This page focuses on salts precipitated in the higher-temperature subsurface realm, generally flushed by igneous or mantle fluids. Most of the precipitates can be considered hydrothermal salts, which is a broader descriptor than burial salts (Warren 2016; Chapter 8), that encompasses a higher temperature range compared to the diagenetic realm. One group of such hydrothermal salts, mostly composed of anhydrite, with lesser baryte, typically develop along oceanic seafloor ridges within heated subsurface fractures or at seafloor vents. There seawater-derived hydrothermal waters are heating, mixing, degassing, escaping and ultimately cooling. Active deep seafloor hydrothermal hydrologies create a specific group of sulphide ore deposits known as volcanic-hosted massive sulphide deposits (VHMS), with anhydrite as the primary-salt driving mineralisation. Another non-solar hydrothermal salt grouping is salting-out precipitates, mostly halite, created when brines reach supercritical temperatures of 400-500°C. Some proponents of this mechanism postulate this halite sources much of the hydrothermal halite in rifts such as the Red Sea or the Danakhil Depression (Hovland et al., 2006a, b). The third group of hydrothermal salts we shall discuss is hydrothermal gypsum, characterised by the growth of giant gypsum crystals, generally in cavities in a volcanogenic host. For a more detailed discussion see Salty Matters June 16, 2019.
Volcanogenic-hosted massive sulphide deposits are forged by the thermal circulation of seawater through newly-formed oceanic crust, in close temporal association with submarine volcanism and white smokers. This milieu is characterised by active hydrothermal circulation and exhalation of metal sulphides, driven by mantle-induced geothermal gradients in oceanic basalt (Piercey et al., 2015). Being hosted in fractured basalts sets apart VHMS deposits from sedimentary exhalative (SedEx) and most sial-hosted Iron-Oxide-Copper-Gold (IOCG) deposits (Warren, 2016; Chapter 16). Hydrothermal anhydrite crystallises within a matrix of submarine volcanics and volcaniclastics via the heating of fissure-bound seawater (Figure 1a). Anhydrite’s retrograde solubility across a range of salinities means the solubility of anhydrite decreases rapidly with increasing temperature in circulating seawater brines (Figure 1b; Blount and Dickson, 1969).
Hydrothermal anhydrite. A) Formation of hydrothermal anhydrite from heated seawater circulaing and heating in oceanic crust. B) Anhydrite solubility (retrograde) as a function of temperature at 1000 bars in H2O and NaCl-H2O solutions of various concentrations. Dashed lines indicate extrapolated values (after Blount and Dickson, 1969).
Retrograde solubility also explains why anhydrite is most evident in the upper portions of vent mounds and black and white "smokers." Anhydrite's heating response is the opposite of baryte, another typical hydrothermal sulphate precipitate. Simple heating of seawater adjacent to seafloor vents, even without fluid mixing, will precipitate anhydrite, while simple cooling of hydrothermal waters will precipitate baryte. Once buried, hydrothermal calcium sulphate, in the presence of organic matter or hydrocarbons and circulating hydrothermal brines, acts as a sulphur source to create H2S, which then interacts with metal-carrying pore waters to co-precipitate metal sulphides. Thus hydrothermal anhydrite, or more typically indicators of its former presence, are commonplace within volcanogenic hosted massive sulphide (VHMS) deposits. VHMS deposits usually form in submarine depressions as circulating seawater becomes an ore-forming hydrothermal fluid during interaction with the heated upper crustal rocks. Submarine depressions, especially those created by submarine calderas or by large-scale tectonic activity in median ocean-ridge rift valleys, are favourable sites and are often the home of an endemic chemosynthetic vent biota (Holden et al., 2012)
A white smoker is a hydrothermal vent composed mostly of anhydrite and emitting alkaline high-pH hydrothermal fluid on the ocean floor. These fluids are cooler (260–300°C) than those emitted by black smokers (360°C) and are general sited away, or “off-axis,” from the mid-ocean ridges. A notable example is Lost City, a hydrothermal vent field in the central Atlantic Ocean. Interaction of downward seeping seawater with mafic or ultramafic rocks produces an alkaline fluid that when it heats seawater precipitates anhydrite, along with silica and Ba sulphates, hence the white colour.
