3. Basic aspects of skarn deposits
This chapter is divided into two paragraphs in order to provide useful informa6on concerning skarn deposits. The first sec6on summarizes some of the historical papers that are the scien6fic basis for the actual gene6c model for skarn deposits. The 6tle “Skarn deposits: Historical bibliography through 2013” is clearly influenced by Burt (1982) that shows as “many observa6ons and ideas in recent publica6ons are actually 50 or even 100 years old”. Another purpose of this chapter is to show the scien6fic relevance of Campiglia MariPma skarn deposit on the actual skarn forma6on model. The second sec6on describes the main features of the actual gene6c model for skarn forma6on and ore deposi6on.
3.1. Skarn deposits: Historical bibliography through 2013
1847 -‐ Descrip6on of garnet masses between diorite and carbona6c rocks at Turjinsk copper mine, Urals (Beamount, 1847). The garnet is produced by the ac6on of the diorite on the limestone and then compares these deposits with similar deposits in Tuscany.
1864 -‐ First descrip6on of contact deposit. Von Co[a proposes that the garnet rocks “are probably, for the most part, the results of the combina6on of the lime in the limestone with the silicates of the Bana6tes, by mel6ng under a high pressure, and subsequent cooling-‐off in enclosed places”. In this paper is depicted an image of Cava del Piombo (Miniera dei Lanzi, Campiglia MariPma) (Von Co[a, 1864).
1868 -‐ First descrip6on on a mineralogical zoning in a metasoma6c rock. The zoning descripted was: magma6c rock ⇒ ilvaite ⇒ hedenbergite ⇒ marble host-‐rock (for more details see chapter 7) (Vom Rath, 1868).
1875 -‐Törnebohm used for the first 6me the term skarn (grönskarn to indicate pyroxene-‐ garnet rocks) (Törnebohm, 1875).
1895 -‐ One of the first examples in the United States. Packard proposed a gene6c rela6on between igneous rock and copper deposits at Seven Devils, Idaho (Packard, 1895).
1900 -‐ Beck cited the examples of Campiglia MariPma and Elba Island as the “contact-‐ metamorphic ore deposits” (Beck, 1900).
1902 -‐ “The genesis of the contact-‐deposits of the Kris6ana type thus seems to be due to the aqueous gas above the cri6cal temperature, which was more or less laden with metallic compounds, and, under heavy pressure, penetrated the limestone adjacent to the igneous intrusive body”. The origin of metals derived from the cooling of the magma6c body and not from the carbona6c rocks (Lindgren, 1902).
1903 -‐ First classifica6ons of contact deposits on the basis of contained ores (Weed, 1903).
1905 -‐ Lindgren recognized the possible role of halides in transpor6ng metals (Lindgren, 1905).
1907 -‐ The “endometamorphism” of intrusive rocks develops by the addi6on of Ca from limestone. The skarns could not form by direct reac6on between magma and limestone, because the magma6c rock has too low Fe content (Kemp, 1907).
1911 -‐ Goldschmidt generalized the use of the term skarn (Goldschmidt, 1911).
1912 -‐ The volume does not change during metasoma6c processes (aier know with the name of Lindgren’s Law) (Lindgren, 1912).
1917 -‐ Kato concluded, based on the Japanese skarn, that “the greater part of the elements composing the lime-‐silicate minerals and the en6rety of ore-‐minerals ... have been derived from the emana6ons from the magma”. Furthermore, he proposed a temporal model for the forma6on of skarn minerals with the “lime-‐silicates containing li[le or no iron precedes the forma6on of those skarn minerals rich in iron” and sulfides are s6ll later or, in part, contemporaneous (Kato, 1917).
1932 -‐ Eskola explained the “principle of enrichment in the stablest cons6tuents”. During mass transfer (metasoma6sm), the most soluble cons6tuents, such as lime and alkalies, are extracted and the least soluble, such as alumina and magnesia, are lei behind, as in aluminosilicate rocks and magnesian skarns (Eskola, 1932).
