Background & Summary

Concentrating solar power (CSP) coupled with thermal energy storage (TES) has attracted significant attention in recent decades as a promising renewable energy technology1,2. CSP systems generate electricity by concentrating sunlight to heat molten salt, which serves as both an efficient heat transfer fluids and a high-capacity thermal energy storage media, and subsequently converts thermal energy into electrical power3,4. Currently, sodium nitrate (NaNO3) and potassium nitrate (KNO3), known for their high energy densities and low vapor pressures, are typically blended in a 60:40 mass ratio to form the commercially available “solar salt”5,6. This binary nitrate mixture has become one of the most widely used heat transfer fluids and thermal energy storage media in CSP plants. Although compatible with the operating range of existing steam turbines, these commercial nitrate salt mixtures face thermal stability challenges, as they begin to decompose when operating temperatures exceed 823 K7,8.

To increase the thermal energy efficiency of CSP systems, which is fundamentally governed by the Carnot principle, operation at higher temperatures is essential. This drives the exploration of alternative molten salts with superior thermal stability. Among these, NaCl-based molten salt mixtures, such as NaCl-MgCl₂9,10, NaCl-CaCl₂11,12, and NaCl-KCl-MgCl₂-LaCl₃13, have emerged as highly promising candidates for next-generation TES applications. These chloride-based systems exhibit superior thermal stability and thermal energy storage potential at elevated temperatures, significantly exceeding the thermal stability limit of conventional nitrate-based solar salt (~823 K). For instance, the thermal conductivity of the mixtures typically ranges from 0.30 to 0.60 W·m−1·K−1 across a wide temperature range, from 1000 K to 1500 K, highlighting their enhanced high-temperature performance. However, a key challenge in optimizing these materials lies in understanding the fundamental relationships between their microscopic ionic interactions, microstructure, and macroscopic physicochemical properties4,14. Experimental characterization of molten salts at extreme temperatures is often limited by technical constraints, leading to sparse and sometimes inconsistent data. To address this gap, computational approaches, including the molecular dynamics simulations15 and the ab-initio calculations16, have become indispensable tools for elucidating the behaviour of molten salts and guiding the development of advanced TES materials.

This work consolidates and analyses published data on NaCl-based molten salt mixtures, providing a dataset on their microstructure and physicochemical properties. The dataset includes thermal expansion coefficient, thermal conductivity, specific enthalpy of fusion, specific heat capacity, density, and viscosity of mixtures, ionic self-diffusion coefficient, coordination bond angle and coordination bond length of ion pairs, and coordination number of ions across varying elemental compositions and temperature ranges from 556 K to 1400 K. The compiled information serves as a valuable resource for researchers and engineers, facilitating data-driven material design, machine learning model training, and the development of optimized molten salt formulations for advanced thermal storage applications. Furthermore, this work supports the integration of CSP industries with other renewable energy technologies, a critical step toward maximizing its role in global climate change mitigation efforts.

Methods

Boundary definition and research strategies

We conducted a systematic and comprehensive public publication search to identify relevant studies on the NaCl-based molten salt mixtures. The Web of Science database (https://webofscience.clarivate.cn/wos/alldb/basic-search) was selected as the primary search platform due to its extensive coverage of high-quality scientific publications. Academic interest in the microstructure and physicochemical properties of molten salt mixtures for the CSP was notably lacking prior to 201514,17. This search was restricted to Article and Review published from 2015-01-01 to 2025-03-31, to focus on recent advances in computational and experimental methods for the molten salt mixtures.

The search strategy employed a combination of keywords, including “NaCl-based molten salts”, “high-temperature molten salts”, “molecular dynamics simulation”, “first-principles calculation”, and “computational thermodynamics” to ensure broad retrieval of potentially relevant studies. The backward citation tracking, examining references of the key publications, and forward citation tracking, identifying newer studies citing the retrieved works, were employed to identify the additional relevant publications, to minimize selection bias and to capture important contributions that might have been missed in the initial database search.