The interface between alkaline hydrothermal fluids and seawater are believed by some to have provided the conditions required for the emergence of life.
The Champagne vent field at NW Eifuku seamount emits droplets of liquid CO2 from the area around these white-smoker hydrothermal vents composed of anhydrite. Image courtesy of Submarine Ring of Fire 2014 - Ironman, NOAA/PMEL, NSF.
Mineralised seafloor vent chimneys - Temperature controls the white smoker to mineralised vent position while cessation of fluid escape creates an oxidised ferrous vent (ultimately a collapse feature)
A mid-oceanic seafloor ridge region with significant documented volumes of anhydrite is located in the sediment-hosted Grimsey hydrothermal field in the Tjörnes fracture zone on the seafloor, north of Iceland (Figure 3a: Kuhn et al., 2003). There an active fracture zone is located at a ridge jump of 75 km, which caused widespread extension of the oceanic crust in this area. Hydrothermal activity in Grimsey field is spread over a 300 m by 1000 m area, at a water depth of 400 m. Active and inactive anhydrite chimneys up to 3 meters high, and hydrothermal anhydrite mounds, are typical of the seafloor in this area (Figure 3b-f). Clear, metal-depleted shimmering hydrothermal fluids, with temperatures up to 250°C, are venting from active chimneys and fluid inclusion in the precipitated anhydrites show the same homogenisation temperature range (Figure 3g). Anhydrite samples collected from the Grimsey vent field average 21.6 wt.% Ca, 1475 ppm Sr and 3.47 wt.% Mg. The average molar Sr/Ca ratio is 3.3x10-3. Sulphur isotopes from vent anhydrites have typical δ34S seawater values of 22±0.7‰, indicating a seawater source for the SO4. Strontium isotopic ratios average 0.70662±0.00033, suggesting precipitation of anhydrite from a hydrothermal-seawater mixture (Figure 3f). The endmember of the venting hydrothermal fluids, calculated on an Mg-zero basis, contains 59.8 µmol/kg Sr, 13.2 mmol/kg Ca and a 87Sr/86Sr ratio of 0.70634. The average Sr/Ca partition coefficient between the hydrothermal fluids and anhydrite is about 0.67, implying precipitation from a non-evolved fluid. In combination, this suggests anhydrite forms in a zone of mixing between upwelling more deeply-seated hydrothermal fluids and shallowly circulating heated seawater (with a mixing ratio of 40:60). Before and during mixing, seawater is heated to 200-250°C, which drives anhydrite precipitation and the likely formation of an extensive anhydrite-rich zone beneath the seafloor, as in Hokuroko Basin. Once hydrothermal circulation slows or stops on a ridge or mound, and the “in-mound” temperature falls below 150°C, and anhydrite in that region tend to dissolve. During inactive periods, the dissolution leads to the collapse of sulphide chimneys and the internal dissolution of mound anhydrite. Additional ongoing disruption by faulting combine, so driving pervasive internal brecciation of the deposit. Through dissolution, former zones of hydrothermal anhydrite evolve into intervals of enhanced porosity and cavities in the mound. Such intervals initiate further fracture and collapse in the adjacent lithologies, which become permeable pathways during later renewed fluid circulation episodes. The alternating “coming and going” role of hydrothermal anhydrite creating precipitation space within the mound hydrology is similar to that of sedimentary evaporites in the sedimentary mineralising systems (Warren, 2016; Chapter 15).