1936 -‐ Magnusson divided the skarn in reac6on skarn (formed during regional metamorphism) and in metasoma6c skarn (all the others). The term skarn has been
extended also for ore minerals (Magnusson, 1936).
1936 -‐ First defini6on of “perfect mobility of components” during metasoma6sm (Korzhinskii, 1936).
1942 -‐ Knopf proposed a strong structural control on mineraliza6on (Knopf, 1942).
1945 -‐ Korzhinskii explained the metasoma6c zoning with the concept of mobile components previously proposed (Korzhinskii, 1945). The skarns are formed as the result of interac6on of three media: limestone, silicate rocks, and postmagma(c (italic from Burt, 1982) solu6on. Chemical poten6al diagrams are used to explain the zoning observed.
1965 -‐ Theory of systems with perfectly mobile components and processes of mineral forma6on (Korzhinskii, 1965).
This paper proposed two principals items for understanding skarn deposits that strongly influenced the modern skarn forma6on model. The first item is the behavior of an element during a metamorphic and/or metasoma6c processes that can be only “inert” or “perfectly mobile”. In metasoma6c processes, two limi6ng cases are to be dis6nguished: 1_ infiltra6on; 2_ diffusion. In infiltra6on metasoma6sm, “the rate of percola6on being constant, every front of replacement moves with a constant speed, and consequently the zones of replacement grow uniformly in the direc6on of the flow, by the replacement of forward zones by those behind” similar to a “laboratory chromatographic column”. In diffusion metasoma6sm the diffusion of components proceed through a stagnant pore solu6on. “Because the concentra6on of the pore solu6on changes within the limits of every zone, the composi6on of a mineral solid solu6on must change as well”. This is important criterion for dis6nguishing diffusion from infiltra6on processes in which “a gradual change of the mineral composi6on cannot take place”.
The second item inspect the interac6on between “bases” and “acids” with the formula6on of the principle of acid-‐base interac6on. “Increasing the acidity of a melt or an acqueous solu6on decreases the bulk ac6vity coefficients of all the bases and increases those of all the acids”. “An increase in the basicity causes an opposite effect”. These phenomena explained for example because “carbonates and other basic rocks commonly cause the precipita6on of the ore minerals from hydrothermal solu6ons” (Korzhinskii, 1965).
1966 -‐ Skarn deposits are produced by postmagma6c hydrothermal fluids similar to those that produce other types of ore deposits. A limestone rock near a pluton is the main factor
for skarn development (Buseck, 1966).
1968 -‐ Korzhinskii explained “The Theory of Metasoma6c Zoning” star6ng from the Lindgren’s volume Law and from the two end-‐members in metasoma6c processes (diffusion vs infiltra6on) (Korzhinskii, 1968).
1968 -‐ Zharikov described a rela6on between chemical composi6on of skarn minerals (Fe/ Mg ra6o in pyroxene and Fe/Al ra6o in garnet) and the metals recovered. He a[ributes this to varia6on in the acidity of skarn-‐forming fluids (Zharikov, 1968).
1969 -‐ Perry studied calcic and magnesian skarn near porphyry copper at Christmas mine, Arizona. He found a volume loss during skarn forma6on (Perry, 1969).
1970 -‐ Bartholomé cannot exclude the possibility that much iron could be leached from earlier crystallized rocks (Bartholomé, 1970).
1970 -‐ Bartholomé & Evrard analyzed the Temperino mine skarn and they suggested that ilvaite-‐bearing skarn are formed under higher values of f(CO2) than the more common skarn
consis6ng of an inner andradite and outer hedenbergite zones (Bartholomé & Evrard, 1970). 1974 -‐ Burt described the metasoma6c zoning sequences commonly found in Ca-‐Fe-‐Si exoskarns on observa6ons also of italian skarns. In agreement with Korzhinskii, he explained the forma6on of skarn as the result of the chemical poten6al between host-‐rocks and skarn-‐ forming solu6on. The relevance of the end-‐member model (diffusion vs infiltra6on) is difficult to assess (Burt, 1974).
