Small quantity but large effect — How minor phases control strain localization in upper mantle shear zones

Tectonophysics 643 (2015) 26–43

Contents lists available at ScienceDirect

Tectonophysics journal homepage: www.elsevier.com/locate/tecto

Small quantity but large effect — How minor phases control strain localization in upper mantle shear zones Jolien Linckens a,⁎, Marco Herwegh b, Othmar Müntener c a b c

Institute of Geoscience, Goethe-University, Altenhoeferallee 1, 604381 Frankfurt, Germany Institute of Geological Sciences, University of Bern, Baltzerstrasse 1 + 3, 3012 Bern, Switzerland Institute of Earth Sciences, University of Lausanne, Quartier UNIL-Mouline, Bâtiment Géopolis, 1015 Lausanne, Switzerland

a r t i c l e

i n f o

Article history: Received 17 June 2014 Received in revised form 11 November 2014 Accepted 21 December 2014 Available online 3 January 2015 Keywords: Strain localization Zener parameter Shear zone Mantle deformation Polymineralic

a b s t r a c t Low viscosity domains such as localized shear zones exert an important control on the geodynamics of the uppermost mantle. Grain size reduction and subsequent strain localization related to a switch from dislocation to diffusion creep, is one mechanism to form low viscosity domains. To sustain strain localization, the grain size of mantle minerals needs to be kept small over geological timescales. One way to keep olivine grain sizes small is by pinning of mobile grain boundaries during grain growth by other minerals (second phases). Detailed microstructural studies based on natural samples from three shear zones formed at different geodynamic settings allowed the derivation of the olivine grain-size dependence on the second-phase content. The polymineralic olivine grain-size evolution with increasing strain is similar in the three shear zones. If the second phases are to pin the mobile olivine grain boundary the phases need to be well mixed before grain growth. We suggest that melt–rock and metamorphic reactions are crucial for the initial phase mixing in mantle rocks. With ongoing deformation and increasing strain, grain boundary sliding combined with mass transfer processes and nucleation of grains promotes phase mixing resulting in fine-grained polymineralic mixtures that deform by diffusion creep. Strain localization due to the presence of volumetrically minor minerals in polymineralic mantle rocks is only important at high strain deformation (ultramylonites) at low temperatures (b~800 °C). At smaller strain and stress conditions and/or higher temperatures other parameters like overall energy available to deform a given rock volume, the inheritance of mechanical anisotropies or the presence of water or melts needs to be considered to explain strain localization in the upper mantle. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Experimental studies on the most important rock forming minerals of the crust and upper mantle such as quartz, feldspar and olivine have led to the formulation of flow laws for monomineralic aggregates that are of fundamental importance for rheological models of the Earth's crust and mantle (e.g. Bai and Kohlstedt, 1992; Bai et al., 1991; Gleason and Tullis, 1995; Hansen et al., 2011; Hirth and Kohlstedt, 1995; Mei and Kohlstedt, 2000a, 2000b; Rybacki and Dresen, 2000). These flow laws show that deformation can occur by various mechanisms, which can be grain size insensitive (dislocation-accommodated creep) or grain size sensitive (diffusion accommodated creep) as described by the constitutive equation:


−m

ε¼ Ad

n

σ expð−Q=RTÞ




ð1Þ

where ε is the strain rate, A is a constant, d is the grain size, m is the grain size exponent, σ is the differential stress, n is the stress ⁎ Corresponding author. E-mail address: [email protected] (J. Linckens).

http://dx.doi.org/10.1016/j.tecto.2014.12.008 0040-1951/© 2015 Elsevier B.V. All rights reserved.

exponent, Q is the activation energy, R is the gas constant and T is the temperature. The most recent development in experimental deformation of olivine shows that dislocation-accommodated grain boundary sliding (disGBS) is an important olivine deformation mechanism at a range of temperature conditions in the dry upper mantle (Hansen et al., 2011). The activity of grain boundary sliding has not yet been observed in olivine deformation experiments at “wet” conditions (Hirth and Kohlstedt, 2003). DisGBS, in contrast to dislocation creep, has a slight grain size dependence (grain size exponent of 0.7 ± 0.1). In addition to dislocations, disclinations (rotational defects) are shown to be an important grain boundary accommodation mechanism during olivine deformation (Cordier et al., 2014). These flow laws are all based on a monomineralic upper mantle, which is only an approximation of natural systems given the generally polymineralic nature of most upper mantle rocks. It has been shown in experimentally (Farla et al., 2013; Hiraga et al., 2010a; Ji et al., 2001; Mcdonnell et al., 2000; Sundberg and Cooper, 2008; Tasaka et al., 2013) and naturally deformed upper mantle rocks (Linckens et al., 2011b; Skemer et al., 2010; Tasaka et al., 2014; Toy et al., 2010; Warren and Hirth, 2006), as well as in theoretical models (Bercovici

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

and Ricard, 2012) that the presence of secondary minerals may lead to substantially different rheological behavior. When the extrinsic physical conditions are similar (e.g. temperature, stress, strain rate) a pronounced decrease of olivine grain size, by recrystallization or brittle deformation, induces a change from dislocation via disGBS to diffusion creep, resulting in a reduced viscosity of the upper mantle (Hansen et al., 2011; Hirth and Kohlstedt, 2003; Precigout and Gueydan, 2009). This reduction in viscosity due to the activation of diffusion creep is thought to be short lived as olivine grain growth kinetics are fast (Karato, 1989), and grain growth is expected in the diffusion creep field, forcing the grains to grow until becoming stabilized by dynamic recrystallization in the disGBS or dislocation creep field (De Bresser et al., 2001). However, second-phases help to preserve the small grain sizes by inhibiting grain growth (Evans et al., 2001; Herwegh et al. 2011; Hiraga et al., 2010b; Ohuchi and Nakamura, 2007; Olgaard and Evans, 1986) and can induce and maintain a change in the dominant deformation mechanism (Bercovici and Ricard, 2012; Farla et al., 2013; Hiraga et al., 2010b; Linckens et al., 2011b; Skemer et al., 2010). Based on naturally deformed peridotites from two different mantle shear zones, and comparing the results with a previously analyzed mantle shear zone (Linckens et al., 2011b), we determine the effect of second phases on the grain size of olivine, the dominant deformation mechanisms and the resulting strain localization behavior. In addition, we discuss the possible processes that can lead to the mixing of the different phases. The data set allows for a more accurate microstructural and therefore rheological description of mantle shear zones.

27

2011a, 2011b). For easy comparison, the data of the Oman shear zone will be included in the result figures. The shear zones formed either under a strike-slip (Othris, Semail) or extensional (Lanzo) geodynamic framework. Their microstructures represent different deformation intensities, from relatively undeformed porphyroclastic tectonites to extensively deformed ultramylonites (Fig. 3). The width of the deformation zones decreases with increasing deformation intensity (Figs. 1 and 2, Boudier et al., 1988; Dijkstra et al., 2002b; Kaczmarek and Müntener, 2008; Linckens et al., 2011b). 2.1. Othris The Othris ophiolite is part of the Hellenic Tethyan ophiolite belt. The ophiolite is thought to have formed near a transform fault at a slowspreading ridge (Barth, 2003; Dijkstra, 2001) or above an intraoceanic subduction zone at a mid-oceanic ridge (Barth and Gluhak, 2009; Barth et al., 2008; Rassios and Smith, 2000). It consists mainly of spinel–harzburgites and plagioclase–lherzolites with minor amounts of spinel–lherzolites and dunites (Dijkstra, 2001; Dijkstra et al., 2002b; Menzies and Allen, 1974). The samples analyzed in this study are from a mylonitic harzburgite (Fig. 1), which locally turns into ultramylonites. The mylonite is part of a km-wide N–S trending shear zone (Fig. 1, Dijkstra (2001)). The mylonites consist of monomineralic olivine and polymineralic domains on a thin section scale. Polymineralic domains formed by melt–rock reactions (Dijkstra et al., 2002b), after which solid-state deformation took place. With increasing deformation the fine-grained bands coalesced and are increasingly elongated resulting in continuous bands of fine-grained polymineralic aggregates in the mylonite (Dijkstra et al., 2002b).

2. Geological settings and sample descriptions In order to quantify the grain size evolution with progressive deformation in mantle shear zones, we studied two different largescale mantle shear zones (Fig. 1; Othris, Greece; Lanzo, Italy). These results will be compared and discussed with previous results derived from a third shear zone located at Semail, Oman (Fig. 2, Linckens et al.,

2.1.1. Mylonite The grains in the monomineralic olivine domains are strongly elongated (axial ratio of up to 1:10, see Fig. 4A). They have subgrain boundaries oriented subvertically with respect to the foliation. Some fine-grained (b50 μm) spinels and pyroxenes occur between the olivine grains in these aggregates (Fig. 4A). Porphyroclasts of orthopyroxene (0.4 to

Fig. 1. Geological maps of (A) Othris (Greece) and (B) Lanzo (Italy) with sample locations, after Dijkstra et al. (2002a, 2002b) and Kaczmarek and Müntener (2008), respectively. PFG = porphyroclastic fine-grained texture, UMB = ultramylonitic bands.

28

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

Fig. 2. Geological map of Oman, after Linckens et al. (2011b) indicating the sample locations.

6 mm) show exsolution lamellae of clinopyroxene. Some smaller porphyroclasts have very irregular boundaries and are often surrounded by fine-grained mixtures of mainly olivine and orthopyroxene (≤~200 μm) as well as some spinel (≤50 μm). Finer grained mixtures between the domains are not continuous on thin section scale but form domains (Fig. 4A). 2.1.2. Ultramylonite bands The ultramylonitic bands make an angle with the mylonitic foliation (~15°, Fig. 4A). The width of the bands is variable (0.1 to 2 mm). Most of the bands are continuous and straight on a thin section scale. There is a sharp boundary between the bands and the mylonite, with an abrupt grain size decrease in the ultramylonites (Fig. 4A). Small brittle structures occur perpendicular to the bands and monomineralic domains are absent in these bands (Fig. 4A). Due to the small grain sizes, individual minerals are difficult to distinguish under the optical microscope (transmitted and crossed-nichols light), although some larger grains (~30 μm) of olivine or orthopyroxene can be recognized in the fine-grained matrix. Some grain boundaries in the polymineralic bands are aligned over a distance of a few hundred microns.

fine-grained texture (PFG), protomylonite, mylonite and mylonite with ultramylonite bands (mylonite–UMB, see Fig. 1). In this study we analyzed the protomylonite, mylonite and ultramylonite in detail. There is a sharp transition from mylonites to porphyroclastic tectonites at the north-eastern part of the shear zone (see also Kaczmarek and Tommasi, 2011). The lherzolite in the shear zone is inhomogeneous and finegrained due to melt–rock reactions at an early stage of deformation, forming olivine, plagioclase and pyroxene mixtures (Higgie and Tommasi, 2014; Kaczmarek and Müntener, 2008). 2.2.1. Protomylonite The orthopyroxene porphyroclasts (1–5 mm), in the protomylonite are mainly round shaped but there are some strongly elongated porphyroclasts (axial ratio 1:11). These elongated clasts have exsolution lamellae of clinopyroxene. The clinopyroxene porphyroclasts are smaller (0.5–1 mm) and are not elongated. The olivine grain size is larger (500 μm) in the monomineralic olivine domains than in the polymineralic domains (Fig. 4B). The large olivine grains have subgrain boundaries and are not elongated (Fig. 4B). Most polymineralic mixtures contain olivine, pyroxenes and spinel and some include additionally plagioclase.

