| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 44 | Number 5 | Pages 901-927 | 2003
© Oxford University Press 2003
Conditions of Formation and Crystallization Kinetics of Highly Oxidized Pseudotachylytes from the High Tatras (Slovakia)
AN PLAIENKA1
1 GEOLOGICAL INSTITUTE OF THE SLOVAK ACADEMY OF SCIENCES, DÚBRAVSKÁ 9, PO BOX 106, 84005 BRATISLAVA, SLOVAKIA
2 DEPARTMENT OF GEOLOGICAL SCIENCES, UNIVERSITY OF MISSOURI–COLUMBIA, COLUMBIA, MO 65211, USA
E-mail: geolpetr{at}savba.sk
RECEIVED FEBRUARY 7, 2002; ACCEPTED NOVEMBER 20, 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Tectonic activity during the Miocene exhumation of the Tatra granitoid basement resulted in frictional melting of granite. The activity marks the early stages of the faulting that is responsible for uplift of the High Tatras. As indicated by pre-existing cataclasite metamorphic mineral assemblages, the ambient pressure was about 250–300 MPa, corresponding to depths between 10 and 12 km. The pseudotachylytes are Fe rich, highly oxidized and crystalline. The matrix composition suggests disequilibrium partial melting of a biotite-dominated assemblage. Oxygen isotopic compositions of a pseudotachylyte sample and its constituent minerals show equilibration with the host granodiorite and allow for the introduction of oxidizing external water rather than oxidation as a result of the dissociation of free water liberated during melting. The kinetic information extracted from hematite crystal-size distributions (CSDs) that are preserved in feldspathic matrices shows that crystals in most places were accumulated while the system was open. The melt was highly mobile and prone to strong differentiation. The hematite crystals reach a maximum of 26 vol. % (45 wt % Fe2O3). In rare places where the flow ceased, the system became closed and produced distinct CSDs. The longest apparent crystallization times (90 s) are recorded mostly in pools in the central parts of the pseudotachylytes whereas the shortest times (10 s) come from rims and tips of fractures. The estimated hematite growth rate was about five orders of magnitude higher than that of ilmenite in lava lakes. Such extreme crystallization rates result from high undercoolings associated with high cooling rates. Very high cooling rates are promoted by the extremely high surface/volume ratios of the pseudotachylyte sheets.
KEY WORDS: pseudotachylyte; hematite; kinetics; High Tatra; Slovakia
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Pseudotachylytes are a common response of a brittle rock to frictional heating associated with fault-slip during an earthquake (Philpotts, 1964
; Sibson, 1975; Maddock, 1983). They often occur in brittle and brittle–ductile zones of deformation and fracturing where they are invariably associated with cataclasite or mylonite. Fault-related pseudotachylytes have been recognized in many exhumed orogenic belts, including the Caledonides, Alps, and Appalachians.
Most previous studies of pseudotachylytes have focused on their tectonic associations, structure, or geochemistry (Reimold, 1991; Maddock, 1992; O'Hara, 1992). Only relatively few have discussed the mineralogy and pressure–temperature (P–T) conditions of their formation (Spray, 1992; Thompson & Spray, 1996; O'Hara & Sharp, 2001). Here we report on several petrological aspects of well-crystallized, hematite-bearing pseudotachylyte veins that occur in a cataclasite within a two-mica granodiorite in the Tatra Mountains of the Western Carpathians. The petrological and kinetic information is derived from both pseudotachylyte matrices and cataclasite mineral assemblages. Friction-induced melts have hematite crystal populations in various textural contexts that allow extraction of unique kinetic information. We show by quantitative modelling that, although melts have cooled rapidly, the times of solidification must have been long enough to allow the hematite crystal populations to develop steady-state crystal size distributions (CSDs) in an open-system crystallization mode. The open-system crystallization resulted in strong differentiation of the primary melt, which is therefore rarely preserved. The differentiation requires a high melt mobility. The highly oxidized nature of the pseudotachylyte is attributed to interaction with oxidized external water rather than to dissociation of the locally produced water.
|
GEOLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The Western Carpathians are part of the trans-European Alpine orogenic system formed by Meso-Cenozoic collisional processes. The central Western Carpathians south of the Pieniny Klippen Belt are correlative with the Austroalpine system of the Eastern Alps (Fig. 1a). They represent a Cretaceous thrust stack composed of three northward-verging, thick-skinned, crustal-scale sheets, including the Tatric, Veporic and Gemeric superunits from the north to the south, and several cover nappe systems (e.g. Pla
ienka et al., 1997). Variscan crystalline basement complexes of the Tatric superunit are exposed in several so-called ‘core mountains’ in northern and western Slovakia. These Late Tertiary mountain ranges have asymmetrical horst and domal horst structures (see Ková
et al., 1994) that are surrounded by Tertiary basins. The highest of these ranges are the Tatra Mts.
|
The crystalline basement of the Tatra Mts (Fig. 1b) is composed of pre-Mesozoic rocks that belong to two tectonic units separated by a Variscan thrust fault (Kahan, 1969
; Janák, 1994). The lower tectonic unit is composed of medium-grade mica schists, whereas the upper unit is composed of high-grade metamorphic rocks and granites (Janák et al., 1996, 1999; Hurai et al., 2000). The basement rocks are overlain by Mesozoic and Cenozoic sedimentary cover sequences and nappes.
Major granite magmatism took place in late Devonian and Carboniferous times (
370–315 Ma), according to Rb–Sr (Gaw
da, 1995) and single-zircon U–Pb data (Poller et al., 2000; Poller & Todt, 2001). In the High Tatra Mountains, the granitoid rocks form a sheet-like pluton. Their compositions range from leucogranite to biotite tonalite and amphibole diorite. Pseudotachylytes occur in a two-mica, S-type granodiorite (Kohút & Janák, 1994; Petrík et al., 1994), one of the dominant granite types in the area. 40Ar–39Ar ages of biotite and muscovite from the granitoids and the metamorphic rocks (330–300 Ma), obtained by both step-heating (Maluski et al., 1993) and laser ablation (Janák, 1994), record mostly cooling and uplift during late Variscan time. Apatite fission-track measurements from the Tatra granitoids yield ages between 20 and 10 Ma (Burchart, 1972; Král', 1977). These data were used to infer the Early to Middle Miocene age of the Tatra Mts uplift (e.g. Ková et al., 1994).
Exhumation of the Tatra Mts and the age of the pseudotachylyte-forming event
The exhumation mechanism of the Tatra crystalline core is not precisely known. Because the Tatra horst is asymmetrical (Mesozoic cover and nappe units occur on the northern slopes only, where they are moderately to steeply north-dipping), the main fault system controlling uplift is considered to have been the southern, steeply dipping, sub-Tatra boundary fault system (Fig. 1b). The system consists of several cross-cutting, arcuate segments that trend east–west to NE–SW. Because the sub-Tatra fault system is largely covered by Quaternary fluvio-glacial deposits, its kinematics are poorly understood. On the basis of the very complex pattern of Neogene block and paleostress field rotations postulated for other segments of the Western Carpathians (e.g. Bac-Moszaszwili, 1993; Marko et al., 1995), the sub-Tatra fault was probably reactivated several times during the Neogene as an oblique-reverse, strike-slip and normal fault system. In the vicinity of the Tatra horst, the reconstructed paleostress field was dominated by vertical flattening during the Late Paleogene and by roughly north–south contraction during the Early Miocene (Sperner, 1996), during the time of the presumed maximum rock uplift and exhumation rates of the Tatra basement. The relative surface uplift and topographic expression was accentuated by the Late Miocene NW–SE extension and Quaternary glaciation of the Tatra Mts.
Plaienka et al. (2001) applied the concept of ‘basement-involved compressive structures’ of Narr & Suppe (1994) to explain the uplift of the Tatra horst. The model is in good agreement with geometrical constraints. According to this model, elevation of the Tatra Mts may have occurred above a south-verging contractional, upward-steepening detachment fault that branches off the ductile–brittle transition within the crust and reaches the surface as the steep sub-Tatra fault bordering the Tatra horst in the south. The studied pseudotachylytes are exposed within the Variscan basement of the Tatra horst in the hanging wall of the sub-Tatra fault. Assuming that the total thickness of the Mesozoic cover and nappe complexes was at least 5 km, and the Paleogene sediments 3–4 km, the inferred pre-Miocene burial depth of the pseudotachylyte-hosting granitoid rocks was 11–12 km (Plaienka et al., 2001), i.e. slightly above the base of the seismogenic crustal layer in a normal geothermal gradient. This inferred depth corresponds well to P–T estimates based upon thermobarometric studies of mineral assemblages (see subsequent discussion).
