Quantitative spectroscopy of photospheric-phase type II supernovae

We present first results on the quantitative spectroscopic analysis of the photospheric-phase of type II supernovae (SN). The analyses are based on the model atmosphere code, CMFGEN, of Hillier & Miller (1998) which solves the radiative transfer and statistical equilibrium equations in expanding outflows under the constraint of radiative equilibrium. A key asset of CMFGEN is its thorough treatment of line-blanketing due to metal species. From its applicability to hot star environments, the main modifications to the source code were to allow a linear velocity law, a power-law density distribution, an adaptive grid to handle the steep H recombination/ionization front occurring in some SN models, and a routine to compute the gray temperature structure in the presence of large velocities. In this first paper we demonstrate the ability of CMFGEN to reproduce, with a high level of accuracy, the UV and optical observations of a sample of well observed type II SN, i. e. SN1987A and SN1999em, at representative stages of their photospheric evolution. Two principal stages of SN are modeled - that where hydrogen is fully ionized, and that in which H is only partially ionized. For models with an effective temperature below similar to 8000 K, hydrogen recombines and gives rise to a steep ionization front. The effect of varying the location of the outer grid radius on the spectral energy distribution ( SED) is investigated. We find that going to 5-6 times the optically-thick base radius is optimal, since above that, the model becomes prohibitively large, while below this, significant differences appear because of the reduced line-blanketing ( which persists even far above the photosphere) and the truncation of line-formation regions. To constrain the metallicity and the reddening of SN, the UV spectral region of early-time spectra is essential. We find that the density of the photosphere and effect of line blanketing decline as the spatial scale of the SN increases. The density distribution is found to have a strong impact on the overall flux distribution as well as line profiles. For a given base density, the faster the density drops, the higher the effective temperature of the model. We also find in cool models that the set of Ca II lines, near 8500 angstrom is strongly sensitive to the density gradient. They show a weaker and narrower profile for steeper density distributions. Hydrogen Balmer lines are very well reproduced in fully or partially ionized models, but underestimated when hydrogen recombines. A reduced turbulent velocity or a flatter density layout are found to partially, but not fully, cure this persistent problem in studies of type II SN. He. lines observed in early-time spectra are very well reproduced, even for very modest helium enrichments, likely resulting from treatment of important non-LTE effects. At similar early epochs CMFGEN predicts, unambiguously, the presence of N II lines in the blue-wing of both H beta and He I5875 angstrom. These lines have been observed but so far have generally been associated with peculiar emission, from locations far above the photosphere, in the strong adjacent lines. Finally, we present a pedagogical investigation on P-Cygni profile formation in type II SN. Ha is found to form very close to the photosphere and thus presents a significant flux-deficit in the red, made greater by the rapidly declining density distribution. This provides a clear explanation for the noticeable blue-shift of P-Cygni profiles observed in early-time spectra of type II SN. Future studies based on CMFGEN modeling will focus on using type II SN for the calibration of distances in the Universe, as well as on detailed spectroscopic analyses for the determination of progenitor properties.