中國四川白馬火成雜岩之實驗岩石學研究與鐵-鈦-釩-氧化物礦床成因隱示

全文

(1)國立臺灣師範大學地球科學研究所碩士論文 Department of Earth Sciences. National Taiwan Normal University Master Thesis. 中國四川白馬火成雜岩之實驗岩石學研究與鐵-鈦-釩-氧化物礦床成因隱示 An experimental petrological investigation of the bi-modal gabbro-syenite Baima igneous complex, SW China Implications for the genesis of Fe-Ti-V-oxide ore deposits. Master student : Hsia, Wen-Yu Advisors : Liu, Teh-Ching and Shellnutt, J. Gregory June, 2015.

(2) Acknowledgement “If master is easy then everyone can be a master.” By Dr. Shellnutt. My advisor said those words to me when I finished a really terrible presentation at the first time. He let me know don't be afraid of making mistakes, but you should get up and move on. I wish to thank my advisors, Professors J.G. Shellnutt and T.C. Liu, for their guidance and patience through the course of this study. Dr. Yoshiyuki. Iizuka and assistants of EPMA lab at the Institute of Earth Sciences, Academia Sinica in Taipei, thank you for helping on my SEM and EPMA work. Dr. S.L. Chung and assistant of ICP-MS lab in National Taiwan University, thank you for helping on my LA-ICPMS work. 感謝天主，給予我這特別的時刻，來學習來成長!讓我在這研究所的時間中，體驗到天主給我滿滿的愛和無數的小奇蹟。感謝蕭炎宏教授和楊懷仁教授在論文資料整理分類及寫作上給予的協助。地質組研究室夥伴--雅芬學姊、伯杰、Thuy、Jen、筱君，感謝有你們與我一起奮鬥，願意與我討論遇到的困難以及給予我需要的幫助。聖宗學長、詠然、沅甫、衍豪、炳權，感謝你們特別來協助我的口試，有你們在真好。謝謝我的家人們，在這三年的研究所生涯中，給予我的支持和鼓勵，當我遇到瓶頸挫折時，有你們在我身後，讓我知道我可以擦乾眼淚再次站起來繼續往前進。教會的夥伴們及我的好朋友、好同學們，一路走來，有你們在我身邊，不斷的幫助我，讓我知道要多依靠主的帶領。育珊你真是我的小天使，總是不厭其煩拯救我的菜英文。最後要謝謝宇恩，你的陪伴是很重要的!謝謝你為我祈禱，也一直鼓勵我要堅持不要放棄。這三年有很豐富的收穫，謝謝愛我的大家，我終於完成了！. i.

(3) Abstract The Baima igneous complex (BIC) consists of a cumulate layered gabbroic unit, a thick Fe-Ti oxide ore zone and an isotropic peralkaline quartz syenite. The BIC contains 1150 Mt of Fe, 44.8 Mt of Ti and 2.85 Mt of V and is one of at least five world class orthomagmatic oxide deposits in the Emeishan large igneous province. The formation of the oxide deposit of the BIC is debated. There are two different models on the formation of Fe-Ti-V oxide ore in the Baima intrusion. One model suggests fractional crystallization of a basaltic parental magma led to the early crystallization of Fe-Ti oxide minerals whereas the second model suggests silicate immiscibility during the early stages of the evolution of a ferro-basaltic magma. The purpose of this study is to determine if a parental magma similar to high-Ti Emeishan basalt can produce all three rock-types observed in the BIC by experimental petrology at atmospheric pressure and mid-crustal pressure (i.e. 1 GPa). The experimental results at atmospheric pressure show that the liquidus temperature and solidus temperature of the basaltic melt are estimated to be 1303 oC and 1120 oC. The crystallization sequence is determined as: titanomagnetite, plagioclase (An65) and pyroxene (Wo43-47En32-45Fs11-23). The residual glass composition, represented by the quenched glass, evolves from lower SiO2 (SiO2 = ~45 wt %) values to higher SiO2 values (SiO2 = ~60 wt %) with corresponding decrease in Ti, Fe, Mg, Ca and. ii.

(4) increase of Na and K. The experimental results at 1 GPa show that the liquidus temperature is ~1220 oC, whereas the solidus temperature is estimated to be 980 oC. The crystallization sequence of the basaltic melt at 1 GPa is determined to be: titanomagnetite, pyroxene (Wo37-46En33-38Fs18-25) and plagioclase (An36). The most evolved glass compositions in the low and high pressure experiments are similar to the enclaves of the Baima syenitic unit but only the low pressure experiments could reproduce the mineral compositions observed in the Baima gabbroic unit. Thus, the liquidus mineral is iron-titanium oxide which is consistent with the observation of basal oxide-ore formation in the gabbroic unit. The low pressure results of this study indicate that early crystallization of Fe-Ti oxides will occur assuming a geologically reasonable starting material and that the residual liquid is silicic. The direct implication is that the oxide deposits and spatially associated granitic rocks formed together by crystallization from a basaltic parental magma. Furthermore, it is suggested that crystallization of a typical high-Ti basaltic parental magma can produce world-class giant magmatic oxide ore deposits and that the occurrence of some alkaline granites may be an indicator of their presence. Key words: Baima igneous complex, Fe-Ti-V oxide.. iii.

(5) Table of Contents Acknowledgement ..............................................................................................................i Abstract ............................................................................................................................. ii Table of content ..............................................................................................................iv List of tables .....................................................................................................................vi List of figures ................................................................................................................. vii 1. Introduction ................................................................................................................ 1 1-1 1-1-1. Geology of SW China (South China Block and Songpan Ganze terrane) ........... 4. 1-1-2. Layered gabbroic intrusions of the Panzhihua-Xichang region ........................... 9. 1-1-3. Baima igneous complex ..................................................................................... 13. 1-1-4. Mafic dykes ........................................................................................................ 14. 1-1-5. Origin of the Baima Fe-Ti-V oxide models ....................................................... 18. 1-2. 2. 3. Geological background......................................................................................... 4. Purpose ............................................................................................................... 19. Experimental methods of petrological investigation ................................................ 20 2-1. Anhydrous high-temperature experiment at 0.001 GPa ..................................... 22. 2-2. Anhydrous high pressure high temperature experiment at 1 GPa ...................... 30. 2-3. Scanning electron microscope (SEM) ................................................................ 34. 2-4. Electron probe micro analyzer (EPMA) ............................................................. 35. 2-5. Laser ablation- inductively coupled plasma-mass spectrometry (LA-ICPMS) .. 36. Results 3-1. .................................................................................................................. 38 Anhydrous high-temperature experiment at atmospheric pressure .................... 38. 3-1-1. SEM analysis ...................................................................................................... 41. 3-1-2. Mineral composition of the synthetic phases ..................................................... 46. 3-1-2-1 Major elements ....................................................................... 46 3-1-2-1-1 Fe-Ti oxide mineral ..................................................... 46 3-1-2-1-2 Pyroxene ...................................................................... 48 3-1-2-1-3 Plagioclase ................................................................... 48 3-1-2-1-4 Glass ............................................................................ 53 3-1-2-2 Trace elements ........................................................................ 57 iv.

(6) 3-2. Anhydrous high pressure high temperature experiment at 1 GPa ...................... 63. 3-2-1. SEM analysis ...................................................................................................... 63. 3-2-2. Mineral composition of the synthetic phases ..................................................... 68. 3-2-2-1 3-2-2-2 3-2-2-3 3-2-2-4 4. Fe-Ti oxide.............................................................................. 68 Pyroxene ................................................................................. 70 Plagioclase .............................................................................. 70 Glass ....................................................................................... 76. Discussion ................................................................................................................ 79 4-1. Texture of minerals ............................................................................................ 79. 4-2. Mineral and residual melt composition .............................................................. 81. 4-3. Implication of oxide ore deposit genesis ............................................................ 89. 4-4. Origin of the Baima Fe-Ti-V oxide deposit: silica immiscibility or fractional. crystallization? ........................................................................................................................ 93. 5. Conclusions .............................................................................................................. 95. 6. References ................................................................................................................ 97. v.

(7) List of tables Table 1-1. The main reserve of iron-titanium-vanadium deposits in the Panxi region. .12 Table 2-1. The major and trace element of the starting material in this study. .............. 21 Table 3-1. Run products at atmospheric pressure. ......................................................... 39 Table 3-2. The average compositions of the synthesized Fe-Ti oxide at atmospheric pressure of this study. ............................................................................................. 47 Table 3-3. The average compositions of the synthesized pyroxene at atmospheric pressure of this study. ............................................................................................. 49 Table 3-4. The average compositions of the synthesized plagioclase at atmospheric pressure of this study. ............................................................................................. 51 Table 3-5. The average compositions of the glass at atmospheric pressure of this study. ............................................................................................................................. 54 Table 3-6. The average compositions of trace element of the synthesized pyroxene at atmospheric pressure of this study. ........................................................................58 Table 3-7. The trace element compositions of the. synthesized plagioclase at. atmospheric pressure of this study. ........................................................................59 Table 3-8. The average compositions of trace element of the glass at atmospheric pressure of this study. ............................................................................................. 60 Table 3-9. Run products at 1 GPa. ................................................................................. 64 Table 3-10. The average compositions of the synthesized Fe-Ti oxide at 1 GPa of this study. ...................................................................................................................... 69 Table 3-11. The average compositions of the synthesized pyroxene at 1 GPa of this study. ................................................................................................................................ 71 Table 3-12. The average compositions of the synthesized plagioclase at 1 GPa of this study. ...................................................................................................................... 74 Table 3-13. The average compositions of the glass at 1GPa of this study. .................... 77 Table 4-1. Percentages of the phases in the run products at 1 atm of this study. ...........90 Table 4-2. Percentages of the phases in the run products at 1 GPa of this study. ..........91. vi.

