b'Don Emersons best of Exploration GeophysicsFeaturerocks, provided the initial SO 2content of the magma is greater than 250 ppm. Because this buffer curve lies below the FMQ buffer at high temperatures, but intersects FMQ at ~ 850C and lies well above FMQ at lower temperatures, it represents a relatively oxidising cooling trend that in principle can oxidise ferrous iron in silicates to magnetite via the reaction:9FeO + SO 2+ H 2 O3Fe 3 O 4 + H 2 S.in silicatesTakagi and Tsukimura (1997) calculate that initial SO 2contents of 250-1900 ppm by weight as the dominant sulphurous species are required to precipitate 0.2-1.5 vol % magnetite from granitic melts and show that other fluid buffers, e.g. H 2- H 2 0, CO 2- CH 4 , or CO 2- CO, cannot produce the oxidising trends that are inferred for many calc alkaline granitic rocks. The general relevance of sulphur dioxide buffering of melts is still an open question, however, because the primary contents of sulphur species and other volatile phases in magmas is poorly known(P. Blevin, pers. comm). Reported sulphur contents in granitoids are lower than the values that are required to produce substantial magnetite, but this may reflect significant late-stage loss of sulphur carried away by hydrothermal fluids, which sometimes produce related sulphide ore deposits.Figure 2.Schematic log 10(fO 2) versus Mg/(Fe + Mg) diagram showing effects of adding MgO to the Fe-Si-O system at a fixed temperature (after FrostFrost (1991b) points out that there can be no unique correlation and Lindsley, 1991). OMQ = olivine-magnetite-quartz, QOOp = quartz-olivine- between fO 2during rock formation and Fe3+/Fe2+ of the rock. orthpyroxene, OpMQ = orthopyroxene-magnetite-quartz, OMOp = olivine- For example, rocks that contain the same mineral assemblages magnetite orthopyroxene, OHQ = olivine-haematite-quartz, OHOp = olivine- must have formed at similar fO 2 , but if they have very different haematite-orthopyroxene. The parameter log fO 2is defined as log 10(fO 2 )- log 10 abundances of the iron bearing minerals, they may have very (fO 2 : FMQ). The values of log 10fO 2 )for the FMQ and MH buffers are shown fordifferent absolute and relative abundances of ferrous and ferric reference. Note that these equilibria become displaced towards higher absoluteiron. However, oxygen fugacity of melts and glasses is simply fO 2as Mg/(Mg + Fe) increases, because Mg preferentially enters olivine overdependent on chemical composition, in particular the relative magnetite and magnetite over haematite.abundance of ferrous and ferric iron. Similarly, in the case of saturated). It is evident from Figure 2 that higher Mg content ofvolcanic rocks that are relatively free of cumulate minerals a melt tends to restrict the occurrence of magnetite to higherfO 2can be calculated at a given temperature and pressure, oxygen fugacities. corresponding to crystallisation conditions midway between the liquidus and solidus for the rock, from the whole rock On the other hand, substitution of ferrous iron + titanium forchemical composition, including ferrous and ferric iron. Kress ferric iron in titanomagnetite reduces the activity of magnetiteand Carmichael (1991) show that the Fe3+/Fe2+ ratio is by far and displaces the fayalite-titanomagnetite-quartz equilibriumthe most important term in the relationship between fO 2and downwards with respect to FMQ. Thus, titanomagnetite is stablechemical composition of volcanic rocks.in igneous rocks at lower oxygen fugacities than is end-member magnetite. Similarly, because Ti substitutes even more readilyBlevin (1994) has shown that ferric/ferrous iron ratios in into haematite than into magnetite, addition of Ti to the systemgranitoid rocks are also very highly correlated with oxygen displaces the HM buffer to lower fO 2 . fugacity as calculated from the chemical composition (and confirmed by mineral assemblages that are dependent on For many plutonic rocks that behave essentially as closedoxygen fugacity). Thus Fe3+/Fe2+, often measured as Fe 2 O 3 / systems during their history, in particular many tholeiitic rocks,(FeO + Fe 2 O 3 ), can in practice be used as a proxy for oxygen the Fe/Mg ratio of the silicates, plus the Ti content and thefugacity in granitoids, in spite of the theoretical possibility that ferrous/ferric ratio of the oxides, monitor and, in effect, controlthe nexus between oxidation state and fO 2might be broken for the oxygen fugacity. In this case the cooling history of the rockrocks that formed under very different conditions or that have is characterised by a path in fO 2-T space that is approximatelyexotic compositions.parallel to the standard mineral-buffered curves. In turn the oxygen fugacity influences the composition of the fluid phaseFigure 3(a) plots isopleths in fO 2 - T space for various and the stability of graphite and sulphides in such igneous rocks. titanomagnetite and ilmenite compositions, with the FMQ and HM buffers shown for comparison. Titanomagnetites Fluid buffering, rather than mineral buffering, of oxygenare solid solutions of magnetite, i.e. Fe3+[Fe2+Fe3+]O 4 , and fugacity is evidently important during hydrothermal processesulvospinel, Fe2+[Fe2+Ti4+]O 4 , whereas natural ilmenites invariably that have large fluid-rock ratios. Fluid buffering may also play aincorporate some haematite, Fe3+ 2O 3 , in solid solution with role in some magmatic processes. If the initial volatile contentilmenite, Fe2+Ti4+O 3 . The square brackets indicate octahedral of the magma is sufficiently high, the oxygen fugacity may becations in the spinel phases. Note that the titanomagnetite largely controlled by the fluid phase, rather than by the ferricisopleths are quite oblique to the oxygen buffer curves, and ferrous iron contents of the melt and the crystallisingwhereas the ilmenite isopleths are subparallel to the buffers. minerals. Takagi and Tsukimura (1997) suggest that the SO 2- H 2 SThis implies that as an igneous melt cools and solidifies along buffer may be important in the evolution of oxidised granitica trajectory that is approximately parallel to FMQ, the stable APRIL 2020 PREVIEW 46'