Résumé
The reports of mass-independent sulfur isotope anomalies (MIF-S) in sediments older than 2.45 Ga have been attributed to photolytic reactions involving volcanic SO (sub 2) in an oxygen-poor atmosphere. Photolysis experiments of SO (sub 2) coupled with various UV shielding scenarios provided additional links to the early atmosphere. However, no simple model can reproduce the mismatch in the Delta (super 33) S-delta (super 34) S relationship between the reference Archaean sulfide array (positive Delta (super 33) S-delta (super 34) S correlation) and product sulfate (negative Delta (super 33) S but positive delta (super 34) S). The discrepancy in the temporal and spatial record of sulfur isotope anomalies, with three main sulfate horizons deposited within less than 300 Ma compare to a global distribution of sulfide over more than 1, 500 Ma, is also unexplained. Here we report a new Delta (super 33) S-delta (super 34) S linear trends recovered in two felsic volcanic ash layers of the 3.2 Ga Mapepe Formation in South Africa. This "felsic volcanic array" forms a tight Delta (super 33) S-delta (super 34) S linear correlation that is best approximated by SO (sub 2) photolysis experiments at deep UV wavelength. The perfect match to the Delta (super 33) S-delta (super 34) S values of associated sulfate and equivalent felsic volcanoclastic and sulfate horizons of the 3.5 Ga old Dresser Formation, Western Australia, indicates that the exogenic sulfur cycle that produced this array was linked to felsic volcanism and sulfate precipitation. An emerging scenario for the early Earth atmosphere is a continuous photochemical haze that is perturbed between 3.5 and 3.2 Gyr by massive and optically thick volcanic plumes. This volcanic activity coincides in time with a period of rapid crust formation (Valbaara supercontinent).