Geothermics xxx (xxxx) 102661

Contents lists available at ScienceDirect

Geothermics
journal homepage: www.elsevier.com/locate/geothermics

F
An update on lithium mica thermodynamics and its geothermometrical
application

OO
Tiziano Boschetti
Department of Chemistry, Life Sciences and Environmental Sustainability, University of Parma, Parco Area delle Scienze, 43124 Parma, Italy

ARTICLE INFO ABSTRACT

PR
Keyword: The extraction of lithium from geothermal brines is one of the most promising fields for mining this important el-
Lithium ement, which plays and will play an increasingly important role in the transition to green energy. Despite this, a
Geothermal brines thermodynamic database of lithium minerals –which would be useful in geochemical modeling of Li cycling- is
Thermodynamics
still far from complete. In this short communication, the thermodynamic (in particular molar heat equation) and
Geochemical codes and databases
solubility data of a lithium mica mineral with a generic formula KLi0.5Al1.5Si4O10(OH)2 are revised. Furthermore,
the solubility data as log(K) are inserted in commonly available thermodynamic databases and applied to the ge-
othermometrical study of the lithium-rich geothermal waters of the Upper Rhine Graben. The general agreement
D
between the temperatures obtained from the equilibrium reaction involving lithium mica and the values already
published or from a multicomponent geothermometrical approach confirms the usefulness and importance of
lithium mica in the study of lithium-rich geothermal brines.
TE

1. Introduction were found to be possible solid phases that could be used as geother-
mometers in brackish and brine geothermal waters (Boschetti, 2022a,
With an expected increased demand of approximately 500% in the b). This is basically in agreement with the conclusions of Sanjuan et
next 30 years, lithium is a high-value metal that is the principal compo- al.’s works (Sanjuan et al., 2022, 2014), who proposed a lithium mica
nent of high-energy-density batteries and a strategic resource for clean as LiAl3Si3O10(OH)2 on the basis of the equilibria that regulate the Na/
EC

energy transitions (Boschetti, 2022b; Stringfellow and Dobson, 2021). Li geothermometers. However, a mineral with that composition does
Even though extracting lithium from brines using evaporation ponds is not correspond to the generic formula of lithium micas (Boschetti,
nominally inexpensive, the process is time consuming, land intensive, 2022a).
and wasteful of water. In contrast, geothermal energy production en- The log(K) values of Li-minerals solubility reactions at different tem-
tails that lithium be recovered from brines without significant losses of peratures can be added to the thermodynamic databases of currently
water, mainly because a large volume of brine is processed in a geother- available geochemical modeling software, such as The Geochemist's
RR

mal power plant and typically reinjected. Therefore, geothermal water Workbench® (GWB) (Bethke et al., 2022). The default database file
could be a valuable Li resource. It must be reinjected to maintain en- thermo.dat of GWB is based on an old LLNL (Lawrence Livermore Na-
ergy production at geothermal facilities, and reinjection provides a safe tional Laboratory) database included in the EQ3/6 software package,
means of disposing of the other produced brine constituents back into version 3245, which dates back to 1986 (Bethke et al., 2022; Delany
the reservoir (Stringfellow and Dobson, 2021). In previously published and Lundeen, 1991; Wolery et al., 1988). Unfortunately, the thermo.dat
manuscripts, a strong emphasis has been given to the possibility of us- file of GWB does not contain the specific literature source of thermody-
CO

ing thermodynamic data of lithium minerals to study the lithium geo- namic properties (e.g., ΔfG°, ΔfH°, S°, V°, Cp) for each basis aqueous
chemistry in geothermal brines (Boschetti, 2022a, b). Although the pro- specie and mineral, and these source data for the old LLNL datasets are
posed lithium mineral review was far from complete, it showed how no longer available (T. Wolery, pers. comm.). Most mineral thermody-
deep temperatures can be deduced by calculating the equilibrium con- namic data available between 1985 and 1992 are mainly based on
stant [log(K)] of specific reaction of selected lithium minerals, whose Helgeson et al. (1978). However, the thermodynamic properties of min-
presence in the reservoir was detected or at least hypothesized erals (Blanc et al., 2012; Holland and Powell, 2011) and aqueous basis
(Boschetti, 2022a, b). Among all proposed Li-bearing solid phases, eph- species such as Al+3 and SiO2(aq) or H4SiO4 (Apps and Spycher, 2004;
esite and a lithium mica with a generic formula KLi0.5Al1.5Si4O10(OH)2 Stefansson, 2001; Tagirov and Schott, 2001; Wolery and Colon, 2017)

E-mail address: tiziano.boschetti@unipr.it.

https://doi.org/10.1016/j.geothermics.2023.102661
Received 10 October 2022; Received in revised form 18 December 2022; Accepted 18 January 2023
0375-6505/© 20XX
T. Boschetti Geothermics xxx (xxxx) 102661

have been updated since then. Therefore, the internal consistency of the values of Li-mica solubility reaction has been transferred to the three
thermo.dat file could be broken when calculating log(K) of reactions us- above-mentioned targeted databases. It should be noted that both the
ing more recent source thermodynamic properties for involving new thermo2021.dat of PyGCC and GWB_Thermod-
minerals and/or aqueous species using these new properties. demV1.10_15Dec2020.dat databases have a format compatible with re-
This short note examines the integration of the thermodynamic data lease 7 of the GWB (Bethke and Yeakel, 2007). Furthermore, the log(K)
of lithium mica in more recent thermodynamic databases for which values of specific reactions have been calculated at different tempera-
source thermodynamic properties (ΔfG°, ΔfH°, S°, V°, Cp) are available tures and pressures by the Organobiogeotherm program taking any ad-
and well documented. As an example, these revised data are applied to ditional minerals from the targeted databases to gain consistent results.
study the chemical composition of geothermal brines from the France As in Boschetti (2022a), the activity of dissolved aqueous species
and Germany Upper Rhine Graben (URG) boreholes, most of them was computed from the previously published chemical composition of

