Analytical methods. Electron Microprobe

Transcription

1 Analytical methods Electron Microprobe The Cameca SX-100 at University of Manitoba was used to acquire X-ray maps and quantitative analyses of garnet and rutile. Compositional variations in the garnet porphyroclasts are not evident under back-scattered electron images. Therefore, X-ray maps were used to guide locations of line profiles for the modelling of diffusion profiles. To minimize cut-effect, two or three large garnet grains in each thin-section were selected for X-ray mapping. Mapping for Fe Kα, Mg Kα, Mn Kα, and Ca Kα were done at a 15 kv accelerating voltage, a 100 na beam current, 2 µm beam size and 20 to 40 ms dwell times. Garnet rims not in contact with recrystallized garnet grains were selected for line profile analyses. Care was also taken to avoid garnet rims with rounded edges, which may produce spurious zoning profiles. Some line profiles were acquired perpendicular to cracks developed within garnet porphyroclasts. Line profiles were done at a 1 µm step using a 15 kv accelerating voltage, a 20 na beam current, 1 µm beam size, and 20 seconds counting times on Si Kα, Al Kα, Fe Kα, Mg Kα, Mn Kα and 40 seconds on Ca Kα. These analytical conditions give better than 2% relative standard deviation (1 σ) for all elements analyzed except for Mn Kα. Background counts were measured once in the beginning of line profiles and the same background counts were applied to the rest spot analyses. Zirconium contents in rutile can be used as a proxy of temperature when it is buffered with quartz and zircon (Watson et al., 2006; Tomkins et al., 2007). Zr Lα was measured with a 15 kv accelerating voltage, a 200 na beam current, and 1 micron beam size. Three PET diffracting crystals were dedicated to the determination of Zr Lα with counting times of 4 minutes on the peak using natural zircon as a standard. These analytical conditions resulted in detection limits of about 36 ppm. For each sample, 6-13 spots on separate rutile crystals were analyzed and the weighted average was used to calculate temperature (The results are presented in Table DR1). 40 Ar- 39 Ar Irradiation of samples for 40 Ar- 39 Ar analysis was carried in facility X33 or X34 of the HIFAR reactor of the Australian Nuclear Science and Technology Organisation. The K-feldspar sample was irradiated for 524 hours with biotite standard GA 1550 (97.9 Ma in age; McDougall & Roksandic, 1974) in the central position as described by McDougall & Harrison (1988). Cadmium shielding 0.2 mm thick was used to minimize interference from thermal neutrons. Sample cans were inverted 180 exactly half way through the irradiation in order to lessen the effects of the neutron flux gradient in HIFAR. K-feldspar, was analysed by progressively

2 increasing the temperature from 450 to 1450 C in a resistance heated furnace, with the gas from each extraction step exposed to Zr-Al getters to remove active gases. Isothermal steps at low temperature (below 800 C) were carried out to extract any excess argon present in fluid inclusions while the higher temperature isothermal steps were done to extract radiogenic argon before melting at 1100 C. Argon was subsequently analysed using a VG Isotech MM1200 gas source mass-spectrometer operating in static mode. Corrections for argon produced by neutron interactions with calcium and potassium were made using the factors determined by Tetley et al. (1980). The 40 K abundance and decay constants recommended by the IUGS Subcommission on Geochronology were used (Steiger and Jäger, 1977). REFERENCES CITED McDougall, I. and Harrison, T.M., Geochronology and thermochronology by the 40 Ar- 39 Ar method. Oxford University Press, New York. McDougall, I., and Roksandic, Z., Total 40 Ar- 39 Ar fusion ages using HIFAR reactor. Journal of the Geological Society of Australia, 21, McDougall, I., and Schmincke, H.-U., Geochronology of Gran Canaria, Canary Islands: Age of shield building volcanism and other magmatic phases. Bulletin Volcanologique, 40, Steiger, R., and Jäger, E., Subcommission on geochronology: Convention on the use of decay constants in geo- and cosmochronology. Earth and Planetary Science Letters, 36, Tetley, N.I., McDougall, I. and Heydegger, H.R., Thermal neutron interferences in the 40 Ar/ 39 Ar dating technique. Journal of Geophysical Research, 85, Tomkins, H.S., Powell, R., and Ellis, D.J., 2007, The pressure dependence of the zirconium-inrutile thermometer: Journal of metamorphic Geology, v. 25, p , doi: /j x. Watson, E.B., Wark, D.A., and Thomas, J.B., 2006, Crystallization thermometers for zircon and rutile: Contributions to Mineralogy and Petrology, v. 151, p , doi: /s

