The Current Status of Thermobarometry in Metamorphic Rocks E. J. Essene
Total Page：16
File Type：pdf, Size：1020Kb
Downloaded from http://sp.lyellcollection.org/ by guest on September 24, 2021 The current status of thermobarometry in metamorphic rocks E. J. Essene S U MMA RY: Information on pressure (P) and temperature (T) is a fundamental aspect of research on metamorphic terrains. Unfortunately, many workers employ thermo- barometers that are not experimentally calibrated, are insensitive or too sensitive to P- T changes, depend on a priori assumptions of water pressure (such as most petrogenetic grids), or are rapidly reset on cooling. Many systems are based on inaccurate thermo- dynamic data, involve solids with inadequately characterized structural states, neglect effects of thermal expansion and compressibility, or require long extrapolations in P-T-X space. For instance, application of the widely used garnet-clinopyroxene KD thermometer may require extrapolation to temperatures where current thermodynamic models of pyroxenes and garnets remain uncertain. Current versions of the Mg/Fc exchange thermometer for biotite-garnet involve substantial compositional extrapolations for many applications and the biotite is easily reset while cooling from higher T. The most widely employed barometer is based on dilution of the reaction grossular + kyanite + quartz = anorthite, but failure to correct molar volumes for P- T-X may yield systematic errors of 1-2 kbar for barometry of crustal metamorphites. Application of this barometer to rocks equilibrated at T < 600-650°C is presently unwarranted in view of unknown a-X relations of garnets and plagioclases at these T. However, by careful selections, thermo- barometry may be accurate to +50°C and + 1 kbar in many metamorphic terrains if a variety of different equilibria can be applied. Well-calibrated barometers that are useful for T > 600-650°C rely on continuous reactions based on equilibria such as almandine + rutile = ilmenite + sillimanite + quartz, garnet + quartz = ferrosilite + plagioclase, garnet + futile = ilmenite + anorthite + quartz, and almandine + sillimanite = hercynite + quartz. An extensive survcy of the recent literature on thermobarometry of individual metamorphic facies reveals the range of P-T encountered in each facies. Temperature estimates are in good agreement with the inferences of Turner (1968). Barometry reveals that the blueschist, amphibolite and granulite facies give way to the eclogite facies over the pressure range of 12-16 kbar. This paper provides an update of a review of Thermodynamic data base thermobarometry (Essene 1982). There will be minimal overlap of discussion although thematic When unravelling the effects of P- T-X on the duplication cannot be eliminated entirely. I shall thermodynamics of minerals, it is imperative emphasize the most recent literature and at- that an accurate thermodynamic data base be tempt to provide a critique of present-day petro- available for the minerals that occur in meta- logical practices in acquiring thermobarometric morphic equilibria. While a comprehensive dis- data on rocks of specific metamorphic facies. cussion of this topic is beyond the scope of this Variations in pressure (P) and temperature (T) paper, it must be addressed in order to correct with time (t) will be considered elsewhere adequately for the effects of solid solution in in this volume, although these variations are natural and synthetic systems. For compilation usually deduced by the application of thermo- of thermodynamic properties of solids and barometry to zoned minerals. In the discussion fluids, the reader is referred to Clark (1966), that follows, 60 equilibria that may be useful for Robie et al. (1966, 1978), Burnham et al. (1969), thermobarometry are listed with reactants as Hultgren et al. (1973), Mills (1974), Helgeson the high pressure (and/or low temperature) et al. (1978), Barton & Skinner (1979), Jacobs side, followed by recent experimental refer- & Kerrick (1981), Kerrick & Jacobs (1981), ences on them. At the beginning of each section, Pankratz (1982), G. R. Robinson et al. (1982), a representative set of references is provided Chase et al. (1985), Berman et al. (1986), and for recent (1980-1988) applications of thermo- Berman (1988). There is no doubt that these barometry in a particular facies. Many values are becoming more accurate with time, additional references to the literature of meta- but data from different sources cannot be morphic petrology before 1980 are given combined with impunity because they may be by Mueller & Saxena (1977), Essene (1982), based on different sets of data for entropy, P. Robinson et al. (1982) and Hyndman (1985). enthalpy or volume. It is important to use From DALY, J. S., CLIFF, R. A. & YARDI.