F
drawing brines from the granitic basement (Sanjuan et al., 2016). The the URG geothermal brines (Aquilina et al., 1997; Bosia et al., 2021;
URG represents one of the most promising sites in Europe for the extrac- Dezayes et al., 2013; Pauwels et al., 1993; Sanjuan et al., 2022, 2016,
tion of lithium from geothermal brines. Moreover, previously published 2020) (Supplementary File 1), using the PhreeqcI code, version 3

OO
hydrogeochemical data from this site are well-documented and numer- (Parkhurst and Appelo, 2013), and the pitzer.dat thermodynamic data-
ous (Sanjuan et al., 2022). The boreholes (Bruchsal, Cronenbourg, Lan- base modified with lithium Pitzer parameters recalculated from Lassin
dau, Insheim, Soultz, Rittershoffen, Vendenheim) were drilled to depths et al. (2015). The equilibrium temperature between the URG geother-
of 2540 to 5000 m. The Triassic Buntsandstein sandstone at depths of ≥ mal brines and a selected minerals assemblage was calculated by GeoT,
4 km in the graben center could represent the potential reservoir of the a computer program for multicomponent solute geothermometry
brine at 225 ± 25 °C, even though the maximum deep on-site tempera- (Spycher et al., 2014). The temperature results were compared with the
ture measured so far has only been 200 °C (Sanjuan et al., 2016). How- previously published values obtained both by measurement at the well

PR
ever, the inferred fluid flow is debated – is it from the graben shoulders bottom and chemical geothermometers. In addition, the temperatures
to the rift center, i.e., from the Carboniferous two-mica granite (Vidal were also obtained comparing the equilibrium constant log(K) of possi-
and Genter, 2018) basement into overlaying sediments like the ble reactions involving Li-micas (using the three above-mentioned data-
Buntsandstein (Druppel et al., 2020), or vice-versa (Sanjuan et al., bases) and the ion activity product [log(Q)] of the brine samples calcu-
2022), or both (Komninou and Yardley, 1997)? lated using the Pitzer formalism and the PhreeqcI code, i.e., equilibrium
temperature at log(K) = log(Q).
2. Methods
3. Results and discussion
D
The thermodynamic data of the mica ephesite were taken from
Boschetti (2022b). That dataset mainly consists of Gibbs free energy of The thermodynamic parameters of the Li-mica, along with the log
formation (ΔfG°), enthalpy of formation (ΔfH°), entropy (S°), molar vol- (K) of solubility reaction calculated for the three thermodynamic data-
ume (V°), and the three regression coefficients (a, b, and c) unique to bases (Thermoddem, PyGCC and SUPCRTL), are shown in Table 1a and
TE

the specified mineral to calculate the Maier-Kelly molar heat capacity b, respectively. The standard deviations of the log(K) values between
(Cp) equation at a specified temperature T (Johnson et al., 1992; Maier the different databases is up to ±0.44 (Table 1b). As mentioned in the
and Kelly, 1932): introduction, this is quite normal considering the difference between
the previously published thermodynamic parameters of aluminum and
(1) dissolved silica. However, the differences in log(K) jump up to 2.7 at
EC

300 °C when comparing the values in Table 1b and those previously

For the generic Li-mica KLi0.5Al1.5Si4O10(OH)2, only the logarithm of published (Boschetti, 2022b). Such a difference is mainly because in the
the equilibrium constants (log(K)) of the mineral solubility reactions latter work the original log(K) values of lithium mica were obtained by
were published in Boschetti (2022b). In that study, the log(K) values of using the constant Cp value at 25 °C, i.e., the constant Cp approxima-
Li-mica in the 0–300 °C temperature range were kindly provided by tion according to the ClayTherm code (Blanc et al., 2021). The new Cp
Blanc (pers. comm.), who calculated these values by the ClayTherm values at different temperatures calculated by the three different codes
computing tool (Blanc et al., 2021). In ClayTherm, a constant Cp ap- and the fitting of the Maier-Kelly equation used in this work are pre-
RR

proximation and thermodynamic properties of ions consistent with the sented in Supplementary File 2. The use of the new thermodynamic
Thermoddem database are used (Blanc et al., 2012). In the present data in the study of URG geothermal brines is discussed below.
study, the thermodynamic parameters ΔfG°, ΔfH°, S°, and V° of the theo- The previously published lithium concentrations detected in the
retical Li-Mica were kept as obtained by ClayTherm because they are URG brines (TDS = 105 ± 9 mg/l) are between 159 and 210 mg/l,
substantially similar to those published by Gordiyenko and with a mean of 178 ± 16 mg/l (Supplementary File 1). The calculated
Ponomareva (1989). On the contrary, the Cp values at different temper- temperatures at depth by GeoT to obtain simultaneous equilibrium with
atures of the Li-mica were averaged from the results obtained by three
CO

