Thermal Characterization of [C₂Im][NO₃] and Multivalent Nitrate Salts Mixtures

Abstract

Due to their intrinsic properties, the current applicability of ionic liquids is enormous. In particular, their use in electrochemistry is beyond question. Numerous studies on these compounds and their mixtures, especially with lithium salts, focus on their use as electrolytes for batteries and other energy storage devices. This includes thermal energy storage devices, where 4th generation ionic liquids and their derivatives show a huge potential. Nevertheless, considering the uneven availability of the raw materials, such as lithium, research has extended to mixtures of these compounds with other salts of different metals that are more abundant and widely distributed, such as magnesium or aluminum. This work presents a comprehensive thermal characterization, using differential scanning calorimetry and thermogravimetry, of the protic ionic liquid ethylimidazolium nitrate and its mixture with magnesium and aluminum nitrate salts at different concentrations. Additionally, a comparison between these results and previous studies of mixtures of this ionic liquid with lithium nitrate, as well as mixtures of the protic ionic liquid EAN with the same metal salts, was also performed. The results indicated that the salt addition tends to broaden and reduce crystallization and melting peaks, while the glass transition becomes more visible and shifts to higher temperatures with increasing salt concentration. This is due to the disorder generated by the rearrangement of ions in the polar domains, which erodes the hydrogen bond network of the protic ionic liquid. Nevertheless, the thermal stability of the blended samples does not change significantly compared to the bulk ionic liquid.

1. Introduction

In recent decades, pure Ionic Liquids (ILs) have received significant attention from scientific communities, primarily owing to their excellent physicochemical properties and vast potential. Their applicability spans various fields including electrochemistry [1], biochemistry [2], solvents [3], pharmaceuticals [4,5], CO₂ capture [6], and energy-related fields [7]. Despite extensive study, the complete characterization of these compounds remains ongoing, with numerous manuscripts concerning pure ILs and their mixtures submitted for publication every year. Although ILs are composed entirely of ions, coupling phenomena limit the number of ions that can participate in the electric transport process (ionic conduction). Among the most studied ILs, the subclass known as Protic Ionic Liquids (PILs) has emerged as a promising and environmentally friendly alternative for energy storage and harvesting purposes [8,9]. The main characteristic of these PILs is the proton transfer from the Bronsted acid and Bronsted base framework. This proton transfer occurs in the presence of a proton donor/acceptor, which is related to their ability to form hydrogen bonds.

In a recent paper by Matuszek et al. [10], dedicated to Charles Austen Angell who died in 2021 and was widely known for the “Angell plot”, which allows liquids to be classified by their fragility [11], the authors reviewed the evolution of this field of research in recent decades and explored the multitude of possible applications, particularly focusing on new or improved energy applications for 4th generation ionic liquids and their analogues. Ionic liquid analogues, developed in the last decade, are hybrid materials that possess both ionic and neutral components capable of interacting through hydrogen bonding (deep eutectic solvents) or Lewis acid–base interactions. These interactions give rise to liquid coordination complexes (LCCs), solvated ionic liquids, or concentrated salt–solvent mixtures. While ILs are known for their wide liquid range and high ionic conductivity, making them attractive for electrochemical applications, their ions are not electroactive for use in batteries. Therefore, ILs must be doped with metal salt, usually lithium salt, with similar anions to the pure IL, resulting in a new hybrid material. The new opportunities for the application of these hybrid ILs have expanded not only through their mixture with inorganic salts [12,13,14], but also their mixture with organic solvents [15] or with other ionic liquids [16,17]. However, these new studies raise concerns about how some characteristic properties of primitive ILs change in these new ionic compounds. For instance, decreasing ionic conductivity [13,18,19,20] and increasing toxicity [21] have been observed after doping the ILs with metal salts. In particular, previous studies by our research group, both experimental and theoretical [12,22], have analyzed the effect of the addition of lithium, calcium, magnesium or aluminum nitrate salts to the protic ionic liquids, ethylammonium nitrate (EAN) and butylammonium nitrate (BAN). These studies revealed that the main mechanism of salt solvation was the formation of ionic clusters in the polar nanodomains of the bulk mixture, resulting in the establishment of solid-like structures; that is, a pseudolattice of cation–anion complexes in these polar nanoregions of the IL gives rise to stabilized nanocrystals in solution. This leads to slower diffusion and lower ionic conductivity and higher viscosity, especially for cations with a larger surface charge density (Mg²⁺ and Al³⁺). Additionally, our findings indicate that water molecules are bonded to these highly charged cations, especially in aluminum mixtures, where the hydration water could not be removed using common purification techniques, contributing to additional changes in the polar nanoregions of these blended compounds [23].

