Viscosity of hydrous kimberlite and basaltic melts at high pressures

Available online at www.sciencedirect.com

ScienceDirect Russian Geology and Geophysics 58 (2017) 1093–1100 www.elsevier.com/locate/rgg

Viscosity of hydrous kimberlite and basaltic melts at high pressures E.S. Persikov a,*, P.G. Bukhtiyarov a, A.G. Sokol b a

Institute of Experimental Mineralogy, Russian Academy of Sciences, ul. Akademika Osip’yana 4, Chernogolovka, 142432, Moscow Region, Russia b V.S. Sobolev Institute of Geology and Mineralogy, Siberian Branch of the Russian Academy of Sciences, pr. Akademika Koptyuga 3, Novosibirsk, 630090, Russia c Novosibirsk State University, ul. Pirogova 2, Novosibirsk, 630090, Russia Received 20 April 2016; accepted 1 September 2016

Abstract New experimental data on the temperature and pressure dependences of the viscosity of synthetic hydrous kimberlite melts (82 wt.% silicate + 18 wt.% carbonate; degree of depolymerization: 100⋅NBO/T = 313 for anhydrous melts and 100⋅NBO/T = 247 for melts with 3 wt.% H2O) were obtained at a water pressure of 100 MPa and at lithostatic pressures of 5.5 and 7.5 GPa in the temperature range 1300–1950 °C. The temperature dependence of the viscosity of these melts follows the exponential Arrhenius–Frenkel–Eyring equation in the investigated range of temperatures and pressures. The activation energies of viscous flow for hydrous kimberlite melts were first shown to increase linearly with increasing pressure. Under isothermal conditions (T = 1800 °C), the viscosity of hydrous kimberlite melts increases exponentially by about an order of magnitude as the pressure increases from 100 MPa to 7.5 GPa. The new experimental data on the viscosity of hydrous kimberlite melts (error ±30 rel.%) are compared with forecast viscosity data for anhydrous kimberlite and basaltic melts (100⋅NBO/T = 51.5) and for hydrous basaltic melts (100⋅NBO/T = 80). It is shown that at comparable temperatures, the viscosity of hydrous kimberlite melts at a moderate pressure (100 MPa) is about an order of magnitude lower than the viscosity of hydrous basaltic melts, whereas at a high pressure (7.5 GPa) it is more than twice higher. It is first established that water dissolution in kimberlite melts does not affect seriously their viscosity (within the measurement error) at both moderate (100 MPa) and high (7.5 GPa) pressures, whereas the viscosity of basaltic melts considerably decreases with water dissolution at moderate pressures (100 MPa) and remains unchanged at high pressures (P > 3.5 GPa). © 2017, V.S. Sobolev IGM, Siberian Branch of the RAS. Published by Elsevier B.V. All rights reserved. Keywords: viscosity; kimberlite; basalt; water; temperature; pressure; melt; model; mantle; Earth’s crust

Introduction Magma viscosity largely controls important features of magmatic processes, such as, e.g., the movement of magmatic melts in the crust and the mantle, their origin, evolution, and stabilization in a variable field of temperature, pressure, and composition. Furthermore, the form of extrusive and intrusive rock masses, their textural characteristics, as well as the differentiation of magmas as a result of fluid-magma interaction, gravitational sedimentation, and partial melting are also limited by magma viscosity. The huge variety of major and volatile elements, the heterophase nature, and the wide range of temperatures and pressures are the main features of the existence of magmatic melts in nature. Experimental studies of the viscosity of such systems at high temperatures and

* Corresponding author. E-mail address: [email protected] (E.S. Persikov)

pressures are a complex technical and methodological problem (Persikov, 1991; Lange, 1994; and others). Despite significant advances in this area of petrology and geochemistry (Allwardt et al., 2007; Brearley et al., 1986; Chepurov et al., 2009; Dingwell et al., 2004; Fujii and Kushiro, 1977; Giordano et al., 2004; Kushiro, 1980; Lebedev and Khitarov, 1979; Lange, 1994; Liebske et al., 2003; Mysen et al., 1988; Neuville and Richet, 1991; Persikov, 1984, 1991, 1998, 2007; Persikov and Bukhtiyarov, 2004, 2009; Persikov et al., 1987, 1989, 1990, 2015; Reid et al., 2003; Scarfe, 1986; Scarfe et al., 1987; Shaw et al., 1968; Whittington et al., 2000; Wolf and McMillan, 1985), a number of problems remain unexplored, especially those concerning the viscosity of mafic and ultramafic melts at the TP parameters of their origin in the Earth’s mantle. Experimental and theoretical data on the viscosity of kimberlite melts at the TP parameters of the crust and upper mantle were completely absent prior to our studies (Persikov et al., 2015a,b,c). Note that predicted data on the temperature and pressure dependences of the viscosity of synthetic kimberlite

1068-7971/$ - see front matter D 201 7 , V . S. So bolev IGM, Siberian Branch of the RAS. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.rgg.2017.08.005 +

