Journal of Geophysical Research: Planets

Petrological modeling of basaltic rocks from Venus: A case for the presence of silicic rocks


Corresponding author: J. G. Shellnutt, National Taiwan Normal University, Department of Earth Sciences, 88 Tingzhou Rd., Section 4, Taipei 11677, Taiwan. (


[1] The presence of highly evolved igneous rocks on Venus is a controversial issue. The formations of highland terranes and pancake domes are the two principal tectonic and volcanic features which argue in favor of the presence of silicic igneous rocks; however, the lack of water on Venus casts doubt on whether or not granites and rhyolites can form. Data returned to Earth from the Venera 13 and 14 landers show that the surface of Venus is composed of basaltic rocks similar in composition to those found on Earth. Here it is shown that anhydrous and hydrous fractional crystallization modeling using the Venera 13 and 14 data as starting materials can produce compositions similar to terrestrial phonolites and rhyolites. It is suggested that at shallow crustal levels (i.e., ≤ 0.1 GPa), mafic magmas can differentiate into silicic magmas resembling phonolites or rhyolites which may or may not erupt. Furthermore, the hydrous equilibrium partial melting models can produce rocks similar to terrestrial andesites and rhyolites, whereas anhydrous models suggest that there may be a uniquely Venusian type of silicic rock. The silicic rocks, if present, could act as “continental nucleation” sites and/or their presence may facilitate preferential sites of shearing and deformation of the Venusian crust.

1 Introduction

[2] Of the terrestrial planets, the Earth has many special features including abundant liquid water, thriving ecosystems, and an oxygen-rich atmosphere. The Earth is geologically dynamic due to active plate tectonics and water- and atmosphere-rock processes which ensure that the surface is constantly changing. One major geological difference between the Earth and all other terrestrial bodies (i.e., planets and asteroids) is the presence of evolved silicic igneous rocks (e.g., granite, ignimbrite, and rhyolite). The continental crust of Earth is composed mostly of silicic rocks with small amounts of sedimentary and mafic rocks, whereas the oceanic crust is primarily composed of basaltic rocks, chert, and ~3% silicic igneous rocks [Dixon and Rutherford, 1983; Bonin et al., 2002; Rudnick and Gao, 2003]. The evolved composition of the continental crust is largely due to recycling processes associated with hydrous plate tectonics (i.e., subduction, collision) and thus unique to Earth [Campbell and Taylor, 1983; Taylor and McLennan, 1985; Rudnick and Gao, 2003]. Although volumetrically minor, silicic rocks are commonly found within continental and oceanic mafic large igneous provinces, oceanic islands, and ocean ridges, where there is no active recycling of ancient crustal material [Yoder, 1973; Clague, 1978; Dixon and Rutherford, 1983; Bellieni et al., 1986; Shellnutt and Jahn, 2010; Natali et al., 2011].

[3] Continental mafic large igneous provinces (LIPs) are temporally, spatially, and structurally contiguous regions of the crust which are predominately composed of basaltic rocks and cover large (i.e., > 105 km2) areas [Jerram and Widdowson, 2005; Ernst et al., 2005; Bryan and Ernst, 2008]. Some LIPs are considered to be the physical expression of mantle upwelling where hot, compositionally primitive silicate magmas are transferred to the crust [Ernst and Buchan, 2003]. The presence of silicic rocks within LIPs demonstrates that high-temperature igneous process will produce compositionally evolved magmas [Melluso et al., 2008; Xu et al., 2010; Shellnutt et al., 2011]. The origin of silicic rocks within oceanic crust and mafic continental LIPs is debated, but they may be formed by: (1) fractional crystallization of mafic magmas, (2) partial melting of mafic rocks, (3) partial melting of crustal rocks (in the case of continental LIPs), and (4) silicate liquid immiscibility [Bellieni et al., 1986; Lightfoot et al., 1987; Shellnutt and Jahn, 2010; Xu et al., 2010; Charlier and Grove, 2012].

