This project focusses on low latitude studies of Galactic 21-cm line and continuum emission from the Milky Way. The three observations described here focus on the distribution and motions of the interstellar HI in the Milky Way disk. The GBT is critical for all these projects because of its low stray radiation response due to the unblocked aperture. The first is a mapping project in support of the interferometer surveys of the Galactic Plane which are currently underway. The second is to study the scale height and random velocity distribution of the HI in the inner Galaxy, and the third probes the outer galaxy HI distribution, to test whether the Milky Way disk has an outer edge.

The curse of stray radiation has haunted many fields of astronomy, but none more severely than Galactic 21-cm spectroscopy. The leakage of bright HI line emission through the back sidelobes of the beam limits accurate baseline determination for 21-cm emission spectra at velocities within about 250 km/s of the LSR. For most telescopes the limiting sensitivity due to uncertainty in the baselines caused by stray radiation is between 0.5 and 2 K. For a few telescopes, the beam is known very well in all directions, which allows correction for stray radiation with precision of a few tenths of a Kelvin in brightness temperature. Examples are the Bell Labs horn (Stark et al. 1992), the Dwingeloo 25m (Hartman et al. 1995), the NRAO 140 foot (Murphy et al. 1996), the Penticton 26m (Higgs and Tapping), and, in the past, the Effelsburg 100m (Kalberla et al. 1980). Most of these have beam sizes of 30' or larger, so they can barely resolve the half-thickness of the HI disk (h = 150 pc) at the Galactic center distance (Ro = 8.5 kpc).

The community of Galactic 21-cm astronomers has been waiting for the Green Bank Telescope in order to map the Galactic HI emission with resolution sufficient to map the thickness of the disk and with baselines trustworthy at the level of a few tens of milli-K in brightness temperature as required to study details of Galactic HI spectra. Now at last it may be possible to achieve baseline precision of a few tens of mK or better in low latitude HI spectra. This project is a first attempt to push well beyond the historical stray radiation limit and answer questions about the Galactic interstellar medium which have been "on hold" since the mid-1970's.

The most urgent need for GBT data on Galactic HI emission is to supplement the 21-cm interferometer surveys of the plane which are currently underway. These are the Canadian Galactic Plane Survey (CGPS) covering the second quadrant, the Southern Galactic Plane Survey (SGPS) covering the fourth quadrant, and the VLA Galactic Plane Survey (VGPS) covering a large portion of the first quadrant. All these surveys have roughly similar specifications : to generate maps of the HI line and continuum emission from the galactic plane with resolution 60" to 100" and 0.8 to 1.5 km/s, and brightness temperature sensitivity 1 to 2 K rms. The first two are being done with east-west interferometers at the DRAO (Penticton) and the ATCA (Narrabri), respectively. Because they are east-west arrays, these telescopes are unable to survey the equatorial potion of the Galactic plane in the longitude range 10 to 65 degrees (where abs(delta) < 25 deg). We have used the VLA D array to cover the intervening region from 18 to 67 degrees longitude, but the D array cannot give as complete a coverage of short spacings (due to shadowing problems) as the east-west interferometers can at high declinations. So the short spacing problem in the first quadrant is acute. For all three surveys we need to fill in the short spacings with single dish data (Parkes for the SGPS and the DRAO 26m for the CGPS), but for the VGPS we need the largest possible single dish so as to get the best uv plane overlap with the limited short spacing coverage of the D array. So the GBT is needed for the VGPS longitude range 18 to 67 degrees.

This proposal is for commissioning time on the GBT; it is designed to test and confirm the absence of stray radiation problems in 21-cm spectra, and at the same time to answer some important scientific questions about Galactic structure. The project is in three pieces, all of which focus on low latitude 21-cm line mapping. They address three different science goals, but they also make a progression of more and more stringent stray radiation tests through progressively deeper integrations.

