“…Recently, observations show that there is a positive correlation between Eddington ratio λ and hard X-ray index Γ for λ > 0.01 (Sobolewska et al 2011;Kalemci et al 2006), and there is an anti-correlation between λ and Γ for λ 0.01 (λ = L bol /L Edd , with L bol being the bolometric luminosity, L Edd = 1.26 × 10 38 m erg s −1 being the Eddington luminosity, m being the central black hole mass scaled with solar mass M ⊙ ) (Wu & Gu 2008). It's interesting that this correlation is also found in active galactic nuclei, i.e., when λ 0.01, the hard X-ray spectrum softens with Eddington ratio (Porquet et al 2004;Shemmer et al 2006;Saez et al 2008;Sobolewska & Papadakis 2009;Cao 2009;Liu et al 2012;Zhou et al 2010;Veledina et al 2011), but when λ 0.01 the hard X-ray spectrum hardens with the Eddington ratio (Constantin et al 2009;Gu & Cao 2009;Younes et al 2011Younes et al , 2012Xu 2011). For the Comptonization models, it's argued that the above correlations mean the ratio of Compton luminosity to seed photon luminosity, ℓ h /ℓ s , changes with Eddington ratio, which is consistent with the scenario that the seed photons are from the synchrotron radiation at lower Eddington ratio and from the truncated thin disk at higher Eddington ratio (Sobolewska et al 2011).…”