Bandpass dependence of X-ray temperatures in galaxy clusters

We explore the band dependence of the inferred X-ray temperature of the intracluster medium (ICM) for 192 well-observed galaxy clusters selected from the Chandra Data Archive. If the hot ICM is nearly isothermal in the projected region of interest, the X-ray temperature inferred from a broadband (0.7-7.0 keV) spectrum should be identical to the X-ray temperature inferred from a hard-band (2.0-7.0 keV) spectrum. However, if unresolved cool lumps of gas are contributing soft X-ray emission, the temperature of a best-fit single-component thermal model will be cooler for the broadband spectrum than for the hard-band spectrum. Using this difference as a diagnostic, the ratio of best-fitting hard-band and broadband temperatures may indicate the presence of cooler gas even when the X-ray spectrum itself may not have sufficient signal-to-noise ratio (S/N) to resolve multiple temperature components. To test this possible diagnostic, we extract X-ray spectra from core-excised annular regions for each cluster in our archival sample. We compare the X-ray temperatures inferred from single-temperature fits when the energy range of the fit is 0.7-7.0 keV (broad) and when the energy range is 2.0/(1+z)-7.0 keV (hard). We find that the hard-band temperature is significantly higher, on average, than the broadband temperature. On further exploration, we find this temperature ratio is enhanced preferentially for clusters which are known merging systems. In addition, cool-core clusters tend to have best-fit hard-band temperatures that are in closer agreement with their best-fit broadband temperatures. We show, using simulated spectra, that this diagnostic is sensitive to secondary cool components (TX = 0.5-3.0 keV) with emission measures >= 10-30% of the primary hot component.