Tensile creep and 750 °C heat treatment induce nanovoiding in dental porcelain.
Transmission electron microscopy study for local distribution and size assessment.
Small angle neutron scattering for comparative bulk (mm scale) analysis.
Heat treatment increases void number density from 1.61 μm −2 to 25.4 μm −2 .
Creep results in a void number density of 98.6 μm −2 and promotes void growth
Recent studies of the yttria partially stabilised zirconia–porcelain interface have revealed the presence of near-interface porcelain nanovoiding which reduces toughness and leads to component failure. One potential explanation for these nanoscale features is thermal creep which is induced by the combination of the residual stresses at the interface and sintering temperatures applied during manufacture. The present study provides improved understanding of this important phenomenon.
Transmission electron microscopy and small angle neutron scattering were applied to a sample which was crept at 750 °C and 100 MPa (sample C), a second which was exposed to an identical heat treatment schedule in the absence of applied stress (sample H), and a reference sample in the as-machined state (sample A).
The complementary insights provided by the two techniques were in good agreement and log-normal void size distributions were found in all samples. The void number density was found to be 1.61 μm −2 , 25.4 μm −2 and 98.6 μm −2 in samples A, H and C respectively. The average void diameter in sample A (27.1 nm) was found to be more than twice as large as in samples H (10.2 nm) and C (11.6 nm). The crept data showed the highest skewness parameter (2.35), indicating stress-induced growth of larger voids and void coalescence that has not been previously observed.
The improved insight presented in this study can be integrated into existing models of dental prostheses in order to optimise manufacturing routes and thereby reduce the significant detrimental impact of this nanostructural phenomenon.
Recent decades have seen significant advances in ceramic processing which have facilitated the manufacture of more efficient and practical ceramic dental prostheses . The appealing appearance of these systems compared to traditional metallic based implants has meant that despite questions regarding their mechanical reliability, ceramic prostheses have become popular with patients and dental technicians .
The exceptionally high toughness of Yttria Partially Stabilised Zirconia (YPSZ) in combination with its high strength and biocompatibility has resulted in the widespread use of this ceramic material in dental applications . Previously, one of the main concerns when using YPSZ in prosthesis manufacture has been the interaction between YPSZ and enamel in natural teeth , as well as the need for managing the aesthetic appearance of the completed restoration. When used in dental prosthetics applications, YPSZ is therefore typically veneered with porcelain in order to tailor both the shade and pearlescence of the completed prosthesis and to match the appearance of natural teeth . Dental porcelain is a predominantly amorphous ceramic composed of silica, alumina and a range of other oxides selected to produce a material with optimal mechanical properties and manufacturability .
Despite the many benefits of this manufacturing approach, the origin of failure in YPSZ–porcelain prostheses is primarily located within the porcelain veneer . Brittle porcelain failure may be induced at surfaces through effects such as wear faceting , or may nucleate at defects within the bulk material . Investigations into the influence of ineffective design and manufacturing route have shed light onto the origins of these types of failure. For example, recent modelling and experimental studies have shown that the interaction between defects and transient stresses induced by thermal gradients increase the likelihood of failure of a porcelain veneer .
Fractographic studies of failed prosthesis (induced both by clinical use and simulated mastication loading) have indicated that failure can also be preferentially initiated within a few hundred microns of the YPSZ–porcelain interface . Microscale analysis has demonstrated that the microstructure and mechanical characteristics of this near-interface zone are distinct from the neighbouring regions . For example, one recent spatially resolved micropillar splitting study shows a significant reduction (of up to 90%) of porcelain fracture toughness within the first 100 μm of the interface, as shown in Fig. 1 . The exact origins of this near-interface variation has so far remained unclear, although an opinion has been expressed by some researchers that failure may also be associated with the tensile residual stresses that arose during manufacture . Guided by these considerations, we recently used Transmission Electron Microscopy (TEM) of the YPSZ–porcelain interface region to obtain improved insight into the nanostructure of this near-interface region . This analysis confirmed the presence of porcelain nanovoiding (with void sizes of ∼10 nm) in a band located a few microns from the YPSZ–porcelain interface, where the residual stresses are also believed to be maximum.