Grimsey Hydrothermal Vent field, offshore Iceland (after Kuhn et al., 2003). A) Location map B) Massive and acicular anhydrite from a beehive-structured vent chimney. C) Photomicrograph of anhydrite from an active chimney showing radially fibrous crystals. D) Rectangular anhydrite crystals (scale in C and D: 400 µm; crossed Nicols). E) A spongy talc-like material grows on top of euhedral platey anhydrite. All samples (B-E) are from active chimneys (fluid temperatures about 250 °C). F) Schematic drawing showing the main processes related to Sr-Ca geochemistry. Seawater (87Sr/86Sr = 0.709225) entrained in the seafloor at a shallow depth is heated to >150°C leading to anhydrite precipitation (1). The heating of seawater is either conductively or caused by mixing with upwelling hydrothermal fluid. It is assumed that the latter has equilibrated with sediment or underlying basalt resulting in 87Sr/86Sr = 0.702914 - 0.704512. The temperature of the ascending hydrothermal solution is controlled by phase separation and cannot exceed 250°C at shallow depth. The mixed fluid being further conductively heated to 250°C has 87Sr/86Sr = 0.70634 (2). This fluid rapidly rises to the seafloor (3) and precipitates anhydrite at and beneath the seafloor (4). The circulation portion of this schematic is not to scale. G) Homogenization temperatures of fluid inclusions in anhydrite. Most of the temperatures range between 200 and 280°C and therefore, are consistent with the measured temperature of the venting fluids of 248–250°
To form VHMS deposits on the seafloor, the through-flushing hydrothermal fluids must transport sufficient amounts of metals and reduced sulphur, each at concentration levels > 1 ppm (Ohmoto, 1996). For a hydrothermal fluid with the salinity of normal seawater (≈0.7m ∑Cl) to be capable of transporting this amount of Cu and other base metals, it must be heated to temperatures > 300°C. Fluids with temperatures above 300°C will boil at pressures >200 bars. Under such conditions, the resulting vapour cannot carry sufficient quantities of metals to form a VHMS deposit. Boiling of a metalliferous hydrothermal brine outflow is prevented when the fluid vents into water that is deep enough to generate sufficient confining pressure. At 350°C, a minimum seawater depth of 1550m is necessary to prevent boiling. If the fluid passes through a sedimentary package where it loses temperature and metals (Cu, Ba) before emanating, the water depth beneath which boiling is prevented is less (≈1375m). Once vented, the turbulent mixing of hot hydrothermal waters with cooler seawater causes rapid precipitation of sulphides and calcium and barium sulphate, which produces the familiar black and white smokers (Blum and Puchelt, 1991). In modern oxic oceans, the sulphide-rich hydrothermal mounds are rapidly destroyed after the cessation of the hydrothermal activity (Herzig and Hannington, 1995; Tornos et al., 2015)). When hydrothermal activity at a mound decreases and the hydrothermal fluids cool to below 150 °C, the previously formed vent anhydrite is dissolved (retrograde solubility). This near-surface cooling contributes to the dissolution collapse of the anhydrite supported mound surface, particularly at the mound flanks, and allows the additional influx of cold seawater. As mound flank collapse expands the remaining detrital pyritic sand residues are replaced by oxyhydroxides, and copper sulphides tend to be oxidised and replaced by atacamite (Knott et al., 1998). If seafloor weathering continues to completion, all the metal sulphides become oxidized or dissolved. Only those metalliferous VHMS deposits capped by impermeable volcanic, volcaniclastic, or sedimentary deposits soon after formation are preserved due to shielding from the oxidising conditions at the deep seafloor.
Subsurface brines at supercritical temperatures, especially in the buried hot portions of thermally-active rift basement. Two recent papers Hovland et al. (2018a,b) summarise much of this earlier material and add the notion of serpentinization being a sink for chloride and a driver of halite formation in many evaporite basins. Countering arguments to the notion of a non-evaporite origin for substantial volumes of halite in sedimentary basins are given in Talbot (2008) and Aftabi and Atapour (2018). The notion of the importance of the Wilson cycle in a sedimentary evaporite (megahalite and megasulphate basins) context, rather than a direct igneous-metamorphic source as argued by Hovland, is summarised in Warren (2016, Chapter 5). A Hovland model of a non-evaporite source of halite relies on heated subsurface brines becoming supercritical and so transforming a brine to a fluid that does not dissolve but precipitates salt (within specific temperature and pressure ranges).
A supercritical fluid is defined as any substance at a temperature and pressure above its critical point. In such a state, it can effuse through solids like a gas, and dissolve materials like a liquid. In addition, close to the critical point, small changes in pressure or temperature result in substantial changes in density. The critical point (CP), also called a critical state, specifies the conditions (temperature, pressure and sometimes composition) at which a phase boundary ceases to exist. At particular pressure/temperature conditions, supercritical water is unable to dissolve/retain common sea salts in solution (Josephson, 1982; Bischoff and Pitzer, 1989; Simoneit 1994; Hovland et al., 2006a). When seawater brines are heated in pressure cells in the laboratory, they pass into the supercritical region at a temperature of 405°C and 300 bar pressure (the CP of seawater). A particulate ‘cloud’ then forms via the onset of ‘shock crystallization’ of NaCl and Na2SO4 (Figure a). The sudden phase transition occurs as the solubility of the previously dissolved salts declines to near-zero, across a temperature range of only a few degrees, and is associated with a substantial lowering of density (Figure b). The resulting solids in the “cloud” consist of amorphous microscopic NaCl and Na2SO4 particles with sizes between 10 and 100 mm. The resultant “salting out” can lead to the precipitation of large volumes of subsurface salts in fractures and fissures and perhaps even in the deeper portions of salt structures.
