Cation Distribution and Equilibration Temperature of Amphiboles from the Sittampundi Complex, South India

Amphiboles are utilized as indicators of temperature, pressure, volatile content, and oxidation state by the petrologist; while to a mineralogist the amphiboles provide instructive examples of structural phase transitions, cation ordering over multiple crystallographic sites, and a variety of atomic substitutions (Ghiorso & Evans, 2002). Relating the chemistry of calcic-amphibole to the P-T conditions of crystallization is a daunting task, because the amphibole crystal structure consists of a variety of cation sites, accommodating elements of a wide range of ionic radii and valences.


An Introduction to the Study of Mineralogy 40
The present work was undertaken to investigate the chemical characteristic of calcic and magnesian amphiboles from upper amphibolite to granulite facies rocks in the Sittampundi complex. In addition, main objective of this work is to characterize iron ions of different valence states and their distribution within the crystal lattice, both in calcic and magnesian amphiboles of amphibolite, meta-anorthosite and chromitite respectively, using 57 Fe Mössbauer spectroscopy and estimate the equilibration temperature attained by the host rock.

Geological setting
Sittampundi Anorthosite Complex occurs as a layered igneous body. The area forms a part of the granulite terrain of South India. Major rock types are chromitite bearing meta-anorthosite, amphibolite, basic granulite, two pyroxene granulite, leptynite, biotite gneiss and pink granite. In this complex, amphibolite and chromitite occur as small discontinuous bands/lenses within the meta-anorthosite. Basic granulite occurs as discontinuous but conformable layers within meta-anorthosite. Two-pyroxene granulite and pink granite occur within the biotite-gneiss. The former two are the later intrusive phases (Subramanium, 1956).
In the basic granulite and two-pyroxene granulite, calcic amphibole occurs as rim surrounding diopside. In meta-anorthosite, tschermakitic hornblende forms triple junction with plagioclase; and triple junctions are common between amphibole crystals in amphibolite.

Experimental methods
Electron microprobe analysis of the samples has been obtained by JEOL-733 superprobe (wavelength dispersive method) at 15 kV with beam current of 0.01 A and 2 m beam diameter and Cameca CAMEBAX SX 50 probe with digital computer PDP 11.53. Operating condition was 15 kV accelerating voltage with a beam current of 15 nA and beam size ~1m. The average spectrum count was compared with the natural standards.
The room temperature (RT) Mössbauer spectra were recorded in a conventional constant acceleration spectrometer (Wissel GmbH, Germany) with a 25 mCi Co/Rh source. Two mirror-imaged spectra were obtained in 512 out of 1024 channels. The finely powdered samples were pressed tight in plexi holders, and based on the iron content (known from EPMA) determined amounts of the samples were taken to optimise the thickness effect. The amount used for the samples are: 60 mg (sample 3c), 68 mg (sample 15b) and 92 mg (sample 30a). The velocity calibration was performed with respect to pure metallic iron (99.99%). The spectra were then fitted to the sum of Lorentzian functions with a non-linear least square fitting programme of Meerwall (1975) on an IBM personal computer. For each doublet the intensity and line-width of the peaks were constrained to be equal.

Chemical composition of Sittampundi amphiboles
Amphiboles have been classified according to Leake (1978). For this classification empirical Fe 3+ in calcic amphibole was estimated from EPMA chemistry following Schumacher (1991). Site occupancy of different elements in amphiboles (viz. calcic amphiboles, anthophyllite) was assigned on the basis of 23 oxygens following Papike (1988). We present here the composition of only three amphiboles (3c, 15b, 30a), the chemistry of which were determined employing both the EPMA and Mössbauer spectroscopy (Table 1)

