We originally fit the X-ray spectrum as a black body, as reported by Walter, Wolk, & Neuhauser (Nature 379, 233). The best fit has kT=57eV and n_{H}=1.4x10^{20} cm^{-2}. While better than any standard one-component model (power law, thermal bremsstrahlung, thermal plasma), the reduced chi square of 2.2 is formally unacceptable. This blackbody fit has a radius R_{}<8 km. The fit underestimates the observed optical fluxes by factors of 2 at U and 3.7 at V.
The Greenstein & Hartke (ApJ 271, 283) model provides for a sinusoidal temperature variation with latitude. A model with T_{pole} = 62eV and T_{equator}=0 fits the optical data better. R_{}<14 km.
No neutron star can be a pure black body: all must have an atmosphere. Fits of atmospheric models may be able to yield the surface composition, the angular diameter and may suggest whether or not magnetic fields are strong. Pavlov et al. (ApJL 472, L33) generated a grid of atmospheric models. Unfortunately, they do not reproduce the optical fluxes.
We are constructing non-magnetic atmospheric models, for a variety of surface compositions. In the figure at left, we show four atmospheric models plus the best fit black body. The atmospheric models use opacity tables from both Los Alamos and the Opacity project. We generate a standard model atmosphere, beginning at an optical depth of 100, and integrate out to =10^{-5}.