Reversible formation of alcohol solvates and their potential use for heat storage

Abstract

In this study, CaCl₂- and MgCl₂–alcohol solvates of different stoichiometric quantities of ethyl alcohol (EtOH) and methyl alcohol (MeOH) were synthesized and characterized via coupled thermogravimetric–differential scanning calorimetry, thermogravimetric–mass spectrometric evolved gas analysis (TG-MS), spectroscopic analysis (Raman) methods as well as by X-ray diffraction. Correlations between the obtained calorimetric, thermodynamic, kinetic, and crystallographic data were carried out. The CaCl₂–alcohol systems seem suitable for heat storage based on the feasible recovery of the salt. However, Raman spectroscopic results revealed that the MgCl₂–EtOH solvates were instable compounds. Irreversible transformation of MgCl₂–alcohol solvates related to the formation of alkyl chloride appeared upon heating, as proven by TG-MS and Raman spectroscopic measurements. Pure salt–alcohol solvates could not be prepared under technically applicable conditions. The samples contained at least traces of water. Appearance of side reactions resulting in magnesium oxychlorides, oxyhydroxides, and possible release of HCl with cycling may contribute to corrosion of reactor components. Based on these considerations, MgCl₂–alcohol solvate systems are not recommended for heat storage.

Introduction

Storage and utilization of thermal energy are of utmost importance for a sustainable energy future. There are three main approaches in this field, e.g., sensible heat storage in water [1, 2], utilization of latent heat by phase change materials (PCM) [3,4,5,6,7,8,9,10,11,12,13,14,15], and conversion to chemical energy by different thermochemical reactions [16,17,18,19,20,21,22]. Heat storage via reversible thermochemical reactions has several benefits, such as high energy density per volume material and low long-term losses. Space can be a limiting factor for many practical applications. Most mobile applications, for instance, demand high energy densities in small volumes.

To date, the reaction of inorganic salts with water to form salt hydrates under heat release (1) has mainly been studied and optimized [23,24,25,26]. Alcohols are known to undergo similar reactions with salts (2). These reactions have only been studied in a casual manner in terms of their application in thermal energy storage [27, 28], although the high vapor pressure and low freezing point of EtOH and MeOH facilitate low-temperature applications unlike the aqueous adducts.

$${\text{MX}} \cdot (y - z){\text{H}}_{2} {\text{O}}\,({\text{s}}) + z{\text{H}}_{2} {\text{O}}\,({\text{g}}) \rightleftharpoons {\text{MX}} \cdot y{\text{H}}_{2} {\text{O}}\,({\text{s}}) + \Delta_{\text{r}} H_{1}$$

(1)

$${\text{MX}} ({\text{s}})+ z{\text{R-OH}}\,({\text{g}}) \rightleftharpoons {\text{MX}} \cdot z{\text{R-OH}}\,({\text{s}}) + \Delta_{\text{r}} H_{2}$$

(2)

During the exothermic association reaction (discharging mode), the solid inorganic salt MX reacts with the gaseous alcohol R-OH forming the salt–alcohol solvate (alcoholate) MX·zR-OH, where z is the stoichiometric coefficient. The chemical energy stored in the reaction is released as heat of reaction Δ_rH₂. The reverse reaction (charging mode) is endothermic. Heat energy must be supplied to initiate the dissociation of the MX·zR-OH adduct. The gaseous product R-OH can be collected and condensed. Therefore, the storage volume can be reduced and the reaction products can be stored separately. As a result, a long-term storage without sensible heat losses is technically feasible. A schematic diagram of the operating principle of a closed thermochemical heat storage system is shown in Fig. 1.

Fig. 1

Schematic diagram of the operating principle of a closed thermochemical heat storage system

Fast reaction kinetics, precise thermodynamic control, and low regeneration temperatures make salt–alcohol pairs an advantageous option. Recently, the suitability of EtOH solvates of CaCl₂, MgCl₂, and their mixtures as heat storage materials for practical implementations has been studied [29]. The EtOH sorption ability of CaCl₂ was better than that of MgCl₂. At high EtOH vapor pressures, overstoichiometric EtOH uptake occurred. The CaCl₂–EtOH reaction system exhibited convenient sorption properties coupled with good multi-cyclic stability. Consequently, it has a great potential for low-grade thermal energy storage. However, poor reversibility appeared in case of the MgCl₂–EtOH system with increasing number of cycles, probably caused by irreversible decomposition that strongly reduced the material’s sorption performance [29].

Due to the variety of synthesis procedures and conditions, an apparent discrepancy of the stoichiometry of salt alcoholates is found in the published literature. Moreover, present work on the energy storage density is limited. The aim of this work was to prepare and characterize salt–alcohol solvates of different stoichiometry based on anhydrous CaCl₂ and MgCl₂ by use of commercially available, specially non-purified chemicals with methods that can be relatively easily transposed into practice. The salt–alcohol solvates were prepared by direct synthesis from CaCl₂/MgCl₂ and EtOH/MeOH. This preparation method is based on the operating principle of a thermochemical heat store and can easily be adapted to practice. The focus was laid not only on the detailed comparative study of the calorimetric behavior, but also on the mapping of the heat-induced changes in these solvates. Besides coupled thermogravimetric–mass spectrometric (TG-MS) [30, 31], thermogravimetric analysis/differential scanning calorimetric (TGA/DSC) [32] techniques are generally applied in heat storage studies. We believed that Raman spectroscopy [33, 34] and X-ray powder diffraction (XRD) [35] methods are informative in sample characterization. Use of above techniques on structural changes in salt alcoholates is quite unique. We tried to find relationships between structural changes in the salt–alcohol solvates and the cyclic stability during repeated alcoholation/dealcoholation reactions observed earlier [29].

In solid state, the solvent molecules could be in a well-localized form and in solution they could be located in the first, second, and so on solvation shell. The interaction energies between the solvent molecules and solvated ions can vary over a broad range. Conducting single-crystal diffraction experiments for each composition was not possible and may be the subject of another work. Throughout this article, the term built-in form refers to the solvent molecules coordinated to the salt ions and also to the molecules loosely bound to the complexes studied, which dissociate upon crystallization or slightly elevating the temperature.

Experimental

Synthesis of salt–alcohol solvates

Anhydrous CaCl₂ powder (Merck, Ph Eur), anhydrous MgCl₂ powder (Roth, ≥ 98.5%), absolute MeOH (max. 0.003% H₂O, Merck), and absolute EtOH (max. 0.01% H₂O, Merck) were used as starting materials. The H₂O content of pure CaCl₂ and MgCl₂ samples as received was determined by TGA. The powdered salts contained only traces of H₂O.

Alcohol-salt pairs in four combinations, i.e., MeOH/CaCl₂, EtOH/CaCl₂, MeOH/MgCl₂, EtOH/MgCl₂, were prepared setting different alcohol/salt molar ratios (Table 1). The alcohol/salt stoichiometric ratios varied between 2 and 8 mol of EtOH or MeOH per mole anhydrous CaCl₂ or MgCl₂. Samples of 10 g of the respective salt were placed into a round bottom flask that was purged with dry nitrogen (N₂) prior to each alcoholate synthesis reaction. The flask was equipped with a magnetic stirrer and two thermometers. A stoichiometric amount of liquid alcohol was dropped onto the vigorously agitated salt under dry N₂ atmosphere, while both the temperature of the sample and the vapor phase were measured. Changes in textural and optical properties of the alcoholated salts were recorded. Preparation and denomination of the samples obtained are summarized in Table 1. The prepared samples were stored in closed glass vessels in a vacuum desiccator over P₂O₅.