1094

E.S. Persikov et al. / Russian Geology and Geophysics 58 (2017) 1093–1100

melts at pressures up to 7.5 GPa were first obtained in (Persikov et al., 2015), and the effective viscosity of heterogeneous (liquid + crystals + fluid phase bubbles) kimberlite magmas was first predicted in (Persikov et al., 2015) over a wide range of TP parameters corresponding to the conditions of origin of these magmas in the Earth’s mantle and the conditions of hypabyssal facies, i.e., the formation of kimberlite dikes, sills, and diatremes. In recent experimental studies of the viscosity of natural kimberlite melts, Chepurov and Pokhilenko (2015) estimated the influence of the crystalline phase (up to 30 wt.%) on the effective viscosity of a heterogeneous kimberlite melt at a pressure of 4 GPa and temperatures of 1400 and 1600 °C. In the present work, the effect of water dissolved in melts, temperature, and pressure on the viscosity of synthetic kimberlite melts was studied in a wide temperature range of 1300–1950 °C and at pressures of 100 MPa–7.5GPa. New experimental data on the viscosity of hydrous kimberlite melts (100⋅NBO/T = 247) were compared with the calculated dependences for such melts, and the calculated dependences of the viscosity of anhydrous kimberlite melts (100⋅NBO/T = 313), and the calculated dependences of the viscosity of anhydrous (100⋅NBO/T = 51.5) and hydrous basaltic melts (100⋅NBO/T = 80) which were obtained using a physicochemical model for predicting the viscosity of magmatic melts (Persikov, 1998, 2007; Persikov and Bukhtiyarov, 2009; Persikov et al., 2015)

Experimental and analytical methods Selecting a representative composition of kimberlite magmas is not a simple problem. Despite the significant advances made in the petrology and geochemistry of kimberlite magmatism (Sparks et al., 2006, 2009; Wyllie, 1980), the problem of determining the composition of the melt from which kimberlites crystallize in the hypabyssal facies and especially the primary composition of the kimberlite magmas formed in the mantle, remains controversial (Kamenetsky and Yaxley, 2015; Kamenetsky et al., 2009; Sharygin et al., 2013; Sokol et al., 2013; Sparks et al., 2006, 2009; Wyllie, 1980). For example, it is assumed that the nucleation of kimberlite melts occurs through the reaction of carbonatite melts with peridotite near the mantle solidus and subsequent partial melting of carbonated peridotite at very low degrees of melting (≤1.0%) at pressures (6–10 GPa), depths of ≈150–300 km, and a temperature ≤1500 °C (Dalton and Presnall, 1998; Dasgupta and Hirschmann, 2006; Kamenetsky and Yaxley, 2015; Kamenetsky et al., 2009; Karanagh and Sparks, 2009; Kopylova et al., 2007; Michell, 2008; Price et al., 2000; Sharygin et al., 2013; Sparks et al., 2006, 2009; Wyllie, 1980). Basaltic magmas can be generated in the asthenosphere though partial melting of garnet peridotite (≤25 vol.%) at pressures of ~5.5 GPa, depths of ~100 km, temperatures of ~1350 °C, and a water concentration in them ≤1.0 wt.% (Yoder, 1976; and others). Therefore, in the Table 1, the composition of the synthetic kimberlite melt used in our experiments is compared

with the average composition of unaltered hypabyssal kimberlite of the Udachnaya-East kimberlite pipe (Siberian Platform, Yakutia) selected as the representative composition of kimberlite magmas. The criterion for the comparison is not the concentration of the main rock-forming melt components, but the overall basicity of such melts, which has a decisive influence on their viscosity and is numerically determined using the structural-chemical parameter: the degree of depolymerization or the basicity coefficient K = 100⋅NBO/T. This structural-chemical parameter of melts adeuqatly reflects the features of the overall chemical composition and structure of silicate and magmatic melts (Kopylova et al., 2007; Mysen, 1988; Persikov, 1984, 1991; Persikov et al., 1990, 2015). Obviously, the concentrations of rock-forming components in the synthetic kimberlite used in the present study are not strictly similar to those in natural kimberlites. However, the validity of the approach used to study the viscosity of magmatic melts has previously been tested in sufficient detail (Persikov, 1984, 1991; Persikov et al., 1990, 2015). Furthermore, this has made it possible for the first time to obtain ultramafic glasses with a high content (18 wt.%) of the molten carbonate phase which are stable (without degassing) both at moderate (100 MPa) and ultrahigh (7.5 GPa) pressures. The average basalt composition obtained in (Le Maitre, 1976) based on a statistical analysis of more than 3500 basalt samples from almost all regions of the world (Table 1) was selected as a representative composition of basaltic melts. The starting materials for viscosity measurements of kimberlite melts were stoichiometric mixtures of natural minerals: albite (Ab, NaAlSi3O8) from the Kolba massif (Kazakhstan) (Persikov et al., 1990) and calcite (Cal, CaCO3) from the Yoko-Dovyren layered intrusion (North-Baikal region, Russia) (Persikov et al., 2010b). It has been found in previous experiments that in contrast to hydrous silicate systems, in carbonate-silicate systems at increasing temperature and at moderate pressures (up to ~3.0 GPa), solid-state reactions between carbonates and silicates with a release of a large amount of CO2 occur well before the melting of silicate and carbonate phases. As a result, the melts that formed differ from the starting mixtures in basicity and the content of CO2 and the carbonate phase; the concentration of the latter in the melt can vary widely, depending on the initial composition of the mixture (Fig. 1) (Persikov et al., 2012). The required synthetic kimberlite melt (82 wt.% silicate + 18 wt.% carbonate, Table 1) was synthesized by melting the starting mixture of powders of albite and calcite (Ab38Cal62, wt.%) in platinum capsules 6 mm in diameter and 60 mm high open at one end at T = 1300 °C and at a CO2 pressure of 100 MPa, with a mass balance control in each experiment. The experiments were carried out using a high-pressure gas vessel with internal heating, equipped with a unique internal device which was described previously (Persikov and Bukhtiyarov, 2004; Persikov et al., 2010a,b) and which allows experiments at a fluid (CO2, H2O) pressure to be carried out without changing the original geometry of the capsule with the melt. The time of the experiments was 4 h at T = 1300 °C with a 1-h pre-exposure of the samples at T = 850 °C to complete the