1974 -‐ The C, H, O and S isotope were used to understand the genesis of Darwin mine, California (Rye et al., 1974).
1977 -‐ First review paper on skarn deposits (Burt, 1977). Burt proposed different skarn classifica6on based on the dimension, structure, host-‐rock and ore metals. Described the forma6on of a typical skarn in 5 different stage. 1) Intrusion of an intermediate to grani6c magma probably at shallow depth. 2) Contact metamorphism (dehydrata6on and decarbona6on) of the country rocks, and results in a volume decrease (ground prepara6on). Crystalliza6on of the intrusive proceeds to comple6on. 3) Early anhydrous zoned skarn
forma6on due to release of Fe and Si rich fluids from the magma, or to the arrival of fluids from a deeper source. 4) Metalliferous ore deposi6on. Ore deposi6on is confined to earlier-‐ formed skarn, some of which remains barren. 5) Late hydrothermal altera6on with destruc6on of early anhydrous skarn minerals and con6nued ore deposi6on (for example ilvaite replace garnet and clinopyroxene). During stages 4 and 5 above the ore-‐bearing fluid presumably changes from magma6c to convec6ve meteoric. He described a systema6c chemical composi6on varia6on in the skarn minerals as a func6on of the distance from an intrusion. Burt cited also the “classic exoskarn zoning sequence” of Temperino mine (Campiglia MariPma) (Burt, 1977).
1981 -‐ Another review paper on skarn deposits (Einaudi et al., 1981). In this paper are discussed the classifica6on of skarn deposits, the forma6on stages of different skarn types and the chemical characteris6cs of the skarn minerals. Furthermore, the authors summarized the new data on fluid inclusion studies and mineral equilibria for different skarn deposits (Einaudi et al., 1981).
1986 -‐ Mathema6cal trea6ses and numerical simula6ons of a typical self-‐organiza6on phenomena in the skarn: the fingering instabili6es (Chadam et al., 1986). They demonstrated that a fluid flow with planar front can spontaneously development in elongated finger with a favoured amplitude and shape.
1987 -‐ Geochemical self-‐organiza6on in reac6on-‐transport systems (Ortoleva et al., 1987). The genera6on of a self-‐organiza6on pa[ern comes about when a given feedback is triggered by noise and driven by disequilibrium. The authors proposed different examples, two of them regard skarn deposits (the reac6ve infiltra6on instability and the Ostwald-‐ Liesegang cycle). The first process cause the development of finger or scallop, while the second can produce banded skarn (Ortoleva et al., 1987).
1992 -‐ Review paper that resumed the typologies, the chemical characteris6cs and the mechanism of forma6on for skarn deposits (Meinert, 1992).
1995 -‐ Systema6c major-‐ and trace-‐element varia6ons in plutons associated with different skarn types. This paper suggest that the composi6on and petrologic evolu6on of a magma are the primary controls on skarn type mineraliza6on and metal content (Meinert, 1995).
1995 -‐ During skarn forma6on the external zone at the contact with skarn and the carbona6c host-‐rock is characterized by the reac6on of decarbona6on causing increase in permeability with the movement of the fluids parallel to the contact. The so-‐called metasoma6c fronts are parallel to the direc6on of flow, and so are be[er termed metasoma6c sides (Yardley & Lloyd, 1995).
1999 -‐ LA-‐ICP-‐MS analysis of fluid inclusions from a range of magma6c-‐hydrothermal ore deposits (Heinrich et al., 1999). These data showed different behavior between Na, K, Fe, Mn, Zn, Rb, Cs, Ag, Sn, Pb and Tl that preferen6ally par66on into the brine (probably as Cl complexes) whereas Cu, As, Au (probably as S complexes) and B selec6vely par66on into the vapor phase.
1999 -‐ The processes for skarn development can also generate peculiar carbona6te melts (Lentz, 1999).
2005 -‐ Another review paper resumed the typologies, the chemical characteris6cs and the mechanism of forma6on of skarn deposits (Meinert et al., 2005).
