Descriptions of crack growth behaviors in glass–ZrO 2bilayers under thermal residual stresses

Abstract

Objectives

This study was intended to separate residual stresses arising from the mismatch in coefficients of thermal expansion between glass and zirconia (ZrO 2 ) from those stresses arising solely from the cooling process. Slow crack growth experimentes were undertaken to demonstrate how cracks grow in different residual stress fields.

Methods

Aluminosilicate glass discs were sintered onto ZrO 2 to form glass–ZrO 2 bilayers. Glass discs were allowed to bond to the ZrO 2 substrate during sintering or prevented from bonding by means of coating the ZrO 2 with a thin boron nitrade coating. Residual stress gradients on “bonded” and “unbonded” bilayers were assessed using birefringence measurements. Unbonded glass discs were further tested under biaxial flexure in dynamic fatigue conditions in order to evaluate the effect of residual stress on the slow crack growth behavior.

Results

When fast-ccoling was induced, residual tensile stresses on the glass increased significantly on the side toward the ZrO 2 substrate. By allowing the bond between glass and ZrO 2 , those tensile stresses observed in unbonded specimens are overwhelmed by the contraction mismatch stresses between the ZrO 2 substrate and the glassy overlayer. Specimens containing residual tensile stresses on the bending surface showed a time-dependent strength increase in relation to stress-free annealed samples in the dynamic biaxial bending test, with this effect being dependent on the magnitude of the residual tensile stress. The phenomenon observed is explained here on the basis of the water toughening effect, in which water diffuses into the glass promoting local swelling. An additional residual tensile stress at the crack tip adds an applied-stress-independent ( K res ) term to the total tip stress intensity factor ( K tip ), increasing the stress-enhanced diffusion and the shielding of the crack tip through swelling of the crack faces.

Significance

Residual stresses in the glass influence the crack growth behavior of veneered-ZrO 2 bilayered dental prostheses. The role of water in crack growth might be of higher complexity when residual stresses are present in the glass layer.

Introduction

The fast-cooling of glass plates is known to induce a parabolic residual stress profile, with both surfaces put in compression and the interior experiencing compensating residual tensile stresses . The compressive outer layers strengthen the glass and benefit structural applications of glass monoliths, but cracks reaching the inner region under tension cause the glass to shatter. Residual tensile stresses are therefore believed to be highly detrimental to the crack resistance and lifetime of glass-based structures.

In dentistry partially crystallized glasses (usually silicate glasses reinforced with feldspar, leucite or other crystal) are widely used in the monolithic form, but are mostly fused onto high-strength polycrystalline ceramics like aluminum oxide (Al 2 O 3 ) and zirconium dioxide (ZrO 2 ). In such bilayer constructs the framework material is deliberately set to contract more in order not to crack the fragile glass overlayer during the cooling phase. For the purpose of further strengthening the glass, bench-cooling processes have been inherited from the era of metal–glass systems in order to induce “tempering” to the piece by quickly removing it from the oven and exposing it to ambient air. As this practice was transferred to zirconia-based structures in the early 2000s, clinical studies sounded the alarm shortly thereafter. While the ZrO 2 component of bilayer prosthetic constructs has shown to survive clinical challenges satisfactorily, the fracture of the overlaying veneering material was shown to occur in unexpectedly earlier stages of clinical usage when compared to well-established conventional metal–glass systems . These fractures are usually limited to the glassy layer, which chip off, leaving some remaining material attached to the ZrO 2 substrate (see Fig. 1 ), but not rarely lead to full prosthetic replacement. Detailed descriptions of such shell-like fracture patterns occurring in vivo can be found in , and in for those reproduced in vitro .

Fig. 1
Chipping of the glassy veneer layer in two veneer-ZrO 2 dental prosthetic constructs in vivo . Detailed fractographic analyses of both cases are found in .

It turns out that thermomechanical processes of fused glass/ceramic bilayers are a bit more complex. The thickness ratio has been known from the early elastic deformation theories to determine whether the (coating) glass layer would be fully engulfed in a compressive state or develop tensile stress on the surface. Fast-cooling protocols, in turn, have shown to induce a much more severe temperature gradient within the glass layer in ZrO 2 -based than in metal-based systems . Compared to metals, ZrO 2 acts rather as a thermal insulator, supplying the glass layer with further heat from the interior, and detrimental tensile stresses may develop therein during cooling . Some evidence has been provided to support this, in that residual tensile stresses have been accounted for the lower lifetime of fast-cooled constructs under cyclic contact loading and peculiar fractographic patterns . By and large, how cooling stresses and contraction stresses develop in isolation and how they synergistically contribute to the global stress state of bilayer systems is still a poorly understood problem. The final stress state is usually assessed in a cooled, bonded condition, and masks their weighted contributions.