2.2. Lanzo The Lanzo peridotite is situated west of Torino (north western Italy, Fig. 1) and is part of the high-pressure belt of the Western Alps. Nonetheless, initial structures and microfabrics survived in lenses surrounded by serpentinized peridotites (Kaczmarek and Müntener, 2008). The peridotite consists mainly of plagioclase–lherzolite and some dunites (Boudier, 1978). It evolved in an ocean–continent transition or (ultra-)slow-spreading ridge (Kaczmarek and Müntener, 2008) during the formation of the Liguro–Piemontese Ocean (Bodinier, 1988). The samples investigated are from a sinistral extensional shear zone in the north of the Lanzo massif. Detailed mapping and thin section studies allowed the discrimination of five different tectonites; porphyroclastic tectonite, porphyroclastic

2.2.2. Mylonite The mylonites contain a large amount of very fine-grained (b10 μm) matrix in which olivine and pyroxene porphyroclasts occur. Orthopyroxene clasts can be extremely elongated (1:50, Fig. 4C) and contain exsolution lamellae, some clasts are folded. Round shaped orthopyroxene porphyroclasts (0.7–1.5 mm) also occur; many have exsolution lamellae of clinopyroxene and all show undulose extinction. Small orthopyroxene (up to 150 μm) is present in polymineralic bands, the larger grains (100–150 μm) have undulose extinction. There are small amounts of olivine porphyroclasts (1–2 mm), which are clustered in lenses with smaller grains (~400 μm) in the tail of the lenses. Other monomineralic olivine domains, that do not include a porphyroclast, form layers in the fine-grained matrix. These layers are

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

29

et al., 2000; Tasaka et al., 2014) as well as the shear zone at Wadi Al Wasit (Linckens et al., 2011b) have been analyzed in detail. In contrast to the other two ophiolites, the Semail ophiolite formed at a fast spreading ridge. The detailed mapping of the peridotites (Fig. 2) shows that the largest part of the mantle was deformed at medium-high temperatures (1100–1200 °C) during ridge-related flow (Ceuleneer et al., 1988; Gnos and Nicolas, 1996; Nicolas et al., 2000). During compression, the ophiolite was obducted onto the Arabian shield, and deformation occurred at lower temperatures (b900 °C, Linckens et al., 2011a) and higher stresses, resulting in the formation of shear zones ranging from a few meters to 1–2 km (Boudier et al., 1988). The different deformation types of the Wadi Al Wasit shear zone have been previously described in detail by Linckens et al. (2011b). For comparison with the other two shear zones, we focus in this study on three deformation types: porphyroclastic tectonite, mylonite and ultramylonite. 2.3.1. Porphyroclastic tectonite The porphyroclastic tectonites show a bimodal grain size distribution, with large sized porphyroclasts of olivine and orthopyroxene (Fig. 4E) and fine-grained domains, which are mainly polymineralic. The olivine porphyroclasts have a grain size of 2–6 mm. The porphyroclast grains show undulose extinction and some are elongated (aspect ratio ~ 4). Orthopyroxene porphyroclasts (1–4 mm) show undulose extinction and contain exsolution lamellae of clinopyroxene. The orthopyroxene porphyroclasts have no preferred orientation and are not elongated. Large spinel grains (200–500 μm) are aligned and define the lineation (Fig. 4E). There are no clinopyroxene porphyroclasts in the samples. The fine-grained domains contain orthopyroxene (~100 μm), olivine (~300 μm), spinel (~60 μm) and clinopyroxene (~150 μm). The domains are not interconnected throughout the thin section and are concentrated around orthopyroxene porphyroclasts. Locally, olivine is altered to serpentine and orthopyroxene to talc and tremolite.

Fig. 3. Crossed nichols optical photomicrographs illustrating the microstructural evolution from porphyroclastic tectonite (Oman, A), mylonite (Lanzo, B) to ultramylonite (Lanzo, C), as well as variations in olivine grain size in mono- (M) and polymineralic (P) areas. Scale bars: 2 mm, the polymineralic regions consist of olivine (bright colors), orthopyroxene (gray) and spinel (black).

not continuous on the thin section scale and include some small grains of spinel and pyroxenes. The grain size transition between the coarser grained monomineralic olivine layers and the fine-grained matrix is abrupt. The fine-grained (b 10 μm) polymineralic matrix contains olivine, pyroxenes, plagioclase and spinel (Fig. 4C). The most abundant second phase in many domains is orthopyroxene, however, in few domains in the mylonite sample, clinopyroxene may predominate. 2.2.3. Ultramylonite The polymineralic matrix of the ultramylonite is very fine-grained (b10 μm, Fig. 4D). The clasts in the matrix are mainly olivine (0.5– 2 mm) with smaller amounts of orthopyroxene (300–700 μm) and clinopyroxene (200–700 μm). Only olivine has recrystallized grains in tails that surround the olivine clasts, parallel to the foliation. The grain size decreases from the center of these lenses (~50 μm) to the border (~ 20 μm). Bands of monomineralic olivine, not continuous on a thin section scale, have a larger grain size (~20 μm) than the matrix. 2.3. Oman The shear zone in Oman is part of the Hilti massif, a segment of the Semail ophiolite. This massif (Ceuleneer et al., 1988; Dijkstra, 2001; Dijkstra et al., 2002a; Michibayashi and Mainprice, 2004; Michibayashi

2.3.2. Mylonites The mylonites consist of alternating layers of polymineralic mixtures and monomineralic olivine that are continuous on a thin section scale. These polymineralic domains are mixtures of orthopyroxene, clinopyroxene, olivine and spinel, with orthopyroxene as the most dominant second phase. The orthopyroxene and clinopyroxene have similar grain sizes in these domains, in contrast to the spinel, which is often very fine-grained (b20 μm). The minerals in the polymineralic domains do not have a shape-preferred orientation. The olivine in the monomineralic domains is larger (~300 μm) than in the polymineralic domains. Late serpentine overprint of olivine (20–30%) affects the monomineralic olivine domains. The serpentine forms a mesh that dissects olivine grains and grain boundaries. The spinel grains form the lineation in the samples and have a variable grain size (0.15– 1 mm). The grain size of orthopyroxene porphyroclasts ranges from 0.5 to 3 mm. They are often elongated and have exsolution lamellae of clinopyroxene. Clinopyroxene does not occur as large porphyroclasts and is only found in the polymineralic domains and as inclusions in the orthopyroxene porphyroclasts. 2.3.3. Ultramylonites Two different ultramylonite types are found in the mantle rocks, which are distinguished based on the amount and arrangement of polymineralic fine-grained material. In the first type, occurring in the Wadi Al Wasit shear zone, the ultramylonite forms 300–400 μm wide bands within the mylonite (Fig. 4F). The grain size of the ultramylonite layers within the mylonite is much smaller compared to the mylonite grain size, and the layers are mostly polymineralic. There is a sharp transition to the mylonite microstructures indicating strongly localized deformation in the ultramylonites (Fig. 4F). This ultramylonite type is similar to the one found in Othris (Fig. 4A). At some locations, the polymineralic ultramylonitic layer dissects orthopyroxene clasts. The

30

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

Fig. 4. (A) Crossed nichols optical photomicrograph of the mylonite and ultramylonite bands of the shear zone in Othris. The ultramylonitic band is around 750 μm thick, there is a sharp boundary between the ultramylonite and mylonite with an abrupt change in grain size. (B) Crossed nichols optical photomicrograph of the Lanzo protomylonite, with coarse-grained monomineralic olivine domains and finer-grained polymineralic domains. There are some relict orthopyroxene clasts (opx). (C) Crossed nichols optical photomicrograph of the Lanzo mylonite showing an elongated orthopyroxene clast. Above the clast is a fine-grained polymineralic layer, and beneath a coarser grained olivine layer. (D) Crossed nichols optical photomicrograph of the Lanzo ultramylonite showing relict sigmoidal-shaped olivine clasts with tails of coarse-grained recrystallized olivine. The polymineralic layers are very fine-grained and no individual grains can be distinguished at this scale. (E) Crossed nichols optical photomicrograph of the Oman porphyroclastic fabric, showing coarse grained olivine, orthopyroxene and spinel clasts. (F) Transmitted light optical photomicrograph of the Oman ultramylonite. There is an abrupt change in grain size between the ultramylonite and host mylonite. The ultramylonitic band is around 200 μm thick.

grain size in the ultramylonites is too small to be determined under the optical microscope. The second ultramylonite type is characterized by a matrix of finegrained polymineralic layers and lenses of stronger mainly monomineralic olivine (see also Michibayashi and Mainprice (2004)). This type is found more to the south of the Hilti massif. In this microstructure the polymineralic ultramylonitic layers are dominant and there is no relict of the mylonitic deformation. This type is similar to the ultramylonites found in Lanzo (Fig. 4D). 3. Methods From each shear zone, thin sections, parallel to the lineation and perpendicular to the foliation, of different tectonites were analyzed (i.e.

porphyroclastic tectonites, (proto)mylonites, ultramylonites) using scanning electron microscopy (Zeiss EVO 50) in combination with Electron backscatter diffraction (EBSD) (DigiViewII and OIM 5.31 software package) and EDS (energy-dispersive X-ray spectroscopy, with an AMETEK Sapphire light element detector). For these analyses an acceleration voltage of 20 kV, a beam current of 10–20 nA and a frame rate of 20 frames/s were chosen. Element maps (Silicon, Aluminium, Chromium and Calcium) and orientation maps were generated from domains in the thin sections with variable second-phase content. The step size depends on the microstructure and ranged from 6 μm (porphyroclastic tectonite) to 1 μm (ultramylonites). These maps were combined in order to trace the grain and phase boundaries. Fig. 5 shows some examples of olivine orientation maps generated with EBSD overlain on a Silicon element map generated by EDS. The

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

resulting phase boundary maps were quantitatively analyzed with the image analysis program image SXM (http://www.liv.ac.uk/~sdb/ ImageSXM/). The area was measured for each grain and based on these data, volume fraction and grain sizes of the olivine and second phases were calculated. For the calculation of the average grain size, the Equivalent Circular Diameter (ECD) was determined for each grain. In case of strongly elongated grains like the olivine in the Othris mylonite, the ECD underestimates the grain size. When the axial ratio is up to 10, the ECD is about 70% smaller than the long axis of the grain. The area weighting of the olivine grain size was used to determine the average grain size, as not to underestimate the large grains. For the average grain size of second phases, the average number weighted ECD was calculated. With these calculations, the dependence of the olivine grain size on the presence of the other minerals (secondary phases)

31

was determined. The second-phase content is described by the Zener parameter (Z) which is defined as the second-phase grain size (dp) divided by the second-phase volume fraction (fp) (Smith, 1948). This parameter has been used in previous studies on olivine and calcite dominant rocks (Herwegh et al., 2005, 2011; Linckens et al., 2011b). The crystallographic preferred orientation (CPO) of olivine was calculated with the software OIM analysis 5.31. For the contour calculations, harmonic binning was used with a maximum expansion of L = 16 within the harmonic calculus and a Gaussian smoothing of 5°. To determine the strength of the fabric the misorientation (M) index was calculated (Skemer et al., 2005), an m-index of one represent a single crystal, zero is a complete random fabric. Chemical analyses were performed with a JEOL JXA 8200 electron microprobe using an acceleration voltage of 15 kV and a beam current of 20 nA. A spot size of ~ 1 μm was chosen to measure small orthopyroxene grains (b250 μm), which have no visible exsolution lamellae. A spot size of 5 μm was used to integrate sub-micron size clinopyroxene exsolution lamellae in large orthopyroxenes. Element peak and background counting time were 20 s and 10 s, respectively, except for Cr (30/15 s), and Na (10/5 s). Natural and synthetic silicates were used as standards. For the temperature calculations two different orthopyroxene geothermometers were used: Ca-in opx (Brey and Koehler, 1990) and Al–Cr in opx (Witt-Eickschen and Seck, 1991). For the calculations, only analyses of the center of recrystallized orthopyroxenes were used. As the Al–Cr in opx geothermometer assumes equilibrium between spinel, olivine and orthopyroxene, it cannot be applied to the plagioclase peridotites of Lanzo. 4. Results 4.1. Olivine and second-phase grain size 4.1.1. Othris The second-phase grain size in Othris decreases from the mylonite (13–30 μm) to the ultramylonites (6–10 μm, Fig. 6A). The grain size of the different second phases is similar, but the spinel grain sizes are the smallest. The dominant second phase is orthopyroxene, clinopyroxenes and spinel represent only small second-phase volume fractions (Fig. 6B) The average second-phase grain size in the mylonite varies from 16 to 21 μm and there is no dependence on the second-phase volume fraction (Fig. 6C). The average second-phase grain size varies more in the ultramylonites (8 to 20 μm) but there is also no relation with secondphase volume fraction (Fig. 6C). The olivine grain size vs. Zener parameter (dp/fp) log–log plot (Fig. 7) shows the dependence of the olivine grain size of Othris mylonite on the second-phase content at small Z (b250 μm). At a Z N250 μm, the olivine grain size increases only slightly (from 175 to 300 μm) with decreasing second-phase content (Fig. 7). The trend line of the polymineralic domains of the Othris mylonite is similar to the one through the Oman mylonite data. In contrast, the olivine grain size in the monomineralic domains is smaller than the Oman grain size. From the Othris ultramylonites only polymineralic domains could be measured (Fig. 4A). The analyzed domains lie on the trend line of the Oman ultramylonite data (Fig. 7).

