Update of the solid reference substances

The mean molar concentration of the elements in the upper continental crustε_j of Eq. 5.3used in Szargut[336], was the recommended by Polanski and Smulikowski

[268]. Ranz [276] used updated values mainly from Taylor and McLennan [354],

[353]. For the elements: Br, C, Cl, F, S, P t, Pu, Ra, Rh, Ru, Te, I, H g and N,

Taylor and McLennan did not provide any information, therefore, Ranz used the values given by Wedepohl [404] for S, Br, C, F, I, H g, N and for the remaining

elements, the values used by Szargut[336]. Some authors like Plank and Langmuir

[267] basing on their studies on marine sediments, suggested already in 1998 some

revisions of the estimated values by Taylor and McLennan [354], [353] for N b, Cs,T i, Ta. As a consequence, McLennan[215] published in year 2001 new mean

molar concentrations of the upper continental crust for the elements: Sc, T i, V , C o,

N i, N b, Cs, P b, Ta. The most recent data about the chemical composition of the upper continental crust has been published by Rudnick and Gao[292], taking into

account the studies published so far.

The recent values provided by Rudnick and Gao will be used for the update of Szargut’s R.E. Nevertheless, values for Pu and Ra that are not provided in their tables, will be assumed to be the ones given by Polanski[268].

As explained in chapter 3, Grigor’ev published in year 2000[125] the average mi-

neral content of the upper continental crust obtained through a great number of quantitative mineralogical analysis of important rocks. In 2007, Grigor’ev updated this information; the new analysis comprises 265 minerals, their varieties and their non-mineral materials, corresponding to 99,13% of the total mineral content of the upper continental crust. With this valuable information, we have been able to pro- pose a new model of the continental crust, based on Grigor’ev’s composition, but assuring the mass balance of the earth. This information allows to obtain directly the standard molar concentration of the following 14 reference substances in the solid environment without using Eq. 5.3: Al2SiO5, BaSO4, Be2SiO4, C aCO3, Au,

Fe₂O₃, M g₃Si₄O₁₀(OH)₂, M nO₂, SiO₂, S r CO₃, T hO₂, SnO₂, T iO₂, Z rSiO₄. For the rest substances, Eq. 5.3must be used, takingε_j from the latest geochemical publi- cations explained before.

For the fraction of the j-th element appearing in the form of reference species (coeffi- cient c_j), Szargut[335] associates values comprised between 0,5 for more abundant

substances and 0,001 for less frequent substances from geochemical data given by Polanski and Smulikowski[268]. Ranz [276] obtained more accurate c_jcoefficients for solid R.S. containing the most abundant elements in the upper continental crust. For this purpose, she used the mineralogical composition of the earth’s upper layer obtained with the CIPW norm before and updated geochemical information, mainly from Taylor and McLennan[354]. For minority elements, due to the lack of informa-

tion, Szargut’s[335] values were used. As long as a better mineralogical composition

of the earth’s crust is not developed and the c_j coefficients are recalculated with this information, we will assume the c_j values obtained by Ranz[276].

The mean molecular mass of the upper layer of the continental part of the earth’s crust, was first estimated by Szargut[334]. The obtained value was MW_{c r}= 135,5

kg/kmole, applying the following estimation method: according to the geochemical data, the mean concentration values (in mole/kg) of particular chemical groups or elements in the external layer of the continental earth’s crust and the chemical compound formed from these groups were assumed. The first considered group was

CO2, which appears in the earth’s crust mainly as the carbonates of C a, M g and Fe.

Per 1 mole of (C aO+ M gO + FeO) 0,035 mole of CO₂ is present. The group CO₂ was partitioned between the mentioned groups and elements Z n, Cu, P b and C d, appearing also in the form of carbonates. The group SO₃ was partitioned between

C aOand M gO forming sulphates. It was assumed that a prevailing part of metals (Sn, C o, M n, Fe, N i) appears in the form of different oxides (C o2O3, C o3O4, Fe2O3,

Fe3O4). It was also assumed that 8% of Fe appears in the form of the free oxide

Fe2O3. The remaining part appears in the form of FeT iO3, FeC r2O3and silicates. For

example, the following silicates were assumed: N aAlSi3O8, KAlSi3O8, N aFeSi2O6,

M gSiO₃, C aO.Al₂Si₂O₇. Because of the large content of SiO₂, a considerable part of it was assumed in the free form. After estimating the composition of a mean sample of the lithosphere, its molecular mass was calculated.

Ranz [276] updated the molecular mass of the upper continental crust using more

recent geochemical information and adopting not only a geochemical approach, but also a geological one. The methodology used was as follows: the international ac- cepted norm CIPW [262] was applied to the mass fractions of the principal oxide

groups obtained by Carmichael[49] for the cratonic and sedimentary layers, in order

to redistribute the chemical components from the oxides to the mineral molecules that are representative in real minerals appearing in a rock. Next, the minerals of the norm and their respective relative masses were modified to adjust them to the real volumes of the principal groups of each rock. Finally, their molar fractions were calculated and the mean molecular mass of the whole was obtained. The result- ing M W_{c r} was equal to 145,5 g/mole. Even though this methodology used better geochemical values than the ones in Szargut [334], and included the geological

approach, we cannot forget that the CIPW norm is an artificial way to obtain the possible minerals that can appear in a rock. It is therefore only an approximation as well.

In the light of Grigor’ev’s analysis, a more accurate molecular weight of the upper continental crust, based on experimental results rather than assumptions, can be easily obtained. The new calculated value is M W_{c r}= 142,1 g/mole, which is very close to the estimation done by Ranz. Our model threw up a mean molecular weight of the upper crust of M W_{c r}=155,2 g/mole.