Solubility and speciation of carbon in silicate and carbonate melts

Get Complete Project Material File(s) Now! »

Short-term carbon cycle

Short-term carbon cycle involves fluxes between the atmosphere, biosphere and hydrosphere, and may be divided into terrestrial and oceanic carbon sub-cycles. The carbon uptake from the atmosphere in the terrestrial ecosystem is regulated by photosynthesis, and carbon release into the atmosphere is regulated by respiration of living organisms (both autotrophic and heterotrophic) and microbial decomposition of dead organic matter (Fig. 1.4). The total carbon uptake by plants through photosynthesis is about 120 Gt yr-1, but after excluding both autotrophic (60 Gt C yr-1) and heterotrophic respiration (60 Gt C yr-1), as well as natural disturbances (such as fires and drought), net biome production is known to be near the zero in 1980s (0.2±0.7 Gt C yr-1) and positive in 2000s (2.2±0.5 Gt C yr-1). In a summary report of IPCC (Intergovernmental Panel on Climate Change, 2000) net biome production was estimated at 0.7±1.0 Gt C yr-1. Part of the net biome production is released from the biosphere into the lithosphere (burial). This process was estimated to be at the rate of 0.06±0.03 Gt C yr-1 (Berner, 2003). Thus, the current carbon burial process is three orders of magnitude slower than the short-term carbon exchange between terrestrial biosphere and atmosphere, and two orders of magnitude slower than the storage of carbon in net biome production. Additionally, the burial rate is one order of magnitude smaller than reasonable error bars of the land-atmosphere flux rate estimation (±0.5-0.7 Gt C yr-1) (Table 1.4).
Fig. 1.4. Contemporary short-term carbon cycle, NASA, 2001.
In seawater, carbon is present in various forms, namely dissolved inorganic carbon (DIC: HCO3- and CO32-), dissolved organic carbon (DOC), and particulate organic carbon (POC), with an approximate ratio DIC:DOC:POC of 2000:38:1 (Denman et al., 2007). Heinze et al. (1991) identified three key ocean carbon pumps that define possible ways for the atmospheric CO2 to exchange with the oceanic pool: solubility pump, organic carbon pump, and CaCO3 ‘counter pump’. Alkaline surface seawater (pH between 7.5 and 8.4) absorbs vast amounts of atmospheric CO2 due to its buffering capacity to neutralize the acidity of CO2 (reactions 1.1, 1.2).
Table 1.4. Global carbon budget (Gt C yr-1) in the context of short-term fluxes of carbon cycle (error bars were estimated in this work from Prentice et al., 2001, Achard et al., 2002, DeFries et al., 2002, Houghton, 2003, Le Quéré et al., 2003, Sabine et al., 2004, Denman et al., 2007).
Bicarbonate-rich oceanic water is generally found at high latitudes during the cold season because of higher solubility of carbonates in cold water. Cold CO2-rich water sinks from surface layer to the deep ocean via the Meridional Overturning Circulation. This process is referred to as ‘solubility pump’. Oceanic buffering capacities are not infinite and depend on the addition of Ca2+ and Mg2+ ions from weathering of silicate rocks (equation 1.3). Atmospheric CO2 can also be taken up for photosynthesis by the phytoplankton, which afterwards either sinks as dead organic matter or transforms into DOC. Some fraction of DOC is respired by bacteria and is consequently recycled in the surface ocean as DIC. The remaining small-particle fraction reaches the deep ocean; some of the particles are then re-suspended, and some are buried in the sea floor (known as ‘organic carbon pump’). Lots of oceanic organisms form their shells, consuming calcium and carbonate ions. This process is called ‘CaCO3 counter pump’.
The exchange of carbon between the atmosphere and oceans amounts to ~90 Gt C yr-1 in each direction, annually ocean stores about !2.0 Gt C yr-1 (Table 4). For comparison, the flux of carbonate sediments stored in the deep ocean is ~0.2 Gt C yr-1 (long-term carbon cycle).
At the beginning of the 21th century, humans have liberated 7.2±0.3 Gt C yr-1 to the atmosphere due to combustion of fossil fuels. Natural short-term carbon cycle regulating processes are not sufficient to establish equilibrium between CO2 input to the atmosphere enhanced by humans and natural CO2 output. As a result, a constant increase of carbon content in the atmosphere is observed (Table 1.4). Rates of long-term carbon storage mechanisms are one-two orders of magnitude slower than short-term ones. That is why injection of carbon dioxide in basalts (see equation 1.3 of long-term carbon cycle) seems not to be a good way of CO2 storage compared to other ones like enhanced oil recovery, plantation, or storage in deep aquifers.