The granitoid core of the High Tatra Mts is crosscut by numerous brittle to brittle–ductile, Alpine-age shear zones. In the area of the High Tatra pseudotachylyte occurrences, two basic fabric elements are present. Moderately to steeply south-dipping, ductile– brittle shear zones include various proto-mylonitic types; these are indistinctly foliated rocks that are obviously the oldest deformation structures in the area. Low-temperature steps in the 40Ar/39Ar spectra of white mica from the mylonitic zones indicate their paleo-Alpine, latest Cretaceous age (75–66 Ma, Maluski et al., 1993). Therefore, these mylonites are probably related to Late Cretaceous nappe stacking, and not to the later exhumation of the Tatra basement. The mylonitic zones are crosscut by a younger, rather irregular system of brittle fractures, slickensides and cataclastic zones. Pseudotachylyte veinlets are associated with subvertical, roughly north–south-trending dislocations, which may be interpreted as tear or transfer faults associated with the main sub-Tatra fault. Accordingly, we infer that the pseudotachylytes were probably generated by seismic slip along subsidiary faults that accompanied the early stages of the development of the large-scale detachment fault in the proximity of the ductile–brittle transition in the crust. Based on preliminary 40Ar/30Ar dating of a pseudotachylyte matrix by laser probe, which yielded ages between 36 and 24 Ma (Kohút & Sherlock, 2002), a Late Oligocene age is presumed for the initiation of thrust activity along the sub-Tatra fault.
|
PETROGRAPHY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Host rock
The host rock for the pseudotachylytes is a medium-grained two-mica granodiorite composed of plagioclase, quartz, K-feldspar, biotite and muscovite (Kohút & Janák, 1994
). The undeformed rock is slightly porphyritic with euhedral plagioclase (50 vol. %) and K-feldspar (10–15 vol. %) that reach 5–10 mm in size. Plagioclase (An26–32 in cores) is partly or wholly sericitized and commonly contains late clinozoisite. Poikilitic, perthitic K-feldspar phenocrysts enclose small, euhedral plagioclase and biotite crystals aligned with growth zones. The only mafic phase is sub- to anhedral biotite (5–12 vol. %). It is dark brown to pale yellow and commonly shows incipient chloritization along cleavage planes that occasionally include epidote. Slightly pleochroic muscovite (2 vol. %) forms large flakes intergrown with biotite. Both micas commonly enclose prismatic apatite, zircon and monazite. Euhedral ilmenite is the only oxide. Slight cataclasis is ubiquitous, especially in quartz, which shows undulose extinction and recrystallization, and in microcline.
Pseudotachylyte–host-rock relationships
Pseudotachylyte veins occur either closely associated with fractured wall-rock, cataclasite, or as injections in less intensely comminuted granodiorite. The pseudotachylytes cut pre-existing mylonites at high angles, suggesting that the brittle, breccia-forming event post-dated a ductile shearing event.
Injection veins that sharply cut the granodiorite are represented by sample Pt-222 (Figs 2a and 3a). The main vein is 3–7 mm thick with thin, <1 mm, branches. There is a 1–2 cm thick zone of altered granodiorite along either side of the vein. Within this zone, biotite is replaced by chlorite, quartz is strongly undulose or recrystallized, and feldspars are commonly fractured, strained or comminuted. This zone is thought to have formed during interaction with fluid liberated from the crystallizing pseudotachylyte.
|
|
A system of tension fractures anastomosing from the pseudotachylyte generation surface in the hanging wall of a fault is represented by sample Pt-226 (Fig 2b). Zones of cataclastic granodiorite of 1–3 cm thickness are developed along the generation and fracture planes. The tension fractures, 4–10 cm long, are filled with clast-laden melt (Fig. 3b) up to their tips, which narrow to <0·2 mm (Fig. 3c). A thick (>10 cm) breccia zone invaded by wavy, irregular pseudotachylyte veins is represented by sample Pt-650 (Figs 2c and 3d). The completely crushed granitoid rock within at least 10 cm from the vein has been transformed to a cohesive breccia. The vein has a wavy contact with the cataclasite, where it swells from about 2 to 20 mm. Thin, 1–2 mm, wavy veins branch from the main vein and reach a few tens of centimetres in length.
Pseudotachylytes
Three pseudotachylyte samples were studied in detail. The Pt-222 vein cutting granodiorite has the lowest contents of hematite (1–8%, average 3·4 vol. %, Table 1) of all three pseudotachylyte samples. Its crystals have grown to slightly larger sizes than those in the two other samples (apparent lengths 3–9 µm, maximum 25 µm). The vein displays two types of matrices that differ both in clast and hematite contents. A lighter matrix (Figs 3a and 4a) is preserved mostly along vein boundaries forming protrusions at contacts with chloritized biotite (Fig. 3a, ‘bt’). This matrix is characterized by a low content of clasts. A later, darker, clast-laden melt invaded the former one, occasionally with the clasts filtered out where the veins narrow (Fig. 3a, ‘f’). Chlorite and calcite occur as rare secondary phases (Fig. 4a and b).
|
|
Hematite contents in the matrices of sample Pt-226 vary widely, depending on the position in a tension fracture (Figs 3b and c, 4c–f). The hematite contents become larger with increasing distance from the fault plane (Fig. 3c, Table 1). This suggests that more Fe-rich melts were efficiently extracted into more distant tension cracks up to their very tips, where hematite reaches the highest concentrations (Fig. 4c). The hematite size ranges on average between 2 and 8 µm (maximum 16–20 µm). Other matrix minerals are albite, quartz and K-feldspar. Rare phengite occurs as a late phase.
Pseudotachylyte Pt-650 (Fig. 3d) is a clast-laden vein with irregular boundaries where the melt forms embayments with the fine-grained cataclasite matrix. The matrix is composed of finely dispersed hematite (length 2–5 µm) together with anhedral alkali feldspars of similar size. The subangular clasts are usually larger than melt minerals (Fig. 4g), but the smallest clasts may be reduced to sizes comparable with those of newly crystallized feldspar. The modal amount of hematite ranges from 11 to 29 vol. %. The feldspars are dominated by albite (see Table 1). Phengitic muscovite, chlorite and epidote occur as late minerals (Fig. 4 h).
In summary, the matrices of all samples consist of three primary phases: albite, K-feldspar, and hematite, accompanied by minor quartz. Whereas the feldspars and quartz show anhedral grain shapes (10–80 µm in size), the hematite is probably tabular (acicular in 2D), forming subhedral prisms (tablets) 1–30 µm in apparent length. Some hematite crystals have skeletal forms with small voids within the prisms. Phengite, chlorite, epidote and calcite occur as late phases. No zircon was found in the matrix.
Clasts
Clasts comprise a substantial part of the pseudotachylytes (0–60 vol. %). They are angular to subangular and range in size from 5–8 mm down to the size of matrix minerals (<0·01 mm). They comprise mainly lithic fragments, plagioclase, K-feldspar and minor quartz. Plagioclase is the same as in the cataclasite, sodic (An1·3–3·1) with low potassium content (Or0·4–3). Ilmeno-hematite and apatite, 0·4–1 mm in size, originating from the granite, also occur among the clastic minerals. Clast sizes measured in Pt-226 (Fig. 4f) obey a power-law distribution with a fractal dimension of 1·3–1·5 (see Shimamoto & Nagahama, 1992).
|
SAMPLING STRATEGY AND ANALYTICAL METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Pebbles and boulders of pseudotachylyte were first observed in the area of Batizovské Lake on the southern slopes of the High Tatras (Petrík & Reichwalder, 1996
). There are scattered outcrops along a path between the top of Such
Mount and Klin Peak over a distance of 3·5 km. The pseudotachylyte occurrences (Fig. 1c) are concentrated at three locations: (1) Batizovské Lake; (2) Batizovská valley; (3) SW slope of Klin Peak. Fault zones marked by strong cataclasis are apparent in all these locations. The faults strike NNE with steep dip of 75° both to ESE and WNW. The pseudotachylytes form networks with vein thickness usually not exceeding 1–2 cm. However, pockets of pseudotachylyte up to 4 cm thick were found in 20 cm pebbles.
The studied samples are represented by Pt-222, a pebble from the top of Such Mount (Fig. 2a). Pt-226, also a pebble, is from Such Mount (Fig. 2b). Pt-650 is from the Batizovské Lake locality, in situ (Fig. 2c). Six whole-rock samples from these two localities were studied, each including a pseudotachylyte, the adjacent cataclasite, and the granite host rock. Samples (0·5 kg) were powdered and analysed by ACME analytical laboratories by the inductively coupled plasma emission (ICP) method for major and trace elements. The pseudotachylyte veins were separated from the crushed rock by hand picking, and the concentrate was then comminuted to <0·1 mm and purified in an isodynamic magnetic separator at 1 A current. The non-magnetic fractions comprise feldspar concentrates. The magnetic fractions were further subdivided in bromoform (
= 2·9 g/cm3). The heavy fractions were hematite concentrates and the light fractions chlorite concentrates. Hematite and chlorite were pulverized in an agate mortar under acetone and their Fe3+/Fe2+ ratios were determined by Mössbauer spectrometry (Slovak Technical University, Department of Nuclear Physics).
Pseudotachylyte matrices were analysed in polished thin sections using a JEOL JCXA-733 electron microprobe at the Slovak Geological Survey in Bratislava using a standardized energy dispersive spectrometry (EDS) mode and raster beams (15 kV accelerating voltage). The areas analysed contained 0–50% of recognizable clasts. At least 20 areas were analysed per thin section. The oxide totals varied between 94 and 98%, not accounting for the high ferric iron and H2O contents. Comparison of microprobe analyses with analyses of bulk pseudotachylyte (5·81 and 5·96 wt % Na2O for bulk and average matrix compositions, respectively) indicates no detectable Na2O loss. Hematite was analysed using a CAMECA SX-100 electron microprobe by wavelength-dispersive spectrometry (WDS) at the Slovak Geological Survey. Additional mineral analyses were obtained by WDS using the JEOL JXA-8600 electron microprobe at the Mineralogisch-Petrographisches Institut, University of Basel. The operating conditions were 15 kV and 10 nA; natural and synthetic standards were used, and the data were reduced using a ZAF correction routine.
