%gemini; ]> We propose to observe a small sample of weak-continuum , dwarf galaxies to investigate the excitation of molecular hydrogen in massive star-forming complexes. In the usable fraction of our previous allocation we were able to observe one of our targets, NGC5461. This dataset unambiguously shows that the gas is excited in low density photo-dissociation regions, contrary to the widespread assumption in the literature that the H2 in galaxies is predominantly shock excited. The weakness of the dwarf galaxy continua permits detection of the higher level H2 transitions which are essential to determine the gas excitation and relative contributions of thermal and UV-excited gas. Background H2 Observations of Galaxies A number of studies of molecular hydrogen and its excitation in galaxies have been made over the last decade (e.g. Joseph et al. 1984; Puxley, Hawarden & Mountain 1988, 1990; Moorwood & Oliva 1990; Kawara & Taniguchi 1993; van der Werf et al. 1993). Early indications from low spatial and spectral resolution H2 line ratios that the gas was always shock excited were challenged by parallel theoretical developments showing that in dense, warm photo-dissociation regions heated by UV-photons the low-lying (and brightest) transitions could be thermalised. Moreover, Puxley et al. (1990) argued that given the observed Br-gamma recombination line emission, the copious non-ionizing UV photons would be expected to produce substantial H2 emission when absorbed by molecular gas. Indeed, the different dependences of H and H2 emission on incident UV flux provided a diagnostic of the relative spatial distributions of exciting stars and gas. Despite these arguments, however, it remains widespread in the current literature that H2 in galaxies is shock excited (by supernova remnants and stellar outflows), although these studies are invariably without detection or reliable analysis of the critical higher level line fluxes. One of the principal difficulties in determining accurate fluxes for H2 lines above the lowest v=1-0 transitions is the strength of the underlying stellar continuum and the presence of metal absorption features in the cool stellar atmospheres. Even with very high S/N data, local variations in stellar population and element abundance produces systematic features after subtraction of empirical or theoretical stellar templates. This restricts the detectability in galaxies with strong continua to transitions brighter than those required to definitively determine the H2 excitation. The wider astrophysical context for studying H2 emission from galaxies is in understanding the interaction between massive stars and their environment (e.g. the relative importance of the excitation mechanisms, the effect of the formation of massive star clusters on the local ISM and its feedback into the star-formation process), the distribution of stars and interstellar material and the subsequent evolution of the starburst. Paradoxically, within our own galaxy it can be difficult to establish, say, the H and H2 emission and excitation mechanisms for a single star-forming complex. For example, although the peak H2 flux from shocks is 100 times that from the diffuse component in Orion, only by integrating over a 7 x 9 arcmin2 region is it revealed that the shocked and UV-excited luminosities are equal (Burton & Puxley 1990; Usuda et al. 1996). Observation of a galaxy provides such a census of the entire emission in a single measurement. During an IR spectroscopic study of several Wolf-Rayet galaxies, four lines of H2 were serendipitously detected in the nearby dwarf galaxy NGC5253 (Lumsden, Puxley & Doherty 1994) : 1-0 S(1), 1-0 S(0) and marginally 3-2 S(3) and 2-1 S(1). This observation represented the first suggestive evidence for the dominance of UV-photon excitation of H2 in at least some galaxies, now strongly supported by the spectrum of NGC5461 in M101 (Fig. 1). Proposed Observations The NGC5253 and NGC5461 observations raise several interesting questions: Is UV-excitation of H2 actually a general phenomenon in galaxies? What are the relative fractions of UV- and thermally-excited gas? What does the UV-excited component and hydrogen recombination line emission imply about the physical relationship of sources in the emitting region e.g. are they consistent with the models described by Puxley, Hawarden & Mountain (1988, 1990)? To address these questions we require measurements of the H2 line excitations sufficient to determine the contributions from UV irradiated low and high density gas, and from shocks. Hence we propose observations of a sample of weak continuum, dwarf galaxies. All of the galaxies have been detected in the 1-0 S(1) H2 transition (Doherty, Puxley, Lumsden & Doyon 1995). The weakness of the continua provides two benefits (i) a relatively large H2-to-continuum ration and (ii) weak underlying stellar absorption features. Dark Matter Absorption Lines PJPuxley ppuxley@gemini.edu 808-974-2501 808-935-9802 Gemini Observatory Northern Operations