This work tries to advance the analysis of the solvation processes involving salts of electrochemical interest in protic ionic liquids and their effect on thermal behavior, by Differential Scanning Calorimetry (DSC) and Thermogravimetry (TG). By examining the endothermic and exothermic peaks of DSC curves during heating and cooling, which are attributed to phase transitions like melting and crystallization, respectively, insights into the internal structure of the materials can be obtained. Additionally, the thermogravimetric curves allow an analysis of the thermal stability changes resulting from salt addition. The present study includes a comparative analysis of the thermal behavior of pure ethylimidazolium nitrate and its mixtures with lithium, magnesium and aluminum nitrate salts at different concentrations. The objective is to evaluate their potential applications in energy storage, encompassing both electrochemistry and thermal aspects, such as batteries, fuel cells, cryogenic energy storage, building thermal regulation, and solar water heating systems [10].

The structure of the paper is as follows: following this introduction, experimental details are provided and, in the subsequent section, we include the discussion of our results. Finally, the main conclusions of the study are summarized.

2. Materials and Methods

2.1. Chemicals

The selected ionic liquid, ethylimidazolium nitrate ([C₂Im][NO₃]), was purchased from IOLITEC, and the inorganic salts of electrochemical interest, namely LiNO₃, Mg(NO₃)₂·6H₂O and Al(NO₃)₃·9H₂O, were supplied by Sigma Aldrich, Panreac, and Scharlau, respectively.

Table 1. Identification of the chemicals used in this work.

Name	Molecular Weight (g·mol−1)	Structure	CAS Number	Purity
Ethyl Imidazolium Nitrate	159.14		[C2Im][NO3]   501,693—38—5	>0.98 a
Lithium Nitrate	68.946		LiNO3   7790—69—4	>0.999 a
Magnesium Nitrate Hexahydrate	256.41		Mg(NO3)2·6H2O   13,446—18—9	>0.98 a
Aluminum Nitrate Nonahydrate	375.13		Al(NO3)3·9H2O   7784—27—2	>0.98 a

^a Indicated by provider.

summarizes the main characteristics of the compounds used in this work. Solutions of ionic liquid and salts at different concentrations, in the molality range from 0.5 mol·kg⁻¹ up to the saturation limit (3, 2 and 1 mol·kg⁻¹ for lithium [24], magnesium and aluminum nitrate salt, respectively) [14,25], were prepared.

The [C₂Im][NO₃] sample was previously dried under vacuum for a minimum of 48 h. The residual amount of water was tested with a Karl Fischer titrator, and this content was shown to be less than 100 ppm. The different salt solutions were prepared by mixing both components with an ultrasound bath for 24 to 48 h. After homogeneous mixtures were obtained, all samples were dried under vacuum for at least 48 h.

2.2. Experimental Section

A DSC Q1000 (Waters-TA Instruments-New Castle-DE-USA) was used to determine the thermal transitions of pure IL and its mixtures with magnesium and aluminum salts. Sealed aluminum pans with a hole that allows water to escape when it evaporates were used. The experiments were performed under nitrogen atmosphere. The melting temperature of indium and zinc and the melting point of indium were used to perform the temperature and heat calibration, respectively. To distinguish between apparently identical peaks, four thermal ramps were applied to all samples, two in the cooling mode and two in the heating mode, with an isothermal step between them.

(a)

heating from 25 to 120 °C at 40 °C min⁻¹;

(b)

isothermal step at 120 °C for 45 min with the objective of removing impurities, the free water content and to erase the thermal history of the sample [25];

(c)

cooling from 120 to −85 °C at 10 °C min⁻¹;

(d)

heating from −85 to 100 °C at 10 °C min⁻¹;

(e)

cooling from 100 to −85 °C at 5 °C min⁻¹;

(f)

heating from −85 to 100 °C at 5 °C min⁻¹.