Separated mineral fractions of sample Pt-222 and its host granite were analysed for their oxygen isotope composition to determine its equilibration temperature and to see if there was any evidence for interaction with an external fluid source. A fresh granite sample, a cataclastic granite sample and the bulk (matrix + clasts) pseudotachylyte (86% quartz plus feldspars, 5% chlorite, 9% hematite) and its feldspar-rich fraction (65% plagioclase, 25% K-feldspar, 7% quartz, 3% hematite), chlorite-rich fraction (70% chlorite, 20% hematite, 10% quartz), and hematite-rich fraction (70% hematite, 20% chlorite, 10% quartz) were also analysed. Oxygen from the samples was extracted using ClF3 and converted to CO2 on a conventional extraction line using the technique of Clayton & Mayeda (1963) before analysis on the Finnigan Mat Delta-E mass spectrometer at the University of Missouri. The precision on 
18O values is better than ±0·2
. The values are reported relative to SMOW.
Image analysis
All interpretations of crystallization kinetics of the pseudotachylytes are based on measurements of the length of the major axis of hematite, a parameter commonly chosen as the most characteristic crystal property (e.g. Cashman & Ferry, 1988; Cashman & Marsh, 1988; but see Higgins, 2000). Digital backscattered electron (BSE) imagery was performed using a JCXA-733 electron microprobe in Bratislava. The UTHSCSA Image Tool v. 2 was used to generate binary images from the backscattered images and to measure crystal properties. The size of images varied depending on the amount and size of clasts present, most frequently between 0·04 and 0·12 mm2 (total range from 1·2 to 0·01 mm2). Numbers of measured crystals varied between 83 and 1878, with a typical range of 300–600 crystals per image. Numbers below 300 resulted either from low hematite density domains or from a very small measured area (as a result of very high hematite density). The low numbers were used only for comparative purposes. The smallest measurable crystal size was 0·5 µm (Fig. 4e). Matrices with very high number densities show that the hematite crystals commonly form clusters, which the software interpreted as single crystals. Therefore, a ‘declustering’ procedure, similar to the ‘decoalescing’ procedure developed for vesicle size measurement in pumice clasts (e.g. Klug & Cashman, 1994), was used to separate single crystals within apparent clusters. The clustered crystals were separated by lines of minimum thickness (1 pixel). Although the procedure is subjective, it produces more consistent CSDs (Fig. 5).
|
2D to 3D conversion
The areal (2D) data are most simply converted to volume (3D) data using the simple equation
 | (1) |
v and a are numbers of crystals in the volume and the area, respectively. This conversion is commonly used in CSD analysis (e.g. Cashman & Marsh, 1988; Cashman & Ferry, 1998). Some investigators use the unfolding technique of Saltikov (1967) for spherical particles, mainly for bubble-size distributions (Sarda & Graham, 1990; Cashman & Mangan, 1994) but also for CSDs (Armienti et al., 1994). The technique employs calculated probabilities of sphere intersections (cut section effect), and the relation
 | (2) |
and references therein]. Higgins (1994) and Sahagian & Proussevitch (1998) showed that intersection probabilities for prisms and tablets—the more commonly occurring crystal shapes—are very different from those of spheres. The prisms are most likely to be intersected through their intermediate dimension. Sahagian & Proussevitch (1998) calculated the intersection probabilities for various parallelopipeds and ellipsoids. Higgins (2000) has developed a software tool for CSD corrections (available from http://www.uqac.uquebec.ca/mhiggins/CSD.html) in which this technique is employed. This program was used here for reducing the raw data obtained by image analysis. The procedure, however, obliterates some features of the CSDs converted by equation (1) (see the section ‘Interpretation of the CSDs’). |
MAJOR ELEMENT GEOCHEMISTRY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Major element compositions were obtained for bulk-rock samples and pseudotachylyte matrices. Compositions of the bulk pseudotachylyte, the adjacent cataclasite, and the host granite were obtained for samples Pt-222 and Pt-650. The analyses are given in Table 2. Compared with the leucocratic cataclasites with compositions dominated by alkali feldspar (± quartz and chlorite), the pseudotachylytes are more mafic with high ferric iron contents. The high Fe2O3 content is reflected by the presence of hematite and is among the most significant features of the pseudotachylytes. Another characteristic of the pseudotachylyte compositions is their depletion in CaO, MgO, MnO, and H2O+ relative to the host granite. Because of the high amount of clasts, microprobe analyses of selected domains were obtained using a rastered beam in EDS mode (Table 2). Clast-free domains are found only in Pt-222 and in hematite-rich portions of Pt-226. Matrices in Pt-650 almost always contain small amounts of clasts (5–20%). The three samples show different trends in Fig. 6, indicating that they are controlled by different minerals.
|
|
Pt-222
Domains representing light and dark phases were analysed, giving a spread of data points around the bulk pseudotachylyte composition. Correlations of K2O with Na2O and Fe2O3 indicate that the matrix composition of Pt-222 is largely controlled by biotite and feldspars (shaded fields in Fig. 6a and b). Because the matrix compositions vary between plagioclase, biotite and K-feldspar end-members, they can be expressed in terms of granite minerals, instead of pseudotachylyte minerals.
Pt-226
Analysed domains of the vein system show more complicated relationships than the simple Pt-222 vein. The highest K2O concentrations are in domains rich in K-feldspar clasts but poor in hematite that occur near the vein–cataclasite contact (Figs 4f and 6d). Domains with few clasts (e.g. 226/2/01, Table 2) may approach the light Pt-222 domains in terms of K2O and Na2O contents, although with lower Fe2O3 contents (1·9 wt %, compared with the typical range of 4–11 wt % in Pt-222). By contrast, the matrices occurring in more distal parts of the injections reach very high to extreme Fe2O3 contents (up to 73 wt % Fe2O3, ‘5’ in Fig. 3b) without accompanying K2O increases.
Pt-650
This sample differs from the previous ones by a lack of correlation between albite and biotite or K-feldspar (Fig. 6e and f). Most of the compositional scatter is due to a variable Fe2O3 proportion. Nevertheless, the Fe2O3-poor matrices fall within the same Plg–Bt–Kfs field as the matrix of Pt-222 (Fig. 6b and f). Fe2O3-rich domains reach similar concentrations as in Pt-226 (43·7 wt %, Fig. 6d and f, Table 2). The major element variations are consistent with concentration–depletion of hematite crystals in an albite-dominated melt.
|
PETROLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Pressure–temperature conditions
As products of disequilibrium melting, pseudotachylytes do not contain appropriate mineral assemblages for thermobarometry; their P–T conditions of formation are, therefore, inferred indirectly from the compositions of metamorphic minerals within the surrounding rocks and oxygen isotope fractionations. On the basis of the microstructures within the host cataclastic rocks, the ambient deformation temperature probably did not exceed 400°C. All samples of cataclasites contain newly formed metamorphic minerals: chlorite, phengitic muscovite, epidote, albite, K-feldspar, hematite, titanite and calcite (Table 3). A late origin of these minerals, subsequent to pseudotachylyte crystallization, is inferred from their occurrence in close proximity to, or within, the pseudotachylyte veins (Fig. 4 h). These minerals could have been formed by metamorphic reactions triggered by shear heating and fluid release from the rapidly quenched pseudotachylyte melt. Mineral assemblages were analysed in two samples, Pt-222 and Pt-650, located >1 km apart. Equilibria relevant to the observed metamorphic assemblage were calculated using the program THERMOCALC v3.1 and the internally consistent thermodynamic dataset of Holland & Powell (1998)
. The average P–T method (Powell & Holland, 1994) yields a temperature of 400°C and pressure of 250–300 MPa (corresponding to 10–12 km). These P–T values are interpreted to reflect equilibrium conditions between the metamorphic mineral assemblages and fluid released during crystallization of the pseudotachylyte melt.
|
The crystalline nature of the pseudotachylyte matrices indicates that undercooling did not exceed 200–300°C, as no glass or spherulites have formed. Thus, we infer the minimum depth of pseudotachylyte formation to have been 10–12 km. If the ambient deformation temperature was 350–400°C, then the geothermal gradient must have been elevated to
30–40°C/km, which is in accord with the Oligocene– Lower Miocene temperature gradient estimated from diagenetic transformation, fluid inclusion and thermal maturity measurements on Paleogene sediments surrounding the Tatra horst (Hurai et al., 1995; Nemok et al., 1996; Biro
et al., 2000). Altogether, these data indicate that the pseudotachylyte veins formed close to the base of the seismogenic upper crust.
Stable isotope compositions and their implications
The sample of the host granite to Pt-222 has 18O value of 9·6, the cataclasite 9·8, the bulk pseudotachylyte 9·6, the feldspar-rich fraction 9·9, the chlorite-rich fraction 7·2, and the hematite-rich fraction 8·8. The 9·6–9·8 values are typical of the orogenic High Tatra granitoids (Kohút et al., 1999). The slightly elevated 18O value of the feldspar-rich fraction and the lower 18O values of the hematite-rich fractions can be easily demonstrated to represent an approach to isotopic equilibrium at elevated temperatures. Figure 7 shows calculated 18O values of coexisting feldspar, chlorite, and hematite-rich fractions in the pseudotachylyte as a function of temperature, assuming bulk 18O value of 9·6.
|
The trajectory of the feldspar-rich portion is given by
 | (3) |
 | (4) |
18Ofld–chl/hem is the isotopic fractionation between the feldspathic fraction and the chlorite-rich fraction (18Ofld–chl), or the feldspathic fraction and the hematite-rich fraction (18Ofld–hem). A 18O value between two minerals as a function of temperature is given by
 | (5) |
for the fractionation between albite–biotite was used for the feldspar–chlorite fractionation. The coefficient for albite–hematite was obtained by combining the albite–quartz coefficient of Bottinga & Javoy (1975) and the quartz–hematite coefficient of Agrinier & Javoy (1988). The calculated isotopic fractionations were adjusted for the observed proportions of chlorite and hematite.