To address this issue we designed this study to isolate the stresses arising solely due to the cooling process from the contraction mismatch stresses in glass–ZrO 2 bilayers. For that, we created an unbonded condition between layers by isolating the ZrO 2 substrate using a boron nitride (BN) coating. This allowed free contraction of both layers while enabling thermal exchange between parts, once heat flow from the ZrO 2 substrate during cooling is thought to influence the solidification rate of the overlaying glass and stress build-up therein. We subjected the two opposite sides of the glass specimens to fracture toughness testing and dynamic biaxial bending. Dynamic fatigue testing in silica-based materials is usually performed in water to promote the environmentally-assisted subcritical growth of a crack, in that water molecules hydrolise Si O bonds at the crack tip under applied stress . The more time is given for this process to occur, the farther the crack grows before instability and strength drops with decreasing stress rates. From that, the crack velocity exponent in the subcritical stress range can be calculated, and the influence of residual stresses on crack growth can be determined.

Birefringence measurements revealed peculiar stress profiles for the unbonded models, where tensile stresses were found on the surface and on the interface sides of the glass discs sintered over BN-coated ZrO 2 . Finally, we show here that tensile residual stresses play the unexpected role of increasing the resistance to subcritical crack growth, theoretically by means of adding to the hydrostatic stress at the crack tip governing the stress-enhanced diffusion of water molecules into the glass. The crack faces swell and create a negative stress intensity factor term that shields the crack tip from the applied stress, a mechanism that has been recently proposed by . To the authors knowledge this is the first description of the so-called water toughening mechanism in a glass-based material presenting a global residual tensile stress field not induced by contact damage. This contribution highlights the instrumental role of residual tensile stress in glass-based structures in increasing the resistance to fracture under subcritical loads (low crack velocities), and its role in the rapid degradation of strength above the critical stress range for fracture.

Experimental procedure

Preparation of veneer specimens

Four hundred and twenty-five discs (with final dimensions: ∅ = 12.5 mm, 1.2 mm-thick) were produced using a commercial leucite-reinforced aluminosilicate glass (VM ® 9, Vita Zahnfabrik, Bad Säckingen, Germany; Shade ENL, Lot 19469) recommended for veneering ZrO 2 infrastructures for single-unit to long-span dental prostheses. This veneering material has approx. 95 wt% of glass phase of the system (mol%) 62SiO 2 + 14Al 2 O 3 + 8K 2 O + 5Na 2 O + 4B 2 O 3 + 2BaO + 1.5CaO (TiO 2 , CeO 2 , ZrO 2 , SnO 2 and Mg-, Fe-, P-oxides in concentrations <0.5 mol%) and approx. 5 wt% dispersed 1.4 μm-leucite crystals (K[AlSi 2 O 6 ]) after sintering. Other thermal, physical and mechanical properties of the glass and ZrO 2 materials are depicted in Table 1 . Glass discs were produced by condensing 0.450 g of powder mixed to 120 μl of ethanol in a cilindric teflon mold and immediately sintered under vacuum over a refractory tray (Vacumat 4000, Vita Zahnfabrik) with an initial drying temperature of 500 °C, heating rate of 80 °C/min, final temperature of 940 °C and holding time of 2 min at the final sintering temperature. The sintered discs were reduced to the final thickness by grinding both sides with a diamond wheel in a grinding machine (MPS 2 120, G&N, Germany) and subsequently polished (320 SiC Grit paper down to 4000 Grit) on the side to be posteriorly subjected to tension.

