1 Identification and interpretation of ferruginous rock in tropical environment

1.2 Weathering of sulfidic rocks

  Some recalls are necessary since the topic is very well defined by numerous synthetic or special studies. The purpose is to emphasize the peculiar features of the alteration of sulfides which superimposes to tropical climatic conditions, with reference to C.R.M. Butt, R.E. Smith, C.F. Blain, R.L. Andrew, W.R. Ryall, G. F. Taylor, A. Kosakevitch, K.M. Scott, P.M. Ashley, D.C. Lawie, providing themselves a more important bibliography and to more general treatises, like those by P. Routhier, A.A.Levinson, J.M. Guilbert and C.F.Jr. Park, A.A. Beus and S.V. Grigorian, V. Smirnov.


Sulfidic rocks are represented in all the groups of rocks and are characterized by the presence of sulfidic minerals which are no minor constituents, should they occur under dispersed or massive form. Their final alteration stage consists of a supergene ferruginous rock, called gossan (“chapeau de fer”). Iron oxides are the principal components of the gossans, the study of which will be our main purpose


a Mechanisms and definition of alteration environments


The mechanisms governing the high reactivity of sulfidic minerals and ore deposits to weathering refer to the principles and laws of chemistry and electrochemistry. From the latters results a selective mobility of elements released by hydrolysis according to oxidation (Eh) and acidity (pH) conditions in the created media (table 2). Here will be considered the consequences rather than mechanisms, i.e. constitutive minerals, materials and rocks.


Table 2 

Mobility of some elements as related to pH and Eh (from Andrews-Jones, in Levinson, 1974)


The conditions in which metallic sulfides alter are particularly agressive, with formation of sulfuric acid by hydrolysis of sulfides succintly speaking. This phenomenon can be observed even if S is in a dispersed form, even if the parent rock cannot be qualified of sulfidic one properly speaking, even under little aggressive climate. The only contact with climatic agents do, create conditions of rock destabilization, such as its disaggregation and neogenesis of secondary sulfates and carbonates (Van Oort and Robert, 1988). As sulfides are major minerals of the rock, thus as S is more concentrated, the tendency described above develops on a large scale resulting in a complete destabilization of the rock and frequently of its surrounding.


Numerous endogeneous and exogeneous factors govern this evolution. Moreover, the notion of “climatic stress” applied to sulfidic mineralizations must obviously be considered in a larger way and specificity does not exclusively depend on climate. Smirnov (1936) characterized the multiple factors acting during supergene alteration of a sulfidic ore deposit. These principal factors depend on climatic conditions, on morphological and tectonic features, on criteria typical of the ore and the surrounding rocks, on the type of ore deposit, on physico-chemical and dynamic properties of ground waters…


From this, it comes that humid and warm climates are more favorable to sulfides weathering than dry or cold ones. Weathering develops more easily on vertical bodies than on horizontal ones, because of a more intense percolation. The abundance of pyrite is an essential factor of acidification (production of higher contents of H2 SO4) whilst a carbonate – rich environment makes neutralization easier by stopping release of mobile elements. A little loaded and easily renewed water is likely to be a better vector, whereas the variations of the piezometric level of a water table are capital to characterize a zonation of the altered ore deposit toward the surface. From these water-ore deposit interactions develop very contrasted and changing media which account for pH and Eh conditions and result in zonation of alteration.


The principal successive stages are as follows: cementation of sulfides (supergene secondary sulfides), supergene oxidized zone structured by sulfatation, carbonation, phosphatization and oxidation of the sulfidic rock. The examples described by Scott et al. (2001) in volcanic massive mineralizations (VMS) are rather common in numerous mineralizations (figure 3).


Fig. 3 

Weathering zonation in the Cu-Pb-Zn-bearing VMS from Scott et al. (2001).


Thermodynamic models are multiple and well established for the principal mineral filiations of metals.


The final product is a oxidized neogenetic rock, ferruginous for the major part, that is what is denominated a gossan (Blanchard, 1939 ; Kosakevitch, 1979 ; Butt and Smith, 1980). In the 19th century, von Groddeck (1879) already wrote : “The gossans are frequent on stratiform sulfide, sulfidic vein and siderite deposits. They constitute the ore mined (for example the iron ore devonian of Nassau, the Ag ores of Potosi, Oruro, Charnacillo …). They (modifications ol filling) produce a gossan from deposit is made at depth by ferruginous pyrites, pyrite bearing copper, magnetic pyrite, arsenopyrite, etc, or by spatic iron carbonate».


