Journal A of thin Mineralogical section scale and original Petrological inhomogeneity Sciences, of bulk Volume rock 103, chemistry page 135 140, inferred 2008-135 LETTER A thin-section scale original inhomogeneity of bulk rock chemistry inferred from compositional zoning of garnet in the Sambagawa metamorphic rocks, central Shikoku, Japan Mutsuko Inui School of Science and Engineering, Kokushikan University, 4-28 - 1, Setagaya, Setagaya - ku, Tokyo 154-8515, Japan The EPMA analysis of a pelitic sample from the Besshi district, central Shikoku, in the Sambagawa metamorphic belt reveals that two types of garnet porphyroblasts with different sizes and chemical compositions are present within a single thin section. One type consisting of medium - sized (d ~ < 1 mm) grains has a high Mn content at the core and shows a pyrope - rich profile. The other type is comparatively larger in size (d ~ 2 mm). The rim compositions of both types are different. The influence of the initial bulk rock chemistry on the garnet is calculated by the forward model. A comparison of the observations with the calculated results indicates that the difference between the two types of garnet grains is attributed to the initial Fe/Mg ratio of the local bulk rocks. Keywords: Garnet, Chemical zoning, Bulk rock chemistry, Sambagawa metamorphic belt INTRODUCTION Pelitic schists found in the Besshi district, central Shikoku, Japan, are well known to contain various types of garnet grains with different sizes, shapes, and chemical profile patterns (e.g., Banno et al., 1986; Takasu, 1986). Garnet grains with normal chemical zoning (i.e., a high - Mn core and bell - shaped zoning profile, e.g., Banno et al., 1986) have been used to construct a P - T trajectory during garnet growth (Inui and Toriumi, 2002). On the other hand, garnet grains with abnormal, complex chemical zoning has been found in several samples and is often used to infer the structural or deformational history of the rocks (e.g., Takasu, 1986; Inui, 2004). In either case, the rim composition of the garnet is consistent for different zoning within a hand specimen. Garnet grains with normal zoning have been reported to show the same compositional profile from the core to the rim within a single hand specimen (Banno et al., 1986). We found two types of garnet with normal zoning. Both occur in a pelitic sample from the Besshi district, central Shikoku. Each represents a different chemical - doi:10.2465/jmps.071022d M. Inui, inui@kokushikan.ac.jp Corresponding author composition trajectory from the core to the rim. Such a discrepancy would generally be ascribed to the variance in bulk rock chemistry without any quantitative consideration. In this study, the compositional trends of the two types of garnet are analyzed and compared with the calculated garnet zoning using the forward model of Inui and Toriumi (2004). The aim of this report is to describe the occurrence of the natural garnet grains and discuss the influence of the compositional variances of bulk rock chemistry in relation to rock textures. SAMPLE DESCRIPTION The Sambagawa metamorphic belt is a high - pressure intermediate - type metamorphic belt in the southwestern part of Japan, comprising mainly Mesozoic trench sediments accreted to the eastern margin of the Eurasian continent. The rocks were metamorphosed around the Late Cretaceous age (116 ± 10 Ma) and exhumed by around 50 Ma (Isozaki and Itaya, 1990). The highest grade rocks of the Sambagawa metamorphic belt are exposed in central Shikoku, where the pelitic schists are accompanied by a small amount of mafic and quartz schists. Four metamorphic zones chlorite, garnet, albite - biotite, and
136 M. Inui Figure 1. Metamorphic zonation map of the Sambagawa belt, Besshi district, central Shikoku, modified from the study by Higashino (1990). The sampling locality is indicated on the map by the number 1. oligoclase - biotite zones are defined on the basis of the mineral assemblages in pelitic rocks in the increasing order of metamorphic grade, as shown in Figure 1 (Higashino, 1990). The oligoclase - biotite zone is located in the central part of structural sequence. The lenticular rock masses of ultramafic and gabbroic origin, some of which contain eclogite, are found in the albite - and oligoclase - biotite zones (Kunugiza et al., 1986). The estimated peak metamorphic conditions are 520 ± 25 C at 0.8-0.95 GPa and 610 ± 25 C at around 0.9-1.1 GPa for the albite - and oligoclase - biotite zones, respectively (Enami, 1983; Enami et