Table 1 Preparation of salt–alcohol solvates by direct synthesis from liquid phase alcohols under neat conditions

Characterization methods

Simultaneous thermogravimetric analysis and differential scanning calorimetry

Experimental alcohol/salt ratios of the salt–alcohol solvates prepared at different theoretical stoichiometric molar ratios, according to Table 1, and associated heat absorptions were identified by simultaneous TGA/DSC. The horizontal furnace TGA/DSC 1 from Mettler Toledo was used. Specimens of 10–15 mg were uniformly placed into alumina crucibles of 70 µL volume and kept isothermally at 30 °C for 15 min to stabilize the sample. Dynamic runs were carried out by scanning the specimen from 30 to 180 °C with a temperature ramp of 3 °C min⁻¹ at atmospheric pressure. During subsequent isothermal stabilization at the final temperature, the mass was recorded until the reaction went to completion. The TGA/DSC instrument was purged with N₂ at a flow rate of 50 mL min⁻¹ during the entire measurement. Effects of buoyancy forces and temperature changes on the TGA and DSC signal were eliminated by automatic blank curve correction. For subtraction of the blank curve from the measurement curve, a blank curve was recorded under the same temperature conditions as the measurement curve, but using empty reference and sample crucibles. All measurements were performed under well-controlled laboratory conditions. The TGA/DSC device was calibrated with high purity metal standards (gallium, indium, lead, aluminum, and gold) over the temperature range of interest. For temperature and enthalpy calibration, the calibration substances were subjected to the same temperature and heating conditions as the samples analyzed. The measured onset temperature, which is assigned the start of the melting process, was compared with the reference melting point of the respective standard. The determined enthalpy of fusion was also validated by comparison with the reference value. The blank curve reproducibility was better than ± 10 µg over the whole temperature range. The standard deviation of the enthalpy reproducibility was given as < 5%. Mass changes and heat powers were determined with a precision of ± 0.1 µg and ± 1 mW. The Mettler Toledo STARe software 11.00a was used for data processing.

Selected samples of the series of salt–alcohol solvates were characterized complimentary by Raman spectroscopy, TG-MS, and XRD. In case of the first two methods, N₂ atmosphere was used for sample handling.

Raman spectroscopy

Raman spectra of samples in closed glass ampoules held at room temperature were recorded with a dynamically aligned Bio-Rad (Digilab) dedicated FT-Raman spectrometer equipped with a Spectra Physics Nd–YAG laser (1064 nm) and high sensitivity liquid-N₂-cooled Ge detector. The excitation laser power used was about 250 mW at the samples. Increase in laser power up to 500 mW does not result in a significant change of the spectra that indicates the stability of the samples during the measurement. The resolution of the Raman instrument was ca. 4 cm⁻¹, and a backscattered geometry was used. For each spectrum, 256 individual spectra were averaged.

Thermogravimetric–mass spectrometric evolved gas analysis

The simultaneous thermogravimetric and mass spectrometric evolved gas analyses (TG-MS) were recorded on a Setaram LabsysEvo thermal analyzer, in high purity (99.9999%) helium atmosphere, with a flow rate of 80 mL min⁻¹. The measurements were done with a heating rate of 20 °C min⁻¹; the samples were weighed into 100 μL aluminum crucibles in inert (dry N₂) atmosphere. Exposure to moisture during sample transfer was prevented, and MS recording of evolved volatiles was enabled by closing the crucibles with aluminum lids pierced with a 700 micron hole by crimping. The measurements were performed in the 25–500 °C temperature range. The obtained results were baseline corrected and then evaluated with the thermal analyzer’s processing software (AKTS Calisto Processing, ver. 1.41). Parallel with the thermogravimetric measurements, the analysis of the evolved gases/volatiles was carried out on a Pfeiffer Vacuum OmniStar™ gas analysis system coupled to the above-described TGA. The gas splitter and transfer line to the mass spectrometer were preheated to 250 °C. The scanned m/z interval was 5–79 amu, with a scan speed of 20 ms amu⁻¹. The mass spectrometer was operated in electron impact mode. During a measurement, the total ion current (TIC), the discrete ion current of all scanned masses (75 masses) and the analog spectra on each scan cycle (1 scan cycle was 1.5 s long) were obtained in parallel.

X-ray powder diffraction

XRD patterns were obtained in a Philips model PW 3710 based PW 1050 Bragg–Brentano parafocusing goniometer using CuK_α radiation (λ = 0.15418 nm), graphite monochromator and proportional counter. Samples were placed in an Anton-Paar HTK2000 high-temperature oven chamber. Prior to measurements, the chamber was flushed with high purity N₂ (99.9999%) and during XRD measurements a continuous slow N₂ flow was ensured. The temperature steps chosen for the “temperature programmed” XRD were based on the results of the TG-MS measurement of the appropriate sample. First, XRD patterns were recorded at room temperature, which was followed by stepwise in situ heating. The heating rate was 5 °C min⁻¹. XRD patterns were recorded at every temperature step. Finally, the samples were allowed to cool down (20 °C min⁻¹) and the room temperature XRD pattern was again measured. For the phase analysis, reference cards from the ICDD PDF-4 (2010) database were used.

Results and discussion

Preliminary results of the direct synthesis of salt–alcohol solvate

During the formation of the different salt alcoholate solvates, heat was evolved and resulted in a change in the temperature. The temperature change in the sample and the surrounding vapor phase was calculated from the difference between the respective initial and final temperatures. The initial sample temperatures ranged from 24 to 27 °C. The increase in temperature in both the sample and vapor chamber laid in the interval of 15–25 °C for almost all salt–alcohol pairs (Fig. 2). Rather fast reactions occurred and maximum temperatures of 40–56 °C were reached within a time of less than 3 min. In general, the temperature rise was higher in the solid phase than in the vapor phase, except for some outliers. For CaCl₂–alcohol solvates, higher temperature differences were measured when compared with that for MgCl₂–alcohol solvates. Since the temperature difference of both CaCl₂–MeOH and CaCl₂–EtOH solvates varied between 20 °C and 25 °C, irrespective of the alcohol/salt molar ratio, similar stoichiometric compositions of the different salt–alcohol solvates were considered. It is assumed that some heat dissipated into the ambient.

Fig. 2

Temperature increase during the synthesis of salt–alcohol solvates, according to Table 1. MeOH methanol; EtOH ethanol; diamond—sample, square—vapor chamber

Lower temperature differences compared with that of CaCl₂ hydrates [36, 37] indicate lower heat outputs and corresponding enthalpies of reaction. Heterogeneous compounds were observed particularly for samples of higher stoichiometry. Some compositions consisted of two phases: a solid salt alcoholate phase and an excess liquid alcohol phase containing dissolved salt particles. Thus, further analysis was carried out using the solid salt–alcohol solvates only.

Results of TGA/DSC measurements of all samples

The composition of the different salt–alcohol solvates synthesized according to Table 1 was analyzed by TGA/DSC technique. The respective alcohol/salt molar ratios were derived from the change in sample mass recorded as a function of both time and temperature.

Thermal analysis of the CaCl₂–MeOH system

The total amount of MeOH uptake of CaCl₂ was calculated to be 4 (Table 2). Samples decomposed in 2–3 overlapping stages over a temperature domain of 33–145 °C (Fig. 3). The majority of MeOH was given off below 100 °C. Peak temperatures of 57–94 °C, 80–114 °C, and 95–118 °C were measured (Table 2). As depicted in Fig. 3, the mass loss rate of the first decomposition step, also referred to as rate of dealcoholation, increased with increasing stoichiometric ratio despite similar sample masses. At theoretical alcohol/salt ratios of 2 and 3, a compound doubly coordinated with MeOH was the only CaCl₂–MeOH solvate found. This result is consistent with the findings of other authors [27, 38,39,40,41,42]. With excess of absolute MeOH, even a CaCl₂ trimethanolate and CaCl₂ tetramethanolate could be synthesized (Table 2). Note that samples prepared at higher theoretical alcohol/salt ratios were a heterogeneous mixture of two phases: a solid phase and a liquid phase. The latter was discarded from analysis. The formation of CaCl₂–MeOH complexes of different stoichiometry has been claimed by other researchers. Gmelin [43] has reported the existence of CaCl₂ monomethanolates and CaCl₂ trimethanolates that have been identified in a study conducted by Gerhold and Kahovec. Bonnell [44] and Menschutkin [45] have obtained CaCl₂ trimethanolates, too. MeOH solvates of CaCl₂ with a molar ratio of 4 [43,44,45,46,47,48] and 6 [49] have been characterized by different analysis methods. The standard reaction enthalpies were deduced from the DSC curve by peak integration and varied between 101 and 151 kJ mol⁻¹ (Table 2).