This evolution was commonly described in the 54 examples of Australian mines quoted by Butt and Smith (1980), among which 32 display secondary ferruginisations. They concern Cu, Pb, Zn, Ag, Ni, (Au) resources and are denominated gossans, pseudogossans or ironstones. This evolution is global, except in zones where climates have been always dry or frozen. However, present-day desertic or quasi desertic zones like Arabia, exhibit gossans (Ryall and Taylor, 1981


b Transformations and mineralogical evolution


From the stress zonation results a succession of structured mineralogical transformations, in which are observed Fe, Zn, Pb, Cu … silicates, halides, tungstates, molybdates, arseniates, vanadates, carbonates, sulfates, phosphates as well as native elements, distributed according to a zonation of oxidized minerals, succeeding to a cementation zone. The later is well identified in Ni-Cu and Fe-Ni-Cu sulfidic mineralizations (Nickel et al., 1974 ; Wilmshurst, 1976 ; Imbernon, 1998) and in massive Zn-Pb-Cu mineralizations (Scott et al., 2001). In the cementation zone of the Ni-Cu-Pt O’Toole mine (MG, Brazil), sulfides are replaced by other supergene sulfides, with filiations from pentlandite, pyrrhotite and chalcopyrite. In that case, oxidation is a direct transition from primary and secondary sulfides to iron hydroxides without intermediary sulfates and carbonates.


Stage 1 

  • pentlandite --- > violarite 1
  • pyrrhotite and chalcopyrite preserved

Stage 2 

  • pyrrhotite --- > Ni pyrite + violarite 2
  • violarite 1 and chalcopyrite preserved

Stage 3 

  • violarite 1 + violarite 2 + Ni pyrite + chalcopyrite ---> (Ni, Cu) goethite
  • pyrrhotite + Ni pyrite + violarite 2 --- > Ni goethite


In the oxidation zone, zonation is expressed by successive horizons being first sulfated, then carbonated, phosphatic and last oxidized, and ending by a gossan. For VMS, Scott et al. (2001) defined a mature ideal zonation. There are obviously important variations around this model according to the characteristic of the parameters proposed by Smirnov (1936). For example, it can be noticed that all the patterns proposed are dealing with a vertical or strongly inclined mineralized body, which is of course a commonly wide-spread form favoring a special geometric form: the examples selected by Blain and Andrew (1977) or Ryall and Taylor (1981) concern only subvertical mineralized bodies.


This partly masks the reality of complexity in filiations which often display secondary minerals which can be ephemeral. However, as in all of weatherings, there are elements redistributions some of them being relatively preserved, the major part of the others being removed. In the Pb-Zn-Ba mine of Canoas (PR, Brazil), the mineralized body weakly inclined (10-15°) is expressed by horizons being little marked in terms of mineral association (table 3). In this framework, the itineraries of the economic Pb and Zn metals evolve in distinct filiations observed under microscope and by electron microprobe. These itineraries implicate participation of elements from sulfides, sulfates and silicates, the latter being the main components of the host rock of the ore (table 4). In that case, despite the silicate-bearing environment and the occurrence of several secondary minerals, no supergene clay minerals have been observed (Imbernon et al., 1999).


Table 3 

Mineralogical composition of the weathering levels of the Canoas mine (after Imbernon et al., 1999)


Table 4 

Place of Pb and Zn in the weathered products of Canoas ore (after Blot et al., 1995, Oliveira et al., 1996 ; Imbernon, 1998 ; Imbernon et al., 1999)


In general the oxidation zone of mines is rich in various secondary minerals, which is a feature emphasized by all the examples of gossans described, and Broken Hill Lode is quoted because of its some sixty different secondary minerals. It can be reminded that the weathering zone was the first ore worked for numerous metals before the industrial era. Frequently, gossans preserve many secondary minerals belonging to various families such as metal, sulfate, arseniate, oxide, carbonate, silicate, halid, phosphate, vanadate and molybdate (Blain and Andrew, 1977 ; Scott et al., 2001). In a quasi desertic zone like Arabia, where weathered products can no more evolve, Ryall and Taylor (1981) observed notwithstanding a great variety of secondary minerals families. However, it can be emphasized that clay minerals, which are typical products of silicated rocks weathering, are rare. They exist, like in Broken Hill, but their origin cannot be specified, since kaolinite is there only one of the very numerous secondary minerals. Among 54 samples of gossans from Arabia, clay minerals, only identified five times, are quite accessory minerals. Clay minerals are thus little specific and testify to weathering of walls or resistant blocks rather than to the mineralized body’s one.