al., 1994). The prograde P - T trajectory of the albite - biotite zone rocks has been estimated by Inui and Toriumi (2002), which is a linear heating and compression path starting from approximately 470 C at 0.6 GPa. Sample F2212 (Fig. 2) collected from the albite - biotite zone near the eclogitic rock mass called the Eastern Iratsu metagabbro body (Fig. 1) contains garnet, muscovite, chlorite, albite, quartz, and epidote. The dominant minerals defining schistosity are muscovite and epidote. Muscovite schistosity anastomoses around large garnet porphyroblasts. Most chlorite is either clustered in contact with garnet rim or exists in the cracks of garnet. Garnet in sample F2212 occurs as subhedral porphyroblasts with a grain size of 0.5-2.5 mm containing quartz inclusions. On the basis of occurrence of garnet, two zones are recognized within the thin section: zone A, where small (d ~ < 1 mm) grains are scattered, and zone B, with a few large (d ~ 2 mm) grains (Fig. 2). The assemblage and texture of the matrix minerals are similar for the two zones, and the boundary between them is not clear. The two zones alternate within a single thin section, and they are nearly parallel to the schistosity. Distinctly small euhedral garnet grains with diameters around 100 μm are also observed in zone A. The occurrence of the small grains is similar to that of garnet with intrasectoral chemical zoning occasionally found in the Besshi district Figure 2. Photograph of the thin section of sample F2212. The garnet grains are easily identified as porphyroblasts. Grain numbers are mentioned in Figure 5. As mentioned in the text, zones A and B are defined by the occurrence of garnet porphyroblasts, and the boundary between the two zones is not clear. Grains L1, L6, L8, and L9 belong to type A, whereas grains L3, L5, and L10 belong to type B. Refer to the text for details. (Shirahata and Hirajima, 1995; Inui and Toriumi, 2002). It is known that the chemical composition of such small garnet is not always similar to that of the porphyroblastic garnet, indicating that they did not grow simultaneously in equilibrium. Because of the uncertainty of their origin, this study does not deal with the small garnet grains in the analysis and discussion. METHOD Chemical mapping analysis was performed on 15 garnet porphyroblasts in a thin section of sample F2212. Of these, seven grains with a comparatively high Mn concentration at the center were selected (since a larger Mn value indicated that the analyzed section was closer to the core of the grain), and their chemical composition was determined. Four of these grains were located in zone A, while the other three were in zone B. The chemical analysis was performed using a JEOL JXA - 8200 WD/ED combined microprobe analyzer at the Graduate School of Frontier Sciences, the University of Tokyo. The acceleration voltage was 15 kv and the beam current was 12 na. The beam diameter was set to be 2-3 µm. Correction procedures followed the methods of Bence and Albee (1968). RESULTS The garnet porphyroblasts in sample F2212 exhibit normal zoning being Mn - rich at the center and Mg - and Fe - rich at the rim (Figs. 3 and 4). The grains found in the same zone show roughly the same chemical trend from the core to the rim. On the other hand, the garnet grains in zone A show a slightly different chemical trend from
A thin - section scale original inhomogeneity of bulk rock chemistry inferred 137 Figure 3. X - ray intensity maps of Mg, Ca, Mn, and Fe of a representative garnet grain (L8) of type A. High/low brightness indicates high/ low concentration of the element. those of zone B. Hereafter, the grains from zones A and B are referred to as type A and type B, respectively. Type A grains have a profile that is richer in pyrope than that of type B grains, as shown in the Mn - Fe - Mg triangular diagram (Fig. 5). The Mn content at the center of most of the type A grains is higher than that of type B grains. In the outer part of type A grains, the Mn content increases slightly and decreases again toward the rim. Type B grains exhibit a pyrope - poor chemical profile in the Mn - Fe - Mg triangular diagram (Fig. 5), generally with a smaller Mn content at the center as compared to that of type A grains. The rim compositions of the two types of grains are different. DISCUSSION Comparison with calculated garnet growth Inui and Toriumi (2004) have modeled the formation of Mg - Fe - Mn garnet from Mg - Fe - Mn chlorite in a MnO - FeO - MgO - Al 2 O 3 -SiO 2 -H 2 O system, involving excess quartz and water to balance the stoichiometry. Garnet and chlorite are treated as solid solutions consisting of three (pyrope, almandine, and spessartine) and four (clinochlore, amesite, daphnite, and Mn - chlorite) end - members, respectively. Grossular content is not considered. Giving constraint on the bulk chemical composition, they have calculated the molar amounts of garnet and chlorite and their chemical compositions in equilibrium for a series of pressures and temperatures along the assumed P - T path of the prograde Sambagawa metamorphism. The model is suitable for the reconstruction of garnet growth in the
