The desert surface is characterized by periodical outbreaks of mineral dust storms. These dust aerosols have an impact on the upwelling infra red radiance. The desert surface has a low emissivity in the 8.7μm channel, compared to the 10.8μm channel. This characteristic can be used to separate the TOA brightness temperatures of clear-sky and aerosol events. In order to do so, we have to take into account the surface contribution to the measured satellite signal. To separate this from the aerosol signal, we calculate the Top Of the Atmosphere (TOA) clear-sky radiances (as brightness temperatures) of the 8.7μm, 10.8μm and 12.0μm channels. The clear-sky brightness temperatures in those channels for a fixed time of the day are selected using a common clear-sky criterion over the different days of a month. Different clear-sky criterions are tested, e.g. taking the maximum 10.8 - 8.7μm brightness temperature difference over the month. Empirical analyses of the 8.7μm and 10.8μm channels shows that the satellite measured radiance over desert decreases in the presence of aerosols or clouds. The brightness temperature value of both channels is very similar for aerosol pixels, whereas for clear-sky it differs significantly. The resulting clear-sky brightness temperatures are valid during both night and day. During day time we verify this method with the clear-sky detection by the RMIB-team in the context of the GERB-processing. The RMIB GERB clear-sky detection onldy uses visible channels. We validate the clear-sky brightness temperature values derived with the infrared method by comparison with the brightness temperature values for the times identified as clear-sky by the RMIB GERB method. The TOA clear-sky brightness temperature for desert will serve as reference image when calculating the optical properties of aerosols. It will prove valuable for the implementation of a climatology of African mineral dust, based on Meteosat-8 imagery.