THE PHYSICAL PARAMETERS THAT CONTROL VIOLENT STELLAR BURSTS
I.P.: F. Garzón
CoIs: grupo GEFE del IAC.
Version 2.1; 1-July-1993
It is our aim to use innovative techniques, made possible by the use of ISO instrumentation, to investigate the parameters that control the formation of stars in stellar systems. To this end, estimates of the stellar mass, in several regions of recent star formation, as well as a variety of measurements revealing the physical properties (density, metal content, gas to dust ratio, temperature, mass, etc) of the star forming clouds are envisaged.
The project demands a well coordinated investigation of regions of recent star formation in a variety of galaxies, in order to determine the various factors that control the strength of a star formation burst. Thus, the aim is to infer the total mass in newly formed stars (M*) as well as the properties of the star forming cloud, to determine in this way a definite correlation. With this aim in mind, we propose to investigate in detail the far IR properties of star forming regions.
The star formation efficiency (e = M*/M_total = M*/(M*+M_gas)) together with the Initial Mass Function (IMF) and the star formation rate (SFR) are the three key ingredients that lead towards the understanding of star formation in stellar systems. The last two issues have been largely investigated both observationally and theoretically (see for example the reviews by Kennicutt 1989, and by Silk 1985). The star formation efficiency however, seems to have been left behind, in need of well defined observational strategies, perhaps motivated by the lack of a refined theory that could reveal the physical process(es) that controls, or stops, the formation of stars in a given cloud and defines the maximum and final value of M*. It was Hoyle (1953) who first pointed out the likely inhibition of the gravitational collapse in a cloud upon the sudden release of energy from recently formed massive stars. In the same vein, Cox (1983), Tenorio-Tagle et al. (1986), Larson (1987), and Franco and Tenorio-Tagle (1991), have shown that photoionization is the likely primary mechanism that limits the formation of stars in a cloud, fixing at the same time the efficiency of the process. Furthermore, given a violent burst of star formation the size and sound speed of the resultant HII region imply that the physical conditions that led to the stellar burst may only be found within the HII region volume.
Without doubt, one of the most important open problems in Astrophysics is the lack of knowledge about the detailed physics of star and cluster formation. Nevertheless, it has been possible to infer some of the global properties of the star formation process, such as the SFR, the IMF and the efficiency of stellar formation, for a well defined entity such as a star forming cloud or a galaxy as a whole. Information on the three properties is however not complete for any given system. From the observational viewpoint, there are several major dificulties resulting from: first, the inability of a direct detection of the IMF in all systems. Second, because of the confusion between local and global efficiencies, the latter of which demands a well defined formation time interval, or duration of burst, and/or finally because of the difficulties in defining both the extent of the star forming region and how coeval a burst may be across the full mass range of resultant stars. Further problems araise from contradicting interpretations such as the tight correlation between the radio emission of all types of galaxies with their infrared brightness over more than four decades (De Jong et al. 1985). The so called F(6cm)/F(60 µm) relation reflects an intimate connection between the rate of stellar formation and non-thermal synchrotron radiation, and thus pronounces the balance between formation and death of massive stars. From the slope of the relation in nearby galaxies however, Beck and Golla (1988) deduced that the non-thermal radiation comes from the diffused interstellar medium and further that the above mentioned relation results from a consorted action of star formation and interstellar magnetic fields. It has been proposed (Völk, 1989) that that relation should break down at small scales. A determination of the critical spatial scale where the break-down occurs would be very interesting.
The programme outlined in this proposal addresses the important topic of the study of the parameters that control the formation of stars in extragalactic HII regions as well as in external galaxies. This study aims at the physical properties of the ionized and neutral gas around young stellar bursts, where the properties of the clouds are still representative of the conditions previous to star formation. ISO is the ideal observatory for the project, as young HII regions are generally associated with dense molecular dark clouds, which are largely opaque to UV or optical observations. Far IR photometry and spectroscopy, in the 50 to 150 mu region is ideally suited for the study of young HII regions and their parental dark clouds, as the maximun of their emission is produced in this spectral range. This range is unaccesible from ground based optical or IR telescopes. As for the Kuipper Airborne Observatory (KAO), that also operates in this wavelength domain, only a couple of very bright sources and only a couple of very prominent fine structure lines from them could reasonably be observed, thus rendering the KAO definitely not suited for this project.
As an example, the number of 19 HII regions found in M33 with IRAS can be enlarged with ISOPHOT by a factor of 3, due to its better angular resolution. Also using PHT-P with an apperture of 52" (equivalent to say 300pc in M33, and about 2.5kpc in M101) for the wavelengths 25 µm, 60µm and 100 µm the integration time needed to achieve a S/N of 5 for the faintest object is about 20 sec with each filter. While some of the bright regions may only requiere half the integration time (8 sec), some fainter ones may need twice as much.
Each object will be mapped with ISOPHOT-C at wavelengths sensitive to the presence of dust, like 60, 90, 130 and 160 µm. This will provide information about the mass of dust, extinction and the distribution of ionizing sources.
