Title: Evolution of mechanism of parton energy loss with transverse momentum at RHIC and LHC in relativistic collision of heavy nuclei

Abstract: We analyze the suppression of particle production at large transverse momenta in ($0-5%$ most) central collisions of gold nuclei at $\sqrt{s_\textrm{NN}}=$ 200 GeV and lead nuclei at $\sqrt{s_{\textrm{NN}}}=$ 2.76 TeV. Full next-to-leading order radiative corrections at ${\cal{O}}(\alpha_s^3)$, and nuclear effects like shadowing and parton energy loss are included. The parton energy loss is implemented in a simple multiple scattering model, where the partons lose an energy $\epsilon=\lambda \times dE/dx$ per collision, where $\lambda$ is their mean free path. We take $\epsilon=\kappa E$ for a treatment which is suggestive of the Bethe Heitler (BH) mechanism of incoherent scatterings, $\epsilon = \sqrt{\alpha E}$ for LPM mechanism, and $\epsilon=$ constant for a mechanism which suggests that the rate of energy loss ($dE/dx$) of the partons is proportional to total path length ($L$) of the parton in the plasma, as the formation time of the radiated gluon becomes much larger than $L$.
We find that while the BH mechanism describes the nuclear modification factor $R_{\textrm{AA}}$ for $p_T \leq$ 5 GeV/$c$ (especially at RHIC energy), the LPM and more so the constant $dE/dx$ mechanism provides a good description at larger $p_T$. This confirms the earlier expectation that the energy loss mechanism for partons changes from BH to LPM for $p_T \ge \lambda <k_T^2>$, where $\lambda \approx$ 1 fm and $<k_T^2> \approx$ 1 GeV$^2$ is the average transverse kick-squared received by the parton per collision. The energy loss per collision at the $\sqrt{s_\textrm{NN}}$ =2.76 TeV is found to be about twice of that at 0.2 TeV.