Sources of carbon in geological fluid

The primary sources of carbon in geological fluids are melts, carbonate rocks, and organic matter. Carbonate-rock carbon pool is 4.5 times bigger than the kerogen one (Table 1.2). Oceanic and meteoric waters may also contribute to crustal fluids, but their carbon concentrations are modest (e.g. 0.002% C in seawater, Fig. 1.1). The essential processes leading to an effective carbon release into geological fluids are magmatic activity and metamorphism (both regional metamorphism and more locally skarns, i.e., interaction of silicate magmas with carbonate rocks). Subduction zones and orogenic belts are the most favorable geological settings for carbon volatile enrichment of crustal fluids. Hydrothermal processes in oceanic crust (alteration of oceanic basalts) also involve carbon from seawater during circulation of fluids in convective cells that operate in seafloor due to a magmatic source of heat. At the same time, all these geological environments with substantial tectonic activity are favorable for ore deposit formation. Therefore, a systematic study of the interactions of carbon volatiles with metals in relevant natural conditions is essential for the progress of economic geology.

Solubility and speciation of carbon in silicate and carbonate melts

Silicate melts are the main agent for transporting carbon from Earth’s interior to the surface. Carbon is dissolved in silicate melts as molecular CO2 or in the form of carbonate groups (CO32-), depending on temperature, pressure, and melt composition (Ni and Keppler, 2013 and references therein). Under strongly reducing conditions CH4 and CO may prevail in C-O-H fluids (Manning et al., 2013). Solubility of CO2 in silicate melts systematically increases with increasing pressure. However, in contrast to other volatiles like water, chloride, and sulfur, CO2 has low solubility in silicate melts and there is no mineral phases capable of retaining it in magmatic rocks. These properties are responsible for early degassing of CO2 from magmas (Lowenstern, 2001). Early degassing of CO2 during magma ascent appears to explain why CO2 is a dominant compound of fluid inclusions in xenoliths of mantle olivine (e.g. Stango, 2011 and references therein). The effect of temperature on CO2 solubility is complex and depends on the type of melt. Furthermore, the direction of temperature effect may be reversed upon pressure change. The solubility of CO2 increases from 1000 ppm to 2000 ppm in a series of basalts with increase of their alkalinity and weakly decreases with an increase of SiO2 content in the calcalkaline magmatic series excluding basalt: from andesite melt (~1600 ppm) to dacite melt (~1300 ppm) and to rhyolite melt (~1000 ppm), with basalt melt having the same CO2 solubility as rhyolite in the same alkaline series (Ni and Keppler, 2013 for a recent review).
Infrared and Raman spectroscopic studies of silicate glasses (e.g., Brey 1976; Fine and Stolper 1985, 1986; Stolper et al. 1987) showed that all CO2 is dissolved as carbonate in basaltic glasses, whereas rhyolite, albite and other silica-rich glasses contain molecular CO2 coexisting with minor amounts of carbonate. In andesite and phonolite glasses, molecular CO2 and carbonate coexist (e.g., Brooker et al. 2001). Depolymerization of the melt, expressed by the increase of quantity of non-bridging oxygen atoms per tetrahedron, favors the formation of carbonate in the glasses at the expense of molecular CO2 according to the reaction:
where “O2-react” is a non-bridging oxygen atom (e.g., Eggler and Rosenhauer, 1978). However, the degree of polymerization is certainly not the only parameter that controls the carbonate/CO2 ratio in glasses. For example, replacing of sodium cations by calcium in melt composition appears to strongly enhance carbonate at the expense of CO2. Increasing temperature shifts the equilibrium of reaction (1.5) towards molecular CO2 and the enthalpy of the reaction (which determines the reaction sign versus temperature) increases with the depolymerization of the melt (Morizet et al., 2001, Nowak et al., 2003).
Carbonatite melts are a relatively rare phenomenon compared to silicate melts. Oldoinyo Lengai in Tanzania is the only active volcano on Earth directly producing alkali-carbonatite lavas (Krafft and Keller, 1989). However, carbonatites, defined as magmatic rocks with >50 wt% of carbonate minerals, are widely spread in the word. There are now 527 recognized carbonatite occurrences, ranging in age from Archean to present (Jones et al., 2013 for a recent review). Carbonatites are located within both continental and oceanic lithosphere, mostly in intraplate settings, in large igneous provinces. This distribution precludes a direct link of carbonatites with mantle plumes and favors a fundamental link to the same underling mantle source of carbon, which is manifested in kimberlites (Woolley and Bailey, 2012).
Carbonatite melts are ionic liquids consisting of carbonate CO32- anions and metal cations that interact principally due to coulombic attraction and are thus very different from silicate melts, which have structures characterized by a polymerized O-Si-O-Si network (Mysen, 1983). As a result, carbonatite melts have low viscosity and the highest known melt capacities for dissolving water and halogens; they show very high solubilities of many elements that are usually rare in silicate magmas (REE, Nb, U, Ta, Cu, P, F, Ba, PGE, Ag, Au; Richardson and Birkett, 1996; Ni and Keppler, 2013). The transformation of carbon from the CO32- triangular to the CO44- tetrahedral group is predicted at lower mantle pressure by quantum-chemical modeling (Boulard et al., 2011).
Another form of carbonate material is that found in kimberlites. Kimberlitic melts, notorious for their capacity of transporting diamonds to the surface great depth, differ in many of their geochemical and mineralogical properties from carbonatites. Kimberlites are devoid of alkalis, and lack the typical association with sodic-potassic igneous rocks (Le Bas, 1981). Kimberlites and carbonatites also differ in their styles of eruption: explosive – for kimberlite pipes, and effusive or extrusive – for carbonatites. Kimberlite magmas are believed to carry high concentrations of volatiles, primarily CO2 and H2O, which help to explain their explosive eruption, high ascent rates, and the carrot-like shape of kimberlite bodies (Mitchell, 1986). For example, initial content of CO2 in magma before eruption of the Udachnaya-East kimberlite pipe was estimated to be 23 mol % (20 wt%) and during the eruption the kimberlite magma lost almost half of the CO2 budget (Shatskiy et al., 2014). For comparison, kimberlites and MORB samples have very different CO2 contents: generally ~0.1-0.2 wt% CO2 for MORBs and up to 15 wt% or more for kimberlites (Guillot et al., 2013).

