One of the major goals of ground, airborne, and space-based
observational astronomy is the identification and study of
protostellar objects. A study of protostars will provide insight into
the initial conditions required for star formation, protostellar
evolution, and the formation of solar systems like our own. An object
can be identified as a candidate protostar from continuum
observations in the far-infrared and submillimeter. However, to
unambiguously identify an object as a protostar requires the direct
observation of infalling gas. This in turn requires spectroscopic
observations. Using millimeter and submillimeter transitions of
density-sensitive molecules such as CS, H 
CO, HCO+ and NH 
,
Walker et al. (1986, 1993, 1994), Zhou et al. (1993), and Narayanan,
Walker and Buckley (1997) have made tentative detections of infalling
gas toward protostellar objects. However, these studies were most
sensitive to cold gas in the extended envelopes of these objects. With
the SOFIA instrument proposed here, it will be possible to
unequivocally detect infalling gas in the FIR and probe physical
conditions much closer to the protostar.
Recently, Ceccarelli, Hollenbach, and Tielens (1995, hereafter CHT)
have modeled the FIR spectrum of gas freely-falling onto a protostar
in a self-consistent manner. With decreasing distance from the
protostar, the gas is mainly heated by collisions with warmer dust,
then compression, and finally through absorption of IR photons by
H O and CO. The gas cools by radiating in IR lines of [OI]\
(principally at 63 
m) and in rotational transitions of CO (in the
outer regions) and H O (in the inner regions). A great deal of
useful information can be derived from observations of mid-J CO
lines. Although the transitions accessible to the heterodyne
array - 
to 
- do not lie at
the peak of the distribution of CO line fluxes as a function of J
( 
usually lying between J=6 and J=12), they have the
advantage of being very insensitive to the time since the start of the
collapse. Furthermore, the transitions at the peak of the distribution
vary only weakly with mass accretion rate, while the higher-J
transitions which can be observed with the array (i.e., those
with 
) are much more sensitive to 
, and are
therefore more useful for estimating the mass accretion rate.
Most importantly, CHT find that the most definitive detection of
infall would come from high spectral resolution observations of
mid-J transitions of CO. The higher-lying transitions are
preferentially excited in the warmer, higher-velocity gas closer to
the protostar, resulting in a correlation between linewidth and
J. By observing several FIR CO transitions, it will be possible to
determine the velocity field - and therefore both demonstrate the
presence of infall and measure the mass of the protostar - and
estimate the excitation conditions in the infalling gas. The
sensitive, broadband, high resolution spectrometer to be used here is
ideally suited for this task. While most of the infall region
will lie within 1 beam, the rest of the array can be
used to probe conditions in the associated molecular outflow. Being
able to observe the infall region, extended gas envelope and the
outflow simultaneously reduces pointing uncertainties - which is
crucially important for studying the line profiles - and makes the
disentangling of outflow, rotation, and infall velocity fields much
more tractable. In addition to using the CO lines, the proposed
instrument can also probe conditions in the outer infall region by
observing the [OI] line at 143 m. CHT also predict that strong OH
rotational emission would arise in the accretion shock. The proposed
instrument will be able to probe this emission in the 163.4 and
163.1 m lines of OH.
Finally, we note that the high spectral resolution of the array may
make it possible to probe the internal dynamics of protostellar
disks. Walker, Maloney, & Serabyn (1994) detected luminous emission
in the v=1, 
and 
lines of CS
towards the young binary or protobinary system IRAS 16293 - 2422 in
Ophiuchus. The most plausible explanation of this
vibrationally-excited emission appears to be an internal shock
propagating through the disk, possibly due to self-gravity or to the
effect of the companion object. The resolution of the array
spectrometer in its currently proposed configuration is approximately
0.15 km s 
; however, an increase in the resolution by an order of
magnitude, to the 
m s level (at the cost of reduced
bandwidth) is easily achievable.
This raises the possibility of probing the dynamics of protostellar
disks by looking for systematic velocity shifts of spectral lines with
time, due, e.g., to the propagation of internal shock waves or the
gravitational effects of binary companions or protoplanets. For this
purpose the array nature of the spectrometer is crucial, as precise
determination of the source pointing will be absolutely necessary to
eliminate apparent velocity shifts caused simply by variations in the
location of the source with respect to the beam.