[4] A large portion (70%–80%) of the surface of Venus is featureless lava plains which lie within ± 1 km of the mean planetary radius [Fegley, 2003; Hansen and Young, 2007]. The remainder of the surface comprises mesolands and highlands. The mesolands have a median elevation (1–2 km) between the highlands and lowlands and contain tectonomagmatic features, such as coronae and chasmata (troughs). The highland regions represent 8%–10% of the surface and consist of crustal plateaux, tesserae terrane, volcanic edifices, and large-scale compression-related mountains [Vorder Bruegge et al., 1990; Suppe and Connors, 1992; Ansan and Vergely, 1995]. The formation of the highlands is debated as different tectonic models have been proposed to explain the compressional features (i.e., montes), extensional features (i.e., tesserae), and their high elevations (> 3 km). The tectonic models are broadly focused on whether mantle upwelling or downwelling is controlling the formation and maintenance of the highland terranes [Bindschadler et al., 1992; Bindschadler, 1995; Jull and Arkani-Hamed, 1995; Phillips and Hansen, 1998; Hansen and Willis, 1998; Hansen et al., 1999; Turcotte, 1995; Romeo and Turcotte, 2008].

[5] Venus is known to have large igneous provinces, and therefore, it stands to reason that silicic rocks derived by fractional crystallization of mafic magmas or partial melting of mafic rocks may be found [Ernst et al., 1995; Bonin et al., 2002; Hansen, 2007; Bonin, 2012]. Moreover, the compressional tectonics on Venus may induce partial melting of the crust which could produce evolved silicic magmas [Vorder Bruegger et al., 1990; Ansen and Vergely, 1995]. The highlands, chiefly Ishtar Terra and Aphrodite Terra, are considered to be analogous to “continents” on Earth in the sense that they are elevated with respect to the surrounding lands and that they may contain silicic rocks [Petford, 2000; Bonin et al., 2002; Romeo and Turcotte, 2008; Hashimoto et al., 2008; Bonin, 2012]. It is suggested by Hashimoto et al. [2008] and Basilevsky et al. [2012] that Venus may contain silicic rocks within the highland and tesserae regions and by extension implies that water was present in large concentrations (i.e., ocean) at some point in the past. Furthermore, the formation of pancake domes is thought to be related to lavas which may be silicic [Mckenzie et al., 1992; Pavri et al., 1992]. The presence of silicic rocks would have considerable implications especially for the isostatic and tectonic processes which operate on Venus [Jull and Arkani-Hamed, 1995; Smrekar et al., 2007; Romeo and Turcotte, 2008].

[6] In this paper, it is demonstrated using thermodynamically calibrated petrological software that silicic melts can be produced from magma compositions equal to basaltic rocks that were analyzed from the surface of Venus. In addition, the possible significance of the silicic rocks in accommodating tectonics and the growth of evolved “continental crust” within the highlands of Venus is discussed.

2 Geochemistry and Source Origin of the Basaltic Rocks From Venus

[7] The X-ray fluorescence chemical data of the surface rocks of Venus was reported by Surkov et al. [1984] from the Venera 13 and 14 landing probes, and additional data were reported from the Vega 2 lander [Surkov et al., 1986; Kargel et al., 1993]. The compositions of the rocks from Venera 13 and 14 are summarized in Table 1. The Venera 13 and 14 landing sites are located at equatorial latitudes at western Navka Planitia which is to the south-south-east of the Beta Regio highland [Surkov et al., 1984; Kargel et al., 1993]. The compositions of the Venera 13 and 14 landing sites are both basaltic, although the Venera 13 site is more alkaline than the tholeiitic compositions found at Venera 14 and even Vega 2 (Table 1). In comparison to terrestrial samples, the rocks analyzed at the Venera 14 and Vega 2 landing sites are similar to mid-ocean ridge basalt, whereas Venera 13 has more in common with mafic leucitic rocks, although the precision of the data is low [Kargel et al., 1993]. It is suggested that the rocks were generated by partial melting and subsequent fractionation from both high (i.e., > 1.8 GPa) pressure (i.e., Venera 13) and low (< 0.2 GPa) pressure (i.e., Venera 14 and Vega 2) mantle conditions [Fegley, 2003; Smrekar et al., 2007; Kiefer and Filiberto, 2009; Filiberto, 2009].

Table 1. Surface Compositions of Venus and Compositions Used for Modeling
SampleVenera 13aVenera 14aVega 2bMORBcVenera 13(anhydrous)Venera 13 (hydrous)Venera 14 (anhydrous)Venera 14 (hydrous)
  1. a

    Venera 13 and 14 data from Surkov et al. [1984].

  2. b

    Vega 2 data from Kargel et al. [1993].

  3. c

    The average mid-ocean ridge basalt (MORB) from Melson and Thompson [1971].

  4. d

    The Na2O content is calculated for the Venera 13, 14, and Vega 2 data [Surkov et al., 1984; Barsukov, 1992]. Error of the Venera 13 and 14 data reported at the 1σ level.