~V, B. W. D. (eds) 1989, Evolution of Metamorphic Belts, Geological Society Special Publication No. 43, pp. 1-44. Downloaded from http://sp.lyellcollection.org/ by guest on September 24, 2021 2 E.J. Essene self-consistent data sets to avoid introducing Ca3AI2Si3012 + SiO2 = CaAI2Si2Os systematic errors. However, serious errors may + 2CaSiO3 persist even in self-consistent thermodynamic (WAGS, Newton 1966, Gasparik 1984, data sets because of reliance on a single erroneous Chatterjee et al. 1984) data set. Thermodynamic parameters should quartz = coesite (3) always be tested against carefully reversed experiments to evaluate their ability to reproduce SiO 2 = SiO 2 (Mirwald & Massonne 1980, Bohlen & the reaction of interest, and thermodynamically Boettcher 1982) calibrated thermobarometers must be regarded with scepticism until evaluated against ex- jadeite + quartz = albite (4) perimentally based equilibria. NaAISi206 + SiO2 = NaAISi3Os (Johannes et al. 1971, Holland 1980) Volume changes of solids almandine + rutile = ilmenite + kyanite or (5) sillimanite + quartz Calculations of solid-solid equilibria must Fe3A12Si3012 + 3TIO2 = 3FeTiO3 + AI2SiO5 include corrections of the volumes for changes + 2SIO2 in pressure (or compressibility: Birch 1966, (GRAIL, Bohlen et al. 1983a; Fig. 3) Vaidya et al. 1974, Hazen & Finger 1982) and temperature (or thermal expansion: Skinner The assumption of constant AVs may be inad- 1966, Hazen & Finger 1982), even if these data equate for calculation of equilibria, e.g. it causes need be approximated (Helgeson et al. 1978, errors of as much as 1-2 kbar in the location of Powell & Holland 1985). Systematic errors may some reactions. Disparate thermodynamic data be generated in calculations assuming that AV~ will thus be generated from experimentally de- is constant at all P- T. The use of equilibria that rived equilibria depending upon the assump- have been derived with a constant AVP = AV~298 tions used to determine the volumes of solids, (e.g. Chatterjee et al. 1984, Berman et al. 1986) and use of these same equilibria for barometry must be regarded as suspect, especially for may yield significantly different results de- thermobarometry employing solid-solid reac- pending on the details of the calculation of AVs. tions with small changes in entropy and/or vol- ume. Even when volume has been corrected for Thermodynamic properties of solid solutions the effects of pressure and temperature, small errors may still persist, because compressibility Thermodynamic models of activity-com- is usually measured at room temperature and position relations are requisite for application thermal expansion at room pressure, and the of barometers involving solid solutions and for volume calculated at P-T will depend on the correction of petrogenetic grids for components path chosen for the calculation. For consistency, encountered in natural systems. Unfortunately, it is recommended that volume be calculated at there is still no consensus concerning the a-X 1 bar and T followed by volume at P and T relations of even the common anhydrous using the compressibility data measured at 25°C. mineral groups such as spinel, ilmenite, plagio- Even though this is an arbitrary procedure, it is clase, alkali feldspar, garnet, orthopyroxene, convenient for calculations of different press- clinopyroxene, olivine and carbonate. Ganguly ures at constant temperature and is congruent & Saxena (1987) reviewed mixing models and with a 1 bar, T standard state for fluids. Any thermodynamic properties of selected mineral errors produced are unlikely to be significant at solutions. Spencer & Lindsley (1981), Sack P < 20 kbar for most reconstructive trans- (1982), Engi (1983), Lehman & Roux (1984), formations. Examples of the result of different Oka et al. (1984), O'Neill & Navrotsky (1984), assumptions for AVs upon the placement of Mattioli et al. (1987) and Shulters & Bohlen calculated equilibria are illustrated in Fig. 1 for (1987) presented mixing models for spinel solid the following equilibria: solutions. Spencer & Lindsley (1981) and Pownceby et al. (1987) evaluated the mixing grossular + kyanite + quartz = anorthite (1) properties of ilmenite solid solutions based Ca3A12Si3012 + 2A12SIO5 + SiO2 = on experimental data. Orville (1972), Kerrick 3CaAl2Si208 & Darken (1975), Newton et al. (1980), (GASP, Goldsmith 1980, Gasparik 1984, Kotel'nikov et al. (1981), Newton & Haselton Chatterjee et al. 1984, Koziol & Newton 1988a; (1981) and Blencoe et al. (1982) assessed plagio- Fig. 2) clases, although Carpenter & Ferry (1984) grossular + quartz = anorthite (2) questioned the standard state of CaAI2Si208 + wollastonite appropriate for intermediate plagioclases. Downloaded from http://sp.lyellcollection.org/ by guest on September 24, 2021 Thermobarometry in metamorphic rocks 3 Many authors, including Haselton et al. (1983),