Designing new ultra-high temperature phase change materials from the Fe-Si-B system for thermal energy storage density above 1 MWh/m3

Designing new ultra-high temperature phase change materials from the Fe-Si-B system for thermal energy storage density above 1 MWh/m 3 Wojciech Polkowski a,* , Paolo Lai Zhong Lo Biundo a , Jianmeng Jiao a , Maria Wallin a , Adelajda Polkowska b , Filip Kateusz b , Merete Tangstad a a Department of Materials Science and Engineering, Norwegian University of Science and Technology, Alfred Getz vei 2, Trondheim 7491, Norway b Łukasiewicz Research Network – Krakow Institute of Technology, Zakopia´ nska 73 Str, Krakow 30-418, Poland ARTICLE INFO Keywords: Fe-Si-B alloys Latent heat thermal energy storage Phase change materials Electron microscopy Phase diagrams ABSTRACT New ternary Fe-Si-B alloys were developed as potential metallic phase change materials (PCMs) for application in ultra-high temperature latent heat thermal energy storage systems (LHTES). By a thermodynamic re-assessment of the Fe-Si-B system, three new Si-rich (approx. 40–50 wt%) compositions were pre-selected for further analyses. The new alloys can provide very high energy density (even above 1 MWh/m 3 ) at melting temperatures around 1150–1200 ◦C. Furthermore, new PCM candidates have twice lower B content than already developed Fe26Si-9B alloy, what is beneficial from the economic point of view. For the sake of experimental validation, the PCM candidates were produced by the arc melting technique. The outputs of thermodynamic calculations were verified in detailed microstructural studies and differential scanning calorimetry experiments. In the case of the newly developed Fe-46Si-5B and already designed Fe-26Si-9B alloy, the energy density higher than 1 MWh/m 3 was experimentally confirmed. 1. Introduction Latent heat thermal energy storage (LHTES) has been widely accepted as distributed and site-independent energy storage solution for both domestic and industrial applications [1]. Nowadays, this technology is mostly considered as a support of concentrated solar power (CSP) systems, giving a desired backup in sunless periods or in steam or coalfired ultra-supercritical power plants. The overall performance of each LHTES system is driven by thermophysical properties of an applied phase change material (PCM). Among various organic and inorganic substances, high temperature metallic PCMs provide several benefits, namely a high specific enthalpy of fusion at high operational temperatures, a high thermal conductivity, a low volumetric expansion, and good durability. So far, the most common metallic alloys considered as PCMs are these based on Al or Al-Si based near eutectic alloys. In this regard, it is worth to mention contribution of the research group led by Professor Takahiro Nomura [2–5]. Beside of using conventional as-cast Al-based alloys, the authors have also developed a microencapsulation technique [6,7] of aluminum based PCMs that brings new opportunities for safe and efficient long-term LHTES operations (up to 3000 cycles) [8]. Nevertheless, the main application of Al-based PCMs is focused on the Carnot battery and the reuse of a conventional ultra-supercritical power plants with a maximum operating temperature of approximately 650 ◦C [9]. Furthermore, the Al-based PCMs show latent heat in the range of 250–400 J/g, and by taking their relatively low specific weight, a total stored energy capacity is not higher than 0.3 MWh/m 3 . Why do we need ultra-high temperature materials (i.e. these operating at temperatures beyond 1000 ◦C, as it is defined by Zhou et al. [10])? Potential receivers of such materials are recently developed costeffective latent heat thermophotovoltaic (LHTPV) batteries that are a kind of power-to-heat-to-power storage systems that store electricity in the form of latent heat at very high temperatures (>1,000 ◦C) and converts it back to electricity on demand, using thermophotovoltaics [11]. As this technology is mostly dedicated for space industry applications [12], it is very important to ensure that a PCM introduced to LHTPV batteries combines: melting point above 1000 ◦C; and a high latent heat and high density to ensure a storage capacity of >1 MWh/m 3 (so, compared to that of pressurized hydrogen). Regarding ultra-high temperature applications Cu and