Gossan properly speaking represents the term of sulfidic weathering into a residual supergene ferruginous rock in which iron minerals are predominant. Blanchard and Boswell (1925), Boswell and Blanchard (1927, 1929) defined two types of iron oxides : native oxides formed in situ in the cavity previously occupied by sulfids ; exotic oxides transported and precipitated in another place. According to the authors, the formation of either of the two types depends on the sulfidic nature. For example, oxidation of chalcopyrite and pyrite would result in native and exotic oxides respectively. Secondary carbonates are interpreted either as resulting either from interaction of sulfated solutions with the carbonated gangue (Loughlin, 1914 ; Blanchard and Boswell, 1925 ; Boswell and Blanchard, 1927, 1929), or as consequence of a particularly high CO2 pressure (Trischka et al., 1929). As minerals, which are non-secondary ones are abundant, the gossan is said immature, whilst as they are very subordinate, they characterize a mature gossan (Blain and Andrew, 1977 ; Scott et al., 2001). The principal secondary mineral is goethite in the major part of the cases studied. Hematite which may be predominant occurs frequently ; lepidocrocite is rather rare. Collomorphous little crystallized forms are known, whereas magnetite and maghemite are more rarely found. This assemblage of minerals is usually denominated iron oxides or limonite and includes informations inherited from parent rock such as textural and geochemical features. The iron minerals are associated in many cases to secondary metallic minerals inherited from the oxidation zone (table 5).



Table 5 

Principal oxidation-born minerals in gossans (after Blain and Andrew, 1977)


Further we shall see that the gossans studied in Togo or Burkina do not display any other marker secondary minerals than iron oxides and manganese and iron hydroxides. These iron minerals are the only minerals bearing elements specific of a given mineralization whatever the form in which they are found. A regional climatic specificity may be referred to, in order to explain this mineralogical poverty of tropical gossans.


c Facies and microfacies


The occurrence of a iron-bearing supergene rock capping the weathered zone of sulfidic mineralizations is common in sum. Its textural and structural description was thoroughly improved because of its specificity as compared to the other rocks and of the importance of possible diagnosis owing to the original ore. The observations and descriptions are ancient and numerous in many mines of all the mining regions in the world although strong differences in maturation and evolution of gossans have been pointed out.


The direct relation between gossans and sulfidic mineralizations extended to ferriferous carbonate rocks, and its identification are important diagnostic factors for the prospecting of sulfide metallic resources (Kosakevitch, 1979, 1983).


These surficial formations frequently display morphologic and faciologic features directly inherited from the mineralized rock, or weathering textures typical of primary minerals and a composition strongly marked by the mineralized rock’s one ; this constitutes the secondary halo of the mineralization, direct or indirect heir of the sulfidic body and primary halo, at the time of mineralization setting (Beus and Grigorian, 1975 ; Butt and Smith, 1980).


Prospectors of mineral resources paid special attention to gossans as markers of sulfidic mineralization. Blain and Andrew (1977) and Kosakevitch (1979) reported the history of the studies, the beginning of the 20th century being dedicated to observation of facies and microfacies as ferruginized inherited features from primary and secondary metallic minerals to which neogenetic filling material adds.


After oxidation, primary minerals generate typical replicas in the form of supergene iron oxide pseudomorphoses. According to Blanchard and Boswell (1925-1934), there are two principal morphologies: cellular morphologies or “boxworks” characterized by regular and angular forms of the cavities and spongy morphologies corresponding to rounded cavities with very variable sizes. During the first half of the 20th century, a pragmatic diagnosis mode developed from a classification of the gossans textures, by Blanchard, Boswell, Emmons, Lindgren, Locke, Sales, Schneiderhohn, Smirnov. The principal boxwoks are defined in table 6.


Table 6 

Diagnostic boxworks textures commonly observed in gossans (Ryall et Taylor, 1981)


Kosakevitch (1983) emphasized the difference between inheritage principally characterized by pseudomorphoses developed from altered minerals and displaying a specific porosity and various types of concretion formation involving transfer of iron considered as exotic. This exotism may concern restricted environments and not necessarily long itineraries.


Out of mineralized regions, there is little information about the oxide forms of eventual mineralizations, first since the prospection records are rarely accessible but also, since supergene ferruginisations are little or not studied at all, which is the case in the whole West Africa where the prospecting of gossans was never a privileged prospection method.


d From facies to outcrops


On the scale of field observation, the important and characteristic features are facies and outcrops. In some way, the summation of the microfacies generate the facies, but macroscopic organisation is still more complex. Some specific facies do only exist in gossans and are observed neither in other surficial rocks nor in other ferruginous rocks. Several criteria such as homogeneity, organization, colour or similarities with other rocks define the facies. For example massive sulfides display facies of massive gossans, whilst boxworks developed more readily from scattered mineralizations. There is thus a transposition of facies inherited from the mineralization, coupled with the general organisation of this mineralization on the scale of the outcrop. Beyond massive facies, some other ones exist such as facies of oriented rock, geodic closed structures, brecciated facies, facies characterized as jasperoidal. These facies are rarely confounded with the other outcropping or supergene ferruginous formations and are thus a good criterion of discrimination from ironcrusts. The confusion with breccias without any relation with mineralizations seems more frequent, particularly because of the importance of structural studies which settle the geodynamic frame of prospectives.


