Introduction

There is an increasing realization that the core regions of clusters of galaxies are complex and dynamic environments. For some time now, it has been argued on the basis of data from imaging X-ray telescopes that the hot intracluster medium (ICM) of the core regions of rich clusters is radiatively-cooling on timescales shorter than the age of the cluster. This gives rise to the phenomenon known as a cooling flow. Fabian (1994) gives an extensive review of cooling flows up to and including constraints from the ROSAT observatory. Prior to the launches of the Chandra X-ray Observatory and XMM-Newton, the X-ray data strongly argued for the inhomogeneous cooling flow model in which gas in cluster cores cools from X-ray emitting temperatures down to unobservable temperatures as part of a multi-phase ICM over a spatially distributed region of the cluster core. An obvious mystery, and a strong hint that the real situation is more complex, was the lack of cool gas (including significant star formation) observed at other wavebands.

Not surprisingly, the X-ray view of cluster cores became appreciably more complicated with the launch of Chandra and XMM-Newton. With the very high dispersions possible using the reflection grating spectrometer (RGS) on XMM-Newton, detailed emission line spectroscopy of cluster cores became possible for the first time. Using these techniques, observations of Abell 1795 (Tamura et al. 2001) and Abell 1835 (Peterson et al. 2001) both revealed clear evidence for gas cooling from the virial temperature $kT>4{\rm\thinspace keV}$ down to 1-2keV. In particular, one could isolate and identify the L-shell emission lines of iron corresponding to gas spanning this temperature range. However, very tight upper limits were set on the amount of gas below 1-2keV which were in strong disagreement with the standard cooling flow model. In other words, there is evidence for gas cooling from the virial temperature down to 1-2keV, whence it disappears. This result has been generalized to a sample of clusters by Peterson et al. (2003). The explanation for the temperature floor is still far from clear. Strong (i.e., order of magnitude) metallicity inhomogeneities will skew the apparent cooling function such that gas below 1keV cools extremely rapidly, thereby eluding detection (Fabian et al. 2001). However, it is not known how such inhomogeneities will be formed or maintained. Thermal conduction and the action of a central radio galaxy may also be important in producing these temperature floors (Fabian, Voigt & Morris 2002; Voigt et al. 2002; Ruszkowski & Begelman 2002).

In addition to spectral complexity, Chandra has revealed that many clusters possess morphological complexities that are thought to arise due to the interaction of the ICM with a central radio galaxy. In some cases, the association is clear. For example, Perseus A (Fabian et al. 2000), Hydra A (McNamara et al. 2000; David et al. 2001; Nulsen et al. 2002), Abell 2052 (Blanton et al. 2001), and Cygnus A (Smith et al. 2001) all show well defined cavities in the X-ray emitting gas which are coincident with the current radio lobes of the central radio galaxy. In these sources, it is clear that the radio lobes have displaced the X-ray emitting gas producing the observed X-ray/radio anti-coincidence.

Chandra has also revealed the presence of ``ghost'' cavities, i.e., X-ray cavities that are not coincident with the active radio lobes. Examples include the outer cavity of Perseus A (Fabian et al. 2000), Abell 2597 (McNamara et al. 2001), NGC 4636 (Jones et al. 2002), and Abell 4059 (Heinz et al. 2002). In these sources, it is believed that the cavities are associated with old radio lobes (related to previous cycles of AGN activity). The low-frequency (74MHz) synchrotron radio emission expected within this scenario from these old radio lobes has been observed from the ghost cavity of Perseus-A (Fabian et al. 2002).

Collectively, these observations give rise to several questions. Most hydrodynamic models for the formation of these cavities (e.g., Clarke et al. 1997; Heinz, Reynolds & Begelman 1998; Reynolds, Heinz & Begelman 2001; Reynolds, Heinz & Begelman 2002, and references therein) involve the pressure-driven growth of a shock-bounded cocoon. However, in almost all cases (with NGC 4636 being a notable exception; Jones et al. 2002), the X-ray shells that bound the observed cavities are cooler than the ambient ICM, seemingly at odds with the shock scenario. It is plausible that the cool shell arises due to the ``lifting'' of lower-entropy material from the cluster core by the radio galaxy activity (Böhringer et al. 1995; Reynolds et al. 2001; Nulsen et al. 2002), but we would still expect to see some fraction of the sources in the shock-bounded phase. More generally, we need to assess the implications of such data for models of radio galaxy evolution. To achieve this goal requires the detailed analysis of more Chandra data together with directed numerical simulations.

Abell 4059 was one of the first clusters known to possess X-ray cavities on the basis of data from the ROSAT high-resolution imager (Huang & Sarazin 1998). These cavities were approximately coincident with the radio lobes of the FRI radio galaxy PKS 2354-35, which is hosted by the cD galaxy at the center of the cluster. Huang & Sarazin (1998) also noted an interesting bar-like feature in the central regions of the cluster perpendicular to the radio axis. The Chandra Advanced CCD Imaging Camera (ACIS) observation of this cluster has been previously described by us in Heinz et al. (2002). In that paper, it was shown that the coincidence between the radio lobes and the X-ray cavities is not exact, leading to the conclusion that these are actually ``ghost'' cavities. It was also suggested that the complex X-ray morphology (including the central bar) arises from an interaction of a radio-galaxy driven expanding cocoon and a pre-existing bulk ICM flow. Such an ICM flow may result from the accretion of a galaxy group by the cluster.

In this paper, we present a detailed reanalysis of the Chandra-ACIS observation of the core regions of Abell 4059. We present a spatially-resolved spectral study of the core regions of this cluster. We also present new 1.4GHz and 4.7GHz radio data from the Very Large Array (VLA) taken with the CnB configuration, thereby providing a better match providing a better match to the typical spatial scales characterizing the X-ray cavities. We confirm that the arcmin scale radio lobes indeed do not coincide precisely with the X-ray cavities, especially to the south-east of the center. We also find that the ridge of emission to the SW of the center is cooler and denser, but probably in pressure equilibrium, with the surrounding ICM. Furthermore, it is determined that the thermal evolution of this structure must be dominated by radiative cooling. We discuss various models for the SW ridge, but prefer an explanation in which it corresponds to shock/compression induced cooling of ICM caused by interaction of the radio-galaxy driven disturbance with a bulk ICM flow -- however, such a model may suffer fine tuning problems. Finally, we also present an archival Hubble Space Telescope (HST) Wide Field Planetary Camera 2 (WFPC2) image of the cD galaxy ESO 349-G010. The presence of a significant dust lane in this elliptical galaxy suggests that it has accreted a gas rich companion galaxy within the past $\sim 10^8$yrs. This provides further circumstantial evidence for the putative cluster/group merger required to produce the bulk ICM flow. Section 2 details the Chandra, VLA and HST data reduction, and Section 3 describes our imaging and spectroscopy investigations. The observational results are summarized, and possible models discussed, in Section 4.

Throughout this paper we assume $H_{0}=65$ km s$^{-1}$ Mpc$^{-1}$ and $q_{0}=0.5$. Given a redshift of $z=0.049$, this cosmology places PKS 2354-35 and Abell 4059 at a luminosity distance of 226Mpc.