SiO2 (wt.%)45.1 ± 3.048.7 ± 3.650.8 ± 1.649.2146.74 ± 3.146.64 ± 3.149.79 ± 3.749.69 ± 3.7
TiO21.6 ± 0.451.25 ± 0.40.22 ± 0.051.391.66 ± 0.471.65 ± 0.471.28 ± 0.41.28 ± 0.4
Al2O315.8 ± 3.017.9 ± 2.617.8 ± 0.915.8116.37 ± 3.116.34 ± 3.118.30 ± 2.718.26 ± 2.7
FeO9.3 ± 2.28.8 ± 1.87.7 ± 0.659.189.64 ± 2.39.62 ± 2.39.00 ± 1.88.98 ± 1.8
MnO0.2 ± 0.10.16 ± 0.080.16 ± ± 0.10.21 ± 0.10.16 ± 0.080.16 ± 0.08
MgO11.4 ± 6.28.1 ± 3.312.8 ± 1.858.5311.81 ± 6.211.79 ± 6.28.28 ± 3.48.26 ± 3.4
CaO7.1 ± 1.010.3 ± 1.28.3 ± 0.3511.147.36 ± 1.07.34 ± 1.010.53 ± 1.210.51 ± 1.2
Na2O2.0 ± 0.5d2.4 ± 0.4d2.0 ± 0.5d2.712.07 ± 0.52.07 ± 0.52.45 ± 0.42.45 ± 0.4
K2O4.0 ± 0.60.2 ± 0.070.11 ± ± 0.64.14 ± 0.60.2 ± 0.070.2 ± 0.07
SO31.6 ± 1.00.35 ± 0.3 
Cl< 0.3< 0.4 
H2O   0.20.2

3 Starting Compositions and Modeling Conditions

[8] The thermodynamic evolution of mafic magmas can be modeled using the program pHMELTS [Ghiorso and Sack, 1995; Smith and Asimow, 2005]. pHMELTS and its derivative software are calibrated to bulk compositions encompassing SiO2-TiO2-Al2O3-Fe2O3-Cr2O3-FeO-MnO-MgO-CaO-Na2O-K2O-P2O5-H2O which represents the bulk composition of nearly all igneous rocks. The software allows the user to control important parameters such as relative oxidation state (fO2), pressure, and water (wt.%) content of the magma that is being modeled and can therefore determine the likely magmatic conditions which control the system.

[9] For this study, the basaltic compositions analyzed at the Venera 13 and 14 landing sites are used to assess whether or not silicic rocks can form on Venus. The Vega 2 composition was not selected due to its original low sum total of major elements in comparison with the Venera probe data [Surkov et al., 1986]. The absolute compositions reported by Surkov et al. [1984] and normalized to 100% are used for the starting compositions (Table 1). The Na2O content of the rocks was not determined by X-ray fluorescence however the values in Table 1 are calculated and are therefore a “best guess” as to the true values for the rocks [Surkov et al., 1984, 1986; Kargel et al., 1993; Treiman, 2007]. The data precision is carried through the models according the analytical results initially reported by Surkov et al. [1984]. MELTS or the MELTS family of the thermodynamic calculators do not induce further data uncertainty; however, there are inherent phase equilibrium uncertainties when using computer simulations because the program is based on experimental results and thus the verisimilitude of the results may be debated. The grey fields in Figures 1-8 are the graphical representation of the maximum 1σ error of the data at 0.1 GPa.

Figure 1.

Results of Venera 13 anhydrous fractional crystallization models. Phonolite (>52 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area. The grey zones represent the maximum 1σ error of the elements found in Table 1 at 0.1 GPa. Classification scheme of LeMaitre et al. [1989], 1 = foidite, 2 = picro-basalt, 3 = basalt, 4 = basaltic andesite, 5 = andesite, 6 = dacite, 7 = rhyolite, 8 = trachyte (quartz < 20%), trachydacite (quartz > 20%), 9 = trachyandesite, 10 = basaltic trachyandesite, 11 = trachybasalt, 12 = tephrite (olivine < 10%), basanite (olivine > 10%), 13 = phonotephrite, 14 = tephriphonolite, 15 = phonolite.

Figure 2.

Results of Venera 13 hydrous (0.2 wt.% H2O) fractional crystallization models. The grey zones represent the maximum 1σ error of the elements found in Table 1 at 0.1 GPa. Phonolite (>52 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area.

