Glass Matrix Composite

Abstract

Glass and glass–ceramic matrix composites represent a very particular type of composite materials. As in most ceramic matrix composites, the reinforcements are mainly intended to increase the resistance to crack propagation, that is, the most significant weakness of the matrices, due to several toughening mechanisms. Unlike polycrystalline ceramic matrices, however, glasses offer the distinctive feature of viscous flow, being readily deformed and flowed in their low viscosity state, with the possibility of incorporating secondary phases at moderately high temperatures. Fibrous reinforcements (mainly C and SiC fibers) are at the basis of materials with impressive properties, for example, strength above 1 GPa, fracture toughness up to 35 MPa m 0.5 , almost zero thermal expansion, and service temperature up to 1500°C, but still facing high costs, although known for many years. Recent applications, such as that of “optomechanical materials,” that is, transparent materials with improved crack resistance, may stimulate research investments and production. As an alternative, dispersion‐reinforced composites, first seen as lower value category, owing to the limited improvements in fracture toughness, are currently the candidates for extensive industrial production, since the low cost treatments applied may be combined with complex combinations of different functionalities.

Bayesian Hierarchical Modelling for Inferring Genetic Interactions in Yeast

“…As an alternative to the maximum likelihood approach that was presented by Addinall et al (2011) and the REM, we present a Bayesian hierarchical methodology where a priori uncertainty about each parameter value is described by probability distributions (Bernardo and Smith, 2007) and information about parameter distributions is shared across orf Δs and conditions. Plausible frequentist estimates from across 10 independent, unpublished QFA data sets, including a wide range of background mutations and treatments were summarized to establish and quantify our a priori uncertainty in model parameters.…”

“…With the Bayesian approach (Bernardo and Smith, 2007) that we adopt in this paper, we have more flexibility of model choice, allowing us to match model structure more closely to experimental design. Bayesian analysis allows us to use binary indicators to describe the evidence that each orf Δ interacts with the query mutation in terms of probability.…”