You are on page 1of 16

Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011.

© 2011 IUPAC, Publication date (Web): 4 April 2011

Chemistry of salts in aqueous solutions:

Applications, experiments, and theory*
Wolfgang Voigt‡

TU Bergakademie Freiberg, Institut für Anorganische Chemie, Leipziger Strasse

29, D-09596 Freiberg, Germany

Abstract: Salts comprise a very large and important group of chemical compounds. Natural
occurrence of salts and industrial processes of their recovery, conversion, purification, and
use depend on solubility phenomena and their chemistry in aqueous solutions, mostly in
multi-ion systems. Modeling of these processes as well as developing new ones requires
knowledge of the properties of the aqueous salt solutions in extended T-p-x ranges including
a growing number of components in solutions (CO2, SO2, lithium salts, salts of rare earth
metals, actinides, etc.). At least for the thermodynamic properties, the general accepted
methodology is to use thermodynamic databases of aqueous species and solids in combina-
tion with an appropriate ion-interaction model to perform equilibrium calculations for
species distributions in solution and phase equilibria. The situation in respect to available
thermodynamic models and data for their parameterization is discussed at selected examples.
Thereby, the importance of accurate experimental determinations of phase equilibrium data
for derivation of model parameters is emphasized. Furthermore, it is concluded that experi-
mental investigations should follow a chemical systematic. Simple physical models or quan-
tum chemical calculations cannot predict unknown quantities in the databases with sufficient
Finally, solubility changes in salt–water systems at enhanced temperatures are consid-
ered. Systems, which can be considered as molten hydrates, display interesting phase behav-
ior and chemical reactivity as protic ionic liquids. The latter can be exploited in chemical syn-
theses to substitute mixtures of concentrated acids like HNO3/H2SO4 by simple salts like
ferric nitrate.
For an understanding of the chemical and phase behavior of water–salt systems in
terms of structure–property relations, a renaissance of chemical guided basic investigations
of such systems would be worthwhile.

Keywords: Brunauer–Emmett–Teller (BET) model; molten salt hydrates; oceanic salts; poly-
halite; thermodynamic models.

The chemistry of salts in aqueous solutions concerns first of all processes of its dissolution and crys-
tallization, hydrolysis, and formation of complex ions in solution. The combination of metal cations
with typical inorganic anions as halide, sulfate, nitrate, phosphate, etc. yields a very large number of

*Paper based on a presentation made at the 14th International Symposium on Solubility Phenomena and Related Equilibrium
Processes (ISSP-14), Leoben, Austria, 25–30 July 2010. Other presentations are published in this issue, pp. 1015–1128.

1016 W. VOIGT

compounds with salt-like properties, that is properties dominated by ions. The class of salts covers a
wide spectrum: from NaCl, gypsum (CaSO4ⴢ2H2O), anhydrite (CaSO4), saltpeter (NaNO3), and cryo-
lite (Na3AlF6) to salts of heavy metals or of noble metals. In geology, environmental sciences, medi-
cine, and industrial processes many problems are related to the question of the solubility of salts. This
question asks for the ion combination, which will form a solid salt from an aqueous solution of certain
composition and temperature. Every textbook on general and inorganic chemistry describes salts as
ionic compounds crystallizing in ionic lattices with a certain lattice energy. Dissolving these salts in
water requires that the lattice energy is overcome by the process of ionic hydration. Both quantities are
then used to explain the tendencies in the solubility for simple salts like alkaline metal halides or alka-
line earth metal sulfates often in relation to the Haber–Born cycle [1]. This picture is only suited to get
a rough idea about factors like ion charge/radius ratio having influence on the solubility of salts. Besides
the intrinsic properties of the solids (lattice enthalpy, entropy) the solubility of salts in water or in solu-
tions containing other ions depends on the ion–water and ion–ion interactions. These interactions are
changing very sensitively with concentration, composition, and temperature. At present, chemistry can
still not give a quantitative answer by means of definite structural models [2]. Solubility relations of dif-
ferent salts in mixed ionic solutions form also the basis for an understanding of their existence and sta-
bility in contact with aqueous solutions.

Fig. 1 Jacobus Henricus van’t Hoff and the document of the Swedish Academy of Science awarding him the Nobel
Prize in Chemistry.

The first systematic investigation of this kind was realized by J. H. van’t Hoff, the first winner of
a Nobel prize in chemistry in 1901 (Fig. 1), whose centenary of death we remember in 2011. By con-
sequently applying state-of-the-art theory of chemical thermodynamics and equilibrium, he and his co-
workers developed within 12 years research a nearly complete scheme of the composition and temper-
ature conditions for the formation of the various salt phases (hydrated and anhydrous simple salts,
double salts, triple salts) from aqueous solutions of the oceanic system, which comprises the ions Na+,
K+, Mg2+, Ca2+, Cl–, and SO42–. The results of the 52 papers were summarized in the famous book
Investigations on the Conditions of Formation of the Oceanic Salt Deposits, especially of the Stassfurt
Deposit (see Fig. 2). At the same time, this was the first systematic investigation of a multicomponent
salt–water system, which was highly needed at that time for a more effective exploration of potash
deposits and its recovery to produce potash fertilizers. The work of van’t Hoff was continued by D’Ans
[3], Autenrieth [4,5], and Emons [6,7] and is continued to the present [8–10]. More than 100 years

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
Chemistry of salts in aqueous solutions 1017

Fig. 2 Cover of the book Untersuchungen über die Bildungsverhältnisse der Ozeanischen Salzablagerungen
insbesondere des Stassfurter Salzlagers, summarizing the papers of van’t Hoff edited by Precht and Cohen, 1912.

research on this salt system has steadily improved the technology in recovery of mineral fertilizers, the
understanding of formation of salt deposits, and the production of basic chemicals as part of the
extended oceanic system including carbonates, H+ and OH–.