1. The first stage of the project is simply to map two standard areas which have been studied by several telescopes with stray radiation removal algorithms. These areas will be mapped quickly, several times, and the results compared with the data from the other surveys. With on-the-fly mapping, driving at one degree per minute, we can cover each area in about 3 hours, with sensitivity of about 200 mK (over 1 km/s) and full sampling (row spacing 3.5'). We will repeat these maps three times spaced at maximum hour angle intervals, in order to vary the part of the galactic plane which is above the horizon, so as to get different stray radiation signatures, if there is any stray radiation at all. We will then repeat the maps one time six months later, again to see if there is any variation in the baseline which can be attributed to stray radiation. Each pass over the area will give us spectra with rms of about 200 mK spaced at 3.5'. The sum of the four passes will have rms of 100 mK, which is fine for combination with the aperture synthesis data. Altogether this stage requires about 20 hours of telescope time, plus calibration and setup time, or about 25 hours total.

2. The second stage of the project makes deeper integrations on four short, (4 degree long) lines of pointings across the Galactic plane (constant longitude) in the inner Galaxy to study the profile shape at the terminal velocity as a function of latitude. For these we will integrate 15 minutes per point, spaced at 10' (i.e. somewhat undersampled). These integrations will reach about 12 mK rms in 1 km/s channels. The objective here is to study both the random velocity distribution of the HI [in the azimuthal direction, i.e. sigma(v-theta)] and the variation of the HI density with height above the plane, n(z). This stage requires about 25 hours of telescope time, plus calibration.

3. The third stage is the most stringent test of baseline stability. Here we integrate for one hour per point, down to about 4 mK rms in 2 km/s channels, in the outer galaxy, to study the radial variation of the HI density outside the solar circle. For this we space the pointings by 30', taking three short rows of pointings across the plane four degrees long centered on the (warped) midplane at longitudes 90, 135, and 225 degrees. We will study the profile shape at the extreme negative velocities (positive at longitude 225), which correspond to the outer edge of the Galactic disk. Each row requires nine hours of integration; we will repeat the maps after a delay of six months to check the baseline stability at the mK level. This stage requires two sessions of 27 hours, plus calibration, for a total of 60 hours.

To cover the entire VGPS area of 18 to 67 degrees longitude, -1.5 to +1.5 latitude (roughly) would take about 60 hours of telescope time for the GBT, assuming on-the-fly mapping driving at 1 degree per minute with tracks spaced at 3.5'. We are not requesting to do this yet; we hope to propose that project after we have done the preliminary mapping described here. With these small area maps we will be able to determine the best observing parameters for the larger survey.

As a test of the techniques for combining GBT and VLA data for the 21-cm line and continuum, in this project we will map two areas of overlap in the VGPS where several single dish telescopes have mapped the same region. The first region we have chosen is the overlap area between the Canadian Galactic Plane Survey (CGPS) and the VLA Galactic Plane Survey (VGPS) at l=65 to 67, b=-2 to +2. Here we have 21-cm maps from the DRAO interferometer, the DRAO 26m (stray radiation corrected, see Higgs and Tapping, 2000), and the VLA D array. We badly need a single dish map of this region (with resolution finer than the 30' of the Penticton and Dwingeloo single dishes) in order to fill in the short spacing information for the interferometer surveys. At low longitudes we have VLA, Parkes multibeam (McClure-Griffiths et al. 2000) and stray radiation corrected 140-foot maps (Murphy et al. 1996) for a field at longitude 27 to 29 degrees, latitude -1.5 to +1.5 degrees. The comparison among data from the GBT, the 140-foot, and Parkes will be a good test of the reliability of weak spectral features near the level of baseline uncertainty.

In both areas, we will combine the single dish and interferometer data (typically we use the Miriad task IMMERG), and also compare the single dish data from the different telescopes. We expect that the GBT data will be vastly superior to the maps from the other single dishes, both because of its better resolution and much better baselines, but this needs to be confirmed. The comparison will help us to determine at what level we can trust the maps from the other telescopes; this is particularly important for the Parkes multibeam, which is the only large single dish available for most of the SGPS area, which is south of declination -35 degrees.

