Why ice is so slippery

Sigbjørn Løland Bore, B.N.J. Persson, Henrik Andersen Sveinsson · 2026

Read Why ice is so slippery by Sigbjørn Løland Bore, B.N.J. Persson, Henrik Andersen Sveinsson (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-11539.txt.

Introduction

Bore, Persson, and Sveinsson settle an old question with fresh simulations: why is ice so slippery? Nanoscale runs alone overstate friction and miss how it changes with speed. The missing piece is frictional heating, which warms the contact toward the melting point. Heat thickens the premelting film tenfold and cuts its viscosity a hundredfold. Friction then drops with velocity, matching experiments. Premelting forms the film with no sliding or heating, yet slipperiness itself hinges on that heating, as Bowden and Hughes proposed in 1939.

Catalog record

CREATORSigbjørn Løland Bore, B.N.J. Persson, Henrik Andersen Sveinsson
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34344
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.11539

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Why ice is so slippery Sigbjørn Løland Bore, B.N.J. Persson, Henrik Andersen Sveinsson Abstract The origin of ice’s slipperiness has long puzzled scientists. To resolve this question, we simulate ice–glass (amorphous silica) friction at the nanoscale from first principles and upscale to the macroscale using a frictional heating model. We find that nanoscale simulations alone cannot capture the correct velocity dependence of ice friction, resulting in an overestimated coefficient of friction. By properly accounting for frictional heating, we find a strong increase in contact temperature toward the melting point, even under modest motion of 1 millimeter with velocities above 0.1 m/s, yielding excellent agreement with experimental friction data across a wide range of velocities. While the initial formation of a lubricating film on ice may occur without heating, the ultimate slipperiness of ice hinges on frictional heating, as proposed by Bowden and Hughes in 1939, but without incorporating melting. The interest in ice friction has ancient roots, with early human activity involving sliding on ice from around 10000 B.C. using sleds Gramly 2024 . On this historical timescale, it was only very recently that the first scientific explanation for ice slipperiness was put forward in the 1800s by Michael Faraday’s observations of premelting Faraday 1859 ; Jellinek 1967 : a liquid premelting film spontaneously forms on ice and provides lubrication. This explanation was followed by John Joly’s pressure-melting theory Joly 1887 : pressure causes a melting point depression, creating a lubricating premelting film. Subsequently, Bowden and Hughes proposed that the low friction of ice results from frictional heating Bowden and Hughes 1939 : friction between a moving object on ice, such as a skate blade, generates heat, creating a thin film of liquid water on the ice surface. This meltwater film acts as a lubricant, significantly reducing friction and enabling the smooth gliding motion characteristic of ice skating. Modern science continues to debate. On the experimental side, Canale et al. emphasized the nanorheology of ice friction, finding a lubricating layer consisting of a viscous slurry of ice nanoparticles in water Canale et al. 2019 . Weber et al. combined friction experiments with MD simulations, attributing ice slipperiness to highly mobile water molecules at the ice surface Weber et al. 2018 . Liefferink et al. followed up, investigating the roles of temperature, pressure, and velocity, proposing a ploughing mechanism Liefferink et al. 2021 . They concluded that the friction coefficient is not very sensitive to frictional heating. On the theoretical side, advances in computational efficiency have enabled molecular dynamics (MD) studies of nanoscale ice friction. Using force-field MD simulations, Baran et al. found the premelting film to be a robust phenomenon, reporting shear thinning and large slip lengths at hydrophobic interfaces as crucial for low ice friction, arguing for a minor role of frictional heating due to low film viscosity Baran et al. 2022 ; Baran and MacDowell 2024 . More recently, Atıla et al. performed ice sliding on ice MD simulations, proposing a mechanism for displacement-driven amorphization at low temperatures Atila et al. 2025 . Their MD shows that a relatively thick ( 10\text{\,}\mathrm{n}\mathrm{m} ) liquid-like film can form at ice interfaces during slip with negligible temperature rise, through a competition between shear-induced amorphization and recrystallization. When sliding stops, the film recrystallizes, leaving only a few interfacial monolayers in non-ice configurations set by the