2001ApJ...548L.197C

We present two Keck High Resolution Echelle Spectrometer spectra (λ/Δλ=40,000) of a bulge K3 giant taken on successive nights during the second caustic crossing of the binary microlensing event EROS BLG-2000-5. This caustic crossing served effectively to resolve the surface of the source star; the spectrum from the second night is dominated by the limb, while the spectrum from the first night is comprised of light from a broader range of radii. To demonstrate that the spectra are adequate to resolve the differences between them, we analyze the Hα line. The equivalent width is ∼8% smaller on the second night, and the signal-to-noise ratio per resolution element (165 and 75 on the two nights, respectively) is sufficient to show that the difference is approximately constant over the ∼2 Å (∼15 resolution element) extent of the line. The sign of the difference is in the expected direction since the limb is the coolest part of the star and therefore should have the weakest Hα. We invite atmosphere modelers to predict the difference spectrum from the entire spectral range 5500 Å<λ<7900 Å so that these predictions can be compared to our observations.