its alloys are also considered as potential PCM because of its high melting point (T mCu * Corresponding author. E-mail address: [email protected] (W. Polkowski). Contents lists available at ScienceDirect Materials & Design journal homepage: www.elsevier.com/locate/matdes https://doi.org/10.1016/j.matdes.2025.114958 Received 1 July 2025; Received in revised form 10 October 2025; Accepted 14 October 2025 Materials & Design 260 (2025) 114958 Available online 22 October 2025 0264-1275/© 2025 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). =1083 ◦C), high thermal conductivity. In recent times, Cu-based PCMs have been studied extensively, mostly in their macro-encaspulated form. The reported examples include works by Akiyama et al. [13], on encapsulation of pure Cu or Sugo et al. [14] on Cu–Fe miscibility gap alloys produced by a powder metallurgy process. Although, the preliminary results seems to be promising, unfortunately latent heat of Cu is only 208.7 kJ/kg (or 0.52 MWh/m 3 ), thus Cu does not fullfil the requirement of high stored energy density. Recently, we have proposed to use the Fe-26Si-9B (wt %) ternary eutectic alloy as a new metallic PCM combining excellent functional properties in terms of application in highly efficient LHTES systems combined with thermophotovoltaic or thermionic energy converters that typically operate at temperatures above 1000 ◦C [15–17]. It was reported that this eutectic alloy exhibits a high thermal conductivity (10–30 W/mK) and a theoretical energy density above >1 MWh/m 3 for a melting temperature of 1150 ◦C [18]. This involves 10 times less volume of storing container than current molten salt-based TES, only comparable to pressurized H 2 storage and provides a great potential to exceed the energy density of most current electrochemical storage technologies [17]. The alloy was designed based on the concept of combining silicon and boron that possess the highest heat of fusion among all metallic and non-metallic elements (ΔH =1800 J/g and ΔH = 4640 J/g, respectively [19,20]). However, applicability of pure Si and eutectic Si-B alloys is quite strongly limited by (i) their high volumetric expansions during the phase change (around 10 % [21]) and by (ii) significantly high reactivity with most of existed refractories [22,23]. Thus, upon designing the Fe-26Si-9B alloy, the main role of iron addition was to reduce these adverse effects taking place during solid ↔ liquid phase changes [24]. Currently, we are working on a large-scale fabrication of this material by utilizing a special thermochemical treatment of B-rich ores, ferrosilicons and ferroborons as raw materials. During technological trials at a metallurgical plant, we were able to produce around 700 kg of the alloy (the results have not been published yet as they are subjected to patent protection, while a relevant paper is now under review). The process that we have developed might be easily adopted for a large-scale production of other Fe-Si-B alloys. In this work, we make a thermodynamical re-assessment of the Fe-SiB system to find out, if there are any other ternary near eutectic or/and congruently melting compositions having even higher heat of fusion. We have assumed that increasing Si content should enhance the heat of fusion of the Fe-Si-B alloys. Furthermore, by considering economical aspects of the usage of Fe-Si-B alloys as PCMs, and by taking into account very high price and geographically limited availability of boron [25,26], the other justification for re-designing the Fe-26Si-9B alloy is to reduce the boron content. Thus, the research question put in this work is as follows: “Can we re-arrange a chemical composition of the Fe-26Si-9B alloy in order to enhance its heat of fusion and at the same time, to move towards increased content of more abundant elements?”. To provide answer to this question, the thermodynamic calculations were followed up with a lab-scale fabrication of alloys and experimental validation of their thermodynamic properties. Finally, the main novelty of this work is based on a re-designing of Fe-Si-B alloys towards mores sustainable compositions that offer the energy storage density above 1 MWh/m 3 at ultra-high temperatures. 