3 COOLING HISTORY PRIOR TO THE PETERMANN OROGENY (~550 MA) In the Musgrave Block there is generally a lack of pressure indicators (e.g. decompression textures) that might aid the reconstruction of an exhumation path after granulite facies metamorphism. In the west, similar granulites show decompression textures (Clarke & Powell, 1991). In addition, some granulites were overlain by felsic volcanics at ~1070 Ma (Glikson et al., 1995). Hence, there is an indication that some parts of the Musgrave Block were exhumed and exposed at the surface by ~1070 Ma. In the Amata-Ernabella region, there is no compelling evidence for the rocks being at the surface prior to the Petermann Orogeny at ~550 Ma. Indirect evidence for the level of the crust at which the granulites were residing is arrived at by modelling the rock compositions of the Alcurra (~1070 Ma;, 1991; Zhao & McCulloch, 1993; Schmidt et al., 2006) and Amata Dyke (~830 Ma; Zhao et al., 1994; Wingate et al., 1998) Swarms using the program MELTS (Ghiorso & Sack, 1995). The dyke compositions are assumed to represent liquid compositions, an assumption that is plausible as chilled margins, which are free of phenocrysts, have the same compositions as the coarser grained dyke interiors (, 1990). This suggests that the dykes did not fractionate or assimilate country rock during crystallisation, an observation consistent with quench textures, and by the lack of felsic segregations and xenoliths in the dykes. The mineral assemblages and chemical compositions of the dykes used for modelling with some examples from the output from MELTS, are presented in Tables 1 & 2. Although, the H 2 O contents of the dykes are not known, similar rocks may contain up to ~0.5 wt% (Sobolev & Chaussidon, 1996). In the model, the melt was allowed to crystallise rapidly (100 C steps) until a stable solid assemblage in equilibrium with a very small proportion (<5%) of felsic residual liquid, inferred to represent the mesostasis fraction, was produced. Crystallisation was modelled at a range of pressures and the modelled assemblage was then compared with the mineralogy present in the rocks. The Alcurra Dyke Swarm would not crystallise olivine (according to MELTS) unless it intruded at shallow levels (<0.5 GPa). Modelling of the cooling of the Amata dykes with 0.2 & 0.3 wt% H 2 O indicated crystallisation of small amounts of garnet at P 0.6 GPa (Table 2). Since garnet is absent from these rocks, they are inferred to have crystallised at pressures <0.6 GPa. Thus, thermodynamic crystallisation modelling using MELTS suggests that the mafic dykes intruded at levels in the crust of 0.5 GPa, at slightly shallower levels than that at which granulite-facies metamorphism took place. The granulites resided at this level until at least ~800 Ma. 1