(8) List of figures Fig. 1-1. The Emeishan large igneous province ............................................................... 6 Fig. 1-2. The distribution of the Emeishan large igneous province, the Ti-V oxide and Ni-Cu-PGE sulphide deposits .................................................................................. 7 Fig. 1-3. Geological map of the Panzhihua – Xi Chang region ..................................... 11 Fig. 1-4. The geological map of Baima igneous complex ............................................. 16 Fig. 1-5. Stratigraphy of Baima layered intrusion.......................................................... 17 Fig. 2-1. R-type thermocouple. ...................................................................................... 24 Fig. 2-2. Procedure for making a platinum envelope. .................................................... 25 Fig. 2-3. High temperature experimental setup at atmospheric pressure. ...................... 26 Fig. 2-4. The high temperature tube furnace. .................................................................27 Fig. 2-5. Diagram of the high temperature tube furnace setup with sample. ................. 28 Fig. 2-6. Illustrated procedure of sample mounted in epoxy and section polishing. .....29 Fig. 2-7. Schematic diagram of the main setting of the Quick press 3.0. ...................... 31 Fig. 2-8. Furnace assembly for the high pressure experiment. ....................................32 Fig. 3-1. The crystallization sequence of the BIC melt at atmospheric pressure. ..........40 Fig. 3-2. The BEI of run BM-023 (1310 oC).................................................................. 42 Fig. 3-3. The BEI of run BM-005 (1300 oC)..................................................................42 Fig. 3-4. The SEI of run BM-026 (1180 oC). .................................................................43 Fig. 3-5. The SEI of run BM-027 (1159 oC). .................................................................44 Fig. 3-6. The SEI of run BM-007 (1125 oC). .................................................................45 Fig. 3-7. Variations of synthesized pyroxene composition compared with those of natural pyroxene in BIC. ........................................................................................ 50 Fig. 3-8. Variations of synthesized plagioclase composition compared with those of natural plagioclases in BIC.. .................................................................................. 52 Fig. 3-9. Variation of glass composition versus temperature at atmospheric pressure. .55 Fig. 3-10. The residual glass composition at 1 atm compared with natural rock in BIC56 Fig. 3-11. The residual glass trace elements composition at atmospheric pressure of this study compared with natural rock in BIC. ............................................................. 61 Fig. 3-12. Variation of trace elements of glass compositions versus temperatures at atmospheric pressure. ............................................................................................. 62 Fig. 3-13. The crystallization sequence of the BIC melt at 1 GPa. ................................ 65 Fig. 3-14. Phase relations of a BIC gabbro at 0 to 1.0 GPa. ..........................................65 vii.

(9) Fig. 3-15. The BEI of run BM-H20 (1240 oC).. ............................................................. 66 Fig. 3-16. The BEI of run BM-H16 (1140 oC).. ............................................................. 66 Fig. 3-17. The BEI of run BM-H09 (1120 oC).. ............................................................. 67 Fig. 3-18. The BEI of run BM-H05 (1000 oC). .............................................................. 67 Fig. 3-19. Variations of synthesized pyroxene composition at 1 GPa and compared with those of natural pyroxene in BIC ...........................................................................73 Fig. 3-20. Variations of synthesized plagioclase composition at 1 GPa are compared with natural plagioclases in BIC. ...........................................................................75 Fig. 3-21. Variation of glass compositions versus temperatures at 1 GPa. .................... 78 Fig. 4-1. The BEI of run BM-027, BM-007 and thin section ........................................80 Fig. 4-2. The barium versus strontium contents of the residual melts at atmospheric pressure compared with natural rock data. ............................................................. 84 Fig. 4-3. The residual glass composition of this study compared with natural rock in BIC and Baima layered series selected chemical compositions. ........................... 85 Fig. 4-4. Estimated oxygen fugacity and temperature of the Baima .............................. 88 Fig. 4-5. Generalized stratigraphic section, showing the composition in the Panzhihua, Baima, and Taihe intrusions ................................................................................... 92. viii.

(10) 1. Introduction. Many mafic-ultramafic layered intrusions are considered to be the frozen remnants of dynamic magma chambers that record the processes of in situ crystallization of silicate, oxide and other minerals (e.g. sulphide and phosphate minerals). The concept of crystal settling and sorting within magma bodies was given by the seminal study by Wager and Deer's classic 1939 memoir on the Skaergaard Intrusion in East Greenland. From that point on their study is recognized as a major influence on the differentiation of basaltic magmas. Layered intrusions are very important hosts of base and precious metals as many contain significant reserves of platinum-group elements (PGEs), base metal sulphides, chromite, magnetite, and ilmenite. For example, meter-thick layers of pegmatitic pyroxenite in the Bushveld and Stillwater are major world repositories of platinum-group elements, i.e. Pt, Pd, Rh, Ir, Ru, and Os (c.f. Eales and Cawthorn, 1996). Layered intrusions are considered to represent the simple concept of a magma chamber undergoing differentiation as a result of early formed crystals settling out of a magma and accumulating in layers on the floor of the chamber. This classical view has since been discarded by most petrologists in favor of models involving in situ crystallization, in which magma chambers are thought to have the general form of a central mass of nearly crystal-free magma that gradually loses heat and crystallizes inwards from its 1.

(11) margins. The Emeishan large igneous province (ELIP) located in southwest China consists of voluminous mafic, ultramafic and silicic volcanic rocks, and layered mafic-ultramafic intrusions some of which are spatially associated with felsic plutonic rocks. A number of the mafic-ultramafic intrusions host economic to sub-economic Ni-Cu-(PGE) sulphide deposits whereas others host world class Fe-Ti-V oxide deposits. The layered intrusions can be divided into two groups based on their compositions and stratigraphy. One group of deposits at Hongge and Xinjie may originate from a more primitive parental magma (i.e. ultramafic) but subjected to open system magmatic processes (i.e. loss of residual liquid, significant crustal contamination). The other group consisting of the Baima, Panzhihua and Taihe deposits are likely derived from more evolved parental magmas (i.e. basaltic) and were comparatively less influenced by the open system magmatic processes (Shellnutt, 2014; Zhong et al., 2005). The layered gabbroic intrusions are all spatially and temporally associated with peralkaline granitic plutons suggesting there may be a petrogenetic link between the two rock types. The mechanism for the formation of the Panzhihua, Baima and Taihe gabbro-granitoid-ore complexes, like many aspects of the ELIP, is a highly debated issue. For example, some consider the gabbro-ore to be exclusive from the formation of the neighboring granitic rocks; whereas the others suggest that they are part of the same 2.

(12) intrusion or they are a rare example of large scale silicate liquid immiscibility (Zhong et al., 2011; Shellnutt et al., 2011, Liu et al., 2014). The major features (e.g. composition, stratigraphy and age) of these three gabbroic intrusions are similar enough to suggest that they likely underwent similar processes of formation (Zhong et al., 2005). Undoubtedly each intrusion had some unique developments that are more of a function of initial parental magma composition and location of emplacement. There are two different models on the formation of Fe-Ti-V oxide ore in the Baima intrusion. One model suggests fractional crystallization of a basaltic parental magma led to the early crystallization of Fe-Ti oxide minerals (Shellnutt et al., 2009) whereas the second model suggests silicate immiscibility at the early stage of the evolution of a ferro-basaltic magma (Zhou et al., 2013). In this study, origin of the Baima gabbro-granite-ore complex was investigated by experimental petrological methods. A starting composition similar to Emeishan basalt (i.e. high-Ti basalt) with a similar composition as the theorized parental magma of the Baima igneous complex (Xu et al., 2001) was melted at atmospheric pressure and at high pressure conditions in order to determine the crystallization order, chemical evolution of the residual liquid composition and if such a single parental magma can produce a liquid composition similar to the spatially associated silicic pluton.. 3.

(13) 1-1 Geological background 1-1-1 Geology of SW China (South China Block and Songpan Ganze terrane) The Yangtze Block consists of Mesoproterozoic granitic gneisses and metasedimentry rocks which have been intruded by Neoproterozoic (~800 Ma) granites (Li, 1999; Zhou et al., 2002b). The granites are overlain by a series of marine and terrestrial strata from the Late Neoproterozoic (~600 Ma) to the Early Permian. The eastern part of the Tibetan Plateau, known as the Songpan–Ganze terrane, is composed primarily of late Triassic–early Jurassic marine sediments (Bruguier et al., 1997). The Emeishan large igneous province (ELIP) is located in southwestern China, along the western margin of the Yangtze Block to the east and the eastern most part of the Tibetan Plateau to the west (Fig. 1-1). It covers an area of ~0.3x106 km2 of SW China and northern Vietnam (Ali et al., 2005; Shellnutt, 2014). The ELIP is not particularly large LIPs (c.f. Siberian Traps and Ontong-Java Plateau) but it contains a number. of. economic. magmatic. Cu-Ni-(PGE)-bearing. sulfide. deposits. and. mafic-ultramafic layered intrusions that host giant Fe-Ti-V deposits (Fig. 1-2) (Zhong et al., 2002; Zhou et al., 2005; Shellnutt, 2014). The volcanic sequences range from 1.0 to 5.0 km in thickness in the western half of ELIP and 0.2 to 2.6 km in the eastern half and consist mostly of flood basalts with picrites in the lower half and basaltic-andesites and silicic volcanic rocks in the upper 4.

(14) half. The volcanic rocks erupted at equatorial latitude of eastern Pangaea within one normal-polarity cycle suggesting rapid emplacement (Shellnutt, 2014). The age of the ELIP is geologically constrained as the Late Permian having an emplacement age of ~260 Ma as determined by high precision zircon U–Pb geochronological analysis of the layered, mafic–ultramafic intrusions and felsic plutons (Zhou et al., 2002a, 2005, 2006; Shellnutt and Zhou, 2007; Shellnutt et al., 2012). .. 5.

(15) Fig. 1-1. The Emeishan large igneous province is located in southwest China (modified from Shellnutt and Pang (2012)) (BIC: Baima igneous complex).. 6.