1095

E.S. Persikov et al. / Russian Geology and Geophysics 58 (2017) 1093–1100 Table 1. Chemical composition (wt.%) and chemical structural parameters (100⋅NBO/T) of kimberlite, basaltic, and synthetic kimberlite melts Component

Kimberlite1

Anhydrous kimberlite melt2

Hydrous kimberlite melt3

Basalt4

Basalt + 3 wt.% H2O5

SiO2

29.64

34.4

33.06

49.2

48.44

TiO2

1.29

–

–

1.84

1.84

Al2O3

1.81

10.31

9.91

15.74

15.5

Fe2O3

3.18

–

–

3.79

3.73

FeO

5.26

–

–

7.13

7.02

MnO

0

–

–

0.20

0.2

MgO

33.54

–

–

6.73

6.63

CaO

11.98

39.8

38.24

9.47

9.32

Na2O

0.31

5.05

4.85

2.91

2.87

K2O

1.02

–

–

1.1

1.08

P2O5

0.45

–

–

0.35

0.34

SO3

0.17

–

–

–

–

H2O (mol.)

0.85

–

2.04

0.95

–

–

OH (acid)

2.02

–

1.86

–

–

OH– (base)

–

–

–

0.48

2.95

CO2

–

0.39

0.32

0.11

0.11

CO2– 3

8.06

10.4

10.0

–

–

F

0.10

–

–

–

–

Cl

0.38

–

–

–

–

Sum

100

100.0

100.0

100.0

100.0

100⋅NBO/T

359

313

247

58

80.6

Note. The dash indicates no data. Average composition of kimberlite from the Udachnaya-East kimberlite pipe, Yakutia, H2O ≥ 1 wt.% (Kamenetsky et al., 2009). 2 Synthetic kimberlite melt after melting of the starting mixture (Ab Cal , wt.%) at T = 1300 K and at a CO pressure of 100 MPa (this paper). 38 62 2 3 Hydrous synthetic kimberlite melt (this paper). 4 Average composition of basalt from (Le Maitre, 1976). 5 Hydrous average composition of basalt (this paper). 1

decarbonization reaction with subsequent isobaric melt quenching. The thus obtained samples of quenched synthetic kimberlite melt (glass) 5.8 mm in diameter and ~3 cm high were removed from platinum capsules and crushed in an agate mortar, and the resulting powder was used for the synthesis of hydrous kimberlite melts (glasses) and for subsequent experimental measurements of the viscosity of hydrous kimberlite melts at moderate and high pressures. The chemical composition of the obtained synthetic kimberlite glasses (Table 1) and their homogeneity were determined using a Vega TS 5130 MM (CamScan MV2300) digital X-ray electron microscope equipped with an INCA Energy 200 energy dispersive microanalysis system at the Institute of Experimental Mineralogy (IEM), RAS. The data of microprobe analyses were standardized using the INCA Energy 200 program and the INCA program developed by A.N. Nekrasov. The concentration of the carbonate phase in the synthesized kimberlite melts (18 wt.%) and hence the CO−2 3 concentration (10.4 wt.%) in them were determined by calculating the mass balance for the decarbonization reaction in each experiment and by analyzing the IR spectra of quenched melts (glasses) (Fig. 1).

Hydrous kimberlite melts were synthesized by saturating water with the melt at a temperature of 1300 °C and a vapor pressure of 100 MPa. For this, anhydrous kimberlite glass powder was again placed in a platinum capsule 25–30 mm high with an open end, which was placed in the reactor of the above-mentioned internal device of the high-pressure gas vessel initially filled with distilled water. Hydration of the kimberlite melt at the indicated parameters was carried out for 3 h. After isobaric quenching of hydrous melts at a rate of ~300 °C/min to T = 850 °C and at a rate of ~100 °C/min to room temperature, the synthesized hydrous kimberlite glass columns were removed from the platinum capsules. The concentration of water dissolved in them (3 wt.%) was determined by Karl–Fischer titration using a KFT AQUA 40.00 instrument. These columns were used to prepare 6 mm glass columns with flat polished ends, which were inserted in platinum capsules with a flat bottom. At the open end of the column, small holes were made, in which platinum–rhodium spheres (typically 2) were placed. The end was covered with a thin plate (about 0.5 mm) of the same glass, then covered with a platinum lid, and sealed. The samples prepared in this

1096

E.S. Persikov et al. / Russian Geology and Geophysics 58 (2017) 1093–1100

Fig. 1. IR absorption spectra in synthesized glasses (synthetic kimberlite melts) with the carbonate peaks at 2520 and 2920 cm–1.