4 INFERRED P-T TRAJECTORY The inferred P-T trajectory for the Fregon Subdomain, starting during granulite facies metamorphism and prior to the ~550 Ma Petermann Orogeny, can be summarized in Figure 1 after combining the cooling and exhumation histories derived above. REFERENCES CITED Camacho, A., An isotopic study of deep-crustal orogenic processes: Musgrave Block, central Australia: Canberra, Australian National University. Camacho, A., Simons, B. and Schmidt, P.W., Geological and palaeomagnetic significance of the Kulgera Dyke Swarm, Musgrave Block, N.T., Australia. Geophysical Journal International, 107, Camacho, A., Simons, B., Schmidt, P.W. and Gray, C.M., Middle to late Proterozoic mafic dykes of central Australia. 2 nd International Dyke Conference, IGCP project 257, Adelaide, Excursion Guide. Clarke, G.L. and Powell, R., Decompressional coronas and symplectites in granulites of the Musgrave Complex, central Australia. Journal of metamorphic Geology, 9, Ghiorso, M.S. and Sack, R.O., Chemical mass transfer in magmatic process IV. A revised and internally consistent thermodynamic model for the interpolation and extrapolation of liquid-solid equilibria in magmatic systems at elevated temperatures and pressures. Contributions to Mineralogy and Petrology, 119, Glikson, A.Y., Ballhaus, C.G., Clarke, G.L., Sheraton, J.W., Stewart, A.J. and Sun, S.-S., Geological framework and crustal evolution of the Giles mafic-ultramafic complex and environs, western Musgrave Block, central Australia. Australian Geological Survey Organisation Journal of Australian Geology and Geophysics, 16, Schmidt, P.W., Williams, G.E., Camacho, A. and Lee, J.K.W., Assembly of Proterozoic Australia: implications of a revised pole for the ~1070 Ma Alcurra Dyke Swarm, central Australia. Geophysical Journal International, 167, Sovolev, A.V. and Chaussidon, M., H 2 O concentrations in primary melts from suprasubduction zones and nid-ocean ridges: Implications for H 2 O storage and recycling in the mantle. Earth and Planetary Science Letters, 137, Zhao, J.X. and McCulloch, M.T., Sm-Nd mineral isochron ages of Late Proterozoic dyke swarms in Australia: evidence for two distinctive events of mafic magmatism and crustal extension. Chemical Geology, 109, Zhao, J.X., McCulloch, M.T. and Korsch, R.J., Characterisation of a plume-related ~800 Ma magmatic event and its implications for basin formation in central-southern Australia. 2

5 Earth and Planetary Science Letters, 121, Wingate, M. T. D., Campbell, I. H., Compston, W., and Gibson, G. M., Ion microprobe U Pb ages for Neoproterozoic basaltic magmatism in south-central Australia and implications for the breakup of Rodinia. Precambrian Research, 87,

6 FIGURE 0.2 Pressure (GPa) Greenschist-facies shear zones Eclogite-facies shear zones Mafic dykes Deformation at granulite-facies and granitoid intrusions Time (Ma) Figure 1. Pressure versus time diagram showing the path granulite-facies rocks in the Musgrave Block that underwent eclogite-facies metamorphism took from ~1170 Ma to the present. After Camacho (1997). 4

7 TABLES Table 1. Geochemistry of selected mafic dykes from the Musgrave Block. Amata Suite Alcurra Dyke Swarm (~800 Ma) (~1050 Ma) Sample W-7 W S-400c W-153a SiO TiO Al2O FeOtotal MgO CaO Na2O K2O P2O Trace elements (ppm) Ba Rb Sr Pb Zr Nb Y La Ce Nd Cr Ni Ga CIPW norm Q Ab An Di Hy Ol Ap

8 Samples W-7, W-98 and W-153a were analysed at the AGSO XRF laboratory. Sample from Zhao & McCulloch (1993) and Sample S-400c from (1991). Table 2. Summary of output from MELTS for mafic dykes from the Musgrave Block. Sample number S-400c S-400c S-400c Wt% H 2 O Pressure (GPa) Liquidus temp ( C) Liquid Clinopyroxene Orthopyroxene Feldspar Quartz Olivine Garnet Apatite Spinel