2. Materials and methods 2.1. Thermodynamic assessment The Phase Diagram module of FactSage 8.1 software powered by the FTLite database, was used to calculate the ternary Fe-Si-B diagram. The liquidus projection was calculated to extract chemical compositions of Fe-Si-B alloys having in-variant points with the liquid phase (so, being formed either in eutectic or peritectic reactions). For these alloys, the Equilib module was used to predict the evolution of enthalpy over temperature range of 20–1800 ◦C for assumed mass of 100 g of each alloy, and to further evaluate the heat of fusion (ΔH f ), temperature range of the phase transformation (ΔT) and density. For the sake of conversion from of heat of fusion units from J/g tom MWh/m 3 , a FactSage calculated density at liquidus temperature of each alloy was used. It should be mentioned that, as density at liquidus is for all alloys lower than that at room temperature, this is the “worst case scenario”, while all the predicted values are rather underestimated. 2.2. Experimental works Small samples (mass of ~ 1 g) of the Fe-Si-B alloys having compositions predicted by the FactSage calculations were produced by electric arc melting (Buehler MAM-1, Germany) of properly weighted mixtures of pure elements (99.99 % Fe, ThermoFisher; solar grade Si, Sintef Industry; 99.97 % B, Shanghai Aladdin Biochemical Technology Co Ltd, China). The melting process was carried out under non-oxidizing conditions by using an initial high vacuum of 10 -5 mbar, multiple purging the chamber with high purity Ar (99.999 %) and pre-melting of Zr oxygen getter. A consumable tungsten electrode was used to produce electric arc having a temperature around 3500 ◦C, as declared by the producer of the furnace (the real temperature of the arc was not measured). Each alloy piece was remelted 4 times to ensure a better structure and chemical composition homogeneity. Chemical compositions of the arc-melted alloys were experimentally examined by ICPSFMS method according to SS-EN ISO 17294-2:2023, US EPA Method 200.8:1994 (ALS Scandinavia AB Luleå, Sweden). The microscopic analyses of arc-melted samples were made on mechanically polished crosssections of alloy pieces. Finally, microstructural features of the alloys were revealed by inspections with a scanning electron microscope (FEI Scios, USA) equipped with X-Ray energy dispersive spectroscopy detector and electron backscatter diffraction system (SEM/EDS/EBSD). For each sample, several EBSD scans were performed using various stepsizes (0.2–2 μ m) to collect around 30 000 measurement points in each scan. The TSL OIM Analysis 5 software uses Confidence Index >0.1 as the threshold value, for which the point is basically classified as correctly indexed. The heat of fusion of the Fe-Si-B alloys was measured by differential scanning calorimetry technique (DSC, Netzsch STA 449 F3 Jupiter, Germany) and quantified by a dedicated software (Netzsch Proteus Thermal Analysis 6.1.0). For each alloy a DSC scan was performed within a temperature range of 20–1400 ◦C, by using heating/ cooling rates of 10 ◦C⋅min −1 , and under flowing argon atmosphere. To protect the alloys against undesired interaction, commercial alumina containers were spray-coated with a boron nitride slurry (as it was previously documented that this ceramic material shows exceptional inertness towards Si-based melts [27]). 3. Results and discussion 3.1. Calphad-based alloys design, heat of fusion and melting behavior The Fe-Si-B alloys compositions having three-point intersections with a liquid phase are marked in isothermal projections of the FactSage calculated ternary diagram (Fig. 1a and b) and further listed in Table 1. The software predicted 12 different ternary compositions. For the sake of comparison, the already developed Fe-26Si-9B alloy (Alloy#8, Table 1) was taken as the reference material. It should be noted that the presently used FactSage version predicts slightly different composition of this alloy (Fe-24Si-11B), however in order to avoid misunderstanding, we decided to keep and use the previous notation throughout the manuscript. The results of calculations on the enthalpy evolution as a function of temperature for Si-based alloys, as well as Feand B-rich ones, are given in Fig. 1c and d, respectively. It is observed that for all chosen Si-rich alloys (Fig. 1c), an additional enthalpy step increase is observed due to a solid-state reaction taking place around 950–1050 ◦C (depending