Table 1
Physical, mechanical and thermal properties of the glass and the ZrO 2 used in this study.
Material property (unit) Veneer ZrO 2
Elastic modulus (GPa) 66.5 200
Hardness (GPa) 7.4 13.2
Density ( ρ ) (g/cm 3 ) 2.44 6.0
Poisson’s ratio ( υ ) 0.21 0.32
T g (°C) 603 1170
CTE (α) (25–600 °C) (ppm K −1 ) 9.3 10.8
Conductivity (W/mm °C) 1.5 × 10 −3 2.0 × 10 −3
Specific heat (J K −1 kg −1 ) 840 450
Crystal Leucite 3 mol%Y 2 O 3 –ZrO 2
Crystalline phase (vol%) 4.6 99.9
Crystal/grain size (μm) 1.4 0.5

Unbonded condition: sintering veneer specimens over coated ZrO 2 substrates

Isostatically pressed green blocks of a commercial 3 mol% Y 2 O 3 partially-stabilized ZrO 2 polycrystalline (3Y-TZP) material indicated for dental infrastructure constructions (IPS e.max ZirCAD, Ivoclar Vivadent, Schaan, Liechtenstein) were sequentially cut into plates and sintered for 2 h at 1530 °C (final dimensions: 12.5 mm × 12.5 mm × 1.2 mm). The sintered zirconia plates were spray-coated with boron nitride (BN) (HeBoCoat ® 401E, Henze BNP GmbH, Kempten, Germany), a common material used for isolating glass and ceramic sheets, which has a high temperature stability in air and vacuum conditions, high thermal conductivity and low thermal expansion. The polished glass disk specimens were again fired (same protocol, 1 min holding time) over the coated zirconia plates and cooled from final sintering temperature following a Fast or a Slow cooling program. Fast-cooling consisted of the immediate opening of the oven’s door (Vacumat 4000) after the end of the firing schedule, exposing the specimens to contact with convective ambient air (∼23 °C). After the end of the firing schedule in the Slow-cooling program, the door of the oven opened only 10%, following a completely opening as soon as the temperature inside the oven reached 300 °C. Between the maximum sintering temperature and 500 °C, cooling rates of 45 °C/s and 0.5 °C/s were recorded with a K-type thermocouple for the Fast – and Slow-cooling protocols, respectively ( Fig. 2 ).

Fig. 2
Temperature versus time plot for the Fast and Slow cooling programs undertaken during the last firing.

The application of the BN layer was standardized and measured to be 10–15 μm thick using a white-light optical profilometer (CT100, Cyber Technologies, Ingolstadt, Germany). The BN coat prevented the bonding between the zirconia and the glass disk, while allowing for heat exchange between both parts. This way the glass discs were influenced by the thermal properties of the zirconia during cooling, and the effects of cooling were isolated from the effects of contraction mismatches that normally take place when both layers are allowed to adhere during sintering of the glass.

In order to evaluate the eventual build-up of cooling residual stresses on the surface side and on the interface side of the glass discs, the polished side of the specimens was directed opposite or toward the coated zirconia substrate, respectively, for each of the cooling protocols. Following this design, experimental groups were labeled Surface or Interface / Fast – or Slow-cooling (see illustration in Fig. 3 ). Because the BN-layer promoted some slight roughening on the previously polished surface of the glass discs, all specimens were sintered twice over the substrate so that the surface groups also received the same surface roughness alteration: during the first sintering, the polished side of the specimens from the Interface group faced up, and only during the second sintering the polished side faced the BN-layer. For the specimens from the Surface group, the opposite sequence was employed. Following this procedure, specimens from all groups received the same number of sintering firings and presented the same surface roughness. A control group (stress-free) was also produced from specimens sintered just like the ones described above, and by annealing them in a furnace for 10 h at 503 °C according with a protocol developed by Fischer et al. , ( T g – 100 °C) and cooled overnight inside the oven. This group was termed Annealed . After fabrication all specimens were sealed in vials containing silica gel until testing.

Fig. 3
Schematic illustration of bonded (a) and unbonded (b) glass–ZrO 2 bilayers. (c) The piston-on-three-balls biaxial test of the unbonded glass specimens. During testing the surface or the interface of the specimens were directed to the bending side.

The superscript notation * will hereafter designate Annealed samples, such as <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='Kc*’>KcKc*
K c *
, or n *; the absence of * will designate any Unannealed group.