9 Garnet Modeling The dependence of the garnet self-diffusion coefficient ( ) of an element i, on temperature (T), pressure (P), and oxygen fugacity ( ) is given by: where is the pre-exponential constant, Q is activation energy for diffusion at 1 bar, is activation volume, and is the oxygen fugacity of the graphite-oxygen equilibrium (Chakraborty and Ganguly, 1992). The coexistence of rutile, ilmenite and magnetite (Fig. DR2) constrains the oxygen fugacity (fo 2 ) in the eclogite-facies shear zones of the Musgrave Block to be ~ (Fig. DR4; Berman, 2007). Using a thermal spike of T = 657 C (pooled weighted mean temperature obtained from rutiles), P = 1.2 GPa and log fo 2 = 14.9, we calculated the diffusion coefficients for garnet in two ways, depending on the number of diffusion parameters determined in each study. 1/ Single diffusion coefficient: DCa-(Fe-Mg) interdiffusion (Vielzeuf et al., 2007). The experimental data were treated in terms of an interdiffusion coefficient that is independent of composition. Because the garnet composition are within the range of those used to derive the coefficients, realistic durations can be estimated using the DCa-(Fe-Mg) interdiffusion coefficient. 2/ Multicomponent diffusion: Diffusion in minerals commonly involves the simultaneous flow of more than two components, with the flux of any component depending not only on its concentration, but also on those of the other diffusing species. The interdiffusion coefficients ( ), can be calculated, using Carlson s (2006) and Perchuk et al. (2009) self-diffusion

10 coefficients, with the following equation developed by Lasaga (1979) simplified for ideal ionic solutions and for species of the same electrical charge where is the self-diffusion coefficient for component i, is the Kronecker delta ( = 0 when i j and 1 when i = j) and n is the total number of components (four in this instance). Numerical solution of the diffusion equation Assuming a planar grain boundary, the multi-component diffusion equation may be posed for the ith component as which forms a partial differential equation containing one spatial dimension, plus time (Spear, 1993). We solved the equation numerically using the method of lines (Schiesser, 1991). In this approach, the spatial variable is digitized using finite-difference methods, forming a system of ordinary differential equations solvable by Runge-Kutta integration. The spatial variable is sampled at equal intervals δx for points k = 1,2,,N at which the spatial derivatives of C i are posed as: where, i is the component index and k is the index of spatial points. In practice, we used a sample interval δx of 1 µm, and a total model length of 200 µm, with the grain boundary at the

11 center of the model (thus, concentration varies on both sides of the interface). The resulting large system of ordinary differential equations was solved using a Runge-Kutta solver included in the MATLAB software package. Simulations were run for up to 1,500,000 yr, with concentration values recorded for each time steps. Simulated concentrated curves were then compared to data to find the diffusion time most consistent with the data set. The fits were made by a modified misfit function to fit all data points within the error bars. In this scheme, the misfit was taken to be zero if the diffusion curve intersected the error bar around an individual measurement, and non-zero only outside of the error bar, making the misfit insensitive to the position of the curve within the error tolerance. If it were possible to fit all points within the error bars, the minimum misfit using this method would be zero. In practice, fitting all points within the estimated error was not achievable, and the times obtained using the modified misfit did not significantly differ from the least-square minimization of the deviations in CaO, MgO and FeO (wt %). The curves we calculated using an isothermal pulse fit well to the measured concentrations (Fig.3; Fig. DR5) suggesting that temperature and thus, concentration were not varying greatly at the grain boundary (e.g. Liermann & Ganguly 2001; Ganguly et al., 2001). REFERENCES CITED Berman, R.G., 2007, wintwq (version 2.32): a software package for performing internallyconsistent thermobarometric calculations: Geological Survey of Canada, Open File 5462, p. 41. Carlson, W.D., 2006, Rates of Fe, Mg, Mn and Ca diffusion in garnet: American Mineralogist, v. 91, p. 1-11, doi: /am Chakraborty, S., and Ganguly, J., 1992, Action diffusion in aluminosilicate garnets: experimental determination in spessartine-almandine diffusion couples, evaluation of effective binary