Hydrothermal halite derived from supercritical seawater. A) P-T projection of the monovariant solid-liquid-vapour saturation curve (solidus) for the NaCl-water system. Arrows indicate the two points of intersection with the section at 250 bar (defining the lower and upper solidus boundaries). B) Density of water and brine as a function of temperature along the 300 bar isobar. The two-phase region (or “out- salting region”) is indicated by the shaded region with onset indicated by a drastic fall in density over a narrow temperature range. C) Ionic and hydrocarbon solubility in heated water at pressures of 200-300 bars. (A-B after Hovland et al., 2006a, b; C after Josephson, 1982; Simoneit, 1994).
The same supercritical conditions improve the ability of brines to carry high volumes of hydrothermal hydrocarbons before the onset of supercritical conditions (Josephson, 1982; McDermott et al., 2018). Supercritical water has enhanced solvent capacity for organic compounds and reduced solvation properties for ionic species due to its loss of aqueous hydrogen bonding (Figure c; Simoneit, 1994). Hovland et al. (2006a, b) predict that some of the large volumes of deep subsurface salt found in the Red Sea, in the Mediterranean Sea and the Danakil depression, formed via the forced magmatically-driven hydrothermal circulation of seawater down to depths where it became supercritical. This salt, they argue, was precipitated deep under-ground via “shock crystallisation” from a supercritical effusive phase and so formed massive accumulations (mostly halite) typically in crustal fractures that facilitated the deep circulation. NaCl then flowed upwards in solution in dense, hot hydrothermal brine plumes, precipitating more solid salt beds upon cooling nearer or on the surface/seafloor (halite is a prograde salt). More recently, Scribano et al. (2017) and Hovland et al. (2018a, b) have theorised that serpentinisation is the dominant source of halite in the Messinian succession of the Mediterranean. To date, the Hovland et al. model of hydrothermal sourcing for widespread halite from a supercritical brine source (in active magmatic settings) has not been widely accepted by the geological community (Talbot, 2008; Warren., 2016; Aftabi and Atapour, 2018).
To date, no direct indications of the formation of masses of halite formed by this process have been sampled. In contrast to the theories of Hovland et al. (2018b), textures in the potash and halite salts in the Danakhil depression are evaporitic with only small volumes of hydrothermal overprint driven by the escape of saline volatiles derived thermal decomposition of hydrated salts. The postulated diapiric structures are not present in seismic, nor are any other buried hydrothermal/halokinetic structures visible in seismic (Bastow et al., 2018; Salty Matters; Warren 2016). Likewise, all the features seen in core and seismic in the Messinian of the Mediterranean are layered with classic sedimentary and halokinetic textures. The seismic across the Red Sea salt structures and the layering in the brine deeps are easily explained by current sedimentary and layered deep seafloor ponded brine (DHAL) models.
Some of the most visually striking examples of hydrothermal gypsum precipitation occur the Naica mine, Chihuahua, Mexico. There several natural caverns, such as Cave of Swords (Cueva de la Espades discovered in 1975) and Cave of Crystals (Cueva de los Cristales discovered in 2000), contain giant, faceted, and transparent single crystals of gypsum as long as 11 m (García-Ruiz et al., 2007; Garofalo et al., 2010). Crystals in Cueva de los Cristales are the largest documented gypsum crystals in the world. These huge crystals grew slowly at very low supersaturation levels from thermal phreatic waters with temperatures near the gypsum-anhydrite boundary. Gypsum still precipitates today on mine walls. According to García-Ruiz et al., 2007, the sulphur and oxygen isotopic compositions of these gypsum crystals are compatible with growth from solutions resulting from the dissolution of anhydrite, which was previously precipitated during late hydrothermal mineralisation in a volcanogenic matrix.
