A Multiwavelength Study of A3266

Introduction

Abell 3266 is a rich cluster of galaxies, and one of the first bright X-ray sources discovered in the southern sky. Its high velocity dispersion ( $\sim 1000 \mbox{ kms}^{-1}$) and the morphology of its hot X-ray emission has led to the hypothesis that A3266 is undergoing a cluster merger. A3266 contains a higher than average number of radio sources, including a region of diffuse emission aligned tangential to the X-ray emission.

These properties make it an interesting choice for a detailed study combining optical, radio and X-ray observations. In the context of this project, A3266 can be used to demonstrate the properties of a cluster which contains a diffuse radio source. This chapter brings together recent studies of A3266 in the literature and our data from recent observations of the cluster. We have good radio images of the cluster at three frequencies, 2368 MHz (13 cm), 1472 MHz (20 cm) and 843 MHz (36 cm) as well as an optical image from the DSS and an X-ray image from ROSAT.

ATCA Observations and Reduction

The Australia Telescope Compact Array (ATCA) is a synthesis array telescope consisting of six 22 m radio antennas which lie along an east-west oriented railway track. We observed A3266 at 3, 6, 13 and 20 cm on the 17th of January 1999 with the ATCA in the 750 m configuration. We observed over a 13 hour period, alternating between the four frequencies. The primary flux calibrator 1934-638 and the secondary calibrator 0438-436 were observed every hour.

The data were reduced using the radio interferometry package Miriad [Sault et al.1995]. The main steps in the process of creating images from raw visibility datasets are shown in Figure 5.1 and described below:

Flagging: each baseline is examined for bad data caused by interference in phase or amplitude. The data is edited, removing the affected frequency channels or time periods.

Calibrating: the secondary calibrator source is used to determine the complex antenna gains as a function of time and hence a relative flux density scale for the data. The primary calibrator is then used to correct for the flux density of the secondary calibrator.

Inverting: an image is created from the edited dataset.

CLEANing: the CLEAN algorithm [Högbom1974] represents a radio source by a number of point sources in an empty field of view. This point source model is convolved with a Gaussian beam, removing artifacts which are caused by the discrete sampling of the synthesis array telescope.

Self-calibrating: the antenna-based gain due to the source itself is determined, using an iterative procedure. This allows a more accurate estimate of the flux density scale and improves the signal-to-noise of the image.

  

Figure 5.1: Outline of the data reduction process.

r4.5cm

Restoring: the final images, Figure 5.2, are produced.

Polarization: an additional task that we have done is polarization analysis. In general, the detected emission is a mixture of linearly and circularly polarized waves. Any electromagnetic wave, with components $E_x = e_x(t)cos[wt+\delta_x]$ and $E_y = e_y(t)cos[wt+\delta_y]$ can be completely characterised by the Stokes Parameters

I = <e_x²(t)> + <e_y²(t)> (5.1)

Q = <e_x²(t)> - <e_y²(t)> (5.2)

$$2<e_x(t) e_y(t) cos[\delta_x - \delta_y]>$$ U = (5.3)

$$2<e_x(t) e_y(t) sin[\delta_x - \delta_y]>$$ V = (5.4)

where I is the total power, Q and U are the linearly polarized components and V is the circularly polarized component. The ATCA can detect linear combinations of two different polarizations and so by observing with different combinations, all of the Stokes Parameters can be determined.

  

Figure: Left: 13 cm image of A3266. The RMS noise in this image was reduced to about 0.26 mJy/beam using iterative self-calibration. Right: 20 cm image of A3266. The RMS noise in this image was reduced to about 0.45 mJy/beam.The circles in both of these images are of radius $\frac{1}{3}R_A$.

Individual Galaxies

Our study is the first deep radio survey of A3266. In this section we outline some of the properties of the individual radio sources in A3266, several of which have interesting radio morphologies. As well as being intersting in their own right, the individual radio sources in a cluster provide information about the dynamics of the cluster enviroment.