Dynamic strength testing

In order to test the effect of the potential build-up of cooling stresses at the Surface and Interface sides of the disk-shaped glass specimens (and this way be able to state on the stress gradient within the veneer layer), their fracture behavior was evaluated in terms of quasi-static strength and dynamic strength testing. For the measurement of strength, 30 specimens were loaded until fracture in biaxial flexure according to the piston-on-3-balls set-up using a universal testing machine (Z 2.5, Zwick, Ulm, Germany) in water at 37 °C. A 6.5 × 10 1 MPa/s stress rate was chosen based on the pretesting of 5 specimens of each group to reach fracture within 1.5 s. The fracture stress ( σ f ) at the center of the discs was calculated according to :

<SPAN role=presentation tabIndex=0 id=MathJax-Element-2-Frame class=MathJax style="POSITION: relative" data-mathml='σf=−0.2387F(X−Y)w2′>σf=0.2387F(XY)w2σf=−0.2387F(X−Y)w2
σ f = − 0.2387 F ( X − Y ) w 2

where F is the load at fracture, w is the specimen thickness and X and Y are determined by:

<SPAN role=presentation tabIndex=0 id=MathJax-Element-3-Frame class=MathJax style="POSITION: relative" data-mathml='X=(1+υ)lnEC2+(1−υ)2EC2′>X=(1+υ)ln(EC)2+((1υ)2)(EC)2X=(1+υ)lnEC2+(1−υ)2EC2
X = ( 1 + υ ) ln E C 2 + ( 1 − υ ) 2 E C 2

and

<SPAN role=presentation tabIndex=0 id=MathJax-Element-4-Frame class=MathJax style="POSITION: relative" data-mathml='Y=(1+υ)1+lnAC2+(1−υ)AC2′>Y=(1+υ)(1+ln(AC)2)+(1υ)(AC)2Y=(1+υ)1+lnAC2+(1−υ)AC2
Y = ( 1 + υ ) 1 + ln A C 2 + ( 1 − υ ) A C 2

where υ is the Poisson’s ratio (taken to be 0.21, measured in ), A is the radius of the support circle (4.5 mm), E is the radius of the tip of the piston (0.6 mm), and C is the radius of the specimen (6.25 mm).

In order to assess the effect of the residual stress on the crack growth behavior, both sides ( Surface or Interface ) were subjected to dynamic testing and compared to the stress-free Annealed group. To determine the slow crack growth parameters and crack velocity during the subcritical growth regime using dynamic bending tests, the time dependence of strength is calculated from the strength values obtained at different stress rates ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-5-Frame class=MathJax style="POSITION: relative" data-mathml='σ˙=dσ/dt’>σ˙=dσ/dtσ˙=dσ/dt
σ ˙ = d σ / d t
). Thus, 10 additional specimens of each group were tested at stress rates of 6.5 × 10 −1 , 6.5 × 10 −2 and 6.5 × 10 −3 MPa/s also in water maintained at 37 °C throughout the test. For the inert strength measurement 10 discs were stressed at 6.5 × 10 1 MPa/s rate in silicon oil (before oil immersion the specimens were dried in an oven at 150 °C for 2 h). For every trial series, the Weibull shape parameter m and characteristic strength parameter σ 0 were determined in a ln σ − ln(1/1 − P F ) plot. The slope of the regression gives the m value from ln(ln 1/1 − P F ) = m ln σ m ln σ 0 , and σ 0 represents the strength at a failure probability P F of 0.632, or 63.2% when ln(ln 1/(1 − P F )) = 0. Weibull modulus m , σ 0 and their respective 90% confidence intervals were corrected by a factor corresponding to the number of specimens, according to the European standard EN 843-5 .

Fatigue parameters n and A can be evaluated by the application of the relation between fracture strength and stress rate:

<SPAN role=presentation tabIndex=0 id=MathJax-Element-6-Frame class=MathJax style="POSITION: relative" data-mathml='σfn+1=Bσin−2σ˙(n+1)’>σn+1f=Bσn2iσ˙(n+1)σfn+1=Bσin−2σ˙(n+1)
σ f n + 1 = B σ i n − 2 σ ˙ ( n + 1 )