Table 2 Data obtained from TGA/DSC measurements

Fig. 3

Mass (TGA) and mass loss rate (DTG) curves of the dealcoholation of CaCl₂–MeOH solvates prepared in neat at different MeOH/CaCl₂ molar ratios (see Table 1), (measurement temperature interval: 30–180 °C; heating rate of 3 °C min⁻¹; 20 min isotherm at 180 °C)

Thermal analysis of the CaCl₂–EtOH system

In CaCl₂–EtOH solvate complexes (Fig. 4), the EtOH was evolved in up to four inseparable steps with peak temperature around 68–72 °C, 87–92 °C, 108–121 °C, and 150–177 °C, respectively (Table 2). The decomposition started roughly at 33 °C and was completed at around 178 °C. The major amount of EtOH was desorbed below 100 °C. The rate of deethanolation varied over a broad range and was not reproducible, as shown in Fig. 4. The calculated percentage of mass loss (Table 2) clearly showed that in each sample only 2 mol of EtOH was adsorbed per mole of anhydrous CaCl₂. However, a variation in the standard enthalpy of reaction with values between 92 and 130 kJ mol⁻¹ was observed. We suppose that CaCl₂ could only hold up to two molecules of EtOH under the experimental conditions studied, irrespective of the initial stoichiometric ratio we used during sample preparation. These data are consistent with the findings described in the literature [27, 50]. According to published literature data, EtOH forms also a number of other CaCl₂ ethanolates. Monoethanolates [43, 50, 51], triethanolates [43,44,45, 52, 53], and tetraethanolates [53,54,55] of CaCl₂ have been reported previously. Their existence was not proven in this study.

Fig. 4

Mass (TGA) and mass loss rate (DTG) curves of the dealcoholation of CaCl₂–EtOH solvates prepared in neat at different EtOH/CaCl₂ molar ratios (see Table 1), (measurement temperature interval: 30–180 °C; heating rate of 3 °C min⁻¹; 20 min isotherm at 180 °C)

Thermal analysis of the MgCl₂–MeOH system

The experimental alcohol/salt ratios of the MgCl₂ methanolates prepared, as listed in Table 1, varied significantly from the theoretical ones. Methanolates of lower methanolation states were formed when the absolute MeOH was added in excess under the conditions studied. According to the obtained experimental data, about 1, 2, 3, and 4 MeOH molecules were sorbed by one molecule of MgCl₂ (Table 2). Samples 13 and 14 were deliquescent and therefore we presume that both physically and chemically attached MeOH was present. Depending on the amount of MeOH, the dissociation reaction proceeded in 3–4 incomplete steps (Fig. 5) over the temperature region of 33–177 °C with peak temperatures of 65–94 °C, 76–103 °C, 120–136 °C, and 158 °C, respectively (Table 2). Trace amounts of MeOH continued to evolve during isothermal stabilization at 180 °C until the demethanolation went to completion. As plotted in Fig. 5, the mass loss rates are in general lower compared with that of the CaCl₂–MeOH and CaCl₂–EtOH solvates. The DSC peak integration yielded standard reaction enthalpy values of 69–183 kJ mol⁻¹ (Table 2). The obtained values were raised by increasing the alcohol/salt molar ratio. Recently, MgCl₂ monomethanolates have only been reported by Iyimen-Schwarz [51]. Furthermore, MgCl₂–MeOH species with stoichiometric numbers of coordinated MeOH molecules of 2 [42, 51, 56], 3 [51, 57], 4 [28, 42, 56, 58, 59], and 6 [47, 49, 56, 57, 60,61,62,63,64,65,66] have been identified in the literature.

Fig. 5

Mass (TGA) and mass loss rate (DTG) curves of the dealcoholation of MgCl₂–MeOH solvates prepared in neat at different MeOH/MgCl₂ molar ratios (Table 1), (measurement temperature interval: 30–180 °C; heating rate of 3 °C min⁻¹; 20 min isotherm at 180 °C)

Thermal analysis of the MgCl₂–EtOH system

MgCl₂ formed ethanolates of different stoichiometry: stoichiometric ratios of 1.9, 2.3, 3.5, and 4.6 were identified under the experimental conditions studied (Table 2). The latter MgCl₂–4.6EtOH is doubtable as the maximum quantity of absolute EtOH added to anhydrous MgCl₂ was 4. The dissociation reaction occurred in 2–4 steps (Fig. 6) between 33 and 177 °C with peak temperatures at 61–67 °C, 101–140 °C, and 136–153 °C (Table 2). Trace amounts of EtOH were desorbed during the subsequent isothermal stabilization segment until the reaction went to completion, as observed for MgCl₂ methanolates (Fig. 5). The EtOH released in a stepwise manner was probably uncoordinated and coordinated EtOH. The samples 17 and 18 consisted of a two phase mixtures: liquid and solid phase. The latter was used for characterization studies. Mass loss rates between 0.2 and 0.4 mg min⁻¹ were measured (Fig. 6). The calculated standard reaction enthalpies amounted to 65–158 kJ mol⁻¹ (Table 2); the standard reaction enthalpies per mole EtOH were surprisingly low (30–45 kJ mol⁻¹), partly lower than the standard enthalpy of vaporization of EtOH (41.680 kJ mol⁻¹) [67]. MgCl₂–EtOH solvates have gained considerable interest among researchers for the synthesis of MgCl₂-supported Ziegler–Natta catalysts and subsequently a plethora of the literature exists. Besides well-defined compounds, addition compounds of non-integer stoichiometry have also been identified in several studies. Iyimen-Schwarz [51], Tewell et al. [68], Chadwick and Severn [69], and Bart and Roovers [70] have prepared MgCl₂ ethanolates with a molar ratio of 0.47, 1, 1.1, and 1.25. The existence of MgCl₂ diethanolates has been proven by Multani [56] and Di Noto et al. [71]. Furthermore, EtOH solvates of MgCl₂ with molar ratios of 1.5 [72], 1.67 [68, 70], 2.05 and 2.1 [68, 69], 2.5 [70, 73], 2.8 [69, 72], and 3.33 [28, 70, 72] have been synthesized and investigated in detail. Complex compounds with higher stoichiometry of 4, 4.5, and 5 have been obtained by Bart and Roovers [70], Multani [56], and Tewell et al. [68]. The maximum coordination number observed in MgCl₂–EtOH complexes has been 6 according to various authors [56, 57, 63, 65,66,67, 70,71,72, 74, 75].

Fig. 6

Mass (TGA) and mass loss rate (DTG) curves of the dealcoholation of MgCl₂–EtOH solvates prepared in neat at different EtOH/MgCl₂ molar ratios (Table 1), (measurement temperature interval: 30–180 °C; heating rate of 3 °C min⁻¹; 20 min isotherm at 180 °C)

Comparing the decompositions temperatures of CaCl₂ alcohol solvates indicates that obviously MeOH was stronger associated with the metal chloride than EtOH. For instance, the EtOH evolution from CaCl₂–EtOH solvates started at lower temperatures than the desorption of MeOH from CaCl₂–MeOH solvates; the last MeOH molecule was liberated at higher temperatures than the last EtOH molecule. This means that during thermal cycling, EtOH can be liberated with lower energy than MeOH. Analogous reaction behavior was observed for MgCl₂ alcohol solvates. Moreover, MgCl₂–alcohol solvates showed lower dealcoholation rates and thus slower alcoholation/dealcoholation reactions are assumed in comparison with CaCl₂ alcohol solvates.