Speciation of carbon in hydrothermal fluids

Within the wide range of geologically relevant redox conditions, carbon may exhibit the whole variety of chemically allowed valence states (+4, +2, 0, -4). Information on the valence and speciation of carbon in hydrothermal fluids is provided by analyses of fluid inclusions in minerals, sampling of active hydrothermal systems, experimental solubility measurements and spectroscopic experiments done in-situ. Thermodynamic modelling put our empirical knowledge in order. In addition, the slow kinetics of some carbon reactions may also be important to consider for interpreting the compositions of natural fluids.
Analyses of fluid inclusions using in-situ Raman spectroscopy, crush-leach technique coupled with gas chromatography and microthermometric observations reveal that minerals can trap a plethora of different carbon species under a broad range of geological settings: e.g., solid CO2 in diamonds under 5 GPa pressure trapped at 220 to 270 km depth (Schrauder and Navon, 1993), liquid CO2 in olivine-bearing nodules and phenocrysts from basalts (Roedder, 1965), bicarbonate (HCO3-) and carbonate (CO32-) in granite pegmatites (Thomas et al., 2011), CO2–COS in ruby from marble-hosted deposits (Giuliani et al., 2003), CO2-CO in andesine-amphibole veins from mantle peridotite xenoliths (Bergman and Dudessy, 1984), CH4 in olivine from ophiolite complexes (Sachana et al., 2007), CH4 and higher hydrocarbons (C2H4, C3H6) in peralkaline igneous rocks (Salvi and Williams-Jones, 1992; Potter et al., 2013), and even ethanol (C2H5OH) in diamonds from Africa, Brazil and Russia (e.g. Melton and Giardini, 1974; Tomilenko et al., 1995). This diversity of carbon species occurring in fluid inclusions in minerals is illustrated in Fig. 1.5. However, the information from fluid inclusions is mostly qualitative and very scarce, as compared to the large amount and diversity of natural fluids, due to limitations of analytical techniques (e.g., molecular species have generally much stronger Raman signals than ionic ones), rarity of fluid inclusions, and artifacts of their preservation (e.g. Roedder, 1965, 1971). In the great majority of fluid inclusions, CO2 (gas) is the most frequently detected and likely the most abundant C-bearing component.
Thermodynamic modelling, in conjunction with experimental studies at controlled laboratory conditions, can help providing additional quantitative information on the abundance of carbonic species in hydrothermal fluids. Under high temperatures, pressures and typical pH and redox conditions of the Earth’s crust and upper mantle, thermodynamic calculations suggest that carbon dioxide is by far the major carbon form (Pawley et al. 1992; Holloway and Blank 1994; Manning et al. 2013), which is consistent with the observations mentioned above. The distribution of C-species is illustrated in Fig. 1.6, where I calculated the domains of predominance of carbon species in the C-H-O system at 450ûC and 700 bar as a function of pH and oxygen fugacity in the fluid. The boundaries between fields of CO2, HCO3-, CO32- and CH4 correspond to equal concentrations of the corresponding major species. As a result, minor species like C2H6, CO(aq), Ca(HCO3)+, CaCO30, HCOO-, CH3COO- cannot be shown on this type of diagram. Minor species concentrations are reported for several points shown in the Figure: e.g. in equilibrium with graphite (reduced conditions) species such as C2H6 and CO do not exceed 0.02 mol% of total carbon. However, carbon speciation different from that shown in Fig. 1.6 may occur in particular geological environments. For example, species such as NaHCO30 and NaCO3- may dominate in fluids in contact with carbonatite rocks.
At very high pressures (e.g. 5 GPa, 600ºC) this speciation may change. Sverjensky et al. (2014) predicted, using the modified HKF model, a dramatic increase in the thermodynamic stability and abundance of aqueous organic species at the expense of CO2 and CH4, such as the propionate ion (CH3CH2COO-) (Fig. 1.7). However their estimations need in-situ spectroscopic verification because the thermodynamic properties of organic species at high temperatures and pressures are poorly known and are based exclusively on extrapolations from low T-P conditions.
Fig. 1.7. pH-fO2 diagram of carbon speciation in the system C-O-H at 600 °C and 5.0 GPa, according to the modified HKF model predictions. It can be seen that equilibrium stability fields appear for aqueous organic species such as acetic acid and acetate below oxygen fugacity of the QFM mineral buffer (Sverjensky et al., 2014).
Recent experimental studies at subduction zone conditions (Facq et al., 2014) suggest that the relative abundance of carbonate species may also change at high pressure. Facq et al. (2014) performed an in-situ Raman spectroscopic study of calcium carbonate dissolution in a diamond anvil cell at 300-400ûC and pressures extending up to 70 kbar (7 GPa). Their experimental results show an increasing dissociation of carbonic acid with increasing pressure: HCO3- predominates over CO32- at pressures below ~40 kbar, whereas CO32- becomes dominant at higher pressures. In addition, this change favors the formation of ion pairs such as CaHCO3+ and CaCO30aq. No in-situ studies are available, however, in the CO2-CH4 coexistence domain (Fig. 1.6 and 1.7), which is the only direct way to reveal carbon species with intermediate oxidation states (like those in Fig. 1.7).
It should be noted, however, that reactions between carbon species of different redox states like CO2 and CH4 are very slow (Ohmoto and Goldhaber, 1997), as illustrated in Fig. 1.8. Therefore, attainment of thermodynamic equilibrium requires very long durations, of the order of 10 000s years, at moderate temperatures (<300ûC). This issue should be taken into account while interpreting carbon speciation in geothermal fluids (Giggenbach, 1982) and during laboratory experiments. In contrast to redox reactions, protonation and ion pairing reactions such as those between carbonate species (e.g. CO2 and HCO3-) are very fast processes reaching equilibrium within seconds to minutes (Martell and Hancock, 1996; Roughton, 1941).
Fig. 1.8. Time (in years) necessary to reach 90% of equilibrium for the ‘sulfate – sulfide’ (solid lines) and ‘carbon dioxide – methane’ (dashed line) pairs in aqueous solution with S and C = 0.01m/kg, as a function of temperature and pH (Ohmoto and Goldhaber, 1997).

