Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 318, L51-L54 (1997)

Previous Section Next Section Title Page Table of Contents

2. Observations and data analysis

2.1. Sample selection and properties

We selected 29 galaxies (cf. Table 1) from a large sample of starburst galaxies (Contini 1996). They share the following properties: 1) they host a young nuclear starburst (age ranging from 3 to 12 Myr) and 2) they have the largest far infrared fluxes. The latter criterion was to optimize our chances of detecting CO.


Table 1. CO linewidths and starburst ages of the sample of barred Markarian galaxies. W(CO) is the width of the line estimated at 20 and 50% of the maximum intensity of the CO profiles. Multiple entries correspond to different HII regions with comparable H [FORMULA] fluxes

All galaxies of the sample are barred galaxies. This classification comes from the LEDA database and/or from our CCD images at a resolution of [FORMULA] (Contini 1996). Most galaxies have morphological peculiarities, either an asymmetric spiral pattern (Mrk 133, 213, 353, 691, 799, 1466 and, to a lesser extent, Mrk 759 and 1050), several bright HII regions along the bar (Mrk 13, 281, 306, 710, 712, 731, 1341), or a polar ring (Mrk 306). Some galaxies also have a close companion (Mrk 306, 602, 691, 1379). Mrk 2 has a companion [FORMULA] away and Mrk 617 is a merger. The only galaxies without marked peculiarities are Mrk 52, 412, 545, 575, 708, 1088, 1194 and 1485. Two galaxies lacked high resolution images (Mrk 1365 and 1379).

2.2. CO line profiles

The radio observations were obtained at the IRAM 30 meter radiotelescope on August 20 to 24, 1995. We observed in single sideband at two frequencies simultaneously, CO(1 [FORMULA] 0) at 115.3 Ghz with the 1.3mm SIS receiver, and CO(2 [FORMULA] 1) at 230.5 Ghz with the 3mm (230G1) SIS receiver. The beamwidth of the 30 meter antenna is [FORMULA] and [FORMULA], and the main beam efficiency 0.75 and 0.39 at 115 and 230 Ghz respectively. The observing procedure and data reduction are detailed in Contini (1996).

We observed 18 galaxies of the sample; CO data for the other 11 galaxies were taken from the literature. Mrk 712 is the only undetected galaxy at 115 Ghz; this is not surprizing, as its far infrared flux turns out to be the lowest of the whole sample. Mrk 412, which also has a low far infrared flux, was barely detected at 115 GHz. Several galaxies, Mrk 13, 759, 1050, 1341, were ambiguously detected at 230 Ghz, mainly because the noise at that frequency was a factor 2 or 3 higher than at 115 GHz. Mrk 213 was probably detected at 230 Ghz, but the adopted recession velocity was too low, and, consequently, part of the profile was outside the observing frequency window.

The CO(1 [FORMULA] 0) and CO(2 [FORMULA] 1) line profiles are displayed in Contini (1996). The spectra were smoothed to a final velocity resolution of 10 km s-1 for both transitions. One of the main characteristics of the profiles is that they are generally not symmetric ; this cannot be attributed to noise, because of the long integration times (typically between 50 and 330 minutes) and the good resulting signal-to-noise ratio. Most profiles can be adjusted by two or three gaussians; this suggests the existence of separate molecular clouds of uneven sizes and/or with distinct velocities. Only six profiles (those of Mrk 13, 133, 412, 1050, 1379 and 1485) are well fitted by one gaussian.

The CO linewidths were measured directly on the profiles rather than by a multi-gaussian fit, because, when several components are present in the profile, the gaussians overlap and the FWHM thus does not reflect the width of the profile. The linewidths were determined both at 20 and 50% of the maximum intensity for our sample as well as for the data taken from the literature. We used these two estimates of the linewidths to check the consistency of our results. For the galaxies with high signal-to-noise ratio, the uncertainty on the linewidth is equal to the velocity resolution (10 km s-1). The linewidths of CO(1 [FORMULA] 0) and CO(2 [FORMULA] 1) profiles at 20 and 50% are given in Table 1. As the observing conditions (e.g. the beamsize) and the calibration procedures for the 11 galaxies from the literature were in general different from ours, these data are rather heterogeneous and have been used with caution; they are thus listed separately in Table 1 and plotted with different symbols in Fig. 1.