Optical spectra at the peaks of the emission maps will be taken in order to determine and model the relative contribution of young stars in heating the dust. This will allow an estimate as accurate as posible of the number of ionizing sources.
For bright objects, IR emission lines from the ionized gas will be observed at selected areas within the star forming regions. Some interesting features required for obtaining physical properties of the ionized gas are given in the table below:
| TABLE 1 | |
| Infrared Lines | |
| OIII | 51.69 µm |
| OIII | 88.1 µm |
| NIII | 57.29 µm |
The emission lines in Table 1 together with optical line measurements will allow the determination of temperature, density and chemical abundances for the regions. For instance, a useful diagnostic diagramme for the determination of density is [OIII]5007Å/[OIII]52µm vs. [OIII]52µm/[OIII]88µm. Likewise the ratio [NIII]57µm/[OIII]52µm gives values for the abundance ratio of these elements, that ratio being independent of other physical parameters. Comparisons of these IR line diagnostics with pure optical line diagnostics will also reveal whether the optically determined parameters are affected by dust extinction in the line of sight. Moreover the luminosity of these fine structure lines imposes constraints in the stellar temperature of the ionizing sources allowing a more accurate sampling of the IMF.
The apertures for the spectra to be obtained will be centred on the maximum of Ha emission, taken from images from the ground. As for the photometry, we will use standard techniques, centred on the emission peaks, as well as surrounding areas.
The massive star formation process, is widely believed to depend on environmental conditions. We start, however, from the premise that the strength of a burst must be determined by the physical properties (density, metal content, etc.) of the star forming cloud and thus, the number of resultant stars should not depend on the formation history of the cloud. Therefore within the sample of selected regions of recent star formation, one could have members belonging to spiral arms, irregular galaxies, as well as within interacting galaxies, dwarfs and/or HII galaxies. The sample of selected objects should therefore cover a wide range of luminosity (and mass), going from small galactic ``Orions'' to extragalactic regions where violent stellar formation is taking place. This point is taken into consideration when selecting the sample.
There are several ways to infer stellar masses: for instance, the measurement of the number of ionizations in the resultant HII region, when related to the number of required stars and to a given initial mass function, leads to an estimate of M*. Two other methods can be used, under the assumption of the Virial theorem, to infer M* together with the mass enclosed within a certain volume.These involve accurate measurements of the gas velocity dispersion (s_gas) or , although technically more difficult, direct measurements of the stellar velocity dispersion (s*); perhaps by means of the CaII ``triplet'' absorption line. The inferred mass minus M_HII within the observed volume will yield an estimate of M*.
All possible methods of inferring M* should be used and compared in every object of the sample. In this way, appart from the star formation efficiency, one should be able to test wether s* = s_gas in such systems. This is an unresolved issue of greatest importance given the correlations found between sigg and size of the emitting regions, as well as with their luminosity, and metal content (Terlevich and Melnick 1981, Gallagher and Hunter 1983; and Melnick et al. 1987), which taken at face value, could be used as a powerful distance indicator. Naturally, if s* = s_gas, one would be able to infer M* pin a larger sample by means of the gas velocity dispersion which implies an easier task. If s* differs from s_gas however, one could still see if their discrepancy correlates with one (or a combination) of the observables.
The project thus proposes an operational strategy to derive the local efficiency of stellar formation, avoiding the usual problems of having to determine the total gaseous mass, or size of the star forming region, burdened by the fact that star formation often occurs only on one side of the cloud and that usually the various gaseous components show different morphologies. Such difficulties become acute when insufficient observing time inhibits the exploration of various other sources in a systematic manner, and lead to the present incomplete, scattered and fragmented knowledge in the subject.
A coordinated investigation, with sufficient telescope time, will permit a complete study. That will allow for a detailed comparison of the observed objects leading to the relative star formation efficiency of the sample. This could then be correlated with the various observables. Knowledge of eps and its controlling parameters will naturally have a direct impact on many aspects of star formation, particularly on the history of star formation and chemical evolution of galaxies, and on the early stages of galaxy formation. Clearly, the deduced eps and the measurements here proposed will also provide detailed information on the IMF and the stellar formation rate in the sample of selected objects.
Further constraints in the selection of the sample arise from the likelihood of accurate determination of the stellar mass as well as of the parameters that describe the physical properties of the star forming cloud. The latter issue should take into account the energy released by the stellar burst, which rapidly modifies the cloud parameters, setting them to values in the affected region, which are not representative of the conditions prior to the star forming process. This implies therefore the inclusion in the sample of very recent bursts (eg regions without the signature of WR stars or supernovae) and/or of regions which having a clumpy structure, clearly display an area still unaffected by the energy deposited by the massive stars. There, accurate determinations of density, metal content, temperature, gas to dust ratio, etc. are imperative to determine the original conditions in the star forming cloud. The method differs from earlier studies centered only on the properties of the resultant HII region, out of which the stellar formation rate could be derived, without the possibility however of establishing a causal trend.
The selected regions together with some information taken from the literature are given in Table 2 anexed at the end of this document. The Table contains both extragalactic HII regions and external galaxies. This allows for objects with various degrees of star formation, with a range of liminosities and of metallicities.