138 M. Inui Figure 4. X - ray intensity maps of Mg, Ca, Mn, and Fe of a representative garnet grain (L3) of type B. High/low brightness indicates high/ low concentration of the element. Table 1. Two different initial chlorite chemical compositions used in the forward model calculation Figure 5. Mn - Fe - Mg and Ca - Fe - Mg ternary diagrams showing the chemical trends from the central part to the rim of type A and type B garnet grains in sample F2212. The grain numbers are as shown in Figure 2. Mn concentration decreases and the Mg/Fe ratio increases from the central part to the rim in all grains. It should be noted that all type A grains follow the Mg - rich trend in the Mn - Fe - Mg diagram. Sambagawa metamorphic belt since the garnet in this belt is considered to have formed mostly at the expense of chlorite. Assuming that there is no cation diffusion within Cch, clinochlore; Ame, amesite; Dph, daphnite; Mnc, Mn - chlorite. * The Fe - poor composition is the representative chemical composition of chlorite in the chlorite zone rocks compiled from the study by Higashino (1975). the garnet, Inui and Toriumi (2004) have reconstructed the mode and growth zoning pattern of garnet during the prograde Sambagawa metamorphism. Two different calculations were performed by the forward model of Inui and Toriumi (2004) using different initial chlorite chemical compositions in order to evaluate the influence of the bulk rock chemistry on garnet composition. The Fe - poor composition (Table 1) is the repre-
A thin - section scale original inhomogeneity of bulk rock chemistry inferred 139 Figure 7. Mn - Fe - Mg triangular diagram showing the chemical trends of garnet calculated for the Fe - rich and Fe - poor conditions. Garnet produced by the Fe - poor condition has a Mn - rich core and a Fe - poor profile toward the rim. Figure 6. Calculated chemical composition and growth amount of garnet during prograde P - T path. Fe - rich and Fe - poor indicate the results obtained using Fe - richer and Fe - poorer chlorite compositions, respectively. ΔM Grt /ΔT is the amount of growth of garnet per degree of temperature change. It should be noted that the Fe - rich condition produces a larger amount of garnet. sentative chemical composition of chlorite in the chlorite zone rocks compiled from the study by Higashino (1975). The Fe - rich composition comprises increased daphnite content and decreased clinochlore + amesite content with the same cch/ame ratio. A linear P - T path analogous to the P - T trajectory deduced by Inui and Toriumi (2002) was given which starts from approximately 480 C at 0.65 GPa and increases at a rate of 9.5 C/0.06 GPa. As a result, it is predicted that the Fe - poor initial chlorite would produce a Fe - poor garnet with a Mn - rich core, while the volume of the produced garnet would be reduced (Figs. 6 and 7). The predicted zoning profiles of the resulting garnet grains are shown in Figure 8. Type A garnet grains are small, their centers are Mn - rich, and the entire chemical path is Fe - poor; these results are consistent with those expected from a Fe - poor bulk rock chemistry. The amplitude of the difference between the Mg/Fe ratio of type A and type B grains is also within the predicted range. The abovementioned discussion on the grain size is valid only if the nucleation rate remains approximately the same. In this study, the chemical composition data of the actual Figure 8. Predicted zoning profile of the calculated garnet. Standardized radius is the cube root of the calculated amount of growth of garnet. The origin (r = 0) represents the center of the hypothetical grain. cores of garnet are ambiguous because the analysis was performed only in the thin section. Influence of bulk rock chemistry The two different chemical compositional trends of garnet observed in a single thin section are indicative of the pres-
