Pre-main sequence and protostellar evolution observed with ISO

One of the main objectives of the ISO-LWS pre Main Sequence programme was to study the earliest stages of stellar formation and evolution using 3 - 200 mm spectrophotometric observations of a large sample of Class 0, I and II objects, and related Herbig-Haro (HH) objects. This work will concentrates on the interpretaion of already existing data, and the development and refinement of diagnostic tools to study the environment of the nearby gas (i.e. photodissociation regions, ionisation state, shocks, and the global relationship of the IR source properties to star formation), as well as related ground-based follow-up observations.

The mid- and far-IR emission from these objects consists of a continuum, due to greybody emission from heated dust grains around the protostar, and line emission from either ionised gas near the newly forming star, shocked gas driven excited by an outflow, or ambient gas close to the protostar.  The far-IR line emission originates mainly from two physical mechanisms: excitation in photoionised (or photon-dominated regions - PDRs) regions, or: shock excitation resulting from the interaction of supersonic winds with the ambient medium. Depending on the wind velocity, magnetic field and ion density, two kinds of shocks with different FIR spectra are expected: high velocity J-shocks, in which temperature, density and velocity have a discontinuous jump will dissociate the molecules, leading to a spectrum dominated by atomic line emission; and C-shocks, where the Alfven velocity is > vshock and the magnetic field transmits energy at a speed > vshock; in this case the temperature, density and velocity have a continuous variation, and molecules play an important role in the cooling processes. The key lines to discriminate between these processes can be readily observed within the ISO LWS and spectral ranges.

The early stages of protostellar evolution are characterised by a period where material in the central core and placental material is expelled in a supersonic outflow. Non-dissociative shocks are expected to be prodigious sources of far-infrared molecules and water emission, in the latter case because H2O forms very efficiently at temperatures > 400 K - which is typical of post-shock gas that we have detected with ISO. In contrast to the predictions of early models, the early ISO results suggest that water plays somewhat less of an important role in cloud cooling than had been thought, and the CO lines, more so.  The observed discrepancy indicates a lack of water production in C-type shocks, which may be explained if the oxygen is more locked onto grain or in the molecular form, than in the atomic form.

ISO observations of high-J CO lines towards preMS objects with LVG model fits. These measurements show unambiguously that strong shocks are present - not previously recognisable from ground-based mm/submm spectra

Finally, our sample covers a very wide range of sources and protostellar classes, the global characteristics of the star formation mechanism can be examined, by comparing the continuum and spectra line characteristics. The long-term objective of the preMS programme is to show how the characteristics of the individual sources which have been examined so far, can be used to examine preMS formation, physics and evolutionary processes. This is likely to be a key result of the SED's analysis of the sources in the preMS sample - some initial results from this kind of analysis are shown in the next figure, where ISO derived propreties are fitted to evolutionary tracks.


The ISO sample bolometric luminosity vs 1300 mm flux, with the ‘outflow strip’ indicated - which shows the boundaries within which molecular outflows are seen, b) computed tracks for various PMS objects, and different mass loss rates c) From pressure and mass estimates, we infer that the star-formation efficieny in the Rho Oph cloud is 4%, significantly lower than previous studies have suggested - and a consequence of our better ability to determine cloud masses, using radiative transfer models constrained by the new far-IR and submm data becoming available.

a) SED Data analysis

The ISO mission is now ended, and the data is starting to enter the public domain at the time of writing. The main objective is now to work on the analysis and modelling of our ISO Guaranteed and Open Time preMS database. This work will concentrate on modelling the environment close to a sample of 62 preMS sources. The tasks are a) to model the SED's of the central point sources; b) the reduction and analysis of mapping data of fine structure and molecular gas distributions - with emphasis on modelling these with photoionisation and shock models; c) the use the data in this uniquely large sample to study the global relationship of the IR source properties to star formation (see above figure). It is likely that c) may include some ground-based observational follow up.

b) Chemical Modelling

The ISO observations have provided an unique view of the chemistry in the ISM; leading to the first detections of mid-IR lines of water, CH+ and CO+ in the ISM which dominate the cooling of the ISM, and a large number of as yet unidentified lines are present in a lot of spectra [38]; detection of high J-transitions of common molecules such as CO, OH, H2O, OI in shocked gas; and given new insight into cooling processes in the ISM. The results on gas-phase molecules can be combined with information on solid state features available from other ISO SWS programs to obtain insight into the physical processes affecting the gas/solid state ratios [82]. We have developed two codes of use in this work i) the use of the photoionisation clode CLOUDY - but with our own enhancements including updates on cross sections, improved physics, and the introduction of clumping, and ii) a 3D radiative transfer code for the solution of the frequency dependent radiative transfer equation (Men'shchikov et al A&A, 318, 879, 1997, White et al L1551) [54]. We have already invested considerable time and effort in developing this range of diagnostic software, and we are now use this software and extending it by incorporating additional molecular processes to improve the treatment of cooling processes, ultraviolet radiation and shocks due to bipolar outflows - which are currently treated at a basic level.


a) LWS spectrum of the PMS object  IC 1396   b) ISO SWS + LWS spectrum + simple model fit    c) full 3D model fits

The aim of these models has not only been to examine the ISO LWS and SWS data, but also to link them to ground based mm and submm data, and to motivate and plan follow-up observations. The model fit c) above successfully matches observations extending from optical to radio wavelenghts, for any chosen geometry, and for a wide range of observing beamsizes. The power of this model is it's ability to handle everything from large beam infrared data to submm interferometric data simultaneously within the model, in an entirely self-consistent way. By using all of the available data, it is possible to strongly constrain the more simple, traditional model fitting such as shown in b) which uses spherical greybodies with only an elementary radiative transfer treatment.

We have developed various diagnostic tools, including a) LVG codes capable of modelling simple molecules such as H2O, CO, OH, for idealised source structures and geometries: this has been very useful in the order of magnitude estimation of source properties, and a Monte Carlo radiative transfer code, which is capable of handling a more complex clumpy source geometry. This is ideally suited to deal with the global radiative transfer problem in sources with both systematic and microturbulent velocity fields. Our intention is to integrate this with clumpy geometries, and to provide a more realistic shock treatment giving self-consistent model which includes both heating and cooling processes for the ISO data. We will use the new codes to solve a set of time dependent chemical equations and simultaneously to solve the UV radiative transfer problem for H2 and CO, extending this work to clumpy media. At present the code is applicable to a clumpy medium only in a rather primitive way, and models and strategies for handling inhomogeneities in a more realistic manner will have to be developed as part of the work..