where σ f denotes the flexural strength in MPa, <SPAN role=presentation tabIndex=0 id=MathJax-Element-7-Frame class=MathJax style="POSITION: relative" data-mathml='σ˙’>σ˙σ˙
σ ˙
is the stress rate, n is the slow crack growth susceptibility coefficient, σ i is the inert strength and the quantity <SPAN role=presentation tabIndex=0 id=MathJax-Element-8-Frame class=MathJax style="POSITION: relative" data-mathml='Bσin−2′>Bσn2iBσin−2
B σ i n − 2
is a scaling parameter. The logarithmic form of Eq. (4) was plotted, from where the slope of the curve of <SPAN role=presentation tabIndex=0 id=MathJax-Element-9-Frame class=MathJax style="POSITION: relative" data-mathml='log(σf)−log(σ˙)’>log(σf)log(σ˙)log(σf)−log(σ˙)
log ( σ f ) − log ( σ ˙ )
rendered n (all values were used for the least-square regression curve), and <SPAN role=presentation tabIndex=0 id=MathJax-Element-10-Frame class=MathJax style="POSITION: relative" data-mathml='Bσin−2′>Bσn2iBσin−2
B σ i n − 2
was determined by the location of the straight line. At the moment the applied stress intensity factor K reaches a critical value K = K c , the crack becomes unstable and the parameter A is taken as the crack velocity v , according to the Paris relation v = A ( K/K c ) n , and is obtained from :

<SPAN role=presentation tabIndex=0 id=MathJax-Element-11-Frame class=MathJax style="POSITION: relative" data-mathml='A=2Kcn−2BY2(n−2)’>A=2Kn2cBY2(n2)A=2Kcn−2BY2(n−2)
A = 2 K c n − 2 B Y 2 ( n − 2 )

where Y is a geometrical factor determined using fractography and K c is the fracture toughness determined by the indentation flexural strength test (see Section 2.4 ).

Fracture toughness ( K c ) determination

To determine the effect of thermal residual stresses on the local (surface) fracture toughness ( K c ) of the veneer specimens, two different methods were used: fractography ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-12-Frame class=MathJax style="POSITION: relative" data-mathml='KcF’>KFcKcF
K c F
) and indentation strength ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-13-Frame class=MathJax style="POSITION: relative" data-mathml='KcS’>KScKcS
K c S
). For the fractographic method, 7–10 specimens that were tested in water at the stress rate of 6.5 × 10 1 MPa/s and fractured in 2 parts were observed in a confocal laser scanning microscope (CLSM) (TCS SL, Leica, Heidelberg, Germany) in reflectance mode at high-magnification (200×). Specimens tested in oil usually fractured in more than 2 pieces, making it difficult for fractographic evaluation. The origin of the critical crack was located and the depth a , and the half width b of the semi-elliptical flaw were measured on the fractured surfaces to calculate the critical flaw size c , as c = ( ab ) 1/2 . This measure was then used in the Griffith/Irwin relation that relates the strength, critical flaw size and the critical stress intensity factor K c , as <SPAN role=presentation tabIndex=0 id=MathJax-Element-14-Frame class=MathJax style="POSITION: relative" data-mathml='KcF=Yσfc1/2′>KFc=Yσfc1/2KcF=Yσfc1/2
K c F = Y σ f c 1 / 2
. The geometrical factor Y was determined based on the crack dimensions using the Newman and Raju equations for semi-elliptical surface cracks ( Y varied between 1.34 and 1.36; these values were used throughout this study).

Using 5 intact specimens for each group, K c was additionally measured using a Vickers indenter (9.8 N, 15 s, in air) to produce median-radial cracks systems. One indentation was performed at the center of the polished surface for each specimen. The specimens were covered with silicon oil and tested under biaxial flexure within 10 min. Eq. (1) was used to calculate the strength σ f , which was used in the solution supplied by Chantikul et al. to determine <SPAN role=presentation tabIndex=0 id=MathJax-Element-15-Frame class=MathJax style="POSITION: relative" data-mathml='KcS’>KScKcS
K c S
.

Values were analyzed statistically using one-way ANOVA and multiple Student’s t -tests at a significance level α = 0.05.

Assessment of stress gradients in bonded and unbonded bilayers

Residual stresses in translucent materials can be made visible using the principle of photoelasticity. While stress-free glasses and ceramics are optically isotropic, any applied mechanical stress or residual stresses in the material cause the two decomposed waves from a polarized light beam to follow parallel orientations to the principal stress vectors ( σ 1 , σ 2 ). This is accompanied by a phase delay between these waves, which would otherwise travel in the same velocity. Different wave velocities in birefringent materials result in two indexes of refraction ( n 1 , n 2 ). The retardation ( δ ) between waves can be expressed by δ = Δ n · t , where t is the thickness of the material. According to the Stress-Optic Law , the difference between refraction indexes is proportional to the difference between principal stresses, that is, Δ n = C el ( σ 1 σ 2 ). Using the coefficient of photoelasticity C el , the magnitude of the residual stresses across the thickness of the glass layer can be obtained using the relation σ res = δ / t · C el .