140 M. Inui ence of effective bulk rock inhomogeneities, although the difference is not visible in the matrix mineral assemblage. Zones A and B may represent the relicts of a primary banding structure of the protolith or may be the result of some deformation structure. In either case, it is likely that the bulk rock chemistry had some influence on the formation of the petrographical texture of the rocks. It is known that the lithologic layering in the Sambagawa metamorphic belt is nearly parallel to the metamorphic schistosity, and both have a gentle northward dip. The spatial distribution of type B grains is roughly parallel to the schistosity of the rock defined by muscovite and epidote, which is consistent with the macroscopic parallelism in the Sambagawa metamorphic belt. Previous studies indicate that the schistosity in the Sambagawa metamorphic rocks is mostly formed during the deformation during the uplift of the metamorphic belt (Toriumi and Masui, 1986; Hara et al., 1990). It is conceivable, therefore, that the initial rock layers with different bulk chemistry have considerably stretched and flattened during the ductile deformation after the main phase of garnet growth. ACKNOWLEDGMENTS This research was partly supported by a Grant - in - Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, via grant number 16740295. The author is grateful to M. Obata for the advice and encouragement. The author would like to express her sincere thanks to M. Tagiri and T. Ikeda for their advice and critical reading of the manuscript. The author is also grateful to T. Kuwatani for the chemical analyses. REFERENCES Banno, S., Sakai, C. and Higashino, T. (1986) Pressure - temperature trajectory of the Sanbagawa metamorphism deduced from garnet zoning. Lithos, 19, 51-63. Bence, A.E. and Albee, A.L. (1968) Empirical correction factors for the electron microanalysis of silicates and oxides. Journal of Geology, 76, 382-403. Enami, M. (1983) Petrology of pelitic schists in the oligoclase - biotite zone of the Sanbagawa metamorphic terrain, Japan: phase equilibria in the highest grade zone of a high - pressure intermediate type of metamorphic belt. Journal of Metamorphic Geology, 1, 141-161. Enami, M., Wallis, S. and Banno, Y. (1994) Paragenesis of sodic pyroxene - bearing quartz schists: implications for the P - T history of the Sanbagawa belt. Contributions to Mineralogy and Petrology, 116, 182-198. Hara, I., Shiota, T., Hide, K., Okamoto, K., Takeda, K., Hayasaka, Y. and Sakurai, Y. (1990) Nappe structure of the Sambagawa belt. Journal of Metamorphic Geology, 8, 441-456. Higashino, T. (1975). Biotite zone of Sanbagawa metamorphic terrain in the Siragayama area, central Sikoku, Japan. The Journal of the Geological Society of Japan, 81, 653-670 (in Japanese with English abstract). Higashino, T. (1990) The higher grade metamorphic zonation of the Sanbagawa metamorphic belt in central Shikoku, Japan. Journal of Metamorphic Geology, 8, 413-423. Inui, M. (2004) Pressure - Temperature paths deduced from garnet zoning of the Sambagawa metamorphic rocks, central Shikoku, Japan. Journal of Geography, 113, 571-586 (in Japanese with English abstract). Inui, M. and Toriumi, M. (2002) Prograde pressure - temperature paths in the pelitic schists of the Sambagawa metamorphic belt, SW Japan. Journal of Metamorphic Geology, 20, 563-580. Inui, M. and Toriumi, M. (2004) A theoretical study on the formation of growth zoning in garnet consuming chlorite. Journal of Petrology, 45, 1369-1392. Isozaki, Y. and Itaya, T. (1990) Chronology of Sanbagawa metamorphism. Journal of Metamorphic Geology, 8, 401-411. Kunugiza, K., Takasu, A. and Banno, S. (1986) The origin and metamorphic history of the ultramafic and metagabbro bodies in the Sanbagawa belt. Geological Society of America Memoir, 164, 375-385. Shirahata, K. and Hirajima, T. (1995) Chemically sector - zoned garnet in Sanbagawa schists; its mode of occurence and growth timing. Journal of Mineralogy, Petrology and Economic Geology, 90, 69-79. Takasu, A. (1986): Resorption - overgrowth of garnet from the Sambagawa pelitic schists in the contact aureole of the Sebadani metagabbro mass, Shikoku, Japan. The Journal of the Geological Society of Japan, 92, 781-792. Toriumi, M. and Masui, M. (1986) Strain patterns in the Sanbagawa and Ryoke paired metamorphic belts, Japan. Geological Society of America Memoir, 164, 387-394. Manuscript received October 22, 2007 Manuscript accepted December 22, 2007 Published online March 22, 2008 Manuscript handled by Masaaki Obata