Fig. 5. Representative olivine orientation maps overlain on a silicon element map, derived from combined EBSD and EDS mapping on the SEM. These maps, in addition to clinopyroxene, orthopyroxene and spinel orientation maps and calcium and chromium element maps were used to trace the grain and phase boundaries.

4.1.2. Lanzo The second-phase grain size of Lanzo peridotites decreases with increasing deformation from protomylonite (24–50 μm), mylonite (7– 11 μm) to ultramylonite (6–9 μm). Spinel displays the smallest and orthopyroxene and plagioclase the largest grain size (Fig. 6A). The dominant second phase is orthopyroxene with variable volume fractions of clinopyroxene and plagioclase (Fig. 6C). The average second-phase grain size variation in the protomylonite (33–58 μm) and ultramylonites (7–12 μm) shows no dependence on the volume fraction, whereas the mylonite second-phase grain size (9–15 μm) decreases slightly with increasing volume fraction (Fig. 6C).

32

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

Fig. 6. Plots showing second-phase grain size and volume fractions in the different tectonites. Data for the Oman shear zone are from Linckens et al. (2011b). (A) Plot of the logarithm of the grain size of the different secondary phases for the three shear zones. The two pyroxenes have similar average grain sizes, whereas spinel usually has a smaller grain size than the pyroxenes. The second phase grain size decreases from porphyroclastic tectonite to ultramylonite. (B) Frequency diagram displaying the volume fraction of the different second phases in the different tectonites. Orthopyroxene is the dominant second phase, followed by clinopyroxene. Spinel is the volumetric smallest second phase. (C) Log–log plot of the second-phase grain size vs. volume fraction (fp). Most tectonites show no relation between second-phase grain size and volume fraction. Some show a slight increase in second-phase grain size with increasing volume fraction. The second-phase grain size decreases from porphyroclastic tectonite to ultramylonite.

The Zener plot (Fig. 7) shows an olivine grain-size decrease in the monomineralic domain from protomylonite (560 μm) to the mylonite (230 μm). The monomineralic olivine lenses of the ultramylonite were not analyzed as the deformation is localized in the polymineralic finegrained matrix. The Lanzo protomylonite olivine grain sizes lie in between the Oman porphyroclast and the Oman mylonite grain size (Fig. 7). The Lanzo mylonite and ultramylonite olivine grain sizes lie between the Oman mylonite and ultramylonite grain sizes. The trend lines of the Lanzo protomylonite and mylonite of the polymineralic data are slightly flatter and steeper than the Oman trend lines, respectively. The trend line through the Lanzo ultramylonites data is parallel to the ones of Oman. 4.1.3. Comparing the three shear zones With increasing localization of deformation (Figs. 1 and 2) we observe a decrease in olivine grain size in all three shear zones, from

porphyroclastic tectonite (400–3000 μm) to mylonite (30–200 μm) and ultramylonite (10–50 μm) (Fig. 7). The olivine grain size is always smaller (up to an order of magnitude) when a large amount of second phases is present (polymineralic), compared to domains with no or small amounts of second phases (volume fraction b 0.05, from here on called monomineralic). The domains where the olivine grain size is strongly dependent on the Zener parameter are referred to as secondphase controlled (Linckens et al., 2011b; inclined lines in Fig. 7). In the case of large Z values, the olivine grain size is only slightly depending on second phases (i.e. subhorizontal lines of Fig. 7). Here the olivine grain size is controlled by dynamic recrystallization. For the quantification of the olivine grain-size dependence on second-phase content we use the Zener relation (e.g., Evans et al., 2001), as previously defined by Herwegh and Berger (2004) and Herwegh et al. (2005):  m Dol ¼ c dp =f p

ð2Þ

where Dol is the olivine grain size (μm), c (μm1 − m) and m are constants, and dp/fp is the ratio of second-phase grain size and volume fraction (μm), defined as the Zener parameter (Z). The three shear zones show roughly the same dependence on Zener parameter in second-phase controlled and dynamically recrystallization controlled domains, where m is 0.61 ± 0.06 and 0.10 ± 0.02, respectively. The derived m value is higher than the ones derived for different carbonate mylonites (0.24–0.45; Ebert et al., 2008; Herwegh et al. 2005) indicating a stronger dependence of the olivine grain size on the presence of second phases than for calcite. The constant c depends on the shear zone and deformation type and varies in the second-phase controlled domain from 9 μm0.39 (Oman porphyroclastic tectonite) to 3 μm0.39 (Oman ultramylonite) and in the dynamically recrystallized domain from 1090 μm0.9 (Oman porphyroclastic tectonite) to 109 μm0.9 (Othris mylonite). 4.2. Olivine CPO Fig. 7. Olivine grain size vs. Z = dp/fp (Zener parameter (Z), second-phase grain size (dp) and volume fraction (fp)) data of the three mantle shear zones on a log–log plot. The data for the Oman shear zone are from Linckens et al. (2011b). A second-phase and dynamic recrystallization controlled domain can be discriminated. The average slopes are 0.61 ± 0.06 and 0.10 ± 0.02 for the second-phase controlled domain and the dynamic recrystallization domain, respectively. Within a given tectonite the variation in olivine grain size can be up to one order of magnitude owing to the effect of second phases. For a given Zener value the olivine grain size decreases from the porphyroclastic tectonite to the ultramylonite. Note that for ultramylonite no dynamic recrystallization controlled domains existed.

Fig. 8 shows the CPO for different tectonites of the three shear zones, the Oman CPOs are from Linckens et al. (2011b) and are added for comparison. In the mylonite of Othris the a-axes are parallel to the lineation, the b-axes lie in the foliation, and are perpendicular to the lineation. The c-axes are parallel to the poles of the foliation. The axis orientations are the same for both low and high Z (i.e. polymineralic and monomineralic), however, the CPO is weaker in the polymineralic domain than in the monomineralic domain (see M-indices in Fig. 8). The Othris ultramylonite has a random CPO.

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

33

Fig. 8. Crystallographic preferred orientations (CPOs) of olivine for all the tectonites at a low and high Zener parameter. The M represents the misorientation index for each CPO, which is a measure for the fabric strength (one is a single crystal and zero is a random fabric). The data for the Oman tectonites is from Linckens et al. (2011b). The CPO in the Othris mylonite is strong at high and low Z values. The pattern strength decreases from high to low Z. The [100] axes are parallel to the lineation, the [001] axes are perpendicular to the foliation. The same fabric is observed in the Oman mylonite (Linckens et al., 2011b). In contrast, the [010] axes are perpendicular to the foliation in the porphyroclastic tectonite of Oman. The Othris ultramylonite shows a random CPO. At high Z, the Lanzo protomylonite and ultramylonite show girdles of [010] and [001] axes. In contrast, the Lanzo mylonite shows clusters of these two axes, and the same pattern as in the Othris and Oman mylonite is observed. The CPO in the Lanzo ultra- and mylonite at low Z values is random.

The Lanzo protomylonite shows a strong CPO in the monomineralic domains (M = 0.29). The a-axis orientations are the same as in the monomineralic Othris mylonite CPO, however, the b- and c-axes form girdles instead of point maxima. In the polymineralic domains of the protomylonite, the CPO is weaker (M = 0.18), and the girdles are discontinuous. The monomineralic domains in the Lanzo mylonite show a clear CPO, however, the CPO is weaker than the CPO of the monomineralic domains in the Othris mylonite. The axes have the same point maxima orientations as in the Othris mylonite. The polymineralic mylonite has the weakest CPO of all analyzed fabrics (M = 0.06). The monomineralic domains in the Lanzo ultramylonite have a relatively strong CPO, with the a-axes parallel to the lineation. The b-axes

form again girdles and the c-axes are relatively random. In the polymineralic domains of the Lanzo ultramylonite, the olivine shows no CPO. 4.3. Deformation temperature calculations The calculated temperatures are displayed in frequency diagrams to compare the temperatures of the three shear zones (Fig. 9), the average temperature calculated with the Ca-in-opx thermometer (Brey and Koehler, 1990) for the different tectonites and shear zones are displayed in Table 1. For the Oman porphyroclastic tectonite we derived a minimum temperature (Linckens et al., 2011a), which is similar as those obtained for the medium temperature deformation microstructures

34

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

Fig. 9. Frequency diagrams of the temperatures calculated with the Ca-in-opx for the three shear zones and Al,Cr-in-opx geothermometer only for Oman and Othris. As Lanzo is a plagioclase lherzolite, the Al,Cr-in-opx geothermometer cannot be applied to these tectonites. The data from the Oman mylonite and ultramylonites are from Linckens et al. (2011a). To determine the deformation temperatures, only the data of small-recrystallized orthopyroxene grains are used. The calculated temperatures for the Othris mylonite are lower than those calculated for the Oman mylonite for both geothermometers. The Oman and Othris ultramylonite have similar calculated temperatures. The tectonites of Lanzo show a decrease in temperature with increasing deformation intensity. The calculated temperatures for the protomylonite are higher than the ones calculated for the Oman mylonite. The calculated temperatures for the Lanzo mylonite are higher than the ones calculated for the Oman mylonite. However, when only the data of orthopyroxene grains near clinopyroxene grains are displayed, the average calculated temperature is lower than in the Oman mylonite (see Table 1). The calculated temperatures for the Lanzo ultramylonite are higher than the calculated temperatures for the Oman ultramylonite. The average calculated temperature for grains smaller than 25 μm is lower than when all grains are included. The average calculated temperature is still higher than the Oman ultramylonite (Table 1).

outside the shear zone (1100 °C) by Gnos and Nicolas (1996). The results of the small grains in the mylonite and ultramylonite in the Oman shear zone (Linckens et al., 2011a) are shown in Fig. 9 for comparison. Both the Ca-in-opx and Al–Cr-in-opx geothermometer give similar and lower temperatures for the Othris ultramylonite and mylonite, compared to the Oman ultramylonite and mylonite, respectively (Fig. 9). The Lanzo tectonites show a decrease in calculated temperature with increasing deformation intensity (i.e. from protomylonite to ultramylonite, Fig. 9). Both the calculated temperatures of mylonite and ultramylonite of Lanzo are lower than the Oman mylonite and ultramylonite. For the Lanzo mylonite only opx grains in the vicinity of cpx grains were used to calculate the

deformation temperature (Fig. 9). There is a correlation between obtained temperature and the interparticle distance between the measured grains and surrounding clinopyroxene. When clinopyroxene grains are in close vicinity of the measured opx grain, the temperatures are lower than when other minerals (e.g. olivine) are surrounding the grain. This observation suggests that for larger diffusion distances between the two pyroxenes, equilibrium was not reached, as calcium diffusion is too slow at low temperatures (800 °C). The data of the Lanzo ultramylonite is split up into two grain-size classes, representing grains above and below 25 μm (Fig. 9). The small grains of the ultramylonites give lower calculated temperatures.

