As a part of the Boston University-Five College Radio Astronomy Observatory Galactic Ring Survey, we have mapped both ¹³CO(1-0) and CS(2-1) emission over two square degrees (44.25 < l < 46.25, -0.5 < b < 0.5 degrees). Because CS has a higher critical density than ¹³CO, the CS to ¹³CO brightness ratio is sensitive to volume density. The GRS offers an unbiased, high-resolution dataset to examine the CS/¹³CO ratio toward a variety of objects, ranging from dense, star-forming regions to diffuse clouds. We find that this ratio varies by factors of (~7) between star-forming cores and diffuse clouds, and that this variation most likely reflects density changes. We also find that ¹³CO data alone are sufficient to identify star forming cores.
 

Figure 1: Is CS the best star-formation tracer? The top image is a map of the integrated intensity of CS(2-1) emission for the VLSR range -5 to 80 km/s. The bottom image shows the same region, where the color image now shows the integrated intensity of 13CO(1-0) emission and the contours are the CS(2-1) data.  We find that 13CO distinguishes star-forming cores as easily as CS.
 
 

    Figure 1 shows maps of two square degrees of sky in CS and ¹³CO.  There is a striking similarity in these two maps; with few exceptions, the CS map resembles a noisy version of the ¹³CO map.  The CS(2-1) line is commonly observed as a dense gas tracer.  With a high critical density, n ~ 10⁵ cm^-3 this line is thermalized in the dense gas of star-forming cores.  This has led many (including the GRS team) to observe in CS as a method of finding dense, star-forming (SF) cores.  The observing strategy of the GRS was to survey in two transitions, one a column density tracer (¹³CO) and the other a volume density tracer (CS), in the hope that each would provide a unique, yet complimentary view of the galactic ring.

   As Figure 1 shows, CS does not give a significantly different view of SF cores from that of ¹³CO.  Any peak in CS emission is seen clearly as a peak in ¹³CO.  (The advantages of observing a high volume density tracer are countered by dense cores having a large column density, as well.)  We easily find star-forming cores in the optically thin line of ¹³CO.  For these reasons, and due to the intrinsic faintness of CS, the GRS has decided to halt surveying in CS.
 

Figure 2:  Shown are CS and ¹³CO spectra together with their CS/¹³CO intensity ratio, for two regions in the Figure 1  image.  The top set is the average of all spectra of an expansive, diffuse cloud, the ``25 km/s Filament''. The bottom set is the average of all spectra of the core of an intense, star-forming region in a portion of the "Klingon Warbird."  The error bars refer to two standard deviations. The large intensity ratio seen in the SF region is due to its high volume density.

   We used the UMass-Stony Brook ¹²CO survey to estimate the kinetic temperature of gas in the two square degrees studied here.  The ¹²CO line is typically thermalized and optically thick, providing a direct measure of gas kinetic temperature.  The two regions represented in Figure 2 span the range of conditions seen in the ISM:  calm and diffuse to active and dense.  We find typical kinetic temperatures in these two regions to be roughly equal (T_mb(¹²CO)  from 8 to 15 K).  If we assume that the CS and ¹³CO emission in these regions is optically thin, then the ratio of their intensities is most sensitive to the H₂ volume density.  With these assumptions, we use the CS to ¹³CO brightness ratio as a measure of volume density.

   Figure 3 shows the dependence of this intensity ratio on ¹³CO intensity.  Future work will use excitation models to quantify the densities implied by these ratios.  Here we note that densities of molecular clouds range from diffuse (n ~ 10² cm^-3, where CS is subthermal) to dense (n ~ 10⁵ cm^-3, CS beginning to be thermalized). Figure 4 shows CS and ¹³CO spectra, as well as the CS/¹³CO intensity ratio for the average of all spectra.  The mean value of the intensity ratio for this "whole field'' spectrum is similar to the ratio seen in diffuse clouds.  This suggests that most of the CS emission comes from faint, somewhat diffuse, subthermally excited gas.

Figure 3: Ratio of CS and ¹³CO emission intensity as a function of I(¹³CO) for selected objects and flux ranges.

   Each point in figure 3 is calculated for a unique set of spectra selected by (1) association with a previously defined object, and (2) flux limits.  HII regions by Lockman (1989), and "diffuse'' objects are defined by general morphology.  The values of the intensity ratio (r = CS/¹³CO) range from r = 0.037/+-0.007 in the faint (3 sigma < I < 9 sigma) emission of a diffuse cloud, to r = 0.257/ +-0.004 in the intense ($I > 27 sigma) emission in the core of a SF (H II) region.  The dotted line shows the intensity ratio calculated for all spectra from VLSR: -5 to 80 km/s.  This velocity range covers all emission seen in the survey.  The ratio calculated over the whole field is r = 0.0436/+-0.0008.  This value for the intensity ratio is similar to the ratio seen in diffuse regions, suggesting that CS is typically subthermally excited.

Figure 4: The average spectra and intensity ratio for the entire figure 1 field.  The intensity ratio suggests that most CS emission is subthermally excited.
 
 

We have mapped two square degrees of the galactic plane in CS(2-1) and ¹³CO(1-0), with high spatial and velocity resolution on a fully-sampled grid.  We have compared the maps of these two species and found the following: ¹³CO is as useful as CS at finding star-forming cores.  In terms of efficiency, ¹³CO is more useful than CS, since its line is typically 20 times brighter. CS emission is not restricted to the densest regions of the ISM.  The average CS emission observed by the GRS is subthermally excited from widespread, diffuse gas.
 

REFERENCES:
Helfer, T.T., Blitz, L., 1997, ApJ, 478, 233
Liszt, H.S., 1995, ApJ, 443, 163
Lockman, F.J., 1989, ApJS, 71, 469
Sanders, D.B., Clemens, D.P., Scoville, N.Z., Solomon, P.M., 1986, ApJS, 60,1

A full size version of the poster can be obtained here (gzipped postscript file)

The GRS is supported by the NSF via grant AST-9800334 and AST-0098562