MOLECULAR GAS PROPERTIES OF THE STARBURST NUCLEUS OF IC 342: High Resolution 13CO(2-l) Imaging David S. Meier, Jean L. Turner Department of Physics and Astronomy, University of California, Los Angeles, CA 90095-1562 email: meierd,[email protected] and Robert L. Hurt Infrared Processing and Analysis Center, MS 100-22, California Institute of Technology, Pasadena, CA 91125 email:h [email protected] Received ; accepted to appear in the Astrophysical Journal 2- ABSTRACT We present a map of the J = 2 - 1 transition of 13C0 in the starburst nucleus of IC 342 made with the Owens Valley Millimeter Array. The interferometric 13C0 map allows us to directly compare with l2C0(1 -0), l2CO(2-1), and 13CO(l-O) mapso f nearly identical (- 4.5”) resolution. While all four transitions show a similar basic morphology, there are spatial differences between the l2C0 and 13C0 transitions which show up in ratio maps. In particular, the 13CO(2-l)/13CO(l-O) ratio has a markedly different distribution across the nuclear region than does 12C0(2-1)/12C0(1-0), indicating that l2C0 and 13C0 trace different components of the molecular gas. These differences are explained if l2C0t races the warm, PDR “skins” of the clouds and 13C0 traces thec ooler, interior portions constituting the bulk of the molecular gas. We derive excitation t-em peratures for the bulk of the cloud mass to be -10 - 20 K and densities of 103.7c m-2 along the molecular mini-spiral. Subject headings: galaxies:individual(IC342)-galaxies:ISM“galaxies: nuclei-ga1axies:starburst-radio 1ines:galaxies -3- 1. Introduction The l2C0(1-O) line is a universal and sensitive tracer of molecular gas in galaxies (Young & Scoville 1991). Because it is easilye xciteda ndt hermalized,l 2C0(1-O) is observed to be optically thick almost everywhere it is seen. The high opacity of the CO line constrains the information that can be derived from it, although it does appear to be a good temperature and mass tracer under certain well-defined conditions (Solomon et al. 1987). But it is important to study the behavior of this molecule in unusual conditions, such as starburst regions, since it is so ubiquitous and easily detected. IC 342 is an excellent subject to study since it is one of the closest spiral galaxies (D -1.8 Mpc; Ables 1971; McCall 1989; Madore, & Freedman 1992; Karachentsev & Tikhonov 1993) with substantial nuclear star formation (Becklin et al. 1980; Turner & Ho 1983)(Table 1). The molecular clouds in the center of IC 342 have been studied extensively (Morris & Lo 1978; Young & Scoville 1982; Young & Sanders 1986; Eckart et al. 1990; Sage & Solomon 1991; Rieu et al. 1992; Xie, Young, & Schloerb 1994; Paglione et al. 1995; Jackson et al. 1995). Est-im ated temperatures of the molecular clouds vary widely, ranging from values of T,, 10 K to more than 100 K ( Rieu et al. 1992; Martin & Ho 1979;1986; Ho, Turner & Martin 1987; Wall & Jaffe 1990; Steppe et al. 1990, Harris et al. 1991; Irwin & Avery 1992; Gusten et al. 1993) Recently, IC 342 has been studied at a high resolution, using aperture synthesis, in l2C0(1-O) (Lo et al. 1984; Ishuzuki et al. 1990; Levine et al. 1993), 13CO(l-O) (Turner & Hurt 1992 [TH92]; Wright et al. 1993), l2CO(2-1) Turner, Hurt, & Hudson 1993 [THH93], NH3 (Ho et al. 1990), and HCN (Downes et al. 1992). Both single-dish and interferometric observations have shown l2C0 line ratios that indicate the l2C0e mission is optically thin despite the large column densities implied by the l2C0 intensities (Ho, Turner & Martin 1987; Eckart et al. 1990; THH93). To reconcile this apparently contradictory results, we have made aperture synthesis - 4 - maps of 13CO(2-l). Due to lower optical depth, the 13C0 transitions potentially provide deeper and more representative probes of the gas than the low J l2C0 lines. We are now able to obtain interferometric maps of 13CO(2-l) and 13CO(l-O) at nearly identical resolutions 4.5") to accurately determine line ratios. By comparing ratio maps of the (w thinner 13C0 transitions with theirth icker l2C0c ounterparts, we investigate how molecular gas temperatures, densities, and general gas properties vary across the starburst region. 