As water diffuses into silica glass, it reacts with the silica network to generate hydroxyl groups that cause a volume swelling of the glass. We discuss this effect quantitatively for cases in which hydroxyl groups were produced either by gamma-irradiation of hydrogen-saturated silica glass or by high-temperature diffusion of water into silica glass. From volume swelling measurements, the partial molar volume of hydroxyl groups for both treatments is determined to be 7.5 cm(3)/mol. Since two moles of SiOH are generated for each mole of water reacting with the glass, the reaction volume, Delta(V) over bar , is 15.0 cm(3)/mol.

Degradation of glass under zero applied load in the presence of humidity at ambient temperature is of great interest to the container and fiber glass industries. The phenomenon is well documented for fused silica used in optical fibers, but has not been studied in detail for multi-component glasses. In this work notches of varying length (500-1500 nm) were placed with a focused ion beam into two types of multi-component glass fibers, E-glass (48 mu m diameter) and soda-lime-silicate (35 mu m diameter). Notched specimens were exposed to dry and humid conditions for up to 32 days. Transmission electron microscopy revealed the presence of extensive reaction products within the root of the notch, even after only 1 day of aging in the nominally dry environment for the soda-lime-silicate glass. Surprisingly, the extensive reactions have no measurable effect on the fiber strength. The uniaxial tensile strength of the notched glass fibers, measured with the fracture surface mirror radius method, does not follow a classic fracture mechanics prediction, implying that the notches are not classic Griffith flaws. Fracture mechanics is applied to show that sharpness at the notch base may be important, especially when subcritical crack growth is present during the strength measurement.

Crack opening displacements were evaluated on semi-elliptical indentation cracks in lead-free (1-x)(Na1/2Bi1/2)TiO3-xBaTiO(3) piezoceramics and a commercially available PZT ceramic. The observed crack-tip toughness of NBT-xBT was found to be substantially higher than for PZT. Two evaluations for the crack opening displacement were demonstrated and contrasted: A more elaborate three-term-approximation and a pragmatic utilization of the Irwin parabola.

When water diffuses into silica glass, a chemical reaction between the water and the glass damages the ring structure of the glass, causing a reduction in the Young's modulus, a lowering of the intrinsic strength and a reduction in the crack-growth resistance of the glass. In the absence of swelling of the glass within the water penetrated zone, the damage caused by hydroxyl group generation will have no effect on the strength of the glass fibers, provided the water-affected surface zone is much thinner than the bulk material. In contrast, swelling stresses within the water penetrated zone at the fiber glass surface will increase the strength of the glass fibers.

Here we study the diffusive transfer of water into silica glass in the presence of externally applied stresses, and stresses caused by water induced swelling of the glass. By considering the simultaneous action of water penetration into the surface of silica glass and the development of swelling stresses in the water-penetrated zone, several experimental findings previously published in the literature can be interpreted quantitatively. These include an apparent decrease of the diffusivity with time, an increase of water solubility in the surface region under compressive loading, the opposite effect under tensile loading and a reversal of these two effects deeper within the glass. These expectations are fully met in published experiments carried out to date.

Cracks terminating at free surfaces are affected by local stresses in the surface region. Under residual compression the crack front must retard compared with the crack contour in the absence of stresses. This effect can be used for an identification of residual stresses at glass surfaces. For an illustration of the procedure, Vickers indentation tests in soda-lime glass are considered. Specimens treated by ion exchange and chemically toughening showed reduced terminating angles compared with untreated glass.

In this paper, we consider the diffusion of water vapor into silica glass at relatively low temperatures, 25 degrees C to approximate to 500 degrees C. Extensive studies of such diffusion by others have shown that water diffusion from inert gases behaves differently than diffusion from liquid water. In liquid water, the concentration of water at the surface quickly achieves saturation value. In gas, the concentration of water at the surface does not achieve a constant value, but builds up with time eventually reaching a steady state value. Here, we show that published data can be explained by assuming the existence of a surface barrier to the diffusion of water into the silica glass. We develop equations that quantify the diffusion rate through the barrier; the diffusion results are consistent with published data. We suggest that barrier formation is a consequence of the structure of the silica glass at the free surface and the way that water reacts at the surface in contrast to the way it reacts in the bulk glass. This description gives a quantitative explanation of the development of a diffusion barrier and provides a contrast between the way liquid water and water vapor behave at a free surface.

From the work of Le Chatelier [1884], it is well known that chemical reactions that exhibit a change in volume are sensitive to the ambient pressure of the reaction. Increasing the pressure will alter the ratio of reaction products to reactants. If the change in volume is constrained to occur at a surface, then such reactions can result in residual stresses that affect the strength of the solid. These effects are applicable to silica glass, which increases in volume when reacting with water. In this paper, we discuss the possibility of using this effect to strengthen silica glass. Using a modification of Le Chatelier's theory to handle applied stresses, we show that water penetration into the surface of silica glass can yield sufficient residual stress to increase the strength of silica glass into the GPa range. Applying these ideas to recent data published by Lezzi et al., we are able to attribute the strengthening they observe to a water/silica reaction under an applied tensile stress.

