Imaging Survey suggestion for Megacam at CFHT starting ~ 2001 Semester II Wolf-Rayet Stars in Nearby Galaxies: Search for Pre-Supernovae A.F.J. Moffat and M.M. Shara Background It is currently believed that, for a solar ambient medium, all stars with initial masses above some 30 M(Sun) will evolve to the exotic Wolf-Rayet stage, before exploding as supernovae of type Ib. The compact remnant is likely a stellar-mass Black Hole. Occasionally, the death process might lead to a gamma-ray burst. For sub-solar metallicity, the lower mass limit required to produce a WR star increases, due to weakening of the stellar winds needed to expose the He-rich cores of massive stars, i.e. WR stars. Project A lot of arguments and assumptions go into this kind of scenario. To prove it (or otherwise), we need hard evidence - ideally, catching a star of a priori known type in the act of going SN. With the advent of Megacam at CFHT in 2 years, we propose to take advantage of its large collecting area (over a square degree), its fine resolution (tip-tilt) and location (CFHT with 0.5" seeing relatively frequent), to survey a respectable number of nearby galaxies for WR stars as precursors of SN Ib. To reach a tractable interval of time before the next WR Ib SN, say ~ 10 years, one needs a "bank" of some 10^4 WR stars, each with mean lifetime of ~ 10^5 years. Note that because of their strong emission lines and relative rarity, WR stars are relatively easy to detect, even in crowded regions. Of course, the WR stars detected will also enable us to track massive star formation and metallicity (from the frequency of different WN and WC subtypes) within different types of galaxies. Feasibility Is it feasible to find this many individual WR stars? Apart from hopelessly unresolved WR stars in distant galaxies, we only know some 600 WR stars in the LG and perhaps a few dozen in other randomly studied galaxies out to some 3 Mpc. On the other hand, our Galaxy is estimated to contain a total of between 1000 and 2000 WR stars, mostly blocked by strong IS extinction, that only future IR surveys will be able to locate. M33 contains some 200 spectroscopically confirmed WR stars, quite reasonable given its lower mass cf. the Galaxy. M31 has only some 50 known WR and deserves more attention; with its higher metallicity, M31 could contain at least some 10^3 WR stars. Assuming that a typical giant spiral galaxy like our own contains 1000 WR stars, we need to survey just ~ 10 such galaxies. We estimate very roughly that by going out to d = 10 Mpc, we should be able to find at least 10^4 WR stars in spiral/irr galaxies observable from CFHT. Many of these galaxies have had multiple supernovae over the past century (M83, d = 6.9 Mpc is a notable example with 5 recorded SNe). For the faintest WR stars, Mv = -4, we need to reach V = 26 for d = 10 Mpc. Especially the intrinsically fainter WR stars tend to have stronger emission lines, making an on-line - off-line search more effective and thus offsetting their faintness. The best detection strategy is to compare on-line and off-line (continuum) images of the same fields. With sufficient S/N, subtraction of the two images reveals the WR stars, even in moderately crowded areas. The trick is to go deep with good spatial resolution in a large field, something for which Megacam will be ideal. The galaxies range in size from the order of a degree for the closest, to 5-20' for the farthest. Once we find candidates, e.g. with strong HeII 4686 emission, we will have to obtain confirming and defining spectra. This will be done in multi-mode on 8m telescopes. We have in our possession already a narrow on-line (4686) filter, 7.75" square, used successfully to survey for Galactic WR stars using the UK Schmidt telescope + photographic plates. As "continuum" filter, we would use a combination of broadband B and V images (rather than narrow-band continuum, in order to save time); this is necessary to deal with colour effects. With bandpass FWHM ~ 40A, we should not be bothered significantly by non-zero redshifts out to 10 Mpc; RVs reach only ~ 500 km/s = 8A at 4686A. With some 20 fields to observe, we expect to accomplish this program in some 10 clear nights of 0.5" (i.e. median CFHT) tip-tilt corrected seeing. One of us will be delighted to serve on the working group.