manner were used to measure melt viscosity in high-pressure devices of two types. The viscosity of hydrous kimberlite melts was measured by the quenching method of a falling sphere using high-pressure devices: (1) a high-pressure gas radiation viscometer (Persikov, 1991; Persikov and Bukhtiyarov, 2004); (2) a BARS type ultrahigh-pressure split-sphere multianvil apparatus (Sokol and Palyanov, 2008; Sokol et al., 2013). However, the capabilities of the radiation viscometer for measuring the melt viscosity could not be used in in situ experiments due to the extremely low viscosity of the synthetic kimberlite melts and hence the high falling speed of platinum spheres filled with 60Co isotope with a minimum diameter of 1.5 mm in the melt (Persikov, 1991, 1998). Therefore, the quenching version of the falling-ball method was used at both moderate (radiation viscometer) and high pressures (BARS type apparatus). In the experiments, platinum–rhodium spheres (Pt60 Rh40) with a diameter of 40–100 µm were used. Their falling speed in melts was determined from the time it takes for each sphere to travel a certain distance in the melt at the required temperature and pressure in the experiment. The time was measured from the moment the required TP parameters were obtained to the moment of isobaric quenching of the melt, and the distance traveled by the sphere in the melt during this time was determined in the quenched samples under a microscope. Temperature was measured by Pt70%Rh30%–Pt94%Rh6% thermocouples, which were used in both types of high-pressure devices. The temperature measurement errors did not exceed ±5 °C for a radiation viscometer and ±20 ºC (at 1300 ºC) and ±50 ºC (at 1950 ºC) for a lot of BARS type apparatus. As usual, the correction for the pressure effect on the EMF of the thermocouples was not evaluated when using the BARS apparatus. The relative error of pressure measurements on the radiation viscometer did not exceed ±1% and the error in determining the pressure on the BARS apparatus did not exceed ±0.1 GPa. Spheres (Pt60Rh40) 40–100 µm in diameter were produced by melting a fine platinum–rhodium wire with water quenching. The diameter and sphericity of the obtained

spheres were measured using a microscope before and after melt viscosity measurements. Melt viscosity at given experimental TP parameters was calculated from the well-known Stokes law with the Faxen correction for the wall effect (Persikov, 1991, 1998): η = 2gr 2∆ρ / 9v (1 + 3.3 r / h) × 1 − 2.104r /ra  + 2.09(r/ra)3 – 0.95(r/ra)5 , (1) where r is the radius of the sphere, cm; ra is the inner radius of the platinum capsule with melt, cm; h is the height of the capsule, cm; g is the acceleration due to gravity, 981 cm/s2; ∆ρ is the difference in density between the sphere and the melt, g/cm3; v is the falling speed of the sphere in the melt, cm/s; η is the melt viscosity at the TP parameters of the experiment, 0.1 Pa⋅s or Poise. The density of the melts was assumed to be equal to the densities of isobarically quenched melts (glasses), whose densities were obtained by hydrostatic weighing after the experiments at moderate (100 MPa) and high pressures (5.5 and 7.7 GPa). The following densities of the melts (glasses) were obtained: ρ = 2.85 g/cm3 (P = 0.1 GPa), ρ = 3.16 g/cm3 (P = 5.5 GPa), and ρ = 3.25 g/cm3 (P = 7.5 GPa). The subsequent correction of the melt density for temperature did not exceed 1% (Persikov and Bukhtiyarov, 2004). Considering the substantially higher density of the platinum–rhodium spheres (17.85 g/cm3), the estimated error of ∆ρ in equation (2) does not exceed ±5%, which agrees well the results of similar calculations in (Liebske et al., 2005). The calculated total measurement error of the melt viscosity did not exceed ±30 rel.%.

Results and discussion Currently available experimental-theoretical data on the temperature and pressure dependences of the viscosity of anhydrous depolymerized mafic (basalt) and ultramafic (pyroxenite, peridotite) melts at the TP parameters of their origin in the mantle are very limited (Liebske et al., 2005; Persikov

E.S. Persikov et al. / Russian Geology and Geophysics 58 (2017) 1093–1100

1097

Fig. 2. Temperature dependence of the viscosity of anhydrous and hydrous kimberlite and basaltic melts. The error of experimental and calculated data is ±30 rel.%; the predicted data on the temperature dependences of the viscosity of dry melts are taken from (Persikov et al., 2015). a, H2O pressure = 100 MPa; b, lithostatic pressure = 5.5 GPa. 1, anhydrous kimberlite (calculated data); 2, anhydrous basalt (calculated data), 3, hydrous basalt (calculated data), 4, hydrous kimberlite (experimental data, this study).

et al., 1989; Reid et al., 2003; Suzuki et al., 2005; Wan et al., 2007). Recently, we obtained the first experimental-theoretical data on the temperature and pressure dependences of the viscosity of anhydrous kimberlite and basaltic melts at the crustal and mantle thermodynamic parameters (Persikov et al., 2015). However, there are no experimental and theoretical data on the temperature and pressure dependences of the viscosity of hydrous mafic (basalt) and ultramafic (pyroxenite, kimberlite, peridotite) melts at the TP parameters of their origin in the mantle. Temperature dependence of the viscosity of hydrous kimberlite and basaltic melts. The temperature dependence of the viscosity is one of the most important characteristics of magmas since knowledge of this characteristic allows one to quantify their transport capacities and heat and masses transfer processes under different geodynamic settings in the Earth’s interior. The experimental data on the temperature dependence of viscosity for hydrous kimberlite melts (error ±30 rel.%) obtained in the present work were compared with calculated temperature dependences of viscosity for hydrous basaltic melts obtained with a nearly experimental error using our improved physicochemical model for predicting the viscosity of magmatic melts (Fig. 1) (Persikov and Bukhtiyarov, 2009; Persikov et al., 2015). The temperature dependence of the viscosity of hydrous kimberlite and basaltic melts is described by the Arrhenius–Frenkel–Eyring equation (equation 2) in the investigated range of temperatures and pressures: at 1300– 1800 °C and a moderate water pressure of 100 MPa (Fig. 2a) and at 1750–1950 °C and high lithostatic pressures of 5.5 GPa (Fig. 2b) and 7.5 GPa: ηPT = η0 exp (EPX/RT),

