Comment on "Optical Imaging of Light-Induced Thermopower in Semiconductor" [Phys. Rev. Applied 5, 024005 (2016)]

In a recent article [Phys. Rev. Applied 5, 024005 (2016)], Gibelli and coworkers proposed a method to determine the thermopower, i.e. the Seebeck coefficient, using photoluminescence measurements. The photoluminescence spectra are used to obtain the local gradients of both the electrochemical potential difference between electron and holes and the temperature of the electron-hole plasma. However, the definition of the thermopower given in that article seems erroneous due to a confusion between the different physical quantities needed to derive this parameter.