New aspects of applications with the need for detailed knowledge of solubility, solution equilibria, and
properties in the oceanic salt system are the storage of nuclear wastes in geological salt formations as
in northern Germany (Gorleben, Morsleben) or in New Mexico, USA (Carlsbad WIPP site), storage of
inorganic toxic waste in old salt mines (Teutschenthal, Germany), CO2 sequestration in highly saline
aquifers of large depths, and exploitation of geothermal brines. These new tasks require knowledge of
the phase behavior and solution properties at extended temperature and pressure ranges, the inclusion
of more and more components as CO2, heavy metals, actinides, acids and bases, and Li+. The latter ion
recently became a very important issue for the recovery of lithium salts from brines of salt flats and
lakes in South America (Chile, Argentina, Bolivia) and Asia (West China) to ensure a sufficient pro-
duction of Li2CO3 for large lithium batteries in automotive and other applications. The brines located
in the salt crusts contain between 300–4000 ppm Li+. The concentration is enhanced by solar evapora-
tion. At the same time, huge amounts of other salts have to be crystallized [11] and separated exploit-
ing the multicomponent solubility equilibria (Fig. 3).
Another topic concerns weathering of historical and modern buildings. It is often controlled by
crystallization and dissolution of salts at changing humidity of the surrounding air. To understand
these processes, detailed data on phase diagrams and properties (water activity, partial molar volumes)
of mixed solutions of alkaline and alkaline earth metal chlorides, nitrates, and sulfates are needed

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
1018 W. VOIGT

Fig. 3 Salar de Uyuni (Bolivia, 10.000 km2) with an average lithium content of 500–600 ppm Li+/L in the inter-
granular brine below the salt surface. Inset: brine cavity 40 cm below surface and main ions in solution with
amounts of salts to be crystallized for the production of 1 t of Li2CO3.

Another topic where the chemistry of salt–water systems is asked for concerns interplanetary sci-
ence. New observations of Mars missions gave evidence for aqueous salt solutions existing at very low
temperatures (180–250 K). The questions arise, which cryotectic water–salt mixtures can be formed
down to about 200 K, what properties can we expect from these solutions, and could it be a possible
environment for life? To answer these questions, a large number of salt–water systems have to be con-
sidered [14].
Modern applications concern of course all the hydrometallurgical processes for recovery and
purification of zinc, aluminum, and heavy metals like copper, nickel, cobalt, etc. The control of leach-
ing and solid–liquid separation processes require a more accurate description of phase and chemical
equilibria and the kinetics to reach these equilibria [15].
Environmental sciences need more accurate equilibrium data in order to make use of the sensi-
tivity and accuracy of modern methods of concentration monitoring.
Summarizing, it can be stated that modern needs in the chemistry of salt–water systems consist
on one hand in a fine-tuning of the models for property descriptions and on the other hand in predic-
tions when extending temperature, pressure, or system components.


Starting with van’t Hoff, further development in the theory of salt solutions or more general electrolyte
solutions focused on the physical description of the long-range ion–ion interactions [16]. Combined
with ion-specific, short-range interaction terms, equations were formulated describing the concentration
dependence of activity coefficients of salt components [17], osmotic coefficients, or more general the
excess Gibbs energy of mixing [18–21]. Still the most popular approach for the thermodynamic treat-
ment of concentrated electrolyte solutions represents Pitzer’s ion interaction model [22]. It is based on
some fundamental considerations on virial equations for electrolyte solutions, a restriction to the sec-
ond and third virial coefficient with a suited choice of the mathematical form of ionic strength depend-
ence. Thus, the model is able to describe the thermodynamic properties of electrolyte solutions up to

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
Chemistry of salts in aqueous solutions 1019

ionic strengths of approximately 6–10 mol/kg H2O, when model parameters are estimated from appro-
priate binary and ternary salt–water systems. For a given temperature and a typical system like
NaCl–KCl–H2O, 3 parameters for every binary system (NaCl–H2O, KCl–H2O) and 2 additional empir-
ical parameters are required for the ternary system, that is 8 parameters in this case. All interactions in
the higher-component systems are calculated using only parameters from the corresponding binary and
ternary systems. Nevertheless, for the hexary oceanic salt system the parameters have to be estimated
from experimental data of 8 binary and 16 ternary systems, which gives 56 isothermal parameters and
2 additional parameters for the 2-2 valent electrolytes MgSO4 and CaSO4. Using these parameters, the
activity coefficients of saturated salt solutions can be calculated and thus the thermodynamic solubility
constants. For the most important 35 salt phases it has been done for the oceanic salts [23]. Pitzer
parameters and solubility constants form a data set stored in a thermodynamic database, which in con-
nection with appropriate equilibrium calculation codes (ChemSage, ChemApp, FACTSAGE, EQ3/6,
PHREEQC, GEMS, MINEQL+) allow simulation of phase equilibria of complex composed salt–water
systems. In a similar way, also parameters for speciation equilibria in homogenous solutions (complex
formation) with interactions of the corresponding complex ions and their equilibrium constants are
extracted from experimental data [24].
Pitzer’s theory gives no hint for the temperature dependence of the model parameters and thus
entirely empirical equations are used. For a temperature interval from 0 to about 200 °C, up to 8 tem-
perature coefficients are necessary for every Pitzer parameter. The same is valid for the solubility con-
stants. Databases containing such temperature-dependent parameters enable calculation also for
enthalpic effects. However, the predictive power of the Pitzer model is limited, especially when the
ionic strengths are higher than 6 molal.
The computing power of personal computers or laptops is sufficient to perform equilibrium cal-
culations with hundreds of species and phases in seconds. Thus, the current state of the art to study geo-
chemical and industrial processes consists in thermodynamic modeling and simulation using a suited
database and program code. Efforts are being made to couple equilibrium calculations with transport
codes [25]. In this case, computing time for the equilibrium states becomes critical.
The large number of parameters is not a limiting factor in computer calculations. The limitation
has to be seen in the availability of experimental data for their estimation. Thus, for solutions formed
from combinations of the most important of 28 cations and 16 anions, activity data for 448 binary and
9408 ternary systems have to be provided and approximately 22.000 parameters have to be extracted
from the available experimental data for one temperature. Only for a limited number of ternary systems
are such data available for one temperature, even worse is the situation when temperature ranges have
to be considered. Database projects, where existent data from literature are gathered to extract model
parameters, have been reviewed recently [26].
Therefore, constant efforts are directed to develop models with a smaller number of adjustable
parameters. However, applicability is mostly tested only for binary salt–water systems [27–31]. One
general approach combines typical nonelectrolyte models such as NRTL, UNIQUAC, or UNIFAC mod-
els with an extended Debye–Hückel term [32–39].
Extensive demonstrations for application of an extended NRTL model for salt–water systems are
given by Iliuta and Thomsen [35]. As long as temperature-dependent models are considered the advan-
tage of an apparently smaller number of parameters in the solution model is often counteracted by
abnormal many temperature coefficients of the solubility constant. Often the existence range of a salt
hydrate or double salt is less than 15 K, however, the solubility constant is modeled with 5 temperature
coefficients [35]. Other types of model use physical reference models like MSA (mean spherical
approximation), integral equation approach, or others: reviewed in [40].
Another way to derive missing interaction coefficients, for instance, in a Pitzer model is to gen-
erate “quasi-experimental” data by the use of mixing rules like those of Zdanovskii or Young [41–43]
Thermodynamic standard-state properties of ions in water at 298 K can be found in [44–47]. For
extended temperatures and pressures the correlation from Helgeson, Kirkham, and Flowers (HKF