Figure 3.

Results of Venera 13 anhydrous equilibrium partial melting models. The grey zones represent the maximum 1σ error of the elements found in Table 1 at 0.1 GPa. Phonolite (> 52 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area.

Figure 4.

Results of Venera 13 hydrous (0.2 wt.%) equilibrium partial melting models. The grey zones represent the maximum 1σ error of the elements found in Table 1 at 0.1 GPa. Phonolite (> 52 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area.

Figure 5.

Results of Venera 14 anhydrous fractional crystallization models. The grey zones represent the maximum 1σ error of the elements found in Table 1 at 0.1 GPa. Rhyolite (> 69 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area.

Figure 6.

Results of Venera 14 hydrous (0.2 wt.%) fractional crystallization models. The grey zones represent the maximum 1σ error of the elements found in Table 1 at 0.1 GPa. Rhyolite (> 69 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area.

Figure 7.

Results of Venera 14 anhydrous equilibrium partial melting models. Andesite (>50 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area. Error zones are not shown for these models because they do not produce results similar to terrestrial samples and thus may be unique to Venus.

Figure 8.

Results of Venera 14 hydrous (0.2 wt.%) equilibrium partial melting models. The grey zones represent the maximum 1σ error of the elements found in Table 1 at 0.1 GPa. Andesite (>50 wt.% SiO2) data taken from the GEOROC database ( and is represented by the hatched area.

[10] The oxidation state of the Venusian mantle is unknown. The lower atmospheric conditions of Venus suggest that the surface is at or near the hematite-magnetite buffer which is considerably oxidizing [Fegley et al., 1997]. However, the atmospheric conditions are not necessarily indicative of the relative mantle oxidation state. In comparison, the relative oxidation state of the Earth's upper mantle is considered to be within ± 2 log units of the FMQ buffer but it is not uniform and has likely changed over time [Arculus, 1985; Kasting et al., 1993; Frost and McCammon, 2008]. Venus currently does not appear to have a direct exchange mechanism between the surface and mantle (i.e., subduction) but volcanism has likely occurred in the recent past, and it is possible that the redox conditions may have evolved accordingly. The modeling results of Filiberto [2009] indicate that terrestrial mantle conditions are suitable to produce rocks with similar compositions as those found on Venus; therefore, the FMQ buffer is used as the relative oxidation state for the models.

[11] Models were run at pressure conditions of 0.01, 0.1, and 1 GPa at the FMQ buffer. Since there is no surface water on Venus, it is likely that the mantle does not contain substantial volatiles (e.g., H2O), but the atmosphere contains 30 ± 15 ppm water vapor, 150 ± 30 ppm SO2, 0.6 ± 0.12 ppm HCl, and is primarily (i.e., ~96%) composed of CO2 [Fegley, 2003; Sandor and Clancy, 2005]. It is possible that volcanic degassing is the source of the atmospheric water vapor, SO2, Cl, and CO2, and thus they are likely to be found in the mantle. Furthermore, Filiberto [2009] suggested that the Venusian mantle may be hydrous (~0.2 wt.%) because basalts similar to those found at the Venera 13 landing site require a hydrous source region on Earth. Because of the uncertainty of the water content of the starting compositions, both anhydrous and hydrous (0.2 wt.%) models were run.

4 Results

4.1 Venera 13 Fractionation and Equilibrium Partial Melt Models

[12] Fractional crystallization models were run for anhydrous and hydrous conditions for the Venera 13 composition. The two models show that the starting composition can produce liquids which contain ~56 wt.% SiO2 and classify as phonolites according to the total alkalies versus silica plot of LeMaitre et al. [1989] (Figures 1a, 2a). The compositional evolution curves of the residual liquids at various pressures are shown in Figures 1, 2 with the range of terrestrial phonolitic rocks for comparison. It is clear that in the anhydrous and hydrous low-pressure models, the liquid evolution curves cross into the terrestrial phonolitic field, whereas the high-pressure curves do not (i.e., TiO2). At 56 wt.% SiO2, the modeled phonolitic liquid compositions have temperatures between 1060°C and 1070°C and densities of ~2.44 g/cm3 (Table 2).

Table 2. Physical Parameters of the Modeled Silicic Liquidsa
SampleProcessPressure (GPa)Temperature (°C)Viscosity (Pa s)Density (g/cm3)Rock TypeWater State
  1. a

    FC = fractional crystallization, PM = partial melting.