Their shapes give a more or less deformed picture of the surficial track of concentration they come from. As the shapes of mineralized bodies are strongly anisotropic (vein, bed, lens…), the latter very often appear on exposure as more or less delimited lenses, continuous or not, displaying frequently orientations at different scales.


In the Australian examples of Butt and Smith (1980) the anisotropy of the outcrop is rather systematically inherited from the mineralization’s one:

  • At Woodcutters, the Zn-Pb mineralized body (180m x 3.75) corresponds to a (150 x 1to 7m) gossan ;
  • At Mary River, to the Zn-Pb mineralized body (140m x 19) corresponds a gossan of 150 m, displaying extensions of same direction over more than 300 m ;
  • The Cu lens of Esperanza (4500m x 10) is surficially expressed by numerous scattered ferruginous blocks, but the diagnosis of gossan is not laid down ;
  • At Ti Tree Well, pseudogossans are evoked which spread over 3km on the Pb-Zn mineralization ;


  • At Canoas (PR, Brazil) on the Pb-Zn-Ag-Ba mineralization, a gossan corresponds to a plurimetric track of the ore lens, whilst two others are dislocated blocks without clear orientation (Imbernon et al., 1999) 
  • At Pagala (Togo) gossans have an apparent thickness of several meters and are several hectometres long.


This morphology of the outcrops is noticeable enough for being a first criterion of discrimination between gossans and ironcrusts in tropical zone with very thick lateritic ironcrust covers.


e Rich geochemical composition


The mineralization defined by its useful metals gives way to a surficial picture which is in some way its replica, non-related to environment. Gossans are ferruginous outcrops rich in metallic or non-metallic elements. Elements usually in traces are clearly more represented here and become eventually major elements and so much because some of major elements are removed during weathering. It is useful to evoke the direct economic role of these weathered formations up to the explosion of metals demand at the beginning of industrial era (von Groddeck, 1879): all the resources worked for basic metals came practically from weathering horizons.


This absolute or relative richness is an essential fact for the identification of gossans and a discriminating criterion in typology (Besnus, 1977 ; Wilhelm and Kosakevitch, 1979 ; Pouit, 1987 ; Chauris and Garreau, 1990 ; Blot, 1990, 2002, 2004). Otherwise, this richness is primordial for modern prospections, with the help of numerous multi-elemental analyses, and of mighty computer calculations for data processing. During the weathering processes, this geochemical memory can be carried by residual minerals or by neogenetic ones being protected up to the surface. A particular place will be further dedicated to iron oxides which are traps for metals at every step of the mineral filiations.


There is a very variant contrast between mineralization and gossan displaying relative or residual elements accumulations like what is observed during weathering also characterized by release of some elements and accumulation of others.


In some cases, supergene enrichments develop, which is rather common in the cementation zone. All the elements are affected by these phenomena (table 7) but they are variably affected and no assured model can be advanced.


Wadi Wassat, VMS, 120 Mt, 80 % pyrite, 21,5 ppm Ag, 0,62 ppm Au
Jabal Sayid, VMS, 27-37 Mt, 1,8 % Cu, 1,3 % Zn

Al Masane, VMS, > 5 Mt, 1,5 % Cu, 6,3 % Zn, 93 ppm Ag, 2,8 ppm Au

Ash Sha'ib, Zn-Cu stratiform, 1, 7 Mt, 6 % Zn, 0,3 % Cu Ar

Ridaniyah West, Zn-Pb stratiform, 0,4 Mt, 4 % Zn

Table 7 

Evolution of principal elements between ores and gossans from Arabia (after Ryall and Taylor, 1981)


From geochemical specificities, Andrew, Blain, Butt, Ryall, Smith, Taylor proposed geochemical classifications of gossans in a genetic purpose, for diverse mineralizations in different regions of Australia (table 8), South Africa and Arabia (table 9).