This methodology is widely described in previous papers [25,26]. The transition temperatures and associated heat were determined during the reheating and recooling steps.

A Mettler Toledo TGA/DSC1 analyzer operating under nitrogen atmosphere was used to analyze the thermal stability of the pure IL and mixtures. The mass samples ranged from 3 to 5 mg and were placed in an open platinum pan. Experiments were performed at temperatures of 100 to 800 °C, at 10 °C min⁻¹ and with a nitrogen purge gas flow of 20 cm³ min⁻¹. Each analysis was repeated three times. The Curie points of alumel, nickel and perkalloy were used for the temperature calibration, as is extensively explained in the previous literature [25].

3. Results

3.1. DSC Results

The thermal behavior of the pure IL, [C₂Im][NO₃], was previously analyzed [24]; this revealed a distinctive exothermic peak at −20 °C upon cooling, associated with the crystallization, and a corresponding peak upon heating, melting at 35 °C. This suggests a supercooled effect at 55 °C, as usually happens in these compounds. Figure 1 shows the DSC curves for the mixtures of [C₂Im][NO₃] with the selected multivalent salts at various concentrations, namely Mg(NO₃)₂ and Al(NO₃)₃, along with the DSC of the pure IL for comparison purposes. The transition temperatures, defined as the onset of every peak, and the enthalpies, defined as the area of the peak, of these mixtures are calculated and summarized in

Table 2. Temperature, molar enthalpy and molar entropy of melting (t_m, ∆_mH in kJ·mol⁻¹ and ∆_mS in J·mol⁻¹·K⁻¹), freezing temperature (t_f), glass transition temperature (t_g), cold crystallization temperature (t_cc), and solid–solid temperature (t_ss) obtained from DSC curves.

Compound	tm/°C	tf/°C	tcc/°C	tg (*)/°C	∆mH/J·g−1	∆mH/kJ·mol−1	∆mS/J·mol−1·K−1
[C2Im][NO3] [24]	35	−20			82.7	13.2	43.0
0.5 m IL + LiNO3 [24]	32	16			74.8	12.3	41.2
1.0 m IL + LiNO3 [24]	21	−3			51.3	8.7	28.9
2.0 m IL + LiNO3 [24]	4		−15	−65	18.6	3.6	11.5
3.0 m IL + LiNO3 [24]				−69
0.5 m IL + Mg(NO3)2	28		−34	−68	51.5	8.8	29.1
1.0 m IL + Mg(NO3)2				−65
2.0 m IL + Mg(NO3)2				−56
0.5 m IL + Al(NO3)3	24	−21	−44		41.2	7.3	24.7
1.0 m IL + Al(NO3)3	11		−4	−70	14.4	2.8	9.7

* Measured at the inflection point of the DSC curve.

. To facilitate a more comprehensive analysis, the results of the DSC curves of mixtures of IL + lithium salt, which were analyzed in a previous work [24], have also been included in Table 2.

A consistent trend towards amorphous behavior with increasing salt concentration is observed in all three IL + salt mixtures. This statement is supported by the following common findings:

• The peak of crystallization in the cooling ramps disappears for the blended samples, except for the lowest concentrations of lithium and aluminum mixtures. At intermediate doses of all the mixtures, an exothermic peak upon heating, known as cold crystallization (see further details), emerges; this is followed by the corresponding melting peak.
• The glass transition (t_g) is present in all saturated mixtures. Particularly in the DSC curves of IL + Mg(NO₃)₂ mixtures, the glass transition is evident in all studied concentrations, showing a tendency to increase with salt addition, as indicated in Table 2; this is consistent with findings reported in the literature [20,25]. We would like to highlight that this is likely not exclusive to mixtures with Mg, and may also occur in the rest of the salt mixtures. However, the t_g is not detected in the rest of the samples since it can appear beyond the lowest temperature range of the equipment used in this work (−80 °C) [27].
• Wider melting peaks shifted to lower temperatures are observed for IL + salt mixtures, which completely disappear at the highest concentrations of the mono and divalent salts mixtures, i.e., the 3 mol·kg⁻¹ IL + LiNO₃ mixture [24] and 1 mol·kg⁻¹ and 2 mol·kg⁻¹ IL + Mg(NO₃)₂ mixtures. Although the DSC curve of the saturated IL + Al(NO₃)₃ mixture shows clear exothermic and endothermic peaks upon heating (cold crystallization and melting process), the heat associated with these peaks is very small (Table 2). This, together with the appearance of the glass transition temperature at −70 °C, underlines the loss of crystallinity upon salt addition.