[FIGURE] Fig. 1. CO linewidth versus starburst age estimated from the models of LH95 (middle panel) and CMH94 (bottom panel). The top panel is for the mean of the ages predicted by the two models. [FORMULA]: CO(1 [FORMULA] 0) for our sample; [FORMULA]: CO(1 [FORMULA] 0) for data from literature; [FORMULA]: CO(2 [FORMULA] 1), when it is different from CO(1 [FORMULA] 0). Different data for the same object are joined by a line. Linewidths are measured at 20% (left panels) and 50% (right panels) of the maximum intensity of the CO profile

2.3. Optical spectra and starburst ages

The CCD spectra were obtained during several runs at the 1.93 meter telescope of Observatoire de Haute-Provence, with the Carelec spectrograph. We took long-slit spectra at 260Å/mm. We observed standard stars for flux calibration. The spectra were reduced with MIDAS. The deblending of the H [FORMULA] +[NII] and [SII] [FORMULA] 6716,6731 lines was done by fitting multigaussian profiles, with constraints on the relative positions of the individual lines.

The starburst ages were estimated from the equivalent widths W(H [FORMULA]) and metallicities (estimated from the oxygen abundance) measured on our spectra and the evolutionary synthesis models of Cerviño & Mas-Hesse (1994, CMH94) and Leitherer & Heckman (1995, LH95). We used a standard Salpeter Initial Mass Function (IMF) with [FORMULA], [FORMULA] to 2 [FORMULA] and [FORMULA] to 200 [FORMULA]. W(H [FORMULA]) was corrected for Balmer absorption (assuming [FORMULA] (H [FORMULA]) [FORMULA] 2 Å) which is a contamination from the underlying stellar population. The procedure for estimating the starburst ages is fully described in Contini et al. (1995) and Contini (1996). When several HII regions (or even a fraction of a giant HII region) fall inside the beam of the radiotelescope, the age is that of the brightest region in H [FORMULA]. HII regions outside the beam were not taken into account. Incidentally, the brightest HII regions are all nuclear, except for Mrk 13 ([FORMULA] from the center) and Mrk 306 ([FORMULA] from the center).

Three HII regions of Mrk 306 have similar (within a factor 2) H [FORMULA] luminosities. But the various ages are comparable for both models (7.3 to 7.4 Myr using CMH94 and 7.0 to 8.5 Myr using LH95), so that we adopted a mean value for the age. The three HII regions of Mrk 710 also have similar H [FORMULA] luminosities, but the ages range from 4 to 5.5 Myr (CMH94) and 6.3 to 8.4 Myr (LH95). We thus kept three entries in our data for this object. For Mrk 1379, two HII regions lie near the nucleus with very different ages (3.9, 5.7 using CMH94; 4.5, 8.6 using LH95). Although the ratio of H [FORMULA] luminosities is close to 9, we decided to keep both regions in our subsequent analysis.

A conservative estimate of the error on the age of young starbursts is 1 Myr for both models, taking into account the possible uncertainties on the slope of the IMF, the contamination by older or younger HII regions on the borderline of the beam, etc. For older starbursts, H [FORMULA] is weaker and the uncertainty on the age increases, and reaches 2 Myr for 10 Myr old bursts. However, there are several galaxies for which the ages differ between CMH94 and LH95 models by more than the sum of the age errors. We thus decided to keep only the objects showing a difference in age less than 2.5 Myr. The final age is thus the mean of the two estimates. This criterion does not reduce the number of young starbursts displayed in Fig. 1 with respect to those listed in Table 1. But fewer old starbursts are displayed, since there the error on the age difference may reach [FORMULA] 4 Myr.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: July 8, 1998