(2)

where η0 = 10–3.5 ± 100.1 (0.1 Pa⋅s or Poise) is the pre-exponential constant characterizing the melt viscosity at T → ∞ (Persikov, 1991); T is the temperature (K); R = 1.987 (cal/mol⋅K) is the universal gas constant; EPX is the activation energy of viscous flow (cal/mol), which is a function of the pressure and melt composition, including volatiles; ηPT is the melt viscosity at a given temperature and pressure (0.1 Pa⋅s, Poise). In accordance with equation (2), the temperature dependence of the viscosity of the studied melts is exponential, i.e., the viscosity of these melt decreases exponentially as the temperature increases and, vice versa, increases exponentially as the temperature decreases both at moderate water pressure (100 MPa) and at high lithostatic pressures (5.5 and 7.5 GPa). It is also found that the pre-exponential constant (η0) of equation (2) is constant and independent of melt composition, temperature, and pressure; this is consistent with theoretical findings (Frenkel, 1975) and previous data at more moderate pressures (Persikov, 1991, 1998; Persikov and Bukhtiyarov, 2004). On this basis, correct and temperature-independent values of the viscous-flow activation energy of these melts— their major structural-sensitive characteristic—were first obtained with high accuracy. For kimberlite melts, E = 138 ± 1.4 kJ/mol (P = 100 MPa) and E = 172 ± 1.7 kJ/mol (P = 7.5 GPa); for hydrous basaltic melts, E = 190 ± 1.9 kJ/mol (P = 100 MPa) and E = 150 ± 1.5 kJ/mol (P = 7.5 GPa). Furthermore, it is proved that the viscous-flow activation energy of hydrous kimberlite melts increases linearly with increasing pressure and that the viscous-flow activation energy of hydrous basaltic melts is inversely dependent on pressure (Fig. 3). Obviously, the activation energies of viscous flow cannot be determined with such high accuracy if (η0) in

1098

E.S. Persikov et al. / Russian Geology and Geophysics 58 (2017) 1093–1100

Fig. 3. Activation energies of viscous flow of anhydrous and hydrous kimberlite and basaltic melts as a function of pressure. The error of experimental and predicted data ±1 rel.%. The predicted data for dry melts are taken from (Persikov et al., 2015). For symbols see Fig. 2.

equation (2) is not constant, which is usually the case in many studies of the viscosity of magmatic melts in which (η0) in equation (2) is assumed to be not constant. For example, Liebske et al. (2005) measured the viscosity of synthetic peridotite melt in the pressure range of 2.8–13.0 GPa using a multianvil apparatus combined with synchrotron X-ray radiation. This interesting paper presents the calculated and constant value of the activation energy of anhydrous peridotite melts (108 ± 23 kJ/mol) in a wide range of lithostatic pressures (0–8 GPa), which contradicts our previous work (Persikov, 1991, 1998; Persikov and Bukhtiyarov, 2004) and the results of the present work on the linear pressure dependence of the activation energy for hydrous kimberlite and the inverse nature of this dependence for basaltic melts (Fig. 3). Pressure dependence of the viscosity of hydrous kimberlite and basaltic melts. The obtained results on the pressure dependence of the viscosity of hydrous and anhydrous kimberlite and basaltic melts are presented in Fig. 4. Analysis of these results shows that under isothermal conditions (T = 1800 °C), the viscosity of hydrous kimberlite melts increases exponentially by almost an order of magnitude as the pressure increases from 100 MPa to 7.5 GPa. New experimental data on the pressure dependence of the viscosity of hydrous kimberlite melts (error ±30 rel.%) are compared with similar dependences of the viscosity of hydrous basaltic melts (Fig. 4) obtained using our structural-chemical model for predicting and calculating the viscosity of magmatic melts (Persikov and Bukhtiyarov, 2009). It was first established that the viscosity of hydrous kimberlite melts at moderate pressure (P = 100 MPa) are more than one order of magnitude lower than the viscosity of anhydrous basaltic melts and nearly a half-order of magnitude lower than the viscosity of hydrous basaltic melts at comparable temperatures; at high pressure (P = 5.5–7.5 GPa), the pressure dependence of the viscosity of basaltic melts changes and the viscosity of hydrous

Fig. 4. Isothermal (1800 °C) pressure dependences of the viscosity of anhydrous and hydrous kimberlite and basaltic melts. The error of the experimental and predicted data is ±30 rel.%; the predicted data on the pressure dependences of viscosity for dry melts are taken from (Persikov et al., 2015). For symbols see Fig. 2.

kimberlite melts, in contrast, becomes 0.5 order of magnitude greater than the viscosity of dry and hydrous basaltic melts. Furthermore, the pressure dependence of the viscosity of hydrous basaltic melts under isothermal conditions (T = 1800 °C) has an inversion nature with a minimum at P ≈ 4.5 GPa, which is almost 1.0 GPa lower than the minimum viscosity for dry basaltic melts (Fig. 4). Effect of dissolved water on the viscosity of kimberlite and basaltic melts. Experimental-theoretical data on the viscosity of hydrous kimberlite melts, including at moderate and high pressures were not known. Recently, we obtained the first predicted data on the temperature and pressure dependences of the viscosity of anhydrous synthetic kimberlite melts at moderate (100 MPa) and high (7.5 GPa) pressures (Persikov et al., 2015). Previously, we have also obtained experimental-theoretical data at moderate pressures, according to which the viscosity and activation energy of viscous flow of mafic (basalt) melts decrease and those of pyroxenite (diopside) melts increase with increasing water pressure and hence with increasing water concentrations in such melts, which is related to the mechanism of H2O dissolution in melts of different basicity (Persikov, 1991, 1998; Persikov and Bukhtiyarov, 2004). It has been shown that in the range of mafic-ultramafic melts, water dissolution in melts occurs by two mechanisms: (1) chemical dissolution (OH– hydroxyl) and (2) physical dissolution (molecular H2O), and the numerical ratio of OH–/H2O depends on the basicity of the melts and the total concentration of H2O in them (Persikov, 1984, 1991, 1998; Persikov and Bukhtiyarov, 2004; Stolper, 1982). In accordance with the principles of acid-base interaction and the maximum polarity of chemical bonds (Persikov, 1998), water dissolving in basaltic melts is a base with respect to them as it is a donor of free oxygen, and its dissolution leads to their depolymerization, i.e., to in an increase in their basicity (K increases) and hence to a decrease in viscosity. Furthermore,