Characterization of samples
Studied amphiboles were selected from different metamorphic rocks, such as amphibolite (sample 3c), meta-anorthosite (sample 15b), chromitite (sample 17c, 18b, 25, 29, 30a) and basic granulite (sample 6b, 13). Mössbauer analysis of two calcic amphiboles (sample 3c and 15b) and one anthophyllite (sample 30a) were considered in the present study. Sample 3c was taken from amphibolite occurring in the eastern part of the Sittampundi complex, near the Cauvery river bed of Kettuvelanpalaiyam. Sample 15b was taken from meta-anorthosite in the western side of Pamandapalaiyam, near the road between Cholasiramani and Tottiyamthottam. Sample 30a was taken from chromitite in the western side of Karungalpatti. The crystal chemistry of the last three amphiboles was determined from microprobe analyses combined with the results of Mössbauer spectroscopy. The studied amphibole species present in different rock types are identified as follows: i. Meta-anorthosite: ferroan pargasite ii. Chromitite: edenitic hornblende, magnesio-hornblende, tchermakitic hornblende, anthophyllite iii. Basic granulite: edenite, edenitic hornblende, pargasite, pargasitic hornblende, ferrohornblende, actinolite iv. Two-pyroxene granulite: edenitic hornblende, magnesio-hornblende v. Amphibolite: ferroan pargasitic hornblende , TiO 2 and Cr 2 O 3 of calcic amphiboles and anthophyllite in different rock types are given in Table 2. TiO 2 content of calcic amphibole in meta-anorthosite is low (0.21-0.36 wt%) but it is high in some basic granulite ( 2.07 wt%). TiO 2 content of calc amphiboles in two-pyroxene granulite, amphibolite and chromitite range between 0.90-1.42 wt%, 1.02-1.47 wt%, 0.47-1.09 wt%, respectively. Cr 2 O 3 content of calcic-amphiboles and anthophyllite in chromitite show wide variation (0.00-1.12 and 0.38-0.83 wt%, respectively), but in other rock types it is low (<0.34 wt%). The plot of Si vs. (Ca+Na+K) of Sittampundi calcic amphiboles, following Shido & Miyashiro (1959) and Binns (1965), show a cluster in the field of green and brown hornblende (belonging to the high amphibolite and granulite facies). In anthophyllite, the ratio of (Na+K) to tetrahedral Al is remarkably constant and it lies close to ¼. Thus, for each 0.1 Na ion per formula unit, there are 0.4 tetrahedral Al ions substituting for Si.

Amphiboles
Host

Mössbauer study and assignment
The room temperature (298 K) Mössbauer spectrum of the calcic amphiboles and anthophyllite show a close range of isomer shift (IS) of octahedral Fe 2+ ion (Table 3) and the assignment is made from the quadrupole splitting (QS) values. The quadrupole splitting of M1, M2 and M3 sites which are regular octahedra, completely overlap in the case of Mg-Fe amphibole, but chemistry becoming more complex with Ca, Na, Al and Fe 3+ the QS values tend to decrease (Goldman, 1979 (Skogby, 1987). With incorporation of more Al and Fe 3+ in amphibole structure the MO bond length of M2 site become smaller (Papike et al., 1969) and the QS for Fe 2+ in the M2 site would manifest a reduction compared to that of the M1 and M3 sites. However, Della Ventura et al. (2005) showed that because the M2 site is adjacent to the M4 site, the QS value change correlated directly with the LiNa substitution. In aluminous amphibole Fe 2+ is enriched in M3 site relative to M1, which host Mg. In calcic amphibole, the M4 octahedral site is most distorted and is preferentially occupied by larger cation such as Ca, Na, Mn although occurrence of Fe 2+ and Mg at M4 site has been postulated by some authors (Goldman & Rossman, 1977;Goldman, 1979;Skogby & Annersten, 1985;Skogby, 1987). Difficulty in determining Fe 2+ at M4 site arises because Ca occupies 85-95% of this site (Leake, 1968). However, a high degree of distortion of the M4 site is expected to produce large Fe 2+ crystal field splitting analogous to that seen in the distorted M2 site of orthopyroxene (Goldman & Rossman, 1977).