12 diffusion coefficients, and applications: Contributions to Mineralogy and Petrology, v. 111, p , doi: /BF Ganguly, J., Hensen, B. J. & Cheng, W Reaction texture and Fe-Mg zoning in granulite garnet from Søstrene Island, Antarctica: Modeling and constraint on the time scale of metamorphism during the Pan-African collisional event. Prodeedings of the Indian Academy of Science (Earth and Planetary Science), 110, Lasaga, A.C., 1979, Multicomponent exchange and diffusion in silicates: Geochimica et Cosmochimica Acta, v. 43, p , doi: / (79) Liermann, H-P. and Ganguly, J., Compositional properties of coexisting orthopyroxene and spinel in some Antarctic diogenites: Implications for thermal history. Earth and Planetary Science Letters, 36, Perchuk, A.L., Burchard, M., Schertl, H.-P., Maresch, W.V., Gerya T.V., Bernhardt, H.-J., and Vidal, O., 2009, Diffusion of divalent cations in garnet: multi-couple experiments: Contributions to Mineralogy and Petrology, v. 157, p , doi: doi: /s Spear, F., 1993, Metamorphic Phase Equilibria and Pressure-Temperature-Time Paths: Washington, Mineralogical Society of America, 799 p. Vielzeuf, D., Baronnet, A., Perchuck, A.L., Laporte, D., and Baker, M.B., 2007, Calcium diffusivity in alumino-silicate garnets: an experimental and ATEM study: Contributions to Mineralogy and Petrology, v. 154, p , doi: /s x

13 Supplementary Figures Fig. DR1. Photomicrographs of the same garnet grains shown in Figure 2. Mylonitic foliation deflected around relict granulite-facies garnet (Grt1) with fringes of new garnet growth. New garnet (Grt2), rutile (Rt), magnetite (Mt) and trace biotite (Bt2) form in pressure shadows and in the foliation. Sample W-97 contains a relict granulite-facies biotite inclusion (Bt1) and the labelled relict plagioclase (Pl) contains a few inclusions of clinozoisite. Plane polarised light. Fig. DR2. Backscattered electron (BSE) image of Neoproterozoic eclogite facies garnet (Grt2), ilmenite (Ilm), magnetite (Mag) and rutile (Rt) crystallizing (a) in pressure shadows of relict Proterozoic garnet (Grt1; sample W-97). (b) Close up showing rutile (with thin intergrown needles of ilmenite) and magnetite (with ilmenite exsolution lamellae) in textural equilibrium. Abbreviations after Kretz (1983). Fig. DR3. Quartz-feldspar mylonite from the Davenport shear zone (sample W-97). (a) Relict granulite-facies plagioclase clasts (Pl1) show subgrain development. Small new grains (Pl2) formed along the margins of relict grains during dynamic recrystallization. Section parallel to the stretching lineation and normal to the foliation. Cross polarized light. (b) and (c) X-ray maps of the same area for Ca and Na, respectively. Note that the relict grains are higher in Ca and lower in Na (Ab 70 ) than the new grains (Ab 80 ) that crystallized at eclogite facies. Rainbow color scheme, with red representing higher concentrations than blue. Fig. DR4. Oxygen fugacity at 1.2 GPa for graphite - O 2 CO 2, ilmenite-o 2 magnetite + rutile, and ilmenite-o 2 hematite + rutile. Curves calculated using TWQ (Berman, 2007). The filled circle best approximates the fo 2 conditions experienced by the quartzofeldspathic rocks during high-pressure metamorphism. Fig. DR5. MgO and MnO (wt %) profiles across relict granulite-facies garnet and multicomponent (FeO-MgO-CaO) diffusion modelling results in (a) sample W-68 from the centre of the Davenport shear zone. Garnet composition: x = 0 (Alm 53 Prp 29 Grs 16 Sps 2 ); x = 50, (Alm 59 Prp 34 Grs 5 Sps 2 ), and (b) sample W-97 from the edge of the Davenport shear zone. Garnet composition: x = 0 (Alm 52 Prp 28 Grs 28 Sps 1 ); x = 49, (Alm 59 Prp 35 Grs 5 Sps 1 ). Grey solid lines represent the best-fit solution of 1.3 Myr and 0.26 Myr, respectively. Error bars are 2σ.

