A new approach to modeling dust emission in classical novae

This is an exploratory work to determine the feasibility of modeling the 3D geometry of classical novae dust shells using DIRTY, a Monte Carlo radiative transfer dust emission code. We apply DIRTY to optical and infrared photometry of Nova Cen 1991 (V868 Cen) and Nova Her 1991 (V838 Her) and investigate different basic dust shell geometries supported by observations of novae remnants: a homogeneous spherical shell, an inhomogeneous spherical shell, and an equatorial torus. For V868 Cen we find both inhomogeneous spherical shell and equatorial torus fit the data, however the latter provides slightly better fits to J and H-band observations at later times in the outburst; while for V838 Her both models fit the photometry equally well. The worst fit to the photometry involved the homogeneous shell, the classic model used for the past thirty years to model nova dust. The failure of the homogeneous spherical shell model highlights the importance of geometry in computing dust shell parameters. We discuss how our models provide possible explanations for the so-called "isothermal dust" stage and support for the recent discovery that not all novae enter a phase of constant bolometric luminosity. Our modeling method would work hand-in-hand with high-resolution spectroscopy to better understand nova ejecta. Our results stress the importance of taking advantage of the power of modern dust modeling codes such as DIRTY to more properly model the radiative properties of circumstellar dust distributions. We also include in an appendix our work studying the infrared emission of another binary star system: GX 17+2, a neutron star low-mass X-ray binary. We found it unexpectedly undergoes K-band brightening episodes of at least 3.5 magnitudes. The source of these episodes is not known. Prior published observations and new measurements we acquired between 2006 and 2008 suggest that the episodes last at least four hours and have a period of three days. Future bright episodes can be predicted using the ephemeris JDmax(n) = 2454550.79829+(3.01254+/-0.00002)(n) days. A growing body of evidence suggests that this emission could be caused by a synchrotron jet.