J. Linckens et al. / Tectonophysics 643 (2015) 26–43 Table 1 Average temperatures (°C) calculated with the Ca-in-opx geothermometers for all the tectonites. Ca-in-opx (°C)

Al-Cr in opx (°C)

Othris Mylonite Ultramylonite

779 ± 39 673 ± 41

742 ± 42 687 ± 22

Lanzo Protomylonite Mylonite Ultramylonite

915 ± 41 818 ± 26 740 ± 16

Oman Porphyroclastic tectonitea Myloniteb Ultramyloniteb

1100a 818 ± 46b 681 ± 34b

a b

784 ± 42b 689 ± 20b

Temperature from Gnos and Nicolas (1996). Temperature from Linckens et al. (2011a).

5. Discussion Fig. 3 illustrates qualitatively the microstructural evolution in mantle shear zones with photomicrographs. The microstructural evolution is quantified in the Zener plot, showing the variation of olivine grain size within different tectonites depending on second-phase content (Fig. 7). Furthermore, the Zener plot quantifies the olivine grain-size evolution within the shear zones with increasing deformation and strain localization at decreasing temperatures (Table 1). Fig. 3 also illustrates the occurrence of well-mixed fine-grained layers in the ultramylonites. In the following sections, we will discuss the grain size evolution and the phase mixing in the shear zones combining our Zener data and microstructural observations with previously published microstructure studies on deformed mantle rocks. 5.1. Grain size evolution of monomineralic olivine The steady-state monomineralic olivine grain size is defined by a balance between grain growth and reduction processes during dynamic recrystallization (Herwegh et al., 1997; Means, 1981). Various theoretical and empirical relations for the steady-state grain size of monomineralic minerals have been derived (Austin and Evans, 2007; De Bresser et al., 2001; Herwegh et al., 2014; Karato et al., 1980; Shimizu, 1998; Van der Wal et al., 1993). Deformation experiments on single crystal olivine (Karato et al., 1980) and dunites (Van der Wal et al., 1993) have shown that the dynamically recrystallized olivine grain size is dependent mainly on stress, and an empirical relation between grain size and differential stress has been derived. This dependence of dynamic recrystallized grain size solely on stress has theoretically been determined by Twiss (1977). However, other studies suggest a dependence of the steady state dynamically recrystallized grain size on recrystallization processes (Shimizu, 2008; Stipp et al., 2010), on the balance between dislocation and diffusion deformation mechanisms (i.e. field boundary hypothesis, De Bresser et al. (1998, 2001)), on the balance between nucleation and growth during dynamic recrystallization (Shimizu, 1998) or on the available deformation energy (i.e. paleowattmeter, Austin and Evans (2007)). The field boundary hypothesis is based on the assumption that olivine will grow in the diffusion creep field, and in the dislocation creep/disGBS field the grain size is reduced by dynamic recrystallization. Therefore, the steady state olivine grain size lies on the boundary between the two deformation mechanisms and depends on stress, temperature and strain rate. The theoretical model of Shimizu (1998) predicts a steady state olivine grain size that depends on stress and to a lesser degree on temperature. The paleowattmeter assumes that the grain size evolution is determined by the rate of mechanical work and depends on stress and strain rate. In the case of olivine, the stress calculated with the paleowattmeter is similar to the one calculated with the paleopiezometer (see Fig. 3 from Austin and Evans (2007)),

35

and therefore relatively independent of temperature. Recent microstructural models considering a thermodynamic-based elasto-visco-plastic theoretical background (e.g., Regenauer-Lieb and Yuen, 2004; Regenauer-Lieb et al., 2012) demonstrate that elastically stored energy and dissipation of energy into microstructural work and heat are important for controlling the deformation mechanisms activated. Furthermore, it is important for the associated evolution in dynamically recrystallized grain size (Herwegh et al., 2014) as well as the local generation of heat by microstructural work as well as the positive feedback. The different theoretical and empirical relations indicate that the factors controlling the olivine steady state grain size during dynamic recrystallization are yet to be resolved. In addition, the field observations of decreasing grain size with decreasing temperature in many mylonites suggest that temperature influences the available deformation energy, diffusion rates, slip systems, grain sizes and reactions (White et al., 1980). In the analyzed shear zones, the deformation parameters (i.e. stress, strain and strain rate) change with continuous deformation. The olivine grain size and the widths of the shear zones decrease with ongoing deformation, indicating an increase in stress and strain, respectively. In addition, the detailed temperature analysis clearly shows a decrease in deformation temperature with increasing deformation intensity and smaller grain size (Fig. 9 and Table 1). The relation between grain size and temperature in the studied shear zones could be due to the stress evolution with decreasing temperature (i.e. it follows from olivine flow laws that when the strain rate is constant, the stress increases with decreasing temperature). Olivine grain size in the monomineralic domains of the three shear zones shows the same relation with the calculated temperature (Fig. 10A). This could indicate that the stress or deformation energy evolution with temperature was the same for the three shear zones, or that temperature, in addition to stress and strain rate, has a large influence on the dynamically recrystallized grain size. The paleopiezometer of Van der Wal et al. (1993) can be used to calculate the differential stresses using the grain sizes of the monomineralic, dynamic recrystallization controlled olivine microstructures (Figs. 7 and 10B, Table 2). Note that by this paleopiezometric approach the possible temperature, strain and strain rate effects on the monomineralic olivine grain size are neglected. 5.2. Grain size evolution of polymineralic domains In the case of the mylonites and less deformed tectonites (i.e. porphyroclastic tectonite and protomylonite) we can assume, as a first order approximation, that the stress and strain conditions were similar in the poly- and monomineralic layers as we see no microstructural evidence on the thin section scale for strain localization in the polymineralic layers (Figs. 3 and 4). In the ultramylonites the differential stresses and strains are different in the poly- and monomineralic domains due to strain localization in the polymineralic layers as can be concluded from the microstructures (Figs. 3 and 4A, D and F). In addition, the olivine grain size of the ultramylonitic polymineralic layers depends not only on stress but also on the second-phase content (Fig. 7). Therefore, the calculated stresses are probably too low, because stresses are higher in the finegrained domains due to the strain localization. The theoretical monomineralic olivine of these fine-grained layers would be smaller than the grain size used for the paleopiezometer calculations. The calculated differential stress is thus a lower limit. From our data, we can derive the olivine grain-size variation for a given differential stress in a mantle shear zone as a function of different second-phase contents, but constant deformation temperature (vertical arrow in Fig. 10B). The olivine grain-size variation with second-phase content is largest in the porphyroclastic tectonite (Figs. 7 and 10B). The olivine grain-size variation within a tectonite decreases with increasing stress, i.e. with decreasing deformation temperature. The stippled lines in Fig. 10B show the evolution of the olivine grain size

36

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

Fig. 10. (A) Log log plot of olivine grain size vs. calculated temperature. As the temperature decreases with increasing deformation the plot looks similar to the log log plot of olivine grain size vs. differential stress (B). The monomineralic domains of the tectonites (Z = 104) lie on the same trend line, indicating a similar dependence of olivine grain size on deformation temperature. The dependence of olivine grain size on temperature in the polymineralic domain (Z = 200 μm) is smaller. The vertical spread in data at similar temperatures indicates the olivine grain-size variation within one tectonite due to the presence of second phases. (B) Log–log plot of the olivine grain size vs. the calculated differential stress using the olivine grain size at a Z value of 104 μm and the paleopiezometer of van der Wal et al. (1993). The differential stress is assumed to be the same in the polymineralic domains (at Z values of 200 and 50). The olivine grain size at the different Z values is determined by using the regression lines of Fig. 7. For the ultramylonites, differential stress is calculated using the olivine grain size from the domains that have the lowest amount of second phases (see Fig. 7, Z ~ 200 μm). The vertical arrow indicates the grain size variation within the tectonites at a constant differential stress, due to the presence of second phases. The olivine grain size variation due to the second phases within the porphyroclastic tectonite is the largest and decreases towards the ultramylonites. The inclined arrow underlines the olivine grain-size evolution with increasing differential stress and shear zone evolution. This is also indicated by the stippled lines which shows this evolution in the monomineralic (Z = 104 μm) and polymineralic (Z = 200 and 50 μm) domains. The olivine grain-size decrease with shear zone evolution is larger in the monomineralic domains than in the polymineralic domains. The olivine grain-size decrease with differential stress is similar for both polymineralic domains (i.e. Z = 200 and 50 μm). (C) Log–log plot of the secondphase grain size vs. calculated differential stress. The second-phase grain size decreases with increasing differential stress (differential stresses are calculated with the monomineralic olivine grain size). (D) Log–log plot of second-phase grain size vs. deformation temperature. The second-phase grain size decreases with decreasing calculated deformation temperature.

with increasing deformation at constant second-phased content (i.e. for a Z value of 50, 100 and 104). In the polymineralic domains, the stressdependent olivine grain-size variation from porphyroclastic tectonite to ultramylonites is smaller than in the monomineralic domains as illustrated by the slope of the trend lines (Fig. 10B). With our data we can add a second-phase dependence of the olivine grain size to the paleopiezometer relation (Van der Wal et al., 1993), the derived equation for the olivine grain-size dependence on second phase grain size and stress is: m

Dol ¼ aZ σ

−n

ð3Þ

where a is a constant (27), Z is the ratio between second-phase grain size (dp) and volume fraction (fp), σ is the differential stress in MPa, and m (0.61) and n (−0.5) are constants. With this equation, the olivine grain-size variation due to the second-phase content can be determined for a specific differential stress. The polymineralic olivine grain-size variation is also smaller than the monomineralic olivine grain-size variation when plotted against temperature (Fig. 10A). The second-phase grain-size dependence on differential stress as calculated by olivine piezometry indicates a decrease of second-phase grain size with increasing stress (Fig. 10C). The slope is similar to the one of the olivine monomineralic domain and smaller than for the olivine polymineralic domains (Fig. 10C). This indicates that the stress dependence of the second-phase grain sizes in the polymineralic domains is larger than for olivine grain sizes. We speculate that the recrystallized second-phase grain size in polymineralic layers shows a stronger stress/strain rate controlled component, while

the olivine grain size in these layers is controlled by the interparticle spacing of the second phases, i.e. is geometrically pinned. Both effects change with temperature due to the temperature dependence of rheology and the pinning controlled by coupled grain coarsening/grain size reduction in the polymineralic layers (see Herwegh et al., 2011 and references therein). The relation between olivine grain size and second-phase content indicates a control on olivine grain growth by second phases. However, in order for the second-phases to have an influence on the olivine grain growth, two steps are required: i) the formation of fine-grained olivine, with a grain size smaller than the dynamically recrystallized steadystate olivine grain size and ii) the mixing of second phases and olivine. 5.3. Coupled grain size reduction and phase mixing There are different processes that can lead to grain size reduction: dynamic recrystallization, metamorphic or melt–rock reactions, and cataclasis. In addition, several processes can lead to the mixing of different phases; grain boundary sliding, metamorphic and melt–rock reactions (Fig. 12). In the following section we will discuss the interplay of the different processes in the mantle shear zones and their influence on the grain size evolution during shear zone development. 5.3.1. Dynamic recrystallization, grain nucleation and grain boundary sliding During dynamic recrystallization, three different processes can occur: subgrain rotation, grain boundary bulging and grain boundary migration (Drury and Urai, 1990; Poirier, 1985). Where dynamic

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

37

Table 2 Compilation of the Zener data (grain sizes and volume fractions). Wh - Wadi hilti, Wa - Wasdi al Wasit, La - Lanzo, OK - Othris, Dol - olivine grain size, dp - second-phase grain size, fp - second-phase volume fraction, Z - Zener parameter (dp/fp). Sample

Dol

dp

(μm)

(μm)

