We study the X-ray spectra of tenuous , two-temperature accretion flows using a model involving an exact , Monte Carlo computation of the global Comptonization effect as well as general relativistic description of both the flow structure and radiative processes .
In our previous work we found that in flows surrounding supermassive black holes , thermal synchrotron radiation is not capable of providing a sufficient seed photons flux to explain the X-ray spectral indices as well as the cut-off energies measured in several best-studied AGNs .
In this work we complete the model by including seed photons provided by nonthermal synchrotron radiation and we find that it allows to reconcile the hot flow model with the AGN data .
We take into account two possible sources of nonthermal electrons .
First , we consider e ^ { \pm } produced by charged-pions decay , which should be always present in the innermost part of a two-temperature flow due to proton-proton interactions .
We find that for a weak heating of thermal electrons ( small \delta ) the synchrotron emission of pion-decay e ^ { \pm } is much stronger than the thermal synchrotron emission in the considered range of bolometric luminosities , L \sim ( 10 ^ { -4 } -10 ^ { -2 } ) L _ { Edd } .
The small- \delta model including hadronic effects in general agrees with the AGN data , except for the case of a slowly rotating black hole and a thermal distribution of protons .
For large \delta , the pion-decay e ^ { \pm } have a negligible effect and then in this model we consider nonthermal electrons produced by direct acceleration .
We find an approximate agreement with the AGN data for the fraction of the heating power of electrons which is used for the nonthermal acceleration \eta \sim 0.1 .
However , for constant \eta and \delta , the model predicts a positive correlation of the X-ray spectral index with the Eddington ratio , and hence a fine tuning of \eta and/or \delta with the accretion rate is required to explain the negative correlation observed at low luminosities .
We note a significant difference between the dependence of plasma parameters , T _ { e } and \tau , on the Eddington ratio that is predicted by the large- and small- \delta models .
This may be the key property allowing for estimation of the value of \delta .
However , a precise measurement of the spectral cut-off is required and we note that differences between results available in literature are similar in magnitude to the difference between the model predictions .
In flows surrounding stellar-mass black holes , the synchrotron emission of pion-decay e ^ { \pm } exceeds the thermal synchrotron only above \sim 0.01 L _ { Edd } .
Furthermore , in such flows the nonthermal synchrotron radiation is emitted at energies \gtrsim 1 keV , and therefore the Compton cooling is less efficient than in flows surrounding supermassive black holes .
This may explain spectral differences between AGNs and black-hole transients around \sim 0.01 L _ { Edd } ( the latter being typically much harder ) .