E.S. Persikov et al. / Russian Geology and Geophysics 58 (2017) 1093–1100

it has been shown that the maximum possible amount of chemically dissolved water (OH–) in basaltic melts is ~4.0 wt.%. In contrast, in ultramafic and pyroxenite melts, H2O has an amphoteric nature and is an acid with respect to such melts since H2O dissolution in them causes melt polymerization, i.e., a reduction in melt basicity (K decreases, Table 1) and hence an increase in melt viscosity. The maximum possible amount of chemically bound water (OH–) dissolved in ultramafic melts is ~1.5–2.0 wt.% (Persikov, 1998; Persikov and Bukhtiyarov, 2004). According to the new data obtained, water dissolution (~3 wt.%) in kimberlite melts has no significant effect on the change in the viscosity of these melts within the error (±30 rel.%) (Figs. 2 and 4) at both moderate (100 MPa) and high pressures (7.5 GPa). In contrast, the viscosity of basaltic melts decreases considerably with water dissolution in them at moderate pressures (100 MPa), whereas at high pressures (P > 3.5 GPa), the effect of dissolved water is eliminated. Furthermore, the inversion pressure dependence of the viscosity of basaltic melts is retained and the pressure at the minimum points of their viscosity and activation energy is considerably reduced: Pmin = 5.5 GPa for anhydrous melts and Pmin = 4.5 for hydrous melts (Figs. 2 and 4). The mechanism of this new and quite unexpected phenomenon is due to the significantly greater effect of pressure on viscosity and viscous-flow activation energy compared to the effect of dissolved water for kimberlite melts over the entire investigated range of pressure and water content, and for basaltic melts, at pressures exceeding the pressure at the point of minima of their viscosity and viscous-flow activation energy (Figs. 3 and 4).

Conclusions The viscosities of hydrous kimberlite and basaltic melts were found to depend exponentially on temperature and pressure over a wide range of temperatures (1300–1950 °C) and at pressures up to 7.5 GPa. Reliable values of the activation energy for hydrous kimberlite and basaltic melts at high pressures were first established. It is proved that the activation energies of hydrous kimberlite melts increase linearly with increasing pressure and that the activation energy of hydrous basaltic melts are inversely dependent on pressure with a minimum at P = 4.5 GPa. It is first established that water dissolution (up to about 3 wt.%) in kimberlite melts within measurement errors does not have a significant effect on the change in their viscosity at both moderate (100 MPa) and high pressures (7.5 GPa). In contrast, the viscosity of basaltic melts considerably decreases with water dissolution in them at moderate pressures (100 MPa), but at high pressures (P > 3.5 GPa), the effect of dissolved water also becomes insignificant. It is found that at comparable temperatures, the viscosity of hydrous ultramafic kimberlite melts at moderate pressures (100 MPa) is nearly an order of magnitude lower than the viscosity of mafic hydrous basaltic melts, whereas at high pressures (5.0–7.5 GPa), it is more than 2 times higher, due

1099

to the features of the effect of pressure on the viscosity of such melts. We are grateful to A.N. Nekrasov and G.V. Bondarenko (IEM RAS) for help in the microprobe analyses and IR spectroscopy of samples. We are also grateful to V.N. Sharapov and an anonymous reviewer for helpful comments which improved the manuscript. The study was supported by the Russian Foundation for Basic Research, grant No. 15-05-01318, and by the Russian Science Foundation, grant No. 14-27-00054.

