The alignment of radio and optical axes

One of the most interesting discoveries in the study of high redshift galaxies is the almost universal alignment between the major axis of the optical emission and the radio axis (Chambers, Miley, van Breugel, 1987). A similar alignment has been seen independendly in high redshift 3CR galaxies by McCarthy et al. (1987). This phenomenon is more significant than the known opposite tendency for optically bright radio sources to be aligned with the minor axis of giant elliptical galaxies. This suggests that there is a fundamental difference between low and high redshift galaxies.

The radio maps were used to observe 33 4C-sources with a CCD in the R band with the 2.1 m KPNO (Kitt Peak National Observatory) telescope in August 1986 and February 1987. Accurate astrometry has identified 19 extended objects with . For 25 extended objects the position angles of the extended emission were measured within 3 arcsec of the central maxima. The images were smoothed to 3 pixels (1.14 arcsec) to improve the signal to noise ratio. Although the objects in this sample are at the limit of detectability, their gross structure does not appear to resemble that of giant ellipticals.

There is no evidence that the alignment that has been found is specifically related to the selection criterion of ultra steep radio spectra, rather than the fact that this criterion is efficient at selecting intrinsically powerful radio sources. The most important questions are:

• what is the nature of the extended optical emission.
• why does it coincide with the presumed energy pathway into the radiolobes.

Several models have been tested to give answers to these questions.

1. The alignment is produced because the optical images include continuum emission from the synchrotron jets. The optical magnitudes can be calculated by extapolation of the radio flux into the optical band. Using a spectral index of -0.6 gives (in the most optimistic case) a R magnitude greater than 20. We don't say that this is impossible, but it is highly unlikely that optical synchrotron jets are responsible for the alignment effect, except maybe for a few cases.
2. The observed elongated morphologies are caused by dust lanes perpendicular to the radio axes. Only a small number of low redshift giant ellipticals with powerful radio sources show a band of absorption across the galaxy, but a study of eight low redshift radio galaxies with dust lanes shows that in seven cases the dust lanes are perpendicular to the axis of the associated radio sources. There is evidence that many powerful radio galaxies are indeed dusty, because these sources often have a far infrared (60 m) excess.
3. The alignment is due to ionized gas associated with the radio source. Morphological association between extended radio and optical line emission regions have been observed in radio galaxies and in Seyfert galaxies. It is also known that high luminosity radio sources are capable of accelerating significant amounts of material (Breugel, Heckman, Miley, 1984).
4. The optical morphology is a result of continuum starlight from starforming regions associated with the radio sources. There is evidence that starformation in one gas-rich dwarf may have been triggered by the radio jet of a nearby ellitical. If this jet-enhanced starformation was more common at earlier epochs then this could cause the alignment found in the survey.
5. The alignment of the optical continuum and radio axis is due to the optical flux being the electron scattered radiation from a central quasar which is beamed out of our line of sight. The electrons are in the hot intracluster gas surrounding the quasar host galaxy. This model has been introduced by Fabian (1989).

One of the best ways to check which mechanism is most likely responsible for the alignment effect is to examine the spectral energy distribution (SED) of the extended emission across the widest possible spectral range. Especially the near infrared is an important diagnostic tool in discriminating between the various alternatives. The alignment effect has been studied extensively with 3C368, one of the brightest examples of the radio-optical alignment. This galaxy has a redshift of . Looking at figure 9 it is quite remarkable that also the 2.2 m image (K band infrared) is aligned along the major axis. There is significant infrared flux extending at least 32 kpc from the radio core, spatially aligned along the radio axis, and located just inside the radio hot spots.

Figure 9: The alignment effect in 3C368

What can these observations tell us about the different explanations for the alignment effect:

1. 3C368 is not a general case and the fact that we observe the alignment is just (bad) luck. The reason why 3C368 was imaged in the infrared was because it's one of the brightest galaxies in the samples of Chambers, Miley, and van Breugel (1987) and McCarthy et al. (1987) as well as having one of the best defined optical axes. There is no reason to doubt that the alignment of infrared flux observed in 3C368 is due to similar processes that produce the optical alignment in other powerful radio galaxies at high redshift.
2. The jets happens to be orientated along the stellar axis and therefore enhances the radio emission causing a statistical bias. We know that interaction with matter is an important method of making jets radiate at radio frequencies, but the alignment effect is only observed at high redshift. A statistical bias would also give similar results for low redshift objects, but this is not seen.
3. The broad-band emission is dominated by emission lines from an ionized gas with which the radio source is interacting. Figure 9d (from Djorgovski et al., 1987) shows the emission of [OII] 3727 in 3C368. There is no doubt that it is aligned with the radio axis. Although the broad band R image can be affected by the redshifted [OII] 3727 emission and similarly the J band measurement by large H plus [NII] flux, there are no strong redshifted emission lines in the K band that could distort the 2.2 m image. Hence, we can rule out the extended line emission as producing the observed infrared morphology.
4. The optical/infrared radio emission is enhanced by a gravitational lens formed by a foreground galaxy. Figure 9b (the high resolution radio map) shows a typical edge brightened galaxy and there is no evidence of a lensed radio core. By looking at the correlation between radio spectral index and radio luminosity (Blumenthal and Miley, 1979) and the fact that 3C368 has an ultrasteep radiospectrum () it is most likely that the source is a distant high luminosity radio galaxy. Other evidence that rule out this option are long-slit spectra of Djorgovski et al. (1987) that show spacial variations in the velocities and velocity widths in a least two spectral lines. This can not be accounted for by a gravitational lens.
5. The infrared and optical morphologies are dominated by non-thermal emission related to the radio source. This is not very likely because there is no spatial coincidence between the radio and infrared emission. If the radio synchrotron spectrum is extrapolated at the location of the infrared emission, the optical/infrared fluxes are underestimated by more than four orders of magnitude and therefore it's unlikely that we see the high frequency tail of the radio synchrotron emission. Nonthermal infrared/optical emission can also be produced by inverse Compton scattering of the microwave background radiation by the radio emitting relativistic plasma. If there is a low-frequency bridge in the region that emits radio flux at 6 cm below the detection limit ( mJy per beam) and has a similar spectral index as the observed infrared/optical measurements , then a magnetic field strength of G would be needed to produce the observed infrared/optical flux. This is a factor of smaller than the corresponding equipartition value. This difference is too large to be realistic.
6. The most likely interpretation is that the infrared/optical flux is starlight. This means that there is a relationship between the stellar emission and the radio activity. Like in the low-redshift radio galaxies you would expect that an old ( yr) population of red giants will produce the infrared emission, or a relatively young ( yr) population of red supergiants. These two interpretations require totally different minimum masses to explain the observations. It would require K5 III red giants with (with and ) to produce the infrared flux in region from the red giants alone. But with K5 Ia supergiants ( and ) you only need red supergiants or .
7. An option we cannot rule out is the hypothesis that the extended optical emission is electron-scattered radiation from a quasar. The strong emission lines also observed along the radio axis are explained as being due to cooled blobs of gas which are photo-ionized by the quasar radiation. The only argument against this model is the fact that large values for the gas density are needed in order to explain the observed luminosity.

If you think of high redshift radio galaxies as less evolved giant ellipticals, the spectral energy destributions (SED) have been interpreted as being dominated in the infrared by a population of red giants and in the optical and ultraviolet by young stars due to ongoing bursts of starformation. If you use this interpretation to explain the infrared emission in the K band you will need two extremely large giant ellipticals at the regions and . This is highly unlikely. There is no good reason why a merger would preferentially occur along the radio axis. If there is a preferred direction during the production of the radio sources and the formation of their associated galaxies, how can such a structure be maintained for yr? Several authors (Chambers, Miley, and van Breugel (1987) and McCarthy (1987)) have suggested that the starformation is caused by the interaction of the radio jet with its environment. This means that the light is dominated by relatively young massive stars. This hypothesis implies that powerful distant radio galaxies have an initial mass function (IMF) which is strongly biased against the formation of solar-mass stars as compared with the solar neighbourhood. Detection of strong CO absorption bands at a restwavelength of m produced by an extreme population of late-type red supergiants (Rieke et al., 1980) would strongly support this interpretation. The unusual IMF can exist if the formation of low mass stars could be suppressed in active galactic nuclei, because less massive clouds do not have enough self-gravity to overcome the forces resisting contraction, such as turbulence from supernovae, tidal effects or large magnetic fields. If the infrared flux is dominated by red supergiants then there are more consequenses for the SED. Red supergiants have 4000 breaks which are as strong or stronger than those of red giants which are dominant in low redshift galaxies and giant ellipticals. Differences in strength between these galaxies can give different broad-band infrared/optical colours. A detailed spectroscopic study of this infrared band around 4000 in high redshift galaxies might give some answers. For the last few years it looked like the discussion was in favor of the stellar population synthesis models, but recently several authors were claiming that the observed features are due to scattered quasar light.