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
1020 W. VOIGT

model) [48,49] is often used, where the values generated by the HKF model can easily be accessed by
means of the software package SUPCR92 [50]. Also, more simple correlations as proposed by Cobble
and co-workers [51,52] or the coulombic compensation principle [24,53] can be used to estimate stan-
dard data for high temperatures.
In online databases these data are often easily accessible [26]. However, the user has to ensure
(sometimes supported by the database itself as in THEREDA) that standard data, speciation, and activ-
ity models are consistent with each other. Despite consistency issues and uncertainties in validity ranges
of model parameters, etc. the community of users routinely applying packages consisting of solution
model, databases, and calculation codes to handle aqueous chemistry problems is ever growing.
Therefore, there is a strong focus to expand the databases and to improve the codes and their user inter-


The most important types of experiments from which parameters for thermodynamic models are esti-
mated are
a) isopiestic measurements [54]
b) vapor pressure measurements [55–57]
c) galvanic cells in dilute aqueous solutions [58–61]
d) measurements of heat capacities [62–68] and heat of dilution [69–71] or dissolution [72]
e) pH [73], spectrophotometric mesurements [74–77], Raman and NMR spectroscopy for speciation
equilibria [78,79]
f) determination of solid–liquid equilibria (solubilities) [80]
Experiments of types a–e are normally performed in undersaturated solutions providing proper-
ties of the solution only. Combined with determinations of solid–liquid equilibria thermodynamic quan-
tities of the solid phases can be derived. In principle (and most basic) the latter information can be
obtained from enthalpy of dissolution of the solid-phase and low-temperature heat capacity measure-
ments. However, the overwhelming majority of Gibbs energy of formation data of salts and its hydrates
tabulated in standard data books [46,47] have been derived from solubility data. Without accurate sol-
ubility data also most of the formation data would be inaccurate.
Particularly in the case of solubility determinations often very old papers have to be studied,
because the study of solubility in aqueous salt systems in contemporary chemistry is not considered as
scientifically interesting. One actual example represents the solubility equilibria of polyhalite, a wide
spread in rock salt formations triple salt mineral, K2SO4ⴢMgSO4ⴢ2CaSO4ⴢ2H2O. At the standard tem-
perature of 298 K, data only from van’t Hoff [81] or [82] are available. In order to accomplish a thermo-
dynamic model to describe the compositions of solutions, where polyhalite can be in equilibrium
besides the solubility product of polyhalite itself also the Pitzer parameters for the Mg2+–Ca2+ and
Mg2+–SO42––Ca2+ interaction have to be known at high concentrations of MgSO4. The latter can be
best determined from solubilities of gypsum in MgSO4 solutions. Harvie, Möller, and Weare estab-
lished the well-known in the geochemical community Pitzer equation based seawater model for T =
298 K [83]. In their data compilation they had to rely on solubility determinations of gypsum by
Cameron [84] from the year 1906 for the description of the Mg2+–SO42––Ca2+ interaction. At that time
no suited analytical method was available for the determination of Ca2+ concentrations in high concen-
trated magnesium salt solutions. The difference between the old and new data is illustrated in Fig. 4.
The consequences of new data are illustrated for the existence range of polyhalite in solutions of potas-
sium and magnesium sulfate at 298 K as shown in Fig. 5. Implemented in a new Pitzer model of the
hexary oceanic salt system (available from <>), a much larger composition range is cal-
culated for the existence of polyhalite at 298 K. This has consequences for the geochemical interpreta-
tion of brine analyses. Also, we conclude that the lower formation temperature of polyhalite in contact

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
Chemistry of salts in aqueous solutions 1021

Fig. 4 Solubility of gypsum according to Cameron 1906 [84] (○), Kolosov [85] () and Wollmann 2010 [86] (),
 error bars.