In order to directly probe the residual stress gradients within the glass layer resulting from the Fast – and Slow-cooling protocols, birefringence measurements were conducted on unbonded specimens produced as described above. Additional bonded bilayers were fabricated by letting the glass fuse to the ZrO 2 substrate (no BN layer was applied onto the ZrO 2 , see Fig. 3 a). Birefringence measurements were performed using an automatic imaging measuring system (StrainMatic M4/120, ilis GmbH, Germany), consisting of an automatized polarimeter with xy -resolution of 11 μm connected to a computer where the data was analysed by a software (StrainAnalyser v.2015.1). Details of the technique can be found in . For that, 1 mm-thick cross-sections of the bonded bilayers and unbonded glass discs were prepared for transillumination ( xy -direction). Intact unbonded discs were also transilluminated from above ( z -direction), dispensing destructive preparation procedures and preserving the original stress state in the discs. Residual stresses were calculated using the value for C el previously measured for this specific glass in to 10.4 T Pa −1 .

Experimental procedure

Preparation of veneer specimens

Four hundred and twenty-five discs (with final dimensions: ∅ = 12.5 mm, 1.2 mm-thick) were produced using a commercial leucite-reinforced aluminosilicate glass (VM ® 9, Vita Zahnfabrik, Bad Säckingen, Germany; Shade ENL, Lot 19469) recommended for veneering ZrO 2 infrastructures for single-unit to long-span dental prostheses. This veneering material has approx. 95 wt% of glass phase of the system (mol%) 62SiO 2 + 14Al 2 O 3 + 8K 2 O + 5Na 2 O + 4B 2 O 3 + 2BaO + 1.5CaO (TiO 2 , CeO 2 , ZrO 2 , SnO 2 and Mg-, Fe-, P-oxides in concentrations <0.5 mol%) and approx. 5 wt% dispersed 1.4 μm-leucite crystals (K[AlSi 2 O 6 ]) after sintering. Other thermal, physical and mechanical properties of the glass and ZrO 2 materials are depicted in Table 1 . Glass discs were produced by condensing 0.450 g of powder mixed to 120 μl of ethanol in a cilindric teflon mold and immediately sintered under vacuum over a refractory tray (Vacumat 4000, Vita Zahnfabrik) with an initial drying temperature of 500 °C, heating rate of 80 °C/min, final temperature of 940 °C and holding time of 2 min at the final sintering temperature. The sintered discs were reduced to the final thickness by grinding both sides with a diamond wheel in a grinding machine (MPS 2 120, G&N, Germany) and subsequently polished (320 SiC Grit paper down to 4000 Grit) on the side to be posteriorly subjected to tension.

Table 1
Physical, mechanical and thermal properties of the glass and the ZrO 2 used in this study.
Material property (unit) Veneer ZrO 2
Elastic modulus (GPa) 66.5 200
Hardness (GPa) 7.4 13.2
Density ( ρ ) (g/cm 3 ) 2.44 6.0
Poisson’s ratio ( υ ) 0.21 0.32
T g (°C) 603 1170
CTE (α) (25–600 °C) (ppm K −1 ) 9.3 10.8
Conductivity (W/mm °C) 1.5 × 10 −3 2.0 × 10 −3
Specific heat (J K −1 kg −1 ) 840 450
Crystal Leucite 3 mol%Y 2 O 3 –ZrO 2
Crystalline phase (vol%) 4.6 99.9
Crystal/grain size (μm) 1.4 0.5

