3.1. X-ray data

ASCA and ROSAT studies of CL 0016+16 showed the cluster to have a somewhat elliptical shape and a gas temperature of about 7.6 keV (Hughes and Birkinshaw 1998). However the errors on the structural parameters of the gas ( and _c) in (25) and on the temperature remained a limiting factor in the interpretation of the cluster. XMM-Newton observations of the cluster were therefore made to provide better measurements of the properties of the atmosphere. A full description of the treatment of the data is given in Worrall and Birkinshaw (2003).

The XMM-Newton observation of CL 0016+16 took 37 ks of data of which rather little was lost to particle flares. An image of the cluster formed from the combined dataset from the three cameras on XMM is shown in Fig. 13. The overall structure of the cluster is somewhat elliptical, but relatively smooth. To first order in the ellipticity we can take the cluster to be circularly-symmetric, and fit the radial profile using the isothermal model, eq. (25). The fit is good (Fig. 14), and yields structural parameters = 0.70 ± 0.01 and a core radius _c = 36 ± 1 arcsec (corresponding to a linear core radius r_c = 240 ± 10 kpc in the standard cosmological model).

Figure 13. Combined MOS-1, MOS-2, and pn image of CL 0016+16 in 0.3 to 5.0 keV, corrected for vignetting but without background subtraction. Artefacts associated with the edges of the chips are evident. CL 0016+16 is the dominant central extended source. The associated quasar and one of the associated companion clusters lie just north and to the south-west, respectively, of CL 0016+16.

Figure 14. The 0.3 to 5.0 keV radial profile of CL 0016+16 after subtraction of local background. The curve shows the best-fit circularly symmetric isothermal model convolved with the point spread function of XMM. The horizontal dotted line shows the range of radii used for local background, and the level of that background.

A spectrum of CL 0016+16 extracted from the central part of the cluster is shown in Fig. 15. This spectrum can be well fitted by a single-temperature plasma, with k_B T_gas = 9.1 ± 0.2 keV and an abundance of 0.22 ± 0.04 times the solar abundance. There are sufficient counts in this spectrum (32600 net counts from the three cameras in 0.3 - 10.0 keV) that the redshift of the emitting plasma can be determined, and shown to be consistent with the optically-derived redshift of the cluster. It should be noted that there remain uncertainties in the spectral extraction and fitting procedure because of the significant backgrounds in the X-ray images and the complicated distributions of background counts in energy and position on the cameras, and because of residual uncertainties in the calibration.

Figure 15. The XMM spectrum of CL 0016+16 from a circle of radius 90 arcsec. The upper spectrum is derived from the pn data, while the lower is from the MOS-1 and MOS-2 data. The fit is to an isothermal gas with kBT = 9.1 keV, and abundance of 0.22 times the solar value, at the optical redshift of the cluster.

A further issue is that of structural or thermal substructure in the cluster. A close examination of the central part of the cluster's X-ray image reveals a small central sub-structure that is also seen in the Chandra data. The change in central brightness in the radial profile Fig. 14 due to this central structure is small, so the overall parameters of the model shouldn't be affected by its presence, but there is some evidence that the central structure is slightly (~ 0.5 keV) hotter than the bulk of the cluster, although the significance is not high.

There are some other variations from the model profile: in particular there is a region of low surface-brightness emission to the SW of the cluster that is also seen in an earlier ROSAT image (Neumann and Böhringer 1997), but again these brightness structures are too small to affect the overall structural fits.

With the aid of the gas temperature and abundance that are provided by the analysis of the X-ray spectrum, the detected counts can be converted into the central density of the gas in the cluster (conveniently characterised by the central proton density, n_p0) if some assumption is made about the cosmology. The derived n_p0 and k_B T_gas are subject to systematic errors from background subtraction uncertainties in both the radial profile and the spectrum, and from our ignorance about small-scale temperature variations in the cluster.