(16) Fig. 1-2. Fig. 1-2. The distribution of the volcanic rocks of the Emeishan large igneous province (green area) and the Ti-V oxide and Ni-Cu-PGE sulphide deposits (from Shellnutt (2014)).. 7.

(17) The Emeishan basalts can be classified into two major magma types (Xu et al., 2001; Zhou et al., 2008). They are: 1) a low-Ti (LT) type that exhibits low Ti/Y (<500), εNd(t) (−4.8 to +1.4), 87Sr/86Sr (0.7046 to 0.7064), Mg# (0.52 to 0.64) and 2) a high-Ti (HT) type that has high Ti/Y (>500), εNd(t) (−3.6 to +4.8), 87Sr/86Sr (0.7039 to 0.7059), Mg# (0.32 to 0.61) (Xu et al., 2001). The εNd values were calculated relative to present-day chondrite values of. 143. Nd/144Nd = 0.512638, Nd isotopic ratios could. certainly help to constrain the source(s) of granites and Mg# = Mg/(Mg+Fe2+). (Nb/Th)PM is best used to indicate the extent of Nb anomaly whereas (Th/Yb)PM is a sensitive indicator of crustal contamination. The oxide-bearing intrusions have very low (Th/Yb)PM ratios and high (Nb/Th)PM ratios, high εNd values (−0.1 to +4.6), low 87. Sr/86Sr ratios (0.7039 to 0.7054) and were probably derived from a parental magma. more similar to the high-Ti basalt (Zhou et al., 2008). The high-Ti basaltic rocks were considered as the parental magmas for the Fe–Ti oxide bearing layered intrusions of the Panxi region, whereas the Ni–Cu sulphide bearing intrusions belong to the low-Ti series (Zhou et al., 2008). The trace element characteristics suggest that high-Ti basalts experienced extensive crystal fractionation from parental magmas either in magma chambers or en route to the surface (Xu et al., 2001).. 8.

(18) 1-1-2 Layered gabbroic intrusions of the Panzhihua-Xichang region. The ELIP is an important host of sulphide and oxide deposits many of which are located between the cities of Panzhihua and Xichang (Panxi) in southern Sichuan. In addition to numerous orthomagmatic deposits the ELIP contains a diverse array of within-plate A-type granitic rocks (Fig. 1-3). A-type granites are considered to silicic rocks from an anorogenic tectonic setting that is not associated with regional metamorphism or convergent plate tectonics. They are typically of peralkaline and contain sodic amphibole or pyroxene (Blatt et al., 2006). The deposits can be divided into two groups based on their compositions and stratigraphy. One group of the deposits at Hongge and Xinjie may have originated from more primitive parental magmas but were certainly subject to open system magmatic processes (i.e. recharge, assimilation). Another group is the Panzhihua, Baima and Taihe deposits likely originated from more evolved parental magmas (i.e. basaltic) and were comparatively less influenced by open system magmatic process. Furthermore, the Panzhihua, Baima and Taihe gabbros are petrogenetically related to isotropic peralkaline quartz-bearing granitoids whereas the others do not appear to be so. The Fe-Ti-V oxide deposits of the ELIP are a substantial economic resource. Collectively the Panzhihua, Baima and Taihe deposits contain a total known metal. 9.

(19) reserve of ~4980 Mt of Fe, ~480 Mt of Ti, and ~14 Mt of V (Zhong et al., 2005) (Table 1-1).. 10.

(20) Fig. 1-3. Geological map of the Panzhihua – Xi Chang region (Panxi) showing the distribution of Late Permian granitic rocks (modified from Shellnutt et al., 2012).. 11.

(21) Table 1-1. The main reserve of iron-titanium-vanadium deposits in the Panxi region (Zhong et al., 2005). Intrusion. Fe (Mt)*. Ti (Mt). V(Mt). Panzhihua. 2050. 237. 6.01. Taihe. 1780. 200. 5.18. Baima. 1150. 44.8. 2.85. Total. 4980. 481.8. 14.04. *Mt: million tons.. 12.

(22) 1-1-3 Baima igneous complex. This research is focused on the Baima igneous complex (BIC), which is located in the central part of the Panxi region, ~100 km northeast of Panzhihua, and consists of a gabbroic unit, a syenitic unit and oxide ore deposit (Fig. 1-4) (Chen, 1990; Yang et al., 1997). The layered gabbroic unit lies to the east of the syenitic unit, dipping ~25° to the west. The two units cover an area of similar size, but their respective volumes are unknown. Yang et al. (1997) suggested that the REE and Sr isotope of BIC data show that the parental magma from which the Baima igneous complex crystallized is a basaltic magma. The U–Pb SHRIMP zircon ages of 258 ± 4 Ma and 259 ± 5 Ma of the syenitic unit are within error of the 261 ± 2 Ma of the gabbroic unit (Shellnutt et al., 2009). It is thought that the layered intrusive rocks and the syenites were comagmatic (Yang et al., 1997). The gabbroic unit can be divided into four lithologic zones from the bottom to the top: 1. a lower cumulate zone; 2. an oxide ore zone; 3. an olivine gabbro zone; and 4. an upper gabbro zone (Chen, 1990) (Fig. 1-5). The characteristics of mineralogy, petrography, petrochemistry and trace element geochemistry indicate that fractional crystallization of the intrusion began at the bottom and finished at the top (Chen, 1990). The rocks consists of coarse grained cumulate olivine, plagioclase, clinopyroxene and. 13.

(23) interstitial Fe–Ti oxide minerals with minor amounts of sulphide minerals, spinel, and apatite (Shellnutt and Pang, 2012). The ore deposits are hosted within the lower portions of layered cumulate gabbroic intrusions (Shellnutt et al., 2009). The syenitic unit is structurally above the gabbroic unit and contains abundant ellipsoidal mafic enclaves varying in size from a few centimeters to 10s of centimeters in length (Shellnutt et al., 2010). The syenitic unit is granular except for a few small bands of finer grained textures which appear to be more siliceous in composition. The syenites are coarse grained and consist of perthitic alkali feldspar, amphibole, quartz, aegirine and accessory (≤ 5 vol %) apatite, titanite, zircon, plagioclase, fluorite, biotite, ilmenite, magnetite and pyrite (Shellnutt and Iizuka, 2011). The BIC was subsequently intruded by two metaluminous syenitic plutons and numerous alkaline mafic dykes (Shellnutt and Zhou, 2007; Shellnutt and Iizuka, 2011; Shellnutt et al., 2012). 1-1-4 Mafic dykes There are abundant felsic and mafic plutonic rocks in Panxi region. Intruding these plutons are narrow (generally ≤ 5 m wide) mafic dykes. The dykes differ considerably from other mafic intrusions of the Panxi region because they have diabasic textures, while the other mafic intrusions are granular and have cumulate textures. The dykes consist primarily of plagioclase (50%), clinopyroxene (35%), biotite (5–10%), and Fe–Ti oxides (5–10%), with minor olivine (< 5%), apatite (< 5%) and sulphides 14.

(24) (Shellnutt et al., 2008).. 15.

(25) Fig. 1-4. The geological map of Baima igneous complex, including syenite (blue), gabbro (red) and oxide ore deposit (black) (modified from Shellnutt and Pang, 2012).. 16.

(26) Fig. 1-5. Stratigraphy of Baima layered intrusion showing the subdivisions of the Layered Series (modified from Chen, 1990; Shellnutt and Pang, 2012).. 17.

(27) 1-1-5 Origin of the Baima Fe-Ti-V oxide models The formation of Fe-Ti-V oxide is considered to form by either silicate liquid immiscibility or fractional crystallization (Liu, 2014; Shellnutt et al., 2009; Zhou et al., 2008). Under suitable conditions (e.g. lower pressure, alkalic composition) two-silicate liquid immiscibility will produce two end-members, Fe-rich and Si-rich members, of approximately equal proportions (Roedder, 1979). The composition of the Fe-rich end-member may resemble the bulk composition of a mafic cumulate rock whereas the Si-rich end-member will be similar to some granitic rocks. Liu et al. (2014) suggested that plagioclase and olivine began to crystallize during cooling of the Baima parental magma. Olivine sinks to the floor of the magma chamber whereas the plagioclase floats upward. The residual magma becomes denser and more enriched in Fe and Ti. The dense Fe-Ti-rich melt droplets with low viscosity tended to sink to the bottom of the chamber where they infiltrated downwards through the unconsolidated crystal pile and filled in the interstitial spaces between the silicate minerals, forming the net-textured and disseminated ores in the bottom side. The immiscible Fe-Ti-rich melts were trapped in the growing silicate crystals and crystallized titanomagnetite, ilmenite, apatite, phlogopite, hornblende and pyrrhotite. On the other hand, Shellnutt et al. (2009, 2011) suggested that a mafic magma was injected into the crust and crystallized olivine, plagioclase and clinopyroxene. During fractionation, the relative oxidation state (i.e. 18.

(28) O2) increased and led to en masse crystallization of oxide minerals. The removal of plagioclase from the parental magma likely caused a „plagioclase effect‟ which resulted in the residual magma to become peralkaline. Continued fractionation and decreasing O2 produced an ilmenite dominated upper gabbro near the boundary with the syenitic unit.. 1-2 Purpose The formation of the oxide ore deposits is debated, the three main problems are as follows: (1) formation of the oxide ore deposits by silicate-liquid immiscibility or fractional crystallization, (2) composition of parental magma, and (3) relationship between A-type granitoids and ore-bearing gabbroic intrusions. Therefore, the purpose of this study is to determine if a parental magma similar to high-Ti Emeishan basalt can produce all three rock-types observed in the BIC by experimental petrology. If so, the results will have significant implications for the genesis of orthomagmatic oxide ore deposits and the formation of A-type granites.. 19.

