Fracture initiation in silicate glasses via a universal shear localization mechanism

TL;DR

This study reevaluates fracture initiation in silicate glasses, arguing that localized shear-flow and shear-band formation—not densification alone—primarily control indentation cracking. By systematically varying composition in Ca/Mg and B-doped silicates and combining indentation tests with cross-sectional roughness analysis and molecular dynamics simulations, the authors show that higher boron content and stronger network modifiers suppress shear localization and markedly increase crack resistance. MD results corroborate that boron lowers yield softening, promoting more homogeneous deformation and reducing localization propensity. Collectively, the work suggests a universal fracture-initiation paradigm in amorphous materials where the morphology and stability of shear bands dictate crack initiation, with practical implications for designing tougher glasses by diffusing localization.

Abstract

Shear bands lie at the root of fracture initiation in bulk metallic glasses and amorphous polymers. For silicate glasses, in contrast, studies have largely emphasized permanent volumetric strain, commonly referred to as densification. Here we systematically investigate indentation-induced fracture in two distinct families of aluminoborosilicate glasses. The results demonstrate that plastic shear flow plays a decisive role in governing fracture initiation. In addition, molecular dynamics simulations reveal a pronounced composition dependence of softening associated with plastic shear flow, closely mirroring the experimentally observed propensity for strain localization. We conclude that silicate glasses conform to a universal pattern of rupture initiation governed by localization of shear-deformation, aligning with a broad range of amorphous materials, including bulk metallic glasses and glassy polymers.

Figures (6)

• Figure 1: (a-c) single alkaline earth (Ca or Mg) ABS glasses with varying B$_2$O$_3$ content: a) Young's Modulus $E$ and Poisson's ratio $\nu$; b) micro-indentation Vickers hardness H$_V$; c) densification (CABS only) at 25 GPa (Multi Anvil Cell) Gomes25 - (d-e) mixed alkaline earth ABS glasses (5% or 15% B$_2$O$_3$) as a function of Mg vs Ca content: d) Young's Modulus $E$ and Poisson ratio $\nu$; e) micro-indentation Vickers hardness H$_V$ - (f) all glass compositions: RID as a function of Poisson ratio including some (LS, SLS, ABS) glasses from Kato et alKato11 and from Barlet et al.Barlet15. Error bars are not shown if smaller than the marker size.
• Figure 2: (a-b) crack resistance $CR$ of (a) single alkaline earth network modifier ABS glasses and (b) mixed alkaline earth network modifiers ABS glasses. If no error bar is shown, it is less than the markers size - (c) crack resistance ($CR$) as a function of recovered indentation depth ($RID$) - (d) recovered indentation depth for different silicate glasses compositions including some (LS, SLS, ABS) from Kato et al kato2011load. Arrows show how crack resistance increases with compositions.
• Figure 3: Sum up of all cross-sections through 1 kgf Vickers indents
• Figure 4: Cross-sections through 1 kgf Vickers indents in (a) CABS2, (b) CMABS4, (c) CABS3 and (d) CMABS9.
• Figure 5: (a) Scanning Electron image of 1kgf Vickers indentation cross-section in soda-lime silicate glass and (b) magnified image on shear bands from the cross-section. The sample has been tilted to 35° to observe shear bands from the bottom. (c-d) Relationship between crack resistance ($CR$) and roughness ($R_a$) for Vickers indentation cross-sections at 1 kgf (c) and 500 gf (d) including some commercial compositions (black markers) from Kato et al. Kato10. Vertical dashed line is determined as the bulk roughness. (e) Crack resistance as a function of the average spacing RS$_m$ between shear bands from Vickers indentation cross-sections at 1kgf for all glass compositions.