Oman Porphyroclastic tectonite wh 3 h 360.7 1869.6 1768.1 wh 3 d 353.1 285.0 3187.2 2137.6 1740.4 Dijkstrab 3920.7 Mylonite wa-7 115.9 109.4 190.7 201.4 283.4 wa-85 77.1 104.1 156.4 224.7 160.3 193.9 298.4 wa-120 78.5 183.6 114.8 262.5 268.6 319.9 Ultramylonite wa-93 12.1 27.6 24.4 37.2 28.7 R5 25.6 42.7 56.6 58.6

Stressa fp

Z (μm)

89.4 770.6 768.5 63.8 50.5 490.4 413.4 457.6 365.0

0.39 0.14 0.13 0.19 0.13 0.06 0.01 0.20 0.01

232.2 5351.5 6099.4 339.1 377.2 7821.1 55337.3 2346.6 37361.3

38.1 18.0 49.1 41.3 47.9 26.8 18.6 50.1 68.2 69.0 17.3 21.3 30.1 83.6 32.0 75.8 50.2 65.2

0.43 0.14 0.16 0.12 0.02 0.41 0.20 0.29 0.33 0.33 0.03 0.01 0.32 0.42 0.13 0.19 0.07 0.07

88.6 126.1 317.0 350.3 1954.9 65.8 92.1 173.9 208.0 212.2 557.2 3127.0 94.8 199.1 244.5 392.7 683.3 946.2

Sample

(MPa) Lanzo Protomylonite La-18 6 6

4 6 4

Mylonite La-19

25 Ultramylonite La-11

34 24

27 23

Othris Mylonite OK73A Ok74A

9.6 12.0 11.8 13.2 11.0 12.0 15.3 15.5 15.3

0.44 0.42 0.37 0.24 0.15 0.30 0.16 0.11 0.07

22.0 28.8 31.6 55.1 74.5 39.9 95.3 144.3 208.5

83

Ultramylonite Ok73B Ok74A OK74B a b

Stressa

Dol

dp

(μm)

(μm)

187.2 210.4 253.1 202.2 313.1 631.2 560.4

53.6 48.4 58.2 33.4 50.0 41.9 43.1

0.35 0.31 0.27 0.14 0.14 0.03 0.02

153.3 156.2 215.5 238.2 367.3 1550.0 2535.3

14 15

28.6 28.3 96.8 110.5 136.7 253.5 225.7 218.1

10.4 9.5 13.7 17.6 17.1 114.0 94.0 105.9

0.50 0.32 0.11 0.11 0.08 0.07 0.03 0.02

20.7 29.3 124.5 160.3 214.0 1583.3 3133.7 4527.4

28 30 31

28.4 40.2 29.3 25.5 43.6 49.6 93.8

8.5 7.1 12.5 8.8 9.4 9.9 9.6

0.34 0.21 0.34 0.24 0.23 0.11 0.05

25.0 34.1 36.3 37.0 40.9 94.2 180.6

58

92.5 296.7 50.1 52.9 60.5 107.2 142.7 156.8 175.6 197.5 219.6 183.2 347.8 168.7

23.1 10.8 18.0 16.5 21.9 23.0 23.6 21.7 18.6 20.3 25.2 21.0 26.3 18.6

0.26 0.0002 0.43 0.38 0.36 0.28 0.20 0.14 0.05 0.02 0.02 0.01 0.004 0.001

88.9 54444.4 42.2 43.9 60.3 80.9 119.2 157.2 381.9 1240.2 1317.7 2184.4 7461.8 14904.0

29.1 41.2 15.6 13.9

12.7 20.2 8.8 8.0

0.25 0.36 0.41 0.41

51.5 55.7 42115 42113

fp

Z (μm)

(MPa)

25

36 33 31 35 22 38

Differential stress is calculated only for the monomineralic domains with the equation from (Van der Wal et al., 1993). Grain sizes and volume fraction calculated from tracings of Fig. 7 from Dijkstra et al. (2002a).

recrystallization of olivine and orthopyroxene clasts can be observed in the analyzed shear zones, mainly evidence for subgrain rotation recrystallization (Fig. 11A) is found (Dijkstra et al., 2002a, 2002b). This recrystallization mechanism has also been observed in other shear zones (Poirier and Nicolas, 1975) and in deformation experiments (Bystricky, 2000; Farla et al., 2013; Linckens et al., 2014). Therefore, we will focus the discussion on this dynamic recrystallization process. During subgrain rotation recrystallization the grain boundary is relatively stationary and no mixing of phases is expected to occur (Fig. 12). However, grain boundary migration can be activated after the formation of new grains (Poirier, 1985). Moreover, a certain component of grain boundary migration is necessary to preserve a steady state microstructure with progressive deformation even under conditions of subgrain rotation recrystallization as dominant deformation mechanism (e.g. Haertel and Herwegh (2014); Herwegh et al. (1997)). In order for second phases to have an effect on the olivine grain size during grain boundary migration, mixing has to occur during or after the

subgrain rotation recrystallization and before extensive grain boundary migration and grain growth. If we ignore melt–rock and metamorphic reactions, phase mixing can only occur by grain boundary sliding. During grain boundary sliding, grains slide past each other which induces neighbor switching (Ashby and Verrall, 1973) as this has been observed in deformed, dynamically recrystallized quartz (Halfpenny et al., 2012) and in material science (Kashyap et al., 1985, and references therein). Although incipient phase mixing was observed in recent deformation experiments on coarse grained olivine–orthopyroxene mixtures, extensive mixing might need large strain perturbations (Linckens et al., 2014). There is also evidence from tensile deformation experiments on very fine grained (b500 nm) forsterite bearing mixtures, that grain boundary sliding leads to phase aggregation instead of phase mixing (Hiraga et al., 2010b, 2013). In contrast, high strain experiments on olivine–opx aggregates show extensive mixing of the two phases due to grain boundary sliding (Farla et al., 2013). In addition, as shown by rock analogue experiments (Ree, 1991) and suggested for the

38

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

grain boundaries in the Lanzo and Othris mylonite (Fig. 11A, B) could indicate that some grain boundary sliding occurred in the mylonites. This alignment of grain boundaries was previously observed in a finegrained olivine–orthopyroxene band in a mylonite by Dijkstra et al. (2002b). These authors suggested that grain boundary sliding occurred in these bands. After the mixing of phases due to grain boundary sliding combined with phase nucleation, grain growth of the olivine grain size will be inhibited by the secondary phases, whereas the grain size can grow unrestricted until a steady state grain size in the monomineralic olivine domains. This can lead to the large variation in olivine grain size observed in the tectonites (Figs. 7, 12). However, in order for olivine grain growth to occur after dynamic recrystallization, the newly formed grains must be smaller than the olivine steady-state grain size. The large grain size spread in the porphyroclastic tectonite, which experienced small amounts of strain, is unlikely to be the result of the coupled process of dynamically recrystallization and grain boundary sliding. This suggest that grain boundary sliding was not the initial phase mixing process in the case of the studied mantle peridotites. Therefore, melt–rock or metamorphic reactions must have occurred as phase mixing process during high temperature deformation in the mantle shear zones.

Fig. 11. (A) Crossed nichols optical micrograph of a fine-grained polymineralic layer in the Othris mylonite showing aligned grain boundaries (white arrows) indicating grain boundary sliding. The scale bar in all microphotographs is 100 μm. (B) Crossed nichols optical micrograph of a fine-grained polymineralic layer in the Lanzo mylonite with aligned grain boundaries (white arrow), indicating grain boundary sliding. (C) Crossed nichols optical micrograph of a recrystallized orthopyroxene clast in the protomylonite of Lanzo. The neoblasts are formed by subgrain rotation recrystallization. In between the newly formed grains small sized olivine grains occur. Note that there is no recrystallized olivine clast in the vicinity implying that the small olivine grain nucleated at this site.

generation of new mica grains as second phases in calcite mylonites (Herwegh and Jenni, 2001), grain boundary sliding generates voids particularly at grain triple junctions. Under the presence of fluids, pore fluid pressure in the new voids drops and new phases can be precipitated/ nucleated. Once nucleated, the second phases are capable to influence the microstructure of the matrix phase by pinning of grain boundaries (Berger et al., 2010). The link between subgrain rotation and an increase in grain boundary sliding has been postulated in the past due to CPO weakening with progressive subgrain rotation recrystallization (e.g. Bestmann and Prior (2003); Stipp and Kunze (2008)). The aligned

5.3.2. Melt–rock reactions Melt–rock reactions are observed in Othris, where in the tectonites adjacent to the mylonites a fine-grained mixture of olivine and orthopyroxene is observed around orthopyroxene porphyroclasts (Dijkstra et al., 2002b). This observation indicates that low-Si melt percolated through the peridotite that reacted with the orthopyroxene porphyroclasts to form olivine and a high-Si melt (Dijkstra et al., 2002b). The fine-grained regions surrounding the opx porphyroclasts become more stretched with increasing deformation. Furthermore, Dijkstra et al. (2002b) also found fine-grained orthopyroxene in recrystallized olivine aggregates that they suggest to have precipitated from a melt. Dijkstra et al. (2002b) concluded that the melt–rock reactions lead to fine-grained mixtures that can deform by diffusion creep. These fine-grained mixtures remain small due to the hindering of olivine grain growth by the second phases. As in Othris, melt–rock reactions occurred in the Lanzo shear zone, which lead to grain size reduction and phase mixtures resulting in strain localization (Kaczmarek and Müntener, 2008). In the Oman shear zone it has been inferred that melt–rock interaction formed finer-grained orthopyroxene and olivine (Dijkstra et al., 2002a). The detailed thin section study of the Oman shear zone also shows the fine-grained mixtures originate around orthopyroxene clasts, which get more elongated with ongoing deformation (Linckens et al., 2011b). At some locations, small irregular shaped olivine grains occur between recrystallized orthopyroxene grains (Fig. 11C). The irregular grain shape, the small amount of olivine grains and the fact that no relict olivine porphyroclast is observed indicate that these olivine grain were not formed by subgrain rotation recrystallization but by melt–rock reactions. As previously stated, the porphyroclastic tectonite shows the largest olivine grain-size variation (Fig. 7). We suggest that this initial (at low strain) large difference between the olivine grain size in mono- and polymineralic domains is related to melt–rock reactions. Another indication for this process is the data gap in Zener space, between relatively monomineralic olivine grain size (high Z) and fine-grained polymineralic olivine grain sizes, in contrast to the continuous (ultra)mylonite data (Fig. 7). This gap indicates that not a gradual phase mixing process occurred in the porphyroclastic tectonites but one that only formed fine-grained polymineralic domains. The melt–rock reactions occur before shear zone development and the deformation occurs at a solid state in the three shear zones (Dijkstra et al., 2002a, 2002b; Kaczmarek and Müntener, 2008). During subsequent deformation, dynamic recrystallization will occur and

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

39

Fig. 12. Schematic model of the microstructural evolution during different processes that could have occurred in the mantle rocks (modified from Piazolo and Passchier (2002)). The three rows show the grain size reduction and phase mixing during ongoing deformation combining dynamic recrystallization with grain boundary sliding (upper row), metamorphic reactions (middle row) and melt–rock reactions (bottom row). The upper row shows a first step of grain size reduction by dynamic recrystallization due to subgrain rotation. This leads to monomineralic mixtures of neoblasts around relict porphyroclasts. The steady state grain size at equal differential stress of orthopyroxene is smaller than the olivine grain size (Farla et al., 2013; Linckens et al., 2014). However, for simplicity we used similar grain sizes for olivine and orthopyroxene. After this first step of dynamic recrystallization grain boundary sliding promoting displacements of grains relative to each other has to occur to generate polymineralic mixtures. During these processes no olivine grain size variation with second-phase content is expected, as no pinning occurred (the grain boundaries are relatively stationary). After the phase mixing by grain boundary sliding a phase of grain growth must occur, for instance due to a change in deformation conditions, that in turn changes the steady state olivine grain size. The olivine in the monomineralic layers is able to grow to a larger grain size resulting in a grain size variation. The middle row shows the microstructural evolution during retrograde metamorphic reactions with ongoing deformation. Dynamic recrystallization and heterogeneous nucleation form polymineralic fine-grained domains around orthopyroxene porphyroclasts. These domains are increasingly elongated with increasing deformation. No grain boundary sliding is necessary in order to obtain fine-grained phase mixtures. The bottom column shows the microstructural evolution when melt–rock reactions take place before deformation. The mantle rocks are already heterogeneous before deformation due to melt–rock reactions. Fine-grained mixtures can form around orthopyroxene porphyroclasts and small sized orthopyroxene may form within monomineralic olivine domains. During deformation, dynamic recrystallization of both phases leads to finer grained mixtures. With increasing strain and ongoing deformation, grain boundary sliding combined with dynamic recrystallization is expected, leading to fine-grained polymineralic layers.