The calculated 18O vs T trajectories of the feldspathic fraction, assuming its equilibrium with either the chlorite-rich or hematite-rich fractions, above 950°C match its 18O value (Fig. 7). Below 900°C, the calculated values start to become significantly heavier. The observed value of the hematite-rich fraction lies above its calculated trajectory. The isotopic composition of the fraction may have been in part kinetically controlled by rapid growth of hematite from the melt, such that hematite acquired more of the isotopic characteristics from the melt than it should have under equilibrium conditions. The data indicate an isotopic reversal between the chlorite-rich and hematite-rich fractions, and the isotopic composition of the chlorite-rich fraction indicates equilibration at 770°C. The isotopic reversal and the lower apparent equilibration temperature of the chlorite-rich fraction may be the result of continued equilibration of chlorite with water that was present in the pseudotachylite to lower temperatures or growth of chlorite as a replacement mineral. Nevertheless, the fractionation between feldspar and hematite in the pseudotachylyte vein indicates rapid crystallization of the melt at >1000°C.
This conclusion is consistent with data for other pseudotachylyte occurrences that were analysed for oxygen isotopes. O'Hara & Sharp (2001) also found that oxygen isotope ratios of coexisting melt and clasts indicate equilibration near 1000°C. In only one occurrence they found anomalously low 18O values, which they attributed to isotopic exchange between a low-temperature pore fluid and the high-temperature melt. The fact that the High Tatra pseudotachylyte is not isotopically shifted compared with the original rock indicates that either (1) there was no interaction of the melt with an external fluid or (2) the isotopic composition of the fluid equilibrated with the host granite before interaction with the pseudotachylite. The second option allows for introduction of an oxidized fluid into the pseudotachylyte melt without leaving an isotopic signature.
Oxygen fugacity and water contents
Hematite is the most characteristic mineral of the High Tatra pseudotachylytes. As the sole oxide mineral, it is also indicative of a highly oxidized regime. Hematite is normally absent in granitoids (except as ilmenohematite). In pseudotachylyte matrices, magnetite is usually the typical oxide (Philpotts, 1964; Macaudière et al., 1985). Only recently have O'Hara & Sharp (2001) reported the presence of hematite in a pseudotachylyte. In the absence of magnetite, hematite cannot be used for oxygen barometry; therefore, the Fe2O3/FeO ratio of the bulk pseudotachylyte Pt-222 was determined by Mössbauer spectrometry to calculate fO2 using the empirical equations of Sack et al. (1980) and Mo et al. (1982). Rapid cooling of the melt could have preserved its original Fe2O3/FeO ratio, although subsequent chlorite crystallization may have changed it slightly. The heavy magnetic fraction of the bulk sample yielded an Fe3+/Fe2+ ratio of 7·48 (88·2% Fe3+), which was applied to the mean composition of the Pt-222 light matrix (n = 5, Table 2). The log fO2 calculated at 1000°C is for P = 250–300 MPa between -2·55 and -2·22, which corresponds to NNO = 7·7–8·0 (where NNO is nickel–nickel oxide), values significantly above the hematite–magnetite (HM) buffer. A possible reason for the high oxygen fugacity in the High Tatra pseudotachylyte melts could have been dissociation of water to H2 and O2 followed by hydrogen escape through the semipermeable host rock. Such oxidation requires, however, the presence of a separate H2O phase (Sato & Wright, 1966). The low 18O values (6–8) and very high water/rock ratio (0·2) of the Long Ridge fault pseudotachylyte enabled O'Hara & Sharp (2001) to interpret a high oxygen fugacity in this hematite-bearing pseudotachylyte as the consequence of the dissociation of water. The High Tatra pseudotachylyte does not show a shift towards lighter oxygen isotopic compositions, because, as noted above, external water probably equilibrated with the host rock at or above the relatively high ambient temperature of 400°C. Because the dissociation of water at this temperature is negligible (log Kf = -15·5), the oxidation of the pseudotachylyte melt probably was caused by oxygen dissolved in this water.
|
CRYSTALLIZATION KINETICS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Quantitative textural data obtained by measuring the number (N) and the size of crystals [L (cm)], can be related to fundamental kinetic parameters, including crystal growth rate [G0 (cm/s)] and nucleation rate [J0 (cm-3/s)], through the equations (Marsh 1998
)
 | (6) |
 | (7) |
introduced the equations
 | (8) |
 | (9) |
; Marsh, 1988). The population density, n, as a function of L (represented by the number of crystals) is the slope of the cumulative frequency, N(L), curve (Marsh, 1988; Cashman, 1990):
 | (10) |
|
) starts after a period of open-system crystallization (
), ln(n0) = 30·6 (i, j). (See text for details.)Open vs closed system
There are two fundamentally different ways in which closed and open systems generate log-linear shapes of CSDs (Fig. 9). In an open system, where the filling of a reservoir is constantly changing by partial emptying and recharge and the nucleation rate is constant, the log-linear pattern is achieved by continuous discharge of larger (=older) crystals (Marsh, 1988
) (Fig. 9a). Many interesting features of open systems have been observed using industrial mixed-suspension and mixed product removal (MSMPR) crystallizers. It can be shown (Randolph & Larson, 1971; Marsh, 1988; Cashman, 1990) that a system with constant volume (V) and flux rate (Q) has a characteristic residence time 
= V/Q and its mass conservation equation has a steady-state analytical solution (Marsh, 1988)
 | (11) |
is the slope of the log-linear CSD, where G = LD, the dominant length. LD corresponds to the most frequent crystal size in the system as a function of residence time . In an open system, the nucleation rate is constant, which, if G is also constant, results in a linear pattern of the CSD. The size actually reflects the probability of a crystal being retained in the system (Marsh, 1988).
|
In a closed system (Fig. 9b), the volume of the crystallizing liquid magma is constant and log-linear CSDs result from the exponential increase of J(t) with time, whereas G(t) may be considered approximately constant. The slopes in closed systems depend on the exponential constant a in equation (8), which determines the change in the nucleation rate. Steeper slopes are generated at higher supersaturations, regardless of the way supersaturation is achieved. Because during crystallization the volume of magma diminishes, the effective late-stage nucleation rate becomes vanishingly small and a maximum is generated in the closed CSD (‘humped’ pattern) as all crystals progress with time to higher size classes (Marsh, 1988
).
Hematite size distribution
The large abundance of hematite crystals in the matrices of the High Tatra pseudotachylytes makes hematite an appropriate mineral for CSD measurements. The measurements are easily made using BSE images because of the extreme brightness of this mineral (Fig. 4). The CSDs are based on 46 images of various positions in the veins of the studied samples. The pseudotachylytes exhibit mostly linear CSDs in ln(n) vs L plots, with the notable exceptions of several CSDs in samples Pt-226 and Pt-650 that have ‘humped’ patterns (Figs 8f–i).
Pt-222
The images from this sample generally show well-defined log-linear CSD patterns, except for images of matrices representing the primary melt, which have poorly defined CSDs because of low numbers of crystals. In comparison with other samples, the CSDs of Pt-222 have the lowest intercepts and slopes. Images of the primary matrix give the lowest values (Fig. 8d). Images of the darker phase at the rims of the vein show steeper slopes than in the centre of the vein, indicating higher degrees of undercooling at the vein margins. The CSDs are linear without any distinct maxima.
Pt-226
Sixteen images of four thin sections of this sample were made. The CSDs have a range of slopes and intercepts that covers almost the total range for all samples. The lowest values were recorded in thicker parts of the fractures filled in by a slurry of clasts, which seems to have been liquefied by an interconnected melt network. In contrast, the highest slope and intercept values occur in CSDs of thin terminations of the fractures. The CSDs with steep slopes have humped patterns with distinct maxima, indicating crystallization in a closed system (Fig. 8f).
Pt-650
Generally high values of both slopes and intercepts of six CSDs of this sample (Fig. 8g–i) also indicate a liquefied slurry. The highest values were recorded near the vein boundary (34 in Fig. 3d). The CSDs with the highest slope have also humped patterns that are indicative of closed-system crystallization.
The results are summarized in Fig. 10 and Table 4, where the slopes, intercepts and other parameters of the measured CSDs are given.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Derivation of the growth rate
Crystal size is the product of crystallization time and growth rate. The growth rates of various minerals, including plagioclase, magnetite and ilmenite, are known from experiments (e.g. Swanson, 1977
) or measurements of CSDs in basaltic lava lakes (e.g. Cashman & Marsh, 1988). Growth rates vary by several orders of magnitude from 10-10 cm/s at low undercoolings (T), reaching a maximum of 10-6 cm/s at about T = 100–250°C. The plagioclase and ilmenite growth rates in the basaltic Makaopuhi lava lake, Hawaii, were 5·4 x 10-11 to 10-10 cm/s, and (4–5) x 10-10 cm/s, respectively, both at very low undercoolings (T = 0·01–0·001°C; Cashman & Marsh, 1988; Cashman, 1990). Applying the ilmenite value to hematite in the pseudotachylyte melt yields crystallization times of 2–23 days. These times are much longer than the short times commonly assumed for the solidification of frictional melts. For example, solidification times of 0·4–40 s were calculated for veins of 1 mm and 1 cm thickness by Sibson (1975) (based on conductive cooling without including a latent heat of crystallization). Assuming (1) the temperature recorded by a metamorphic cataclasite assemblage to be Tcc 400°C, initial magma temperature T0 = 1000°C and (2) the temperature profile of a cooling sheet, where the temperature at the wall-rock contact is 
(T0 + Tcc) (Jaeger, 1968), the undercooling of the pseudotachylytes is estimated to have been 300°C.