The indexed publications were first filtered through the title and abstract relevance screening to identify studies providing both sufficient methodological details and quantitative data on either microstructure parameters or physicochemical properties of NaCl-based molten salt mixtures. And then, the publications with extractable and reproducible results were further filtered by the full-text evaluation. With this rigorous screening process, as shown in Fig. 1, only 21 high-quality publications, detailed in Table 1, were retained for in-depth analysis.

Fig. 1
figure 1

Flowchart of the literature screening and selection process.

Full size image
Table 1 Details of remaining 21 available publications for in-depth analysis.
Full size table

Data collection and management

The NaCl-based molten salt mixtures in this dataset encompass a variety of compositions, including the binary mixtures, such as NaCl-MgCl2 salt, NaCl-CaCl2 salt, NaCl-LiCl salt, NaCl-KCl salt, and NaCl-ZnCl2 salt, the ternary mixtures, like NaCl-MgCl2-KCl salt, NaCl-MgCl2-CaCl2 salt, NaCl-KCl-NaF salt, and NaCl-KCl-LiCl salt, as well as the quaternary mixture, such as NaCl-KCl-MgCl2-LaCl3 salt. The elemental composition of the above-mentioned NaCl-based molten salt mixtures is systematically summarized in Fig. 2, providing a clear overview of their ionic constituents.

Fig. 2
figure 2

Elemental composition of (a) binary mixtures and (b) ternary & quaternary mixtures in present dataset.

Full size image

The computational or experimental results in the publications for the molten salt mixtures are typically presented in two formats: tabular data and graphical representations. Numerical values listed in tables can be directly extracted for analysis, whereas data presented in figures often require digitization to ensure the data can be further processed. For this purpose, tools such as WebPlotDigitizer (https://plotdigitizer.com) are widely used to accurately extract numerical values from plotted curves or scatter points.

We systemically extracted available data from the publications, including the thermal expansion coefficients, thermal conductivity, specific enthalpy of fusion, specific heat capacity, density, and viscosity of the molten salt mixtures, the ionic self-diffusion coefficient, the coordination bond angle and coordination bond length of the ion pairs, and the coordination number of ions, with various elemental compositions in different temperature ranges. The explicit definitions and units of those parameters for molten salt mixtures were provided in Table 2. And two distinct datasets were meticulously compiled to facilitate: the physicochemical properties dataset and the microstructure dataset, which illustrated in Fig. 3. It is important to note that certain data records may exhibit missing parameters due to limitations in the source literature, including instances where specific properties were not computed or reported incompletely. The entire processing work of the present dataset is described in Fig. 4.

Table 2 Explicit definitions and units of parameters for molten salt mixtures.
Full size table
Fig. 3
figure 3

Classification of available data extracted from the publications.

Full size image
Fig. 4
figure 4

Flowchart of entire processing procedure of the present dataset.

Full size image

The relative uncertainties of those parameter data are below 1.0%, as reported in the referenced publications18,19.

Data Records

Two Excel-based datasets were created and are presented in Table 3. The details of the thermal expansion coefficient, thermal conductivity, specific enthalpy of fusion, specific heat capacity, density, and viscosity of the NaCl-based molten salt mixtures in different temperature ranges are recorded in an Excel file named physicochemical properties dataset. And another Excel file named microstructure parameters dataset contains the details of the ionic self-diffusion coefficient, coordination bond angle, coordination bond length, and coordination number of different ion pairs for the mixtures in different temperature ranges. The database was uploaded and publicly available at the Figshare repository20 and is available for download in Excel format.