All these observations suggest that salt addition leads to an increase in the supercooled state of the sample, extending even until the end of the cooling ramp (−80 °C) for the most concentrated mixtures. This phenomenon, particularly crucial in thermal energy storage materials, as some previous papers have pointed out [28,29], is commonly observed in materials with slow crystallization rates from the liquid to the solid phase, such as ionic liquids, due to their strong ionic character and the molecular structure’s degree of freedom [28]. The liquid phase maintains its bulky conformation when cooling down and it becomes disordered, allowing for a variety of ion displacements. During the supercooled phase, the material stores excess thermal energy in a non-stable state, which can be released and evolved into a more stable and ordered state, resulting in an exothermic peak upon heating in cold crystallization; for the lowest and intermediate concentrated mixtures of IL + LiNO₃, the lowest concentrated mixture of IL + Mg(NO₃)₂ and for both mixtures of IL + Al(NO₃)₃, their ions stay in the disordered state at all temperatures. It is also important to highlight the differences between the two mixtures of IL and aluminum salt. While exothermic peaks are observed in both ramps in the case of the less concentrated mixture, with cooling at −21 °C and heating (small and wide) at −44 °C, in the case of the more concentrated mixture, only an exothermic peak is observed upon heating at −4 °C. That means that the less concentrated aluminum mixture cannot crystallize completely during cooling (at −21 °C), remaining a small part in the amorphous phase until the end of the cooling process and crystallizing in the heating ramp at −44 °C; meanwhile, the most concentrated sample was unable to crystallize upon cooling and the exothermic peak previous to the melting is only observed in the heating ramp. In this last sample, the glass transition is clearly visible at −70 °C; however, for the less concentrated sample, this transition was not observed, probably due to it being out of the studied temperature range and, in agreement with the previous literature, a shift being observed in the glass transition to higher temperatures with the concentration of salt added to the mixture [30].

These results are consistent with our previous findings analyzing the effect of adding monovalent, divalent and trivalent nitrate salts in the protic IL ethyl ammonium nitrate (EAN) [25], which complements the nanostructured solvation paradigm. In this paradigm, the amphiphilic nature of ILs leads to polar nanoregions (nitrate anions and head group of the cation) and an alkyl chain forming apolar domains. The added salt becomes incorporated into the polar region, forming stable solvation complexes [23]. The cation of the salt in the polar nanodomains can disrupt the hydrogen bond network of the PIL, reducing hydrogen bonding and inducing orientational disorder in the polar part [31]. The disorder is conditioned by the type of solvation complexes formed; thus, the Li⁺ forms tetrahedral kinetic stable complexes with four NO₃⁻ anions (Li(NO₃)₄)³⁻), whereas the aluminum cations bind water, even in vacuum-purified mixtures, and diffuse as a kinetic entity in an octahedral complex [23]. This aspect seems to be the reason for the lower solubility of aluminum nitrate in [C₂Im][NO₃]. As we have previously reported [12], the water content in multivalent salt mixtures is essential for achieving their solubility in ILs.

Beyond this common behavior in protic ILs, similar findings have been observed in aprotic ILs; Ngo et al. [32] reported the inhibition of crystallization in highly concentrated mixtures of [C₄Im][TFSI] with mono and divalent TFSI salts. They also noted a shift to higher glass transition temperatures as the salt concentration increased, attributing this to the significant size difference between salt and IL cations, which prevents efficient packing and leads to a homogeneous fluid. Additionally, these authors observed that the glass transition temperature decreases with an increasing salt cation size. This is in good agreement with our results for mixtures of 1 m of IL + Mg(NO₃)₂ and IL + Al(NO₃)₃, which showed glass transition temperatures at −65 °C and −70 °C, respectively. Ngo et al. [32] associated this phenomenon with the reduction in the lattice energy for the smaller ions, which reduces ion packing in the mixture and forms more asymmetric systems. Similarly, Girard et al. [20], studying trimethyl(isobutyl)phosphonium bis(fluorosulfonyl)imide and lithium bis(fluorosulfonyl)imide mixtures, and Montanino et al. [33], studying pyrrolidinium TFSI–LiTFSI mixtures, also found the vanishing of the crystallization exothermic peaks upon cooling, with a reduction in the heat of crystallization and melting as the salt concentration increases (with the total disappearance at the highest concentrations) and a glass transition that shifts to higher temperatures according to the salt concentration. This common behavior indicates that the IL + salt mixtures behave as pure glass-forming liquids with low glass transition temperatures. The interactions between the metal ions and the IL disrupt the crystallization, increasing the disorder and reducing the efficiency of ion packing, making the system less mobile.