orthopyroxene can hinder the growth of the olivine grains. This results in olivine grain sizes smaller than the steady state recrystallized olivine grain size. An additional mixing process (e.g. metamorphic reactions and/or grain boundary sliding) is required when solid state deformation continues under retrograde conditions (see Fig. 12). 5.3.3. Metamorphic reactions Metamorphic reactions in mantle shear zones have been proposed to lead to stress weakening (de Ronde et al., 2005; Hidas et al., 2013; Newman et al., 1999; Toy et al., 2010). The mineral reactions can occur due to decompression of the peridotites during uplift, leading to the reaction of spinel to plagioclase (Hidas et al., 2013; Newman et al., 1999). These reactions lead to fine-grained mixtures of olivine, pyroxenes, plagioclase ± spinel. In addition, during shear zone development, temperature controls the composition of the mantle minerals that can participate in metamorphic reactions (Linckens et al., 2011a; Toy et al., 2010). The reaction products are finer grained and have a different chemical composition compared to host rock assemblages. In contrast to the melt–rock reactions, the reactions can be continuous and occur throughout the shear zone evolution because the mantle minerals are solid solutions (Linckens et al., 2011a; Newman et al., 1999; Toy et al., 2008). Oman and Othris are both spinel peridotites but spinel did not react to plagioclase during decompression. In the Lanzo shear zone, the reaction of spinel to plagioclase did not form polymineralic mixtures

around spinel porphyroclasts, as observed by Newman et al. (1999). The detailed chemical study for the different tectonites in the three shear zones does show a change in chemical composition with increasing deformation intensity as reflected in the temperature calculations (Fig. 9 and, Linckens et al., 2011a). The heterogeneous nucleation of metamorphic new phases leads to fine-grained polyphase mixtures (Fig. 12, de Ronde et al., 2004; Kruse and Stünitz, 1999). Although at the time this nucleation occurs, grain boundary sliding is not necessarily required to mix the phases, it will enhance it by providing the nucleation sites in voids between the sliding grains (Herwegh and Jenni, 2001) In this way, retrograde grain size reduction in olivine by subgrain rotation with an increasing grain boundary sliding component will (i) result in smaller olivine grain sizes, providing smaller interparticle distances with grain boundary voids as preferential nucleation site for second phases. (ii) Consequently, the interparticle distances of newly nucleated second phases decrease with decreasing temperature. (iii) Particularly non-isotropic stress distributions manifest by stress concentrations around mechanically rigid second-phase grains (e.g. Mancktelow (2011)) might favor spatial variations in sizes of dynamically recrystallized olivine grains. Combined with chemical driving forces for recrystallization, this behavior can explain, how in polymineralic layers, the olivine grain size can become much smaller than the theoretical steady state grain size of a monomineralic olivine layer, accounting for a continuous grain size decrease with decreasing temperatures.

40

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

5.4. Summary In contrast to the monomineralic olivine grain size, the polymineralic olivine grain size depends not only on the deformation conditions but also on the second-phase content. The three shear zones show the same relation between olivine grain size and second-phase content even though they were formed at different geodynamic settings. Based on the microstructures of the three shear zones we propose that a first stage of phase mixing in the mantle rocks already occurred by melt– rock reactions before deformation. During ongoing retrograde deformation, heterogeneous nucleation of second phases during continuous metamorphic solid-state net transfer reactions results in fine-grained polymineralic mixtures. In addition, grain boundary sliding is likely to be an important supportive mechanism in forming fine-grained wellmixed layers in the low temperature mylonite and ultramylonite. We suggest that all three phase mixing process are important in the mantle shear zones but at different stages of the shear zone evolution. 5.5. CPO, deformation mechanisms and strain localization The olivine CPO indicates which deformation mechanism was active. In the case of dislocation creep and disGBS a CPO develops, in contrast to when diffusion creep is the dominant deformation mechanism. During deformation in the disGBS regime the type of fabric depends on the amount of strain and changes from girdles to clusters of [010] and [001] axes with increasing strain and a steady state is reached only at high strain (γ N 10) (Hansen et al., 2014). The girdles seen in the Lanzo protomylonite could indicate relatively low strain deformation during deformation in the disGBS regime. The other CPO patterns in the monomineralic domains of the three mylonites indicate that dislocation or disGBS creep was active and [100] (001) was the active slip system. This subsidiary slip system, found in other natural shear zones and in experiments, is thought to be related either to (i) an enhanced water content (Katayama et al., 2004; Mackwell et al., 1985; Skemer et al., 2010), (ii) low deformation temperatures (Carter and Ave'Lallemant, 1970) or (iii) pre-existing anisotropies (Michibayashi and Mainprice, 2004; Warren et al., 2008). For the Lanzo and Othris shear zones, the CPO in the monomineralic domains is stronger than in the polymineralic domains (see M-indexes in Fig. 8). This decrease in CPO strength from mono- to polymineralic domains is related to a higher dislocation density because of slower

cycles of dynamic recrystallization and dynamic recovery due to the pinning by second phases. The pinning promotes diffusion creep as strain accommodation mechanism in addition to dislocation creep in these polymineralic domains (Ebert et al., 2007). The ultramylonites of all the three shear zones and the polymineralic domain of the Lanzo mylonite show a weak CPO indicating diffusion creep as dominant deformation mechanism. The monomineralic domain in the Lanzo ultramylonite shows a CPO indicating disGBS or dislocation creep as dominant deformation mechanism. The weak CPOs suggest a dominance of diffusion creep as deformation mechanism in the ultramylonites. When the second-phase content decreases the CPO becomes more pronounced (i.e. see CPO in the Lanzo ultramylonite). The pinning of olivine by second-phases in the ultramylonites keeps the olivine grains in the diffusion creep stability field. The resulting low flow stresses lead to strain localization in the ultramylonitic domains. In the mylonites this switch in deformation mechanism seems to occur only in the Lanzo mylonite. Here the CPO for polymineralic olivine is weak, in contrast to the strong CPO in the monomineralic layers. The Lanzo mylonite shows the smallest olivine grain sizes of the three mylonites (Fig. 7) meaning that the switch in deformation mechanism occurs only at low temperatures (around 800 °C, Table 1) and only when the olivine grain size is relatively small in the polymineralic layers (b40 μm, Fig. 7). In addition to using the CPO to determine the dominant deformation mechanisms, we can apply olivine flow laws and generate olivine deformation mechanism maps (Fig. 13). In the first approach, we use the monomineralic olivine grain size and the deformation temperature as a constraint and the strain rate is set constant at 10−13 s−1 (Fig. 13A). The differential stresses are not determined with the paleopiezometer, but determined from the flow laws. The stresses increase from protomylonite to ultramylonites. The vertical extent of the boxed fields of Fig. 13A is due to the estimated error in temperature measurement. Othris ultramylonite gives unrealistically high differential stresses when the low temperature range is used (i.e. N 500 MPa). Note that following this approach, all the tectonites must be deformed by the dislocation-accommodated grain boundary sliding process (Hansen et al., 2011) that is slightly grain size dependent. In the second approach, the differential stresses are calculated with the olivine paleopiezometer (Van der Wal et al., 1993). Furthermore, the temperature is assumed constant at 800 °C. In this case all the tectonites, just as in the first approach, all lie in the disGBS field

Fig. 13. Olivine deformation mechanism maps based on the flow laws of Hansen et al. (2011) and Hirth and Kohlstedt (2003) showing the diffusion creep, disGBS and dislocation creep fields. (A) Deformation mechanism map using the calculated deformation temperatures and the range in olivine grain sizes for the protomylonite to ultramylonites. The strain rate is set at a constant 10−13 s−1, the differential stresses are derived from the flow laws (not from paleopiezometery). All tectonites lie in the disGBS field, the stress range derived for each tectonite is related to the 2σ error in the temperature calculations. The differential stresses increase from protomylonite to ultramylonite. When using the average temperature and lower estimates for the Othris ultramylonite, unrealistic high stresses are derived using these assumptions. (B) For this deformation mechanism map a constant deformation temperature (800 °C) is assumed for all the tectonites. The differential stress is calculated with the paleopiezometer (Van der Wal et al., 1993). In this case the tectonites also lie in the disGBS field.

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

(Fig. 13B). However, in the ultramylonites no monomineralic layers are analyzed. Therefore, the differential stress calculated with the paleopiezometer is probably too high as this olivine grain size is smaller than the corresponding dynamic recrystallization controlled olivine grain size. Shifting the range of grain sizes to smaller stresses will lead the smaller grain sizes to be in the diffusion creep field (Fig. 13B). Both approaches suggest that disGBS is the dominant deformation mechanism, which corresponds to the strong CPOs measured in the mylonites (Fig. 8). As previously stated the disGBS deformation mechanism is slightly grain size dependent. Therefore, due to the range in olivine grain sizes, strain localization should occur in the polymineralic layers of all the mylonites if disGBS is the dominant deformation mechanism. This was also previously modeled by Precigout and Gueydan (2009) and Precigout et al. (2007) using the disGBS flow laws from Hirth and Kohlstedt (2003). However, olivine is not the only phase in the polymineralic layers and it has been shown in deformation experiments that the viscosity of polyphase rocks depends on the volume fraction and viscosity of all the phases (e.g., Ji et al., 2001, 2003; Tasaka et al., 2013; Tullis et al., 1991). At the experimental conditions during deformation experiments of olivine–opx mixtures, olivine was the stronger phase (Ji et al., 2001; Tasaka et al., 2013). In contrast, clinopyroxene is expected to be the strongest phase in the polyphase mixture (Bystricky and Mackwell, 2001; Dimanov and Dresen, 2005). In order to derive the viscosity of the polymineralic layers the deformation mechanism, appropriate flow laws and volume fraction of each phase needs to be determined in future studies. As we see no strain localization in the polymineralic layers in the (proto)mylonites (Figs. 3B and 4A, B and C), second phases at these deformation conditions not only have a large effect on the olivine grain size but might also effect the total viscosity of these polymineralic layers. In this case, the strengths of second phases result in a viscosity difference between poly- and monomineralic layers small enough so no strain localization develops in the polymineralic layers. Another possibility is that dislocation creep is the dominant deformation mechanism in the (proto)mylonites. This could be the case when deformation occurred under “wet” conditions, as there is no evidence yet from deformation experiments that disGBS is an important olivine deformation mechanism at hydrous conditions. The combined analysis of CPO and deformation mechanism maps shows that there are many uncertainties, and it is difficult to deduce the strain rate and stress during the progressive deformation at retrograde conditions of these natural shear zones.

