IMPLICATIONS FOR THE HISTORY OF SUNSTONES FROM SOUTHWESTERN OREGON FROM NEW CHEMICAL, ISOTOPIC AND GEOCHRONOLOGIC DATA

Plagioclase megacrysts (1-10 cm) that contain macroscopic inclusions of native copper are locally hosted in basalt flows of eastern Oregon. The native Cu in these ‘sunstones’ occurs as thin platelets (‘copper schiller’) with crystallographically-controlled orientations. These Cu platelets appear to have formed via exsolution, and are typically found in the cores of the highest-grade gemstones. Apart from their Cu, the labradorite megacrysts (~An₆₇) have remarkably homogeneous distributions of major and trace elements, and they exhibit internally homogeneous ⁸⁷Sr/⁸⁶Sr ratios of ~ 0.70365, comparable to plagioclase in the Steens Basalt of the Columbia River Basalt Group (CRBG). The trace element compositions for ~10 sunstone megacrysts from the Dust Devil Mine are more chondrite-like than typical basalts. In agreement with the homogeneous ⁸⁷Sr/⁸⁶Sr and major element data, our LA-ICPMS data also show that the megacrysts are unzoned with respect to trace element concentrations from core to rim. The homogeneous nature of ⁸⁷Sr/⁸⁶Sr and the major and trace element data indicate that the megacrysts have generally undergone little chemical transformation (diffusive mass transport, alteration, weathering) following crystallization. A mid-Miocene ⁴⁰Ar/³⁹Ar age has been determined for finely grained matrix phases of basalts hosting the sunstones, comparable to known ages of the CRBG. Conversely, ⁴⁰Ar/³⁹Ar analyses of the megacrysts yield complex spectra, variably affected by radiogenic ⁴⁰Ar loss, recoil and extraneous (non-atmospheric) ⁴⁰Ar, with model ages that are anomalously younger than obtained for the matrix phases. The zoning and distribution of Cu, apparent lack of isotopic and chemical zoning for other elements, and anomalously young ages for these megacrysts are very perplexing. We agree with previous suggestions that Cu was originally distributed homogeneously in the sunstones during their initial crystallization and subsequent cooling was at a rate that allowed Cu to exsolve but also induced lattice damage and pervasive defects. Our new data indicate the rate of cooling was too rapid for substantial diffusion of major and trace elements and Sr isotopes; accompanying lattice damage and defects in the megacrysts promoted ⁴⁰Ar loss resulting in their anomalously young ⁴⁰Ar/³⁹Ar ages.