  

Figure: DSS image overlaid with MOST contours at 2, 8, 20, 40, 100, 300 mJy/beam. The sources studied in this section are numbered from 1-13.

Optical Identifications

We tried to identify an optical host for all the radio sources within a radius of $\frac{1}{3}R_A$. A DSS optical image was overlaid with MOST radio contours, Figure 5.3, and a search carried out in NED for sources within 30 arcsec of the coordinates of any optical host candidates. Sources with an offset of less than 6 arcsec were considered to be positive identifications. Information about these sources was collated and is shown in Table 5.2. The main purpose of identifying the sources is to determine whether they are associated with the cluster or whether they are just background sources which fall within the projected area of the cluster. Table 5.1 has comments on all of the sources, which are shown in Appendix E.

In summary, there are 13 radio sources detected in the region of interest (3 are just outside the $\frac{1}{3}R_A$). Of these, 7 are likely to be associated with the cluster (1, 2, 3, 4, 9, 10, 12), 3 are probably not associated with the cluster (5, 7, 8) and for the others (6, 11, 13) association is uncertain.

 

Table: Comments on each of the radio sources in A3266. Srcs 3, 12 and 13 are discussed further in Section 5.4. The acronyms in the host galaxy names are; GGP90 - [Green et al.1990], IRAS - Infrared Astronomical Satellite [Helou and Walker1988] and MRC - Molonglo Reference Catalogue [Large et al.1981]

Src	Comments
1	No optical identification. This source was not detected at 13 cm (c.f. Figure 5.2) which implies it has a very steep spectral index. The source is slightly extended and possibly associated with the cluster.
2	Identified as GGP90(008). This galaxy has an S0 morphology.
3	A diffuse source. It has a possible optical host GGP90(039) near the SW boundary of the radio emission but this is not confirmed.
4	Identified as IRAS 04311-6129, an irregular spiral galaxy.
5	No optical identification. This is a very faint point source and is likely to be a distant galaxy which was not detected by the DSS.
6	Possibly identified with GGP90(175). However, as Figure E.2(a) shows, the centres of the optical and radio emission are not completely aligned suggesting that this may be a chance coincidence.
7	No optical identification. This is a point source and is likely to be a distant galaxy, too faint for the DSS to detect.
8	No optical identification. This is a point source and is likely to be a distant galaxy, too faint for the DSS to detect.
9	Identified as GGP90(023) at redshift z = 0.062950 [Quintana et al.1996] This is a head-tail source with the peak radio flux centred on the optical galaxy and a fainter tail extending towards the east in the direction of the cluster centre.
10	Identified as GGP90(030).
11	There are no optical galaxies detected in this region.
12	Identified as MRC 0429-616, an S0 galaxy at redshift z = 0.055718 [Quintana et al.1996]. This is the brightest radio source in the field and has a very unusual morphology.
13	No optical identification. This source is very diffuse and is located towards the cluster boundary.

 

Table: Data associated with the optical identifications of the individual galaxies in A3266. The references are; L - [Loveday1996] Q - [Quintana et al.1996] J - [Jones and McAdam1992] G - [Green et al.1990] T - [Teague et al.1990]

Src	NED ID	RA	Dec	S843	Velocity	Redshift	Refs
		(J2000)	(J2000)	(mJy)	(kms-1)
1	no ID			69.6
2	GGP90(008)	04:31:57.3	-61:18:29	72.9	17681	0.058997	L Q G
3	GGP90(039)	04:30:45.5	-61:23:37	200.4	18819	0.062440	Q G T
4	IRAS(04311-6129)	04:31:47.7	-61:22:48	2.9	14426	0.048120	Q G T
5	no ID			4.3
6	GGP90(175)	04:31:12.3	-61:25:20	44.9	15336	0.051155	Q G T
7	no ID			10.5
8	no ID			7.0
9	GGP90(023)	04:30:42.0	-61:27:16	127.0	18872	0.062950	Q G T
10	GGP90(030)	04:30:56.8	-61:29:00	11.0	17325	0.057390	Q G T
11	MRC(0429-615)	04:29:56.1	-61:28:45	44.8			J
12	GGP90(003)	04:30:22.0	-61:32:03	2470.6	16704	0.055718	L Q J G
13	no ID			38.9