Figure 1. The terminal velocity drop-off or HI emission spectra in the inner galaxy. This figure is taken from the SGPS, in the longitude range 326 to 333 deg, at b=0. The contours are drawn at 5 K, 25 K, and 45 K; the grey scale goes from -3 K (white) to about 50 K (black).
 

Another useful application of accurate HI profiles near the terminal velocity in the inner Galaxy is to study the variation in the brightness temperature with latitude. At the sub-central point, the variation of density with z, n(z), translates directly to the variation in brightness temperature, T(b,vt) where vt is the terminal velocity. Thus (assuming cylindrical rotation) we can directly map the z distribution of the density of HI, in the same area where we determine the random velocities of the HI. Since we have independent measures of the shape of the gravitational potential in the inner galaxy from optical and infra-red surveys, we can thus estimate the dynamic contribution of the magnetic fields and cosmic rays (reviewed by Ferriere, 1998). This sort of analysis is one of the scientific drivers for the interferometer surveys of the disk, but the GBT offers the unique advantage of reliable baselines with at least an order of magnitude better precision than any other telescope. Thus we can get a much more accurate trace of the z dependence of the tail of the random velocity distribution.

For the z dependence of the terminal velocity shape we will do scans from -2 to +2 deg in b at longitudes 11.5, 23.6, 36.9, and 53.1. (These longitudes give equally separated tangent point radii, of 0.2 Ro, 0.4Ro, 0.6 Ro and 0.8 Ro.) The integration time of 15 minutes per spectrum goes much deeper than in the two dimensional mapping, but less than the very deep integrations of part three. Spectral rms should be 12 to 15 mK for 1 km/s channels. This should allow us to trace the HI density beyond the terminal velocity drop off down to 50 mK or so, corresponding to column densities of 10**17 cm**(-2) (per channel).

Many spiral galaxies have very extended disks of HI, reaching to two or three times their Holmberg radii. Typically, the HI surface density in the outer regions decreases exponentially with radius, with a similar scale length to that of the light (a nice figure showing this is in Kenney and Young, 1989). Whether or not this is true for the Milky Way is a difficult question, particularly as the flaring of the disk (rapid increase in the scale height with radius) in the outer regions makes it difficult to measure all the HI. Some spiral galaxies show in addition a fairly sharp cutoff in their HI disks at the point where the gas surface density reaches about 2 x 10**19 cm**(-2) (e.g. Corbelli and Salpeter 1993). Whether or not the Milky Way shows this is unknown. The deep 140-foot survey of Burton and te Lintel-Hekkert (1986) shows that the HI is traceble to about 26 kpc Rgal (see Burton 1988). Beyond that the brightness temperature is too faint to measure reliably in the face of uncertain baselines due to stray radiation.

Figure 2. Models of the spectral dependence of the HI emission to be expected from the outer galaxy, for several simple density vs. radius laws. Note that the y axis is logarithmic. We hope to get below 10 mK (smoothed) in the long integrations of part 3 of this project.

One of the most ambitious attempts to determine the outer boundary of the Galactic HI disk was that of Knapp et al. (1978). They attempted several deep integrations with the 140-foot at longitudes 90, 100, and 225. The idea is that, at large galactic radii, the radial velocity approaches just the projection of the LSR velocity onto the line of sight, independent of the rotation curve, since the velocity of the gas is nearly perpendicular to the line of sight. Thus if the disk were very extended (say, to 150 kpc) there should be a pile-up of gas at about -220 km/s (viewed at longitude 90) or +220 km/s (at longitude 270). No such pile-up is seen, the profiles of Knapp et al. drop below about 0.1 K at velocities near -150 (at l=90) and +100 (at l=225). Even with a rotation curve which drops fairly rapidly at large radii, it is impossible to reconcile these results with an LSR velocity of 250 km/s, which contributed to the IAU revision of the standard Galactic model shortly after 1979. But baseline problems limited the possibility of measuring the tail of the emission profile at brightness temperatures less than 0.1 to 0.2 K. Assuming that the rotation curve stays flat, and that the surface density drops as an exponential with radius outside the solar circle (with scale length Rs), we can predict how the HI emission profile from the outer galaxy should drop off with velocity. Figure 2 shows a simple model prediction for the profile expected at longitude 90 degrees with several different values for the scale length Rs (in units of the solar circle radius, Ro). All the curves are normalized to give brightness temperature 1.0 K at velocity -115 km/s, which is what Knapp et al. see at longitude 90. The fourth example is the same as the second (Rs=0.4Ro) but with a very small constant density term added, which reaches out to 50 kpc radius. Only the fourth model shows the expected "pile up" of gas at the extreme negative velocities. Extending the constant density term further would increase this effect, and shift the peak all the way to -220 km/s.