2. Observations The observations of the 13(2-1) transition (220.40 GHz) were made between 1990 December 26, and 1991 January 31 with the 3-element Owens Valley Millimeter Array (OVRO) (Woody et al. 1989; Padin et al. 1991). System temperatures were 800 - 900 K (SSB). A spectrometer with 32, 5 MHz, channels was used for the transition. Phase calibration was done using the quasar 0224+671. The absolute positional uncertainty is estimated to be 5 1'' at 1.3 mm. Absolute flux calibration was done using Uranus as the primary calibrator and the quasar 3C 454.3 as a secondary calibrator. Uncertainties in the absolute 1 mm flux of Uranus and 3C 454.3 give an estimated accuracy of 20% - 25% in absolute flux scale. We compare these 13CO(2-l) observations with the previous OVRO observations of 13CO(l-O), l2C0(1-0) and l2CO(2-1). The details of these observations are summarized in Table 2 (see Levine et al. 1993 ['2CO(l-O)], TH92 ['3CO(l-O)] and THH93 ["C0(2-1)]) for full details). To estimate how much extended flux the interferometer resolved out of the 13CO(2-l) map, it was compared with the fully-sampled, single-dish maps from the 30m IRAM telescope (Eckart et al. 1990) at 14'' resolution, and the FCRAO 14m telescope (Xie et al. 1994) at 23" resolution. The peak integrated intensities in the OVRO map when convolved to the 14" and 23" beamsizes are 25 K kms -l and 15 K km s-l, respectively. This - 5 - corresponds to ~ 8 0 %and~ 1 00% of the peak intensity of the respective single-dish maps. So it appears that little,i f any, of the 13CO(2-l) flux has been missed by the interferometer. This also is the case for 12C0(2-1), (THH93) and 13CO(l-O) (TH92), but not the case for l2C0(1-O), for which ”20% is resolves out. All of the maps were made and CLEANed using the NRAO AIPS package. The maps were naturally weighted and primary beam corrected. The OVRO primary beam is -64” at (1-0) and -34” at (2-1) (FWHM). In making the integrated intensity maps only emission greater than 1.30 was included. Structures extended on scales larger than 20” is resolved N out of the maps by the 10 kX minimum baselines. 3. RESULTS 3.1. Spatial Distribution of 13C0 gas in IC 342 The integrated intensity maps of the four CO transitions are plotted to the same scale in Figure la-d. In Figure 2, we display a schematic representation of the important regions. The basic morphology of the 12CO(l-O) is an open two-armed spiral pattern extending more than 60” in the north-south direction, which we refer to as the “arm” region (Lo et al. 1984; Ishizuki et al. 1990; Levine et al. 1993). The emission from the inner portion of the arm region is dominated by three major peaks lying about 40 PC to the northeast (C), southeast (A), and southwest (B+E) of the dynamical center (we use the letter designations of Downes et al. 1992). This three peak structure is seen in all the low CO transitions as well as in HCN (Downes et al. 1992). The starburst fills the “central trough” in CO, with the strongest region of star-formation coming from the southwestern edge, towards the interface with GMC B (Turner & Ho 1983; Figure 5). This region with the strongest star formation as traced by radio continuum emission is called the “starburst region.” A region - 6 of 12C0(2-1)/12C0(1-0) peaks are found near GMC C (53.3). This region is referred to as the “Eastern Ridge” (Eckart et al. 1990). Although the CO transitions all display a similar basic structure there are significant differences. In this paper we focus on these differences and their implications for the gas in the inner nuclear region. The 13CO(2-l) channel maps are presented in Figure 3. When corrected for the Rayleigh-Jeans approximation, T,=(hy/rC)/Zn(hy/~T+l), the radiation temperatures found for the ”CO(2-1) channels range from a 30 value of T, = 3.4 K to a peak of 7.7 K. The radiation temperatures for the other transitions are: 6 K - 24 K for 12CO(1-O), 5 K - 30 K for l2CO(2-1), and 2 K - 5 K for 13CO(l-O). Emission is present in channels V L ~ R= -5 km s-’ to 75 km s-l with higher velocities in the northeast. Measured antenna temperatures and integrated intensities of each transition are listed in Table 3 for various