(29) 2. Experimental methods of petrological investigation. Previous studies have suggested that the parental magma composition of the Baima igneous complex (BIC) was similar to the high-Ti Emeishan basalt, specifically to the sample EM-81 reported by Xu et al. (2001) (Table 2-1). In this study, we selected a known composition of Emeishan high-Ti basalt from the Baima area to represent the parental magma of the BIC. Sample GS04-026 was collected from a mafic dyke that intruded the syenitic portion of the BIC and was reported by Shellnutt et al. (2008). The bulk composition of GS04-026 is similar to EM-81 but there are minor differences with respect to the Al2O3, Fe2O3 and Na2O contents (Table 2-1). The rock sample (GS04-026) was crushed and ground into powder using a steel jaw crusher and an agate mill which eventually produced a powder size less than 200 mesh. The powder was stored in a desiccator before use in the experiments.. 20.

(30) Table 2-1. The major and trace element of the starting material in this study (High-Ti basalt average data reported by Xu et al., 2001). Element SiO2 (wt. %) TiO2 Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5 Total In ppm Sc V Cr Co Ni Cu Zn Ga Rb Sr Y Zr Nb Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Th U. High-Ti avg. EM-81. GS04-026. 47.52 3.44 13.33 14.20 0.20 5.61 8.96 2.67 1.17 0.43 97.06. 44.51 3.47 13.75 16.37 0.22 6.93 10.15 2.40 1.08 0.35 99.23. 45.01 3.97 12.42 16.48 0.24 5.89 10.16 3.23 1.34 0.30 99.04. -. 438 98 52 86 209 121 26 35 780 35 227 31 590 29.6 63.8 35.9 8.5 2.8 1.1 2.3 0.3 5.8 1.7 0.7 0.3. 33 893 26 57 61 333 158 23 62 401 26 165 21 0.7 310 24.5 54.2 7.2 30.5 6.8 2.2 6.0 1.0 5.5 1.0 2.8 0.4 2.3 0.3 4.6 1.5 3.2 0.8. 21.

(31) 2-1 Anhydrous high-temperature experiment at 0.001 GPa All temperatures were measured using a R-type thermocouple (Pt-Pt87Rh13 thermocouple) (Fig. 2-1). The precision of the temperature measurement is ＋/－ 1 ºC. All temperatures were corrected to the International Practical Temperature Scale of 1968 (Biggar, 1972) based on the calibration with the melting point of synthetic diopside (CaMgSi2O6). The temperatures read from the thermocouple were incremented to 10 ºC to obtain the experimental temperatures. The details of the melting experimental techniques at atmospheric pressure were the same as Liu et al. (1997). The experiment procedures are listed as follows: 1.. A platinum sheet (10 mm × 18 mm) was folded to be an envelope. An approximately 0.06-0.08 grams of the sample powder was loaded into the envelope (Fig. 2-2).. 2.. A small hole was added to the top of the platinum envelope through which a thin platinum wire was threaded in order for the envelope to be hung from the alumina tube (Fig. 2-3).. 3.. A desired temperature was raised and fixed for a specific run. The temperatures were measured using a R-type thermocouple.. 4.. Place the sample into a high temperature vertical-quenched furnace (Fig. 2-4).. 22.

(32) 5.. The range of the experiment duration time in this study is between 6 hours and 99 hours depends on the temperature.. 6.. At the end of the run, the charges were quenched in water (Fig. 2-5).. 7.. The temperatures of the experiments were measured again. The experimental charges were mounted in epoxy and polished to make a section. (Fig. 2-6). The mineral phases in the run products were first identified microscopically in reflected light according to characteristic relief, reflectivity, and crystal habit.. 23.

(33) Fig. 2-1. A R-type thermocouple was used to measure to the temperature.. 24.

(34) A.. Cut the platinum foil (10 mm  18 mm) and fold into two equal portions.. B.. Fold opposite sides of the platinum foil from the outside to inside two times.. C.. Encased the powder in platinum envelope.. D.. Fold the top side and sealed the envelope.. E.. Puncture the top side of the envelope.. F.. Thread thin platinum wire through the envelope hole (~1.5 cm long) and hang on alumina tube.. Fig. 2-2. Procedure for making a platinum envelope. 25.

(35) Thick platinum wire (ϕ = 0.5 mm). Ceramics insulation. Alumina tube. 35 cm. (ϕ = 4 mm). Thin platinum wire (ϕ = 0.1 mm). Alumina insulation with two holes Platinum envelopes (8  8 mm). Fig. 2-3. High temperature experimental setup at atmospheric pressure. 26.

(36) Fig. 2-4. The high temperature tube furnace.. 27.

(37) Ceramics insulation. Heating rods. Bottom stopper. Water. Release current device. Fig. 2-5. Diagram of the high temperature tube furnace setup with sample. 28.

(38) Vaseline A.. Hardener. B.. Smear vaseline around mold. Mix the resin and hardener at a proportion of 25:3.. C.. Epoxy is added to the mold.. D.. Solidification of epoxy within 24 hours.. E.. The epoxy was polished using progressively finer (i.e. 320, 600, 1000 grit) abrasive paper and 6 μm to 1 μm diamond compound paste.. Fig. 2-6. Illustrated procedure of sample mounted in epoxy and section polishing. 29.

(39) 2-2 Anhydrous high pressure high temperature experiment at 1 GPa For the high pressure experiments, the runs were carried out using a Quickpress 3.0 piston cylinder apparatus, Depths of the Earth Company (Fig. 2-7). Pressure is generated by pumping a hydraulic ram to force a piston into a pressure plate that consists of concentric layers of hardened-steel around a tungsten carbide core. The area ratio between the ram and the piston is 100:1. The operating machine generates a pressure range between 0.5 and 2.5 GPa (~75 km depth) and a maximum temperature of 2200 oC. The furnace apparatus is shown in Fig. 2-8. Au75-Pd25 capsules were used as sample containers. The pressure-cell apparatus was set to a pressure ten percent higher than the expected pressure. Then, the experiment temperature was raised by the temperature controller at a rate of 60 oC per minute. After reaching the experiment temperature, the pressure was released to the exact experimental range. The temperature for the high pressure experiments were measured with a C-type thermocouple (W5Re95-W26Re74). The cooling water was circulated through a recirculating chiller (CFT-75, Neslab Company, USA) and the temperature of the water was kept at 25 oC.. 30.

(40) Fig. 2-7. Schematic diagram of the main setting of the Quick press 3.0 (not to scale) (modified from Liu et al., 1997).. 31.

(41) 1. 2. 3. 4. 5. 6. 7. 8.. Graphite cap Glass rod Graphite furnace Pyrex sleeve Salt sleeve Au75Pd25 capsule Magnesia powder Crushed alumina tube. 9. 10. 11. 12. 13. 14. 15. 16.. Wafer-thin alumina disc 4-hole crushable alumina tube Magnesia rod Stainless steel base plug Pyrex collar Insulator Base plug support block Thermocouple (C-type). Fig. 2-8. Furnace assembly for the high pressure experiment. 32.

(42) The high pressure experiment procedure was: 1.. Cut Au75Pd25 tube to make a capsule (3 mm × 8 mm), used hammer and metal rod to close one side of tube, encased the powder in capsule, then used hammer and metal rod again to seal the capsule.. 2.. Used W5Re95 & W26Re74 alloy wire through 4-hole crushable alumina tube and have a cross on the top to make the C-type thermocouple.. 3.. Combine every parts of furnace assembly for the high pressure high temperature experiment.. 4.. Put furnace assembly into pressure plate and push pressure plate into piston cylinder apparatus.. 5.. Raised the pressure-cell assembly to 1.1 GPa, 10% higher than the expected experiment pressure.. 6.. Turn on cooling water machine and set experiment temperature.. 7.. Adjust the experiment pressure to be 1.0 GPa.. 8.. Start the experimental run. Secured the temperature and pressure of the machine.. 9. At the end of the run the machine was turned off and the capsule was chilled to room temperature. The experimental temperature was decreased at a rate of 100 oC per minute.. 33.

(43) 2-3 Scanning electron microscope (SEM) A scanning electron microscope (SEM: JEOL JSM-6360LV) equipped with an energy dispersive X-ray spectrometer (EDS: Oxford Instruments, INCA-300) was used, with 15 kV acceleration voltage and 0.2-0.3 nA in the primary electron beam current under low-vacuum conditions (25 Pascal) in the EPMA lab of the Institute of Earth Sciences, Academia Sinica in Taipei. The points to be analyzed were selected under the back-scattered electron image. Grains of minerals in the quenched products chosen for analysis were usually larger than 10 μm in diameter and the analyzed glass pools were usually larger than 30 µm in diameter. Minerals were identified based on comparisons of the x-ray spectra with those of chemically-known minerals (Iizuka et al., 2005). Chemical compositions of elements present in amounts less than 1 weight percent in relative standard deviations (RSD) were analyzed with a 2 µm electron beam for 30 seconds. All iron content is measured as ferrous iron (Fe2+). The quantitative data were corrected to oxide compositions and were normalized to 100 percent. The Igpet2011 computer program was also used to calculate the CIPW norm of the glasses. The experiment procedures are listed as follows: 1.. After polishing and surface cleaning, a petrographic microscope was used to identify specific mineral targets.. 2.. The section was loaded on the stage of the specimen chamber. 34.

(44) 3.. The load current of the filament was adjusted in two steps to make sure the current is stable.. 4.. The working distance was set to 10 mm and the focus, brightness and contrast were adjusted accordingly.. 5.. Images were captured by scanning electron microscope and the target points were selected under back-scattered electron image.. 6.. Energy dispersive X-ray spectrometer was used to analyze the mineral phases.. 7.. Calculate and interpret the mineral phase of every experimented temperature by the spectra.. 2-4 Electron probe micro analyzer (EPMA) Quantitative and elemental distribution (mapping) analyses were made by a JEOL electron probe micro analyzer (EPMA) JXA-8900R which is equipped with four wave-length dispersive spectrometers (WDS) at the Institute of Earth Sciences, Academia Sinica in Taipei. Secondary and back scattered electron images were used to guide the analysis on target positions of minerals. A 2 µm defocused beam was operated for quantitative analysis at an acceleration voltage of 15 kV with a beam current of 12 nA. The measured X-ray intensities were corrected by ZAF method using the standard calibration of synthetic chemical-known standard minerals with various diffracting crystals, listed as follows; diopside for Si with TAP crystal, wollastonite for Ca (PET), 35.

