The search for diffuse non-thermal inverse Compton (IC) emission from galaxy clusters at hard X-ray energies has been undertaken with many instruments, with most detections being either of low significance or controversial. Because all prior telescopes sensitive at E > 10 keV do not focus light and have degree-scale fields of view, their backgrounds are both high and difficult to characterize. The associated uncertainties result in lower sensitivity to IC emission and a greater chance of false detection. In this work, we present 266 ks NuSTAR observations of the Bullet cluster, which is detected in the energy range 3-30 keV. NuSTAR's unprecedented hard X-ray focusing capability largely eliminates confusion between diffuse IC and point sources; however, at the highest energies, the background still dominates and must be well understood. To this end, we have developed a complete background model constructed of physically inspired components constrained by extragalactic survey field observations, the specific parameters of which are derived locally from data in non-source regions of target observations. Applying the background model to the Bullet cluster data, we find that the spectrum is well-but not perfectly-described as an isothermal plasma with kT = 14.2 ± 0.2 keV. To slightly improve the fit, a second temperature component is added, which appears to account for lower temperature emission from the cool core, pushing the primary component to kT ∼ 15.3 keV. We see no convincing need to invoke an IC component to describe the spectrum of the Bullet cluster, and instead argue that it is dominated at all energies by emission from purely thermal gas. The conservatively derived 90% upper limit on the IC flux of 1.1 × 10 −12 erg s −1 cm −2 (50-100 keV), implying a lower limit on B 0.2 μG, is barely consistent with detected fluxes previously reported. In addition to discussing the possible origin of this discrepancy, we remark on the potential implications of this analysis for the prospects for detecting IC in galaxy clusters in the future.

A3667 is the archetype of a merging cluster with radio relics. The northwest (NW) radio relic is the brightest cluster relic or halo known and is believed to be due to a strong merger shock. We have observed the NW relic for ∼40 ks of net XMM-Newton time. We observe a global decline of temperature across the relic from 6 to 1 keV, similar to the Suzaku results. Our new observations reveal a sharp change of both temperature and surface brightness near the position of the relic. The increased X-ray emission on the relic can be equivalently well described by either a thermal or nonthermal spectral model. The parameters of the thermal model are consistent with a Mach number M ∼ 2 shock and a shock speed of ∼1200 km s −1 . The energy content of the relativistic particles in the radio relic can be explained if they are (re)-accelerated by the shock with an efficiency of ∼0.2%. Comparing the limit on the inverse Compton X-ray emission with the measured radio synchrotron emission, we set a lower limit to the magnetic field in the relic of 3 μG. If the emission from the relic is nonthermal, this lower limit is in fact the required magnetic field.

The brightest cluster radio halo known resides in the Coma cluster of galaxies. The relativistic electrons producing this diffuse synchrotron emission should also produce inverse Compton emission that becomes competitive with thermal emission from the ICM at hard X-ray energies. Thus far, claimed detections of this emission in Coma are controversial (e.g., Fusco-Femiano et al. 2004;Rossetti & Molendi 2004). We present a Suzaku HXD-PIN observation of the Coma cluster in order to nail down its non-thermal hard X-ray content. The contribution of thermal emission to the HXD-PIN spectrum is constrained by simultaneously fitting thermal and non-thermal models to it and a spatially equivalent spectrum derived from an XMM-Newton mosaic of the Coma field (Schuecker et al. 2004). We fail to find statistically significant evidence for non-thermal emission in the spectra, which are better described by only a single or multi-temperature model for the ICM. Including systematic uncertainties, we derive a 90% upper limit on the flux of non-thermal emission of 6.0 × 10 −12 erg s −1 cm −2 (20-80 keV, for Γ = 2.0), which implies a lower limit on the cluster-averaged magnetic field of B > 0.15 µG. Our flux upper limit is 2.5× lower than the detected non-thermal flux from RXTE (Rephaeli & Gruber 2002) and BeppoSAX (Fusco-Femiano et al. 2004). However, if the non-thermal hard X-ray emission in Coma is more spatially extended then the observed radio halo, the Suzaku HXD-PIN may miss some fraction of the emission. A detailed investigation indicates that ∼50-67% of the emission might go undetected, which could make our limit consistent with Rephaeli & Gruber (2002) and Fusco-Femiano et al. (2004). The thermal interpretation of the hard Coma spectrum is consistent with recent analyses of INTEGRAL (Eckert et al. 2007a) and Swift (Okajima et al. 2008; Ajello et al. 2009) data.