6. Conclusions Detailed microstructural studies of three different shear zones formed at different geodynamic settings show that the olivine grain size depends on the second-phase content. This dependence is the same in the three shear zones and is shifted to smaller olivine grain sizes with decreasing temperature and increasing stress and strain (Fig. 7). The monomineralic grain size evolution of olivine in the shear zones can be related to differential stress and decreasing temperatures, while in the case of polymineralic layers, phase mixing and pinning of the olivine grain size by the second phases is crucial. A first stage of phase mixing occurs by melt–rock reactions before deformation. The polymineralic grain-size evolution is then controlled by dynamic recrystallization and metamorphic reactions (phase nucleation) during deformation at retrograde conditions. In a last step, ongoing deformation, metamorphic reaction (phase nucleation) and possibly grain boundary sliding lead to fine-grained well-mixed layers that localize strain in the ultramylonites. The Zener plot (Fig. 7) can be used to determine the conditions (e.g. T, Z, stress) where a switch in olivine deformation mechanism in polymineralic mantle rock can be expected, which in turn leads to strain localization. The microstructural study underlines the importance of

41

second phases for the switch from dislocation-accommodated to diffusion creep at temperatures lower than 800 °C. Our data indicate that second phases only play a subordinate role for strain localization at deformation temperatures exceeding 800 °C. At these higher temperatures, the olivine in both poly- and monomineralic domains deforms by grain size insensitive deformation mechanisms (e.g. dislocation controlled) or by dislocation-accommodated grain boundary sliding. disGBS is a grain size sensitive deformation mechanism (Hansen et al., 2011; Warren and Hirth, 2006) and strain localization due to the olivine grain size variation is expected (Precigout et al., 2007). However, our microstructural observations indicate no strain localization, suggesting that if disGBS was the dominant deformation mechanism, the secondary phases have an important effect, not only on the olivine grain size but also on the overall polymineralic viscosity. These phases might keep the viscosity difference between mono- and polymineralic layers small enough to hinder strain localization. Other parameters like overall energy available to deform a given rock volume (Regenauer-Lieb and Yuen, 2004; Regenauer-Lieb et al. 2012), the inheritance of mechanical anisotropies (Michibayashi and Mainprice, 2004; Tommasi et al., 2009; Vauchez et al., 1998), or the presence of water or melts (Dijkstra, 2001; Mei and Kohlstedt, 2000a) need to be considered to explain strain localization in the upper mantle at elevated temperatures. Acknowledgments This research was financially supported by the Swiss National Science Foundation (200021-113563, 200020-126560, 200021103479/1, 200021-109369), JL acknowledges support by NSF EAR0911289 during finalizing of the paper. We thank Ivan Mercolli for help in the field, Arjan Dijkstra for samples, and Alfons Berger, Bénédicte Cenki-Tok and Martin Robyr for assisting with microprobe analyses. The Ministry of Commerce and Industry of Oman are gratefully acknowledged for their help in organizing the fieldwork in Oman. An anonymous reviewer and Frederic Gueydan are thanked for their detailed and helpful comments. References Ashby, M.F., Verrall, R.A., 1973. Diffusion-accommodated flow and superplasticity. Acta Metall. 21, 149–163. Austin, N.J., Evans, B., 2007. Paleowattmeters: a scaling relation for dynamically recrystallized grain size. Geology 35, 343. Bai, Q., Kohlstedt, D.L., 1992. High-temperature creep of olivine single crystals, 2. Dislocation structures. Tectonophysics 206, 1–29. Bai, Q., Mackwell, S.J., Kohlstedt, D.L., 1991. High-temperature creep of olivine single crystals 1. Mechanical results for buffered samples. J. Geophys. Res. 96, 2441–2463. Barth, M.G., 2003. Geochemistry of the Othris Ophiolite, Greece: evidence for refertilization? J. Petrol. 44, 1759–1785. Barth, M.G., Gluhak, T.M., 2009. Geochemistry and tectonic setting of mafic rocks from the Othris Ophiolite, Greece. Contrib. Mineral. Petrol. 157, 23–40. Barth, M.G., Mason, P.R.D., Davies, G.R., Drury, M.R., 2008. The Othris Ophiolite, Greece: a snapshot of subduction initiation at a mid-ocean ridge. Lithos 100, 234–254. Bercovici, D., Ricard, Y., 2012. Mechanisms for the generation of plate tectonics by twophase grain-damage and pinning. Phys. Earth Planet. Inter. 202-203, 27–55. Berger, A., Brodhag, S.H., Herwegh, M., 2010. Reaction-induced nucleation and growth v. grain coarsening in contact metamorphic, impure carbonates. J. Metamorph. Geol. 28, 809–824. Bestmann, M., Prior, D.J., 2003. Intragranular dynamic recrystallization in naturally deformed calcite marble: diffusion accommodated grain boundary sliding as a result of subgrain rotation recrystallization. J. Struct. Geol. 25, 1597–1613. Bodinier, J.L., 1988. Geochemistry and petrogenesis of the Lanzo peridotite body, western Alps. Tectonophysics 149, 67–88. Boudier, F., 1978. Structure and petrology of the Lanzo peridotite massif (Piedmont Alps). Geol. Soc. Am. Bull. 89, 1574–1591. Boudier, F., Ceuleneer, G., Nicolas, A., 1988. Shear zones, thrusts and related magmatism in the Oman ophiolite: initiation of thrusting on an oceanic ridge. Tectonophysics 151, 275–296. Brey, G.P., Koehler, T., 1990. Geothermobarometry in four-phase lherzolites II. New thermobarometers, and practical assessment of existing thermobarometers. J. Petrol. 31, 1353–1378. Bystricky, M., 2000. High shear strain of olivine aggregates: rheological and seismic consequences. Science 290 (80-), 1564–1567. Bystricky, M., Mackwell, S., 2001. Creep of dry clinopyroxene aggregates. J. Geophys. Res. 106, 13443–13454.

42

J. Linckens et al. / Tectonophysics 643 (2015) 26–43

Carter, N.L., Ave'Lallemant, H., 1970. High temperature flow of dunite and peridotite. Geol. Soc. Am. Bull. 81, 2181–2202. Ceuleneer, G., Nicolas, A., Boudier, F., 1988. Mantle flow patterns at an oceanic spreading center — the Oman peridotites record. Tectonophysics 151, 1–26. Cordier, P., Demouchy, S., Beausir, B., Taupin, V., Barou, F., Fressengeas, C., 2014. Disclinations provide the missing mechanism for deforming olivine-rich rocks in the mantle. Nature 507, 51–56. De Bresser, J.H.P., De Peach, C.J., Reijs, J.P.J., Spiers, C.J., 1998. On dynamic recrystallization during solid state flow: effects of stress and temperature. Geophys. Res. Lett. 25, 3457–3460. De Bresser, J., Ter Heege, J., Spiers, C.J., 2001. Grain size reduction by dynamic recrystallization: can it result in major rheological weakening? Int. J. Earth Sci. 90, 28–45. De Ronde, A.A., Heilbronner, R., Stünitz, H., Tullis, J., 2004. Spatial correlation of deformation and mineral reaction in experimentally deformed plagioclase–olivine aggregates. Tectonophysics 389, 93–109. De Ronde, A.A., Stünitz, H., Tullis, J., Heilbronner, R., 2005. Reaction-induced weakening of plagioclase–olivine composites. 409, 85–106. Dijkstra, A.H., 2001. Deformation and Melt in Natural Mantle Rocks: The Hilti Massif (Oman) and the Othris Massif (Greece). Utrecht University. Dijkstra, A.H., Drury, M.R., Frijhoff, R.M., 2002a. Microstructures and lattice fabrics in the Hilti mantle section (Oman Ophiolite): evidence for shear localization and melt weakening in the crust–mantle transition zone? J. Geophys. Res. 107, 1–18. Dijkstra, A.H., Drury, M.R., Vissers, R.L.M., Newman, J., 2002b. On the role of melt–rock reaction in mantle shear zone formation in the Othris Peridotite Massif (Greece). J. Struct. Geol. 24, 1431–1450. Dimanov, A., Dresen, G., 2005. Rheology of synthetic anorthite–diopside aggregates: implications for ductile shear zones. J. Geophys. Res. 110, B07203. Drury, M.R., Urai, J.L., 1990. Deformation-related recrystallization processes. Tectonophysics 172, 235–253. Ebert, A., Herwegh, M., Evans, B., Pfiffner, A., Austin, N., Vennemann, T., 2007. Microfabrics in carbonate mylonites along a large−scale shear zone (Helvetic. Alps): Tectonophysics. 444 (1–4), 1–26. http://dx.doi.org/10.1016/j.tecto.2007.07.004. Ebert, A., Herwegh, M., Berger, A., Pfiffner, A., 2008. Grain coarsening maps for polymineralic carbonate mylonites: a calibration based on data from different Helvetic nappes (Switzerland). Tectonophysics 457, 128–142. Evans, B., Renner, J., Hirth, G., 2001. A few remarks on the kinetics of static grain growth in rocks. Int. J. Earth Sci. 90, 88–103. Farla, R.J.M., Karato, S.-I., Cai, Z., 2013. Role of orthopyroxene in rheological weakening of the lithosphere via dynamic recrystallization. Proc. Natl. Acad. Sci. U. S. A. 110, 16355–16360. Gleason, G.C., Tullis, J., 1995. A flow law for dislocation creep of quartz aggregates determined with the molten salt cell. Tectonophysics 247, 1–23. Gnos, E., Nicolas, A., 1996. Structural evolution of the northern end of the Oman Ophiolite and enclosed granulites. Tectonophysics 254, 111–137. Haertel, M., Herwegh, M., 2014. Microfabric memory of vein quartz for strain localization in detachment faults: a case study on the Simplon fault zone. J. Struct. Geol. 68, 16–32. Halfpenny, A., Prior, D.J., Wheeler, J., 2012. Electron backscatter diffraction analysis to determine the mechanisms that operated during dynamic recrystallisation of quartzrich rocks. J. Struct. Geol. 36, 2–15. Hansen, L.N., Zimmerman, M.E., Kohlstedt, D.L., 2011. Grain boundary sliding in San Carlos olivine: flow law parameters and crystallographic-preferred orientation. J. Geophys. Res. 116, 1–16. Hansen, L.N., Zhao, Y.-H., Zimmerman, M.E., Kohlstedt, D.L., 2014. Protracted fabric evolution in olivine: implications for the relationship among strain, crystallographic fabric, and seismic anisotropy. Earth Planet. Sci. Lett. 387, 157–168. Herwegh, M., Berger, A., 2004. Deformation mechanisms in second-phase affected microstructures and their energy balance. J. Struct. Geol. 26, 1483–1498. Herwegh, M., Jenni, A., 2001. Granular flow in polymineralic rocks bearing sheet silicates: new evidence from natural examples. Tectonophysics 332, 309–320. Herwegh, M., Handy, M.R., Heilbronner, R., 1997. Temperature- and strain-ratedependent microfabric evolution in monomineralic mylonite: evidence from in situ deformation of norcamphor. Tectonophysics 280, 83–106. Herwegh, M., Berger, A., Ebert, A., 2005. Grain coarsening maps: a new tool to predict microfabric evolution of polymineralic rocks. Geology 33, 801. Herwegh, M., Linckens, J., Ebert, A., Berger, A., Brodhag, S.H., 2011. The role of second phases for controlling microstructural evolution in polymineralic rocks: a review. J. Struct. Geol. 33, 1728–1750. Herwegh, M., Poulet, T., Karrech, A., Regenauer-Lieb, K., 2014. From transient to steady state deformation and grain size: a thermodynamic approach using elasto-viscoplastic numerical modeling. J. Geophys. Res. Solid Earth 119, 900–918. Hidas, K., Garrido, C.J., Tommasi, A., Padron-Navarta, J.A., Thielmann, M., Konc, Z., Frets, E., Marchesi, C., 2013. Strain localization in pyroxenite by reaction-enhanced softening in the shallow subcontinental lithospheric mantle. J. Petrol. 54, 1997–2031. Higgie, K., Tommasi, A., 2014. Deformation in a partially molten mantle: constraints from plagioclase lherzolites from Lanzo, western Alps. Tectonophysics 1–15. Hiraga, T., Miyazaki, T., Tasaka, M., Yoshida, H., 2010a. Mantle superplasticity and its selfmade demise. Nature 468, 1091–1094. Hiraga, T., Tachibana, C., Ohashi, N., Sano, S., 2010b. Grain growth systematics for forsterite ± enstatite aggregates: effect of lithology on grain size in the upper mantle. Earth Planet. Sci. Lett. 291, 10–20. Hiraga, T., Miyazaki, T., Yoshida, H., Zimmerman, M.E., 2013. Comparison of microstructures in superplastically deformed synthetic materials and natural mylonites: mineral aggregation via grain boundary sliding. Geology 41, 959–962. Hirth, G., Kohlstedt, D.L., 1995. Experimental constraints on the dynamics of the partially molten upper mantle, 2. Deformation in the dislocation creep regime. J. Geophys. Res. 100, 15411–15449.