Energy analysis

Variation in the coordination numbers of the salt–alcohol solvates cited in the references and this study could be the result of the diversity in the synthesis procedures and conditions. According to Bart and Roovers [70], the synthesis method plays a crucial role with regard to the nature of the reactants. Salt–alcohol solvates can be prepared by direct synthesis, solution crystallization and elimination of excess solvent, or recrystallization, for example. The alcohol coordinates to the alkaline earth metal cations to form complexes with variable stoichiometry and structural properties that exist either in the solid state or in solution.

Note that the calculated enthalpies of reaction and dissociation (Table 2) tend to be inaccurate as no quantitative differentiation between alcohol molecules and possible H₂O molecules coordinated to one molecule of salt could be made at this stage. Samples contained probably trace amounts of H₂O that were neglected, since the freshly prepared samples were immediately characterized via TGA/DSC.

A linear regression analysis of the experimentally obtained standard enthalpies of dissociation as a function of the number of alcohol molecules revealed that the applied model fitted well the data (Fig. 7). The coefficient of determination was close to one for each reaction system. Evidently, the standard enthalpy of dissociation increased along the homologous series of the alcohols studied. EtOH solvates possessed higher enthalpies of formation compared to MeOH solvates. The Ca²⁺ ion is of higher charge than the Mg²⁺ ion and has a greater polarizing power. Therefore, the standard enthalpies of formation of CaCl₂–alcohol complexes are comparatively higher.

Fig. 7

Experimentally obtained standard enthalpies of dissociation of various salt–alcohol solvates plotted against the number of alcohol molecules evolved during the dissociation reaction, MeOH methanol, EtOH ethanol. Linear regression analysis was applied to determine the coefficient of determination (R-squared)

The best fit (R² = 0.9981) was achieved with the CaCl₂–EtOH reaction system confirming its chemical stability. The calculated enthalpies of dissociation for all CaCl₂–2EtOH solvates analyzed were in the range of 1303–1394 kJ mol⁻¹ (Table 2) and coincided with formation enthalpy and dissociation enthalpy data given in the published literature [27, 50] (Table 3). CaCl₂–EtOH complexes of higher ethanolated states as reported by Parker et al. [53] and those of lower ethanolated states [50, 51] could not be synthesized under the experimental conditions applied.

Table 3 Data available in the literature on experimental and theoretical standard enthalpies of reaction per mole salt alcoholate Δ_rH° and standard enthalpies of formation/dissociation Δ_fH° of CaCl₂–MeOH and CaCl₂–EtOH complexes of different stoichiometry z

The CaCl₂–MeOH system exhibited an acceptable fit with a coefficient of determination of R² = 0.9967. The enthalpies of dissociation of the CaCl₂–2MeOH solvates with values of 1314 kJ mol⁻¹ and 1337 kJ mol⁻¹, respectively, (Table 2) were slightly higher than literature values [27, 38, 42, 51] (Table 3). Enthalpies of reaction and formation/dissociation for CaCl₂–3MeOH and CaCl₂–4MeOH have not been determined yet. The enthalpy of dissociation of CaCl₂–3MeOH and of CaCl₂–4MeOH was estimated at 1520 kJ mol⁻¹ and 1771 kJ mol⁻¹/1754 kJ mol⁻¹, respectively. The enthalpy of dissociation of CaCl₂–MeOH complexes increased with the number of alcohol molecules indicating a linear relationship. However, a nonlinear relationship could be observed when the standard enthalpies of reaction per mole of alcohol $\frac{{\Delta_{\text{r}} H^{0} }}{z}$ were compared (Table 2). At levels of methanolation higher than 2, the enthalpies of reaction per mole MeOH varied between 31 and 38 kJ mol⁻¹ (Table 2), which are close to the standard enthalpy of vaporization of MeOH (37.965 kJ mol⁻¹) [67]. Apparently, up to 2 molecules could be chemically bound as a ligand. In MeOH solvates of CaCl₂ with higher molar ratios, the inclusion of MeOH molecules or physical incorporation of molecules into the crystal lattice could have caused the decrease in the standard enthalpy of reaction per mole of MeOH.

A linear relation between the enthalpy of dissociation and the number of alcohol molecules was also found for MgCl₂–EtOH and MgCl₂–MeOH solvate complexes (Fig. 7). The R-squared values of the MgCl₂–EtOH and MgCl₂–MeOH systems were R² = 0.9979 and R² = 0.9971, respectively. Reference data are rarely available. Iyimen-Schwarz [51] has collected the enthalpies of formation and dissociation of MgCl₂–1EtOH, MgCl₂–1MeOH, and MgCl₂–3MeOH solvates from thermal cycling tests by DSC technique (Table 4). When compared with values determined in this study, a significant deviation appeared. The values given by Iyimen-Schwarz [51] are mean values averaged over 13 cycles for MgCl₂–MeOH solvates and 10 cycles for MgCl₂–EtOH solvates. Maximum formation and dissociation enthalpy values of 31 kJ mol⁻¹/− 34 kJ mol⁻¹ and 111 kJ mol⁻¹/− 103 kJ mol⁻¹ were derived for MgCl₂–1MeOH and MgCl₂–3MeOH. In the present study, the enthalpies of dissociation of MgCl₂–1MeOH, MgCl₂–2MeOH, MgCl₂–3MeOH, and MgCl₂–4MeOH ranged from 891 to 1639 kJ mol⁻¹ (Table 2). The calculated enthalpies of reaction per mole MeOH of 43–77 kJ mol⁻¹ (Table 2) were higher than the standard enthalpy of vaporization of MeOH. However, the enthalpies of reaction per mole EtOH with values of 30–45 kJ mol⁻¹ (Table 2) were lower in comparison with the enthalpy of vaporization of EtOH (42 kJ mol⁻¹) [67]. Iyimen-Schwarz [51] has obtained even lower enthalpies of formation and dissociation and, respectively, enthalpies of reaction per mole of EtOH for MgCl₂–EtOH solvates (Table 3). These inconsistent and unreliable experimental results are ascribed to inherent instability issues of the MgCl₂–alcohol solvates.

Table 4 Data available in the literature on experimental and theoretical standard enthalpies of reaction per mole salt alcoholate Δ_rH° and standard enthalpies of formation/dissociation Δ_fH° of MgCl₂–MeOH and MgCl₂–EtOH complexes of different stoichiometry z

To assess the suitability of the various salt–alcohol solvate systems for low-temperature heat storage, the gravimetric energy density was calculated from the experimentally measured standard enthalpies of reaction. The gravimetric energy density is a thermochemical material characteristic. As a key performance metric, it is used to evaluate and compare the energy storage performance of thermal energy storage systems. The gravimetric energy density of the investigated salt–alcohol systems was similar, irrespective of the reaction pair combination and ranged from 354 to 943 kJ kg⁻¹, respectively. From the energetic point of view, salt–alcohol solvate systems for heat storage are not as good as salt–water solvate systems, as the measured standard enthalpies of reaction and associated gravimetric energy densities are roughly half the respective values of the latter ones [76].