2005 -‐ Experimental studies performed at the vapor-‐liquid equilibrium curve of the system H2O-‐NaCl±HCl with add of chlorides (FeCl2, ZnCl, CuCl, AgCl) and oxides (As2O3, Sb2O3
and SiO2). The results obtained were compared to natural examples (Pokrovski et al., 2005).
In par6cular, the data reported for natural examples porphyry Cu and skarn deposits (see e.g., Audétat et al., 2008; Heinrich et al., 1999; Baker, 2004) implies much higher concentra6ons of sulfur in the vapor phase than in the coexis6ng liquid.
2005 -‐ Review paper on metal transport in hydrothermal systems (Williams-‐Jones & Heinrich, 2005).
2006 -‐ Recogni6on on fractal geometries for two different rhythmic-‐banded metasoma6c rocks (Rusinov et al., 2006). On the basis of fractal analysis the authors dis6nguished diffusion metasoma6c rock (quartz-‐gold-‐silver veins; Dukat deposit, Russian Northeast) from infiltra6on metasoma6c rock (wollastonite-‐hedenbergite skarn; Dal’negorsk deposit, Primorye) (Rusinov et al., 2006).
2007 -‐ The depth of pluton emplacement plays a key role for the forma6on of a skarn deposit. The authors, on the basis of Japanese ore deposits, concluded that Pb-‐Zn and Mo deposits are related to intrusion emplaced at pressure below 1 kbar, Cu-‐Fe and Sn at 1-‐2 kbar, W deposits at 2-‐3 kbar while barren skarn are related to pluton emplaced at pressure above 3 kbar (Uchida et al., 2007).
2008 -‐ Experimental study on synthe6c fluid inclusion of the behaviour of Cu and Zn under boiling condi6ons (550-‐650°C; 35-‐100 MPa) in sulfur-‐bearing and sulfur-‐free systems (Nagaseki & Hayashi, 2008). In sulfur-‐bearing systems, Cu preferen6ally par66ons into the vapour phase, whereas Zn s6ll prefer the liquid phase.
2008 -‐ Theore6cal model for explain the development of rhythmically banded skarn and the rela6onships with the natural example of wollastonite-‐hedenbergite banded skarn (Dal’negorsk Deposit, Primorye) (Rusinov & Zhukov, 2008).
2008 -‐ Fluid inclusion study in garnet, pyroxene and sphalerite in the distal skarn of El Mochito, Honduras (Samson et al., 2008). This paper indicated that the skarn-‐forming fluids, also for distal skarn, have magma6c signature and do not represent a mixture with basinal brine. The Zn and Pb contents in the deposits have similar ra6os that in the fluids.
3.2. The current skarn forma@on model
Classifica@on
Skarn deposits are studied since 1847 (Beamount, 1847) and several classifica6ons have been proposed. The most widely used classifica6on is based on ore type (for review see Meinert et al., 2005). The term Pb-‐Zn skarn (Einaudi et al., 1981) instead of Zn skarn (Meinert et al., 2005) is more used in scien6fic li[erature.
Stages of skarn forma@on
The forma6on of a typical skarn deposit related to a magma6c intrusion involve subsequent (but probably con6nuos) stages.
1.The first stage is the intrusion of magma (typically with temperature of 900-‐700°C but possibly up to >1200°C). The intrusion causes metamorphism in the host rocks with extension and temperature depending on the depht of pluton emplacement (Burt, 1977; Meinert et al., 2005; Fig. 8 A).
2.During crystalliza6on the magma6c body releases hydrothermal fluids. Usually, the skarns develop at the expense of carbona6c rocks (Fig. 8 B). The pluton emplacement depth controls the size and morphology of skarn bodies. The deeper skarns are small and ver6cally oriented compared to the shallower ones that are laterally extensive (Fig. 8 C). Structural sePng can cause varia6ons from this idealised model. These features are influenced by the different behaviour (duc6le vs bri[le) of carbonate rocks at depth. The early anhydrous skarn minerals are garnet and pyroxene with subordinate and variable amounts of wollastonite, olivine, and other phases. During the prograde skarn forma6on the temperature can exceed 700°C except for Cu and Pb-‐Zn skarns (Sn and W skarn: Fluid Inclusion (FI) Th > 700 °C, Kwak & Tan, 1981; Cu, Pb-‐Zn skarns: FI typical 300°C < Th <
550°C; Meinert et al., 2005 and references therein) with salinity up to 50 wt% NaCl equivalent. The ore deposi6on starts during this stage (scheelite and oxides earlier than sulfides). Distal skarn can be produced by hydrothermal fluids from deeper source (e.g., Burt, 1977; Samson et al., 2008).