Figure 3. The Leiden-Dwingeloo baseline limitations. This spectrum is an average over about ten degrees of latitude, at longitude 135, from the Leiden-Dwingeloo survey. It shows that there is indeed galactic HI at the extreme negative velocities discussed here, -200 km/s or so. However, the baselines clearly become a limiting factor in determining how far the disk extends. The emission profile drops smoothly to about -10 mK at -250 km/s, and then rises due to baseline curvature out to the edge of the band at -450 km/s. The GBT baseline should be much better than this.

The GBT offers the potential of superb baselines for Galactic HI spectra, without confusion by stray radiation or other effects of aperture blockage such as standing waves or solar interference. Figure 3 shows how even the best existing data (the Leiden-Dwingeloo survey of Hartmann et al., 1996) is limited by baseline curvature. We can hope that with the GBT it will be possible to measure the shape of the tail of the Galactic HI emission at negative velocities in the second quadrant (and at positive velocities in the third quadrant, which should be symmetrical) down to well below 100 mK, perhaps below 10 mK. If so we should be able to easily distinguish among models for the HI distribution in the outer galaxy such as those shown on figure 2. Knowing the density profile of the HI in the outer disk, and particularly knowing whether the Milky Way HI has an "edge" (which would cause a precipitous drop in the profiles at the corresponding velocity) will help us to understand the ionizing radiation environment and the tidal disruption history of the Local Group.

Burton, W.B., 1988, in Galactic and Extragalactic Radio Astronomy, 2nd ed., eds. G.L. Verschuur and K. Kellerman, (New York : Springer-Verlag) p. 295.

Burton and te Lintel-Hekkert, 1986, Astron. Astrop. Supp. 65, 427.

Corbelli, E. and Salpeter, E.E., 1993, Ap. J. 419, 104.

Ferriere, K. 1998, Ap. J. 503, 700.

Hartmann, D., Kalberla, P.M.W., Burton, W.B., and Mebold, U., 1996, Astron. Astroph. Supp. 119, 115.

Higgs, L.A. and Tapping, K.F., 2000, A.J. 120, 2471.

Kalberla, P.M.W., Mebold, U., and Reich, W., 1980, Astron. Astroph. 82, 275.

Kenney, J.D.P. and Young, J.S. 1989, Ap.J. 344, 171.

Knapp, G.R., Tremaine, S.D., and Gunn, J.E., 1978, Astron. J. 83, 1585.

Lockman, F.J. and Gehman, C.S., 1991, Ap. J. 382, 182.

Malhotra, S., 1995, Ap. J. 448, 138.

McClure-Griffiths, N.M., Dickey, J.M., Gaensler, B.M., Green, A.J., Haynes, R.F., and Wieringa, M.H., 2000, A.J. 119, 2828.

Murphy, E.M., Lockman, F.J., Laor, A., and Elvis, M., 1996, Ap. J. Supp., 105, 369.

Stark, A.A., Gammie, C.F., Wilson, R.W., Bally, J., Linke, R.A., Heiles, C., and Hurwitz, M., 1992, Ap. J. Supp. 79, 77.