locations across the nucleus. The three peaks seen in the 13CO(2-l) map generally coincide with the position of the peaks found in the other CO transitions. However, unlike I2CO(2-1) which peaks in the northeast, the 13C0 emission peaks towards the southwest. Of all the transitions observed, 13CO(2-l) most closely reflects the structure detected in HCN by Downes et al. (1992). The southwest 13CO(2-l) intensity peak is a blend of their peaks, B and E, (as seen in the 11.2 - 31.6 km s-l channels). The field of view of the CO(2-1) maps are smaller than those of the CO(1-0) maps due to smaller primary beam at 1.3 mm. Even so, the detection of 13CO(2-l) is not limited by the primary beam as is the case for 12C0(2-1). In the Arm regions, the sensitivities are high enough to detect 13CO(2-l) if it has the same relative distribution as l2C0. The fact that we don’t implies 13CO(2-l) must be concentrated more strongly to the GMCs than the other CO transitions. 7 - 3.2. Molecular Gas ColumnD ensities Molecular gas column densities can be obtained from l2C0 using a standard conversion factor or from the 13CO(l-O) line assuming the transition is optically thin: Adopting isotopic abundance ratios for IC 342 of ['2CO]/['3CO] = 40 and [12C0]/[H2] = 8.5~10"f~or IC 342 (Henkel & Mauersberger 1993; Frerking et al. 1982), the observed 13C0 integrated intensities yield column densities ranging from N(H2) E 4x102' to 3 ~ 1 0 ~ ~ cmP2. (We have adopted a value of T,, = 20 K, which is lower than the dust temperature of Rickard & Harvey 1984, but consistent with the excitation temperature implied by the 13C0 line ratios ($4.1.2)) Using the Galactic conversion factor of XcO=2.3~10c~m~P 2 (K km sP1)-l (Strong et al. 1988), derivedc olumnd ensitiesr angef rom 1 . 5 ~ 1 0cm~-2~ - 2 ~ 1 c0m~-2 ~ov er beamsizes of -40 PC (Figure la). Assuming the central peaks are single spherical GMCs, we derive beam averaged densities of < nH2 > N 1.3~10cm~P 3 and individual cloud masses of < Adcloud > N 1.4~10~D.e nsities are likely much higher than the average < nHz > over portions of the cloud, consistent with the presence of HCN emission (Downes et al. 1992). The Hz masses we derive assuming 13C0 is optically thin are a factor of four lower than those derived from l2C0(1-O). However, it is likely that 13C0 column densities are underestimates of the true N(H2) due to optical depthe ffects ($4.3). 3.3. CO Line Ratio Maps In Figure 4a-d, the line ratio maps are presented. We follow the nomenclature of Aalto (1994) and designate = 12C0(2-1)/12C0(1-0), r13 = 13CO(2-l)/13CO(l-O), Rlo 1-12 = 12CO(l-O)/13CO(l-O), and Rzl = 12CO(2-l)/13CO(2-l). In generating the ratio maps, -8- the integrated intensity maps were convolved to matching the beamsizes, divided, and then blanked anywhere emission was < 40 in either map. Errors in the ratio maps are conservatively estimated to be 535% in magnitude and 5 2” in position. The r13 ratio map has relatively constant values of 0.8 - 0.9 over much of the central mapped region (Figure 4b). At dist-an ces greater than about 150 PC towards the northeast, the r13 ratio decreases to values of 0.3. The r13 ratio has only one peak with a value of 1.7. The location of the r13 peak is not coincident with the peak in the rI2 ratio map N associated with starburst region, but northeast of that peak, at the central trough. The 1-12 ratio map, by contrast, shows a fair amount of structure (THH93; Figure 4a). The 1-12 ratio ranges from 0.9 to 3.7, with an ave-ra ge across the region of approximately 1.4. Along the Eastern Ridge peak values of r12 2.2 are found. The r12 peak at the overlap of the nuclear starburst region and the southern CO intensity peak has a value of 2.1. The ratio maps are consistent with the single-dish ratio maps of Eckart et al. (1990), given the differences in beamsizes. The isotopic ratio maps, Rlo and Rzl are presented in Figure 4c & 4d. Values of 15 - 30 are found in the regions behind the spiral arms, both north and south of the nucleus. The ratio in the central trough is Rlo=11.3. The lowest values of Rlo -3 - 6 and R21 -5 - 9 are seen towards the GMCs A, C &L E. In agreement with Wright et al. (1993), we find