The work presented in this thesis is dedicated to the study of the physical properties of photodissociation regions (PDRs), the surface layers of molecular clouds which are irradiated by ultraviolet radiation. The structure of PDRs is investigated with the development of an anlytical model which incorporates the essential heating and cooling mechanisms in a PDR. The main parameters in the model are the density and the incident ultraviolet radiation field, above the ambient value in the solar neighbourhood, impinging on the surface (Go) which dissociates the molecules in theaPDR. It is demonstrated that when the ratio (n/Go) is high (> 100 cm' ) the attenuation of ultraviolet photons is dominated by H2 self shielding which brings the HI/H2 transition zone close to the surface of the cloud (Av < 1). When the ratio is of order unity then the attenuation of ultraviolet photons is dominated by dust grains in the PDR. In this case, the HI/H2 transition zone occurs at a depth of Av ~ 2 - 3.
Images of the PDR in the northern bar of M17 show that there is a spatial coincidence, accurate to ~ 1 arcsec, of the H2 and 3.28 fim emission regions (the 3.28 |im emission being a tracer of the hot edge of the PDR delineated by the Hll/HlC 'ltransition) placing a lower limit to the density in the clumps of 10 cm . This coincidence is also observed in other PDR sources (eg. NGC 2023) and can be readily explained if the sources are clumpy. It is not clear in the northern bar of M l7, where Go ~ 10^, whether shielding by dust or H2 molecules is dominating the attenuation of ultraviolet photons. A uniform, high density PDR model is sufficient to reproduce the observed H2 line intensity, however the images clearly reveal structures at the 2 arcsec level suggesting that a clumpy model is a realistic solution.
Long slit K band spectroscopy measurements were taken in the northern bar of M l7, where up to 16 H2 lines were identified. Analysis of the data suggests that the emission can only be explained if the H2 molecules are being excited radiatively, rather than by shocks. The diagnostic line ratio of the H2 v=l-0 S(l) and 2-1 S(l) transitions is approximately 3 over the region observed implying that the lower levels of the molecule are slightly thermalised by the warm gas. The constancy of the ratio further implies that that the collisional deexcitation rate must be constant along this region. This may be taken to mean that the physical conditons do not change along the region and that we are observing the surface of the PDR. The profile of the molecular emission along the slit can be successfully modelled if it is assumed that the large scale surface geometry in the region can be described by a parabola.
The ortho to para ratio of H2, measured in three PDR sources (M17 northern bar, NGC 2023 and Hubble 12), is less than the expected value for a hot (~ 300 K) gas in thermodynamic equilibrium (ie. 3). The measured values fall in the range 1.3 - 2.3. Modelling the ortho to para ratio using the rates of spin conversion of the H2 molecule do not satisfy the observations. Some other fonn of processing the ortho to para ratio must be occuring and three different models are considered. First, a different reformation mechanism is considered which allows the newly formed molecule to reside on the surface of the grain for a certain amount of time before it evaporates into the gas. The second model describes a dynamic PDR. Hot gas at the surface of the
PDR is allowed to escape setting up an advancing photodissociation front. If the front advances in to the cooler gas before there is enough time for spin changing interactions to take place, then the measured ortho to para ratios will have values characteristic of the cooler (T ~ 50 - 150 K) gas. Both of these models can successfully reproduce the observed ortho to para ratios. The third model assumes that there is some processing of the ortho to para ratio during the fluorescent cascade, after it has been excited by ultraviolet photons. This model is not successful at reproducing the observed ortho to para ratios as the cross sections in the excited states are so large compared to the radiative decay rate that the ortho to para ratio is determined at the gas temperature, giving ratios ~ 3.
Column densities calculated from the emission from NGC 2023 has shown an excess of emission above the pure fluorescent cascade in levels which are highly rotationally (J > 6) and vibrationally (v = 3 - 6) excited. In according with theoretical predictions, this may be the first direct evidence of the formation of H2 occuring in excited levels.