Description:

_Journal of Structural Geology 32 (2010) 59–69_ Contents lists available at ScienceDirect Journal of Structural Geology journal homepage: www.elsevier.com/locate/jsg Grain size distribution and microstructures of experimentally sheared granitoid gouge at coseismic slip rates – Criteria to distinguish seismic and aseismic faults? Holger Stu?nitz a,*, Nynke Keulen b,d, Takehiro Hirose c, Rene?e Heilbronnerb a Department of Geology, University of Troms?, Dramsveien 201, 9037 Troms?, Norway b Dept. of Environmental Geosciences, Basel University, Bernoullistr. 32, 4056 Basel, Switzerland c Kochi Institute for Core Sample Research, JAMSTEC, Kochi, Japan d Geological Survey of Denmark and Greenland, ?ster Voldgade 10, 1350 Copenhagen K, Denmark Article info Article history: Received 9 December 2008 Received in revised form 1 July 2009 Accepted 4 August 2009 Available online 11 August 2009 Keywords: Cataclasite Rock deformation Microstructures Seismic deformation Grain size Abstract Microstructures and grain size distribution from high velocity friction experiments are compared with those of slow deformation experiments of Keulen et al. (2007, 2008) for the same material (Verzasca granitoid). The mechanical behavior of granitoid gouge in fast velocity friction experiments at slip rates of 0.65 and 1.28 m/s and normal stresses of 0.4–0.9 MPa is characterized by slip weakening in a typical exponential friction coefficient vs displacement relationship. The grain size distributions yield similar D-values (slope of frequency versus grain size curve ? 2.2–2.3) as those of slow deformation experiments (D ? 2.0–2.3) for grain sizes larger than 1 mm. These values are independent of the total displacement above a shear strain of about g ? 20. The D-values are also independent of the displacement rates in the range of w1 mm/s to w1.3 m/s and do not vary in the normal stress range between 0.5 MPa and 500 MPa. With increasing displacement, grain shapes evolve towards more rounded and less serrated grains. While the grain size distribution remains constant, the progressive grain shape evolution suggests that grain comminution takes place by attrition at clast boundaries. Attrition produces a range of very small grain sizes by crushing with a D<-value ? 1. The results of the study demonstrate that most cataclastic and gouge fault zones may have resulted from seismic deformation but the distinction of seismic and aseismic deformation cannot be made on the basis of grain size distribution. ? 2009 Elsevier Ltd. All rights reserved. 1. Introduction Fault rocks, and especially fault rocks from seismic events, have received increasing attention over the past years. In the field, it is extremely difficult, if not impossible, to distinguish between seismic and aseismic fault rocks (Cowan, 1999; Sibson, 1989). So far, the occurrence of pseudotachylites appears to be the main reliable criterion for the identification of seismic slip in natural fault rocks (McKenzie and Brune, 1972; Sibson, 1975; Spray, 1992, 1995, 1997). Another criterion is the occurrence of clay-clast-aggregates in clay-rich gouges (Boutareaud et al. 2008). Alternatively, the direct association of fault zones with active faulting, e.g., in the San Andreas fault system or the Nojima fault of the Kobe earthquake may be used to infer that seismic slip has occurred in the fault rocks (Sammis et al., 1987; Wilson et al., 2003; Chester et al., 2005; Murata et al., 2001; Monzawa and Otsuki, 2003). These criteria apply to a small number of faults. For the large part of the natural inventory of fault rocks, which may be available for detailed analysis as potentially seismic zones, seismic slip rates cannot be inferred with certainty. If measurable properties, such as grain size, grain shape, etc., could be used to identify fast slip rates, a large number of natural faults from all geological periods and from many different tectonic settings would be available for studies of processes and mechanisms of the earthquake cycle. One way of obtaining information about deformation processes in fault rocks is to perform experiments employing fast slip rates as these may simulate seismic slip rates (e.g., Spray, 1987; 1993; Tsutsumi and Shimamoto, 1997; Hirose and Shimamoto, 2003, 2005; Mizoguchi et al., 2007; Han et al., 2007a, b; Boutareaud et al., 2008). We have carried out fast slip experiments on granitoid fault gouge because there is data of microstructures and grain size distributions of the same material from slow deformation experiments. The experiments at slow deformation rates have been performed in a Griggs solid medium apparatus by cracking of solid samples of granitoid material followed by sliding of the fragments at high confining pressures and a range of temperatures (Keulen et al., 2007, 2008). The objectives of the present investigation are to study the evolution of the grain size distribution at different slip rates and at different amounts of total slip in order to obtain information about the comminution process, the dependence of sliding on the microstructures, and to potentially quantify the evolution of the grain size distribution for different rates of slip (seismic and aseismic) and for different total amounts of slip. 2. Experimental and analytical procedures The fast slip experiments have been performed on crushed starting material from Verzasca gneiss in a high speed rotary testing apparatus (Shimamoto and Tsutsumi, 1994). The crushed material has been prepared from the same Verzasca granitoid gneiss material as the slow deformation experiments of Keulen et al. (2007, 2008). Verzasca gneiss is a fine-grained granitoid gneiss (grain size w280 mm) with