14 W-68 Rt Grt2 Grt1 Rt Bt2 Pl 500 µm W-97 Bt1 Grt2 Bt2 Grt1 500 µm Rt SuppFigure DR1

15 GSA DATA REPOSITORY (a) Grt2 Grt1 Mag 500 µm Rt (b) Rt Ilm Mag 30 µm Grt1 SuppFigure DR2

16 GSA DATA REPOSITORY (a) Pl1 500 µm (b) (c) Pl2 Grt1 500 µm SuppFigure DR3

17 -10 Log fo 2 (at 12 kbar) log fo 2 = Mt + 6 Rt 6 Ilm + O2 CO2 C + O2 2 Hm + 4 Rt 4 Ilm + O Temperature ( o C) SuppFigure DR4

18 A W68 Garnet 1 LP1 MgO 5 4 MgO (wt%) MnO MnO (wt%) B W97 Garnet 1 LP1 Distance (µm) MgO MgO (wt%) MnO MnO (wt%) Distance (µm) SuppFigure DR5

19 Supplementary Tables Table DR1. Summary of zirconium concentrations and calculated temperatures, using the using Zr-in rutile calibration of Tomkins et al. (2007) for mylonitized rocks in the Musgrave Block, central Australia. Table DR2. Time (in Myr) calculated using a single diffusion coefficient (Vielzeuf et al., 2007) and multicomponent diffusion (Carlson, 2006; Perchuk et al., 2009) approach for the development of the compositional profile developed ~550 Myr ago during high-strain deformation at eclogite-facies in relict Proterozoic granulite-facies garnet. Strain rates (1) calculated using durations using the diffusion coefficient of Vielzeuf et al. (2007) and (2) Perchuck et al. (2009). Table DR3. 40 Ar/ 39 Ar step heating data for K-feldspar from a mylonitized granite (sample W-25b) in the North Davenport shear zone (sample weight = 3.59E-3 g; J = ).

20 Table DR1. Sample Zr* ± (2 sigma) Temp number (ppm) (ppm) N MSWD ( C) Felsic gneiss W-45a W W W-97c W W-197e Mafic dykes W W W-142d W-153a Mafic gneiss W-97b Weighted Mean * Weighted mean Errors are 2 sigma

21 Table DR2. Sample Ca (Fe-Mg) 1 (Multicomponent) 2 (Multicomponent) 3 rate rate Fe-Mg-Ca Fe-Mg-Ca Average strain Average strain number Time (Ma) Time (Ma) Time (Ma) (s -1 ) x (10-12 ) (s -1 ) x (10-12 ) Equal weight Equal weight 2 3 W45 area1 LP area1 LP area1 LP W68 area1 LP area2 LP area2 LP area4 LP area5 LP W88 area1 LP W97 area1 LP area1 LP area2 LP area4 LP W104 area1a W122 area W170 area1 LP area1 LP W-181 area Vielzeuf et al. (2007) 2 Carlson (2006) 3 Perchuk et al., (2009)

22 Table DR3. Temp time Ar36 ± % Ar37 ± % Ar39 ± % Ar40 ± % % Ar40* Ar40*/Ar39(K) Cumulative Age ± Ca/K ( C) (mins) (mol) (1σ) (mol) (1σ) (mol) (1σ) (mol) (1σ) Ar39 (%) (Ma) (Ma; 1σ) E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-02