three different cooling schedules were performed. One sample of each group (VM and VZr) was tempered from 900 °C to room temperature by opening the furnace door, as classically performed in dental laboratories, and removed from the mesh-tray at 200 °C (Classic Cooling, CC). This schedule was the one used during the veneering layering process. In comparison with manufacturer recommendations, the firing temperature was maintained 6 min in place of 1 min in order to reach 900 °C within the framework. The second one was submitted to the slow cooling procedure recently proposed by Vita for VM9 (Vita Zahnfabrik, Bad Säckingen, Germany). This procedure comprised a slow cooling from 900 °C to 600 °C, furnace closed (Modified Cooling, MC). The last one was cooled at 2 °C/min in a special furnace (Carbolite LMF 12/2, Carbolite, Hope Valley, UK), from 900 °C to room temperature (Slow Cooling, SC). All firing schedules are summarized in Table 1 .
Determination of temperature profiles
Temperatures on the surface of the veneering ceramic and of the framework during CC and MC were measured with S-type thermocouples and recorded with NI LabView software (National Instruments, Austin, Texas, USA). The first thermocouple was placed in contact with the center of the veneering ceramic. The second one was placed in contact with the center of the framework surface, through the mesh-tray.
The CTE at temperatures above and below T g for VM9 ( T g ∼ 600 °C, T s ∼ 670 °C, following the manufacturer), VM13 ( T g ∼ 560–565 °C, T s ∼ 635 °C, following the manufacturer), zirconia and CoCr alloy were measured using a single pushrod dilatometer (Netzsch DIL 402C) at a heating rate of 2 K min −1 up to 700 °C. Beam samples were manufactured for each veneering ceramic using a rectangular mold into which powders were condensed and sintered. For Duceralloy C, a cylindrical sample was cast and fired for 30 min at 950 °C to simulate veneering procedure and to eliminate internal stresses due to casting. For Vita In-Ceram YZ, a rectangular sample was manufactured, sintered and submitted to a regeneration firing procedure at 1000 °C. Each sample was tested two times.
Strain gage rosette installation
A specialized six-element Type C rosette (N2K-06-030RR-350/DP, Vishay, Malvern, PA, USA) was installed on the center of the veneering ceramic surface. To promote the strain gage bond, the ceramic surface was prepared by etching with 10% hydrofluoric acid for 1 min, and was then cleaned for 5 min in an ultrasonic bath containing 90% alcohol. The strain gage rosette was installed with M-Bond 200 Adhesive (Vishay, Malvern, PA, USA), following the manufacturer’s instructions. The adhesive was allowed to cure overnight to ensure complete curing. The installation was monitored using an optical microscope.
Electrical measurement chain
The strains expected from the strain gages are very small and cannot be measured with sufficient accuracy using conventional industrial equipment. A specialized data acquisition system was therefore built where each strain gage was connected in a Wheatstone bridge circuit with 3 control gages (identical gages attached to an undisturbed sample). All gages and control rosettes were exposed to identical constant temperature conditions. Finally, the very low voltage measurements were performed with specific custom-built electronic equipment comprising a precision DC and AC current source 6221 (Keithley Instruments, Inc., Cleveland, OH, USA) and 3 Nanovoltmeters 2182A (Keithley Instruments, Inc., Cleveland, OH, USA). Filtered measurements were recorded on a computer using NI LabView software (National Instruments, Austin, TX, USA).
The specimens were placed in an aluminum container. After sample centering in the drilling machine, the container was filled with silicon oil to enhance drilling lubrication, thermal conductivity and electrical insulation. In addition, the silicon oil bath was thermally controlled and maintained at 36 ± 0.1 °C with a Eurotherm 3208 system (Eurotherm Ltd., Worthing, UK) to avoid the effects of any ambient temperature variations. Temperature at the sample contact was recorded with a thermocouple connected to NI LabView data acquisition system.
An Isel CAD-CAM machine (CPM 3020, Houdan, France) was used for the drilling procedure. To increase strain sensitivity, the maximum allowable hole diameter for the strain gage rosette type was performed with a 2.5 mm diameter cylindrical bur (Asahi Diamond Industrial Europe SAS, Chartres, France). The bur rotation speed was 19,000 rpm. A hole was cut at the center of the rosette in steps of 0.1 ± 0.01 mm, as measured by a Digimatic indicator (Mitutoyo Corporation, Kawazaki, Japan). Hole diameter and concentricity were checked after the experiment with an optical microscope and motorized micrometer, Micro Controle CV 78 (Newport, Irvine, CA, USA). The protocol of the hole-drilling method was designed to eliminate/minimize crack initiation through choice of drilling process, drill type and lubricant used. However, in the few cases where cracks had nevertheless occurred, abnormal large variations of microstrains were induced. If these were confirmed by optical microscopy, the sample was eliminated.
Strain measurements and residual stress calculation
Strain measurements were taken continuously during each step of the drilling procedure and for 10 min afterwards. This time allowed stabilization of any temperature fluctuations caused by the drilling process. The strain measurements were recorded in an Excel spreadsheet (Microsoft Corporation, Redmond, WA, USA). Mean values were evaluated for each strain gage based of the final 200 values (1 Hz acquisition) registered for each step. The corresponding profiles of residual stress vs. depth from the specimen surface were then calculated according to ASTM Standard Test Method E837-08 using H-Drill software (Vishay, Malvern, PA, USA). For the rosette size used, the hole-drilling method can measure residual stresses to depths to 1.2 mm.