University of Manchester, Nuffield Radio Astronomy
Laboratories,
Jodrell Bank
These latest results are part of a continuing study of the starburst phenomenon in M82. The compact radio features are now established to be recent supernova remnants with ages of a few hundreds or thousands of years. A new remnant is thought to appear about every 20 years. A previous study at 5GHz showed the M82 remnants are similar to, but younger and smaller than the equvilent remnants in the LMC and our own galaxy. This latest image extends this study to include larger older remnants. Suprisingly many remnants show significant low-frequency spectral turnover which is probably due to free-free absorption in the gas surrounding the remnant. The required emission measure is similar to those of giant Galactic HII regions. By comparison with a VLA 5GHz image of comparible resolution, the central 12 arcsec (180 pc) around the inner core of M82 is seen to possess a steeper radio spectral index than regions further from the core. This steeper region corresponds to that lying within the established ring of molecular gas which surrounds the nucleus of M82.
The MERLIN array has been used to study the nuclear region of M82 at high angular resolution, complimenting the well established VLA studies by Kronberg and Sramek (eg Kronberg et al, 1985). A previous study with the enhanced MERLIN at 5GHz mapped in excess of 40 of the supernova remnants (SNRs) with an angular resolution of 50 mas (Muxlow et al, 1994). It was found that these SNRs follow a surface brightness/diameter relation consistant with those found in the LMC and Galactic remnants. The cumulative number as a function of diameter appears to increase linearly to diameters of at least 3 pc, and indicates a supernova rate of 0.05 per year if the shells are expanding at 5000 km/s .
This latest study attempts to fully image the whole of the nuclear region of M82 at L-Band. In addition to the original 1.42GHz data (April 1993), a second run was obtained (May 1993). Furthermore, in order to enhance the density of the spatial frequency coverage an additional run was obtained during May 1993, cycling every few minutes between 1.612, 1.658, and 1.720GHz. This technique is known as Multi-Frequency Synthesis (MFS). In total around 50 hours of data were collected. The remaining low spatial-frequency coverage was obtained by adding a 4 hour VLA A-array observation (May 1994).
The analysis of the L-Band data is still continuing, however a preliminary image at 0.3 arcsec resolution is shown in Figure 1. The dynamic range of this image is close to 2000:1 and residual image errors are close to this level. Although further improvements are expected, the present image is substantially correct. The flux densities and sizes of the more compact SNRs in Figure 1 agree well with the single frequency analysis of Sanders (1994), however, for the first time the extended emission at L-Band is fully imaged.
A comparison of Figure 1 with a VLA 5GHz A-Array image of M82 (data kindly supplied by Kronberg and Sramek) observed in December 1992 with a angular resolution of 0.33 arcsec show both striking similarities and interesting detailed differences. Figure 1 is more sensitive to the extended emission around the inner 1 kpc of the nucleus of M82 than the 5GHz image (1 kpc = 66 arcsec). However even at higher flux density levels there are signinficant differences in the extended emission between L-Band ant 5GHz which cannot be assigned to imaging errors.
Several of the SNRs show significant spectral flattening, the most extreme case being the remnant 44.01+596 which has strongly rising spectrum between 1.4GHz and 5GHz. By including flux density measurements from VLA 8.4GHz observations (Huang et al. 1993, and recent (May 1994) unpublished A-Array data) it is clear that the spectra for the SNRs are steep between 5GHz and 8.4GHz. Thus we are detecting low-frequency turnovers. Around 40% of the more compact SNRs show evidence for low-frequency turnovers in their spectra. At 50 milliarsec resolution, 5GHz MERLIN resolves virtually all these compact SNRs; the resulting sizes are thus too large to allow sychrotron self-absorption to explain the spectra. We have suggested that free-free absorption in the gas surrounding such SNRs can explain the low-frequency turnovers seen. Using the gradient of the spectrum between 5GHz and 8.4GHz (where we assume there is no absorption) we can extrapolate back to the value of the SNR flux density that we would expect at 1.4GHz. By comparing this with the measured value we can calculate the optical depth and hence the emission measure (E) for a number of SNRs. Assuming an electron temperature of 104K we get values of E of around a few x 106pc cm-6which is consistant with the emission measure in giant Galactic HII regions. It is also comparible with the well documented emission measure of the brightest SNR 41.95+575 which shows a frequency turnover below 1GHz (Wilkinson & deBruyn 1984). This strongly suggests that free-free absorption is the cause of the low-frequency turnover in many of the SNRs. The random distribution of the remnants exhibiting this behaviour implies that the absorption is being caused by local clumps of gas rather than one single cloud. It may also explain the differences between the 1.4GHz and 5GHz images of the extended structure in the nuclear region.A further point to note is that the central 12 arcsec (180 pc) in general shows an overall steeper spectral index than the surrounding regions. The reason for this is not as yet fully understood, but it is interesting to note that starformation in M82 is interpreted as moving outwards from the centre of the galaxy, and that CO observations by Nakai et al. (1986) show a ring of molecular gas with a radius centered on 200pc. This is the region of current starformation and the radio SNRs appear to be situated on its inner rim. The innermost 12 arsec is thus that region lying inside the molecular gas ring.
Kronberg, P.P., Biermann, P., & Schwab, F.R., 1985. Astrophys.
J.,291, 693.
Huang, Z.P., Thuan, T.X., Chavalier, R.A., Condon, J.J., & Yin, Q.F.,
1993. Astrophys. J.: submitted.
Muxlow, T.W.B., Pedlar, A., Wilkinson, P.N., Axon, D.J., Sanders, E.M.,
& deBruyn, A.G., 1994. Mon. Not. R. astron. Soc., 266, 455.
Nakai, N., Hayashi, M., Hasegawa, T., Sofue, Y., Handa, T., & Sasaki,
M., 1986. IAU Symp. 115, 614.
Sanders, E.M., 1994. M. Sc. Thesis, Unversity of Manchester.
Wilkinson, P.N., & deBruyn, A.G., 1990. Mon. Not. R. astr. Soc.,
242, 529.