A Non-LTE Model for the Origin of the CO First Overtone Band Emission in Young Stellar Objects.
Session 32 -- Circumstellar Environments
Oral presentation, Wednesday, January 12, 2:15-3:45, Salon VI Room (Crystal Gateway)

## [32.06] A Non-LTE Model for the Origin of the CO First Overtone Band Emission in Young Stellar Objects.

P. N. Safier (UC Berkeley), S. Martin, A. K\"onigl (U. Chicago)

We propose that the CO first overtone emission detected in YSOs originates in a centrifugally driven disk-wind where the level population of CO is dominated by non-LTE effects. Using a revised set of collisional cross sections, we show that there is no need to invoke an LTE level population to account for the observed fluxes and line ratios. We use the disk-wind models presented by Safier (1993; ApJ 408, 115), and compute for the four lowest $\Delta v=2$ transitions the emergent intensities and line profiles. In contrast to previous work by other authors, the thermal structure of the emitting region is not freely specified to match the data, but is computed self-consistently once the wind's mass outflow rate, the stellar parameters ($M_\ast$, $R_\ast$, and $T_\ast$), and the wind dynamical parameters (ratio of magnetic flux to mass flux and the initial angle between the magnetic field and the disk surface) are specified. These winds are endowed with a robust heating mechanism, ambipolar diffusion heating, which keeps the wind hot ($T_{\rm wind} \sim$ a few thousand K) at large distances (several AU) from the central star. We compute the abundance of CO in the wind using a simplified chemical network that includes the destruction and formation of $\rm H_2$ and the destruction of CO by endothermic reactions with H atoms. In addition, we do not assume that the populations of the vibrational levels of CO are in LTE (as assumed by other workers in the field), but carry a full NLTE calculation which includes IR radiative pumping and the surprisal'' cross sections for collisional excitation. The agreement of our results with the observed fluxes and line profiles for the $v=2\rightarrow 0$ and $v=3\rightarrow 1$ transitions is excellent, but because we compute the level population for only the lowest six vibrational levels, the computed fluxes for the $v=4\rightarrow 2$ and higher transitions do not match the data. A better fit to these transitions requires the inclusion of many more vibrational levels in the model molecule. We have under way calculations which include at least the lowest ten vibrational levels of the ground electronic state of CO.