Fig. 5 Crystallization field of polyhalite (Ph) in the system K2SO4–MgSO4–CaSO4–H2O at 298 K; thin lines =
calc. with Pitzer model based on new solubility data [86], thick-lined polygon = polyhalite field calc. with model
of HMW [83], stars = existence triangle for polyhalite according to experiments of Van Klooster [82].

with solution can be lower than 286 K as was predicted by van’t Hoff [81] and never adjusted until pres-
ent days. First findings of polyhalite from drillings in the Salar de Uyuni, where temperatures are well
below 286 K, seem to confirm this conclusion.


The prediction of the crystallization fields of the various hydrate types of a salt or double salt represents
a long-standing issue. In past times the preferred approach was the direct phase equilibrium experiment.
For a certain class of salts or their combinations, the phases in equilibrium were separated and chemi-

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
1022 W. VOIGT

cally analyzed. Guided by some correlations of chemical parameters like ionic radius, chemical hard-
ness, and softness [87] or ligand field stabilization energies, the comparison of a series of systems and
their phase diagrams led to conclusions about the possible existence of double salts in analogous sys-
tems not investigated at time. Of course, such correlations were more or less qualitative. The situation
has changed dramatically in the last 20 years. Progress in methods for X-ray crystal structure analysis
made it possible to discover a large number of new crystal phases in nature and in the lab. Often also
the composition of known salts are determined more accurately as, for example, for the
“MgSO4ⴢ12H2O”, which is now meridianite, MgSO4ⴢ11H2O, [88,89], similar for kainite [90] or
loeweite [89]. Thus, the situation arose that hundreds of double salt phases are known structurally but
no thermodynamic data of formation are available and the conditions of their formation or crystalliza-
tion are unknown. In order to fill such data gaps in thermodynamic databases, several estimation meth-
ods have been proposed as, for instance, by Jenkins and Glasser [92–94] for ionic hydrates. In essence,
the authors make use of linear relations of the type
{ΔfH°hydrate, ΔfG°hydrate} = {ΔfH°anhydrate, ΔfG°anhydrate} + a ⴢ nH2O (1)
where ΔfH°, ΔfG° denote the respective standard enthalpies and Gibbs energies of formation, nH2O the
number of moles water in the hydrate, and the slope “a” represents the increment per mol water. The
data fulfill the equation within a scatter of 25 kJ/mol. Thus, this method yields quite rough estimates.
Difference rules of this type do not consider structural features of the solids, and due to the linearity of
the formation quantities, hydrates with any number of nH2O would be predicted as stable for a given salt.
Another recently published (but not really new) approach concerns the estimation of solubility
constants of double salts, KL(DS), from their single component salts [95,96] by means of eq. 2.
KL(DS) = KL(1)ⴢKL(2) (2)
Since the relation 3 is valid eq. 2 implies the additivity rule of the Gibbs energies of
–RTⴢln[KL(i)] = ΔLG°(i) (3)
formation of the component salts. This means, however, an absence of a driving force to form the dou-
ble salt. The apparent confirmation of this “rule” by comparing calculated and experimentally found
solution concentrations is a consequence of the form of some solubility diagrams as shown in Fig. 6 as
well as of the accuracy of chemical analysis. In the case of type A, the solution composition in equi-
librium with the double salt is not much different from a solution composition of the mixture of the sin-
gle salts, and hence the “rule” is apparently valid. Not so in cases of types B and C, where the forma-

Fig. 6 Types of solubility isotherms of systems MX–NX–H2O with double salt formation MXⴢNX, thick line =
crystallization branch of double salt, thin lines = crystallization branches of single salts MX, NX.

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
Chemistry of salts in aqueous solutions 1023

tion of the double salt decreases the total salt concentration considerably in relation to the extrapolated
single salt branches in the diagram.
Double salts of the type M2SO4ⴢM'SO4ⴢ6H2O, Tutton’s salts, belong to the most extensively
investigated double salt hydrates. To predict which combination of M and M' form stable salts in con-
tact with aqueous solutions, one could try to estimate the Gibbs energy of formation from aqueous ions
and use one of the above-mentioned programs and solution models to calculate the crystallization
branches of possible double salts. In Figs. 7a,b, the known Gibbs energies of formation for M = K are
correlated with the position in the periodic table (7a) and the radius of ion M'2+ (7b). As can be seen
from the graphs for Mn2+, V2+, and Cr2+ the correlation would yield negative values of Gibbs energies
of formation, however, with an uncertainty of approx. ±3 kJ/mol. The influence of such a variance of
ΔfG°(DS) is demonstrated in Fig. 8 for M'2+ = Mg2+. The changes in the solubility isotherms are dras-
tic, and other phases such as leonite overlap in the field of uncertainty. There are also efforts to apply
quantum chemical methods to predict the existence of solid phases [97,98]. However, for double salt
predictions they are too inaccurate.

Fig. 7 Correlation of the standard Gibbs energies of formation of picromerite, K2SO4ⴢMSO4ⴢ6H2O, from aqueous
ions at 298 K and prediction for M = Mn. Ionic radius with coordination number of 6.

Fig. 8 Effect of an uncertainty of ±3 kJ/mol in the Gibbs energy of formation of picromerite on the solubility curve
at 298 K. Open circles = solubility of picromerite, stars = calculated solubility of picromerite with reduced or
enhanced Gibbs energy of formation, open rectangles = solubility of leonite; picromerite = K2SO4ⴢMgSO4ⴢ6H2O,
leonite = K2SO4ⴢMgSO4ⴢ6H2O.