Venera 13-1FC0.011060103.82.44Phonoliteanhydrous
Venera 13-2FC0.11070103.72.46Phonoliteanhydrous
Venera 13-3FC0.011060103.32.43Phonolitehydrous
Venera 13-4FC0.11060103.52.44Phonolitehydrous
Venera 14-1FC0.01970105.82.42Rhyoliteanhydrous
Venera 14-2FC0.1980105.52.44Rhyoliteanhydrous
Venera 14-3FC0.01960105.02.37Rhyolitehydrous
Venera 14-4FC0.1910104.82.32Rhyolitehydrous
Venera 14-5FC1900103.82.49Rhyolitehydrous
Venera 14-6PM0.011080102.62.54Andesitehydrous
Venera 14-7PM0.1940103.32.27Andesitehydrous
Venera 14-8PM1880104.62.49Rhyolitehydrous

[13] In contrast to the fractionation models, the equilibrium partial melt models fail to produce results which are similar to terrestrial phonolites or tephriphonolites (Figures 3 and 4). The models do not replicate the MgO content of naturally occurring terrestrial phonolitic rocks, although other elements do overlap within the phonolite field in Figures 3 and 4. There are no known rocks observed on Earth with the modeled partial melt compositions, and it is beyond the scope of this paper to verify the existence of uniquely Venusian rocks due to the lack of precision of the original basaltic analysis, but the results indicate that there could be a type of silicic rock that has not been found on Earth (Table 3).

Table 3. Modeled Compositions and Their Physical Parametersa
SampleSiO2 wt.%TiO2Al2O3FeOtMnOMgOCaONa2OK2OASINKA% Residual LiquidDry Melting Temperature (°C)Wet Melting Temperature (°C)
  1. a

    Major element results in weight %. Percentage (%) of residual liquid is the amount of liquid remaining from the original parental magma. Wet melting temperature = 1% H2O in model. ASI = Al3+/Ca2++Na++K+. NKA = Na++K+/Al3+.

Venera 13-156.051.7820.462.510.291.173.303.9610.420.860.8738.4970900
Venera 13-255.671.8720.822.610.
Venera 13-355.351.5921.
Venera 13-455.471.6421.342.860.291.333.483.939.590.920.7942.21020760
Venera 14-175.570.896.674.961.100.354.274.431.650.401.369.1900770
Venera 14-274.461.096.665.001.390.254.394.701.940.371.487.1950820
Venera 14-373.680.628.295.480.940.613.925.001.320.491.1612.6880820
Venera 14-475.640.448.103.991.100.343.305.411.600.491.3110.0930830
Venera 14-576.760.6110.123.502.650.011.795.212.270.711.095.710601000
Venera 14-656.692.8314.479.560.643.517.233.551.340.710.508.01090830
Venera 14-764.330.5517.124.260.951.133.804.473.310.960.64≤ 12.6920730
Venera 14-874.330.6214.840.433.160.342.352.241.681.520.375.3940810

4.2 Venera 14 Fractionation and Equilibrium Partial Melting Models

[14] The fractionation models of the Venera 14 analyses are more compelling than the Venera 13 models. The low-pressure anhydrous models produced rhyolitic compositions with ≥71 wt.% SiO2 (Figures 5a and 6a). The modeled liquid composition evolution curves fall within the range of terrestrial rhyolites (Figures 5 and 6). The Al2O3 content of the low-pressure model curves for the anhydrous conditions is low and just touches the terrestrial range (Figure 6c). Since terrestrial igneous rocks contain some amount of H2O or volatile phase, it is reasonable to suggest that the low Al2O3 content of the Venera 14 models could simply reflect anhydrous conditions on Venus. Low Al2O3 rhyolites are very unlikely to form on Earth, whereas on Venus, it could be a possibility. Nonetheless, the hydrous models replicate terrestrial rhyolites very well including the Al2O3 content for all of the pressure conditions (Figure 6). The anhydrous low-pressure models indicate liquid temperatures of 970°C–980°C, whereas the hydrous models indicate temperatures between 900°C (high pressure) and 960°C (surface pressure). In both models, the rhyolites are similar to peralkaline A-type silicic rocks of Earth and have densities of 2.3 to 2.4 g/cm3 [Frost et al., 2001; Bonin, 2007] (Table 2).

