Résumé
Pedogenic carbonates and their geochemical composition are pertinent proxies of paleoclimates, paleoenvironments,and past Earth surface processes [1]. But dating pedogenic carbonate remains a challenge due to re-precipitationprocesses [2]. Radiocarbon dating is limited to 55 ka, making attractive the use of U-Th dating method (up to 500 ka).However, pedogenic carbonates are impure by nature making U-Th dating a methodological and analytical issue. The TotalSample Dissolution technic has been proved to be the only method capable to properly correct for the detrital Thcomponent [3]. An alternative method based on physical separation of detrital and authigenic phases using heavyliquor has been performed successfully on non-consolidated carbonate marine sediments [4]. Nevertheless, no attempt wasmade on indurated impure carbonates. We tested for the first time the heavy liquor separation on two Holocene pedogeniccarbonate nodules from Far North Cameroon, where the geological settings and the processes that have led to theirformation are well-constrained [5]; this area is thus ideal to test a new method and to compare it with existingradiocarbon ages.Pedogenic carbonate nodules include three main components: i) authigenic calcite; ii) granite saprolite; and iii)Saharan dust. The conventional method of isochrones is based on a two end-member mixing (i.e. authigenic phase anddetrital phase). This model is no longer applicable in the present case. We attempt to separate the differentmineralogical phases by using heavy liquor in order to reduce the system to a two-pole mixture. Results from a first noduledo not provide a plausible age. However, in a second nodule, an isochrone could be applied giving a calculated age quiteconsistent with radiocarbon data.