Effective crystal growth rates can be calculated from solidification times based on heat conduction equations (Jaeger, 1968) with inclusion of the latent heat of crystallization (H). The numerical solution provided by Jaeger with H = 419 J/g yields times for complete solidification of 31–63 s for vein half-thicknesses a = 0·35 and 0·5 cm (thermal diffusivity 
= 0·008 cm2/s, dimensionless time = t/a2 = 2, which corresponds to cooling from 1100 to 600°C). The value of 31 s was used to calculate G together with the longest apparent hematite crystal size 0·0012–0·0019 cm in a Pt-222 vein of 0·7 cm thickness. The resulting growth rate (G = Lmax/t) is (4–6) x 10-5 cm/s. This growth rate is one order of magnitude higher than the maximum obtained by Swanson (1977) for plagioclase at undercoolings of 150–200°C. The value may indicate that the hematite growth rate is indeed faster than the plagioclase growth rate, or may be biased by an underestimate of the crystallization time or overestimate of Lmax (aggregated rather than single crystals). For the model hematite habit 1:2:4 (see below) the ‘true’ lengths (and growth rates) would be doubled because the ratio of the longest to the intermediate dimension equals two (Higgins, 1994). In view of this, the obtained value of time and G should be viewed as apparent. The value of G = 5 x 10-5 cm/s is used below for calculations that require the consideration of growth rate.
Interpretation of the CSDs
The refined unfolding technique of Sahagian & Proussevitch (1998) involved in the program ‘CSD corrections’ of Higgins (2000) has one important effect on reduced data—the smallest crystals show increasingly high population densities. CSDs that are log-linear in 2D [or when converted using equation (1)] tend to become concave upward at small sizes when corrected using the program (Fig. 11a), whereas the humped patterns tend to become log-linear (Fig. 11b). This effect is even stronger in histograms with geometric class sizes when the number densities are converted to population densities (dividing by logarithmically decreasing bin sizes). If the downbending of CSDs was an artefact, e.g. caused by imperfect detection of the smallest crystals, it should occur in all images. However, only six from a total of 46 CSDs are downbent. Moreover, the humped CSDs seem to have developed only for specific textural positions.
|
The measured hematite crystals were appoximated by prisms of dimension 1:2:4 as they show maximum width/length ratio of 0·5 (Higgins, 1994
). Larger crystals when refocused in transmitted light seem to confirm the habit. However, smaller crystals or crystallites (1–3 µm) appear to be isometric or spherical rather than to be basal sections of small prisms as assumed in the calculation procedure. The population is, therefore, multidispersed rather than polydispersed (Sahagian & Proussevitch, 1998) and smaller sizes are better approximated by spherical intersection probablities (P). If the small crystals were treated as a separate polydispersed system, higher P values might be used for the spheres, which would result in less overcorrection. It is suggested that the use of a polydispersed system (e.g 1:2:4 prisms) overcorrects the smallest crystal sizes in assumed multidispersed systems. Log-linear CSDs are well established in 3D (by screening) in chemical technology (Randolph & Larson, 1971), whereas downbending CSDs are the inevitable result of crystallization in closed systems (Marsh, 1998). Therefore, the conversion according to equation (1) was used for the interpretation of humped patterns (Fig. 8), whereas the conversion by the CSD corrections program was used for the interpretation of slopes and intercepts. It is noted that the relationships between samples and between images are the same for both conversions.
Modelling of CSDs
Marsh (1998) developed a series of equations that allow quantitative modelling of crystallization in a closed system. These equations were used to determine if crystallization of the pseudotachylytes is consistent with closed-system crystallization. Model curves to characteristic CSDs are shown in Fig. 8. Of the parameters used [G (growth rate), J0 (initial nucleation rate), 
(overall crystallinity), a (exponential constant)], G was assumed to be constant, whereas the others were varied to follow an observed CSD. J0 varies typically between 3 x 105 and 5 x 107 cm-3/s and a from three to six. About a quarter of the CSDs coincide with closed-system model curves, which are linear only for low degrees of crystallinity, = 0·2–0·5 (Fig. 8b, c and e). Model closed-system curves for about half of the CSDs require higher values (>0·5) to fit the observed slopes, which results in downbending and formation of maxima (dashed curves in Fig. 8a, g and i). Modelling, therefore, leads to the rather counterintuitive conclusion that the majority of linear CSDs cannot be approximated by closed-system crystallization. Instead, their patterns reflect crystallization in an open system.
Any log-linear array may be modelled as an open system simply by extracting G and n0 from its slope and intercept according to equation (11). As the volume crystallizes, a steady-state CSD is developed after some time. The CSD is the same in the crystallizing volume and in the magma that empties the volume (Randolph & Larson, 1971), which means that kinetics alone cannot explain the highly variable hematite contents in the melt. It is not easy to reconcile an open-system crystallization process with the inherently short-lived process responsible for the formation of the pseudotachylyte melt. However, pseudotachylyte melts have been shown to be highly fluid (Spray, 1993) and their migration driven by extension fractures seems inevitable (Camacho et al., 1995). As indicated especially by the CSDs in vein Pt-222, influx of new melt must have persisted for some time. Melt flow is evident in Fig. 3a, where an older primary melt is invaded by a later, more mafic melt. Also distinct is the direction of flow, with clasts being filtered out as the melt approaches a narrower portion of the vein. It is suggested that during melt injection, when the frictional melt flowed into tensional fractures, the crystallizing hematite generated a CSD that approached steady state and after the temperature dropped, was arrested in the solidifying feldspathic matrix.
Open system followed by closed system
Even the observed ‘humped’ CSD patterns can seldom be fitted by simple closed-system curves, because very high values (>0·95) are required to fit the changing slopes and downbending. The model curves tend to be more pronounced than the observed ones (dashed curves in Fig. 8a, f, g and i). It is suggested that the injected melt, which was captured in the tips of fractures or within cracks in larger lithic clasts (Fig. 3c and d), developed most of its CSD during the open-system stage, but continued to crystallize in a closed-system crystallization mode. During this stage, the hematite crystals stopped being flushed out and started to progress in all sizes with new crystals nucleating at an exponentially increasing rate [J(t)] starting with the original n0 inherited from the open-system crystallization (Fig. 8j). It is important to note that the nucleation rate, J, is prevented from increasing in open systems because the feeding hot magma maintains the temperature constant. After the melt influx ceases, J0 begins to increase exponentially in relation to the local cooling rate. Thus, the log-linear character of the CSD is maintained (not necessarily with the same slope). With increasing crystallinity values , the effective late-stage nucleation rate decreases (Marsh, 1998) and the CSD bends down. Figure 8i and j shows that modelling of closed-system crystallization requires that the starting n0 value is higher than that corresponding to the highest crystal size Lmax (n in Fig. 8j). The higher initial n0 of the closed-system stage [ln(n0) = 30·6 compared with ln(n0) = 27·5 for one-stage model, Fig. 8i] may be inherited from n0 final density of the open-system stage [notation after Marsh (1998)]. From the reconstruction of the initial open-system CSD (‘open system‘ in Fig. 8i) it follows that it had a much lower population density ln(n). A low crystal content of hematite of 2·6 vol. % was calculated for this CSD from frequencies of size classes 0·0005–0·001 cm, assuming that crystals in these classes have grown by a factor of 2·25 (and their areas by 2·252). This supports the suggestion that the low crystal-density magma could have moved freely in an open system, but after it stopped it became mostly crystalline (to 22 vol. % of hematite) as a closed system.
Humped CSD patterns
Downbending (humped CSD patterns, Fig. 8d, f, h and i) may be generated either by crystallization in a closed system or by an annealing process (Ostwald ripening). The latter process has been demonstrated for regional metamorphic garnets (Cashman & Ferry, 1988) and for interiors of thick mafic dikes (Marsh, 1998). For the case of annealing in mafic dikes, humped CSDs for clinopyroxene in interiors of dikes have lower slopes and intercepts than linear CSDs from chilled margins. The CSD for the matrix of 222/1/17 (Fig. 8d) has a flat slope and very low intercept, and is represented by the primary melt enclosed by a later, Fe-enriched pulse (Fig. 3a). Insulation of this domain by the later melt may have prolonged the crystallization time sufficiently for annealing to take place (50 s). The humped CSDs in Fig. 8f, h and i contrast with the previous example in having the steepest slopes and highest intercepts. The maxima in these CSDs are considered to be indicative of advanced degrees of closed-system crystallization. This means that, in spite of very short crystallization times (10–25 s), nucleation continued to high crystallinities ( > 0·9) when the diminishing melt volume caused a progressive decrease in the nucleation rate. It is suggested that high contents of network modifiers (Fe2O3, K2O, Na2O) suppressed polymerization in these melts and allowed the protracted nucleation. It appears significant that all matrices with humped CSDs (except 222/1/17) have Si:O ratios (a general melt structure parameter; Watson, 1977) below 0·286, which corresponds to the breakdown of silicate polymers. By contrast, the log-linear CSDs of the primary Pt-222 melts, whose crystallization was arrested while still in the steady-state open system, have Si:O ratios between 0·32 and 0·36.