3.The cooling of the pluton with circula6on of meteoric waters can cause retrograde altera6on with overprints of hydrated phases on previously formed anhydrous skarn minerals. The most common phases in this stage are epidote, amphibole, chlorite, ilvaite, carbonates, etc. Fluid inclusion studies on retrograde minerals show temperature and salinity significantly lower than during prograde skarn stage (several hundred degrees with
salinity < 25 wt% NaCl equivalent; Meinert et al., 2005). Also in this last phase the depth is an important parameter because the retrograde altera6on (Fig. 8 D) is more extensive in shallow environments. In this stage the deposi6on of ore minerals comes to end (Burt, 1977; Einaudi et al., 1981; Ray & Webster, 1991; Meinert, 1992; Meinert et al., 2005; Fig. 8).
Fig. 8: Evolu(onary stages A ➝ B ➝ C ➝ D of pluton-‐associated skarn deposits (Meinert et al., 2005). A: Metamorphism of host-‐rock during emplacement of magma(c body; B: Different types of skarn at the contact between two lithologies; C: Morphologies and size of skarn bodies; D: Late hydrothermal altera(on on
Fluid inclusion composi6on (e.g., KCl/CaCl2, KCl/NaCl) and isotopic data (e.g., C, O, H, S)
are inherently powerful tracers of skarn-‐forming processes (e.g., Rye et al., 1974; Taylor & O'Neil, 1977; Kwak & Tan, 1981, Bowman, 1998; Lentz, 1998; Meinert et al., 2005; Samson et
al., 2008). These data suggest that several skarn deposits form from diverse fluids sources
(from magma6c to meteoric to typical sedimentary values; Meinert et al., 2005 and references therein).
Magma@c rocks related to skarn deposits
Meinert (1992; 1995) has recognised various chemical composi6on in plutons associated with different skarn types iden6fying systema6c major-‐ and trace-‐element varia6ons in plutons associated with Fe, Au, Cu, Pb-‐Zn, W, Mo and Sn skarns (Fig. 9).
Fig. 9: Correla(on between composi(on of plutonic rocks and associated skarn deposit types (modified aLer Meinert, 1992)
Sn skarns are associated with the most silica-‐rich igneous rocks, thus having the strongest crustal signature. Similar features are shown also by Mo and W skarn deposits that are typical exploited for both elements. Iron, Cu, and Au skarns are gene6cally associated with less evolved magma6c rocks. Gold-‐skarn plutons are significantly more reduced than typical Fe-‐ and Cu-‐skarn plutons. The Pb-‐Zn skarn-‐related plutons are intermediate between the Fe-‐ Cu-‐Au group and the Sn-‐W-‐Mo group. The Pb-‐Zn skarn are commonly distal is respect to their causa6ve pluton (Meinert, 1995). Other authors (e.g., Uchida et al., 2007) pointed out the importance of the depth of pluton emplacement for style of ore deposi6on (Fig. 10).
Fig. 10: Al2O3/(CaO+Na2O+K2O) molar ra(o of grani(c rocks vs total Al content in bio(te and types of skarn deposits (modified aLer Uchida et al., 2007). Note: The Al content in bio(te have direct posi(ve correla(on with
the pluton emplacement depth (Uchida et al., 2007).
The metasoma@c column model
Several authors have discussed the mechanism of fluid transport during metasoma6c processes (infiltra6on vs diffusion; e.g., Korzhinskii, 1965, 1968; Chadam et al., 1986; Ortoleva et al., 1987; Rusinov et al., 2006). In recent years, the forma6on of skarn has been considered a process related to the infiltra6on of metasoma6c fluids (Meinert et al., 2005 and references therein) but the debate is s6ll open (Burt, 1977; Rusinov & Zhukov, 2008).