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
1024 W. VOIGT

Thus, much more accurate data of ΔfG°(DS) are needed for solubility predictions of unknown
double salts. Presently available estimation methods are not sufficient for this purpose. Table 1 lists all
known double salts of the type M2SO4ⴢM'SO4ⴢ6H2O. The K–Mn combination could not be prepared
until now [99]. Our own investigation of the system K2SO4–MnSO4–H2O at 298 and 313 K
revealed the formation of the double salts K2SO4ⴢ3MnSO4ⴢ5H2O, K2SO4ⴢMnSO4ⴢ2H2O,
K2SO4ⴢMnSO4ⴢ1.5H2O, K2SO4ⴢMnSO4ⴢ4H2O, and K2SO4ⴢ2MnSO4 [100], but no Tutton’s type.
Maybe the latter can be found at lower temperatures, however, already at ambient temperatures equi-
libration times of 1–2 months had to be applied to obtain the equilibrium solid phases.

Table 1 Evidence for the existence of Tutton’s salts M2SO4ⴢM'SO4ⴢ6H2O

after Nalbandyan [99]; number of PDF entries, “no success” means no
successful attempts for preparation, “no info” means no information at all.
M2+ K Rb Tl NH4 Cs
Ni 3 2 1 7 2
Ru 1 no info no info no info no info
Mg 2 2 1 5 2
Cu 2 2 2 6 2
Co 4 2 1 4 2
Zn 3 2 1 5 2
Fe 1 2 1 2 2
V no info no info no info 2 no info
Cr no info no info no info 1 no info
Mn no success 1 1 2 2
Cd no info no info no success 2 no success


Enhancing the temperature the solubility of salts can increase or decrease, for MgSO4 both can be
observed as shown in Fig. 9. At temperatures near 200 °C MgSO4 becomes a sparingly soluble salt
comparable with CaSO4. The negative temperature coefficient of solubility is caused by strongly

Fig. 9 Temperature dependence of the solubility of MgSO4, symbols: experimental data (filled rectangles:
heptahydrate, rhombus: hexahydrate, stars: monohydrate, line: calculated with THEREDA model).

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
Chemistry of salts in aqueous solutions 1025

decreased ion hydration and ion association. Addition of another electrolyte turns the negative temper-
ature coefficient into a positive one if sufficiently high concentrations are reached.
This topic was thoroughly discussed by Valyashko [101]. This phenomenon was recently also
confirmed the solubility of CaSO4, anhydrite, in solutions of NaCl at enhanced temperatures [102],
which is nicely visible at the crossing point of the solubility isotherms at 7.0 mol/kg H2O (see Fig. 10).

Fig. 10 Solubility isotherms of anhydrite at 373 (), 423 (), and 473 () K in dependence on NaCl concentration;
large symbols: data [102], arrows: direction of temperature coefficient of solubility.

In mixed salt solutions at very high concentrations due to the coordination competition of cations
for water and anions, peculiar behavior can occur as is illustrated for the system KCl–MgCl2–H2O in
Fig. 11. The course of the solubility isotherms of KCl in dependence on the concentration of MgCl2
passes a minimum and at extremely high MgCl2 concentration, in the range of molten hydrates [103]

Fig. 11 Solubility isotherms in the system KCl–MgCl2–H2O at 160 °C (*), 170 °C (○), and 180 °C () [105].
Arrow and hatched area: supersaturation of KCl when water is added to a saturated solution at point A. other solids:
C = carnallite, KClⴢMgCl2ⴢ6H2O, M1 = MgCl2ⴢ4H2O, M2 = solid solution MgCl2ⴢ2H2O–KClⴢMgCl2ⴢ2H2O,
M3 = KClⴢMgCl2ⴢ2H2O, M4 = 1.5 KClⴢMgCl2ⴢ2H2O.

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
1026 W. VOIGT

solubility increases again due to action of KCl as a chloride ion donator for MgCl2, which forms chloro-
complexes [104]. This has the effect that at T ≥ 160 °C and m(MgCl2) > 130 mol/1000 mol H2O
(>7.2 mol/kg H2O) a very soluble salt such as KCl can be crystallized by addition of water (Fig. 11),
which is quite surprising.
The thermodynamic properties in the range of molten salt hydrates are successfully described by
means of the modified Brunauer–Emmett–Teller (BET) model [106–111]. This model uses only two
parameters for every salt component (maximum hydration number and the adsorption energy of water
at the salt component). Often the temperature dependence of these parameters is sufficiently described
by a linear or quadratic function. When salt–salt interactions are strong (complex formation) the sim-
ple BET model cannot be used.
Recently, it was shown that models with a small number of parameters can be formulated, when
the enthalpy and entropy of step-wise formation of hydrated ions and anionic complexes are interrelated
in reaction chains [112]. Formation data of the anhydrous species can be estimated from thermo-
dynamic measurements in anhydrous molten mixtures at much higher temperatures. An example of the
description quality is given for the system KCl–MgCl2–H2O in Fig. 12. For this system, enthalpic and
entropic parameters for the following species were derived from all available thermodynamic data in
the aqueous and molten anhydrous system (for details, see [110]): MgCl2(H2O)n (i = 0,...,6),
Mg(H2O)2+6+n, K2(MgCl2)nCl2 (n = 1,..,5), K(H2O)+, K(H2O)2+, KMgCl3(H2O)3, KMgCl3(H2O)4.
Considering the large interval of composition and temperature a fairly good description of the thermo-
dynamic quantities of liquid phase could be achieved with 20 parameters. For the phase diagrams only
the formation enthalpy and entropy of the solid hydrates had to be adjusted.

Fig. 12 Simultaneous description of the phase diagram of the system KCl–MgCl2–H2O (A) between 373–523 K
and the thermodynamic properties of the anhydrous molten system KCl–MgCl2 at 1073 K (B) applying a reaction
chain model [112].

