2002ApJ...575.1057W

We analyze Lyα fluoresced H₂ lines observed in the UV spectrum of Mira B. We identify 13 different sequences fluoresced by 13 different H₂ transitions within the Lyα line. The observed H₂ line ratios within these sequences imply significant line opacity, so we use a Monte Carlo radiative transfer code to model the line ratios, correcting for opacity effects. We find that the observed line ratios can best be reproduced by assuming that the H₂ is fluoresced in a layer between the observer and Mira B with a temperature and column density of T=3600 K and logN(H₂)=17.3, respectively. The strengths of H₂ absorption features within the Lyα line are roughly consistent with this temperature and column. We use the total flux fluoresced within the 13 sequences to infer the Lyα profile seen by the H₂. In order to explain differences between the shape of this and the observed profile, we have to assume that the observed profile suffers additional interstellar (or circumstellar) H I Lyα absorption with a column density of about logN(HI)=20.35. We also have to assume that the observed profile is about a factor of 2.5 lower in flux than the profile seen by the H₂, and a couple of possible explanations for this behavior are presented. Several lines of evidence lead us to tentatively attribute the fluoresced emission to H₂ that is heated in a photodissociation front within Mira A's wind a few AU from Mira B, although it is possible that interaction between the winds of Mira A and B may also play a role in heating the H₂. We estimate a Mira B mass-loss rate of M{dot}=5x10^–13 M_☉ yr^–1 and a terminal velocity of V_∞=250 km.s^–1, based on wind absorption features in the Mg II h and k lines. We note, however, that the wind is variable and IUE Mg II spectra suggest significantly higher mass-loss rates during the IUE era.