High‐Velocity Friction Experiments Indicate Magnetic Enhancement and Softening of Fault Gouges During Seismic Slip

Abstract

High-velocity friction experiments were conducted on two natural fault gouges retrieved from the Yingxiu-Beichuan fault zone (Sichuan, China), which accommodated the 2008 Wenchuan Mw 7.9 earthquake. The experiments simulate large earthquake slip; the rotary shear apparatus enables to gather information as a function of shear displacement (and slip velocity) in a single experiment. The two starting fault gouges are essentially paramagnetic with one containing goethite. The experimentally sheared gouges both show a significant magnetic enhancement and softening with increasing slip distance. Rock magnetic measurements reveal that magnetite was formed during the experiment due to thermochemical reactions of iron adsorbed on clay minerals (smectite, chlorite). Scanning electron microscope observations showed that the new magnetite occurs as spherical and sintered irregular aggregates. This implies a high-temperature origin. The peak temperatures are estimated numerically to range from~260 to~440°C. Also, during the experiment on the goethite-bearing fault gouge, goethite was altered, reduced to magnetite due to the decomposition of trace organic matter present in the starting material. Magnetic susceptibility and magnetization of sheared samples are linearly increasing with the temperature rise induced by frictional heating; coercivities are decreasing. The grain size of the newly formed magnetite increases with heating temperature. This demonstrates that short-duration frictional heating generated by fast seismic slip induces thermal alteration: neoformation of ferrimagnetic minerals leads to magnetic enhancement and softening of slip zones. Thus, rock magnetic methods are a useful tool for diagnosing the earthquake slip and estimating the temperature rise of coseismic frictional heating.

Faulting Processes Unveiled by Magnetic Properties of Fault Rocks

“…Substantial magnetic enhancement is also observed experimentally in high‐velocity friction experiments on natural fault gouges with increasing slip distance (e.g., Fukuchi et al, 2005; Tanikawa et al, 2007; Yang et al, 2019). Magnetite, occurring as spherical and sintered irregular aggregates (Figure 16), was formed during friction experiments on fault gouges from the Yingxiu‐Beichuan fault.…”

“…On heating they desorb iron which is precipitated as magnetite. Fine‐grained magnetite is produced when heating smectite over 250°C in the laboratory; nucleation and growth of the magnetite grains proceed up to ~450–500°C (Hirt et al, 1993; Yang et al, 2019). Thermochemical reactions of the iron adsorbed to chlorite (particularly chamosite) result in the neoformation of magnetite on heating to 400–700°C (Tanikawa et al, 2008).…”

“…The reactions were induced by seismic frictional heating and related coseismic hot fluids (Figure 12), thus intimately associated with earthquake events (Yang et al, 2018). Examination of the magnetic mineral assemblage in fault zones may therefore be a prospective way for identifying the most recent PSZ (Cai et al, 2019; Chou, Song, Aubourg, et al, 2014; Chou, Song, Aubourg, Song, et al, 2012; Yang et al, 2012a, 2018, 2019).…”

“…Neoformation of ferrimagnetic minerals through thermochemical reactions of Fe‐bearing minerals induced by frictional heating is deemed widespread in seismic fault zones (see section 4.3). This opens up a new perspective in estimating the temperature experienced during seismic slip (e.g., Chou, Song, Aubourg, Song, et al, 2012; Fukuchi et al, 2005, 2007; Han et al, 2007; Hirono et al, 2009; D. Liu et al, 2016; Mishima et al, 2006; Tanikawa et al, 2007, 2008; Yang et al, 2019, 2018; Yang, Dekkers, & Zhang, 2016). For example, a strong magnetic susceptibility anomaly in fault gouge was observed at the Fault Zone FZB1136, in TCDP Hole‐B (Figure 11a).…”

“…(c) Modeled temperature due to frictional heating in the different annuluses as a function of slip time. The reader is referred to Yang et al (2019) for more details on the equations and modeling strategy. (d) Hysteresis loops after and before (insets) correction for the paramagnetic contribution of the starting material and run products after the experiments.…”