almost no secondary alteration minerals and a very weak planar and linear fabric. The location of the sample of the starting material is: Swiss coordinates 704.65 126.30; the composition is: 29% plagioclase, 27% K-feldspar, 35% quartz, 7% mica (mostly biotite). The starting material has been prepared by pounding of the solid rock and large fragments (only once or a couple of times) and sieving the fine fragments after pounding. The larger fragments have been pounded and sieved again until the grain size of the starting material has been smaller than 500 mm. By this procedure, we attempt to produce fragmented material more typical of natural damage zones rather than cataclastic or gouge material which is typically produced in natural slip or process zones after extensive attrition and wear. The starting material (1 g) has been placed between roughened (ground flat using no. 80 SiC powder) cylinders of Verzasca gneiss (24.5–24.6 mm diameter; Fig. 1), resulting in about 1 mm thick layers of the crushed material. No water has been added but some water may be present as a result of adsorption from air humidity. A Teflon sleeve (Fig. 1) holds the crushed material in place during shearing but no confining pressure is applied. The sample is inserted into a high speed rotary shear apparatus (Tsutsumi & Shimamoto, 1994) and deformed at different slip rates and different amounts of total displacement and normal stress (Table 1). The slip rate of the sample varies between 0 at the center and the maximum rate at the periphery. An ‘‘equivalent slip velocity’’ Veq (termed ‘‘slip rate’’ in this text) is defined by obtaining the frictional work rate Wr on the slip surface area S assuming that the shear stress s is constant over the slip surface and does not vary with velocity: Wr ? s VeqS (1) Table 1 Results of fast experiments performed on high speed rotary testing apparatus. Inferred temperature for all experiments 400 x14C (see text). Sample Normal Slip rate Steady state Total Average Number stress _MPa_ _m/s-1_ friction coef?cient displacement D-value of _m_ gouge HVR 833 0.8–0.9 HVR 835 0.8–0.9 HVR 836 0.8–0.9 HVR 840 0.8–0.9 HVR 841 0.8–0.9 HVR 842 0.4–0.5 1.28 1.28 0.65 1.28 1.28 1.27 0.35–0.45 0.35–0.5 0.57–0.7 0.35–0.5 0.45–0.55 0.45–0.65 24 26 20 36 11 39.5 2.24 2.26 2.26 ? 2009 Elsevier Ltd. All rights reserved. After the experiment, the crushed starting material is much finer in grain size and represents a fault gouge typical for extensive slip, attrition, and wear. A minor amount of shortening occurs during the experiment, partly by compaction of the gouge during the rotation, partly by small amounts of gouge moving between the cylinders and the Teflon sleeve. All experiments have been carried out at room temperature, but some shear heating of the samples has occurred. The sample temperature has not been measured, but gouge experiments of Boutareaud et al. (2008) were performed under the same slip rates and the same normal stress. Boutareaud et al. (2008) have measured peak temperatures around 400 x14C, and we infer similar temperatures for our experiments. No melt has been detected in any of the samples. After the experiments, the Teflon ring was removed using a C-clamp to hold the forcing blocks in place. The gouge was vacuum-impregnated with Laromin, a low viscosity epoxy. After impregnation, the samples were cut for thin sections parallel to the rotation axis, about 3–4 mm from the periphery of the cylinder. The shape analysis of grains was carried out on SEM-BSE micrographs using the image analysi... Ключевые слова: fault gouge, keulen, slip rate, process, journal structural, crushed material, small, seismic, slow deformation, cracked grain, shear stress, hirose shimamoto, geology, doi, total slip, marone, friction coef?cient, structural, gouge layer, stress, frequency, mechanical data, murata, mpa, plot, attrition, slip zone, fractal dimension, mm, displacement rate, study, plane, nitz, granitoid gouge, coef?cient, sample hvr, aspect ratio, friction coefcient, cowan, geophysical, gouge, slow, grain comminution, fault zone, seismic slip, panozzo, slope, teon sleeve, mechanical property, data, rate, granitoid, small displacement, shape, nature, natural fault, rk, cracked material, higher, slip experiment, displacement, weakening, slow experiment, material, rotation axis, ?ne, existing gouge, cylinder, friction, stu nitz, heilbronner, gouge material, hirose, structural geology, melt, shimamoto mizoguchi, d-value, clast, slip weakening, distribution, observed, values, slip, sibson, velocity, zone, particle size, strain, high, fast, deformation, normal, island, natural, comminution process, damage zone, large, scholz, normal stress, increasing slip, journal structural geology, letters, journal geophysical, earthquake, shear heating, solid, total, experiment, journal, ?eld, deformed sample, microscopy, microstructural, blenkinsop, fault, grain growth, frequency relationship, stu?nitz, stu?nitz journal, shimamoto, boundary, wilson, relationship, deformation experiment, nk, starting, larger fragment, convex shape, normal gouge, ha, nojima fault, science, fault rock, aseismic, shear, applied, fast slip, result, clast boundary, solid cylinder, size, tectonophysics, microstructures, starting material, solid rock, size distribution, tsutsumi, total displacement, grain size, spalling, gsd, deformed, hvr, shear strain, comminution, temperature, shape analysis, cracked, chester, sammis, grain, rock, wide range, origin, mechanical behavior, verzasca gneiss, grain shape, particle, marone mair, mair, sample, range, d-values plotted, barrett, d-values, layer, fragment, mizoguchi, paris factor, stu, analysis, storti, spray