Raman spectroscopic characterization of selected samples

In a starting series of experiments, we increased the laser energy systematically from 200 mW to 500 mW and we found that spectra of sample 6 were not influenced by the change of the excitation energy. This observation let us to conclude that the alcohol solvate sample was stable enough to be characterized by Raman spectroscopy. Raman spectra of several CaCl₂–alcohol solvates recorded at room temperature can be seen in Fig. 8. Comparing the spectrum of pure CaCl₂, (line d in Fig. 8a) to those of the different alcohol solvates (lines a, b, e, f in Fig. 8), it could be suggested that the very low wavenumber region was characteristic for CaCl₂ and bands from ~ 450 to ~ 3100 cm⁻¹ belonged to the alcohols. Although Raman spectroscopy is not actually sensitive to H₂O, a very weak band at about 3400 cm⁻¹ indicated that CaCl₂ contained a certain amount of coordinated H₂O (line e in Fig. 8b). As a comparison, the Raman spectrum of moisture exposed CaCl₂ is also shown (line h in Fig. 8b).

Fig. 8

Raman spectra of alcohol solvates of CaCl₂. a MeOH solvates; b EtOH solvates; a: sample 1; b: sample 4; c: MeOH; d: CaCl₂; e: sample 6; f: sample 9; g: EtOH; h: moisture exposed CaCl₂

The MeOH in sample 1 (line a in Fig. 8a) was mainly in built-in form as ν_as of CH₃ at 2952 cm⁻¹ and ν_s of CH₃ at 2844 cm⁻¹ and was shifted with ~ 10 cm⁻¹ comparing to those of the MeOH (2942 cm⁻¹ and 2833 cm⁻¹, line c in Fig. 8a). Bands at 3013 cm⁻¹ and 2982 cm⁻¹ in sample 1 indicated the presence of coordinated OCH₃ (line a in Fig. 8a), while the band at ~ 3400 cm⁻¹ indicated the presence of certain built-in H₂O (cf. line a and line d in Fig. 9a) that was somewhat more intensive than in the starting salt. As the Raman technique gives average information about the material, we could not decide, whether the sample consisted of the mixture of non-coordinated, OH-coordinated, and MeOH-coordinated units of the CaCl₂ or units, which had MeOH and H₂O in the same coordination sphere.

Fig. 9

Raman spectra of MeOH solvates of MgCl₂. a overview spectra. b enlargement of CH₃ region; a: sample 12; b: sample 13; c: MeOH; d: MgCl₂; e: sample 10; f: sample 12; g: sample 13; h: MeOH

Contrary to sample 1, sample 4 contained a large amount of free MeOH besides the coordinated MeOH witnessed by the doublet of ν_s CH₃ band at 2844 cm⁻¹ and 2836 cm⁻¹; the latter belongs to non-coordinated MeOH. In addition, a double or triple band in the region ~ 3300 cm⁻¹ indicated that –OH and non-coordinated MeOH existed in this sample (cf. lines b and c in Fig. 8a). The appearance of a significant amount of non-coordinated MeOH in sample 4 (prepared at an alcohol/salt ratio of 6:1) confirmed that the maximum number of the coordinated MeOH could be 4, as it was already indicated previously (Sect. 3.1). It is worth noting that the Raman spectrum of sample 3 was very similar to that of sample 4, apart from the band of the non-coordinated MeOH, which was somewhat weaker in sample 3 than in sample 4.

Regarding the alcohol solvates obtained from CaCl₂ and EtOH, no real difference between the Raman spectra of the samples prepared at different alcohol/salt ratios could be found (cf. line e and f in Fig. 8b) in accordance with the findings described in Sect. 3.1. Certain relative intensity changes in 2973 cm⁻¹/ ~ 2878 cm⁻¹ of ν_asCH₃/ν_sCH₃ in the region of EtOH (cf. lines e, f and g in Fig. 8b) indicated the presence of the EtOH incorporated into the crystal lattice. The shift of the νC-O bands at 1096 cm⁻¹ and 1051 cm⁻¹ to 1088 cm⁻¹ and 1048 cm⁻¹, respectively, could support this idea. However, the appearance of a weak band at 3396 cm⁻¹ confirmed again the presence of coordinated H₂O.

Raman spectra of several MgCl₂–MeOH solvates are depicted in Fig. 9. Samples in these series were very similar to each other. The splitting of the νC–O band at 1033 cm⁻¹ indicated MeOH incorporation into the crystal structure. The intensity ratios of MeOH bands/Mg–Cl bands increased with the amount of introduced MeOH. The ν_as of CH₃ appeared at 2950 cm⁻¹ and ν_s of CH₃ appeared at 2848 cm⁻¹ in the MgCl₂–MeOH solvates. These values were shifted compared to that of free MeOH (cf. lines a, b, c and d in Fig. 9b), which indicated that MeOH definitely existed in built-in form in the samples 10, 12, and 13.

Regarding the MgCl₂–EtOH systems (Fig. 10), the spectra resembled the spectrum of EtOH showing only negligible band shifts; it was difficult to decide, whether the EtOH existed in a real built-in form or not (cf. line a and c in Fig. 10a). Surprisingly, a new C-H stretching band at high wavenumber appeared (better visualized for sample 18, and only as a shoulder for sample 15) that might be assigned to νC-H of halogen substituted methyl group. As a comparison, the library IR spectrum of ethyl chloride is also shown (Fig. 10b). It seems plausible that ethyl chloride was formed in MgCl₂–EtOH systems. Simultaneously, the Ca–Cl vibrational band at 241 cm⁻¹ almost completely disappeared (cf. line b in Fig. 10a).

Fig. 10

Raman spectra of MgCl₂–EtOH solvates. a EtOH solvates; b reference FT-IR spectrum of ethyl chloride from Aldrich Vapor FT-IR Spectral Library; a: sample 15; b: sample 18; c: EtOH; d: MgCl₂

As a conclusion, MgCl₂–EtOH solvates underwent certain decomposition resulting in ethyl chloride formation. Pure salt–alcohol solvates could not be obtained, because of the presence of a certain amount of salt hydrates.

TG-MS behavior of selected samples

As preliminary investigations, the starting materials, i.e., the CaCl₂ and MgCl_2, were analyzed, and it was found that both of them contained roughly around 2% H₂O of crystallization (more precisely, CaCl₂ contained 2.28% H₂O, while the H₂O content of MgCl₂ was 2.36%).

Four samples were chosen for the TG-MS measurements, namely 3 (containing CaCl₂ and MeOH), 7 (which contain CaCl₂ and EtOH), 12 (containing MgCl₂ and MeOH), and finally 18 (composed of MgCl₂ and EtOH). In Fig. 11, the mass loss and some selected ion currents are plotted against the temperature obtained from sample 3, while on the right side of the figure the chemical species corresponding to the chosen m/z values are also shown. On the TG curve, between 55 and 300.5 °C, three mass loss steps can be seen. The first mass loss step is between 55 and 123.5 °C, with a mass loss of 23.3%, the second step is between 123.5 and 178 °C, with 19.2% mass lost, and the last step between 178 and 300.5 °C resulting in 9.5% mass loss; the total mass loss during the measurement was 52.9%. Comparing the shape of the three ion currents, three overlapping peaks can be observed on each curve, which correspond to the appropriate mass loss step. The m/z 32 is the molecular ion of MeOH, the m/z 31 is the base peak of MeOH (and also a characteristic marker of aliphatic alcohols), while the m/z 15 corresponds to the methyl ion (CH₃). From the TG-MS measurement, it can be concluded that the evolution of MeOH starts at very low temperatures (around 30 °C). Some H₂O was also detected during the measurement (the ion current of H₂O is not shown), but its concentration variation was within one order of magnitude compared to the concentration variation of MeOH, which was 2 orders of magnitude (m/z 31), thus confirming that the major volatile component formed during the measurement was MeOH. Moreover, no chlorinated compounds (e.g., methyl chloride) were detected. It can be seen that above 300 °C all three ion current curves are still decreasing, which implies that some MeOH is still lost, but this causes a very small mass loss, not detected by TG.

