Is the star formation rate in $z\sim 6$ quasars overestimated?

Abstract

The large total infrared (TIR) luminosities ($L_{\rm TIR} \gtrsim 10^{12}~L_\odot$) observed in $z \sim 6$ quasars are generally converted into high star formation rates ($SFR \gtrsim 10^2~M_\odot$ yr$^{-1}$) of their host galaxies. However, these estimates rely on the assumption that dust heating is dominated by stellar radiation, neglecting the contribution from the central Active Galactic Nuclei (AGN). We test the validity of this assumption by combining cosmological hydrodynamic simulations with radiative transfer calculations. We find that, when AGN radiation is included in the simulations, the mass (luminosity)-weighted dust temperature in the host galaxies increases from $T\approx 50$ K ($T \approx 70$ K) to $T\approx 80$ K ($T\approx 200$ K), suggesting that AGN effectively heat the bulk of dust in the host galaxy. We compute the AGN-host galaxy $SFR$ from the synthetic spectral energy distribution by using standard $SFR - L_{\rm TIR}$ relations, and compare the results with the "true" values in the simulations. We find that the $SFR$ is overestimated by a factor of $\approx 3$ ($\gtrsim 10$) for AGN bolometric luminosities of $L_{\rm bol} \approx 10^{12}~L_\odot$ ($\gtrsim 10^{13}~ L_\odot$), implying that the star formation rates of $z\sim 6$ quasars can be overestimated by over an order of magnitude.