Polycrystallinity Enhances Stress Buildup around Ice

Damage caused by freezing wet, porous materials is a widespread problem but is hard to predict or control. Here, we show that polycrystallinity significantly speeds up the stress buildup process that underpins this damage. Unfrozen water in grain-boundary grooves feeds ice growth at temperatures below the freezing temperature, leading to fast stress buildup. These stresses can build up to levels that can easily break many brittle materials. The dynamics of the process are very variable, which we ascribe to local differences in ice-grain orientation and to the surprising mobility of many grooves—which further accelerates stress buildup. Our Letter will help understand how freezing damage occurs and in developing accurate models and effective damage-mitigation strategies.

Figure 1

(a) Schematic of polycrystalline ice in the experimental cell. (b) An ice-water interface imaged through crossed polarizers, highlighting individual grains with different crystal orientations. (c) Schematic of the cryosuction process into a premelted layer. (d) Grain boundaries (thin lines) in a bright-field micrograph of ice in the cell. (e) Stresses below the ice in (d). The image is taken 10 min after the start of the experiment. Substrate stiffness, 265 kPa; substrate thickness, $130\mathrm{\mu }\mathrm{m}$; ice thickness, $500\mathrm{\mu }\mathrm{m}$; temperature gradient, $0.4\mathrm{K}/\mathrm{mm}$. Crystal symbols are illustrative.

Figure 2

(a),(b) Bright-field images of ordinary and giant grooves. (c),(d) Substrate indentations caused by ice growth. (e),(f) Cross sections through the data at the lines indicated in (c),(d), with corresponding groove schematics. Giant grooves have basal facets as their walls, while ordinary grooves are unfaceted. Images are 30 min after the experiment starts. Substrate stiffness, 38 kPa; substrate thickness, $100\mathrm{\mu }\mathrm{m}$; ice thickness, $200\mathrm{\mu }\mathrm{m}$; temperature gradient, $0.1\mathrm{K}/\mathrm{mm}$. Substrate stiffness was chosen so as to easily visualize stresses [9].

Figure 3

Evolution of indentations under (continuous curves) and parallel to (dashed curves) two grain-boundary grooves [(a) and (b), respectively] under identical experimental conditions. Insets, top left: two images through crossed polarizers, showing the grain boundaries. ${\theta }_c$ is the angle between the ice-crystal $c$ axis and the substrate, estimated from the birefringence color [9]. Inset in (b): evolution of the maximum indentation. The dashed line shows a power law as a guide to the eye. Substrate stiffness, 280 kPa; substrate thickness, $100\mathrm{\mu }\mathrm{m}$; ice thickness, $500\mathrm{\mu }\mathrm{m}$; temperature gradient, $0.4\mathrm{K}/\mathrm{mm}$.

Figure 4

(a)–(c) Substrate indentations below ice with a stationary (bottom) and a mobile (top) grain boundary moving in a stop-start fashion (Supplemental Video 2 [40]). Dashed curves show the current grain boundary locations. Ice thickness, $\sim 200\mathrm{\mu }\mathrm{m}$; substrate stiffness, 52 kPa; substrate thickness, $45\mathrm{\mu }\mathrm{m}$; temperature gradient, $0.1\mathrm{K}/\mathrm{mm}$. (d) Change in displaced volume below a sporadically moving groove (green sections indicate motion). Approximate local power laws are given as guides to the eye. Top panel: the $y$ position of the grain boundary close to the bulk ice-water interface. Panels I–IV show surface displacements at the corresponding time points. Ice thickness, $\sim 500\mathrm{\mu }\mathrm{m}$; substrate stiffness, 52 kPa; substrate thickness, $45\mathrm{\mu }\mathrm{m}$; temperature gradient, $0.1\mathrm{K}/\mathrm{mm}$.