References Allwardt, J.R., Stebbins, J.F., Terasaki, H., Du, L.S., Frost, D.J., Withers, A.C., Hirschmann, M.M., Suzuki, A., Ohtani, E., 2007. Effect of structural transitions on properties of high-pressure silicate melts: Al-27 NMR, glass densities, and melts viscosities. Am. Mineral. 92, 1093–1104. Brearley, M., Dickinson, J.E. Jr., Scarfe, M., 1986. Pressure dependence of melt viscosities on the join diopside–albite. Geochim. Cosmochim. Acta 30, 2563–2570. Chepurov, A.I., Sonin, V.M., Kirdyashkin, A.A., Zhimulev, E.I., 2009. Use of a pressless multianvil split-sphere apparatus to measure the silicate melt viscosity. J. Appl. Mech. Techn. Phys. 58 (5), 826–830. Chepurov, A.A., Pokhilenko, N.P., 2015. Experimental estimation of the kimberlite melt viscosity. Dokl. Earth Sci. 462 (2), 592–595. Dalton, J.A., Presnall, D.C., 1998. The continuum of primary carbonatitickimberlite melt compositions in equilibrium with lherzolite: data from the system CaO-MgO-Al2O3-SiO2-CO2 at 6 GPa. J. Petrol. 39, 1953–1964. Dasgupta, R., Hirschmann, M.M., 2006. Melting in the Earth’s deep upper mantle caused by carbon dioxide. Nature 440, 659–662. Dingwell, D.B., Copurtial, P., Giordano, D., Nichols, A.R.L., 2004. Viscosity of peridotite liquid. Earth Planet. Sci. Lett. 226, 127–138. Fujii, T., Kushiro, I., 1977. Density, viscosity, and compressibility of basaltic liquid at high pressures. Carnegie Inst. Year Book, Vol. 76, pp. 419–424. Frenkel, Ya.I., 1975. Kinetic Theory of Liquids [in Russian]. Izd. Akad. Nauk SSSR Moscow. Giordano, D., Romano, C., Papale, P., Dingwell, D.B., 2004. The viscosity of trachytes, and comparison with basalts, phonolites, and rhyolites. Chem. Geol. 213, 49–61. Kamenetsky, V.S., Yaxley, G.M., 2015. Carbonate–silicate liquid immiscibility in the mantle propels kimberlite magma ascent. Geochim. Cosmochim. Acta 158, 48–56. Kamenetsky, V.S., Kamenetsky, M.B., Weiss, Y., Navon, O., Nielsen, T.F.D., Mernagh, T.P., 2009. How unique is the Udachnaya-East kimberlite? Comparison with kimberlites from the Slave Craton (Canada) and SW Greenland. Lithos 112S, 334–346. Kavanagh, J.L., Sparks, R.S.J., 2009. Temperature changes in ascending kimberlite magma. Earth Planet. Sci. Lett. 286, 404–413. Kopylova, M.G., Matveev, S., Raudseep, M., 2007. Searching for parental kimberlite melt. Geoch. Cosmoch. Acta 71, 3616–3629. Kushiro, I., 1980. Viscosity, density and structure of silicate melts at high pressures, and their petrological applications, in: Physics of Magmatic Processes. Princeton University Press, New Jersey, pp. 93–120. Lange, R.A., 1994. The effect of H2O, CO2, and F on the density and viscosity of silicate melts, in: Carrol, M.R., Holloway, J.R. (Ed.), Reviews in Mineralogy, Vol. 30: Volatiles in Magmas. MSA, Washington, pp. 331– 369. Le Maitre, R.W., 1976. The Chemical variability of some common igneous rocks. J. Petrol. 17, Part 4, 589–637. Liebske, C., Schmickler, B., Terasaki, H., Poe, B.T., Suzuki, A., Funakoshi, K.I., Ando, R., Rubie, D.C., 2005. The viscosity of peridotite liquid at pressures up to 13 GPa. Earth Planet. Sci. Lett. 240, 589–604. Lebedev, E.B., Khitarov, N.I., 1979. Physical Properties of Magmatic Melts [in Russian]. Nauka, Moscow.

1100

E.S. Persikov et al. / Russian Geology and Geophysics 58 (2017) 1093–1100

Michell, R.H., 2008. Petrology of hypabyssal kimberlites: relevance to primary magma compositions. J. Volcanol. Geotherm. Res. 174, 1–8. Mysen, B.O., 1988. Structure and Properties of Silicate Melts. Elsevier, Amsterdam. Neuville, D.R., Richet, P., 1991. Viscosity and mixing in molten (Ca, Mg) pyroxenes and garnets. Geochim. Cosmochim. Acta 55, 1011–1019. Persikov, E.S., 1984. Viscosity of Magmatic Melts [in Russian]. Nauka, Moscow. Persikov, E.S., 1991. The viscosity of magmatic liquids: experiment, generalized patterns; a model for calculation and prediction: application, in: Perchuk, L.L., Kushiro, I. (Eds.), Physical Chemistry of Magmas. Adv. Phys. Geochem. Springer-Verlag, New York, Vol. 9, pp. 1–40. Persikov, E.S., 1998. Viscosity of model and magmatic melts pressures ant temperatures of the Earth’s crust and upper mantle. Geologiya i Geofizika (Russian Geology and Geophysics) 39 (12), 1793–1804 (1780–1792). Persikov, E.S., 2007. Structural chemical model to calculate and predict the viscosity of magmatic melts in full range of compositions and conditions, in: General Assembly EGU-2007. Vienna. Geophysical Research. Abstracts, A-02262. SRef-ID: 1607-7962/gra/EGU2007-A-02262C European Geosciences Union 2007, Vol. 9. Persikov, E.S., Bukhtiyarov, P.G., 2004. Experimental investigation of the effect of lithostatic and water pressures on the viscosity of silicate and magmatic melts. The new structural-chemical model for calculating and predicting their viscosity, in: Zharikov, V.A., Fed’kin, V.V. (Ed.), Experimental Mineralogy: Some Results at the Turn of the Century [in Russian]. Nauka, Moscow, Vol. 1, pp. 103–122. Persikov, E.S., Bukhtiyarov, P.G., 2009. Interrelated structural chemical model to predict and calculate viscosity of magmatic melts and water diffusion in a wide range of compositions and TP parameters of the Earth’s crust and upper mantle. Russian Geology and Geophysics (Geologiya i Geofizika) 50 (12), 1079–1090 (1393–1408). Persikov, E.S., Bukhtiyarov, P.G., Kalinicheva, T.V., 1987. Effects of composition, temperature, and pressure on the fluidity of magmatic melts. Geokhimiya, No. 4, 483–498. Persikov, E.S., Kushiro, I., Fujii, T., Bukhtiyarov, P.G., Kurita, K., 1989. Anomalous pressure effect on viscosity of magmatic melts, in: Phase Transformation at High Pressures and High Temperatures: Applications to Geophysical and Petrological Problems. Misasa, Tottori-ken, Japan, DELP, International Symposium, pp. 28–30. Persikov, E.S., Zharikov, V.A., Bukhtiyarov, P.G., Pol’skoy, S.F., 1990. The effect of volatiles on the properties of magmatic melts. Eur. J. Mineral. 2, 621–642. Persikov, E.S., Bukhtiyarov, P.G., Nekrasov, A.N., 2010a. Water diffusion in basalt and andesite melts at high pressures. Geokhimiya, No. 3, 3–15. Persikov, E.S., Newman, S., Bukhtiyarov, P.G., Nekrasov, A.N., Stolper, E.M., 2010b. Experimental study of water diffusion in haplobasaltic and haploandesitic melts. Chem. Geol. 276, 241–256. Persikov, E.S., Bukhtiyarov, P.G., Nekrasov, A.N., Bondarenko, G.V., 2012. Experimental study of the melting of silicate-carbonate systems at high temperatures and pressures, in: Proceedings of the XIII International Conference on Physicochemical and Petrophysical Research in the Earth Sciences. Nauka, Moscow, pp. 217–220.