albite-low, K-feldspar and quartz are between 157 and 231 °C, with a
different codes which use different internal thermodynamic datasets: mean of 187 ± 30 °C (TminRMED in Table 2). The mineral assemblage
Thermomin (Van Hinsberg et al., 2005), FactSage 8.2 (Bale et al., 2016) used for calculation is the same as that suggested by Alsemgeest et al.
and HSC Chemistry 7.1 (Roine, 2009). After that, the heat capacity co- (2021), who calculated a temperature of 229 ± 7 °C and 225 ± 25 °C
efficients of Eq. (1) were obtained by OriginPro 2021 (OriginLab Corpo- for Soultz brine (GPK-2 borehole 5000 m depth in the Carboniferous
ration, Northampton, MA, USA) using the non-linear fitting tool. granite) by a multicomponent geothermometrical approach similar to
The Thermoddem GWB_ThermoddemV1.10_15Dec2020.dat (Blanc GeoT, but using the PhreeqcI code and two different thermodynamic
et al., 2012), SUPCRTBL (Zimmer et al., 2016), and speq21.dat/ther- databases. Alsemgeest et al. (2021) also advised that: i) slightly lower
mo2021.dat of the PyGeochemCalc (PyGCC) tool (Awolayo and Tutolo, temperatures with higher standard deviation were obtained by differ-
2022) were the three targeted thermodynamic databases. To produce ent mineral assemblages, and ii) it is challenging to determine equilib-
internally consistent databases and to compare them, the log(K) values rium temperature due to the high ionic strength of the geothermal brine
were calculated using OrganoBioGeoTherm, a GUI program to use and the consequent limitations in the calculation of activity coefficients
SUPCRT92 code (Johnson et al., 1992) in Microsoft Windows. In addi- by an extended Debye-Hückel equation with both GeoT and Phreeqc. In
tion to the thermodynamic parameters of the Li-mica, all the necessary this study, a temperature of 231 °C and 200 °C were obtained by GeoT
parameters of the ions occurring in the solubility reactions were taken using the 2013 (Sanjuan et al., 2016) and 2021 (Bosia et al., 2021)
from the three above mentioned database. After calculations, the log(K)

2
 T. Boschetti
Table 1
– Thermodynamic parameters (1 cal = 4.184 J) (a) and solubility data as the base-ten logarithm of equilibrium constant (log(K)) at different temperature and water saturation pressure (b) of the theoretical
lithium mica. In (b), the aqueous solute species of the three different thermodynamic databases were used for calculations of log(K) values: Thermoddem (Blanc et al., 2012), PyGCC (Awolayo and Tutolo,
2022), and SUPCRTBL (Zimmer et al., 2016). Best-fit parameters of the Maier-Kelly molar heat capacity (Cp) equation as in Supplementary File 2. See the method section for details.
a - thermodynamic parameters

name type formula ΔfH°

G° S° *Cp = a + bT + cT−2
V°

a(100) b(103) c(10−5)

cal mol−1 cal mol−1 cal 3mol−1

cm cal mol−1 cal K
K−1 mol
K−1 −1 K−2 mol−1

ClayTherm (Blanc et al. 2021) this work

Li-mica mica KLi0.5Al1.5Si4O10(OH)2 -1384933.08

-1296883.37 77.182 94.486
144.32 30.360 -21.659
*Cp(25°C) = 80.03 cal mol−1 K−1 by ClayTherm (Blanc et al. 2021), Cp(25°C) = 77.29 ± 1.76 cal mol−1 K−1 in this study (Supplementary File 2).
b - log(K) values at different temperature using different thermodynamic databases; P = 1.01314 bar at 0 ≤ T ≤ 100°C and vapour saturation pressure at 100 < T ≤ 300°C
name reaction T°C 25
0 60 100
150 200 250
300 database
P bar 1.01314 1.01314 4.76000
1.01314 15.54900 39.77600
85.92700
Li-mica KLi0.5Al1.5Si4O10(OH)
6.0483
2 + 6H
+ 1.6382
3.9208 -0.2866 -2.0659
-3.4706 -4.7515 PyGCC
-6.1511
=
=
K+ + 0.5Li+ + 1.5Al+3 + 4SiO2 + 4H2O
Li-mica KLi0.5Al1.5Si4O10(OH)2 + 6H+ = 3.8622
5.2808 1.9319 -1.6258
0.1216 -3.0104 -4.2479
-5.5682 SUPCRTBL
= K+ + 0.5Li+ + 1.5Al+3 + 4SiO2 + 4H2O

F
Li-mica KLi0.5Al1.5Si4O10(OH)2 + 6H+ + 4H2O = 3.9273
5.3057 2.0444 -1.6095
0.2264 -3.1338 -4.5154
-5.9007 Thermoddem

