The occurrence of actinolite in iron deposits of possible magmatic origin has prompted an investigation of the upper-thermal stability of actinolite in the presence of a P-Fe-rich melt. Experiments have been done using (a) natural actinolite with a molar Fe/(Fe+Mg) ratio (=Fe#) of 0.22 (from Pleito Melón, Chile) and phosphoric acid (H₃PO₄); (b) actinolite, magnetite (Mt), and H₃PO₄ and (c) actinolite, Mt, fluorapatite (Fapat) and H₃PO₄ over the range 700 to 950° C, 1 to 4 kbar, for 2-3 days at the NiNiO and Mt-Hm oxygen buffer. The composition of the phases and the melt were analyzed by electron microprobe.

Experiments done with assemblage (a) at 4 kbar produced a P-(Mg,Fe)-rich melt and apatite (Apat) at 950° C or higher, but at these conditions the actinolite was completely transformed into a mixture of orthopyroxene (Opx), clinopyroxene (Cpx), and a Si-rich melt but  no quartz (Qz). At 900° C, instead of a P-Fe-Melt, both a phosphate Ca1.8(Mg,Fe)2.6(PO4)3 and Apat were observed.

Experiments done with assemblage (b) at 4 kbar showed that the actinolite completely breaks down to a mixture of Opx, Cpx, Qz, Apat and a P-Fe-melt at temperatures higher than 900° C. At temperatures lower than 800° C actinolite always reacted to form rims of cummingtonite which coexisted with Mt, Apat, and a P-Fe-rich melt. At 1 kbar, actinolite coexisted with Mt and P-Fe-melt at 750° C without any sign of reaction or re-equilibration.

Similarly, experiments done with assemblage (c) at 4 kbar showed that at 900° C or higher, actinolite breaks down to a mixture of Opx, Cpx, Qz, Fapat and a P-Fe-melt. At temperatures lower than 800° C actinolite reacted to form cummingtonite which can coexist with Fapat, Mt, and a P-Fe-rich melt. Only at oxygen fugacities of the Mt-Ht buffer actinolite did not nucleate cummingtonite.

Based on these set of experiments and the magma immiscibility experiments on andesite AGV1 by Lledo (GSA Abstract, 2003) the estimated temperature for the onset of a P-Fe-rich magma is about 700° C at 4 kbar and 750° C at 1 kbar. These results and the thermodynamic modeling of the stability of actinolite gives a window of possibility of least 200° C at 1 kbar where actinolite with a Fe# £ 0.3 can coexist with a P-Fe-rich melt. These results support the hypothesis that actinolite in the Kiruna-type iron deposits may be magmatic in origin.