The entropy of the melting point is calculated using Equation (1).

$${∆}_{m}S=\frac{{∆}_{m}H}{T_m}$$

(1)

where ${∆}_{m}H$ and $T_m$ are the enthalpy and temperature of melting, respectively. The results are also shown in Table 2, and the trend obtained with the salt concentration is also in very good concordance with that of EAN mixtures [25]. Both ILs showed similar tendencies with salt addition regarding the entropy and enthalpy, following the sequence Li⁺ > Mg²⁺ > Al³⁺. This trend is linked to the electrostatic potential of the salt anion [25].

An important aspect to highlight is that the molar melting enthalpy of both protic ILs, EAN [25] and [C₂Im][NO₃], is exactly the same at 13.2 kJ·mol⁻¹, despite the different melting temperatures, namely 12 °C and 35 °C for EAN and [C₂Im][NO₃], respectively, and the different melting heats per gram of pure IL, namely 122 J·g⁻¹ and 83 J·g⁻¹ for EAN and [C₂Im][NO₃], respectively. Taking into account that the thermal energy absorbed during the phase transition is used to break down strong intermolecular interactions such as Coulombic, hydrogen bond, and van der Waals interactions, which are the basis of the crystalline structure, and considering the similarity between both PILs, given that they share the same anion and alkyl chain and have very similar Coulombic and Van der Walls interactions, this could be an expected result. Furthermore, this coincidence evidences the similarity between the solid and liquid states of both bulk ILs. Beyond the electrochemical applications of the pure and blended ILs, which constitute the primary focus of our research group, this observation also carries significant implications in the emerging field of Phase-Change Materials (PCM), which are utilized in thermal energy storage [10]. In PCM, energy is stored via a reversible phase transition, such as melting, and can be used during the reverse transition. Additionally, owing to the minimal volumetric change, the solid–liquid transition proves more practical for this application than the liquid–vapor transition.

Similar to the melting enthalpy, the entropy of melting decreases as the concentration of salt increases, with more pronounced changes with higher electrostatic potentials at the surface of the metal cation (Li⁺ < Mg²⁺ < Al³⁺), as we have also observed for EAN–metal nitrate salt mixtures [25]. According to the criteria established by Timmermans [34], which correlate plastic crystal behavior with very low entropies of melting (<20 J·K⁻¹·mol⁻¹), the addition of these metal salts promotes plastic crystallization instead of the formation of rigid crystal structures.

3.2. Thermal Stability

The thermogravimetric curves (TG) and their corresponding derivative curves (DTG) for the pure IL and its mixtures are shown in Figure 2, with the corresponding parameters presented in

Table 3. Onset temperature (t_onset) obtained from TG curves, peak temperatures (t_peak) obtained from DTA curves, and Wooster temperature (t_Wooster) for pure [C₂Im][NO₃] and its saturated mixtures with LiNO₃, Mg(NO₃)₂ and Al(NO₃)₃ salt determined from TG and DTG curves. Experiments were performed with (998 ± 8) hPa of atmospheric pressure and relative humidity of (52 ± 7)%.

Compound	tonset/°C	tpeak/°C	tWooster/°C
[C2Im][NO3]	224	232	125
3 m IL + LiNO3	223	230	127
2 m IL + Mg(NO3)2	222	228	133
1 m IL + Al(NO3)3	181	185	97

Expanded uncertainties are U(t) = 5 °C and U(W) = 1% (0.95 level of confidence (k = 2)).