(45) rutile for Ti (PET), corundum for Al (TAP), Chromium oxide for Cr (PET), hematite for Fe (LiF), manganese oxide for Mn (PET), periclase for Mg (TAP), nickel oxide for Ni (LIF), albite for Na (TAP), orthoclase for K (PET). Peak counting for each element and both upper and lower baselines are counted for 10 sec and 5 sec, respectively. Relative standard deviations (RSD) for Si, Na and K were less than 1%, and others were less than 0.5%. Detection limit which based on 3σ of standard calibration were less than 500 ppm for all elements. Mapping analysis was performed by focused beam with 12 kV and 20 nA of the acceleration voltage, and probe current, respectively. X-ray intensities were counted in each 1 µm spot for 0.020 sec.. 2-5 Laser ablation- inductively coupled plasma-mass spectrometry (LA-ICPMS) The trace elements were measured by inductively coupled plasma-mass spectrometry (ICP-MS) using an Agilent 7500s quadrupole spectrometer, which has a stability range within ∼10% variation in the ICP-MS lab of National Taiwan University. Before and during analysis of the samples, the standards NIST610 and BCR-2G were analyzed. The analytical precision was generally better than 5% (2σ) for most trace elements. The laser ablation was performed with a helium carrier gas that can substantially reduce the deposition of ablated material onto the sample surface and greatly improve transport efficiency, and thus increase the signal intensities, as 36.

(46) compared to “conventional” ablation using argon as the carrier gas. The spot size of ablation is 30 μm, with laser repetition rate and laser energy density being 4 Hz and of 10 J/cm2, respectively. The NIST 610 was used as the external standard; The USGS glass and BCR-2G was also used as secondary standard for verification of analytical results. Ca (determined by EPMA) was used as the internal normalization standard. All concentrations were calculated using the GLITTER 4.0 (GEMOC) software (Andersen, 2002). The results of BCR-2G is found in supplementary materials together with literature reference values, indicating that the precision of about 30 trace elements measured is generally better than 10%.. 37.

(47) 3. Results. 3-1 Anhydrous high-temperature experiment at atmospheric pressure The experimental results at atmospheric pressure are listed in Table 3-1. The range of experimental temperature is between 1310 oC and 1125 oC. Liquidus temperature of the gabbroic melt at atmospheric pressure is determined as ~1303 oC, whereas the solidus temperature is estimated to be ~1120 oC and the melting interval is ~183 oC. The crystallization sequence is: the iron-titanium oxide crystallizes at 1303 oC; pyroxene appears at 1184 oC; and plagioclase is present at 1162 oC. The crystallization sequence of the BIC melt at atmospheric pressure is shown in Fig. 3-1.. 38.

(48) Table 3-1. Run products at atmospheric pressure. Run No.. Temp. (oC). Duration (hrs : mins). BM-023. 1310. 6:00. Gl. BM-028. 1305. 6:10. Gl. BM-005. 1300. 6:10. Gl + FTO. BM-018. 1294. 22:00. Gl + FTO. BM-015. 1286. 21:00. Gl + FTO. BM-014. 1275. 26:50. Gl + FTO. BM-004. 1250. 19:30. Gl + FTO. BM-003. 1198. 47:30. Gl + FTO. BM-024. 1187. 17:40. Gl + FTO. BM-026. 1180. 19:00. Gl + FTO + Px. BM-008. 1173. 20:34. Gl + FTO + Px. BM-022. 1165. 38:20. Gl + FTO + Px. BM-027. 1159. 21:00. Gl + FTO + Px + Pl. BM-001. 1151. 17:20. Gl + FTO + Px + Pl. BM-007. 1125. 19:50. Gl + FTO + Px + Pl. Phase(s)*. *Gl = glass; FTO = Fe-Ti oxide; Pl = plagioclase; Px = pyroxene. 39.

(49) Fig. 3-1. The crystallization sequence of the BIC melt at atmospheric pressure.. 40.

(50) 3-1-1. SEM analysis. Mineral phases with different compositions will appear different grayscale in the back-scattered electron image (BEI). It is not only very helpful for us to identify the phases, textures and mineral composition but also can be solid bases before EPMA work. The highest temperature (1310 oC) of experiment run is filled with glass (Fig. 3-2). The glass is gray in BEI. Iron-titanium oxide is present at 1300 oC. It is eubhedral and white in BEI (Fig. 3-3). Pyroxene appears at 1180 oC. It is subhedral to anhedral and gray in BEI (Fig 3-4). Plagioclase appears at 1159 oC. It is subhedral to euhedral and black in BEI (Fig. 3-5). In the run of 1125 oC, there is a trace amount of glass (Fig. 3-6). The solidus of the BIC gabbro at atmospheric pressure is therefore deduced to be 1120 oC.. 41.

(51) Gl. Fig. 3-2. The BEI of run BM-023 (1310 oC). Gl: glass.. FTO. Gl. Fig. 3-3. The BEI of run BM-005 (1300 oC). FTO: iron-titanium oxide. Gl: glass.. 42.

(52) Px Gl. FTO Fig. 3-4. The BEI of run BM-026 (1180 oC). FTO: iron-titanium oxide, Gl: glass, Px: pyroxene. The pyroxene is mainly subhedral to anhedral and usually has inclusion FTO.. 43.

(53) Pl Px. FTO Gl. Fig. 3-5. The BEI of run BM-027 (1159 oC). FTO: iron-titanium oxide, Gl: glass, Pl: plagioclase, Px: pyroxene. FTO is tiny, less than 20 μm, and clustered together. Plagioclase is mainly in lath shape. Pyroxene is subhedral to anhedral.. 44.

(54) Px. FTO Gl Pl. Fig. 3-6. The BEI of run BM-007 (1125 oC). FTO: iron-titanium oxide, Gl: glass, Pl: plagioclase, Px: pyroxene. The glass is trace. The whole space is mostly occupied by the crystals. The temperature is close to the solidus.. 45.

(55) 3-1-2. Mineral composition of the synthetic phases. The major elements of synthesized minerals were analyzed by electron probe micro-analyzer (EPMA) at Academia Sinica. The trace elements of synthesized minerals were analyzed by laser ablation-inductively coupled plasma mass spectrometry (LA-ICPMS) at National Taiwan University. 3-1-2-1. Major elements. 3-1-2-1-1 Fe-Ti oxide mineral The compositions of the synthesized iron-titanium oxides are listed in Table 3-2. It is the near liquidus mineral. The calculated cation exchange FeO content of iron-titanium oxide at atmospheric pressure ranges from 34.92 wt % to 36.56 wt %, Fe2O3 content ranges from 37.94 wt % to 48.43 wt %, the TiO2 content ranges from 10.70 wt % to 16.78 wt %. At the lowest temperature, the iron-titanium oxide exsolves into two mineral phases, one is Ti-rich spinel and the other is similar to ilmenite.. 46.

(56) Table 3-2. The average compositions of the synthesized Fe-Ti oxide at atmospheric pressure of this study. Run No. Temp. (oC) Avg. of SiO2 TiO2 Al2O3 Cr2O3 Fe2O3 FeO MnO MgO NiO CaO Total* Cations O Si Ti Al Cr Fe3+ Fe2+ Mn Mg Ni Ca Total Mg# b *. Normalized to 100%.. a. Standard deviations of the mean.. b. Mg# = 100 × Mg/(Mg＋Fe2+).. BM-005 1305 1 0.01 10.18 2.21 0.06 48.43 35.37 0.11 3.45 0.09 0.09 100.00. BM-026 1180 8 0.03(0.02)a 13.24(0.28) 1.97(0.05) 0.00 43.91(0.49) 34.92(0.23) 0.20(0.04) 5.54(0.12) 0.03(0.03) 0.15(0.04) 100.00. BM-027 1159 9 0.10(0.05) 14.32(0.43) 1.78(0.09) 0.00 42.17(0.71) 34.96(0.52) 0.26(0.06) 6.14(0.26) 0.04(0.02) 0.23(0.12) 100.00. BM-007-a 1125 2 0.00 16.86(0.67) 1.67(0.02) 0.00 37.94(1.30) 36.56(0.68) 0.34(0.03) 6.52(0.02) 0.04(0.04) 0.07(0.05) 100.00. BM-007-b 1125 2 0.00 46.69(0.62) 1.49(0.59) 0.00 14.42(0.16) 31.25(1.15) 0.14(0.07) 5.81(0.98) 0.04(0.01) 0.15(0.09) 100.00. 4. 4. 4. 4. 6. 0.001 0.281 0.095 0.002 1.338 1.086 0.003 0.189 0.003 0.004 3.001 14.83. 0.001 0.359 0.084 0.000 1.192 1.054 0.006 0.298 0.001 0.006 3.001 22.04. 0.004 0.387 0.075 0.000 1.140 1.050 0.008 0.329 0.001 0.009 3.002 23.85. 0.000 0.454 0.070 0.000 1.021 1.094 0.010 0.348 0.001 0.003 3.001 24.12. 0.000 1.696 0.085 0.000 0.524 1.262 0.006 0.418 0.002 0.008 4.000 24.90. 47.

