The evolution of the comoving luminosity density at ultraviolet wavelengths is, effectively, a measure of the history of star formation in the Universe. Similarly, the infrared luminosity density is an indicator of the existent stellar mass. The evolution of the comoving luminosity density can, therefore, be used to reconstruct the history of star formation in the Universe.
A measurement of the luminosity density can be obtained by integrating the galaxy luminosity function. Using the novel technique of photometric redshifts, the galaxy luminosity function is determined over the redshift range 0 < z < 4. Brightening is seen in both the luminosity function and the luminosity density out to z~3; this is followed by a decline in both at z>3.
Since dust shrouds star-forming regions, the amount of dust must be determined in order to accurately measure the amount of star formation at high redshifts. Broadband spectral energy distributions of z>2 galaxies are combined with spectral synthesis models to show that high-z galaxies (Lyman break objects) are shrouded in enough dust to attenuate their ultraviolet fluxes and, hence, star formation rates by a factor of more than 10. It is also found that these high-z objects are unlikely to be direct progenitors of present-day massive galaxies.
A color selection technique is used to identify candidates for ultra-luminous z>5 galaxies. If indeed at high-z, then these ``g-band drop-outs'' have extremely high star formation rates (greater than about 50--1500 Msun/yr) and produce comoving luminosity densities comparable to those observed at the present epoch. These objects may represent the formative stages of present-day massive galaxies.
The above luminosity density and attenuation measurements, together with other data, are used to construct a qualitatively robust picture in which star formation peaked at t~0.2t_o and declined more or less exponentially ever since. While qualitatively this picture is robust, different combinations of IMF, amount of high-z dust, and underlying cosmology result in scenarios of cosmic star formation which are quantitatively different from each other.
Young, massive stars profoundly affect the interstellar medium (ISM) in galaxies. Many of these effects are traceable in the radio continuum. The creation of thermally-emitting ionized gas, and the acceleration of synchrotron electrons in supernova remnants, are two of the most important processes in that their properties can be used to constrain the characteristics of the underlying stellar population.
The dwarf galaxy M82 is undergoing a burst of star formation so intense that its emission properties across the spectrum are dominated by the effects of new stars. Using modern instruments, these effects are resolvable down to parsec (pc) scales. Therefore, M82 has become the prototypical object for study of the starburst phenomenon.
A 5-band radio continuum study of the nucleus of M82 is presented. A method is discussed whereby the data are imaged with the best possible noise characteristics. These low noise, high resolution images at 20, 6, 3.6, 2 and 1.3 cm provide information on the population of bright, compact radio sources and the diffuse, highly nonuniform interstellar continuum emission.
A total of 73 compact sources are catalogued. Their radio emission has been modelled using the combined effects of non-thermal and thermal emission and thermal absorption. Full spectral models are fit to 26 sources, identifying them as either supernova remnants (SNR's) or compact H II regions. A further 21 sources have sufficiently detailed spectra to allow tentative identifications.
The SNR population is modelled as a population of sources slowly decaying in flux as a powerlaw with the passage of time using a continuity equation approach. The best model predicts a supernova rate of 0.090 per year.
The largest remnant is about 16 pc in diameter, and is the only resolved ring-like source in the data. The lack of large remnants coupled with the presence of a high pressure interstellar medium (ISM) could indicate that the SNR's are strongly confined by an ISM of moderate density, n = 10-100 particles per cubic cm. This medium can also account for much of the observed absorption seen at wavelengths longer than about 30 cm.
The wispy, diffuse interstellar radio emission has been imaged at about 1.0 arcsec resolution (17.6 pc) over five frequencies, and spectrally decomposed into images of thermal-only and non-thermal-only emission. The thermal emission is clearly organized into three regions: a central ring around the stellar nucleus, and two ionized lanes extending eastward and westward along the major axis. These regions are further organized into clumps. Each clump is comparable in size and energy output to the largest, locally known extragalactic H II regions.
A comparison of the spectrally decomposed thermal radio continuum image with images of molecular gas suggests that the ionized and molecular material is organized in an alternating arrangement, where dense clouds of cold gas are situated in between H II complexes.
Comparisons with images at optical and infrared wavelengths imply that these ionized regions are also the sites of emission from very small grains (VSG's) and other tracers of ionizing radiation, but that the extinction towards the near-IR is still quite high in localized regions.
The derived supernova rates are used in conjunction with other observations to constrain starburst models. The models cannot reproduce the properties of M82 using a solar neighbourhood stellar initial mass function (IMF). IMF's that are biased against the formation of low mass stars are required.
The integrated radio emission of the galaxy predicts a higher supernova rate than that predicted using either the ``ring source'' or the continuity equation method. Since the integrated radio emission records an historical rate from sources that have already been dispersed into the ISM, the implication is that the supernova rate is decreasing over time and therefore the starburst in M82 in coming to an end. This prediction is in agreement with other evidence from a consideration of star formation rates and the amount of available molecular material.
Graduate Student Page
Tables of Contents