counter-surface. From these conflicting views, it can be concluded that no consensus has been reached on why ice is so slippery . In this paper, we present evidence that indicates frictional heating is the primary cause of ice slipperiness. Applying machine learning interatomic potentials based MD, we simulate nanoscale ice friction against a realistic ice–glass (amorphous silica) interface—a model system for ice–rock contacts in nature—from first principles (see Figure 1 a and the Supplemental Material SM for details). By integrating our nanoscale MD findings into the frictional heating model developed by Persson Persson 2015 , we predict the macroscopic ice friction behavior. Comparing our macroscale predictions with experimental ice–glass friction data, we find excellent agreement, strongly supporting the frictional heating theory of Bowden and Hughes. Figure 1: Nanoscale vs. macroscale ice friction. a Setup for nanoscale MD simulations of amorphous silica sliding on ice. b Friction stress as a function of contact temperature for different sliding velocities from nanoscale MD simulations, with fitted equation ( 1 ) (dashed lines). c Ice friction coefficient as a function of velocity for different temperatures with equation ( 1 ) directly, without frictional heating, comparing with experiments. d Illustration of macroscale friction, where nanoscale represents one asperity contact. e Contact temperature as a function of velocity for different background temperatures using the frictional heating model (equation ( 4 )). f Macroscale ice friction coefficient using the frictional heating model, comparing with experiments. Contact temperatures for nanoscale MD simulations are reported in \mathrm{\SIUnitSymbolCelsius} relative to the DC-R 2 SCAN model melting point of 289\text{\,}\mathrm{K} . Experimental friction coefficient data are from ice–glass friction experiments Canale et al. 2019 ; Miyashita et al. 2023 ; Akkok et al. 1987 and curling stone experiments Penner 2001 ; Nyberg et al. 2013 , which is composed of granite with \sim 70\text{\,}\mathrm{\%} content of silica. The key control parameters in ice friction experiments are the background temperature T_{0} of the ice and the velocity v of the sliding body. Direct prediction of experimental ice friction conditions using MD is impossible due to the large length and time scales involved. However, from nanoscale MD simulations, we can obtain the frictional shear stress \sigma_{\rm f}(v,T) acting in the asperity contact regions for given sliding velocity v and contact temperature T SM . We perform such simulations of the ice–glass contact (Figure 1 a), for v\gtrsim$0.01\text{\,}\mathrm{m}$ , with the friction stress shown in Figure 1 b. The friction stress from nanoscale MD is well-modeled by the following expression: \sigma_{\rm f}(v,T)=\sigma_{\rm k}\left(1-\frac{T}{T_{\rm m}}\right)^{\beta}\left[1-\left(1-\frac{\sigma_{\rm c}}{\sigma_{\rm k}}\right)e^{-\tfrac{v}{v_{k}}}\right], (1) where \sigma_{\rm c}=$60\text{\,}\mathrm{MPa}$ has been chosen based on experimental ice–glass friction data Miyashita et al. 2023 and \sigma_{\rm k}=$391\text{\,}\mathrm{M}\mathrm{P}\mathrm{a}$ , v_{k}=$0.54\text{\,}\mathrm{m}\mathrm{/}\mathrm{s}$ and \beta=0.61 on our nanoscale MD simulation data for v\geq$0.01\text{\,}\mathrm{m}\mathrm{/}\mathrm{s}$ , with parameterization details provided in the Methods below. Equation ( 1 ) can be rationalized as a sigmoid between two limiting behaviors: \displaystyle\sigma_{\rm f}(v,T)=\sigma_{\rm c}\left(1-\dfrac{T}{T_{\rm m}}\right)^{\beta},\quad v\ll v_{k}, (2) \displaystyle\sigma_{\rm f}(v,T)=\sigma_{\rm k}\left(1-\dfrac{T}{T_{\rm m}}\right)^{\beta},\quad v\gg v_{k}, (3) representing lower and upper velocity limits, respectively. Equation ( 2 ) has previously been proposed by Persson for ice–ice friction experiments for low velocities Persson 2015 , albeit with \beta=0.15 . To incorporate heat transport effects, we use a frictional heating model. By assuming that ice and glass make contact at asperities, making up only a small fraction of the apparent contact area (Figure 1 d), the contact temperature can be expressed as Greenwood 1991 ; Persson 2015 (see Methods below for additional details): T=T_{0}+\sigma_{\rm f}(v,T)\frac{s(v)vR}{\kappa}\left(1+\frac{\pi}{8}\frac{Rv}{D}\right)^{-\frac{1}{2}}. (4) Here, R is the asperity contact radius, D is the thermal diffusivity, \kappa is the thermal conductivity, all of which are available from literature data (see Table 1 ). s(v) is the proportion of heat that is absorbed by the slider, and can be explicitly expressed as a function of the sliding velocity (see Methods below), typically in the range from 0.6 to 1 with increasing sliding velocity.

Read the complete text