(57) 3-1-2-1-2 Pyroxene The synthesized pyroxene compositions are listed in Table 3-3. The wollastonite (Wo) component of the synthesized pyroxenes ranges from 45.2 mole % to 46.2 mole %; Enstatite (En) component ranges from 34.8 mole % to 41.5 mole %; and the ferrosilite (Fs) component ranges from 13.3 mole % to 19.1 mole %. They are classified as diopside and augite (Fig. 3-7). The TiO2 content ranges from 0.28 wt % to 1.37wt %. The tFeO content ranges from 8.36 wt % to 11.48 wt %. 3-1-2-1-3 Plagioclase The compositions of synthesized plagioclases are listed in Table 3-4. Most of their An components are from 48.6 mole % to 67.9 mole % and correspond to andesine and labradorite (Fig. 3-8). The Al2O3 component ranges from 27.23 wt % to 27.32 wt %. K2O component ranges from 0.25 wt % to 0.63 wt %.. 48.

(58) Table 3-3. The average compositions of the synthesized pyroxene at atmospheric pressure of this study. Run No. Temp. (oC) Avg. of SiO2 TiO2 Al2O3 Cr2O3 tFeO MnO MgO CaO Na2O K2O Total* Cations O. BM-026 1180 3 46.75(0.50) 1.34(0.08) 5.22(0.51) 0.00 9.69(0.26) 0.15(0.01) 13.73(0.25) 22.43(0.34) 0.68(0.08) 0.01(0.01) 100.00. Si Ti Al Cr Fe Mn Mg Ca Na K Total Mg# Wo En Fs. 1.774 0.038 0.234 0.000 0.307 0.005 0.777 0.912 0.050 0.000 4.096 71.64 45.69 38.91 15.40. 6. BM-027 1159 7 48.26(0.99) 1.37(0.21) 4.35(0.65) 0.00 8.36(1.19) 0.24(0.06) 14.60(1.17) 22.16(0.25) 0.65(0.07) 0.01(0.01) 100.00. BM-007 1125 8 52.42(0.51) 0.28(0.11) 1.13(0.82) 0.00 11.48(0.91) 0.33(0.05) 11.75(0.39) 21.73(0.86) 0.79(0.26) 0.08(0.14) 100.00. 6. 6. 1.814 0.039 0.193 0.000 0.263 0.008 0.818 0.892 0.047 0.000 4.074 75.67 45.23 41.45 13.32. 1.982 0.008 0.050 0.000 0.363 0.011 0.662 0.880 0.058 0.000 4.014 64.59 46.19 34.76 19.05 49.

(59) Fig. 3-7. Variations of synthesized pyroxene composition compared with those of natural pyroxene in BIC. The natural data from Shellnutt and Pang (2012). The boundary lines are based on Morimoto (1988).. 50.

(60) Table 3-4. The average compositions of the synthesized plagioclase at atmospheric pressure of this study. Run No. Temp. (oC) Avg. of SiO2 TiO2 Al2O3 Cr2O3 tFeO MnO MgO CaO Na2O K2O Total* Cations O. BM-027 1159 3 52.56(0.13) 0.18(0.03) 27.23(0.25) 0.00 2.56(0.04) 0.01(0.01) 0.23(0.02) 13.03(0.26) 3.96(0.15) 0.25(0.01) 100.00 8. 8. Si Ti Al Cr Fe Mn Mg Ni Ca Na K Total. 2.417 0.006 1.476 0.000 0.098 0.000 0.016 0.000 0.642 0.350 0.010 5.023. 2.515 0.008 1.455 0.000 0.040 0.002 0.019 0.000 0.465 0.460 0.040 4.995. An Ab Or. 63.58 34.99 1.43. BM-007 1125 5 55.65(1.10) 0.23(0.16) 27.32(1.03) 0.01(0.01) 1.06(0.51) 0.04(0.04) 0.28(0.34) 9.59(1.02) 5.20(0.58) 0.63(0.34) 100.00. 48.57 47.61 3.82 51.

(61) Fig. 3-8. Variations of synthesized plagioclase composition compared with those of natural plagioclases in BIC. The natural plagioclase data are from Shellnutt and Pang (2012). The boundary lines are drawn based on Deer (1963).. 52.

(62) 3-1-2-1-4 Glass At the end of the high-temperature experiment, the residual melts are quenched into glasses. The composition of glasses in the run products at temperatures between 1310 oC to 1125 oC were analyzed and are listed in Table 3-5. The variations of the oxides in the glass compositions versus temperature are plotted in Fig. 3-9. As the temperature decreased, the liquid line of descend demonstrated that SiO2, Al2O3, Na2O and K2O were enriched, whereas MgO, TiO2, and total FeO were depleted. The CaO of the residual liquids became enriched between 1305 oC and 1180 oC, and then depleted in the final stage at lower temperatures. The crystallization trend indicated that the residual melts were considerably depleted in TiO2 and total FeO when the temperature was below 1305 oC. This depletion was triggered by the extensive crystallization of the iron-titanium oxides. When the temperature decreased below 1180 oC, the CaO and MgO contents of the residual liquids became depleted as a result of the crystallization of clinopyroxene. The variation diagrams of the glass compositions versus silica are shown in Fig 3-10. The negative correlations of CaO, TiO2, tFeO, MgO, and positive trends of Al2O3, K2O against SiO2 are shown. The BIC residual melt has highly variable concentrations of compatible trace elements, for example, Sc (15.1 – 30.6 ppm), Ni (16.1 – 54.8 ppm), and Cr (3.8 – 26.1 ppm). 53.

(63) Table 3-5. The average compositions of the glass at atmospheric pressure of this study. Run No. Temp. (oC) Avg. of SiO2 TiO2 Al2O Cr2O3 tFeO MnO MgO CaO Na2O K2O Total* CIPW Quartz Plagioclase Orthoclase Nepheline Diopside Hypersthene Olivine Ilmenite Magnetite Mg#. BM-023 1310 165 47.00(0.36) 3.96(0.41) 12.68(0.26) 0.01(0.01) 15.04(0.21) 0.20(0.04) 6.26(0.10) 10.06(0.12) 3.57(0.07) 1.20(0.04) 100.00. 0.00 36.06 7.21 4.96 28.66 0.00 12.19 7.31 3.61 42.66. BM-005 1300 73 46.94(0.27) 4.09(0.12) 12.92(0.22) 0.01(0.02) 14.84(0.18) 0.21(0.04) 6.24(0.10) 10.16(0.13) 3.40(0.07) 1.19(0.03) 100.00. 0.00 37.41 7.03 4.20 28.01 0.00 12.03 7.75 3.58 42.84. BM-026 1180 6 52.63(0.27) 2.98(0.12) 14.21(0.24) 0.00 7.58(0.21) 0.21(0.04) 6.00(0.10) 11.00(0.15) 3.98(0.07) 1.41(0.03) 100.00. 0.00 47.48 8.33 1.56 30.45 0.00 4.70 5.66 1.83 58.53. 54. BM-027 1159 7 55.15(0.8) 2.62(0.34) 14.14(1.13) 0.00 6.50(0.31) 0.20(0.05) 5.40(0.81) 9.76(1.02) 4.49(0.38) 1.74(0.14) 100.00. 0.00 51.24 10.22 28.18 0.26 0.00 3.55 4.98 1.57 59.70. BM-007 1125 4 59.57(0.52) 2.11(0.22) 15.66(0.55) 0.01(0.01) 4.97(0.31) 0.15(0.04) 3.96(0.46) 6.31(0.59) 4.32(0.29) 2.94(0.35) 100.00. 5.59 51.21 17.32 0.00 13.35 7.32 0.00 4.01 1.20 58.68.

(64) Fig. 3-9. Variation of glass composition versus temperature at atmospheric pressure.. 55.

(65) Fig. 3-10. The residual glass composition at 1 atm compared with natural rock in BIC by Shellnutt et al. (2009, 2010). (a): Silica versus aluminum oxide contents, (b): Silica versus calcium oxide, (c): Silica versus total iron oxide, (d): Silica versus titanium dioxide, (e): Silica versus potassium oxide, (f): Silica versus magnesium oxide. 56.

(66) 3-1-2-2. Trace elements. The trace elements of the synthesized pyroxenes of this study are listed in Table 3-6. The trace element compositions of the synthesized plagioclase of this study are listed in Table 3-7. The trace element compositions of the residual melt of this study are listed in Table 3-8. The concentrations of compatible elements of the residual melts are highly variable; for example: Sc (15.1 – 30.6 ppm), Ni (16.1 – 54.8 ppm), and Cr (3.8 – 26.1 ppm). The evolutions of the trace elements versus silica of the residual melt are illustrated in SiO2-Rb, SiO2-Ba, SiO2-Nd, SiO2-Zr diagrams (Fig 3-11). As the SiO2 content of the residual glass increased, the amount of Ba, Rb, Nd, and Zr increased. The variations of trace elements of glass compositions versus temperature are plotted in Fig. 3-12. As the temperature decreased, the line of descend demonstrated that Nd, Sr, Zr, Nb, Rb, and Ba were enriched. It is consistent with the crystallization of the iron-titanium oxides and pyroxene. Those elements became depleted along with the crystallization of the plagioclase. The Ni and Cr of the residual liquids became depleted between 1305 oC and 1159 oC because of the crystallization of the iron-titanium oxides and pyroxene. They became enriched in the final stage as the plagioclase crystallized at lower temperatures.. 57.

(67) Table 3-6. The average compositions of trace element of the synthesized pyroxene at atmospheric pressure of this study. Run No.. BM26. BM27. 1180. 1159. Avg. of. 2. 2. Sc(ppm). 60.4. 115.4. Cr. 39.2. 51.9. Ni. 74.3. 107.7. Rb. o. Temp. ( C). 148.9. 3.5. 86. 900.6. 70.8. 88. Sr. 878.8. 63.5. Y. 52.2. 23.7. Zr Nb. 320.9 48.8. 84.9 2.0. Ba. 643.0. 13.5. La. 49.0. 4.7. Ce. 113.1. 15.4. Pr. 14.9. 3.2. Nd. 64.5. 18.3. Sm. 13.6. 5.6. Eu. 4.6. 2.0. Gd. 12.6. 6.0. Tb. 1.8. 0.9. Dy. 10.1. 5.7. Ho. 2.1. 1.0. Er. 5.2. 2.4. Tm Yb. 0.7 4.7. 0.3 1.9. Lu. 0.6. 0.2. Hf. 8.2. 3.8. Ta. 2.8. 0.2. Sr. 58.

