Continuous reactions with amphibole, garnet and plagioclase

Abstract End-member, continuous and degenerate reactions are derived for the multisystem with the six components Na₂O, CaO, (Mg/Fe)O, Al₂O₃, SiO₂, H₂O among the phases plagioclase_ss, garnet_ss, amphibole_ss, cpx, opx, olivine, spinel, quartz and an aqueous fluid. The chemography of this system is degenerate due to the co-linearity 2Opx = Ol + Qtz. This co-linearity has its implications both on reaction space and phase equilibria. From a total of 28 reaction systems, reaction space is derived for nine subsystems (phases in parentheses are absent): Case A₁: (Cpx,Ol) (Cpx,Opx) and (Cpx,Qtz), Case A₂: (Spl,Ol) (Spl,Opx) and (Spl,Qtz), Case B: (Ol,Opx) (Ol,Qtz) and (Opx,Qtz). In the absence of either cpx or spl (case A), three reactions form an invariant point, either [Cpx] or [Spl], where the co-linear phases olivine, opx and quartz coexist on the transformation line 2Opx = Ol + Qtz. Changing mineral compositions force invariant points to move along the line with the different reaction curves changing their relative position according to Schreinemakers’rules. Zero contours, i.e. the location where (a) phase(s) disappear(s) in reaction space correspond to singular points in phase diagrams. Two types are distinguished; singular points of indispensable and of substitutable phases. In the first case the phase disappears from the entire bundle while in the second it disappears from a single reaction. In the specific case where the substitutable phases are also the co-linear ones, two of the three co-linear phases disappear simultaneously. Two of the three reaction curves coincide. In the system including Cpx and Spl (Case B) three reactions, (Ol,Opx) (Ol,Qtz) and (Opx,Qtz), oppose three invariant points, [Ol], [Opx] and [Qtz]. Invariant points no longer move along the line 2Opx = Ol + Qtz. The coincidence of the zero contours of all three co-linear phases in reaction space-the result of the chemographic degeneracy-causes the respective singular points to coincide in the phase diagrams. This is the location where curves must be rearranged in a bundle to conform Schreinemakers’rules. The reaction Grs₁Prp₂= 2 Ol + An is fourth order degenerate and part of all nine subsystems (cases A and B). It can be used to relate the different phase diagrams to one another.     

Immiscibility in Ca-amphiboles

The literature data of nine different occurrences of coexistingmineral pairs of Ca-amphibole have been studied and the bulkvectors, spanning the miscibility gap, derived. The additivecomponent is always impure Mg-tremolite accompanied by someglaucophane and cummingtonite component. The four major exchangecomponents required to describe the compositional variationin coexisting mineral pairs are the edenite (ED), tschermak's(TS), FeMg_–1 and Fe³⁺-tschermak's (FeTs) vector. Trivalentiron is postulated on the basis of excess charges in the bulkvector the size of which coincides with residuals in Al^tet,–Si, Fe and –Mg. The four cations have equal sizes,forming the vector Fe³⁺ Al^tet Mg_–1Si_–1. This distributionscheme is consistent for all the different occurrences and setsthe basis for a comparison. Deviations from the scheme wouldradically complicate the proposed exchange pattern. The ratioTS:ED in most mineral samples fluctuates between one and two.Projection of the data points in the vector space TS–EDonto the line 1ED: 2TS (Tr–Hbl) or 1ED:1TS (Tr–Prg)provides the projected tremolite content (= 1–X_Hbl or = 1–X_prg). This parameter,applied to coexisting pairs, and plotted against the ratio Mg/(Mg+ Fe) shows some characteristic features about the miscibilitygap. In the Mg-pure system the solvus is almost symmetric andlocated in the temperature range between 800 and 870C. Smallamounts (0.10 pfu) of Fe²⁺ in the M(4) -sites and replacingCa have a dramatic effect, forcing the solvus to much lowertemperatures of 650C. An increase in the ratio Fe/(Fe + Mg)causes a shift of the solvus towards more tremolitic compositionswith temperatures 500–650C. The maximum asymmetry ofthe solvus is reached where the Al-poor member (tremolite) hasa composition of =1.0 and Mg/ (Mg + Fe) 0.6. The corresponding Al-rich member has =0.5 and Mg/ (Mg + Fe) 0.4. An anomalyof the solous is observed at Mg/ (Mg + Fe)=0.8. It manifestsas a kind of highly asymmetric ‘sub-gap’ in thetremolite-rich composition range. This is explained by the partitioningof Fe²⁺ into the single M(3) -site and is characterized by athermal hump to 650–700C. KEY WORDS: tremolite; hornblende; pargasite; immiscibility; solous