Physics of earthquake source can be investigated by monitoring active faults from borehole observatory in reservoirs (Maxwell et al. 2010) or by interpretation of seismic waves at the earth’s surface (Shearer 2019). Indeed, most information on earthquake mechanics is retrieved from seismology (e.g., Lee et al. 2002). However, the low resolution of these indirect techniques (cm to km scale) yields limited information on the physical and chemical deformation mechanisms active during earthquake rupture nucleation and propagation (Kanamori and Anderson 1975).
Experimental studies of frictional instabilities on fault gouge material or pre-existing surfaces (e.g., Brace and Byerlee 1966) may overcome those limitations (Scholz 1998; Marone 1998; Persson 2013). For instance, friction controls earthquake nucleation and propagation, the static and dynamic stress drops, the frictional heat generated during slip, and consequently the energy budget of earthquakes (Scholz 2019; Di Toro et al. 2011). All these processes can be investigated and monitored through laboratory experiments. In the last decades, rock friction properties have long been investigated using triaxial apparatuses in saw-cut configuration (e.g., Jaeger 1959; Byerlee 1967; Handin 1969), in which the fault is loaded at low velocities, typically orders of µm/s, and accumulates small displacements, typically few mm. In a seminal paper, Brace and Byerlee (1966) suggested that the stick–slip phenomenon observed in these rock friction experiments is analogous to natural earthquakes. Furthermore, to address the problem of earthquakes nucleation, biaxial apparatuses were developed and have long been used to study frictional properties of experimental faults under sub-seismic slip velocities in double-direct shear configuration (e.g., Dieterich 1972; Mair et al. 2002; Collettini et al. 2014; Giorgetti et al. 2015). The biaxial apparatus developed at USGS (USA) is amongst the first biaxial apparatuses used to investigate rock frictional properties (e.g., Dieterich 1972). Other pioneering biaxial apparatuses are the one in the Rock and Sediment Mechanics Laboratory at the Pennsylvania State University (USA) (e.g., Mair et al. 2002) and BRAVA (Brittle Rock deformAtion Versatile Apparatus) installed at INGV in Rome (Italy) (Collettini et al. 2014). Although the biaxial apparatuses developed in the past 50 years are characterized by different boundary conditions in terms of forces, pressures, temperatures and size of the samples, all of them take advantages from the double-direct shear configuration that allows good control of the normal and shear forces acting of the fault, accurate measurements of fault slip and dilation/compaction, and constant contact area.
Friction studies conducted with triaxial and biaxial deformation apparatuses are characterized by sub-seismic slip velocities and a limited amount of slip, 〈 10–3 m/s and few cm, respectively (e.g., Jaeger 1959; Byerlee 1967,1978; Brace and Byerlee 1966; Handin 1969; Paterson and Wong 2005; Lockner and Beeler 2002; Mair et al. 2002; Savage and Marone 2007; Samuelson et al. 2009; Carpenter et al. 2016). These experiments showed that the apparent static friction coefficient μ (i.e., μ = τ/σneff, where τ is the shear stress and σneff the effective normal stress acting on the fault) is between 0.60 and 0.85 for most rocks (Byerlee’s rule; except for phyllosilicates-rich rocks [Byerlee 1978]), for normal stresses up to 2 GPa, and temperatures up to 780 K. The apparent friction can thus be expressed as a function of slip velocity and a state variable, and modelled with the empirical rate- and state-dependent friction law (Dieterich 1979; Ruina 1983). Additionally, at velocities typical of earthquake nucleation phase, the apparent friction varies only a few percents for small changes in slip velocity, determining if a fault is or not prone to nucleate earthquakes.
Although Byerlee’s rule and the rate-and-state law have many applications in earthquake mechanics (inter-seismic and nucleation phase of earthquakes), these experiments were performed at slip velocities and displacements orders of magnitude smaller than those of earthquakes. Therefore, these experiments are unable to characterize the propagation phase of earthquakes. In the last 15 years, the multiplication of the rotary shear apparatus, designed to achieve slip velocities higher than 1 m/s and infinite displacement, overcome those limitations and produced unexpected results (Di Toro et al. 2010). A pioneering rotary shear apparatus capable of achieving seismic slip velocities up to 1.3 m/s were built and installed in Japan (Shimamoto 1994). Amongst others (see Di Toro et al. 2010 and references therein), a state-of-art rotary shear apparatus (SHIVA, Slow to High-Velocity Shear Apparatus) capable of deforming samples at slip rates up to 9 m/s has been installed at INGV in Rome (Italy) (Di Toro et al. 2010). Studies performed with these rotary shear apparatuses have shown a significant decrease in fault strength with increasing slip and slip velocity. They also reveal various dynamic fault‐weakening mechanisms (frictional melting, thermal pressurization, silica gel, elastohydrodynamic lubrication) that are likely active during earthquakes, including mechanisms that were unknown before conducting these experiments.
Though this new frontier is promising, key aspects of earthquake mechanics laboratory investigation, like being able to conduct high slip velocity experiments on rocks under elevated pore fluid pressure and temperatures characteristic of natural and induced earthquakes, remain beyond current experimental capabilities. Furthermore, studying links between pore‐fluid pressure, permeability, and frictional properties remains a challenge. To date, very few high-velocity friction experiments have been performed in presence of pore fluid pressure (Tanikawa 2012a, b, 2014; Violay et al. 2014, 2015, 2019; Cornelio et al. 2019a, b).
In this paper, we present a new state-of-art apparatus combining the advantages of biaxial apparatuses, i.e., simple geometry, high normal forces, confining pressure and pore fluid pressure, and the advantages of the rotary shear apparatuses, i.e. high slip velocity implemented thanks to the presence of electromagnetic motors. Building on the design of recent low-velocity biaxial machines implemented with pressure vessels (Samuelson et al. 2009; Collettini et al. 2014) and implementing the system with powerful linear motors (Di Toro et al. 2010), the new HighSTEPS (High Strain TEmperature Pressure Speed) apparatus is able to reproduce the deformation conditions typical of the seismogenic crust, i.e., confining pressure up to 100 MPa, slip velocity from 10–5 to 0.25 m/s, temperature up to 120 °C, pore pressure up to 100 MPa. Under these unique boundary conditions, the new apparatus allows the investigation of the entire seismic cycle (inter-seismic, nucleation and propagation).
3T. Fisica dei terremoti e Sorgente Sismica
Biaxial friction apparatus
Low to high slip velocity
Deformation conditions of the seismogenic upper crust
Istituto Nazionale di Geofisica e Vulcanologia (INGV)