on the alloy’s chemistry). Based on the results of phase equilibria calculations, this step was associated with a phase W. Polkowski et al. Materials & Design 260 (2025) 114958 2 Fig. 1. The results of FactSage calculations: isothermal projection of the Fe-Si-B system with marked new alloys’ compositions (a); enlarged area of Fe-Si-B system showing Si-rich and Fe-rich compositions (b). Total enthalpy evolution for Si-rich (c), Fe-rich and B-rich alloys (d). Table 1 The results of FactSage calculations: Chemical compositions, liquidus temperatures, heat of fusion and predicted phase composition of ternary Fe-Si-B alloys having three-point intersections with a liquid phase. Chemical composition [% mass] Liquidus T [◦C] Density at Liquidus [g/cm3] Melting Temperature range (ΔT) [◦C] Heat of fusion (ΔH f ) [MWh/m 3 ] Phases in equilibrium with a liquid Alloy # Type Fe B Si 1 Brich 69 23.1 8 1583 4.3849 279.6 1.62 SiBn Si Rhombohedralboron(beta) 2 Brich 72.2 20.8 6.9 1442 4.6085 291.8 1.77 FeB SiBn Si 3 Sirich 35.8 4.6 59.7 1272 3.3496 103 1.29 Si(B) Si SiB 3 4 Ferich 96 3.7 0.3 1179 6.7242 2.3 0.57 BCC solid solution Fe 2 B FCC solid solution 5 Sirich 46 4.4 49.6 1168 3.6953 19 1.12 Si FeSi 2 SiB 3 6 Ferich 86.9 3.2 10 1161 5.865 24 0.64 FeB Fe 3 Si Fe 2 B 7 Sirich 49 5 45.9 1159 3.799 10 1.09 SiB 6 FeSi 2 SiB 3 8 (ref) Ferich 65.1 10.8 24 1150 4.4223 0 1.09 FeB SiB 6 FeSi 9 Sirich 51.1 5.3 43.5 1149 3.8758 0 1.07 SiB 6 FeSi 2 FeSi 10 Ferich 80 2.3 17.7 1147 5.3599 51 0.62 FeB Fe 2 Si FeSi 11 Ferich 85.1 3.2 11.7 1137 5.7333 0.7 0.63 FeB Fe 3 Si Fe 2 B 12 Ferich 83 2.5 14.5 1131 5.5832 1 0.58 FeB Fe 3 Si Fe 2 Si W. Polkowski et al. Materials & Design 260 (2025) 114958 3 transformation between various iron silicides (FeSi 2 ↔ Fe 3 Si 7 ). It should be also noted that except the Alloys #8 and #9, the other Si-rich alloys undergo solid ↔ liquid phase change in some certain temperature range (ΔT). This transformation range in most of the cases is rather narrow (ΔT =0, 10, 19 ◦C), except the Alloy#3 for which is around 103 ◦C. By comparing the results of thermodynamic calculations in terms of the total enthalpy of fusion and melting temperature range it is found that: 1) The Fe-Si-B alloys highly alloyed with boron (namely, Alloys #1 and #2) exhibit high ΔH f values. On the other hand, the phase change takes place over a very wide temperature range (ΔT =280–290 ◦C), what is not acceptable from the point of view of the LHTES application [28]. 2) The Fe-rich alloys (Alloys #4; #6; #10, #11; #12) show rather low heat of fusion, and in terms of energy density, the values are not higher than ΔH f =0.57–0.64 MWh/m 3 . 3) As expected, the Si-rich alloys (Alloys #3; #5; #7, #9) show almost twice the higher enthalpy of fusion than that of Fe-rich counterparts. For all Si-rich alloys ΔH values higher than 1 MWh/m 3 were predicted for a liquidus temperatures around 1150–1170 ◦C. 4) Most of the Si-rich alloys melt over certain temperature range, so they combine both latent and sensible heat. However, except the Alloy #3 (ΔT =103 ◦C), the other materials undergo a full solid ↔ liquid change in a relatively low ΔT (between 0 up to 19 ◦C). The basic thermodynamic properties received from the FactSage indicate that all Si-rich PCM candidates show at least comparable ΔH to that of the reference Fe-24Si-9B alloy (Alloy#8). Therefore, it seems that increased Si content in the new alloys provide the same energy density, despite half the B content (4–5.3 vs. 9 wt%) as compared to the previously developed material. Finally, we narrowed our focus to Si-rich alloys (#3, #5, #7, #9 in Table 1) because of their high calculated enthalpy of fusion (~1.1–1.3 MWh/m 3 ) and relatively low melting ranges. In contrast, B-rich alloys were predicted to melt over an unacceptably wide temperature span (~280–290 ◦C) despite high enthalpy, and Fe-rich alloys had much lower enthalpy (<0.65 MWh/m 3 ). This rationale is sound and in line Fig. 2. A macroscopic view (a) and respective SEM images (a-e) of arc-melted Fe-Si-B alloys. Phase constituents were identified by using local EBSD and EDS microanalyses. W. Polkowski et al. Materials & Design 260 (2025) 114958 4 with expectations: adding Si boosts latent heat, while excessive B or Fe dilutes it or broadens the phase change interval. The calculated values (e.g. ~ 1.1 MWh/m 3 for alloys around 45–50 % Si) are plausible, and indeed similar magnitudes were reported in prior modeling of the eutectic Fe-26Si-9B (~0.87–0.95 MWh/m 3 predicted). 