the Rlo ratio varies systematically across the nucleus. Rlo is uniformly lower on the leading, compressed side of the density wave and becomes higher than is commonly observed in the Galaxy, on the trailing side of the spiral arm. Large R21 ratios are also seen in the “arm” and “off-arm” regions. The precise value ofR 21 is somewhat less certain, due to the weakness of 13CO(2-l) but must be very high (> 25). -9- 4. DISCUSSION 4.1.C O(2-1)/CO(l-O) Ratios (r12& r13): Cloud Temperatures In theory, r12 and r13, can be used to estimate temperatures of molecular clouds. In practice, this is difficult unless one simultaneously studies other molecular transitions including isotopes of different opacities (Sakamoto 1993; Aalto 1995). Assuming LTE: where i denotes the isotopic species, ZT, is the Rayleigh-Jeans corrected brightness temperature of the respective transitions and r is the optical depth of each respective transition. 2TBG(1-0) and iTBG(2-1) aret heR ayleigh-Jeansc orrectedb rightness temperatures of the cosmic microwave background (CMB). We can ignore the TBGt erms here because the uniform CMB is resolved out and will not affect ratios obtained from optically thick or strongly beam-diluted gas. A simplifying assumption often made is that both of the transitions have the same excitation temperature, Tez. Under this assumption, ratios between zero and four are expected, with ratios less than one for optically thick gas. Since the Rayleigh-Jeans approximation does not hold for cool clouds at this frequency, the ratio of the measured RJ brightness temperatures for CO(2-1)/CO(l-O) is not unity, and always less than one, for optically thick gas. This fact can be exploited to estimate the gas excitation temperature. 4.1.1. 7-12: Tracer of CO Photospheres At first glance it would seem that r12 > 1 (Figure 4a) imply optically thin emission. However, the column densities of the gas are much higher than would be implied by optically thin emission (Wall & Jaffe 1990; TH92; Meier & Turner 1999). THH93 suggest the possibility that temperature gradients can explain the higher ratios. They propose that 10 - - heating due to the star-formationco mplex is occurring in photo-dissociation regions (PDRs) at the surfaces of the large molecular clouds. Since the l2C0(2-1) transition saturates more quickly than l2C0(1-O), it would arise preferentially in these warm, outer layers of the externally heated clouds (Crawford et al. 1985; Stacey et al. 1991; Wolfire, Hollenbach, & Tieliens 1993; Hollenbach & Tieliens 1997). The l2CO(2-1) would then have a higher T,, and larger brightness temperatures than '2CO(l-O). The recent detection of 2.12pm H2( 1-0) emission from this region also supports the existence of hot (52 000 K), "PDR" gas (Boker et al. 1997). For the Eastern Ridge, a similar explanation seem likely for the high 1-12. In Figure 5, the HST' Ha image of IC 342 is overlaid on the 13CO(2-l) and2 cm contours (2 cm: Turner & Ho 1983; Ha: Gallagher et a1 1999). HI1 regions extend along the outside edge of both CO arms as well as towards the starburst core associated with the 2 cm peak. The 2 cm radio continuum maps (Figure 4) show a secondary peak at the Eastern Ridge. This is the location of warm NH3 gas (270 K Martin & Ho 1986; Ho et al. 1990) and high excitation l2C0(6-5) (Harris et al. 1991). An increase in the HeI/By ratio is seen at this location, indicating the presence of substantial numbers of hot (T, 2 36,000 K), massive stars, which may provide the ionization and heating for the Eastern Ridge (Boker et al. 1997). Towards the southwest edge of the map r12 4. The emission from this off-arm region appears to be warm and optically thin in the usual sense, since the CO emission is weak here and there is bright Ha emission. Because these are ratios of weak emission and near the edge of the primary beam, we do not discuss them in detail. But it does appear that ~~ ~~~ 'Based on observations made with the NASA/ESA Hubble Space Telescope, obtained from the data archive at the Space Telescope Science Institute. STScI is operated by the Association of Universities for Research in Astronomy, Inc. under the NASA contract NAS 5-26555.