Geothermics xxx (xxxx) 102661

= K+ + 0.5Li+ + 1.5Al+3 + 4H4SiO4

O
std.dev. 0.04
0.44 0.21 0.26
0.27 0.24 0.36
0.41

R O
P
D
T. Boschetti Geothermics xxx (xxxx) 102661

Table 2
– Comparison of the estimated deep temperatures of URG from literature (Sanjuan et al., 2016; Sanjuan et al., 2020) and those obtained in this study by GeoT
(Spycher et al., 2014) and at the equilibrium of reaction 2, i.e.: log(K) = log(Q). In this latter case, different thermodynamic databases such as Thermoddem
(Blanc et al., 2012), speq21.dat of PyGCC (Awolayo and Tutolo, 2022), and SUPCRTBL (Zimmer et al., 2016) were used to determine the log(K) values at differ-
ent temperature (Supplementary File 3).)
Brine sample and sampling aquifer Estimated deep temperature Geo-T thermoddem pygcc supcrtbl
borehole code date *(literature)

a b c d

♠T.min RMED ♣T.min ♥T.weight reaction reaction 2 but with muscovite(ord)

F
SDEV av. 2 instead of muscovite(dis)

equilibrium temperature (°C)

OO
1 GTLA-1 8/9/2011 G 225 208 206 205 233 390 348 555
2 GTLA-1 20/06/2013 G 208 206 205 267 440 397 854
3 GPK-2 6/19/2013 G 225 231 266 247 257 426 383 748
4 GPK-2 2016-2020 G 200 251 222 193 331 292 355
5 INSH 20/06/2013 G 225 191 204 200 206 350 310 409
6 GRT-1 1/10/2013 G 225 175 188 184 110 211 179 146
7 GRT-2 4/23/2018 G 225 190 200 195 268 442 399 869
8 GRT-2 2016-2020 G 202 214 211 184 318 280 324

PR
9 VDH-GT1 3/8/2018 G 225 168 180 176 182 315 277 317
10 CRON 1980 S 225 203 242 218 252 418 375 698
11 GBRU-1 1992 S 190 128 191 165 237 396 354 581
12 GBRU-1 3/4/2012 S 137 153 147 240 401 359 606
221 187 208 198 219 370 329 538
†28 30 32 27 46 68 65 230

♠ T.min RMED: temperature at which the median of scaled log(Q/K) absolute values (RMED) is minimum;
♣ T.min SDEV: temperature at the minimum value of standard deviation (SDEV) of the scaled log(Q/K);
T. weight av.: weight average of all statistically determined temperatures.
D
♥
⁎ mean of different chemical geothermometers from Sanjuan et al. (2016; 2020);
† value obtained combining of standard deviations as between that column values (s1 = 12.37°C) and those reported in *literature (s2 = 25°C).
TE

chemical composition of the GPK-2 borehole, respectively. It should values of reaction 2 at different temperatures are in the Supplementary
also be noted that the latter value is close to the temperature of 205 °C File 3). Apart from the different aquifer lithology (sandstones, granites)
measured at the well-bottom of 5000 m (Sanjuan et al., 2020). The and temperature at the depth of the source brine in the URG, the rela-
mean ionic strength of URG NaCl brines is 2.2 ± 0.2 M. However, even tively high standard deviation could also be due to the activity of hy-
if the ionic strength limit of the extended Debye-Hückel equation for drogen ion at the depth conditions, which is calculated by an iterative
calculating the activity coefficients in the GeoT code is probably ade- mass balance/mass action procedure from the measured pH at the sam-
EC

quate up to 4 M in NaCl-dominated solution (Alsemgeest et al., 2021; pling temperature of the brine. In turn, some of the sampling pH could
Xu et al., 2004), the Pitzer ion-interaction model should be preferred be overestimated because it is affected by CO2 degassing. Despite this,
(Xu, 2008). The activities of selected dissolved species in the 12 geot- in comparison to the value of TminRMED of 187 ± 30 °C obtained by
hermal brine samples from the URG, recalculated at 225 °C using both GeoT, the calculated mean temperature of 219 ± 46 °C obtained by
the Pitzer activity coefficient and the extended Debye-Hückel models equilibrium of reaction 2 is closer to the estimated deep temperature of
are shown in Table 3. 221 ± 28 °C (Table 2) and to the most frequently inferred temperature
RR

Similar to other lithium brines (Boschetti, 2022a), and considering of 225° ± 25 °C for the granite aquifer (Sanjuan et al., 2016, 2020).
that a widespread mica dissolution has affected both the sandstone and Moreover, it is also similar to the mean temperature at the minimum
the granite aquifer formations in the URG (Aquilina et al., 1997), a si- standard deviation (TminSDEV) of 208 ± 32 °C obtained by GeoT (Table
multaneous equilibrium between albite-low and quartz but also Li mica 2). It is also noteworthy that the most significant difference between ac-
and muscovite could be considered: tivities of the dissolved species in reaction 2 calculated by Pitzer and ex-
tended Debye-Hückel are with lithium (Table 3). Despite that, the mean
log(Q) values are quite similar (Table 3). Indeed, the mean tempera-
CO

tures inferred from Pitzer and Debye-Hückel model's logQ values are
(2) 214 and 219 °C, respectively.
The log(K) of reaction 2 was also recalculated at 500 bar (Supple-
mentary File 4). At the same total pressure, the log(Q) of reaction 2 us-
ing the activity of the dissolved species in the GPK-2 borehole was also
According to the law mass action, the ion activity product (Q) of re-
recalculated by PhreeqcI using a CO2 partial pressure of 0.894 bar
action 2 is:
(Kunan et al., 2021) and the temperature of 205 °C measured at the
well-bottom of 5000 m (Sanjuan et al., 2020) (Supplementary File 4).
(3) In those conditions, the calculated equilibrium temperature of GPK-2 is
232 °C, which is quite similar to the temperature of 229 ± 7 °C and
225 ± 25 °C inferred in the previously published studies (Alsemgeest
The mean equilibrium temperature obtained comparing the log(Q)
et al., 2021; Sanjuan et al., 2016).
values of the aqueous ions activities (square brackets in 3, values in
If we suppose that in the reaction 2 the muscovite was an ordered
Table 3) in the URG brines and the log(K) of the reaction 2 obtained us-
phase instead of a disordered one, the computed equilibrium tempera-
ing the Thermoddem database was 219 ± 46 °C (Table 2) (the log(K)
ture (using the Thermoddem dataset) would be greater than 300 °C for