1 - Nickel gossans, 2 - Nickel gossans siliceous, 3 - Nickel gossans siliceous, 4 - Nickel gossans, 5 - Probable Nickel gossans, 6 - Gossanous Tuffs in UM, 7 - Lateritized UM, 8 - Fe Sulphide Gossans in UM, 9 - Copper-Zinc Gossans, 10 - Copper-Zinc Gossans, 11 - Lead-Copper Gossans, 12 - Zinc-Copper Gossans, 13 - Stratiform Fe Sulphide Gossans, 14 - Copper Gossans (Vein type), 15 - Pyrite Gossans (Vein type), 16 - Ferricretes

Table 8 

Geochemistry of Australian gossans characteristic of the different mineralization types (Blain and Andrews, 1977)


Table 9 

Marker elements of the original environment for the different ferruginous formations typical of Arabia (after Ryall and Taylor, 1981)


Besides these geochemical criteria, maturation is an important factor (Scott et al., 2001), but even still more, a classification in a given zone is not the same elsewhere, in other climatic conditions. And it can be immediately emphasized that a tropical region like the whole Africa is lacking in any comparative reference.


The variations result from factors as numerous as those evoked for gossans by Smirnov (1936), i.e. variations of contents from one ore to another, mineralogical composition, different maturation conditions including geometry of the mineralization and climatic conditions s.l.


f Storage of elements by iron oxides


Goethite is the most widespread and the most abundant mineral in gossans in which it is frequently associated with elements inherited from mineralizations : silicium is always present and high contents are frequent for Al, Zn, Cu, Ni, Pb, S, P As. The mechanisms responsible for trapping these elements are complex and the trap form is rarely pointed out. Some deformations in the crystalline lattice can explain the trapping of metals such as aluminum and Al- hematites and Al-goethites are common in the tropical zone.


In the case of basic metals the retention of one or more elements by Fe-Mn oxi-hydroxides can be coupled with other mineralogical forms of this (or these) element(s) (CO3 for Pb and Zn, SO4 for Pb). But iron oxi-hydroxide can be the only supergene trap, as for As at Irécê, whilst sulfophosphates are devoid of this element, or Ni and Cu at Kwademen where there are no other secondary minerals.


In, the example of the filiations in Canoas mine (table 4), it is showed that goethites can originate from different minerals, which foreshows the preservation of the metal released from sulfide, carbonate or silicate replaced by neogenetic goethite. After some examples, Kosakevitch then assumes that there is a relation between the shape and composition of Fe oxi-hydroxide. And he describes pyrite pseudomorphoses enriched in different elements typical of the corresponding ore deposit, for example :

  • At Bois Madame (Hérault, France), iron oxides contain up to 10% ZnO, 2,5-5% PbO, 2-3% Al2O3, 0,3% MgO, 0,5-1% CaO.
  • At l’Argentella (Corse, France), they also contain U.5-2% Pb0, 1% Cu0, 1-1.5% SO3, O.1-1% As2O3.
  • Chalcopyrite pseudomorphoses contain between 5 and 20% CuO, whilst in associated pyrite pseudomorphoses the content is only 2-6%.

At Pagala (Togo), Togbé (1991) and Togbé et al. (1993, 1994) described pseudomorphoses of siderite into goethite with frequently 1-3% ZnO.


In Brazil, different families of goethites were discriminated from their metallic or anionic contents, such as Pb-goethites, Zn-goethites, Pb-Zn goethites, Ni-goethites, Ni-Cu goethites. At Canoas, these families as well as little crystallized forms permit to identify on outcrops sterile goethites as well as Zn-goethites, Pb-goethites or Zn-Pb goethites (Fig. 4). The fixation of these elements could not be explained up to now, but what is sure is that no modification of the crystal lattice has been observed.


Fig. 4 

Zn and Pb contents of goethites from Canoas (Brazil) (n=257).

Zn = ZnO / Fe2O3 ; Pb = PbO / Fe2O3


At the O’Toole mine, as well as on the ferruginous blocks of the Bonga lateritic nickel ore deposit (Burkina), goethite traps nickel or copper or both (Fig. 5), but these small-sized metals are included in the crystalline lattice of goethite (Imbernon, 1998 ; Lavaud, 2002 ; Lavaud et al., 2004).


Fig. 5 

Ni and Cu contents of goethites from O’Toole (Brazil) in atomes % (n = 134)


The mineral filler of goethite solely provides the goechemical reply of gossans – which is obvious since gossans do not contain any other secondary minerals – especially as for Ni-Cu-Pt mineralizations at Yilgarn in Australia and O’Toole in Brazil (Fig. 6) or in the case of extreme maturation.


Fig. 6 

Ni-Cu-Zn diagram of goethites and gossans from O’Toole (MG, Brazil)