Further peculiarities of molten salt hydrates concern the acidity and reactivity. The acidity of sim-
ple solutions like ZnCl2 or Al(NO3)3 become comparable with strong mineral acids like HCl, when the
molar water/salt ratio is diminished below 6 [113]. In these liquids, noble metals as gold can be dis-
solved at temperatures below 100 °C [114]. In essence, molten salt hydrates represent protic ionic liq-
uids, the reactivity of which can be controlled by the water content and specific choice of cation-anion
combination. An idea of the structural situation and resulting reactivity can be gained from Fig, 13 for
molten ferric nitrate pentahydrate. The water molecules are strongly polarized between nearest-neigh-

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
Chemistry of salts in aqueous solutions 1027

Fig. 13 Possible structural situation in molten ferric nitrate pentahydrate to explain the nitrating effect.

bor cations and anions, such that proton transfer is facilitated and an activity of HNO3 results. Direct
coordination of NO3– at Fe3+ can generate a situation, which is comparable with the formation of a
nitryl cation. These effects are obviously responsible for the nitrating ability of this molten hydrate. In
this medium, toluene can be nitrated to mono- and dinitro-toluene [115], a reaction for which, in gen-
eral, mixtures of concentrated sulfuric and nitric acid are needed.

Modern science and technology requires accurate property data of aqueous salt solutions at extending
ranges of temperature, pressure, and composition. At the same time, a growing number of ionic con-
stituents have to be considered. The computing power of personal computers is sufficient to calculate
complex solubility and speciation equilibria of multicomponent water–salt systems by means of con-
temporary equilibrium calculation codes. However, the results are limited by the underlying thermo-
dynamic databases and parameters of solution models. The limitations concern
• the accuracy of the description of experimental data,
• the available model coefficients for the various ion–ion interactions,
• the standard data for solid salts, particularly hydrates and double salts, and
• the ranges of temperature and pressure.
Current physical models cannot predict missing data for solid phases, particularly hydrates and double
salts, with the required accuracy. For extrapolations of mixed solution properties toward higher con-
centrations, simple mixing rules such as Zdanovskii’s or Young’s rules should be preferred to extrapo-
lation of validity ranges of more complex models.
Application-oriented experimental programs are set up to fill some of the data gaps. However, in
parallel a renaissance of systematic physico-chemical investigations of electrolyte systems would accel-
erate the development of structure–property-related models suited for predictions.

Continuous financial support from the German Federal Ministry of Education and Research and the
German Federal Ministry of Economy is acknowledged.

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
1028 W. VOIGT

1. M. Ladd. Chemical Bonding in Solids and Fluids, Ellis Horwood, New York (1994).
2. W. Kunz (Ed.). Specific Ion Effects, World Scientific, Singapore (2010).
3. J. D’Ans. Die Löslichkeitsgleichgewichte der Systeme der Salze ozeanischer Salzablagerungen,
Verlagsgesellschaft für Ackerbau mbH, Berlin (1933).
4. H. Autenrieth. Kali Steinsalz 2, 18 (1958).
5. H. Autenrieth, G. Braune. Kali Steinsalz 3, 395 (1959).
6. Th. Fanghänel, H.-H. Emons. Neue Ergebnisse über die fest-flüssig Gleichgewichte der Systeme
der ozeanischen Salze (T > 100 °C), Abh. Sächs. Akad. Wiss. zu Leipzig 57, 3 (1992).
7. (a) R. Beck, H.-H. Emons, H. Holldorf. Freib. Forsch. A 628, 7 (1981); (b) R. Beck, H.-H. Emons,
H. Holldorf. Freib. Forsch. A 628, 19 (1981).
8. W. Voigt. Freibforsch. A 853, 5 (1999).
9. W. Voigt. Pure Appl. Chem. 73, 831 (2001).
10. D. Freyer, W. Voigt. Geochim. Cosmochim. Acta 68, 307 (2004).
11. D. E. Garrett. Handbook of Lithium and Natural Calcium Chloride, Elsevier, Academic Press
12. M. Steiger. Restor. Build. Monum. 11, 419 (2005).
13. M. Steiger, J. Asmussen. Geochim. Cosmochim. Acta 72, 4291 (2008).
14. D. Möhlmann, K. Thomsen. Icarus 212, 123 (2011).
15. E. Königsberger, G. Hefter, P. May. Pure Appl. Chem. 81, 1537 (2009).
16. P. Debye, E. Hückel. Phys. Z. 24, 185 (1923).
17. J. N. Brönsted. Trans. Faraday Soc. 23, 416 (1927).
18. E. A. Guggenheim. Philos. Mag. 22, 322 (1936).
19. E. A. Guggenheim, J. C. Turgeon. Trans. Faraday Soc. 51, 747 (1955).
20. G. Scatchard, R. M. Rush, J. S. Johnson. J. Phys. Chem. 74, 3786 (1970).
21. P. J. Reilly, R. H. Wood, R. A. Robinson. J. Phys. Chem. 75, 1305 (1971).
22. K. S. Pitzer. In Activity Coefficients in Electrolyte Solutions, 2nd ed., K. S. Pitzer (Ed.), pp.
75–153, CRC Press, Boca Raton (1991).
23. THEREDA project, <>.
24. I. Grenthe, I. Puigdomenech. Modeling in Aquatic Chemistry, NEA-OECD Publications, Paris
25. S. A. Bea, J. Carrera, C. Ayora, F. Battle. Computers Geosci. 36, 526 (2010).
26. W. Voigt, V. Brendler, K. Marsh, R. Rarey, H. Wanner, M. Gaune-Escard, P. Cloke, Th. Vercouter,
E. Bastrakov, S. Hagemann. Pure Appl. Chem. 79, 883 (2007).
27. X. Xu, E. A. Macedo. Ind. Eng. Chem. Res. 42, 5702 (2003).
28. G. R. Pazuki, F. Arabgol. CALPHAD 29, 125 (2005).
29. A. Haghtalab, M. Joda. Int. J. Thermophys. 28, 876 (2007).
30. A. Haghtalab, A. Shojaeian, S. H. Mazloumi. J. Chem. Thermodyn. 43, 354 (2011).
31. X. Ge, X. Wang. J. Chem. Eng. Data 54, 179 (2009).
32. C. C. Chen, H. J. Britt, J. F. Boston, L. B. Evans. AIChE J. 28, 589 (1982).
33. X. Lu, G. Maurer. AIChE J. 39, 1527 (1993).
34. X. Lu, L. Zhang, Y. Wang, J. Shi. Fluid Phase Equilib. 116, 201 (1996).
35. M. C. Iliuta, K. Thomsen, P. Rasmussen. AIChE J. 48, 2664 (2002).
36. A. Jaretun, G. Aly. Fluid Phase Equilib. 175, 213 (2000).
37. H. Kuramochi, M. Osako, A. Kida, K. Nishimura, K. Kawamoto, Y. Asakuma, K. Fukui,
K. Maeda. Ind. Eng. Chem. Res. 44, 3289 (2005).
38. H.-M. Polka, J. Li, J. Gmehling. Fluid Phase Equilib. 162, 97 (1999).
39. J. Li, Y. Lin, J. Gmehling. Ind. Eng. Chem. Res. 44, 1602 (2005).