4
T. Boschetti Geothermics xxx (xxxx) 102661

Table 3
– Comparison between the calculated activity of the aqueous species and the base-ten logarithm of activity product (log(Q), equation 3) of the URG brine sam-
ples: a) Pitzer ion-interaction model; b). The full physicochemical composition of the brines is in Supplementary File 1. Activity data in a) were obtained by
PhreeqcI, version 3 code (Parkhurst and Appelo, 2013), and a modified version of pitzer.dat thermodynamic database integrated with re-fitted lithium virial coef-
ficients (Boschetti, 2022a; Lassin et al., 2015). Activity data in b) were obtained by GeoT (Spycher et al., 2014) which use an extended Debye-Hückel equation
(Alsemgeest et al., 2021).
a) Borehole code Sampling date (dd/mm/yyyy) †aq.lit. Pitzer (PhreeqcI + Lassin et al. 2015) *Chemical data source

# Ionic strenght activity of aqueous species eq. 3

[H+] [Na+] [Li+] [H2O] logQ

F
(molality) (adimensional)

1 GTLA-1 8/9/2011 G 2.23 1.16E-06 5.89E-01 8.44E-03 9.44E-01 10.46 Sanjuan et al. (2016)

OO
2 GTLA-1 20/06/2013 G 2.23 1.78E-06 6.18E-01 8.43E-03 9.43E-01 10.12 Sanjuan et al. (2016)
3 GPK-2 6/19/2013 G 2.09 1.53E-06 5.87E-01 8.14E-03 9.48E-01 10.22 Sanjuan et al. (2016)
4 GPK-2 2016-2020 G 1.94 6.37E-07 5.16E-01 9.79E-03 9.52E-01 10.94 Bosia et al. (2021)
5 INSH 20/06/2013 G 2.26 8.33E-07 6.24E-01 7.77E-03 9.43E-01 10.77 Sanjuan et al. (2016)
6 GRT-1 1/10/2013 G 2.13 1.35E-07 5.91E-01 8.90E-03 9.47E-01 12.35 Dezayes et al. (2013)
7 GRT-2 4/23/2018 G 2.12 1.70E-06 5.93E-01 7.54E-03 9.47E-01 10.11 Sanjuan et al. (2022)
8 GRT-2 2016-2020 G 2.15 5.76E-07 5.65E-01 8.77E-03 9.47E-01 11.05 Bosia et al. (2021)
9 VDH-GT1 3/8/2018 G 1.97 5.56E-07 5.85E-01 7.73E-03 9.46E-01 11.08 Sanjuan et al. (2020)

PR
10 CRON 1980 S 2.12 1.61E-06 6.50E-01 9.54E-03 9.44E-01 10.27 Pauwels et al. (1993)
11 GBRU-1 1992 S 2.71 1.48E-06 8.22E-01 6.94E-03 9.29E-01 10.42 Pauwels et al. (1993); Aquilina et al. (1997)
12 GBRU-1 3/4/2012 S 2.58 1.44E-06 7.49E-01 6.92E-03 9.32E-01 10.38 Sanjuan et al. (2016)
mean 2.21 1.12E-06 6.24E-01 8.24E-03 9.43E-01 10.68
std.dev. 0.23 5.49E-07 8.35E-02 9.21E-04 6.59E-03 0.63
b) Borehole code Sampling date (dd/mm/yyyy) †aq.lit. extended Debye-Hückel (GeoT) *Chemical data source
# Ionic strenght activity of aqueous species eq. 3
[H+] [Na+] [Li+] [H2O] log(Q)
(molality) (adimensional)
D
1 GTLA-1 8/9/2011 G 1.59 1.22E-06 3.95E-01 1.22E-02 9.64E-01 10.25 Sanjuan et al. (2016)
2 GTLA-1 20/06/2013 G 1.60 1.85E-06 4.16E-01 1.24E-02 9.64E-01 9.92 Sanjuan et al. (2016)
3 GPK-2 6/19/2013 G 1.50 1.65E-06 4.06E-01 1.17E-02 9.66E-01 10.00 Sanjuan et al. (2016)
4 GPK-2 2016-2020 G 1.41 7.73E-07 3.65E-01 1.36E-02 9.68E-01 10.62 Bosia et al. (2021)
TE