Slope vs intercept
When interpreting a number of CSDs, it is convenient to compare them in terms of their two fundamental parameters, slope S and intercept I (in addition to Lmax). For this interpretation, all measured CSDs were converted to 3D using the ‘CSD correction’ program. The S and I calculated from straight segments of CSDs are plotted in Fig. 10a. Compared with conversion by equation (1) they have slopes less steep, as a result of doubled ‘true’ crystal lengths (Fig. 11), but the relationships between samples and between images are the same for both conversions. Error bars (2
), which result from the scatter of samples in ln(n) vs L diagrams, were added to all CSDs. There is a clear tendency of increasing intercept with increasing absolute value of the slope. Whereas the slopes of hematite-poor CSDs in Pt-222 decrease slowly, the slopes of CSDs in hematite-rich Pt-226 and Pt-650 bend down more steeply at the highest intercepts.
The slopes and intercepts of separate parcels of melt are mutually dependent, provided the parcels have the same overall crystallinity (Randolph & Larson, 1971; Marsh, 1988):
 | (12) |
1 > 2 and n0 is nucleation density. G is considered here a constant (G1 = G2). Because I = ln(n0) and S = -1/G, equation (12) upon substitution becomes
 | (13) |
 | (14) |
(i.e. identical masses). Such ‘isodensity’ curves are calculated and compared with the observed CSDs in Fig. 10b. The curves were adjusted arbitrarily (by suitable choice of S2/S1) to follow the CSDs with approximately the same crystallinities. It is clearly seen how the CSDs with increasing are horizontally displaced to higher intercepts, whereas their steep bending-down reflects progressive steepening of the CSD slopes with increasing intercept.
Zieg & Marsh (2002) have recently developed very similar equations relating slope and intercept [ln(I) = 4 ln(S) - ln(C)], which includes a constant C combined from constants CL and CN [equations (6) and (7)]. This constant can be calibrated against modal amount of the measured crystals and the curves, such those in Fig. 10b, are then fixed. The values of C = 1, 1·8, 4, 10 and 35 correspond approximately to the isodensity curves in Fig. 10b. The steepening of the isodensity curves results from the mass conservation and the curves correspond approximately to the closure limits of Higgins (2002) for pertinent hematite volume fractions.
For the steady state, the nucleation rate J is the product of the kinetic frequency and the concentration of nuclei (Cashman, 1990):
 | (15) |
v is number of atoms per unit volume, Ga is the activation energy and G* is the thermodynamic driving force for the nucleation. Thus, the steady-state J may be considered proportional to the concentration and logarithm of the ‘driving force’. By setting the hematite volume fraction (Xhem) for the concentration v, F for the driving force (
undercooling), and by dropping the constants, equation (15) becomes
 | (16) |
) and rearranging gives
 | (17) |
 | (18) |
The value of F discriminates fairly well both between samples and within samples. The simple vein pseudotachylyte Pt-222 forms the low-F segment of the curve, the thick iron-rich vein Pt-650 forms the high-F segment, and the more complex anastomosing system of Pt-226 covers the whole range (Fig. 10c). The lowest F values are provided by CSDs from clast-poor, low-Fe matrices in Pt-222, interpreted as primary melts, or from thick, clast-laden injections of Pt-226. Intermediate F values belong to linear CSDs of Pt-650 (42, 44) that are inconsistent with closed-system crystallization (Figs 8g and 10c). Here, the flow of hot melt with hematite residence time = 4–5 s may have prolonged the crystallization interval to 30 s, which contrasts with CSDs that have the highest F and humped patterns. These CSDs (650/48, 34) formed within 20 s, mainly in a closed system: 48 is a matrix filling a crack in a larger lithic clast (Fig. 3d), and 34 occurs just at the margin of the vein. The matrices 3 and 58 from Pt-226/2 offshoots (Figs 3b and 8e) have linear CSDs with low F values, suggesting open-system crystallization with residence time = 12–14 s, and crystallized within 80–90 s. Hydrofracturing of the pressurized melt zone followed by flow of the melt into long tensional cracks may explain the longer crystallization times. The highest F values belong to CSDs from a very thin (90 µm) offshoot (Fig. 3c), where the system became closed as suggested by the humped CSD (Figs 8f and 10c). The other is the Pt-226/1/2 matrix from the contact of the generation surface with the footwall. Extreme cooling rates can be explained by the very high surface/volume ratio in the former case, and by possible thin, pre-slip injection (Spray, 1993) into the cooler rupture plane for the latter case.
Published magmatic CSDs mainly come from volcanic rocks, which crystallized under conditions entirely different from those of pseudotachylyte melts. The obvious major factor determining G and J values, undercooling, coupled with extremely short crystallization times, resulted in very steep slopes (-1000 to -8000) of the pseudotachylyte hematite CSDs.
Formation of pseudotachylyte melts
The compositions of the pseudotachylyte melts differ strongly from granitic melts in both their low SiO2 and their high to very high Fe contents. They are obviously controlled by the proportions of minerals that preferentially enter the melt. Thus, pseudotachylytes are disequilibrium melts. Preferential melting of hydrous ferromagnesian phases is a typical process of pseudotachylyte melt generation (Spray, 1987, 1992; Reimold, 1991).
The light matrix phase of Pt-222 is believed to have a composition that is closest to the original melt, because clast-poor, in situ embayments of this phase are observed in contact with biotite (now chlorite) and are cut by a later, darker, clast-laden phase. The darker phase may represent a less viscous, iron-enriched melt injected from a more distal generation surface (Table 2).
Melted mineral assemblage
Assuming that the light phase of Pt-222 represents the primary melt composition, it can be used to estimate the proportions of granite minerals that entered the pseudotachylyte melt. Mineral proportions of a light matrix containing no clasts or hematite clusters (222/1/11, Table 1) were estimated using a modified version of the least-squares approximation program MAGFRAC (Morris, 1984). Because all the matrices are apparently depleted in MgO and TiO2, both oxides were considered mobile in modelling. The resulting mixture (P in Table 5) yields a composition almost identical to the matrix 222/1/11 with 
R2 = 0·006, assuming the loss of 1·3% MgO and 0·2% TiO2. The higher proportion of biotite in the melted assemblage (18·3 wt %) compared with the average proportion in the granodiorite (11·6 wt %) requires a mass of granite 1·6 times greater than that of the pseudotachylyte to supply the required amount of Fe and K. From this mass, 63·5 wt % underwent complete melting and 36·5 wt % was left as residue (Table 5, Fig. 12).
|
|
O'Hara (2001)
suggested that the amount of clasts in a pseudotachylite is a function of the thermodynamic efficiency of fault work and may be calculated using the equation W/M = (Thigh - Tlow)/Thigh, where W/M is the clast/melt ratio, Thigh is the temperature of melting and Tlow is the country rock temperature. Using Tlow = 400°C and Thigh = 1000°C for the High Tatra pseudotachylyte Pt-222, the equation yields the ratio 0·47 corresponding to 32% clasts. This corresponds well to the value 36·7% (by volume) for the residue derived by mass balance, suggesting that the residue is fully represented by the clasts. The mass balance strongly depends on the matrix composition considered primary: the higher matrix Fe contents result in higher residue proportions as a result of the greater granite mass required. It seems, therefore, that only matrices with 2 vol. % of hematite (4–5% Fe2O3) represent a primary melt composition in the sample Pt-222.
The melting occurred in a comminuted rock where localized shearing encountered biotite, which started to melt preferentially (Maddock, 1992; Spray, 1992; Swanson, 1992). Both the water liberated from biotite and oxidized external water allowed subsequent melting of quartz and feldspars (Fig. 4 h). Pools of the primary melt were interconnected by shearing into a melt layer along the shear zone. Thus, the hematite-rich melt originated mostly from the breakdown of biotite and coalesced from a larger volume of cataclastic rock to localized planes as in samples Pt-226 and Pt-650.
The absence of any magmatic Mg-bearing mineral in the pseudotachylytes suggests depletion of MgO in the melts, perhaps because under conditions of instantaneous and non-equilibrium melting and very high oxygen fugacity, no Mg mineral was stable. MgO probably entered the fluid phase, which eventually separated after the melt solidified, and reprecipitated as the late chlorite or phengite, both in the pseudotachylyte and the neighbouring cataclasite (Fig. 4 h). Similarly, because of the high oxygen fugacity, only a limited amount of Ti entered the hematite. An overall Ti (and Ca) balance was approached only locally by accumulations of primary interstitial or subhedral titanite (Fig. 4e). The preference of Mg for the fluid phase is documented by chlorite-bearing or Mg-actinolite-bearing vesicles in some pseudotachylytes (Nockolds, 1940, in Philpotts, 1964; Maddock et al., 1987).
Pseudotachylyte melt differentiation
Whereas the K/Fe ratio of the primary melt in Pt-222 reflects its origin as a disequilibrium melt of a biotite-bearing source, the vein system of Pt-226 shows decoupling of Fe3+ from K. The ferric iron accumulated in very small melt volumes in the most distal fracture tips. There is almost no overlap between data from Pt-226 and Pt-222 in the K2O–Fe2O3 plot (Fig. 6b and d). The trend of Pt-226 compositions to higher K2O at low Fe2O3 can be explained by the increasing proportion of K-feldspar clasts in relatively larger analysed domains (0·39–0·45 mm2). Although the primary melt composition does not seem to be preserved in Pt-226, the clast-poor matrices (e.g. 2 in Fig. 3b) probably approach the original melt except for the strong depletion of Fe3+. These melts were able to crystallize only 0·6–0·9 vol. % of hematite in contrast to melts in narrow parts of the veins (5 in Fig. 3b, 13 in Fig. 3c), which have 28 vol. % of hematite (Table 1). The observed Fe3+ distribution is easily explained by hematite accumulation. As demonstrated in the previous section, the great majority of the CSDs are consistent with open-system crystallization, which implies movement of the crystallizing melt. When such a melt approaches a boundary (e.g. a fracture tip), the moving crystals may accumulate as a result of inertial forces. High concentration of hematite along some vein boundaries (Fig. 3c) may be explained by flow differentiation during which clasts become concentrated in centres of the veins.