Recent studies have also revealed that the combination of residual stress and high sintering temperatures applied to the near interface region during manufacture are sufficient to induce localised creep in dental porcelain . Micro-to-nano scale void-like features have previously been observed during creep studies of porcelain and are known to lead to reduced material strength, stiffness and fracture toughness . The origin of this voiding behaviour is believed to be associated with the diffusional motion of vacancies within the atomic structure that induces the early onset of component failure in amorphous materials .
Based on the above considerations we hypothesise that porcelain creep at the YPSZ–porcelain interface leads to the generation of nanovoids that ultimately reduce the mechanical strength of this near interface region and lead to failure. Critical evaluation of this hypothesis is the subject of the present study. To assess and quantify the nanoscale porosity associated with creep, we make a comparison between three porcelain specimens. Sample C was crept at a temperature and stress representative of the conditions present at the YPSZ–porcelain interface , sample H was exposed to the same thermal conditions but without the application of load, and sample A represents the original as-machined state.
Only a limited number of experimental techniques can provide complete characterisation of cavitation phenomena at the nanometre length scale. In the analysis that follows two independent but complementary methods were used. Small Angle Neutron Scattering (SANS) is a reciprocal space technique based on the analysis of diffraction spectra from the samples that provides information regarding the mean shape, number density, size and distribution of voids with sizes in the range ≈1–100 nm . The large gauge volumes (typically >mm length scales) associated with SANS ensure that this technique provides results representative of the average bulk response, but, conversely, the technique cannot provide insight into local features .
To obtain spatially resolved information about nanovoiding, TEM imaging was performed on lamellae extracted from samples C, H and A. This approach benefits from direct 2D visualisation of voids, so that image post-processing can be used to assess their shape, number density, interface characteristics and distribution . These results can be directly compared with the TEM analysis performed at the YPSZ–porcelain interface and can be combined and contrasted with the results of SANS to provide improved confidence in the evaluation of nanovoiding characteristics.
Material and methods
The three samples examined in this study were manufactured from a single batch of I14 Vitablocs ® Mark II for Cerec . Vitabloc ® restorations, although composed of similar ceramics to conventional metallic or ceramic porcelain veneers, are known to have a different manufacturing route, thermal characteristics and mechanical response . Despite these differences, the similarities in elemental composition between these porcelain types, in combination with the carefully controlled industrial sintering process used to manufacture Vitablocs ® , ensures that these components offer a comparable and microstructurally consistent base material. This high level of consistency between samples is crucial in order to identify the subtle nanoscale changes induced by the heat treatment and creep of dental porcelain.
The samples were machined using a diamond-coated grinding lathe and tensile specimen manufacturing procedure outlined by Lunt et al. . Multiple dog-bone samples were produced with the nominal gauge length of 6 mm and diameter of 3.4 mm. Following a multi-stage polishing routine, the gauge diameters and lengths of each specimen were measured using a micrometer screw gauge. The reference sample in the as-machined state (sample A) was selected as one of these specimens.
High temperature creep was induced in sample C using the 50 kN uniaxial servo-hydraulic Instron loading rig at beamline Engin-X at ISIS Spallation Source, UK and the approach previously outlined by Lunt et al. . The sample was placed into high temperature nickel alloy grips, a thermocouple was attached to the surface and then the sample was preloaded to around 0.2–0.7 MPa in a displacement control mode. Radiant heat lamps were then used to heat the specimen up to 750 °C at a rate of 10 °C min −1 . This holding temperature corresponds to the sintering conditions applied to dental veneering porcelains of similar composition during manufacture . A thermal dwell of 3 h was used to stabilise the system fully, ensuring that no further displacement could be observed in the loading rig output. The sample was then loaded to 100 MPa in order to match the residual stress previously observed at the YPSZ–porcelain interface of a veneer of similar thickness . A loading rate of 1 MPa s −1 was used to minimise the likelihood of brittle failure.
The secondary creep rate behaviour of this type of dental porcelain has previously been identified elsewhere as :
ε ˙ = A σ n e − Q / R T ,