5 INSH 20/06/2013 G 1.60 9.12E-07 4.20E-01 1.14E-02 9.64E-01 10.53 Sanjuan et al. (2016)
6 GRT-1 1/10/2013 G 1.53 1.93E-07 4.08E-01 1.29E-02 9.64E-01 11.88 Dezayes et al. (2013)
7 GRT-2 4/23/2018 G 1.53 1.91E-06 4.07E-01 1.09E-02 9.66E-01 9.86 Sanjuan et al. (2022)
8 GRT-2 2016-2020 G 1.51 6.67E-07 3.49E-01 1.25E-02 9.66E-01 10.70 Bosia et al. (2021)
9 VDH-GT1 3/8/2018 G 1.42 5.14E-07 4.03E-01 1.10E-02 9.64E-01 10.99 Sanjuan et al. (2020)
10 CRON 1980 S 1.52 1.48E-07 4.49E-01 1.43E-02 9.62E-01 12.20 Pauwels et al. (1993)
11 GBRU-1 1992 S 1.88 1.37E-06 5.20E-01 1.13E-02 9.59E-01 10.31 Pauwels et al. (1993); Aquilina et al. (1997)
EC

12 GBRU-1 3/4/2012 S 1.80 1.38E-06 4.80E-01 1.08E-02 9.59E-01 10.24 Sanjuan et al. (2016)
mean 1.57 1.05E-06 4.18E-01 1.21E-02 9.64E-01 10.62
std.dev. 0.14 6.08E-07 4.66E-02 1.10E-03 2.68E-03 0.74
† aquifer lithology:G: Carboniferous granite basement;S: sedimentary rock (Triassic - Permian or Triassic Buntsandstein sandstones);
⁎ see Supplementary File 1 for details.

most of the URG brine samples (Table 2). That is also confirmed by us- hypothetical equilibrium reaction with albite-low, quartz, and disor-
RR

ing the PyGCC and SUPCRTL thermodynamic datasets, which do not dered muscovite has provided deep temperature values of the URG
have the disordered muscovite in the mineral list (Table 2) (Supplemen- brines (219 ± 46 °C), compatible with both the previously published
tary File 3). In the URG, a dominant temperature range from values (221 ± 28 °C) and slightly higher than the multiparametric ge-
140–170 °C to approximatively 230 °C was inferred from clay mineral othermometrical approach (TminRED187 ± 30 °C by GeoT code). The
formation and transformation (Nguyen-Thanh et al., 2021). However, a equilibrium temperature closest to the calculated and measured tem-
temperature peak of 340 °C was associated to the Permian illite at perature value was obtained by integrating the generic Li-mica with the
CO

Soultz (Schleicher, 2005). Considering that the white mica/illite (poly- aqueous solute species and minerals thermodynamics of the Thermod-
type composition 1Md, 1 M and 2M1, which include mixed and disor- dem dataset. Higher temperature values are obtained by other data-
dered structures) is the higher lithium mineral found in the URG brine bases (PyGCC and SUPCRTBL), but this is mainly because other data-
aquifer (Clauer et al., 2008; Druppel et al., 2020; Schleicher, 2005; bases do not have a disordered muscovite but only an ordered form of
Schleicher et al., 2006), then the use of disordered muscovite in the that mineral. In other words, the high lithium concentration of the
equilibrium reaction 2 could be justified, as also supported by the white mica/illite detected in the URG aquifer seems to be reproduced
match between the temperature calculated and measured. by the equilibrium reaction involving the generic Li-mica revised in this
study and the disordered muscovite included in the Thermoddem
4. Conclusions dataset. The application of the generic Li-mica's revised thermodynam-
ics to URG brines provides a first test of the usefulness of this data for
The re-evaluation of the theoretical lithium mica's thermodynamic geothermometric purposes. The new data would be considered prelimi-
data confirms the mineral's usefulness in the geothermometric model- nary and in need of further checks. The new data should be considered
ing of lithium-rich geothermal brine. Although there are significant dif- preliminary and in need of further validation. The study of other
ferences between the solubility log(K) values compared with those pre- lithium-rich geothermal waters in different geological contexts using
viously published ones (Boschetti, 2022b), the use of the Li-mica in the these data is therefore recommended..

5
T. Boschetti Geothermics xxx (xxxx) 102661

Declaration of Competing Interest Holland, T.J.B., Powell, R., 2011. An improved and extended internally consistent
thermodynamic dataset for phases of petrological interest, involving a new equation
of state for solids. J. Metamorph. Geol. 29 (3), 333–383.
The authors declare that they have no known competing financial Johnson, J.W., Oelkers, E.H., Helgeson, H.C., 1992. SUPCRT92: a software package for
interests or personal relationships that could have appeared to influ- calculating the standard molal thermodynamic properties of minerals, gases, aqueous
species, and reactions from 1 to 5000 bar and 0 to 1000 C. Comput. Geosci. UK 18 (7),
ence the work reported in this paper.
899–947.
Komninou, A., Yardley, B.W.D., 1997. Fluid-rock interactions in the Rhine Graben: a
Data Availability thermodynamic model of the hydrothermal alteration observed in deep drilling.
Geochim. Cosmochim. Ac 61 (3), 515–531.
Kunan, P., Ravier, G., Dalmais, E., Ducousso, M., Cezac, P., 2021. Thermodynamic and
Used data in this manuscript were attached or referenced if kinetic modelling of scales formation at the soultz-sous-forets geothermal power
previously publisahed. Moreover, all elaborated data are available as plant. Geosciences 11 (12).

F
table or supplementry file. Lassin, A., Christov, C., André, L., Azaroual, M., 2015. A thermodynamic model of
aqueous electrolyte solution behavior and solid-liquid equilibrium in the Li-H-Na-K-
Cl-OH-H2O system to very high concentrations (40 molal) and from 0 to 250°C. Am.
Acknowledgements J. Sci. 315 (3), 204–256.

