GMOS SV program 69

The halos of galaxies provide a fossil record of their early evolution, and can be used to probe how galaxies were assembled. Monolithic collapse models predict that the mean metallicity and metallicity dispersion will vary in a systematic manner with quantities such as total mass and specific angular momentum, which in turn defines morphology. On the other hand, hierarchal merger models, in which the early chemical evolution occurs in distinct dwarf proto-galaxies, predict that halo metallicities will be subject to stochastic effects (e.g. Cote et al. 2000, AJ, 533, 869). In this case the metallicity distribution function (MDF) of halos is expected to show large variations when comparing galaxies of similar mass and morphological type.

In order to measure the MDF of halos it is necessary to resolve individual stars, and this has heretofore only been possible for Local Group galaxies (but see also Davidge & Pritchet 1990, AJ, 100, 102). With only three spiral galaxies in the Local Group, it has not been possible to determine if there are systematic trends with host galaxy properties. However, with 8 metre telescopes studies of the resolved stellar content in halos can easily be expanded to other nearby groups, and thereby provide insight into the earliest phases of galaxy evolution.

As one of the nearest Sc galaxies outside of the Local Group, NGC 2403 is an obvious target for such an investigation. NGC 2403 is in many respects a twin of M33: these galaxies have similar integrated brightnesses, morphologies, gas phase metallicities (Garnett et al. 1997, ApJ, 489, 63), and disk star-forming histories (Davidge & Courteau 2001, AJ, submitted). M33 and NGC 2403 even have conveniently similar orientations. Do the halos of these galaxies have similar MDFs, as predicted by monolithic collapse models, or are they very different, as predicted by merger models?

I propose to observe halo fields in M33 and NGC 2403 in griz filters. These multi-color data will be used to construct CMDs of the upper RGB in each galaxy, and then infer mean metallicities and metallicity dispersions. The basic goal is to determine if the halos of M33 and NGC 2403 have similar mean metallicities and MDFs; an important aspect of this program is that both galaxies would be observed with the same instrumental set-up, which will remove systematic errors and provide an ironclad differential comparison. Fields with the same galactocentric distance will be observed in each galaxy, and the M33 images can be processed to simulate the increased distance of NGC 2403 to calibrate blending effects. Given that the goal is to go relatively deep, a background field would also be observed to sample contamination from faint foreground stars and background galaxies. This background field, which may be of scientific interest for other programs, would also be used to construct fringe frames. It is also possible that we could use deep images from other imaging programs for this purpose.

This project does not require a large chunk of observing time. The goal is to obtain 10-sigma photometry 1 mag below the RGB-tip in g, r, and i, which occurs at M_I = -4 (e.g. Davidge 1993, ApJ, 409, 190). For M33 this means obtaining 10-sigma photometry at I = 21.5, while in NGC 2403 the corresponding brightness in I = 24.5. For M33 this requires integration times of only minutes per filter, while for NGC 2403 the estimated integration times are on the order of 1 hour per filter. These integration times were derived from the GMOS ITC using 50%ile IQ, 50%ile sky brightness (grey time), and an airmass of 1.5. If need be the program could be restricted to two filters (r and i, say) to save time. Because of the lower qe of the GMOS detectors in z it is not possible to go as faint as in the other filters without a huge expenditure of exposure time. However, observations in the z filter are scientifically interesting, as this bandpass is least susceptible to line blanketing, which can affect the brightnesses of moderately metal-rich stars, and cause them to be missed in surveys of this nature. For the z filter we would obtain a 2700 sec exposure that would allow a 10-sigma detection of objects at the RGB-tip.