Persikov, E.S., Bukhtiyarov, P.G., Sokol, A.G., 2015. Change in the viscosity of kimberlite and basaltic magmas during their origin and evolution (prediction). Russian Geology and Geophysics (Geologiya i Geofizika), 56 (6), 885–892 (1131–1115). Price, S.E., Russell, J.K., Kopylova, M.G., 2000. Primitive magma from the Jericho Pipe, N.W.T., Canada: constraints on primary kimberlite melt chemistry. J. Petrol. 47, 789–808. Reid, J.E., Suzuki, A., Funakoshi, K.I., Terasaki, H., Poe, B.T., Rubie, D.C., Ohtani, E., 2003. The viscosity of CaMgSi2O6 liquid at pressures up to 13 GPa. Phys. Earth Planet. Inter. 139, 45–54. Scarfe, G.M., 1986. Viscosity and density of silicate melts, in: Scarfe, G.M. (Ed.), Silicate Melts, Min. Assoc. Can. Short Course Handbook, Vol. 12, pp. 36–56. Scarfe, C.M., Mysen, B.O., Virgo, D., 1987. Pressure dependence of the viscosity of silicate melts, in: Mysen, B.O. (Ed.), Magmatic Processes: Physicochemical Principles. Geochem. Soc. Spec. Publ. Vol. 1, pp. 59–68. Sharygin, I.S., Litasov, K.D., Shatskii, A.F., Golovin, A.V., Otani, E., Pokhilenko, N.P., 2013. Melting of kimberlite of the Udachnaya-East pipe: Experimental study at 3–6.5 GPa and 900–1500 °C. Dokl. Earth Sci. 448 (2), 200–205. Shaw, H.R., Wright, T.L., Peck, D.L., Okamura, R., 1968. The viscosity of basaltic magma: An analysis of field measurements in Makaopuhi lava lake, Hawaii. Am. J. Sci. 266, 225–264. Sokol, A.G., Palyanov, Y.N., 2008. Diamond formation in the system MgO–SiO2–H2O–C at 7.5 GPa and 1600 °C. Contrib. Mineral. Petrol. 121, 33–43. Sokol, A.G., Kupriyanov, I.N., Palyanov, Y.N., 2013. Partitioning of H2O between olivine and carbonate-silicate melts at 6.3 GPa and 1400 °C: Implications for kimberlite formation. Earth Planet. Sci. Lett. 383, 58–67. Sparks, R.S.J., Baker, L., Brown, R.J., Field,M., Schumacher, J., Stripp, G., Walters, J., 2006. Dynamical constraints of kimberlite volcanism. J. Volcanol. Geotherm. Res. 155, 18–48. Sparks, R.S.J., Brooker, R.A., Field, M., Kavanagh, J., Schumacher, J.C., Walter, M.J., White, J., 2009. The nature of erupting kimberlite melts. Lithos 1125, 429–438. Stolper, E.M., 1982. The speciation of water in silicate melts. Geochim. Cosmochim. Acta. 46, 2609–2620. Suzuki, A., Ohtani, E., Terasaki, H., Funakoshi, K., 2005. Viscosity of silicate melts in CaMgSi2O6_NaAlSi2O6 system at high pressure. Phys. Chem. Miner. 32, 140–145. Yoder, H.S., Jr., 1976. Generation of Basaltic Magmas. National Academy of Sciences, Washington D.C. Wan, J.T.K., Duffy, T.S., Scandolo, S., Car, R., 2007. First-principle study of density, viscosity and diffusion coefficients of liquid MgSiO3 at conditions of the Earth’s deep mantle. J. Geophys. Res. 112, 1–7. Whittington, A., Richet, P., Holtz, F., 2000. Water and viscosity of depolimerized aluminosilicate melts. Geochim. Cosmochim. Acta 64, 3725–3736. Wolf, G.H., McMillan, P.F., 1995. Pressure effects on silicate melt structure and properties, in: Reviews in Mineralogy, Vol. 32: Structure, Dynamics and Properties of Silicate Melts. MSA, Washington, pp. 505–562. Wyllie, P.J., 1980. The origin of kimberlite. J. Geophys. Res. 85, 6902–6910.

Editorial responsibility: O.P. Polyansky