OO
Maier, C.G., Kelly, K.K., 1932. An equation for the representation of high-temperature
heat content data. J. Am. Chem. Soc. 54 (8), 3243–3246.
The author would like to thank N. Spycher and one anonymous re- Nguyen-Thanh, L., Mählmann, R.F., Hoang-Minh, T., Petschick, R., Reischmann, T.,
viewer whose helpful comments and suggestions helped improve and Nesbor, H.D., Ruttmann, F., Fritsche, J.G., 2021. Clay mineral formation in Permian
clarify this manuscript. rocks of a geothermal borehole at Northern Upper Rhine Graben, Germany. Int.
J.Earth Sci. 110 (4), 1415–1438.
Parkhurst, D.L., Appelo, C.A.J., 2013. Description of input and examples for PHREEQC
Supplementary materials version 3–a computer program for speciation, batch-reaction, one-dimensional
transport, and inverse geochemical calculations, US Geological Survey, Techniques

PR
and Methods. U.S.G.S., Reston, VA, p. 497. book 6, chap. A43.
Supplementary material associated with this article can be found, in
Pauwels, H., Fouillac, C., Fouillac, A.M., 1993. Chemistry and isotopes of deep geothermal
the online version, at doi:10.1016/j.geothermics.2023.102661. saline fluids in the Upper Rhine Graben: origin of compounds and water-rock
interactions. Geochim. Cosmochim. Ac 57 (12), 2737–2749.
References Roine, A., 2009. HSC Chemistry® 7.0 User’s Guide – Volume ½ – Chemical Reaction and
Equilibrium Softwarewith Extensive Thermochemical Database and Flowsheet
Simulation. Outotec Information Center, Finland.
Alsemgeest, J., Auque, L.F., Gimeno, M.J., 2021. Verification and comparison of two
Sanjuan, B., Gourcerol, B., Millot, R., Rettenmaier, D., Jeandel, E., Rombaut, A., 2022.
thermodynamic databases through conversion to PHREEQC and multicomponent
Lithium-rich geothermal brines in Europe: an up-date about geochemical
geothermometrical calculations. Geothermics 91.
characteristics and implications for potential Li resources. Geothermics 101.
Apps, J., Spycher, N., 2004. Data Qualification For Thermodynamic Data Used to Support
D
Sanjuan, B., Millot, R., Asmundsson, R., Brach, M., Giroud, N., 2014. Use of two new Na/
THC Calculations. Bechtel SAIC Company, Las Vegas, NV.
Li geothermometric relationships for geothermal fluids in volcanic environments.
Aquilina, L., Pauwels, H., Genter, A., Fouillac, C., 1997. Water-rock interaction processes
Chem. Geol. 389, 60–81.
in the Triassic sandstone and the granitic basement of the Rhine Graben: geochemical
Sanjuan, B., Millot, R., Innocent, C., Dezayes, C., Scheiber, J., Brach, M., 2016. Major
investigation of a geothermal reservoir. Geochim. Cosmochim. Ac 61 (20),
geochemical characteristics of geothermal brines from the Upper Rhine Graben
4281–4295.
TE

granitic basement with constraints on temperature and circulation. Chem. Geol. 428,
Awolayo, A.N., Tutolo, B.M., 2022. PyGeochemCalc: a Python package for geochemical
27–47.
thermodynamic calculations from ambient to deep Earth conditions. Chem. Geol.
Sanjuan, B., Negrel, G., Le Lous, M., Poulmarch, E., Gal, F., Damy, P.C., 2020. In: Main
606.
geochemical characteristics of the deep geothermal brine at Vendenheim (Alsace,
Bale, C.W., Belisle, E., Chartrand, P., Decterov, S.A., Eriksson, G., Gheribi, A.E., Hack, K.,
France) with constraints on temperature and fluid circulation., World Geothermal
Jung, I.H., Kang, Y.B., Melancon, J., Pelton, A.D., Petersen, S., Robelin, C., Sangster,
Congress 2020. Reykjavik, Iceland.
J., Spencer, P., Van Ende, M.A., 2016. FactSage thermochemical software and
Schleicher, A., 2005. Clay Mineral Formation and Fluid-Rock Interaction in Fractured
databases, 2010-2016. Calphad 54, 35–53.
Crystalline Rocks of the Rhine Rift System: Case Studies from the Soultz-Sous-Forêts
EC

Bethke, C.M., Farrell, B., Sharifi, M., 2022. The Geochemist’s Workbench® Release 12
granite (France) and the Schauenburg Fault (Germany) Fakultät für Chemie Und
(five volumes). Aqueous Solutions LLC, Champaign, Illinois.
Geowissenschaften. Institut für Geowissenschaften, Heidelberg.
Bethke, C.M., Yeakel, S., 2007. The Geochemist’s Workbench® Release 7 (Four Volumes).
Schleicher, A.M., Warr, L.N., Kober, B., Laverret, E., Clauer, N., 2006. Episodic
The University of Illinois, Urbana, Illinois.
mineralization of hydrothermal illite in the Soultz-sous-Forets granite (Upper Rhine
Blanc, P., Gherardi, F., Vieillard, P., Marty, N.C.M., Gailhanou, H., Gaboreau, S., Letat, B.,
Graben, France). Contrib. Miner. Petr. 152 (3), 349–364.
Geloni, C., Gaucher, E.C., Made, B., 2021. Thermodynamics for clay minerals:
Spycher, N., Peiffer, L., Sonnenthal, E.L., Saldi, G., Reed, M.H., Kennedy, B.M., 2014.
calculation tools and application to the case of illite/smectite interstratified minerals.
Integrated multicomponent solute geothermometry. Geothermics 51, 113–123.
Appl. Geochem. 130.
Stefansson, A., 2001. Dissolution of primary minerals of basalt in natural waters - I.
Blanc, P., Lassin, A., Piantone, P., Azaroual, M., Jacquemet, N., Fabbri, A., Gaucher, E.C.,
RR