. It is noteworthy that pure IL exhibits a single peak on the DTG curve, which is indicative of a single-step process. However, all studied mixtures exhibit a second mass loss process, attributed to the degradation of the nitrate salts [25]. Due to the water content of magnesium and aluminum nitrate salts, a mass loss step at a lower temperature can be observed in the TGA curves of these mixtures. No results regarding the thermal stability of these mixtures have been found in the previous literature. As can be seen in Table 3, the addition of lithium and magnesium salt does not significantly influence the onset temperature. However, this temperature decreases by approximately 40 °C for the saturated IL + Al(NO₃)₃ mixture. This reduction can be linked to the lower thermal stability of the aluminum salt and the presence of hydrated water in this salt, which cannot be eliminated by vacuum procedures and evaporates at temperatures higher than free water [35], likely overlapping with the degradation phase of the mixture. These results are also in good agreement with a previous paper by our group analyzing the effect of the addition of monovalent (Li⁺), divalent (Ca²⁺, Mg²⁺) and trivalent (Al³⁺) nitrate salts on the thermal stability of EAN [25].

As widely recognized, the criterion based on the onset degradation is commonly used for comparative analyses of different materials. However, it tends to overestimate the long-term thermal stability of materials. Various methodologies utilizing thermogravimetry are available to determine the real thermal stability, but they typically require numerous long experiments (isothermal curves at different temperatures above the onset temperature) to be conducted, leading to extended study durations. However, as we have demonstrated in previous work [26], the approach proposed by Wooster et al. [36] introduces a simple method with results similar to those of previous, time-consuming methodologies. This approach offers a practical and efficient means of assessing thermal stability, facilitating the evaluation of materials for different applications. They propose that the temperature at which 1% degradation occurs in 10 h (T_0.01/10h) serves as a reliable indicator of thermal stability (T_Wooster), which is calculated from dynamic scans by using the following equation:

$$T_{Wooster}=0.82T_{\left(\raisebox{1ex}{$dW$}\!\left/ \!\raisebox{-1ex}{$dt$}\right.\ne 0\right)}$$

The temperature (in Kelvin) corresponding to the first appreciable weight loss is denoted as $T_{\left(\raisebox{1ex}{$dW$}\!\left/ \!\raisebox{-1ex}{$dt$}\right.\ne 0\right)}$. The results of this calculation are also shown in Table 3. As expected, the values obtained are much lower than those corresponding to t_onset. The Wooster criterium is mandatory for long-term exposure, while the stability limit can be extended to the onset temperature for short-term exposure.

While the Wooster temperature for these protic compounds is lower than that of previously studied aprotic ILs [26], it is adequate for the above aforementioned applications. This ensures the compounds’ efficacy and stability over extended periods of use.

4. Conclusions

This work is focused on the thermal characterization of a protic IL, ethylimidazolium nitrate, and its mixtures with monovalent, divalent, and trivalent nitrate salts at different concentrations. The main results of this work can be summarized as follows.

Although pure ionic liquid shows clear crystallization (upon cooling) and melting (upon heating) peaks, the salt addition tends to broaden and reduce both peaks. In fact, these peaks disappear for the saturated samples of lithium and magnesium nitrate salts. Additionally, the melting temperature decreases with an increasing valence and concentration of nitrate salt in the mixtures. However, the glass transition becomes more visible and shifts to higher temperatures with increasing salt concentration. These observations are attributed to the disorder generated by the rearrangement of ions in the polar domains, which erodes the hydrogen bond network of the protic IL.

Furthermore, the molar melting enthalpy and entropy of the mixtures decrease with the valence of the nitrate salt. This is consistent with the fact that the most saturated mixtures develop plastic crystal behavior, characterized by difficulty crystallizing, the presence of glass transitions, and a melting entropy lower than 20 kJ·mol⁻¹·K⁻¹, as observed in these blended samples.

Although the protic nitrate-based ILs studied in this and previous works have different cations, namely imidazolium and ammonium, the same molar melting enthalpy was obtained for pure ILs, consistent with the similar influence of coulombic and Van der Waals interactions in both ILs.

The thermal stability of the blended samples is similar to that of the bulk IL, around 200 °C, except for the IL + Al(NO₃)₃ mixtures, which is slightly lower due to the lowest thermal stability of the aluminum salt and the hydration water in the samples.

Although these compounds present great potential for use in energy storage applications, further studies of these mixtures are necessary before a final conclusion is reached.