Strength, toughness, microstructure, and atomic adsorption arrangement in silicon nitrides with MgO and RE2O3 additions (RE = La, Gd, Y, Lu) were examined. Mechanical properties were high for La, Gd, and equal La–Lu additions, but surprisingly were progressively lower for Y‐ and Lu‐doped samples. The lower strength and toughness were associated with fewer visible crack deflections and grain bridges. Detailed microstructural analysis of the Lu‐doped material revealed a complex intergranular nanostructure with variable Lu content and Si3N4 nanocrystals. Furthermore, the Lu‐rich areas showed an extra Lu‐adsorption site on the Si3N4 prismatic planes not previously observed in other studies. This inhomogeneous structure was attributed to grain growth impingement and higher viscosity of the Lu‐doped oxynitride glass that slows homogenization. The Y‐doped material with nearly identical glass viscosity demonstrates intermediate behavior. Finally, substituting half of the Lu2O3 with La2O3 resulted in a homogenous intergranular structure, attributed to a lower viscosity of the oxynitride glass phase, and high mechanical properties. Overall, care must be taken when adapting Si3N4 processing parameters for the smaller ionic radius rare earth dopants such as Lu and Y.

There is a need for methods that can help predict and avoid fatigue failures of silicon nitride ceramic components. The fatigue threshold R‐curve has been proposed as potential solution to this problem. In this study, the fatigue threshold R‐curve for small, semielliptical surface cracks was calculated for a silicon nitride ceramic using the published bridging stress distribution developed from fatigue threshold tests on macroscopic crack specimens. To test the accuracy of the endurance strengths predicted using the fatigue threshold R‐curve, fatigue tests were conducted using four‐point bend beams of silicon nitride containing semielliptical surface cracks introduced by Knoop indentation. The effectiveness of the methodology was verified; indeed, 77% of the beams tested at stress levels above the predicted endurance strength failed within 107 cycles and 0% of the beams tested below the predicted endurance strength failed within 107 cycles. Furthermore, using the bridging stress distribution, which is thought to be a material property, the need for prohibitively difficult fatigue threshold experiments on small surface cracks is avoided. Accordingly, this methodology is potentially quite practical for use in the engineering design of ceramic mechanical components.

By measuring the curvature of thin disks of vitreous silica that have been penetrated by water from one side only, we determined the volume expansion of the silica and the effect of this volume expansion on its strength. We found that the water-strengthening process depended on crack-size, temperature, and the amount of swelling of the silica. We also evaluated the diffusivity of water in vitreous silica, using the swelling stresses as the diffusion metric. Diffusivity values, so obtained, are close to the accepted values for the diffusion of water in vitreous silica, as is the activation energy for the diffusion process. Our data suggest that swelling and the consequent bending of the disks is caused by silanol group formation in the silica structure; molecular water plays little role in the swelling process.

When Griffith presented his famous theory of crack stability in elastic materials in the early twentieth century, he was unable to provide much detail on the structure of cracks at the nanometer level of resolution. Now, almost 100 years later, techniques such as transmission electron microscopy, atomic force microscope, nuclear reaction analysis, and nuclear reflection are available to achieve this level of resolution. Here, we review the kind of data obtained using these techniques and the implications of the data vis‐á‐vis cracks in silicate glasses. Measurements by atomic force microscopy provide information on the size of the nonlinear zone at crack tips in glass, on environmental conditions at crack tips, and on the possibility of cavity formation as a mechanism of crack growth. Examination by nuclear reaction analysis and neutron reflection of fresh fracture surfaces formed in water has yielded information on water penetration through the glass surrounding the crack tip, to a resolution of 3–5 nm. Improvement of measurement techniques in the coming years should enable us to study crack tips in glasses to even higher levels of resolution and to answer more detailed questions concerning the level of stress and the size of the nonlinear zone at the crack tip.

Fatigue failure is a concern when high‐strength, high‐toughness silicon nitride ceramics are used in mechanical components and the growth of natural flaws will determine the usable upper bound strength. In this study a fracture resistance curve (R‐curve) model is incorporated into an established method for deducing natural flaw growth rates from a combination of strength and fatigue life data for smooth specimens. Experimental data for a commercial silicon nitride, SL200, were examined. When compared with results deduced using a constant fracture toughness model, the new method gives more physically realistic growth rate results. Specifically, by incorporating the R‐curve the deduced fatigue threshold is equal to the reported intrinsic toughness for crack propagation of 2.2 MPa√m, whereas the constant fracture toughness model gives a physically unrealistic threshold value. Furthermore, much better agreement is achieved with the growth rates measured using macroscopic compact‐tension specimens. Overall, it is concluded that the R‐curve effect should not be ignored when deducing the fatigue crack growth rates of natural flaws in high‐toughness silicon nitride ceramics.

Ceramic rolls for wire hot rolling at multi-line rolling mills may fail bycontact overloading. The present paper deals with a refinement of first publications onthis topic. In the first part the relations for stress intensity factor computation via theweight function method are compiled. Then it is shown whether the mixed-mode stressintensity factors of the curved cracks can be applied for the prediction of crack path viathe condition of local symmetry.

Silicon nitride is often used, when high fracture toughness and strength is needed. For a safe and economic structural design with this material, a prediction of its resistance against thermal and mechanical loads is important. The finite element method together with a continuum damage mechanics model allows for such calculations. The parameters of the suggested model have been adjusted to three-dimensional micromechanical finite element simulations, which include models for the microstructure, the thermoelasticity and the fracture. The material model is used for four-point bend test simulations. The results are compared to recent experiments.

Many biological materials are hierarchically structured, with highly anisotropic structures and properties on several length scales. To characterize the mechanical properties of such materials, detailed testing methods are required that allow precise and site-specific measurements on several length scales. We propose a fracture toughness measurement technique based on notched focused ion beam prepared cantilevers of lower and medium micron size scales. Using this approach, site-specific fracture toughness values in dental enamel were determined. The usefulness and challenges of the method are discussed. (C) 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.