Origins of Calc-Alkaline (s.l.) Andesitic Magmas – Where We Stand Today

Abstract:

The mainly igneous continents have intermediate mean SiO₂ (~60 wt%) and Mg# (0.4-0.55 mol). Bowen championed crystallization-differentiation of basalts to silicic daughter liquids, but he encountered difficulties with conflicting evidence for iron enrichment. We now know that as dry reduced basalts crystallize, melt SiO₂ begins climbing steeply at F (melt fraction): 0.25-0.35, reaching intermediate values at F: 0.1-0.2. H₂O changes basalt crystallization by (1) increasing the ratio of olivine to high-Ca pyroxene, (2) delaying onset and reducing the proportion of plagioclase, (3) stabilizing calcic plagioclase, (4) destabilizing low-Ca pyroxene, (5) eventually stabilizing amphibole, and (6) destabilizing silicate minerals more than magnetite. Factors 1-5 enrich melts early in SiO₂, 6 limits decline of Mg#. H₂O of 2-6 wt% is sufficient for early SiO₂ enrichment; fO₂ modestly above Ni-NiO yields daughter liquids with high (calc-alkaline) Mg#s. These values are matched by many natural subduction zone magmas, so seemingly, the problem is solved.

Yet, difficulties remain: Melt SiO₂ enrichment commences at F: 0.6-0.7 in experiments on hydrous Mg basalts, reaching intermediate SiO₂ at F: 0.2-0.4. Cumulates 60-80% of the basalt mass must have returned to the mantle if basalt differentiation formed the continents. Differentiation mainly near the base of the crust has geologic support, and would aid disposal of cumulates, but most experiments on hydrous basalts at such pressures yield peraluminous andesitic liquids, unlike common metaluminous andesites. A composite scenario avoiding peraluminous intermediates is crystallization-differentiation plus mixing with silicic liquids mainly residual or remelted from shortly antecedent intrusions. Even in this case, cumulates and restites exceeding the continental mass must have been lost. The Archean was different: Residual-eclogite-sourced granitoids are widespread, but experiments on hydrous eclogites at plausible temperatures yield melts with SiO₂ too high and Mg# too low. Reaction of such melts with peridotite is a likely solution, but must concur with the somewhat heavy oxygen in zircons from many Archean plutons. If the heavy oxygen derives from slab alteration, it provides a limiting measure of peridotite reaction.