Annu. Rev. Astron. Astrophys. 1997. 35: 309-355 Copyright © 1997 by Annual Reviews. All rights reserved

3.5. Hydrogen in Spectra of SNe Ia

The presence or absence of circumstellar material can shed light on the nature and evolution of SN Ia progenitors, as discussed by Branch et al (1995, and references therein). One way in which this gas can reveal itself is through transient, narrow H emission or absorption lines in early-time spectra of SNe Ia. Branch et al (1983), for instance, gave tentative evidence for a weak, narrow H emission line in a spectrum of SN 1981B obtained six days after maximum brightness, but Cumming et al (1996) showed that this interpretation is unlikely to be correct. Similarly, Polcaro & Viotti (1991) claimed to have detected H absorption in a spectrum of SN 1990M obtained four days after maximum brightness, but Della Valle et al (1996) argued that this was probably an artifact of the reduction procedure.

Calculations by Cumming et al (1996) indicate that circumstellar emission in H will drop rapidly after explosion; detection is not possible unless very early observations are made. Sensitive high-resolution spectroscopy is starting to set useful limits on H absorption or emission andin turn on the amount of circumstellar hydrogen around SN Ia progenitors. Cumming et al (1996) did not detect H in a spectrum of SN 1994D obtained at t = -10 days; under the assumption of spherical symmetry for the progenitor's wind, they find an upper limit of 2.5 × 10^-5 M year^-1 (Lundqvist & Cumming 1997) if the wind speed is 10 km s^-1. Unfortunately, this limit can exclude only the most extreme symbiotic systems as progenitors of SNe Ia. Later (at t 23 days), Ho & Filippenko (1995; see also Filippenko 1997a) used the Keck telescope to carry out a more sensitive search for H in SN 1994D, though they did not detect any absorption or emission features (equivalent width 2 upper limits of ~ 3 mÅ) within ±100 km s^-1 of the SN's systemic velocity. Early-time observations at this sensitivity should be able to reveal narrow H from nearby SNe Ia, if it is present.

Thus, to date there have been no convincing detections of narrow, transient H in early-time spectra of any SNe Ia, though the sample is still very small. (Also, no such helium lines have been reported, but few if any careful searches have been attempted.) Note, however, that weak hydrogen in spectra of SNe Ia, if ever detected, will not necessarily be of circumstellar origin. For example, at the time of explosion the surface of the white dwarf may contain some hydrogen, presumably donated by the secondary star. If so, it should be a broad feature, as it is in the spectra of classical novae (e.g. Williams et al 1994) but much more subtle. In progenitors consisting of main-sequence or subgiant donors (e.g. cataclysmic variables), the ejecta can strip and ablate gas from the secondary star, thereby contaminating the early-time spectrum with hydrogen (Applegate & Terman 1989, Wheeler 1992), but this has never actually been seen.

For certain progenitor models, H emission might be expected in the late-time spectra of SNe Ia. Chugai (1986b) predicted that most of the hydrogen-rich material stripped from a red-giant secondary during the explosion is trapped within the ejecta, subsequently expanding at relatively low speeds. Two-dimensional hydrodynamic calculations supported this hypothesis (Livne et al 1992). The hydrogen becomes visible only after the photosphere recedes substantially, and the expected line width is small: Full width at half maximum (FWHM) 2000 km s^-1. Such a feature may have been detected in SN 1991bg by Ruiz-Lapuente et al (1993; see also Turatto et al 1996, Garnavich & Challis 1997), but there are other possible interpretations if it is real (e.g. [Fe II]; Turatto et al 1996).