The splitting of primary melts into series of both hematite-enriched and -depleted fractions is evident in Pt-226 by a subhorizontal shift in the slope–intercept plot (arrows in Fig. 10a). For the two images with the highest F, crystal accumulation is less probable because no high slope and low intercept of the pertinent CSD occurs. Instead, melt inhomogeneity is invoked here. The inhomogeneity may have resulted from differing rheology (viscosity) of melt portions with various Fe concentrations. Distinct inhomogeneity of artificial pseudotachylyte melts (glasses), presumably owing to varying Fe, was demonstrated experimentally by Spray (1993).
Pseudotachylyte vein Pt-650 resembles Pt-226 in the efficiency of Fe3+ accumulation (Fig. 6d and f), although the matrices maintain a constant K2O/Na2O ratio of 0·2 (Fig. 6e). Low-hematite domains were analysed in cataclasite areas either beyond or in the pseudotachylyte vein. The vein is much thicker (up to 1·5 cm) than in Pt-226; therefore, hematite must have accumulated from a larger volume of cataclasite, from which the Fe-enriched melt was drained to form the localized swell.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Detailed investigation of the High Tatra pseudotachylytes has revealed some distinctive features in comparison with pseudotachylytes in other localities. These include the complete crystallization expressed by the absence of glass and the strong primary oxidation expressed by the presence of hematite. The absence of glass is consistent with the rather high ambient temperature of
400°C of the rocks, as recorded by the metamorphic cataclastic assemblage. Nevertheless, the amount of undercooling was high enough to cause extremely high growth and nucleation rates unseen in any other rock type. The oxygen isotopic compositions of the pseudotachylytes support introduction of meteoric, oxidizing water, which isotopically equilibrated during flow through the host rock. Early preferential fusion of biotite is clearly seen in textural relationships and is evidenced by the chemistry of the light pseudotachylyte matrix. This implies that frictional melting was not inhibited by the free water that was present. The fusion resulted in early precipitation of hematite whose crystal size distributions point to high mobility of melt + crystals (+clasts) in an open system. The hematite CSDs approached steady state with residence times strongly depending on the local rate of heat escape. At the closures of melt pathways, the crystals accumulated to yield the very high iron concentrations, which are unattainable by any melting mechanism. This suggests that the primary melt compositions in pseudotachylytes are only rarely preserved. |
ACKNOWLEDGEMENTS |
|---|
We thank Peter Reichwalder for providing the sample Pt-222 and for inspiring the first author to work on the pseudotachylyte problem. Professor J. Lipka of the Slovak Technical University is thanked for Mössbauer analyses of pseudotachylyte concentrates. Bruce Marsh, John Spray, Anthony Philpotts and an anonymous reviewer are thanked for critical and very useful reviews, and Marjorie Wilson for very detailed editorial work. Free software ImageTool (available from http://ddsdx.uthscsa.edu/dig) of the University of Texas Health Science Center, San Antonio, was used for image analysis. This work was supported by grants GA4078, 7030 and 3167, and the University of Missouri Research Board.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGYPETROGRAPHYSAMPLING STRATEGY AND ANALYTICAL...MAJOR ELEMENT GEOCHEMISTRYPETROLOGYCRYSTALLIZATION KINETICSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Agrinier, P. & Javoy, H. (1988). A natural calibration of the quartz–rutile mineral pair oxygen isotope geothermometer (abstract). Chemical Geology 70, 182.[Web of Science]
Armienti, P., Pareschi, M. T., Innocenti, F. & Pompilio, M. (1994). Effects of magma storage and ascent on the kinetics of crystal growth. Contributions to Mineralogy and Petrology 115, 402–414.[CrossRef][Web of Science]
Bac-Moszaszwili, M. (1993). Structure of the western termination of the Tatra massif (in Polish, English summary). Annales Societae Geologie Poloniae 63, 167–193.
Biro, A., Kotulová, J. & Soták, J. (2000). Burial history of the Levoa Basin in the light of illite/smectite and vitrinite reflectance data. Mineralia Slovaca 32, 332.
Bottinga, Y. & Javoy, M. (1975). Oxygen isotope partitioning among the minerals in igneous and metamorphic rocks. Reviews of Geophysics and Space Physics 13, 401–418.[Web of Science]
Burchart, J. (1972). Fission-track age determination of accessory apatite from the Tatra mountains. Earth and Planetary Science Letters 15, 418–422.[CrossRef][Web of Science]
Camacho, A., Vernon, R. H. & Fitz Gerald, J. D. (1995). Large volumes of anhydrous pseudotachylyte in the Woodroffe Thrust, eastern Musgrave, Australia. Journal of Structural Geology 17, 371–383.[CrossRef][Web of Science]
Cashman, K. V. (1990). Textural constraints on the kinetics of crystallization of igneous rocks. In: Nicholls, J. & Russell, J. K. (eds) Modern Methods of Igneous Petrology. Mineralogical Society of America, Reviews in Mineralogy 24, 259–314.
Cashman, K. V. & Ferry, J. M. (1988). Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization. III. Metamorphic crystallization. Contributions to Mineralogy and Petrology 99, 401–415.[CrossRef][Web of Science]
Cashman, K. V. & Mangan, M. T. (1994). Physical aspects of magmatic degassing II. Constraints on vesiculation processes from textural studies of eruptive products. In: Caroll, M. R. & Holloway, J. R. (eds) Volatiles in Magmas. Mineralogical Society of America, Reviews in Mineralogy 30, 447–478.
Cashman, K. V. & Marsh, B. D. (1988). Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization. II. Makaopuhi lava lake. Contributions to Mineralogy and Petrology 99, 292–305.[CrossRef][Web of Science]
Clayton, R. & Mayeda, T. K. (1963). The use of bromine pentafluoride in the extraction of oxygen from oxides and silicates for isotopic analysis. Geochimica et Cosmochimica Acta 27, 43–52.[CrossRef][Web of Science]
Gaweda, A. (1995). Geochemistry and Rb/Sr isochrone age of pegmatites from the Western Tatra Mts. Geologica Carpathica 46(2), 95–99.[Web of Science]
Higgins, M. D. (1994). Numerical modelling of crystal shapes in thin sections: estimation of crystal habit and true size. American Mineralogist 79, 113–119.[Abstract]
Higgins, M. D. (2000). Measurement of crystal size distributions. American Mineralogist 85, 1105–1116.
Higgins, M. D. (2002). Closure in crystal size distributions (CSD), verification of CSD calculations, and the significance of CSD fans. American Mineralogist 87, 171–175.
Holland, T. J. B. & Powell, R. (1998). An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16, 309–343.[CrossRef][Web of Science]
Hurai, V., iráová, V., Marko, F. & Soták, J. (1995). Hydrocarbons in fluid inclusions from quartz–calcite veins hosted in Paleogene flysch sediments of the Central Western Carpathians. Mineralia Slovaca 27, 383–396.
Hurai, V., Janák, M., Ludhová, L., Horn, E. E., Thomas, R. & Majzlan, J. (2000). Nitrogen-bearing fluids, brines and carbonate liquids in Variscan migmatites of the Tatra Mountains, Western Carpathians—heritage of high-pressure metamorphism. European Journal of Mineralogy 12, 1283–1300.
Jaeger, J. C. (1968). Cooling and solidification of igneous rocks. In: Hess, H. H. & Poldervaart, A. (eds) Basalts, Vol. 2. New York: Interscience, pp. 503–536.
Janák, M. (1994). Variscan uplift of the crystalline basement, Tatra Mts., Central Western Carpathians: evidence from 40Ar/30Ar laser probe dating of biotite and P–T–t paths. Geologica Carpathica 45, 293–300.
Janák, M., O'Brien, P. J., Hurai, V. & Reutel, C. (1996). Metamorphic evolution and fluid composition of garnet–clinopyroxene amphibolites from the Tatra Mountains, Western Carpathians. Lithos 39, 57–79.[CrossRef][Web of Science]
Janák, M., Hurai, V., Ludhová, L., O'Brien, P. J. & Horn, E. E. (1999). Dehydration melting and devolatilization of high-grade metapelites: the Tatra Mountains, Western Carpathians. Journal of Metamorphic Geology 17, 379–396.[CrossRef][Web of Science]
Kahan, . (1969). Eine neue Ansicht über den geologischen Aufbau des Kristallinikums der West Tatra. Acta Geologica et Geographica Universitatis Comenianae 12, 115–122.
Klug, C. & Cashman, K. V. (1994). Vesiculation of May 18, 1980, Mount St. Helens magma. Geology 22, 468–472.
Kohút, M. & Janák, M. (1994). Granitoids of the Tatra Mts., Western Carpathians: field relations and petrogenetic implications. Geologica Carpathica 45, 301–311.
Kohút, M. & Sherlock, S. C. (2002). Laser probe 40Ar/39Ar study of pseudotachylyte and its host rock from the Tatra Mountains (Western Carpathians, Slovakia). Geolines 14, 46–47.
Kohút, M., Poller, U., Todt, W., Nabelek, P. & Janák, M. (1999). The Tatra Mts. granitoid pluton—a collision related Variscan intrusion in the Western Carpathians, Slovakia. Journal of Conference Abstracts EUG X 4, 672.