Fig. 11

TG-MS trace of the sample 3 (mass loss on the upper part, while ion currents on the lower part of the graph), on the right side the formulae of the corresponding fragments/ions

Considering the TG-MS trace of the sample 7 (Fig. 12), only two mass loss steps can be distinguished. The first mass loss step is between 55 and 95 °C, with a mass loss of 1.9%, while the second step is between 95 and 241 °C, with a mass loss of 48%; the total mass loss during the measurement was 50.3%. It can be seen that the formation of EtOH vapors (m/z 46 is the molecular ion of EtOH, m/z 45 is a deprotonated EtOH, while m/z 31 is the base peak of EtOH) begins at very low temperatures, as already observed in the case of sample 3.

Fig. 12

TG-MS trace of the sample 7, on the right side the formulae of the corresponding fragments/ions

A very small amount of H₂O can be detected between 152 and 235 °C, which means that the H₂O is much more strongly bound to the CaCl₂ than the EtOH. No alkyl chloride was detected proving the thermal stability of the CaCl₂–alcohol solvate. Comparing the results of sample 3 and 7, it can be concluded that EtOH (sample 7) is lost in a narrower temperature range than MeOH (sample 3). Moreover, the temperature value, where the alcohol is practically lost decreases from 300.5 °C (sample 3) to 240 °C (sample 7). This means that during cyclic measurements, EtOH can be liberated with less energy than MeOH.

Replacing CaCl₂ with MgCl₂, significant differences appear in the TG-MS trace of sample 12 and 18. On the mass loss curve of the sample 12 (Fig. 13), three mass loss steps can be identified. The first larger step is between 55 and 198 °C, with a mass loss of 40%, the second, smaller mass loss is between 198 and 306 °C (10% mass lost), and the last, very small step is from 306 °C up to the end of the measurement (up to 500 °C, 1.2% mass loss). Evaluating the ion current curves of four masses (m/z 32 molecular ion of MeOH, m/z 31 base peak of MeOH, while m/z 52 is the molecular ion of methyl chloride, with the ³⁷Cl isotope, and m/z 50 is the molecular ion of methyl chloride, with the ³⁵Cl isotope), it can be seen that MeOH is liberated below 180 °C. Above this temperature, besides MeOH, which is still the major component of the volatiles and some traces of H₂O, the formation of methyl chloride begins. This confirms that at higher temperatures, MgCl₂ hydrolyzes and converts the MeOH into methyl chloride. Around the upper end of the temperature scale, a small amount of methyl chloride is still released between 410 and 490 °C (0.7% mass is lost).

Fig. 13

TG-MS trace of the sample 12, on the right side the formulae of the corresponding fragments/ions

Replacing the MeOH in the magnesium salt with EtOH (sample 18, Fig. 14), again a complicated, multi-step decomposition pattern is obtained. On the mass loss curve of sample 18, four decomposition steps can be identified (first step between 50 and 100 °C, mass loss 4.2%; second, larger step between 100 and 170 °C, mass loss 46.3%, third step between 170 and 212 °C, mass loss 13.3% and fourth step between 212 and 288 °C, mass loss 10.3%). It can be seen that below 140 °C, the mass loss is caused mainly by the evaporation of EtOH (m/z 46 is the molecular ion of the EtOH, m/z 45 is a deprotonated EtOH, while m/z 31 is the base peak of EtOH) and some traces of H₂O (ion curves not shown). Above 140 °C, the formation of ethyl chlorides (m/z 66 is the molecular ion of ethyl chloride, with the ³⁷Cl isotope, while m/z 64 is the molecular ion of ethyl chloride, with the ³⁵Cl isotope) can be detected. Ethyl chloride is formed over a broad temperature range (between 140 and 370 °C), additionally the concentration of ethanol decreases above 290 °C, while the ethyl chloride is still formed. The formation of ethyl chloride confirms the fact that MgCl₂ hydrolyzes and converts the EtOH into the corresponding alkyl chloride. Comparing the TG-MS results of sample 12 and 18, it can be seen that the alkyl chloride formation starts at lower temperatures in the case of sample 18 (140 °C). The end temperature value of the mass loss end is shifted toward lower temperatures (mass loss end temperature of sample 12 is 306 °C, while the corresponding value is 288 °C in the case of sample 18).

Fig. 14

TG-MS trace of the sample 18, on the right side the formulae of the corresponding fragments/ions

Temperature programmed XRD characterization of selected samples

Figure 15 shows the XRD patterns of four selected salt–alcohol solvate samples recorded by use of stepwise heating. The temperature steps were also chosen according to the TG curve obtained from the TG-MS measurement.

Fig. 15

XRD patterns of selected alcohol solvates at different temperatures (temperature steps adjusted according to TG curve obtained from the TG-MS measurement; heating rate 5 °C min⁻¹; atmosphere: dry N₂). a MeOH–CaCl₂ (sample 3); b EtOH–CaCl₂ (sample 7); c MeOH–MgCl₂ (sample 12); d EtOH/MgCl₂ (sample 18)

Sample 3 in its starting form (Fig. 15) consisted of MeOH-coordinated phases. The XRD diffraction pattern was somewhat more complicated at 55 °C than at 25 °C, which indicated the formation of a more complex system by the mild heating. These samples consisted of certain CaCl₂xMeOH (H₂O) phases. The most likely composition was the average CaCl₂/MeOH/H₂O = 1:1:3 ratio in the coordination sphere, which could be estimated from TG. Either a mixture of 1:4 and 1:2 coordinated samples of CaCl₂–H₂O and CaCl₂–MeOH or samples with mixed coordination spheres could be imagined. At 120 °C, 177 °C, and 301 °C samples were CaCl₂-like materials having amorphous part. At 500 °C, only the CaCl₂ phase existed. The sample after re-cooling was CaCl₂ with some Ca-hydroxide-carbonate. The XRD patterns at 120 °C, 177 °C, and 301 °C were very similar in spite of the large amount of MeOH that went away in the above temperature steps, which implies that a significant amount of MeOH was not really bound in the coordination sphere of the salt.

The diffraction pattern of sample 7 (Fig. 15b) did not change significantly during the heating up to 53 °C in accordance with the TG-MS results, which shows that no significant amount of EtOH or H₂O was removed during the first some minutes. The starting solvate formed at 1:3 molar ratio could be imagined as a mixture of crystals with coordination sphere 1:2 and 1:4 or crystals with coordination sphere 1:2 and alcohol inclusions. The situation was complicated with the presence of H₂O. Consequently, besides the pure CaCl₂·zEtOH and pure CaCl₂·zH₂O crystals, CaCl₂·(x–y)EtOH·yH₂O structures could exist. Based on the XRD results it was impossible to give a more detailed description. Pure CaCl₂ phase could also be found in the region of 55–500 °C. Probably easily decomposable alcoholic, a relatively stable alcoholic and hydrous phases existed parallel in this sample. The final state was alcohol- and H₂O-free CaCl₂.