These giant gypsum crystal meshworks in the Cueva de los Cristales, Naica region, Mexico were first discovered in 2000 after water was pumped out of the phreatic cavern system as part of a silver mine's expanding operations (image courtesy of Penn State (https://science.howstuffworks.com/environmental/earth/geology/mexico-giant-crystal-cave.htm).
Naica Mine, Mexico. A) Cross-section of Naica mine. The mine exploits a hydrothermal Pb-Zn-Ag deposit with irregular manto and pipe morphologies entirely enclosed in subhorizontally dipping carbonates. Cavities of gypsum crystals are located in carbonates close to primary and secondary faults. Galleries have been excavated down to –760 m, requiring average pumping rate of 55 m3/min to depress groundwater to –580 m with respect to phreatic level located at −120 m; Naica and Gibraltar faults act as main drains. B) Homogenization temperatures of 31 fluid inclusions showing the actual temperature of growth. C) The solubility of gypsum calculated at 55°C and 105 Pa and measured activities of shallow and deep cave fluids from fluid inclusion data. Mixing at equilibrium between these two fluids in any proportion generates a liquid with a composition would consistently supersaturated with respect to gypsum, as shown by the position of the mixing curve, indicated by a dashed line, in the gypsum supersaturation field (after García-Ruiz et al., 2007; Garofolo et al., 2010))
The chemistry suggests that these megacrystals formed via a self-feeding mechanism, driven by a solution-mediated, anhydrite-gypsum phase transition. Nucleation kinetics calculations based on laboratory data show that this mechanism can account for the formation of these giant crystals, yet only when operating within a very narrow range of temperature of a few degrees as identified by the fluid inclusion values. Fluid inclusion analyses show that the giant crystals came from low-salinity solutions at temperatures ≈ 54°C, slightly below the temperature of 58°C where the solubility of anhydrite equals that of gypsum (Figure 9b; García-Ruiz et al., 2007). Van Driessche et al. (2011) argue the slowest gypsum crystal growth in the phreatic cavern occurred when waters were at 55°C. At this temperature, the crystals would take 990,000 years to grow to a diameter of 1 meter. By increasing the temperature in the cave by one degree, to 56° C, the same size crystal could have formed in a little less than half the time, or around 500,000 years. This possible growth rate would work out to about a billionth of a meter of growth per day and is perhaps the slowest growth rate that has ever been measured. Garofolo et al., 2010, accept the need for a limited temperature range during precipitation but argue the precipitating solutions were in part meteorically influenced. Their work focused on Cueva de las Espadas.
As for most other hypogenic caves, before their analytical work, they assumed that caves of the Naica region lacked a direct connection with the land surface and so gypsum precipitation would be unrelated to climate variation. Yet, utilising a combination of fluid inclusion and pollen spectra data from cave and mine gypsum, they concluded climatic changes occurring at Naica exerted an influence on fluid composition in the Espadas caves, and hence on crystal precipitation and growth. Microthermometry and LA-ICP-Mass Spectrometry of fluid inclusions in the gypsum in the Cueva de las Espadas indicate that brine source was a shallow, chemically peculiar, saline fluid (up to 7.7 eq. wt.%NaCl) and that it may have formed via evaporation, during an earlier dry and hot climatic period.
In contrast, the fluid of the deeper caves (Cristales) was of lower salinity (≈3.5 eq. wt.% NaCl) and chemically homogeneous and likely was little affected by evaporation processes. Galofolo et al. (2010) propose that mixing of these two fluids, generated at different depths of the Naica drainage basin, determined the stable supersaturation conditions needed for the gigantic gypsum crystals to grow (Figure 9c). The hydraulic communication between Cueva de las Espadas and the other deep Naica caves controlled fluid mixing. Mixing must have taken place during alternating cycles of warm-dry and fresh-wet climatic periods, which are known to have occurred in the region. Pollen grains from 35 ka-old gypsum crystals from the Cave of Crystals indicates a relatively homogenous catchment basin dominated by a mixed broadleaf wet forest. This suggests precipitation during a fresh-wet climatic period; the debate continues as to whether the gypsum at Naica is a mixing zone or a hydrothermal salt.