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
Chemistry of salts in aqueous solutions 1029

40. J. M. G. Barthel, H. Krienke, W. Kunz. Physical Chemistry of Electrolyte Solutions, Steinkopf

Darmstadt, Springer, New York (1998).
41. Z.-C. Wang. J. Chem. Eng. Data 54, 187 (2009).
42. E. Königsberger, P. May, B. Harris. Hydrometallurgy 90, 177 (2008).
43. E. Königsberger, L.-C. Königsberger, P. May, B. Harris. Hydrometallurgy 90, 168 (2008).
44. Y. Marcus. Pure Appl. Chem. 59, 1093 (1987).
45. Y. Marcus. J. Chem. Soc., Faraday Trans. 87, 2995 (1991).
46. M. W. Chase. J. Phys. Chem. Ref. Data, Monograph No. 9, NIST-JANAF Thermochemical Tables,
4th ed., ACS, AIP, NIST (1998).
47. D. D. Wagman, W. H. Evans, V. B. Parker, R. H. Schumm, I. Halow, S. M. Bailey, K. L. Churney,
R. L. Nuttall. J. Phys. Chem. Ref. Data 11, Suppl. 2 (1982).
48. H. C. Helgeson, D. H. Kirkham, G. C. Flowers. Am. J. Sci. 281, 1249 (1981).
49. J. R. Haas, E. L. Shock, D. C. Sassani. Geochim. Cosmochim. Acta 59, 4329 (1995).
50. J. W. Johnson, E. Oelkers, H. C. Helgeson. Computers Geosci. 18, 899 (1992).
51. (a) E. Djamali, J. W. Cobble. J. Phys. Chem. B 113, 2398 (2009); (b) E. Djamali, J. W. Cobble.
J. Phys. Chem. B 113, 5002 (2009); (c) E. Djamali, J. W. Cobble. J. Phys. Chem. B 113, 10792
52. E. Djamali, J. W. Cobble. J. Phys. Chem. B 114, 3887 (2010).
53. R. J. Fernandez-Prini, H. R. Corti, M. L. Japas. High-Temperature Aqueous Solutions:
Thermodynamic Properties, CRC Press, Boca Raton, London (1992).
54. J. Rard, R. F. Platford. In Activity Coefficients in Electrolyte Solutions, 2nd ed., K. S. Pitzer (Ed.),
pp. 75–153, CRC Press, Boca Raton (1991).
55. A. Apelblat, E. Korin. J. Chem. Thermodyn. 33, 113 (2001).
56. A. Apelblat, E. Korin. J. Chem. Thermodyn. 39, 1065 (2007).
57. M. Gruszkiewicz, J. M. Simonson. J. Chem. Thermodyn. 37, 906 (2005).
58. J. N. Butler, R. N. Roy. In Activity Coefficients in Electrolyte Solutions, 2nd ed., K. S. Pitzer (Ed.),
pp. 75–153, CRC Press, Boca Raton (1991).
59. W. J. Hamer. J. Am. Chem. Soc. 57, 9 (1935).
60. G. M. Giordano, P. Longhi, T. Mussini, S. Rondinini. J. Chem. Thermodyn. 9, 997 (1977).
61. Z. Jun, G. Shi-Yang, X. Shu-Ping. J. Chem. Thermodyn. 35, 1383 (2003).
62. M. J. V. Lourenco, F. J. V. Santos, M. L. V. Ramires, C. A. Nieto de Castro. J. Chem. Thermodyn.
38, 970 (2006).
63. S. Schrödle, E. Königsberger, P. M. May, G. Hefter. Geochim. Cosmochim. Acta 74, 2368 (2010).
64. D. Smith-Magowan, R. H. Wood. J. Chem. Thermodyn. 13, 1047 (1981).
65. J. S. Jones, S. P. Ziemer, B. R. Browns, E. M. Woolley. J. Chem. Thermodyn. 39, 550 (2007).
66. A. W. H. Hakin, J. L. Liu, K. Erickson, J.-V. Munoz, J. A. Rard. J. Chem. Thermodyn. 37, 153
67. Ch. S. Oakes, J. A. Rard, D. G. Archer. J. Chem. Eng. Data 49, 313 (2004).
68. K. Ballerat-Busserolles, M. L. Origlia, E. M. Woolley. Thermochim. Acta 347, 3 (2000).
69. X. Chen, J. L. Oscarson, S. E. Gillespie, R. M. Izatt. Thermochim. Acta 11, 285 (1996).
70. D. A. Polya, E. M. Woolley, J. M. Simonson, R. E. Mesmer. J. Chem. Thermodyn. 33, 205 (2001).
71. G. del Re, G. di Giacomo, F. Fantauzzi. Thermochim. Acta 161, 201 (1990).
72. G. Wolf, H. Jahn. Thermochim. Acta 116, 291 (1987).
73. L.-O. Öhman. Chem. Geol. 151, 41 (1998).
74. J. Szilágyi, E. Königsberger, P. M. May. Dalton Trans. 7717 (2009).
75. J. Bebie, T. M. Seward, J. K. Hovey. Geochim. Cosmochim. Acta 62, 1643 (1998).
76. J. Brugger, D. C. McPhail, J. Black, L. Spiccia. Geochim. Cosmochim. Acta 65, 2691 (2001).
77. T. Gajda, P. Sipos, H. Gamsjäger. Monatsh. Chem. 140, 1293 (2009).
78. P. Sipos, G. Hefter, P. M. May. Talanta 70, 761 (2006).
79. W. W. Rudolph, G. Irmer, E. Königsberger. Dalton Trans. 900 (2008).

