Sen2Cor AC correction at edge wierd


before correction

after correction (also tried directly downloading from Copernicus data hub)
Anyone knows why “clouds” appear along the edge after atmospheric correction?