Calcic amphiboles
The room temperature (298 K) Mössbauer spectra (Figs. 1a, b) of the ferroan pargasitic hornblende (from amphibolite, sample 3c) and ferroan pargasite (from meta-anorthosite, sample 15b), show two asymmetric paramagnetic doublets of which the more intense peak lies at lower velocity region (  0.13 mm/s). The fitting of the Mössbauer spectrum was approached in two ways: viz. (a) four doublet fit (three doublets for octahedral Fe 2+ ions, and the other for octahedral Fe 3+ ), and (b) five doublet fit (three doublets for octahedral Fe 2+ ions, and two for octahedral and tetrahedral Fe 3+ ). The five doublet fit improves the chisquare value and narrows down the line-width (0.35-0.49 mm/s) from wider line widths of the four doublet fit. The Mössbauer parameters as well as the assignment of the doublets are presented in Table 3. The basis of assignment is stated below.
In five doublet fit, the outermost doublets (1-1) with IS 1.23 mm/s (for sample 3c) and 1.22 mm/s (for sample 15b) and corresponding QS values of 2.93 mm/s, 2.84 mm/s are assigned to Fe 2+ in M1 and M3 sites (Goldman, 1979;Schumacher, 1991 (Goldman, 1979;Schumacher, 1991). The innermost doublets (5-5) with IS 1.47 mm/s (for sample 3c), 1.42 m/s (for sample 15b) and corresponding QS 1.41 mm/s, 1.40 mm/s are assigned to Fe 2+ in the M4 site (Goldman, 1979;Skogby & Annersten, 1985;Skogby, 1987;Schumacher, 1991). Differing QS value arises from difference in magnetic field generated by the orbital momentum of Fe 2+ ion in those octahedral (M) sites giving rise to different orbital contributions H orb and H hf and also to the differences in covalency (Linares et al., 1983). The M4 site is bonded more covalently with Ca ion, with consequent decrease in the QS value of Fe 2+ at this site. The decrease can also be accounted by considering the NNN effect where one Fe 2+ is substituted by cations such as Mg 2+ or Ca 2+ and the super exchange in FeOFe bond, between octahedral sites (Goodenough, 1963). The entry of Al 3+ ion in the M2 site could also affect the QS value of the M4 site. The lower QS value of 1.41 mm/s for sample 3c (Al VI = 0.916 p.f.u.) and 1.40 mm/s for sample 15b (Al VI = 1.290 p.f.u.) of M4 site may also signify greater distortion caused by faster rate of its formation (Jenkins, 1987;Skogby, 1987) during cooling.
The IS value of the other Fe 3+ doublet (2-2) suggests a tetrahedral co-ordination rather than Fe 3+ at octahedral site. So the doublets with IS 0.02 mm/s (for sample 3c), 0.07 mm/s (for sample 15b) and the corresponding QS 0.35 mm/s, 0.38 mm/s are assigned to Fe 3+ in the less-distorted tetrahedral site. Extremely low value of IS for sample 3c is due to Fe 3+ at an almost regular tetrahedron.

Anthophyllite
The room temperature (298 K) Mössbauer spectrum (Fig. 1c) of anthophyllite occurring in chromitite (sample 30a) consists of two asymmetric, paramagnetic doublets with more intense peak at lower velocity region ( 0.25 mm/s). The inner, more intense doublet can, on the basis of the crystal structure determination, be assigned to Fe 2+ in M4; and the outer doublet to Fe 2+ in M1, M2 and M3 sites (Seifert & Virgo, 1973). Similar to the calcic amphiboles, the fitting of Mössbauer spectrum was approached likewise in two ways: (a) four doublet fit and (b) five doublet fit. The five doublet fit here again showed better Mössbauer parameter ( 2 ) than those of the four doublet fit. The Mössbauer parameter as well as the assignment of the iron doublet to specific crystallographic site of anthophyllite, following calcic amphibole assignment, is presented in It should be mentioned that the chi-square values of samples 15b ( 2 = 3.84) and 30a ( 2 = 4.94) are not within the usual limits. X-ray study shows that the sample 15b is inhomegenous in character. However, the inhomogeneity could be seen in backscattered electron (BSE) images but the fineness of the lamellae evaded analysis. The spectral fitting of these samples show higher chi-square values, but the fitting logic and Mössbauer parameters are in conformity with the earlier reported values (e.g., Schumacher, 1991).

Equilibration temperature of the amphiboles
Mineral assemblage of the studied calcic amphibole bearing samples indicates that the rock underwent amphibolite to lower granulite facies metamorphism. Amphiboles are good indicators of temperature and pressure over wide range of metamorphic conditions (Shido & Miyashiro, 1959;Engel & Engel, 1962;Binns, 1965;Leake, 1965Leake, , 1971Bard, 1970;Ernst, 2002). We have used Holland & Blundy's (1994) amphibole-plagioclase thermometry to calculate the equilibration temperature of the studied Sittampundi calcic amphiboles. In the studied samples, the magmatic Ca-amphibole probably equilibrated with calcic plagioclase but may not have been silica-saturated, and the mineralogical assemblages of the samples do not show any free quartz, so the thermometer (B) was utilized. A pressure of 10 kbar has been estimated using Newton & Perkin's (1982) geobarometry for the basic granulites and this is considered for calculating the equilibrium temperature of the studied samples. The estimated equilibration temperatures of the calcic amphiboles are presented in Table 4. Due to the absence of suitable coexisting mineral pairs, for the anthophyllite associated with the Sittampundi chromitites, we used the Plyusnina's (1982) curve to calculate the equilibration temperature and found to range from 530-560 o C. We have also used Graham & Powell's (1984) hornblende-garnet geothermometry to calculate the equilibration temperature of the basic granulites. The temperatures calculated from the two methods are falling within the same range.