670 A Ohoku Place

Hilo, HI 96720

USA jpurcell@gemini.edu 808-935-9235 808-974-2500 MattMountain mmountain@gemini.edu 808-974-2523 808-935-9650 Gemini Observatory Northern Operations SamIYam sam@roe.ac.uk 520-318-8246 Royal Observatory, Edinburgh NGC5253 22 V 13:37:05.12 -31:23:13.2 Haro2 24 V 10:29:22.67 54:39:30.8 Haro3 23.2 B 10:42:15.84 56:13:26.0 HZw40 05:53:04.93 03:23:06.5 GSC0726501274 13:36:58.93 -31:19:05.48 GSC0726501323 13:37:07.387 -31:18:45.0 GSC0381600860 10:29:02.335 54:44:17.34 GSC0381601019 10:28:44.474 54:40:29.5 GSC0382600306 10:42:05.969 56:09:32.98 GSC0382600254 10:42:26.261 56:17:49.45 GSC0012100147 5:53:27.173 3:23:58.45 GSC0012100938 5:53:25.073 3:20:15.14 Gemini Observatory

670 A'Ohoku Place

Hilo, HI 96720

USA phase1@gemini.edu 808-974-2500 808-935-9235 11.0 3.0 3.5 2.0 2.5 2.0 2.5 2.0 2.5 none Puxley, Doyon & Ward (1996). The spatial distribution of stellar CO absorption in M83, ApJ, 476, 120. Puxley (1997). "Multi-wavelength hydrogen recombination lines in the compact HII region K3-50a" , Star-formation with ISO, in press." Harrison, Puxley, Brand & Russel. "Molecular hydrogen and the ortho/para ration in NGC253", MNRAS, in press. Puxley & Skinner (1996). "Search for IR positronium emission from the great annihilator", CTIO/ESO conf. on the Galactic Centre, in press. Lurnsden & Puxley (1996). "Near IR spectroscopy of the ultracompact HII region G45.12". MNRAS, 281, 493. UKIRT 97A 2.0 100 Data shown in text, paper in preparation 12.0 200085 2000818 20001214 20001227 Gemini science retreat; SPIE meeting K-window spectra with NIRI are requested of four weak-continuum, blue compact dwarf galaxies (NGC5253, Haro2, Haro3, IIZw40). The giant HII region NGC5461 (in M101) previously observed provides a 'bridge' between studies of the starburst nuclei of galaxies and Orion-like complexes in our own Galaxy. The preferred NIRI configuration is 40l/mm grating and long camera which provides adequate spectral resolution and the wavelength coverage necessary to get accurate line ratios across the K window. This is equivalent to the configuration used for our NGC5461 observation (Fig. 1). (If necessary we can use the 150l/mm grating and observe a 4 grating settings to cover the ranges 2.02-2.1um, 2.09-2.17um, 2.16-2.24um and 2.23-2.31um. At this higher resolution, the line/continuum contrast is improved by a factor of ~5 which largely offsets the increased time required to observe at 4 settings. Taking account of calibration and other overheads, however, leads us to favour the lower resolution gratings as our preferred strategy. At least three lines from three vibrational bands (1-0, 2-1, 3-2) will be measured with a detection limit of 0.03 of the 1-0 S(1) line, plus individual lines from energy levels which can only be excited by UV radiation (e.g. 8-6 O(5)at ~40,000K). From such a spectrum we will determine the relationship between column density and energy level, derive the rotational and vibrational temperatures and ortho/para ratio. For radiative excitation we expect a high vibrational temperature (~5000K) compared with rotational temperature (~1000K) and an ortho/para ratio less than 3. Comparison of the column density/energy level plot for the different vibrational bands will reveal any thermalisation due to collisions in a high density gas. The expected continuum flux density in GNC5253 is 0.5 x 10-15 W/m2/um (K ~ 14.7) is a 0.6 x 2.4 arcsec2 aperture (1-pixel slit, 4 pixels binned along the slit) assuming approximately uniform surface brightness in scaling from our previous 3 x 3 arcsec2 measurement. The 1-0 S(1) line / continuum ration at a resolving power of l/dl = 850 is ~0.3 and we would aim to detect lines which are several percent of the continuum at 10 sigma. Thus the continuum S/N required is 250-300, requiring an integration time of 1-2 hours. Similar previous observations of the rest of the sample gives expected continuum flux densities of 0.5 - 2 x 10-15 W/m2/um (K ~13.3-14.8) and 1-0 S(1) line / continuum ratios of 0.5-1.0. The galaxies with weaker continua have larger line/continuum ratios and thus we expect similar integration times. The total request to observe the four targets, allowing 1 hour per target for calibration and acquisition, and 75% efficiancy on long integrations is 12 hours. Facility Gemini North Instrument Hokupa'a + QUIRC - visiting Adaptive Optics System Filter H (1.65 um) J (1.25 um) K (2.20 um) 2000QS 200061 2001131 2000331 unassigned 200011 100010001A 22 4.0 19991222 We think this proposal should be fully scheduled.