The model of infiltra6on (Fig. 11) explains skarn mineral zoning along the fluid path in agreement with “The Theory of Metasoma6c Zoning” (Korzhinskii, 1968 and references therein). A single fluid-‐flow event can produce mul6ple propaga6ng reac6on fronts, each traveling at different velocity from the fluid source. This process causes an increase in the separa6on of different fronts along the fluid path. The skarn mineral zoning reflects the rela6ve mobility of Si > Fe, Mg and Mn > Al. The inner front (slower traveling reac6on) replaces the external previously formed skarn zone (fastest traveling reac6on). In the peripheral skarn zone at the contact with the carbonate host-‐rock a transient secondary porosity develops and the metasoma6c fluid flows orthogonal to the contact (metasoma6c sides; Yardley & Lloyd, 1995).
Increasing depth of pluton emplacement
1 kbar 2 kbar 3 kbar
Fig. 11: Schema(c illustra(on of the propaga(on of mul(ple reac(on fronts during progressive fluid flow (Meinert et al., 2005).
This mechanism focuses the fluid flow and accelerates reac6ons (reac6ve infiltra6on instability; Ortoleva et al., 1987) and results in fingered (or scalloped) reac6on fronts. This transient aquifer may focus fluid flow even more efficiently into the 6ps of the skarn fingers (or scallops), amplifying the lobate nature of skarn deposits (Meinert et al., 2005).
Transport of metals
The phisico-‐chemical characteris6cs of fluids are cri6cal factors for our understanding of the transport mechanisms genera6ng ore deposits. In the last years the composi6on of natural and synthe6c fluid inclusions were analysed and experimental studies were performed in different magma6c-‐hydrothermal condi6ons (e.g., Drummond & Ohmoto, 1985; Hemley et al., 1986; Hedenquist & Lowenstern, 1994; Audétat et al., 1998; Heinrich et
al., 1999; Ulrich et al., 1999; Baker, 2004; Pokrovski et al., 2005; Williams-‐Jones & Heinrich,
2005; Audétat et al., 2008; Nagaseki & Hayashi, 2008; Pokrovski et al., 2008; Samson et al., 2008; Pokrovski et al., 2013).
The fluids exsolved early from magma6c body can be immiscible brine and vapor or intermediate density fluids (depend of the emplacement depth) with complex chemical composi6on. The most abundant component is H2O with Cl-‐ and S-‐complexes and CO2 as is
shown for the presence of many daughter minerals in fluid inclusions (e.g., halite, sylvite, pyrite, chalcopyrite; Williams-‐Jones & Heinrich, 2005 and references therein; Fig. 12).
Fig. 12: Example of fluid immiscibility illustrated by the model system NaCl + H2O (Heinrich et al., 2004).
Several data exist from porphyry deposits but unfortunately, no complementary data exist for skarn deposits (Baker et al., 2004). For this reason the behaviour of metals and metalloids in skarn-‐forming fluids is discussed. In general, these elements (e.g., Fe, Mn, Pb, Zn) are enriched in the liquid/brine phase and show strongly posi6ve correla6on with Cl concentra6on. Cl and in par6cular S seem to have strong influence on the par66oning of Cu, Au and minor Ag in the vapor phase (e.g., Audétat et al., 1998; Heinrich et al., 1999; Baker et
al., 2004; Pokrovski et al., 2005; Williams-‐Jones & Heinrich, 2005; Audétat et al., 2008;
Nagaseki & Hayashi, 2008; Pokrovski et al., 2008; Samson et al., 2008, Pokrovski et al., 2013; Fig. 13).
Fig. 13: LeL: Par(oning of metals between vapor and coexis(ng brine (Audétat et al., 1998). Right: Comparison between esperimental KD values for Cu and Zn (both S-‐bearing and S-‐free experiments) with natural