The composition of the initial sample 12 was rather complicated (Fig. 15c). A volatile part was removed at 56 °C that resulted in a slight change in the diffraction pattern. One might think that physisorbed solvent was removed during the heating from 25 up to 56 °C and the solid consisted of mainly several H₂O-coordinated MgCl₂ and MeOH-coordinated MgCl₂ forms. Regarding the starting solvate, the TG-MS indicated parallel removing of MeOH and a significant amount of H₂O. The TG-MS gave a MgCl₂/MeOH ratio of 3.12 only instead of 4. Since the molar mass of H₂O is smaller than that of MeOH, the difference might come from the presence of structures with coordination number 4, but both ligands were in the coordination sphere in different variation. Consequently, the starting solvate itself had to be a mixture. The slight changes in the XRD pattern of sample 12 at 56 °C compared to that at 25 °C resulted from the small variation of the individual components. At 200 °C, the broadenings in the XRD patterns indicated the formation of amorphous phases. A less crystalline form observed was supposed to be a mixture of MeOH- and H₂O-coordinated materials. It is conceivable that MeOH and H₂O bound to the salt more or less equally strongly. Some small peaks at about 2θ ~ 32° and 2θ ~ 50° decreased upon increasing the temperature, but we can only say that probably the decomposition of 1:2 complexes appeared. At 307 °C either H₂O or MeOH could not be present according to TG-MS. The XRD pattern shows the presence of a certain amount of MgCl₂ and/or MgOCl besides amorphous phase(s). At 500 °C, the final material was more crystallized; the final state was MgO (with certain hydrate) and probable Mg-hydroxy-hydrate. The presence of certain MgOOH·HCl could be supposed, too. From these results, it is obvious that the MgCl₂ decomposed upon increasing the temperature from 307 to 500 °C. When the sample was cooled to 25 °C, the XRD pattern (and of course the structure) formed at 500 °C was maintained.

The XRD patterns of sample 18 (Fig. 15d) at low temperatures (25 °C, 53 °C) were very similar to those of MgCl₂·zH₂O. However, TG-MS had shown the removal of mainly EtOH (fragments of EtOH). This observation excluded that the starting material was merely MgCl₂.zH₂O. We had to assume that MgCl₂.zH₂O is isostructural with MgCl₂·zEtOH. Only a little shift of the peaks appeared in comparison with MgCl₂.zH₂O, which means that the point group was the same in both cases. Because the shift was small, the cell parameters were not altering significantly. This was very surprising considering the difference in the steric properties of H₂O and EtOH. The very small change in the XRD pattern at 53 °C compared to the XRD pattern at 25 °C could be explained by the removal of a small amount of non-chemisorbed alcohol (physisorbed EtOH or inclusion of EtOH). In the temperature region of 101–306 °C, the sample gradually lost the EtOH. According to TG-MS results, only EtOH was removed up to 171 °C; EtOH and a small amount of H₂O went away between 171 and 306 °C. Nevertheless, ethyl chloride was also detected by TG-MS indicating a decomposition process. Increasing the temperature from 101 °C up to 212 °C resulted in a change in the XRD patterns with gradual formation of amorphous phases. MgOHCl, MgCl₂ (aq) probably existed at higher temperatures. The XRD pattern obtained at 500 °C indicated a cubic crystalline material. The pattern of MgO could be fitted well, but the coexistence of other cubic components could also be suggested. The “temperature programmed” XRD behavior of sample 18 was very similar to that of sample 12.

Results of XRD measurements led us to the following conclusions. Mixtures of different phases containing alcohol and H₂O existed in all of our salt–alcohol solvate samples. CaCl₂ was retrievable from its alcohol solvate by bake out of the alcohol solvate, but irreversible processes appeared in case of MgCl₂–alcohol solvates. Upon increasing the temperature up to 500 °C, MgO was obtained with elimination of hydrogen chloride similarly to the MgCl₂·6H₂O·1,4-C₄H₈O₂ [77] and MgCl₂·6H₂O system. Although the transformation of the MgCl₂ hydrates has long been known [78], clarifying of its mechanism is still in the focus of interest [79].

Possible problems of the applications of salt–alcohol solvates in heat storage systems, based on the results of various techniques used

Our results revealed that the MgCl₂–EtOH solvates were instable compound; decomposition of these samples during storage, handling, and analysis was assumed. Raman spectroscopic measurements proved the appearance of ethyl chloride from MgCl₂–EtOH without heating. Upon heating, both MeOH and EtOH solvates of MgCl₂ were involved in alkyl chloride release, as proven by TG-MS measurements. The alkyl chloride ion current increased slightly from the beginning of the heating and intensified above 140 °C. Although the heating rate can influence on the mechanism of the decomposition, TG-MS measurements by use of high heating rate indicated the formation of alkyl chloride as well as Raman spectroscopy without any heating. On the other hand, XRD measurements showed transformation of MgCl₂ in all MgCl₂–alcohol solvate samples, which released alkyl chloride.

These observations were in accordance with the literature. Micro-calorimetric analysis has been conducted by Iyimen-Schwarz [51], who has made pioneering scientific contributions to the usability of salt–alcohol solvates based on CaCl₂/MgCl₂ and MeOH/EtOH for thermal energy storage. Iyimen-Schwarz [51] calculated the energy density from the forward and reverse reaction enthalpy determined in dynamic DSC measurements under vacuum at controlled temperatures and alcohol vapor pressures. Based on observations on both the shape of the MgCl₂–MeOH solvate’s measurement curve and the change of the physical appearance, Iyimen-Schwarz suspected the decomposition of the MgCl₂–MeOH solvate and the release of CH₃Cl with successive cycling, analogous to the reaction of MgCl₂ with H₂O, affecting the reproducibility of the measurements. According to his work, EtOH can be desorbed easily from the MgCl₂–ethanolate system. Iyimen-Schwarz further assumed that in this system the MgCl₂ is likely to decompose into C₂H₅Cl and HCl, due to unreproducible results of the measured dissociation enthalpy. Results of our TG-MS and Raman spectroscopic measurements fully support the conclusions of Iyimen-Schwarz and explain the poor cyclic stability of MgCl₂–EtOH solvates [51].

For alkyl chloride formation in our system different pathways could be assumed. It has been reported that gas phase reaction of EtOH and HCl with formation of H₂O over ZnCl₂/Al₂O₃ as a catalyst, is a suitable method for ethyl chloride preparation [80]. According to a possible explanation, the starting material (MgCl₂) contains MgCl₂·zH₂O in the presence of traces of H₂O. Heating the MgCl₂ hydrates results in MgOH_xCl_x with probable formation of hydrochloric acid (HCl). At the same time, reaction of alcohols and HCl can lead to the formation of methyl chloride (CH₃Cl), or ethyl chloride (C₂H₅Cl) via MgCl₂ acting as a catalyst. Deliberation of H₂O carries on the decomposition of MgCl₂. A direct interaction between the ligands (i.e., Cl⁻ and alcohol) in the coordination sphere of the Mg²⁺ can also be assumed. This idea does not need the H₂O to assist in the alkyl chloride formation, but it can lead to the gradual hydrolysis of MgCl₂. Although we do not have direct evidence for the first or second pathway, we believe that the direct way is more likely, since ethyl chloride formation appeared even at room temperature, while HCl release from MgCl₂·2H₂O was reported only at 167 °C [79]. It is worth to note that the final temperature in the cyclic stability test of MgCl₂·zEtOH was 180 °C [29].

The results presented in Sect. 3.3–3.5 demonstrate that pure salt–alcohol solvates could not be prepared under technically applicable conditions; the samples contained at least traces of H₂O. H₂O could be introduced by both the preparation procedure and by the starting materials, despite the use of commercial absolutized solvents and N₂ atmosphere for the preparation. The starting salts also contained 1–2% of H₂O, which resulted in the presence of salt hydrates besides the mixture of salt–alcohol solvates. The possibility of the formation of salt hydrates implies the possibility of HCl formation during the thermal treatment above 167 °C [81]. In practical applications, an inert atmosphere cannot be maintained during the whole process, so that traces of H₂O cannot be avoided. Additionally, the use of high purity grade H₂O-free substances might be too expensive at technical scale. The appearance of side reactions and the release of HCl with cycling can contribute to corrosion of the reactor components. The decomposition of MgCl₂–alcohol solvates results further in a degradation of the overall performance of the thermal energy storage system. In conclusion, the reaction system MgCl₂·R-OH is not suitable for practical implementation, due to its instability and irreversibility. From the energetic point of view, this compound is also not favorable as the measured enthalpies of reaction and associated energy densities are lower than that of CaCl₂–alcohol solvates and salt-H₂O systems [76].

