15839

Measuring the bulk composition of exo-planets is crucial for the development of planet formation theories comma and currently the only available method is spectroscopy of white dwarfs accreting planetary debris. The photospheric abundances of these white dwarfs reflect the bulk compositions of the accreted parent bodies comma entirely analogous to meteorite abundance studies. HST ultraviolet spectroscopy has played a pivotal role in this field comma providing abundance measurements of many elements inaccessible from the ground. The bulk compositions measured so far are broadly similar to those found in the solar system comma suggesting that planet formation may be a fairly uniform process. The accuracy of the measured bulk compositions relies on the physically correct analysis of the white dwarf spectroscopy. However comma all studies have so far used an assumption that remains untested - that the metals are uniformly distributed over the stellar surface. Given accretion proceeds via a geometrically thin disk comma the validity of this assumption is far from obvious. We propose deep COS time-series ultraviolet spectroscopy of G29-38 comma the prototype white dwarf accreting from a debris disk. G29-28 undergoes non-radial pulsations comma which heat localized regions on its surface that can briefly dominate the flux of the star. A spectrum obtained during such a heating event reflects the abundances of a small patch on the white dwarf surface. We will analyze a sufficiently large number of pulses to accurately map the distribution of metals across the entire surface of G29-38. Our results will be used to improve the modelling of debris-accreting white dwarfs comma and lead to more reliable bulk compositions.