Calculation of mineral solubilities from 0 degrees C to 350 degrees C. Chem. Geol. 172
2012. Thermoddem: a geochemical database focused on low temperature water/rock
(3–4), 225–250.
interactions and waste materials. Appl. Geochem. 27 (10), 2107–2116.
Stringfellow, W.T., Dobson, P.F., 2021. Technology for the recovery of lithium from
Boschetti, T., 2022a. Comments on “Lithium-rich geothermal brines in Europe: an up-date
geothermal brines. Energies 14 (20).
about geochemical characteristics and implications for potential Li resources” by
Tagirov, B., Schott, J., 2001. Aluminum speciation in crustal fluids revisited. Geochim.
Sanjuan et al. (2022). Geothermics 105.
Cosmochim. Ac 65 (21), 3965–3992.
Boschetti, T., 2022b. A revision of lithium minerals thermodynamics: possible
Van Hinsberg, V.J., Vriend, S.P., Schumacher, J.C., 2005. A new method to calculate end-
implications for fluids geochemistry and geothermometry. Geothermics 98.
member thermodynamic properties of minerals from their constituent polyhedra II:
Bosia, C., Mouchot, J., Ravier, G., Seibt, A., Jähnichen, S., Degering, D., Scheiber, J.,
heat capacity, compressibility and thermal expansion. J. Metamorph. Geol. 23 (8),
CO

Dalmais, E., Baujard, C., Genter, A., 2021. Evolution of brine geochemical
681–693.
composition during operation of EGS geothermal plants (Alsace, France). In: 46th
Vidal, J., Genter, A., 2018. Overview of naturally permeable fractured reservoirs in the
Workshop on Geothermal Reservoir Engineering. Stanford University, Stanford, CA,
central and southern Upper Rhine Graben: insights from geothermal wells.
USA. p. 21.
Geothermics 74, 57–73.
Clauer, N., Liewig, N., Ledesert, B., Zwingmann, H., 2008. Thermal history of Triassic
Wolery, T.J., Colon, C.F.J., 2017. Chemical thermodynamic data. 1. The concept of links
sandstones from the Vosges Mountains-Rhine Graben rifting area, NE France, based
to the chemical elements and the historical development of key thermodynamic data.
on K-Ar illite dating. Clay Miner. 43 (3), 363–379.
Geochim. Cosmochim. Ac 213, 635–676.
Delany, J.M., Lundeen, S.R., 1991. The LLNL Thermochemical Data Base - Revised Data
Wolery, T.J., Jackson, K.J., Bourcier, W.L., Bruton, C.J., Viani, B.E., Delany, J.M., 1988.
and File Format For the EQ316 Package. Lawrence Livermore National Laboratory
The EQ3/6 Software Package For Geochemical Modeling - Current Status. American
(LLNL), Livermore, CA (United States), p. 67.
Chemical Society (ACS) national meeting, Los Angeles, CA (United States). pp. 23-
Dezayes, C., Sanjuan, B., Gal, F., Lerouge, C., Brach, M., 2013. Forage d'exploration
Geoc.
géothermique GRT-1. Suivi géochimique des fluides et caractérisation des zones
Xu, T., 2008. TOUGHREACT Testing in High Ionic Strength Brine Sandstone Systems.
fracturées.
Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States).
Druppel, K., Stober, I., Grimmer, J.C., Mertz-Kraus, R., 2020. Experimental alteration of
Xu, T., Sonnenthal, E., Spycher, N., Pruess, K., 2004. TOUGHREACT User’s Guide: A
granitic rocks: implications for the evolution of geothermal brines in the Upper Rhine
Simulation Program for Non-isothermal Multiphase Reactive Geochemical Transport
Graben. Germany. Geothermics 88.
in Variably Saturated Geologic Media. Lawrence Berkeley National Laboratory
Gordiyenko, V.V., Ponomareva, N.I., 1989. Stability conditions for lithium micas of the
(LBNL), University of California, Berkeley, CA (United States).
lepidolite series. Int. Geol. Rev. 31 (3), 253–257.
Zimmer, K., Zhang, Y.L., Lu, P., Chen, Y.Y., Zhang, G.R., Dalkilic, M., Zhu, C., 2016.
Helgeson, H.C., Delany, J.M., Nesbitt, W.H., Bird, D.K., 1978. Summary and critique of the
SUPCRTBL: a revised and extended thermodynamic dataset and software package of
thermodynamic properties of rock-forming minerals. Am. J. Sci. 278–A, 1–229.
SUPCRT92. Comput. Geosci. UK 90, 97–111.