Table 3 Overview of two Excel files in present dataset.
Full size table

Technical Validation

Physicochemical properties vs. temperature of the NaCl-based molten salt mixtures

A Pearson correlation analysis was conducted in this dataset, accompanied by a heatmap to display the correlations between the physicochemical properties vs. temperature of the NaCl-based molten salt mixtures, as shown in Fig. 5. The heatmap analysis reveals that the thermal expansion coefficient, thermal conductivity, and specific enthalpy of fusion exhibit significantly positive temperature dependence with higher Pearson correlation coefficients of r = 1.00, r = 0.98, and r = 0.99, respectively. In contrast, density and viscosity display significant inverse correlations with temperature with negative Pearson correlation coefficients r = −1.00 and r = −1.00, respectively21. The heat capacity shows moderate positive correlation with an intermediate Pearson correlation coefficient r = 0.53.

Fig. 5
figure 5

Correlations heatmap of physicochemical properties vs. temperature of NaCl-based molten salt mixtures from 556 K to 1400 K.

Full size image

Ionic self-diffusion coefficient vs. temperature of the NaCl-based molten salt mixtures

The ionic self-diffusion coefficients could be calculated via the mean-squared displacement method22. Consistent with fundamental thermodynamic and diffusion principles, the elevated temperatures promote ionic thermal agitation, resulting in a monotonic increase in the ionic self-diffusion coefficients with increasing temperature in Fig. 6, which illustrates the temperature dependence for the ionic self-diffusion coefficients of Na+, Mg2+, Ca2+, K+, F-, and Cl- in molten salt mixtures, respectively. The trends not only align with theoretical expectations but also demonstrate the internal consistency and thermodynamics validity in the present dataset.

Fig. 6
figure 6

Ionic self-diffusion coefficients of (a) Na+, (b) Mg2+ and Ca2+, (c) K+ and F-, and (d) Cl- in molten salt mixtures from 556 K to 1400 K.

Full size image

Coordination bond length of ion pairs vs. temperature of the NaCl-based molten salt mixtures

The coordination bond length of ion pairs servers as a crucial parameter for characterizing the microstructure of the molten salt mixtures23,24. These bond lengths are typically determined through the radial distribution function analysis in molecular calculations. Figures 7, 8 present the temperature dependence of the coordination bond lengths for different ion pairs from 556 K to 1400 K. Notably, the cation-cation pairs (Na+-Na+, K+-K+, and Mg2+-Mg2+) and the anion-anion pairs (Cl–Cl-) exhibit consistent decrease trends in the coordination bond lengths with increasing temperature, as shown in Fig. 7. All examined cation-anion pairs (Na+-Cl-, K+-Cl-, and Mg2+-Cl-) demonstrate opposite trends in Fig. 8, showing gradual elongation of the coordination bond lengths at elevated temperature. These two contrasting trends suggest fundamentally different temperature-dependent interactions between like-charged and oppositely-charged ions in the NaCl-based molten salt mixtures.

Fig. 7
figure 7

Coordination bond lengths of like-charged ion pairs (a) Na+-Na+, (b) K+-K+, (c) Mg2+-Mg2+, and (d) Cl–Cl- in molten salt mixtures from 556 K to 1400 K.

Full size image
Fig. 8
figure 8

Coordination bond lengths of oppositely-charged ion pairs (a) Na+-Cl-, (b) K+-Cl-, and (c) Mg2+-Cl- in molten salt mixtures from 556 K to 1400 K.

Full size image

Usage Note

This dataset serves as a critical resource for researchers and engineers in material screening and design. It enables the rapid identification of promising NaCl-based or MgCl₂-based molten salt compositions with desired thermophysical properties for next-generation CSP plants. Concurrently, it functions as a benchmark for validating the molecular dynamics and the ab-initio simulations, as well as a high-quality training dataset for developing machine learning models in materials informatics. Furthermore, the dataset provides essential input parameters for system-level modeling and simulation of the CSP thermal energy storage and heat transfer loops.

Although the data is structured in two separate Excel files at the Figshare repository to allow for focused analysis, it is recommended that users consult the original publications that provided in Table 1 for more detailed results and to facilitate effective use. The dataset will be periodically updated to include new compositions and properties, maybe not only limited to NaCl-based mixtures.