Hirth, G., Kohlstedt, D.L., 2003. Rheology of the upper mantle and the mantle wedge: a view from the experimentalists. Inside the Subduction Factorypp. 83–105. Ji, S., Wang, Z., Wirth, R., 2001. Bulk flow strength of forsterite–enstatite composites as a function of forsterite content. Tectonophysics 341, 69–93. Ji, S., Zhao, P., Xia, B., 2003. Flow laws of multiphase materials and rocks from endmember flow laws. Tectonophysics 370, 129–145. Kaczmarek, M.-A., Müntener, O., 2008. Juxtaposition of melt impregnation and hightemperature shear zones in the upper mantle; field and petrological constraints from the Lanzo Peridotite (Northern Italy). J. Petrol. 49, 2187–2220. Kaczmarek, M.−A., Tommasi, A., 2011. Anatomy of an extensional shear zone in the mantle, Lanzo massif, Italy: Geochemistry, Geophysics. Geosystems 12 (8). http:// dx.doi.org/10.1029/2011GC003627. Karato, S., 1989. Grain growth kinetics in olivine aggregates. 168, 255–273. Karato, S., Toriumi, M., Fujii, T., 1980. dynamic recrystallization of olivine single crystals during high-temperature creep. Geophys. Res. Lett. 7, 649–652. Kashyap, B.P., Arieli, A., Mukherjee, A.K., 1985. Review microstructural aspects of superplasticity. J. Mater. Sci. 20, 2661–2686. Katayama, I., Jung, H., Karato, S., 2004. New type of olivine fabric from deformation experiments at modest water content and low stress. Geology 32, 1045. Kruse, R., Stünitz, H., 1999. Deformation mechanisms and phase distribution in mafic hightemperature mylonites from the Jotun Nappe, southern Norway. Tectonophysics 303, 223–249. Linckens, J., Herwegh, M., Müntener, O., 2011a. Linking temperature estimates and microstructures in deformed polymineralic mantle rocks. Geochem. Geophys. Geosyst. 12, 1–19. Linckens, J., Herwegh, M., Müntener, O., Mercolli, I., 2011b. Evolution of a polymineralic mantle shear zone and the role of second phases in the localization of deformation. J. Geophys. Res. 116, 1–21. Linckens, J., Bruijn, R.H.C., Skemer, P., 2014. Dynamic recrystallization and phase mixing in experimentally deformed peridotite. Earth Planet. Sci. Lett. 388, 134–142. Mackwell, S.J., Kohlstedt, D.L., Paterson, M.S., 1985. The role of water in the deformation of olivine single-crystals. J. Geophys. Res. Solid Earth 90, 1319–1333. Mancktelow, N.S., 2011. Deformation of an elliptical inclusion in two-dimensional incompressible power-law viscous flow. J. Struct. Geol. 33, 1378–1393. Mcdonnell, R.D., Peach, C.J., Van Roermund, H.L.M., Spiers, C.J., 2000. Effect of varying enstatite content on the deformation behavior of fine-grained synthetic peridotite under wet conditions. J. Geophys. Res. 105, 13535–13553. Means, W.D., 1981. The concept of steady-state foliation. Tectonophysics 78, 179–199. Mei, S., Kohlstedt, D.L., 2000a. Influence of water on plastic deformation of olivine aggregates 1. Diffusion creep regime. J. Geophys. Res. 105, 21457–21469. Mei, S., Kohlstedt, D.L., 2000b. Influence of water on plastic deformation of olivine aggregates 2. Dislocation creep regime. J. Geophys. Res. Solid Earth 105, 21471–21481. Menzies, M., Allen, C., 1974. Plagioclase lherzolite–residual mantle relationships within two eastern Mediterranean ophiolites. Contrib. Mineral. Petrol. 45, 197–213. Michibayashi, K., Mainprice, D., 2004. The role of pre-existing mechanical anisotropy on shear zone development within oceanic mantle lithosphere: an example from the Oman ophiolite. J. Petrol. 45, 405–414. Michibayashi, K., Gerbert-Gaillard, L., Nicolas, A., 2000. Shear sense inversion in the Hilti mantle section (Oman ophiolite) and active mantle uprise. Mar. Geophys. Res. 21, 259–268. Newman, J., Lamb, W.M., Drury, M.R., Vissers, R.L.M., 1999. Deformation processes in a peridotite shear zone: reaction-softening by an H2O-deficient, continuous net transfer reaction. Tectonophysics 303, 193–222. Nicolas, A., Boudier, F., Ildefonse, B., Ball, E., 2000. Accretion of Oman and United Arab Emirates ophiolite — discussion of a new structural map. Mar. Geophys. Res. 21, 147–179. Ohuchi, T., Nakamura, M., 2007. Grain growth in the forsterite–diopside system. Phys. Earth Planet. Inter. 160, 1–21. Olgaard, D.L., Evans, B., 1986. Effect of second-phase particles on grain growth in calcite. J. Am. Ceram. Soc. 69, C-272–C-277. Piazolo, S., Passchier, C.W., 2002. Controls on lineation development in low to medium grade shear zones: a study from the Cap de Creus peninsula, NE Spain. J. Struct. Geol. 24, 25–44. Poirier, J.P., 1985. Creep of Crystals: High-temperature Deformation Processes in Metals, Ceramics and Minerals. Cambridge University Press. Poirier, J.P., Nicolas, A., 1975. Deformation-induced recrystallization due to progressive misorientation of subgrains, with special reference to mantle peridotites. J. Geol. 83, 707–720. Precigout, J., Gueydan, F., 2009. Mantle weakening and strain localization: implications for the long-term strength of the continental lithosphere. Geology 37, 147–150. Precigout, J., Gueydan, F., Gapais, D., Garrido, C., Essaifi, A., 2007. Strain localisation in the subcontinental mantle — a ductile alternative to the brittle mantle. Tectonophysics 445, 318–336. Rassios, A., Smith, A.G., 2000. Constraints on the formation and emplacement age of western Greek ophiolites (Vourinos, Pindos, and Othris) inferred from deformation structures in peridotites: ophiolites and oceanic crust: new insights from field studies and Ocean Drilling Program. GSA Spec. Publ. 349, 473–483. Ree, J.H., 1991. An experimental steady-state foliation. J. Struct. Geol. 11, 1001–1011. Regenauer-Lieb, K., Yuen, D., 2004. Positive feedback of interacting ductile faults from coupling of equation of state, rheology and thermal-mechanics. Phys. Earth Planet. Inter. 142, 113–135. Regenauer-Lieb, K., Weinberg, R.F., Rosenbaum, G., 2012. The role of elastic stored energy in controlling the long term rheological behaviour of the lithosphere. J. Geodyn. 55, 66–75.

J. Linckens et al. / Tectonophysics 643 (2015) 26–43 Rybacki, E., Dresen, G., 2000. Dislocation and diffusion creep of synthetic anorthite aggregates. J. Geophys. Res. 105, 26017–26036. Shimizu, I., 1998. Stress and temperature dependence of recrystallized grain size: a subgrain misorientation model. Geophys. Res. Lett. 25, 4237–4240. Shimizu, I., 2008. Theories and applicability of grain size piezometers: the role of dynamic recrystallization mechanisms. J. Struct. Geol. 30, 899–917. Skemer, P., Katayama, I., Jiang, Z., Karato, S., 2005. The misorientation index: development of a new method for calculating the strength of lattice-preferred orientation. Tectonophysics 411, 157–167. Skemer, P., Warren, J.M., Kelemen, P.B., Hirth, G., 2010. Microstructural and rheological evolution of a mantle shear zone. J. Petrol. 51, 43–53. Smith, C., 1948. Grains, phases, and interfaces an interpretation of microstructure. Trans. Am. Inst. Min. Metall. Eng. 175, 15–51. Stipp, M., Kunze, K., 2008. Dynamic recrystallization near the brittle–plastic transition in naturally and experimentally deformed quartz aggregates. Tectonophysics 448, 77–97. Stipp, M., Tullis, J., Scherwath, M., Behrmann, J.H., 2010. A new perspective on paleopiezometry: dynamically recrystallized grain size distributions indicate mechanism changes. Geology 38, 759–762. Sundberg, M., Cooper, R.F., 2008. Crystallographic preferred orientation produced by diffusional creep of harzburgite: effects of chemical interactions among phases during plastic flow. J. Geophys. Res. 113, 1–16. Tasaka, M., Hiraga, T., Zimmerman, M.E., 2013. Influence of mineral fraction on the rheological properties of forsterite + enstatite during grain-size-sensitive creep: 2. Deformation experiments. J. Geophys. Res. Solid Earth 118, 3991–4012. Tasaka, M., Hiraga, T., Michibayashi, K., 2014. Influence of mineral fraction on the rheological properties of forsterite + enstatite during grain size sensitive creep 3: application of grain growth and flow laws on peridotite ultramylonite. J. Geophys. Res. Solid Earth http://dx.doi.org/10.1002/2013JB010619.

43

Tommasi, A., Knoll, M., Vauchez, A., Signorelli, J.W., Thoraval, C., Logé, R., 2009. Structural reactivation in plate tectonics controlled by olivine crystal anisotropy. Nat. Geosci. 2, 423–427. Toy, V.G., Prior, D.J., Norris, R.J., 2008. Quartz fabrics in the Alpine Fault mylonites: influence of pre-existing preferred orientations on fabric development during progressive uplift. J. Struct. Geol. 30, 602–621. Toy, V.G., Newman, J., Lamb, W., Tikoff, B., 2010. The Role of pyroxenites in formation of shear instabilities in the mantle: evidence from an ultramafic ultramylonite, Twin Sisters Massif, Washington. J. Petrol. 51, 55–80. Tullis, T.E., Horowitz, F.G., Tullis, J., 1991. Flow law of polyphase aggregates from endmember flow laws. J. Geophys. Res. 96, 8081–8096. Twiss, R.J., 1977. Theory and applicability of a recrystallized grain size paleopiezometer. Pure Appl. Geophys. 115. Van der Wal, D., Chopra, P., Drury, M.R., Fitz Gerald, J.D., 1993. relationships between dynamically recrystallized grain size and deformation conditions in experimentally deformed olivine rocks. Geophys. Res. Lett. 20, 1479–1482. Vauchez, A., Tommasi, A., Barruol, G., 1998. Rheological heterogeneity, mechanical anisotropy and deformation of the continental lithosphere. Tectonophysics 296, 61–86. Warren, J.M., Hirth, G., 2006. Grain size sensitive deformation mechanisms in naturally deformed peridotites. Earth Planet. Sci. Lett. 248, 438–450. Warren, J.M., Hirth, G., Kelemen, P.B., 2008. Evolution of olivine lattice preferred orientation during simple shear in the mantle. Earth Planet. Sci. Lett. 272, 501–512. White, S.H., Burrows, S.E., Carreras, J., Shaw, N.D., Humphreys, F.J., 1980. On mylonites in ductile shear zones. J. Struct. Geol. 2, 175–187. Witt-Eickschen, G., Seck, H.A., 1991. Solubility of Ca and Al in orthopyroxene from spinel peridotite: an improved version of an empirical geothermometer. Contrib. Mineral. Petrol. 106, 431–439.