Astron. Astrophys. 323, 323-336 (1997)

3. New radio continuum measurements

Since the publication of DBDKS, a number of additional radio continuum measurements, taken at the NRAO-VLA ³, have been reduced and are summarized here briefly. The data were mostly taken at 8.44 GHz and higher frequencies, with the goal to better determine the thermal radio emission reported in DBDKS. Observations at X (8.44 GHz), U (15.0 GHz) , K (22.4 GHz), as well as additional L (1.49 GHz) band data were undertaken in the C and D -array configurations. An overview of the radio observations and the measured flux densities is given in Table 5. Data already presented in DBDKS are included, this time listing the locations of the flux peak. None of the K band observations produced usable results, due to high atmospheric humidity. Default VLA observing set-ups and standard calibration and mapping procedures were used for all new observations, except for L band, where 8 channel spectral line mode was used to avoid possible external interference. The calibration of the L band data was analogous to the P band observations described in DBDKS. The FWHM of the restoring clean beam has sizes of about for X band in C array and in D array, for U band in D array, and about for L band in D array. The new radio data extend the known radio spectra for several of the galaxies, and the new H images set limits on the sizes of the H II regions. Where flux densities at similar wavelengths had been determined previously (Table 4 in DBDKS), the new observations generally confirmed these.

Table 5. Overview of the radio observations

We also re-evaluated the inverse Compton loss rates, which are now based on the optical diameter , whereas the 1.4 GHz radio diameter was used in DBDKS. This re-evaluation, and the addition of the new radio data, changes the interpretations of the galaxies' radio spectra from DBDKS in minor ways only, except for Haro 15. The bend in its radio spectra is now most likely caused by a 'delta-shaped' relativistic electron injection at a higher age of 6 Myrs, versus 1.2 Myrs previously given. This is due to significantly lower Compton losses from a big difference between its diameter of , and a 1.4 GHz diameter of .

Haro 1 had not been included in DBDKS and high resolution radio data are presented here for the first time. For Mkn 314 the addition of two higher frequency data points made an analysis of its radio spectrum possible. Both galaxies' spectra (Figs. 1a and 1b) can be understood either as combinations of thermal and nonthermal radio emission or as straight power laws. Both possibilities can be fitted to the data with about the same, relatively poor, significance. The numerical results of the fits are given in Table 6a for Haro 1 and Table 6b for Mkn 314. For a description of these tables and of the fits we refer to Table 6 in DBDKS and the associated text.

Fig. 1a and b. a Radio spectrum of Haro 1. The solid line is a fit of a combination of thermal and nonthermal power law spectra, the dashed line is a fit of a straight power-law. The data points for 1.489 GHz and 8.44 GHz are from this work, the others are from Klein, Wielebinski and Beck (1984) and Klein et al. (1991). b Radio spectrum of Mkn 314. The fit shown is a straight power-law. Data points not listed in Table 5 are from Klein et al. (1991).

Table 6a. Results of fits on Haro 1

Table 6b. Results of fits on Mkn 314

The additional 8.44 GHz point for Mkn 527 supports the previous interpretation, that its spectrum is a simple sum of thermal and nonthermal powerlaw spectra, and does not change any of the derived parameters. The extension of III Zw 102's spectrum to 14.9 GHz lowers to a maximum of 15% and limits the maximum size of the H II region to , to be compatible with the interpretation involving free-free absorption and emission (fit function in DBDKS), but this does not change the conclusions in DBDKS otherwise.

3.1. Estimates of the thermal emission

Thermal radio emission serves as a primary indicator for starforming activity. Thermal radio emission and H line emission are both proportional to the intensity of Lyman-continuum photons emitted by hot stars inside the H II region. Although H emission is subject to extinction, it can serve as a useful lower limit to the thermal radio emission. Radio emission is not subject to extinction or absorption, at least at frequencies 2 GHz (DBDKS). Radio measurements at high frequencies ( 8GHz) can be assumed to contain very little nonthermal emission and therefore give stringent upper limits on the thermal emission, , at any frequency. However, the most probable amount of thermal radio emission, , (Table 7, for 1.49 GHz) can only be derived from fits of the radio spectra which estimate the thermal fraction, , of the total radio emission, (DBDKS, Deeg 1993). A comparison between H and thermal radio fluxes is therefore useful. Following Lequeux (1980), the thermal radio flux at a frequency and the H flux, , are related to each other by:

where is the electron temperature of the ionized gas, for which a value of K has been assumed. Using the lower limit to the H flux, , and the fully corrected flux, (from Table 4), the equivalent thermal radio fluxes and at 1.49 GHz have been derived (Table 7).

Table 7. Thermal flux densities at 1.49 GHz, given in mJy

As expected, the fluxes are significantly lower than the fluxes, , which have been derived from fits to the radio spectra (DBDKS; Deeg, 1993). In fact, many authors (e.g. Lequeux (1980), Lequeux et al., (1981), Caplan and Deharveng (1986), Berkhuijsen (1983)) define the optical absorption in H II regions from the ratio . The intrinsic H emission, , based on , is about equal or slightly less than , the worst discrepancy is II Zw 70 with . This result agrees with those of the authors just mentioned, who found that even fully corrected H fluxes are always approximately equal or less than the H -equivalent derived from radio observations. This discrepancy may result from two factors: Although not done in this work, it has been frequent usage to equal the measured total radio flux at 5 GHz with the thermal flux. Unless the radio spectrum of a galaxy is known to be flat, significant nonthermal emission may still be present at 5 GHz and the thermal radio flux quoted from such a measurement is too large. Second, the corrections to obtain the intrinsic H fluxes may be poor, as the H flux may be subject to absorption that deviates from the conditions of uniform interstellar absorption and from the galactic extinction laws.

For the galaxies Haro 1 and Haro 15, Fanelli et al. (1988) performed spectral synthesis based on UV spectra. The UV spectra contain direct information on the intensity of ionizing photons in the H II regions, which Fanelli et al. used to derive a thermal radio flux density at 5 GHz, , for which a conversion to 1.49 GHz is included in Table 7. In the case of Haro 1, agrees well with the measured radio flux, whereas for Haro 15, is barely compatible with . For the adopted thermal flux, , and the adopted thermal fraction , the largest weight was given to the radio spectra if either free-free absorption or thermal-nonthermal separation gave good fits in DBDKS, otherwise all estimates were given about equal weight. These adopted values are used in the discussion of Paper II.

© European Southern Observatory (ESO) 1997

Online publication: June 5, 1998

helpdesk.link@springer.de