@Article{Engrand:2005, author = {C. Engrand and K.D. McKeegan and L.A. Leshin and G.F. Herzog and C. Schnabel and L.A. Nyquist and D.A. Brownlee}, title={Isotopic compositions of oxygen, iron, chromium, and nickel in cosmic spherules: Toward a better comprehension of atmospheric entry heating effects}, journal={Geochim. Cosmochim. Acta.}, year={2005}, volume={69}, number={}, month={Nov}, pages={5365--5385}, note={}, annote={}, keywords={}, url={http://sims.ess.ucla.edu/PDF/Engrand_et_al_2005_GCA.pdf}, doi={10.1016/j.gca.2005.07.002}, abstract={Large, correlated, mass-dependent enrichments in the heavier isotopes of O, Cr, Fe, and Ni are observed in type-I (metal/metal oxide) cosmic spherules collected from the deep sea. Limited intraparticle variability of oxygen isotope abundances, typically <5$\permil$ in $\delta$$^{18}O, indicates good mixing of the melts and supports the application of the Rayleigh equation for the calculation of fractional evaporative losses during atmospheric entry. Fractional losses for oxygen evaporation from w\"{u}stite, assuming a starting isotopic composition equal to that of air (\delta$$^{18}$O = 23.5$\permil$; $\delta$$^{17}O = 11.8\permil), are in the range 55\%-77\%, and are systematically smaller than evaporative losses calculated for Fe (69\%-85\%), Cr (81\%-95\%), and especially Ni (45\%-99\%). However, as \delta$$^{18}$O values increase, fractional losses for oxygen approach those of Fe, Cr, and Ni indicating a shift in the evaporating species from metallic to oxidized forms as the spherules are progressively oxidized during entry heating. The observed unequal fractional losses of O and Fe can be reconciled by allowing for a kinetic isotope mass-dependent fractionation of atmospheric oxygen during the oxidation process and/or that some metallic Fe may have undergone Rayleigh evaporation before oxidation began. In situ measurements of oxygen isotopic abundances were also performed in 14 type-S (silicate) cosmic spherules, 13 from the Antarctic ice and one from the deep sea. Additional bulk Fe and Cr isotopic abundances were determined for two type-S deep-sea spherules. The isotopic fractionation of Cr isotopes suggest appreciable evaporative loss of Cr, perhaps as a sulfide. The oxygen isotopic compositions for the type-S spherules range from $\delta$$^{18}O = -2\permil to + 27\permil. The intraspherule isotopic variations are typically small, \sim5\% relative, except for the less-heated porphyritic spherules which have preserved large isotopic heterogeneities in at least one case. A plot of \delta$$^{17}$O vs. $\delta$$^{18}O values for these spherules defines a broad parallelogram bounded at higher values of \delta$$^{17}$O by the terrestrial fractionation line, and at lower values of $\delta$$^{17}O by a line parallel to it and anchored near the isotopic composition of \delta$$^{18}$O = -2.5$\permil$ and $\delta$$^{17}$O = -5$\permil$. Lack of independent evidence for substantial evaporative losses suggests that much of this variation reflects the starting isotopic composition of the precursor materials, which likely resembled CO, CM, or CI chondrites. However, the enrichments in heavy isotopes indicate that some mixing with atmospheric oxygen was probably involved during atmospheric entry for some of the spherules. Isotopic fractionation due to evaporation of incoming grain is not required to explain most of the oxygen isotopic data for type-S spherules. However spherules with barred olivine textures that are thought to have experienced a more intense heating than the porphyritic ones might have undergone some distillation. Two cosmic spherules, one classified as a radial pyroxene type and the other showing a glassy texture, show unfractionated oxygen isotopic abundances. They are probably chondrule fragments that survived atmospheric entry unmelted.Possible reasons type-I spherules show larger degrees of isotopic fractionation than type-S spherules include: a) the short duration of the heating pulse associated with the high volatile content of the type-S spherule precursors compared to type-I spherules; b) higher evaporation temperatures for at least a refractory portion of the silicates compared to that of iron metal or oxide; c) lower duration of heating of type-S spherules compared to type-I spherules as a consequence of their lower densities.} }