Direct evidence of basic aluminium sulphate minerals in an S-impacted Andosol T. D ELFOSSE a , F. E LSASS b & B. D ELVAUX a a Unite´desSciencesduSol,Universite´catholiquedeLouvain,PlaceCroixduSud,2/10,1348Louvain-la-Neuve,Belgium,and b INRA, Science du Sol, Route de St-Cyr, Versailles 78026, France Summary The formation of basic aluminium sulphate (BAS) minerals [(K,Na) n Al x (OH) y (SO 4 ) z ] has often been invoked to explain sulphate retention in soils. These minerals have not yet, however, been directly observed in the soil. We extracted the clay fractions of Andosols intercepting large inputs of volcanogenic sulphur dioxide and acids (HCl, HF), simply by dispersing clays with Na þ -resins in deionized water without any other chemical treatment. Clay fractions concentrate 39–63% of total sulphur content of soil. Transmission electron microscopy coupled with energy-dispersive analysis revealed the presence of BAS particles, appearing as nodules and spheres. These particles have an equivalent diameter smaller than 0.2 m. They have an Al:S ratio close to 2.2 and 3.8 and are possibly amorphous aluminite or basaluminite, respectively. They seem to have been formed in microenvironments enriched in sulphate, but also in fluoride anions. Their formation seems to have been enhanced by the combination of large inputs of acids and SO 2 and an effective Al supply from weathering of volcanic glass. Introduction As suggested by Adams & Rawajfih (1977), the formation of basic aluminium sulphate (BAS) minerals [(K,Na) n Al x (OH) y (SO 4 ) z ], also named hydroxy aluminium sulphate, may contribute to sulphate retention in soils. Indirect proofs based on soil solution data support this hypothesis. Direct evidence of the occurrence of BAS minerals in soils is, however, still lacking. The presence of BAS minerals in soils has been generally inferred from studies involving geochemical modelling (e.g. Courchesne & Hendershot, 1990; Wolt etal., 1992; Agbenin, 2003) or selective extraction of sulphate (Delfosse etal., 2004a). Alunite [KAl 3 (SO 4 ) 2 (OH) 6 ], basaluminite [Al 4 SO 4 (OH) 10 .5(H 2 O)] and jurbanite [AlSO 4 OH.5(H 2 O)] were predicted to form according to calculations based on soil solution composition from a large variety of soil environ- ments: acid sulphate soils (van Breemen, 1973), forest soils (e.g. Courchesne & Hendershot, 1990; Prietzel & Hirsch, 2000; Mayer etal., 2001), soils fertilized with S (Hue etal., 1985), and ones exposed either to marine aerosols (Hue etal., 1990) or to S from volcanogenic deposition (Delmelle etal., 2003). Despite numerous attempts to identify BAS in soils, these minerals have not yet been detected in them. In soils, SO 4 2– -bearing minerals are generally present as compounds of Ca, Ba, Na, Mg and Fe, e.g. gypsum [CaSO 4 .2H 2 O] in gypsisols, barite [BaSO 4 ] in acid sulphate soils (Hanor, 2000), Na 2 SO 4 and MgSO 4 salts in evaporites on soils (Keller etal., 1986), and schwertmannite [Fe 8 O 8 (OH) 6 SO 4 ] and jarosite [KFe 3 (SO 4 ) 2 (OH) 6 ] in acid sulphate soils or in soils affected by acid mine drainage (e.g. van Breemen & Harmsen, 1975; Bigham etal., 1996; S ˇ ucha etal., 2002). The formation of BAS minerals is essentially determined by pH and the activities of Al and SO 4 2– in soil solution (Nordstrom, 1982), and is favoured by a SO 4 2– -rich solution in contact with a source of Al. These conditions are likely in young Andosols receiving volcanogenic acid and S, which may contain up to 6.6 g of SO 4 2– as BAS minerals per kg of soil (Delfosse etal., 2004a). In this paper, we report the results of a transmission electron microscopy study on clay fractions extracted from Andosols exposed to substantial deposition of acid and sulphur dioxide from Masaya volcano, Nicaragua. Materials and methods Samples of Vitric and Eutric Andosols exposed to prolonged deposition of volcanic sulphur and acid were collected in Nicaragua around Masaya volcano, one of the world’s largest sources of SO 2 . The recent Vitric Andosols (VI for Vitric) are weakly developed, and have large amounts of volcanic glass, Correspondence: B. Delvaux. E-mail: delvaux@sols.ucl.ac.be Received 16 April 2004; revised version accepted 4 August 2004 European Journal of Soil Science, June 2005, 56, 281–286 doi: 10.1111/j.1365-2389.2004.00675.x # 2004 British Society of Soil Science 281