[15] Unlike the Venera 13 models, the Venera 14 equilibrium partial melt models produced composition evolution curves which fall within the range of terrestrial rocks (Figures 7 and 8). Only the hydrous partial melt models were able to replicate terrestrial rhyolitic to andesitic rocks with SiO2 content ranging from ~55 wt.% in the surface model up to 64 wt.% for the moderate pressure model and to ~73 wt.% for the high-pressure model which correspond to temperatures of 1080°C, 940°C, and 880°C and densities of 2.3–2.5 g/cm3 (Table 2). The anhydrous models do not replicate compositions similar to terrestrial rocks, suggesting there may be a type of silicic rock unique to Venus.

5 Discussion

5.1 Formation of Silicic Rocks on Venus

[16] The modeling results presented in Figures 1-8 indicate that silicic rocks can form by fractional crystallization and equilibrium partial melting of starting material equal to anhydrous and hydrous compositions observed on the surface of Venus. The possibility of silicic rocks on Venus has direct implications for the presence of volatiles (i.e., H2O, Cl, F, S, CO2) and understanding how tectonics operates on Venus as well as its ancient planetary development.

[17] It has been known that the surface of Venus is dry since the 1960s; however, studies of the atmosphere have detected volatiles, with very modest amounts of water vapor, which may suggest that volcanic degassing is responsible [Fegley, 2003; Sandor and Clancy, 2005]. The anhydrous fractionation models demonstrate that water is not required to produce silicic magmas (i.e., phonolites and rhyolites), and that they are the likely consequence of high-temperature crystallization processes. The equilibrium partial melting models, on the other hand, reveal a different situation. The hydrous partial melting models of Venera 14 produce compositions which are similar to terrestrial rocks, whereas the anhydrous models did not, thus there could be rocks present on Venus which do not have equivalents on Earth. The presence of non-H2O volatiles in rocks can also reduce melting temperatures and therefore may be an alternative to hydrous partial melting [Bailey, 1977; Manning, 1981; Lowenstern, 2000; Dolejs and Baker, 2007a, 2007b; Dasgupta et al., 2007; Smrekar et al., 2007; Brey et al., 2009; Filiberto and Treiman, 2009; Filiberto et al., 2012]. There are probably traces of F and Cl within the Venera 13 rock and possibly the Venera 14 rock but, at the moment, there is no way to verify their abundance [Fegley, 2003]. The anhydrous fractionation models presented are probably closer to the real situation on Venus; however there is no precise account of volatile phases, such as CO2, SO2, F, or Cl in the mantle or crust [Chassefière et al., 2012].

5.2 Tectonomagmatic Setting of Silicic Rocks on Venus

[18] Large mafic-ultramafic cumulate intrusions are very common within continental and oceanic crust of Earth and tend to host large base metal deposits of Cr, Fe, Ti, V and PGEs [Irvine, 1977; Dixon and Rutherford, 1983; Weibe, 1996; Shellnutt et al., 2011]. The cumulate rocks are very often associated with silicic rocks which represent the residual liquids after crystal fractionation [Bonin, 2007; Shellnutt et al., 2011]. The gabbro-granitoid complexes of the Emeishan large igneous province may be an analog for Venus. The gabbro-granitoid complexes consist of a cumulate mafic portion and a peralkaline quartz-rich silicic portion [Shellnutt and Jahn, 2010; Shellnutt et al., 2011]. The silicic portion of the Panzhihua complex, for example, shows compositional evolution from syenodiorite (~63 wt.% SiO2) to granite (~72 wt.% SiO2) [Shellnutt and Jahn, 2010]. The presence of perthitic feldspar and F-rich apatite and ferrorichterite within the silicic rocks of the Panzhihua complex indicates that it was likely water-poor but F-rich [Shellnutt and Jahn, 2010; Shellnutt and Iizuka, 2011]. Water-poor granitic rocks are also found within oceanic crust [Dixon and Rutherford, 1983]. The estimated crystallization temperature of the Panzhihua granite is ~940°C which is only 30°C–40°C lower than the anhydrous modeled conditions for Venera 14 but overlaps with the hydrous model temperatures. Furthermore, the Panzhihua granite likely fed surface flows as there are temporally and spatially associated columnar-jointed trachytes which have identical compositions as the granites, thus it is possible that a similar process may occur on Venus [Pavri et al., 1992; Shellnutt and Jahn, 2010]. As an alternative, Charlier and Grove [2012] demonstrated that silicate immiscibility can occur in anhydrous, low-pressure conditions. Their experimental evidence shows that the immiscible Fe-rich and Si-rich silicate end-members of compositions similar to common basaltic and rhyolitic rocks found on Earth and could potentially be a large-scale phenomenon (i.e., plutono-volcanic). Whether by fractionation or silicate immiscibility, providing the existence of magma chambers, silicic rocks should be found on Venus either as plutono-hypabyssal complexes or possibly as volcanic rocks.