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011
1030 W. VOIGT

80. G. T. Hefter, R. P. T. Tomkins (Eds.). The Experimental Determination of Solubilities, Wiley

Series of Solution Chemistry, Vol. 6, John Wiley, New York (2003).
81. J. H. Van’t Hoff. Z. Anorg. Chem. 47, 244 (1905).
82. H. S. Van Klooster. J. Phys. Chem. 106, 93 (1917).
83. Ch. E. Harvie, N. Möller, J. H. Weare. Geochim. Cosmochim. Acta 48, 723 (1984).
84. F. K. Cameron, J. M. Bell. J. Phys. Chem. 10, 202 (1906).
85. A. S. Kolosov. Trud. Khim.-Metallug. Inst., Sap.-Sibirsk. Filial. Akad. Nauk. SSSR 35, 35 (1958).
86. G. Wollmann. Crystallization filed of Polyhalite and its Heavy Metal Analogues, Dissertation, TU
Bergakademie Freiberg (2010).
87. Ch. Balarew, R. Duhlev. J. Solid State Chem. 55, 1 (1984).
88. R. C. Peterson, R. Wang. Geology 34, 957 (2006).
89. F. E. Genceli, M. Lutz, A. L. Spek, G.-J. Witkamp. Cryst. Growth Des. 7, 2460 (2007).
90. P. D. Robinson, J. H. Fang, Y. Ohya. Am. Mineral. 57, 1325 (1972).
91. J. H. Fang, P. D. Robinson. Am. Mineral. 55, 378 (1970).
92. H. D. B. Jenkins, L. Glasser, J. F. Liebman. J. Chem. Eng. Data 55, 4369 (2010).
93. H. D. B. Jenkins, L. Glasser. J. Chem. Eng. Data 55, 4231 (2010).
94. H. D. B. Jenkins, L. Glasser. Inorg. Chem. 41, 4378 (2002).
95. C. H. Yoder, N. R. Gotlieb, A. N. Rowand. Am. Mineral. 95, 47 (2010).
96. C. H. Yoder, J. P. Rowand. Am. Mineral. 91, 747 (2006).
97. M. Jansen. Angew. Chem. 114, 3896 (2002).
98. I. V. Pentin, J. C. Schön, M. Jansen. Z. Anorg. Allg. Chem. 636, 1703 (2010).
99. V. B. Nalbandyan. Powder Diffr. 23, 52 (2008).
100. G. Wollmann, W. Voigt. Fluid Phase Equilib. 291, 76 (2010).
101. V. M. Valyashko. Phase Equilibria and Properties of Hydrothermal Systems, Nauka, Moscow
102. D. Freyer, W. Voigt. Geochim. Cosmochim. Acta 68, 307 (2004).
103. H.-H. Emons, Th. Fanghänel, W. Voigt. Salzhydratschmelzen: Ein Bindeglied zwischen
Elektrolytlösungen und Salzschmelzen, Sitzungsber. AdW DDR 3N Akademieverlag, Berlin
104. W. Voigt, P. Klæboe. Acta Chem. Scand. A40, 354 (1986).
105. Th. Fanghänel, H.-H. Emons, K. Köhnke. Z. Anorg. Allg. Chem. 576, 99 (1989).
106. M. R. Ally, J. Braunstein. J. Chem. Thermodyn. 30, 49 (1998).
107. M. R. Ally. J. Chem. Eng. Data 54, 411 (2009).
108. S. L. Clegg, J. M. Simonson. J. Chem. Thermodyn. 33, 1457 (2001).
109. W. Voigt, D. Zeng. Pure Appl. Chem. 74, 1909 (2002).
110. D. Zeng, W. Voigt. CALPHAD 27, 243 (2003).
111. Y. Marcus. J. Solution Chem. 34, 307 (2005).
112. W. Voigt, K. Hettrich, D. Zeng. “Ion coordination and thermodynamic modeling of molten salt
hydrate mixtures”, in Thermodynamic Properties of Complex Fluid Mixtures, G. Maurer (Ed.),
pp. 241–266, Wiley-VCH, Weinheim (2004).
113. S. K. Franzyshen, M. D. Schiavelli, K. D. Stocker, M. D. Ingram. J. Phys. Chem. 94, 2684 (1990).
114. E. J. Sare, C. T. Moynihan, C. A. Angell. J. Phys. Chem. 77, 1869 (1973).
115. W. Mackenroth, J. Büttner, E. Ströfer, W. Voigt, F. Bok. Verfahren zur Herstellung von nitrierten
Aromaten und deren Mischungen, WO 2010/097453 A1.

© 2011, IUPAC Pure Appl. Chem., Vol. 83, No. 5, pp. 1015–1030, 2011