   
Integrated Flux Density

Also shown in Table 5.2 are the integrated flux densities, S₈₄₃, for each source. These are estimates based on calculations done using the cgcurs task in Miriad. We selected a polygon around each of the radio sources from the MOST image, and calculated the integrated flux over that area. This process had to be done manually as most source fitting programs deal only with point sources rather than extended ones. This method is quite accurate for point sources, such as Src 7, as it is relatively straightforward to fit a polygon to a source with well defined boundaries. For the more diffuse sources, such as Src 3, the fitting was harder to do, and the standard deviation after repeating the process five times was $\sim 1$ mJy. We tested this procedure on several strong point sources from the NVSS which had known flux densities. Our measured flux densities tended to slightly underestimate the NVSS values, the ratio being $S_{NVSS}/S_{kview} = 0.9856 \pm 0.0080$ [Azhar1999]. This method was also used to calculated the integrated flux densities for the candidate diffuse sources in Chapter 3.

   
Spectral Index

The spectral index of a source, $\alpha$, is a measure of how the intensity changes with frequency as modelled by the power law distribution $S_\nu \propto \nu^{\alpha}$. Although we have data at three different radio frequencies, the uv-coverage at each frequency is not the same and so the sensitivity to structure of a given angular size is different at each frequency. MOST has essentially continuous uv-coverage, whereas the ATCA is not uniform. Usually the configuration of the ATCA antennas would be changed to compensate for the different coverage at each frequency but this was not possible during our observations.

To partially compensate for the different uv-coverage we smoothed the higher frequency images so that the effective beam size was the same as the low frequency observations. For each of 6 regions in Src 3 we calculated the integrated flux density, as in Section 5.3.2, and determined the spectral index $\alpha$ from the slope of a log-log graph of intensity vs. frequency. When there was no detectable emission at the higher frequencies, a value of $3\sigma$ was used to put an upper limit on the flux density and hence a lower limit on the spectral index.

The overall spectral index of Src 3 is very steep, $\alpha \sim -2.8$ which led to the hypothesis that it is the remains of a radio galaxy, since these are the only sources of this type with such a steep index. Figure 5.4 shows the variation in spectral index with distance from the possible optical host galaxy of Src 3. The progressive steepening of the spectral index with distance from the head of the source strongly supports this hypothesis. This will be discussed further in Section 6.2. A more qualitative analysis also demonstrates that Src 3 is a steep spectrum source. Much of the emission in Figure 5.5 is only visible at the lowest frequency (843 MHz). At higher frequencies the contours show that a comparatively small region of the source is detectable.

  

Figure: Variation in spectral index with distance from the possible optical host galaxy of Src 3. The error bars combine the errors in our method of measuring integrated flux (see Section 5.3.2) and the error associated with which regions are selected to calculate the spectral index.

  

Figure: MOST greyscale (843 MHz) overlaid with ATCA contours at 1384 MHz (blue) and 2496 MHz (red). The contours are at 2 mJy/beam and 4 mJy/beam. This image shows the rapid decrease in intensity with increasing frequency, confirming this is a steep spectrum source. The green boxes label the regions used to calculate the spectral index at various places within the source.

Polarization

Some polarized emission was detected at 20 cm, but the levels at 13 cm were too low to be studied further. The 20 cm polarization image is shown in Figure 5.6.

  

Figure: Linear polarization in A3266 at 1384 MHz (20 cm). This image shows 20 cm polarized flux in greyscale, overlaid with the 20 cm total intensity contours to show the source positions.

Polarization can be seen in the areas of Src 2, Src 3, and Src 12. There is polarized emission in some regions of the image which do not contain a radio source which suggests the error margin associated with the polarized emission is quite high. Despite this, it is clear that there is strong polarized emission from Src 12, and much weaker, though significant ($>3\sigma$) emission from sources 2 and 3.