Unbonded condition: sintering veneer specimens over coated ZrO 2 substrates

Isostatically pressed green blocks of a commercial 3 mol% Y 2 O 3 partially-stabilized ZrO 2 polycrystalline (3Y-TZP) material indicated for dental infrastructure constructions (IPS e.max ZirCAD, Ivoclar Vivadent, Schaan, Liechtenstein) were sequentially cut into plates and sintered for 2 h at 1530 °C (final dimensions: 12.5 mm × 12.5 mm × 1.2 mm). The sintered zirconia plates were spray-coated with boron nitride (BN) (HeBoCoat ® 401E, Henze BNP GmbH, Kempten, Germany), a common material used for isolating glass and ceramic sheets, which has a high temperature stability in air and vacuum conditions, high thermal conductivity and low thermal expansion. The polished glass disk specimens were again fired (same protocol, 1 min holding time) over the coated zirconia plates and cooled from final sintering temperature following a Fast or a Slow cooling program. Fast-cooling consisted of the immediate opening of the oven’s door (Vacumat 4000) after the end of the firing schedule, exposing the specimens to contact with convective ambient air (∼23 °C). After the end of the firing schedule in the Slow-cooling program, the door of the oven opened only 10%, following a completely opening as soon as the temperature inside the oven reached 300 °C. Between the maximum sintering temperature and 500 °C, cooling rates of 45 °C/s and 0.5 °C/s were recorded with a K-type thermocouple for the Fast – and Slow-cooling protocols, respectively ( Fig. 2 ).

Fig. 2
Temperature versus time plot for the Fast and Slow cooling programs undertaken during the last firing.

The application of the BN layer was standardized and measured to be 10–15 μm thick using a white-light optical profilometer (CT100, Cyber Technologies, Ingolstadt, Germany). The BN coat prevented the bonding between the zirconia and the glass disk, while allowing for heat exchange between both parts. This way the glass discs were influenced by the thermal properties of the zirconia during cooling, and the effects of cooling were isolated from the effects of contraction mismatches that normally take place when both layers are allowed to adhere during sintering of the glass.

In order to evaluate the eventual build-up of cooling residual stresses on the surface side and on the interface side of the glass discs, the polished side of the specimens was directed opposite or toward the coated zirconia substrate, respectively, for each of the cooling protocols. Following this design, experimental groups were labeled Surface or Interface / Fast – or Slow-cooling (see illustration in Fig. 3 ). Because the BN-layer promoted some slight roughening on the previously polished surface of the glass discs, all specimens were sintered twice over the substrate so that the surface groups also received the same surface roughness alteration: during the first sintering, the polished side of the specimens from the Interface group faced up, and only during the second sintering the polished side faced the BN-layer. For the specimens from the Surface group, the opposite sequence was employed. Following this procedure, specimens from all groups received the same number of sintering firings and presented the same surface roughness. A control group (stress-free) was also produced from specimens sintered just like the ones described above, and by annealing them in a furnace for 10 h at 503 °C according with a protocol developed by Fischer et al. , ( T g – 100 °C) and cooled overnight inside the oven. This group was termed Annealed . After fabrication all specimens were sealed in vials containing silica gel until testing.

Fig. 3
Schematic illustration of bonded (a) and unbonded (b) glass–ZrO 2 bilayers. (c) The piston-on-three-balls biaxial test of the unbonded glass specimens. During testing the surface or the interface of the specimens were directed to the bending side.

The superscript notation * will hereafter designate Annealed samples, such as K*c
K c *
, or n *; the absence of * will designate any Unannealed group.

Dynamic strength testing

In order to test the effect of the potential build-up of cooling stresses at the Surface and Interface sides of the disk-shaped glass specimens (and this way be able to state on the stress gradient within the veneer layer), their fracture behavior was evaluated in terms of quasi-static strength and dynamic strength testing. For the measurement of strength, 30 specimens were loaded until fracture in biaxial flexure according to the piston-on-3-balls set-up using a universal testing machine (Z 2.5, Zwick, Ulm, Germany) in water at 37 °C. A 6.5 × 10 1 MPa/s stress rate was chosen based on the pretesting of 5 specimens of each group to reach fracture within 1.5 s. The fracture stress ( σ f ) at the center of the discs was calculated according to :

σf=0.2387F(XY)w2
σ f = − 0.2387 F ( X − Y ) w 2

where F is the load at fracture, w is the specimen thickness and X and Y are determined by:

X=(1+υ)ln(EC)2+((1υ)2)(EC)2
X = ( 1 + υ ) ln E C 2 + ( 1 − υ ) 2 E C 2

and

Y=(1+υ)(1+ln(AC)2)+(1υ)(AC)2
Y = ( 1 + υ ) 1 + ln A C 2 + ( 1 − υ ) A C 2
Only gold members can continue reading. Log In or Register to continue

Nov 23, 2017 | Posted by in Dental Materials | Comments Off on Descriptions of crack growth behaviors in glass–ZrO 2bilayers under thermal residual stresses
Premium Wordpress Themes by UFO Themes