Crack initiation and propagation mechanisms in windshield glass

Details

Supervisors

Pardoen, Thomas

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil mécanicien, à finalité spécialisée

Abstract

The increasing integration of technology in car windshields has impacted their price over the years, making use of fracture-resistant glass more appealing. One proposed solution has been to use borosilicate glasses, considered to have greater resistance to cracking, as well as better circularity of impacts. However, the very high glass transition temperature of borosilicate results in higher costs and makes it impractical for large-scale manufacturing. In this context, three different compositions of glasses are analyzed in this study: pure silica (quartz), soda-lime-silica (SLS) and borosilicate. The aim of this thesis is to explore the deformation processes of glass during an impact, as well as its crack initiation and propagation behavior; understanding how its composition and mechanical properties could be used to achieve improved crack control. Initially, we investigate the mechanical properties of glasses, seeking to identify factors influencing their resistance to crack initiation, often resulting from impacts in the real world. This involves indentation, with a particular focus on deformation processes, such as densification and pile-up. Links with composition-structure using the concept of free volume or atomic packing density are then discussed. Moreover, glasses are classified as normal or anomalous depending on their crack pattern. The link between driving forces and material properties is explored. Mechanisms increasing the apparent fracture toughness, such as crack deflection, bridging, branching and surface treatments are also reviewed. Finally, we explore resistance to crack extension of a pre-existing impact in glass. To do so, coaxial double ring tests are made on both bare and pre-indented samples. A very strong linear correlation between the atomic packing density (Cg) and the Poisson's ratio (v) is found. The low v values (hence Cg) of borosilicate and quartz explain their anomalous behavior, which is easily observed in the form of more circular cracking patterns than SLS (i.e. ring-cone cracks). We obtain a very wide range of cracking resistance (CR) values, going from 0.71 to 2.08N from SLS to quartz; mainly explained by Cg decreasing from 48.87 to 44.72%. The very particular atomic self-adaptivity of the boron units are viewed to significantly increase CR of borosilicate (+78.4% w.r.t. quartz). Furthermore, brittleness (B) and CR are observed to be linked. SLS shows the lowest B of 6.51 um^-0.5 w.r.t. borosilicate and quartz (3.99 and 4.15 um^-0.5). These values are linked to toughness, which increases from 0.86 to 1.43 and 1.74 MPam^0.5 with increasing network connectivity (cn). Finally, the surface condition has been shown to be of crucial importance. Roughly polished quartz samples show significantly lower flexural resistance than classical float SLS (-62%) despite their advantageous composition.