EL LACO, CHILE: EVIDENCE FOR THE ERUPTION OF AN IMMISCIBLE FE-O-S-P MELT

Massive magnetite ores can be divided into two types: Nelsonites, which are apatite and TiO2-rich and intrusive; and Kiruna-type ores, which are apatite and TiO2-poor and intrusive or volcanic. El Laco is the best preserved volcanic Kiruna-type deposit known. The El Laco flows contain abundant vesicles and gas escape tubes. SO4-rich fluorapatite from both El Laco andesites and ores, the abundance of fumarolic sulfate at El Laco, and anhydrite in pyroxene fluid inclusions, suggest that S was important in the magmatic system. Fine-grained "chilled margins" to magnetite ores at El Laco are also SO4-rich. In addition, magnetite-hematite ash layers beneath the Laco Sur ore body contain rounded masses of diadochite, 2 to 10 cm in diameter. These Fe-sulfate-phosphate lumps have rounded upper surfaces, and lower contacts that are deformed to fit the underlying surface, suggesting that they were liquid after they landed. These "lumps" may be evidence that the El Laco magma was saturated in SO4 and/or PO4.

The immiscible relationship between silicate liquids and FeS liquids has been long recognized. There is a continuum between FeS liquids and FeOS liquids with O/S > 1.3. Such liquids are characterized by high Fe and low Ti and Si contents. Some S flows erupted at < 170 oC are observed to spontaneously burn and leave very little or no record. An FeOS magma erupted at > 800 oC, should release S as SO2 gas, sequestering O2 and preventing magnetite from converting to hematite upon eruption. If the iron-ore at El Laco formed from 25 km3 of silicate magma, it would only lower the FeO content of the magma by 0.74 %. The S released (based on the Fe/O/S ratio [1.9/1.3/1.0] in experimental melts) would form 214,000,000 tons of SO2 gas. Assuming 25 km3 of silicate magma, the SO2/magma would be similar to that calculated for some historic eruptions. In addition, magmas with high O/S have very strong "wetting" properties and could, therefore, easily separate from a crystal-liquid mush. As the magma continued to evolve, these FeOS liquids would cool, reach gas saturation, and eventually vesiculate and erupt.

El Laco appears to be the volcanic equivalent of silicate-sulfide liquid immiscibility at high fO2. P2O5 also appears to play an important role. Until we witness flaming Fe-O-S-P lava flowing down the side of a volcano, however, we might not know for sure.