[19] Pancake domes are a distinctive volcanic feature of Venus. The pancakes are broad, flat, circular lava domes, ~1 km high and up to 100 km wide which populate the surface of Venus [Head et al., 1992]. The pancake domes are thought to be formed by viscos lavas which may be silicic or andesitic [Pavri et al., 1992; Mckenzie et al., 1992; Fink et al., 1993; Petford, 2000; Treiman, 2007]. However, Sakimoto and Zuber [1993] and Bridges [1995] suggest that the pancake domes are probably a cooling feature of basaltic lavas rather than a morphological feature of silicic lava. Although pancake domes may be a formed by mafic silicate magmas, it is possible they represent eruptions of silicic lavas from shallow level magma chambers (Figure 9). The viscosities calculated for the silicic liquids in this study range between 102.5 and 106.0 Pa s and match the silicic magma viscosities suggested by Pavri et al. [1992] that are favorable for the genesis of pancake domes. Thus, if silicic lavas or magmas form on Venus, the most likely scenario is within large-scale (i.e., ≤ 100 km3) crustal magma chambers which may or may not erupt (Figure 9). The implication is that the presence of a pancake dome may indicate that a shallow level magma chamber is located directly beneath the surface [Pavri et al., 1992].

Figure 9.

Conceptual diagrams of (1) probable, (2) possible, and (3) speculative silicic rock formation and accumulation on Venus.

5.3 The Formation of Siliceous Crust on Venus

[20] The formation of Earth-like continental crust, although not completely understood, is largely a product of rock recycling through plate tectonic processes (e.g., orogenesis and subduction). Many large-scale granitic batholiths on Earth in either collisional or arc environments are produced by partial melting and mixing of mantle- and crust-derived magmas. The formation of silicic rocks on Earth is greatly aided by the presence of water which reduces melting temperatures and affects viscosity and density [Campbell and Taylor, 1983; Thomas and Davidson, 2012]. Since water is not currently pervasive on Venus and Earth-like plate tectonics does not exist, the possibility of Cordilleran-style batholiths such as the trans-Himalayan batholith complex or large-scale juvenile crust formation similar to the Central Asian Orogenic Belt is highly unlikely [Hansen and Phillips, 1993; Nimmo and Mckenzie, 1998; Fegley, 2003; Smrekar et al., 2007].

[21] The highland terranes of Venus are concentrated around Ishtar Terra, Aphrodite Terra, and Beta Regio. Both Ishtar Terra and Aphrodite Terra show evidence for complex compressional features and volcanic edifices. The highlands of Venus make up ~10% of the surface, whereas the lowlands comprise ~80% of the surface [Nimmo and Mckenzie, 1998; Fegley, 2003; Hansen and Young, 2007]. At the present time, the crust of Venus is considered to be derived from basaltic rocks and their weathered products [Kargel et al., 1993; Nimmo and Mckenzie, 1998; Fegley, 2003; Hansen, 2007]. Hashimoto et al. [2008] and Basilevsky et al. [2012], using near-infrared mapping spectrometry data, identified signatures indicative of silicic rocks in the highland regions, whereas the lowlands have higher emissivity indicative of basaltic rocks. As a consequence, Hashimoto et al. [2008] speculated that water was present during the ancient past on Venus in order to facilitate the formation of granites within the highland terranes. Romeo and Turcotte [2008] suggest that the tessera terranes on Venus may represent low-density continental crust which does not participate in the episodic recycling of the lithosphere. The formation of the Venusian continental (i.e., siliceous) crust is not explicitly stated in their model, but it is suggested that that crust is related to downwelling tectonics. The model presupposes that low (i.e., ≤ 2.9 g/cm3) density rocks already exist but would require a mechanism (e.g., partial melting) to generate their compositions. There are other models for the generation of the highland terranes of Venus which involve strictly mafic compositions either directly or indirectly related to upwelling tectonics (i.e., mantle plumes) and crust thickening by magmatic underplating [Phillips and Hansen, 1998; Hansen and Willis, 1998; Hansen et al., 1999].