Typical concentrations of carbon dioxide in fluid inclusions from ore deposits

The above overview shows that carbon dioxide is a major component of most crustal fluids, including those responsible for ore deposit formation at redox conditions of the Earth’s crust, which is generally at or above QFM. However, the quantification of CO2 contents in fluid inclusions is not routine, and requires detailed microthermometric measurements (e.g., Rusk et al., 2008a), and/or in-situ spectroscopic methods such as Fourier Transform Infrared or Raman spectroscopy, and apropriate calibration procedures (e.g., Wopenka and Pasteris, 1986; Dubessy et al., 1989; Burke et al., 2001; Frezzotti et al., 2012). Another difficulty in identifying the impact of CO2 on the formation of ore deposits is the ambiguity in deposit classification. The Wilson cycle (aggregation/dispersal of continental crust, Fig. 1.2) leads to overprinting of processes and complicates reconstruction of ore deposits formation and fluid origin. Below we overview CO2 contents in different types of ore deposits as they were classified in the original literature. Typical CO2 concentrations for various ore deposits are summarized in Table 1.5.

Table of contents :

Chapitre 1. État de l’art
1.1.Abundance of carbon on Earth
1.2.Carbon cycle
1.2.1. Long-term carbon cycle
1.2.2. Short-term carbon cycle
1.3. Sources of carbon in geological fluids
1.4. Solubility and speciation of carbon in silicate and carbonate melts
1.5. Speciation of carbon in hydrothermal fluids
1.6.Typical concentration of carbon dioxide in fluid inclusions
1.7.Metal concentrations in CO2-rich natural fluids
1.8.Role of carbon dioxide on metal transport by geological fluids
1.8.1. Vapor-Liquid equilibria
1.8.2. Changes in the liquid phase
1.8.3. Direct complexing
1.8.4. Changes in solvent properties
1.9.Goals of this thesis
Chapitre 2. Matériaux et méthodes
2.1. Réacteurs hydrothermaux utilisés
2.1.1. Réacteur à trempe
2.1.2. Réacteur à séparation de phase
2.1.3. Réacteur à cellule flexible (type Coretest)
2.2. Traitement des échantillons expérimentaux
2.2.1. Solutions après trempe
2.2.2. Solutions prélevées (autoclaves à séparation de phase et à cellule flexible)
2.3. Méthodes analytiques pour les solutions aqueuses
2.3.1. Spectrométrie d’émission atomique couplée à un plasma inductif (ICP-AES)
2.3.2. Spectrométrie atomique à la flamme (en absorption et en émission, AAFS et AEFS)
2.3.3. Spectrométrie de masse couplée à un plasma inductif (ICP-MS)
2.3.4. Colorimétrie de la silice par la réduction du complexe silicimolybdate
2.3.5. Titrage des formes chimiques de soufre
2.3.6. Chromatographie à haute performance en phase liquide (HPLC)
2.3.7. Résumé des methodes analytiques
2.4. Préparation et caractérisation des phases solides
2.5. Modélisation thermodynamique
2.5.1. Modèle du solvant mixte H2O-CO2
2.5.2. Principe de calcul d’équilibre chimique
2.5.3. Propriétés thermodynamiques des espèces aqueuses
Chapitre 3. Rôle du CO2 dans les transferts et le fractionnement des métaux d’intérêt économique par des fluides géologiques
3.1. Résumé en français de l’article: «The role of carbon dioxide in the transport and fractionation of metals by geological fluids» Maria A. Kokh, Nikolay N. Akinfiev, Gleb S. Pokrovski, Stefano Salvi and Damien Guillaume soumis à Geochimica and Cosmochimica Acta le 27 Octobre 2015
Chapitre 4. L’effet du CO2 et du soufre sur le fractionnement liquide-vapeur des métaux dans les systèmes hydrothermaux
4.1. Résumé de l’article: «Combined effect of carbon dioxide and sulfur on vapor-liquid partitioning
of metals in hydrothermal systems» Kokh M.A., Lopez M., Gisquet P., Lanzanova A., Candaudap F.,
Besson Ph. and Pokrovski G.S. soumis à Geochimica and Cosmochimica Acta
4.2. Article: «Combined effect of carbon dioxide and sulfur on vapor-liquid partitioning of metals in
hydrothermal systems» Kokh M.A., Lopez M., Gisquet P., Lanzanova A., Candaudap F., Besson Ph.
and Pokrovski G.S. soumis à Geochimica and Cosmochimica Acta
Chapitre 5. Conclusions et perspectives
5.1. Conclusion générale
5.2. Perspectives
Références

GET THE COMPLETE PROJECT