Conclusions

Reversible chemical reactions are highly efficient in terms of storage volume, storage period, and sensible heat losses to the environment compared to other energy storage technologies. Different CaCl₂- and MgCl₂–alcohol solvates (EtOH, MeOH) were synthesized and their suitability for heat storage was examined by employing combined thermogravimetric analysis and differential scanning calorimetry (TGA/DSC), spectrometric and spectroscopic analysis (TG-MS, Raman) methods as well as by using X-ray diffraction (XRD). Due to their chemical nature, the CaCl₂–EtOH systems exhibited lower energy densities than CaCl₂-H₂O systems. Decomposition of MgCl₂–EtOH solvates accompanied by ethyl chloride formation started already during storage. Upon heating, both MeOH and EtOH solvates of MgCl₂ were affected by alkyl chloride release, as proven by TG-MS measurements. Our results fully support the assumptions of Iyimen-Schwarz [51] and explain the poor cycle stability of MgCl₂–EtOH solvates reported previously. We also demonstrated that pure salt–alcohol solvates cannot be prepared under technically applicable conditions. Formation of salt hydrates implies the possibility of HCl formation during the thermal treatment. Appearance of side reactions and possible release of HCl with cycling may be conducive to corrosion. Conclusively, MgCl₂–alcohol solvate systems are not recommended for heat storage, whereas CaCl₂–alcohol systems are suitable demonstrating stable cycle performance. The information on the decomposition pattern and associated changes in the structural integrity of the salt alcoholates obtained in this study are essential for selecting and designing efficient thermochemical energy stores. The fundamental data on the heat content of the parent salt provide the basis for the development of new two-component thermochemical materials with advanced properties.

Appendices

Appendix 1: Calculations

The alcohol/salt molar ratios (levels of alcoholation) were derived from experimental data obtained by TGA. The percentage mass loss X is defined as the mass of alcohol m_R-OH desorbed per unit mass of salt alcoholate m_{MX ZR-OH}:

$$X = \frac{{m_{\text{R-OH}} }}{{m_{\text{MX zR - OH}} }}100\%$$

(3)

The level of alcoholation z is defined as the ratio of the number of alcohol molecules evolved during the endothermic dissociation reaction n_R-OH to the amount of anhydrous salt n_MX and was calculated from the following equation:

$$z = \frac{{n_{\text{R-OH}} }}{{n_{\text{MX}} }} = \frac{{m_{\text{R-OH}} M_{\text{MX}} }}{{m_{\text{MX}} M_{\text{R-OH}} }}$$

(4)

wherein m_MX and m_R-OH are the masses of the anhydrous salt and the alcohol desorbed, respectively.

The principle of thermochemical heat storage using reversible gas–solid reactions is based on the conversion of thermal energy into chemical energy required to break the chemical bonds of the reactants. At constant pressure, the amount of heat energy that must be supplied to induce the decomposition reaction equals the endothermic heat of reaction, also designated enthalpy of reaction. An infinitesimal change in the temperature results in a change of the enthalpy by ΔC_pdT. The enthalpy of reaction Δ_rH under non-standard condition, in case of a reaction temperature T₁ different from the standard state temperature T₀ can be estimated from the standard reaction enthalpies and heat capacities of the reactants using Kirchhoff’s law. The heat energy ΔQ is then expressed by:

$$\Delta Q = \Delta_{\text{r}} H(T_{1} ,p_{0} ) = \Delta_{\text{r}} H^{0} (T_{0} ,p_{0} ) + \int\limits_{{T_{0} }}^{{T_{1} }} {\Delta C_{\text{p}} {\text{d}}T + \Delta_{\text{Tr}} H}$$

(5)

where ΔrH⁰(T₀) is the standard enthalpy of reaction at standard state conditions T₀ = 298.15 K and p₀ = 1 bar, C_p is the constant-pressure heat capacity and Δ_TrH is the enthalpy of transformation.

The enthalpy of reaction was obtained from DSC measurement by peak area integration. The energy that is liberated or absorbed as heat during a chemical reaction as a result of a temperature difference ΔT can be quantitatively determined by heat flux DSC. This technique measures the thermally induced heat flux transferred between the sample and an inert reference that are connected by a low-resistance heat flow path. Thermocouples below the symmetrically positioned sample crucible and empty reference crucible detect and compare the temperature of the specimen to the temperature of the reference as a function of time under same conditions. The heat flux between sample and reference is proportional to the temperature difference:

$$\phi = E_{{({\text{T}})}} {\text{DSC}} = E_{{({\text{T}})}} \Delta T$$

(6)

wherein E_(T) and DSC are the calorimetric sensitivity and the measured DSC signal. Integration of the peak area under the baseline-subtracted DSC signal over time yields Δ_rH⁰:

$$\Delta_{\text{r}} H^{0} = \int_{{t_{1} }}^{{t_{2} }} {{\text{DSC}}\,{\text{d}}t} = \int_{{t_{1} }}^{{t_{2} }} {\frac{\phi }{{E_{{({\text{T}})}} }}{\text{d}}t}$$

(7)

According to Hess Law, the value of the standard reaction enthalpy of the forward and reverse reaction must be equal and the same applies for the standard enthalpy of association and standard enthalpy of dissociation. The standard enthalpy of dissociation is the inverse of the standard enthalpy of formation. In general, the standard enthalpy of reaction is calculated from the difference of the standard enthalpy of formation of the products and the standard enthalpy of formation of the reactants:

$$\Delta_{\text{r}} H^{0} = \sum {\Delta_{\text{f}} H_{\text{products}}^{0} } - \sum {\Delta_{\text{f}} H_{\text{reactants}}^{0} }$$

(8)

Solving Eq. (8) for $\Delta_{\text{f}} H_{\text{products}}^{0}$ gives the standard enthalpy of formation and hence the standard enthalpy of dissociation.

Different mathematical approaches are used to determine the specific energy storage density. Two types of energy densities are known: gravimetric and volumetric energy density. The gravimetric energy density E_m is defined as the capacity of heat energy stored at a defined temperature and pressure per unit mass of storage material and can be calculated from the ratio of the standard reaction enthalpy to the molar mass of the reactant M_{MX zR-OH}:

$$E_{{\text{m}}} = \frac{{\Delta _{{\text{r}}} H^{0} }}{{M_{{{\text{MX}}\,{\text{zR-OH}}}} }}$$

(9)

whereas the volumetric energy density E_V is related to the volume of storage material and is described as:

$$E_{{\text{V}}} = \frac{{\Delta _{{\text{r}}} H^{0} }}{{M_{{{\text{MX}}\,{\text{zR-OH}}}} }}\rho$$

(10)

The storage material volume is derived from the material’s mass and bulk density ρ. The higher the level of alcoholation, the lower is the density. The volumetric energy density is an important key energy storage metric for designing and operating storage systems. It is also preferred for performance comparison studies. Space can be a limiting factor for many practical applications. There are no data available for the mass and bulk densities of the salt–alcohol systems studied and hence only the gravimetric energy density could be calculated.

Appendix 2: Data processing—example of a TGA/DSC curve

Figure 16 displays the DSC measurement curve and corresponding TGA signal normalized to the sample mass for sample 6. To determine the heat content of the sample and the associated enthalpy of reaction, the area under the DSC peak was integrated. Similar curves were recorded for all other samples.

Fig. 16

Example of a recorded TGA and DSC curve. The TGA curve or change in mass (%) and DSC curve or heat flow (mW) are given as a function of time (min) and temperature (°C). The sample was heated from 30 to 180 °C at a heating rate of 3 K min⁻¹. Since the signs of the thermal processes were inverted due to software issues, the signed number of the underlying endothermic reaction is negative