[22] The modeling results presented in this paper indicate that low-density silicic rocks should be able to form at shallow crustal levels of Venus. The rocks are consistent with phonolitic (~2.4 g/cm3) and rhyolitic (~2.4–~2.6 g/cm3) compositions which have lower densities than the basaltic crust of Venus (~2.9 g/cm3). Because the models are based on differentiation of mafic magmas, the type of Venusian highland tectonic model (i.e., upwelling or downwelling) does not matter so long as a mantle melt is generated and differentiates. It is conceivable that specific portions of the upper Venusian crust have a higher concentration of mafic-silicic complexes (e.g., magma chambers) which could produce a localized averaged crustal density closer to intermediate (i.e., 2.7 to 2.8 g/cm3) values. If there was a sufficiently sized mixed mafic/silicic crustal region, it could represent a continental crust nucleation site which establishes a low-density buoyant terrane (Figure 9). Any subsequent stress or strain applied to the buoyant terrane could be more readily accommodated by the presence of silicic rocks which have lower melting temperatures and are easier to deform than basaltic rocks [Jull and Arkani-Hamed, 1995]. Furthermore, the silicic rocks could accumulate in compressional regions associated with mantle downwelling and mountain building processes and “resist” subduction or burial (Figure 10). The lower average density of the buoyant terrane means that it would not sink back into the mantle and could therefore be older than the surrounding lowlands. Any additional eruptions of volcanic rocks on top of the of the nucleation site, whether they are mafic or silicic, would be supported by the buoyant crust.

Figure 10.

Conceptual diagram of tectonic accumulation of silicic rocks in a (arrows) compressional regime on Venus.

[23] If the crustal rocks of Venus are anhydrous and volatile-free, then the formation of silicic rocks within compressional regimes is very difficult. Assuming that silicic rocks equal to the compositions modeled in this paper exist and they are within the compressional areas of Venus (i.e., Maxwell Montes), a minimum temperature of 880°C is required to partially melt the rhyolites, assuming they are anhydrous composition. In contrast, temperatures of ~630°C and ~770°C are required to induce melting if the phonolitic and rhyolitic rocks contain ~1% H2O. Therefore, it is unlikely that anhydrous silicic magmas are generated during Venusian orogenesis because the high temperatures required to melt the anhydrous rocks should occur at pressures similar to eclogite facies on Earth (i.e., >1 GPa). Determining the presence of molecular water, identified as OH- ions, or other volatiles (i.e., Cl, F) within the crustal rocks of Venus, would be an indicator of the potential for hydrous melting and thus the possibility of granitic rocks within the compressional regions. However, relatively anhydrous fractional crystallization of mafic magmas is likely the primary process of generating silicic rocks on Venus.

[24] The silicic rocks represent between 6% and 12% of the total parental magma according to the models (Table 2). The fact that the Venusian highland crust and residual liquid rhyolitic compositions both represent ~10% of their respective processes is surprising and at first glance suggests that there may be a link. In other words, the highland terranes could be regions of concentrated silicic rocks after fractional crystallization which have been subsequently deformed, but until silicic rocks are positively identified on Venus, all hypotheses will remain speculative [c.f. Jull and Arkani-Hamed, 1995; Hashimoto et al., 2008].

6 Conclusions

[25] Thermodynamical modeling of the whole rock geochemical data collected from the Venera 13 and 14 landers indicate that silicic melts can form on Venus. Fractional crystallization modeling at shallow pressures (≤ 0.1 GPa) at the FMQ buffer of both anhydrous and hydrous parent compositions can reproduce phonolitic and rhyolitic rocks which are similar to the compositional range observed on Earth. Anhydrous equilibrium partial melt modeling, irrespective of pressure, did not produce results which overlap with known terrestrial rocks, suggesting there could be a type of silicic rock unique to Venus. Equilibrium partial melt modeling of Venera 13 produced results that are similar to andesite and rhyolite but require hydrous or volatile-rich conditions. The results indicate that silicic rocks may form a small portion (≤ 10%) of the Venusian crust similar to large igneous provinces or oceanic crust on Earth. Furthermore, the silicic rocks, if present, could act as “continental nucleation” sites, and/or their presence may facilitate faulting and deformation of the Venusian crust thereby assisting surficial tectonic processes (i.e., compression).


[26] JGS would like to acknowledge support from the National Science Council of Taiwan through grant 100-2116-M-003-006.

[27] The Editor on this paper was Mark Wieczorek. He thanks three reviewers, Justin Filiberto, Etienne Médard, and David Baratoux.