The granular physics contribution to rate- and state-dependent fault friction

摩擦在许多社会利益领域发挥着关键作用，包括运输和制造业。在地球科学中，了解摩擦力对于更好地了解与地震和山体滑坡相关的危险至关重要。材料的摩擦性能已经被研究了几个世纪，但其与时间相关的物理和化学基础仍然不清楚。然而，地震和山体滑坡期间摩擦特性的微小波动可能会对这些事件的规模和速度产生巨大影响。滑动界面上的摩擦，如构造断层，通常用所谓的“速率和状态相关的摩擦”定律来描述。这些经验定律解释了滑动速度(“速率”)和称为“状态”的界面的演化特性；后者是滑动历史的函数，很难直接观察到。速率-状态框架被广泛用于模拟摩擦滑动。但相应的定律未能准确描述与地震相关的一系列条件的实验室观察结果。在这里，该团队的目标是更好地理解岩石摩擦特性背后的物理原理。研究人员使用计算机模拟来模拟沿着构造断层存在的被称为断层泥的细碎岩石颗粒层的行为。其目的是测试岩石摩擦力及其时间依赖性是否受颗粒间相互作用在颗粒尺度上的支配。模拟输出受到实验观察的限制：在许多情况下，它们比最成功的速率-状态摩擦定律更好地描述它们。因此，该团队逐渐揭开了孕育地震的断层行为背后的物理机制。除了其强烈的社会相关性，该项目还为早期职业科学家提供支持，并为本科生提供培训。为了模拟刨削的行为，研究人员使用了离散元方法模拟。他们使用模型几何和加载条件来模拟标准的岩石摩擦实验，如“速度-阶跃”和“滑动-保持-再滑动”协议。他们测试了在实验室中观察到的岩石摩擦力是由颗粒-颗粒接触尺度上的与时间无关的特性控制的假设。这种创新的方法不同于更传统的方法，后者假设微观接触处的依赖于时间的塑性或化学键是摩擦的速率和状态依赖的来源。颗粒模拟与滑动协议中最成功的依赖于速率和状态的摩擦方程是一致的，其中这些方程准确地描述了实验(“速度-步长”和“滑动-保持”协议)。在这些方程失效的地方，它们更好地与滑动方案的实验室数据相匹配(例如，在“滑动-保持”方案之后的重新下滑)。此外，颗粒模拟的输出允许调查模型的与速率和状态相关的类似摩擦行为的来源。该团队发现，如果泥颗粒的动能被围压适当地归一化，它就会产生与模拟一致的速度依赖估计，并在实验室数据的大致范围内。研究人员继续探索颗粒流模型，将其与现有公式不能很好解释的更广泛的滑动协议进行比较(例如，“滑动-保持-再滑动”和“法向应力-台阶”实验)。他们还将模拟中挖沟层的压实/膨胀与实验观察进行比较；目标是评估孔隙度对刨削滑动行为的作用。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Friction plays a critical role in many areas of societal interest, including transportation and manufacturing. In Earth Sciences, understanding friction is critical for a better understanding of the hazards associated with earthquakes and landslides. The friction properties of materials have been studied for centuries, but the physics and chemistry underlying their time dependence remain obscure. Yet, small fluctuations in friction properties during earthquakes and landslides can have tremendous effects on the size and speed of these events. Friction on sliding interfaces such as tectonics faults are usually described by the so called "rate- and state-dependent friction" laws. These empirical laws account for the sliding speed ("rate") and for the evolving properties of the interface termed "state"; this latter, a function of the slip history, is difficult to observe directly. The rate-and-state framework is widely used to model frictional sliding. But the corresponding laws fail to accurately describe laboratory observations for a range of conditions relevant to earthquakes. Here, the team aims to better understand the physics underlying the frictional properties of rocks. The researchers use computer simulations to model the behavior of granular layers of finely-ground rock, called gouge, that are present along tectonic faults. The goal is to test whether rock friction and its time dependence is governed at the grain scale by grain-to-grain interactions. The simulation outputs are constrained by experimental observations: in many cases they describe them better than the most successful rate-and-state friction laws. The team, thus, gradually unveils the physics underlying the behavior of earthquake-generating faults. In addition to its strong societal relevance, this project provides support for an early career scientist as well as training for undergraduate students. To model the behavior of the gouge, the researchers employ Discrete Element Method simulations. They use model geometries and loading conditions designed to mimic standard rock-friction experiments, such as "velocity-step" and "slide-hold-reslide" protocols. They test the hypothesis that rock friction as observed in the laboratory is governed by time-independent properties at the grain-grain contact scale. This innovative approach differs from more traditional ones which assume that time-dependent plasticity or chemical bonding at microscopic contacts are the source of the rate-and-state dependence of friction. The granular simulations are consistent with the most successful rate-and-state-dependent friction equations for sliding protocols where those equations accurately describe experiments ("velocity-step" and “slide-hold” protocols). They better match laboratory data for sliding protocols where those equations fail (e.g., the reslides following "slide-hold" protocols). Furthermore, output of the granular simulations allows investigating the source of the rate-and-state-dependent friction-like behavior of the model. The team finds that if the kinetic energy of the gouge particles is suitably normalized by the confining pressure, it produces an estimate of the velocity dependence that is consistent with the simulations and within the ballpark of laboratory data. The researchers continue exploring the granular flow model by comparing it to a wider range of sliding protocols that are not well explained by existing equations (e.g., "slide-hold-reslide" and "normal-stress-step" experiments). They also compare the compaction/dilation of the gouge layers in the simulations to experimental observations; the goal is to evaluate the role of porosity on the gouge sliding behavior.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.