Substructure in A3266

Optical and X-ray studies indicate that around $30\% - 50\%$ of clusters show substructure in their galaxy and/or mass distribution [Flores et al.1999]. Evidence of a cluster merger in A3266 has been presented in the form of the spatial distribution of galaxies [Quintana et al.1996], N-body simulations [Flores et al.1999] and temperature maps [Henriksen et al.1999]. The merger is thought to be between two systems of comparable mass and to have started about $2 \times 10^9$ years ago.

The velocity distribution of the 317 members of A3266 is approximately Gaussian and in itself does not suggest the presence of substructure. The velocity distribution of all the galaxies with velocities 14000-25000 km/s is shown in Figure 5.7. This histogram is based on selected data from[Quintana et al.1996]. Also shown is the distribution of velocities for the radio sources possibly associated with the cluster. While the number of radio source is too small for statistical conclusions, the radio galaxies do appear to fall in the lower half of the velocity distribution. The galaxies in a cluster are thought to be on plunging orbits towards the strong potential well in the cluster centre. The fact that the radio galaxies are moving with velocities lower than the average probably means that they are moving towards us (since the overall cluster velocity is away from us).

  

Figure: Velocity distribution of galaxies in A3266. The white boxes show all the galaxies from [Quintana et al.1996] and the black boxes show the radio sources which we have identified as possible cluster members.

  

Figure 5.8: ROSAT image of A3266, overlaid with MOST contours at 2, 8, 20, 40, 100, 300 mJy/beam

   
Discussion

Evidence from the literature shows that A3266 is likely to be undergoing a merger between two comparable-mass units. A3266 has a higher than average number of radio sources, many of which have interesting radio morphologies. Our observations of A3266, which are of a higher quality than any previous ones, show that there are two diffuse sources (Srcs 3 and 13), a head-tail source (Src 9), a wide-angle-tail source (Src 12) and two slightly extended sources with no optical hosts that could be associated with the cluster (Srcs 1 and 6). This could support the hypothesis that a cluster-wide shock is occurring in the ICM.

Of particular interest for this project were the diffuse sources in A3266. We made an exciting new find by co-adding three MOST images. In the south-west end of the field, and outside the $\frac{1}{3}R_A$, is another diffuse source which has not been seen before (Src 13). It has very low surface brightness and so was only visible with the improved signal-to-noise of the co-added MOST image. We could not identify an optical host for this source.

Src 3 has a possible optical host at one end of the source -- identified with an arrow in Figure E.1(s). This is a relatively compact, elongated source. The overall spectral index of the source is quite steep and gets steeper with distance from the possible host galaxy. Both regions of diffuse emission in A3266 (Srcs 3 and 13) are aligned approximately tangential to the X-ray emission, see Figure 5.8. These properties suggest that Src 3 is a `relic' radio galaxy and that both Src 3 and Src 13 could be experiencing particle acceleration due to a shock front. These ideas will be discussed in the next chapter.

Src 12, the brightest source in the field is a wide-angle-tail (WAT). [Burns et al.1999] suggest that the morphology of these sources, rather than being due to their movement through the ICM, is due to the bulk flow of the turbulent ICM past the source. N-body simulations support this hypothesis, as does the fact that WAT's are often associated with elongated X-ray emission in the direction of the tails. It is also possible that the X-ray gas pressure influences the path of the jets. The morphology of head-tail sources, such as Src 9, can also be explained by the bulk flow of the ICM. In fact wide-angle-tail and head-tail source may be two different projections of the same phenomenon; a wide-angle-tail source viewed side on could look the same as a head-tail source. This again highlights the difficulty of developing a three dimensional model from two dimensional images.

In this chapter we have presented a detailed multiwavelength study of the cluster A3266. We have attempted to identify an optical host for every radio source in the region of interest and collated data from the literature for each of these sources. We have calculated the integrated flux density for each source, and the spectral index for Src 3, a diffuse source. A possible model to explain the morphology of the radio sources in A3266 is discussed in Section 6.4.