(68) Table 3-7. The trace element compositions of the synthesized plagioclase at atmospheric pressure of this study. Run No. o. Temp. ( C). BM27 1159. Avg. of. 1. Sc(ppm). 25.7. Cr. 4.5. Ni. 18.9. Rb Sr86. 113.5 740.3. Sr88 Y. 726.0 37.7. Zr Nb. 288.5 34.7. Ba. 511.9. La. 38.0. Ce Pr. 85.1 11.3. Nd Sm. 48.9 10.7. Eu Gd. 3.2 9.6. Tb Dy. 1.3 8.1. Ho Er. 1.5 4.0. Tm Yb. 0.5 3.4. Lu Hf. 0.5 7.3. Ta. 2.0. 59.

(69) Table 3-8. The average compositions of trace element of the glass at atmospheric pressure of this study. Run No. Temp. (oC) Avg. of. BM23 1310 9. BM26 1180 6. BM27 1159 2. BM07 1125 3. Sc(ppm) Cr. 30.6 26.1. 27.7 14.9. 22.1 3.8. 15.1 10.7. Ni Rb. 54.8 69.4. 36.6 74.4. 16.1 93.9. 22.1 89.4. Sr86 Sr88. 429.3 429.6. 464.8 460.9. 509.1 505.2. 453.3 450.3. Y. 25.8. 27.3. 31.7. 24.3. Zr Nb. 172.9 24.2. 184.2 25.9. 215.0 29.0. 151.9 21.9. Ba La. 321.8 24.5. 340.7 26.0. 422.0 31.4. 404.6 27.4. Ce Pr. 55.0 7.4. 58.4 7.8. 70.1 9.4. 60.3 7.7. Nd Sm Eu. 31.8 6.9 2.2. 33.4 7.2 2.3. 40.1 8.7 2.8. 31.9 6.4 2.3. Gd Tb. 6.3 0.9. 6.7 0.9. 7.6 1.1. 5.8 0.8. Dy Ho. 5.4 1.0. 5.6 1.1. 6.6 1.3. 5.1 0.9. Er Tm. 2.6 0.3. 2.8 0.4. 3.1 0.4. 2.4 0.3. Yb Lu. 2.2 0.3. 2.4 0.3. 2.6 0.4. 2.1 0.3. Hf Ta. 4.6 1.5. 4.7 1.6. 5.6 1.7. 3.6 1.1. 60.

(70) Fig. 3-11. The residual glass trace elements composition at atmospheric pressure of this study compared with natural rock in BIC (black star is GS04-026 data) by Shellnutt et al. (2008). Silica versus Rb, Ba, Nd and Zr (symbols as in Fig. 3-10).. 61.

(71) Fig. 3-12. Variation of trace elements of glass compositions versus temperatures at atmospheric pressure.. 62.

(72) 3-2 Anhydrous high pressure high temperature experiment at 1 GPa The experimental results at 1 GPa are listed in Table 3-9. The range of experimental temperature is between 1240 oC and 1000 oC. Liquidus temperature of the gabbroic melt is determined to be ~1220 oC, whereas the solidus temperature of it is estimated to be ~980 oC and the melting interval is determined to be ~240 oC. The crystallization sequence of the gabbroic melt is: the iron-titanium oxide and pyroxene appears at 1220 oC; and plagioclase is present at 1130 oC. The crystallization sequence of the BIC melt at 1 GPa is shown in Fig. 3-13. The phase relationship of the gabbroic melt at atmospheric pressure and 1 GPa are plotted in a pressure-temperature diagram Fig. 3-14. 3-2-1. SEM analysis. The highest temperature (1240 oC) of experiment run is filled with glass (Fig. 3-15). Iron-titanium oxide and pyroxene crystallizes at 1220 oC. The glass is gray, iron-titanium oxide is subhedral and white, and pyroxene is subhedral and gray in BEI (Fig. 3-16). Plagioclase appears at 1130 oC. It is subhedral and darker than the phases mentioned above in BEI (Fig. 3-17). In the run at 1000 oC, there is a trace amount of glass (Fig. 3-18). The solidus of the BIC gabbro at 1 GPa is, therefore, deduced to be 980 oC.. 63.

(73) Table 3-9. Run products at 1 GPa.. Run. Temp. Duration Phase(s)*. o. No.. （ C）. （hrs：mins）. BM-H20. 1240. 7:00. Gl. BM-H18. 1200. 6:00. Gl + Px + FTO. BM-H16. 1140. 12:30. Gl + Px + FTO. BM-H09. 1120. 13:30. Gl + Px + FTO + Pl. BM-H06. 1100. 6:00. Gl + Px + FTO + Pl. BM-H05. 1000. 6:00. Gl + Px + FTO + Pl. *FTO = Fe-Ti oxide; Gl = glass; Pl = plagioclase; Px = pyroxene.. 64.

(74) Fig. 3-13. The crystallization sequence of the BIC melt at 1 GPa.. Fig. 3-14. Phase relations of a BIC gabbro at 0 to 1.0 GPa.. 65.

(75) Fig. 3-15. The BEI of run BM-H20 (1240 oC). Gl: glass. This is the highest temperature run of the experiments at 1 GPa. The glass is homogeneous.. Fig. 3-16. The BEI of run BM-H16 (1140 oC). FTO: iron-titanium oxide, Gl: glass, Px: pyroxene. Pyroxene is subhedral to anhedral. 66.

(76) Fig. 3-17. The BEI of run BM-H09 (1120 oC). FTO: iron-titanium oxide, Gl: glass, Pl: plagioclase, Px: pyroxene. Plagioclase is mainly in lath shape. Pyroxene is subhedral to anhedral.. Fig. 3-18. The BEI of run BM-H05 (1000 oC). FTO: iron-titanium oxide, Gl: glass, Pl: plagioclase, Px: pyroxene. Pyroxene is subhedral to anhedral. The glass is of trace amount. 67.

(77) 3-2-2. Mineral composition of the synthetic phases. The major elements of synthesized minerals were analyzed by the electron probe micro-analysis (EPMA) at Academia Sinica. The chemistry of the synthesized minerals will be discussed as follows: 3-2-2-1. Fe-Ti oxide. The compositions of the synthesized iron-titanium oxide at 1 GPa are listed in Table 3-10. It is the near liquidus mineral. The FeO of the iron-titanium oxides at 1GPa ranges from 22.73 wt % to 35.91 wt %, Fe2O3 component ranges from 16.24 wt % to 44.81 wt %. At lower temperatures (1000 oC), the iron-titanium oxides are exsolved into two (a and b) types. The FeO component of a-type iron-titanium oxides ranges from 21.55 wt % to 33.49 wt %, Fe2O3 component ranges is from 14.37 wt % to 39.54 wt % and TiO2 component of a-type ranges from 29.76 wt % to 46.43 wt %, whereas the FeO component of b-type ranges from 33.90 wt % to 34.96 wt %, Fe2O3 component ranges is from 39.67 wt % to 46.24 wt %.. 68.

(78) Table 3-10. The average compositions of the synthesized Fe-Ti oxide at 1 GPa of this study. Run No. Temp. (oC) Avg. of SiO2 TiO2 Al2O3 Cr2O3 Fe2O3 FeO MnO MgO NiO CaO Na2O ZnO Total* Cations O Si Ti Al Cr Fe3+ Fe2+ Mn Mg Ni Ca Na Zn Total Mg#. BM-H16 1140 2 0.16(0.09) 31.44(0.03) 1.82(0.01) 0.03(0.04) 40.48(0.12) 22.73(0.03) 0.04(0.06) 3.00(0.09) 0.17(0.08) 0.14(0.02) 0.00 0.00 100.00 6 0.008 1.179 0.107 0.001 1.518 0.948 0.002 0.223 0.007 0.008 0.000 0.000 4.000 19.04. BM-H09 1120 14 0.05(0.04) 32.70(0.36) 1.64(0.06) 0.02(0.02) 38.56(0.71) 23.67(0.39) 0.10(0.03) 3.09(0.13) 0.00 0.12(0.06) 0.00 0.04(0.04) 100.00 6 0.002 1.226 0.097 0.001 1.446 0.986 0.004 0.230 0.000 0.006 0.000 0.002 4.000 18.88. BM-H05-a 1000 3 0.03(0.02) 45.51(1.02) 0.65(0.11) 0.01(0.01) 16.24(1.72) 32.72(0.75) 0.42(0.07) 4.28(0.24) 0.00 0.14(0.02) 0.00 0.00 100.00 6. BM-H05-b 1000 2 0.11(0.00) 9.94(2.36) 5.15(0.52) 0.18(0.07) 44.81(4.74) 35.91(0.87) 0.37(0.01) 3.20(0.90) 0.00 0.12(0.06) 0.17(0.25) 0.03(0.03) 100.00 4. 0.001 1.680 0.038 0.000 0.600 1.343 0.017 0.313 0.000 0.008 0.000 0.000 4.000 18.92. 0.004 0.271 0.220 0.005 1.220 1.087 0.011 0.173 0.000 0.005 0.012 0.001 3.009 13.71. * Normalized to 100%. 69.