Ková, M., Král, J., Márton, E., Plaienka, D. & Uher, P. (1994). Alpine uplift history of the Central Western Carpathians: geochronological, paleomagnetic, sedimentary and structural data. Geologica Carpathica 45, 83–96.
Král, J. (1977). Fission track ages of apatites from some granitoid rocks in West Carpathians. Geologick Zborník—Geologica Carpathica 28, 269–276.
Macaudière, J., Brown, W. L. & Ohnenstetter, D. (1985). Microcrystalline textures resulting from rapid crystallization in a pseudotachylite melt in a meta-anorthosite. Contributions to Mineralogy and Petrology 89, 39–51.[CrossRef][Web of Science]
Maddock, R. H. (1983). Melt origin of fault-generated pseudotachylytes demonstrated by textures. Geology 11, 105–108.
Maddock, R. H. (1992). Effects of lithology, cataclasis and melting on the composition of fault-generated pseudotachylytes in Lewisian gneiss, Scotland. Tectonophysics 204, 261–278.[CrossRef][Web of Science]
Maddock, R. H., Grocott, J. & Van Nes, M. (1987). Vesicles, amygdales and similar structures in fault-generated pseudotachylyte. Lithos 20, 419–432.[CrossRef][Web of Science]
Maluski, H., Rajlich, P. & Matte, P. (1993). 40Ar–39Ar dating of the Inner Carpathians Variscan basement and Alpine mylonitic overprint. Tectonophysics 223, 313–337.[CrossRef][Web of Science]
Marko, F., Plaienka, D. & Fodor, L. (1995). Meso-Cenozoic tectonic stress fields within the Alpine–Carpathian transition zone: a review. Geologica Carpathica 46, 19–27.[Web of Science]
Marsh, B. D. (1988). Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization. Contributions to Mineralogy and Petrology 99, 277–291.[CrossRef][Web of Science]
Marsh, B. D. (1998). On the interpretation of crystal size distributions in magmatic systems. Journal of Petrology 39, 553–599.[CrossRef][Web of Science]
Mo, X., Carmichael, I. S. E., Rivers, M. & Stebbins, J. (1982). The partial molar volume of Fe2O3 in multicomponent silicate liquids and the pressure dependence of oxygen fugacity in magmas. Mineralogical Magazine 45, 237–245.[Web of Science]
Morris, P. A. (1984). Magfrac: a Basic program for least-squares approximation of fractional crystallization. Computers and Geosciences 10, 437–444.[CrossRef]
Narr, W. & Suppe, J. (1994). Kinematics of basement-involved compressive structures. American Journal of Science 294, 802–860.
Nemok, J., Bezák, V., Janák, M., Kahan, ., Ryka, W., Kohút, M., Lehotsk, I., Wieczorek, J., Zelman, J., Mello, J., Halouzka, R., Raczkowski, W. & Reichwalder, P. (1993). Geological map of the Tatra Mountains, 1:50 000. Bratislava: Geologick ústav D. túra.
Nemok, M., Keith, J. F., Jr & Neese, D. G. (1996). Development and hydrocarbon potential of the Central Carpathian Paleogene Basin, West Carpathians, Slovak Republic. In: Ziegler, P. A. & Horváth, F. (eds) Peri-Tethys Memoir 2: Structure and Prospects of Alpine Basins and Forelands. Mémoires du Muséum National d'Histoire Naturelle, Paris 170, 321–342.
Nockolds, S. R. (1940). Petrology of rocks from Queen Mary Land: Australian Antarctic Expedition 1911–1914. Scientific Reports, Series A 4, 15–86.
O'Hara, K. (1992). Major- and trace-element constraints on the petrogenesis of a fault related pseudotachylyte, western Blue Ridge province, North Carolina. Tectonophysics 204, 279–288.[CrossRef][Web of Science]
O'Hara, K. (2001). A pseudotachylyte geothermometer. Journal of Structural Geology 23, 1345–1357.[CrossRef][Web of Science]
O'Hara, K. & Sharp, Z. D. (2001). Chemical and oxygen isotopic composition of natural and artificial pseudotachylyte: role of water during frictional fusion. Earth and Planetary Science Letters 184, 393–406.[CrossRef][Web of Science]
Petrík, I. & Reichwalder, P. (1996). Pseudotachylyte from the High Tatra Mts. Polskie Towarzystwo Mineralogiczne—Prace Specialne 7, 61–63.
Petrík, I., Broska, I. & Uher, P. (1994). Evolution of the Western Carpathian granite magmatism: age, source rock, geotectonic setting and relation to the Variscan structure. Geologica Carpathica 45, 283–291.
Philpotts, A. R. (1964). Origin of pseudotachylites. American Journal of Sciences 262, 1008–1035.
Plaienka, D., Grecula, P., Puti, M., Hovorka, D. & Ková, M. (1997). Evolution and structure of the Western Carpathians: an overview. In: Grecula, P., Hovorka, D. & Puti, M. (eds) Geological Evolution of the Western Carpathians, Mineralia Slovaca—Monograph, Bratislava 1–24.
Plaienka, D., Janák, M. & Petrík, I. (2001). Excursion to the Tatra Mountains, Central Western Carpathians: tectonometamorphic records of Variscan and Alpine orogeny. Geolines 13, 141–148.
Poller, U. & Todt, W. (2001). U–Pb single zircon data of granitoids from the High Tatra Mountains (Slovakia): implications for the geodynamic evolution. Journal of the Geological Society of Edinburgh, 91, 235–243.
Poller, U., Janák, M., Kohút, M. & Todt, W. (2000). Early Variscan magmatism in the Western Carpathians: U–Pb zircon data from granitoids and orthogneisses of the Tatra Mountains (Slovakia). International Journal of Earth Sciences 89, 336–349.[CrossRef][Web of Science]
Powell, R. & Holland, T. J. B. (1994). Optimal geothermometry and geobarometry. American Mineralogist 79, 120–133.[Abstract]
Randolph, A. D. & Larson, M. A. (1971). Theory of Particulate Processes. New York: Academic Press.
Reimold, W. U. (1991). The geochemistry of pseudotachylites from the Vredefort dome, South Africa. Neues Jahrbuch für Mineralogie, Abhandlungen 162, 151–184.[Web of Science]
Sack, R. O., Carmichael, I. S. E., Rivers, M. & Ghiorso, M. S. (1980). Ferric–ferrous equilibria in natural silicate liquids at 1 bar. Contributions to Mineralogy and Petrology 75, 369–376.[CrossRef]
Sahagian, D. L. & Proussevitch, A. A. (1998). 3D particle size distribution from 2D observations: stereology for natural applications. Journal of Volcanology and Geothermal Research 84, 173–196.[CrossRef][Web of Science]
Saltikov, S. A. (1967). The determination of the size distribution of particles in an opaque material from a measurement of the size distribution of their sections. In: Elias, H. (ed.) Second International Congress for Stereology. Berlin: Springer, pp. 163–173.
Sarda, P. & Graham, D. (1990). Mid-ocean ridge popping rocks: implications for degassing at ridge crests. Earth and Planetary Science Letters 97, 268–289.[CrossRef][Web of Science]
Sato, M. & Wright, T. L. (1966). Oxygen fugacities directly measured in volcanic gases. Science 153, 1103–1105.
Shaw, H. R. (1972). Viscosities of magmatic silicate liquids: an empirical method of prediction. American Journal of Science 272, 870–893.[Abstract]
Shimamoto, T. & Nagahama, H. (1992). An argument against the crush origin of pseudotachylytes based on the clast-size distribution. Journal of Structural Geology 14, 999–1006.[CrossRef][Web of Science]
Sibson, R. H. (1975). Generation of pseudotachylyte by ancient seismic faulting. Geophysical Journal of the Royal Astronomical Society 43, 775–794.[Web of Science]
Sperner, B. (1996). Computer programs for the kinematic analysis of brittle deformation structures and the Tertiary tectonic evolution of the Western Carpathians (Slovakia). Tübinger Geowissenschaftliche Arbeiten, Reihe A 27, 1–120.
Spray, J. G. (1987). Artificial generation of pseudotachylyte using friction welding apparatus: simulation of melting on a fault plane. Journal of Structural Geology 7, 49–60.
Spray, J. G. (1992). A physical basis for the frictional melting of some rock-forming minerals. Tectonophysics 204, 205–221.[CrossRef][Web of Science]
Spray, J. G. (1993). Viscosity determinations of some frictionally generated silicate melts: implications for fault zone rheology at high strain rates. Journal of Geophysical Research 98(B5), 8053–8068.
Swanson, M.T. (1992). Fault structure, wear mechanisms and rupture processes in pseudotachylyte generation. Tectonophysics, 204, 223–242.[CrossRef][Web of Science]
Swanson, S. E. (1977). Relation of nucleation and crystal-growth rate to the development of granitic textures. American Mineralogist 62, 966–978.[Abstract]
Thompson, L. M. & Spray, J. G. (1996). Pseudotachylyte petrogenesis: constraints from the Sudbury impact structure. Contributions to Mineralogy and Petrology 125, 359–374.[CrossRef][Web of Science]
Watson, E. B. (1977). Partitioning of manganese between forsterite and silicate liquid. Geochimica and Cosmochimica Acta 41, 1363–1374.[CrossRef]
Zieg, M. J. & Marsh, B. D. (2002). Crystal size distributions and scaling laws in the quantification of igneous textures. Journal of Petrology 43, 85–101.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||