(79) 3-2-2-2 Pyroxene. The compositions of synthesized pyroxenes at 1 GPa are listed in Table 3-11. The wollastonite (Wo) component of the synthesized pyroxenes ranges from 37.77 mole % to 46.48 mole %; Enstatite (En) component from 33.83 mole % to 38.62 mole %; and the ferrosilite (Fs) component from 18.92 mole % to 25.3 mole %. They are classified as diopside and augite (Fig. 3-19). The synthesized pyroxenes of this study were divided into two types at lower temperature (1120 oC). The TiO2 component ranges from 0.23 wt % to 1.78 wt %. The tFeO component ranges from 11.53 wt % to 14.02 wt %. 3-2-2-3 Plagioclase The compositions of synthesized plagioclases at 1 GPa are listed in Table 3-12. Most of their An components are from 20.8 mole % to 35.9 mole % and are consistent with andesine and oligoclase (Fig. 3-20). The Al2O3 component ranges from 22.8 wt % to 24.77 wt %. K2O component ranges from 0.51 wt % to 0.91 wt %.. 70.

(80) Table 3-11. The average compositions of the synthesized pyroxene at 1 GPa of this study. Run No. Temp. (oC) Avg. of SiO2 TiO2 Al2O3 Cr2O3 tFeO MnO MgO CaO Na2O K2O Total* Cations O Si Ti Al Cr Fe Mn Mg Ca Na K Total Wo En Fs Mg#. BM-H16 1140 4 48.38(0.85) 1.78(0.17) 6.19(0.66) 0.06(0.07) 11.66(0.26) 0.22(0.04) 12.32(0.29) 18.15(0.37) 1.23(0.09) 0.02(0.02) 100.00. BM-H09-a 1120 6 47.51(0.61) 1.22(0.16) 7.74(0.96) 0.01(0.02) 11.82(0.20) 0.20(0.05) 11.68(0.54) 18.56(0.22) 1.26(0.08) 0.00 100.00. BM-H09-b 1120 3 52.22(0.24) 0.23(0.10) 1.03(0.18) 0.00 11.53(0.36) 0.35(0.10) 12.04(0.63) 21.84(0.61) 0.76(0.05) 0.00 100.00. 6 1.821 0.051 0.275 0.002 0.367 0.007 0.691 0.732 0.090 0.001 4.036. 6 1.792 0.035 0.344 0.000 0.373 0.006 0.656 0.750 0.090 0.000 4.047. 6 1.975 0.006 0.046 0.000 0.365 0.011 0.679 0.885 0.056 0.000 4.023. 40.89 38.62 20.50 65.33. 42.14 36.90 20.96 63.77. BM-H05-a 1000 1 49.31 0.88 6.36 0.00 14.02 0.34 11.47 16.32 1.19 0.10 100.00 6 1.861 0.025 0.283 0.000 0.442 0.011 0.645 0.660 0.087 0.000 4.016. 45.89 35.19 18.92 65.04. 37.77 36.92 25.31 59.32 71. BM-H05-b 1000 8 52.47(0.34) 0.24(0.11) 0.94(0.53) 0.00 11.90(0.44) 0.32(0.05) 11.47(0.27) 21.93(0.52) 0.74(0.06) 0.00 100.00 6 1.986 0.007 0.042 0.000 0.377 0.010 0.647 0.889 0.050 0.000 4.013 46.48 33.83 19.69 63.22.

(81) (a) 1120 oC. (b) 1000 oC. 72.

(82) (c) All. Fig. 3-19. (a) The pyroxene is classified as diopside and augite at 1120 oC. (b) The lowest temperature (1000 oC) of experiment run has diopside and augite. (c) Variations of synthesized pyroxene composition at 1 GPa and compared with those of natural pyroxene in BIC. The natural data from Shellnutt and Pang (2012). The boundary lines are based on Morimoto (1988).. 73.

(83) Table 3-12. The average compositions of the synthesized plagioclase at 1 GPa of this study. Run No. Temp. (oC) Avg. of SiO2 TiO2 Al2O3 Cr2O3 tFeO MnO MgO CaO Na2O K2O Total* Cations O Si Ti Al Cr Fe Mn Mg Ca Na K Total An Ab Or. BM-H09 1120 5 59.23(0.80) 0.16(0.08) 24.77(0.23) 0.01(0.01) 0.87(0.25) 0.00 0.07(0.08) 7.33(0.63) 6.64(0.24) 0.91(0.12) 100.00. BM-H05 1000 5 63.35(0.94) 0.01(0.01) 22.80(0.50) 0.00 0.41(0.05) 0.01(0.01) 0.00 4.27(0.75) 8.64(0.43) 0.51(0.22) 100.00. 8 2.659 0.005 1.311 0.000 0.033 0.000 0.005 0.352 0.580 0.050 4.996. 8 2.806 0.000 1.190 0.000 0.015 0.000 0.000 0.202 0.740 0.030 4.985. 35.9 58.8 5.3. 20.8 76.2 3.0 74.

(84) Fig. 3-20. Variations of synthesized plagioclase composition at 1 GPa are compared with natural plagioclases in BIC. The natural plagioclase data are from Shellnutt and Pang (2012). The boundary lines are drawn based on Deer (1963).. 75.

(85) 3-2-2-4 Glass The composition of the glasses in the run products at temperatures between 1240 o. C to 1000 oC were analyzed and are listed in Table 3-13. Figure 3-21 shows that the. variation of glass composition versus temperature. As the temperature decreased, the liquid line of descend demonstrated Al2O3-enriched and K2O-enriched, whereas CaO-depleted and MgO-depleted by the crystallization of pyroxene. The TiO2 and total FeO of the residual liquids became enriched accompanying the presence of iron-titanium oxide from the initial stage to the middle stage between 1240 oC and 1120 o. C, and then depleted from the middle stage to the final stage at lower temperatures. The. SiO2 and Na2O content of the residual liquids became slightly enriched in the beginning then became depleted a little in the middle stage and then enriched in the final stage as the temperature decreased.. 76.

(86) Table 3-13. The average compositions of the glass at 1 GPa of this study. Run No. BM-H20 BM-H16 BM-H09 Temp. (oC) 1240 1140 1120 Avg. of 9 9 2 SiO2 55.17(0.74) 57.18(0.56) 50.61(0.51) TiO2 2.94(0.25) 3.26(0.16) 3.75(0.03) Al2O3 12.75(0.46) 11.28(0.36) 14.75(0.28) Cr2O3 0.01(0.02) 0.00 0.00 tFeO 10.83(0.89) 11.80(0.27) 13.32(0.43) MnO 0.13(0.09) 0.15(0.08) 0.18(0.03) MgO 5.96(0.14) 3.95(0.13) 4.39(0.19) NiO 0.06(0.07) 0.06(0.04) 0.00 CaO 7.36(0.16) 6.38(0.12) 7.46(0.21) Na2O 3.80(0.17) 4.55(0.09) 3.89(0.04) K2O 0.99(0.08) 1.39(0.04) 1.63(0.06) Total* 100.00 100.00 100.00 CIPW Quartz 4.40 6.02 0.00 Plagioclase 46.90 44.71 50.77 Orthoclase 5.85 8.16 9.63 Nepheline 0.00 0.00 0.00 Corundum 0.00 0.00 0.00 Diopside 17.76 21.06 15.84 Hypersthene 16.93 11.02 5.93 Olivine 0.00 0.00 7.52 Acmite 0.00 0.00 0.00 Na2SiO3 0.00 0.00 0.00 Ilmenite 5.56 6.17 7.10 Magnetite 2.61 2.84 3.22 Mg# 49.50 37.40 37.01 77. BM-H05 1000 6 57.22(0.91) 1.46(0.38) 19.25(0.45) 0.01(0.01) 8.43(0.30) 0.14(0.03) 1.96(0.31) 0.00 3.91(0.22) 4.19(0.87) 3.43(0.21) 100.00 3.98 54.77 20.21 0.00 1.53 0.00 14.70 0.00 0.00 0.00 2.77 2.03 29.26.

(87) Fig. 3-21. Variation of glass compositions versus temperatures at 1 GPa.. 78.

(88) 4. Discussion. 4-1 Texture of minerals The Fig. 4-1(a) shows that first crystallized mineral, iron-titanium oxide, encloses glass. The Fig. 4-1(b) shows that pyroxene encloses both glass and iron-titanium oxide. They are similar to the textures of Fe-Ti oxides in the thin-section of the rocks of the Panzhihua intrusion described by Pang et al. (2008). In Fig. 4-1(c), it shows the magnetite inclusions in a clinopyroxene grain enclosed in Fe-Ti oxide ore.. 79.

(89) Fig. 4-1. (a) The BEI of run BM-027 shows that iron-titanium oxide encloses glass. (b) The BEI of run BM-007 shows that pyroxene encloses glass and iron-titanium oxide. (c) The magnetite inclusions in a clinopyroxene grain enclosed in iron-titanium oxide ore (Cpx = clinopyroxene, Opa = opaque oxides) (Pang et al., 2008). 80.

(90) 4-2 Mineral and residual melt composition Figures 3-7 and 3-19 show synthesized pyroxene composition of 1 GPa and 1 atm and the data of natural pyroxene from the BIC. Figures 3-8 and 3-20 show the synthesized plagioclase composition of this study compared with natural plagioclase in BIC. The results of the 1 atm are compositionally more similar to the natural rock data than the results from the 1 GPa. The Ba and Sr contents of the residual melts at atmospheric pressure increased from the highest temperature to the lowest temperature (Fig 4-2b). The chemical variability follows the path outlined by Shellnutt et al. (2009) for the BIC in which crystallization of mafic silicate and Fe-Ti oxide minerals crystallize from the lower portion to upper portion (Fig 4-2a). As iron-titanium oxide and pyroxene crystallized, strontium and barium both increase in the residual liquid. As plagioclase crystallizes, strontium decreases and barium increases. Finally, as the residual liquid becomes more silicic alkali feldspar should crystallize and therefore both should decrease. The residual glass compositions at 1 GPa and 1 atm of this study are plotted in Fig. 4-3 to be compared with the natural rocks of BIC (Shellnutt et al., 2009, 2010) and Baima layered series (Chen, 1990). The „ultramafic series‟ consist of Fe-Ti oxides, olivine, clinopyroxene and spinel which have the highest densities among the minerals. 81.