3.2. Microstructural characterization The macroscopic image of arc-melted alloys is shown in Fig. 2a. It should be noted that a rounded shape of arc-melted alloy pieces (Fig. 2a) suggests a rather low volumetric expansion during solidification (as opposite to arc-melted pure Si and Si-B alloys [27]). Microstructural features of the alloys are revealed in Fig. 2b–e. Generally, the identified phase compositions show a good agreement with these predicted by the results of thermodynamic calculations. The alloys’ microstructure consisted of a matrix of Fe-Si secondary solid solutions; and Si or various iron silicides precipitates accompanied by a primary crystals of silicon borides. The only one exception was previously reported microstructure of the Alloy #8 (Fig. 2d) that also include iron boride phase (FeB +FeSi +SiB 6 ) [29]. Nevertheless, some discrepancies between predicted and experimentally confirmed phases were found for Si-borides (having the SEM revealed morphology of dark regular particles). Although the FactSage results pointed towards a stability of SiB 3 or SiB 6 phases (Table 1), the best fitting of EBSD data (reflected by the highest Confidence Index values) was obtained for a hexagonal SiB n boride (when the n was equal to 36). A boron-rich part of the binary Si-B phase diagram has been discussed for a long time, while few contradicted findings have been presented over the years [30]. The most recent form of this diagram presented by Olesinski and Abbaschian [31] assumes a presence of SiB 3 (the metastable one); SiB 6 and SiB n (the one having the highest melting point) phases in a narrow Si dissolubility range closed to that of rhombohedral boron. In the present case, we should assume that rapid quenching conditions imposed in the arc-melting method facilitate a non-equilibrium solidification path. This finding is supported by the results of previous works [32] documenting a co-existence of various boron-rich Si-B intermetallics in rapidly quenched binary alloys. Exemplary results of the SEM/EDS/EBSD analyses are given in Fig. 3. 3.3. Heat of fusion – DSC results The results of FactSage calculations were verified in the DSC experiments on Fe-Si-B alloys produced by the arc-melting technique (Figs. 4 and 5) (Table 2). By using a peak area observed on DSC heat flow vs. T diagrams, the experimental ΔH was extracted and then compared to that established in the thermodynamic simulations. Before starting the experiments on Fe-Si-B alloys, we used a high purity Si (solar grade) to validate the DSC setup. The presently used DSC procedure and settings were adopted from previously reported paper on measurements of temperature and heat of phase transformation of pure silicon [33]. In the present work, quantification of the DSC run for the pure Si (Fig. 4) proved the feasibility of applied experimental conditions. The received Fig. 3. Exemplary results of the SEM/EDS/EBSD analyses obtained for the Alloy #5 (a); Alloy#7 (b); Alloy #8 (c); Alloy #9 (d): the EBSD phase maps an EDS mapping. W. Polkowski et al. Materials & Design 260 (2025) 114958 5 ΔH f of 1711 J/g was only 5 % lower than the theoretical value of 1800 J/ g [34], while the peak recorded during the continuous heating at 10 ◦C/ min was observed at T =1411–1428 ◦C (the theoretical melting point of Si is T m =1414 ◦C). Thus, these rather small discrepancies may suggest that the DSC experiments were properly designed. A difference between T onset and T end of a DSC peak was used to evaluate melting temperature range (ΔT) of the alloys at least comparatively under applied experimental conditions. However, it should be noted that due to applied continuous heating the obtained values are overestimated in some extent, what is clearly confirmed by the ΔT =17 ◦C for pure Si. The following findings are drawn from the DSC results obtained for the Fe-Si-B alloys (Fig. 5): 1) During a continuous heating at 10 ◦C⋅min −1 the new Fe-Si-B alloys melt in a rather low temperature range (ΔT between 23–68 ◦C). 2) There are some discrepancies between ΔH f values predicted by the FactStage approach and these measured by the DSC technique. Among all investigated materials, the ΔH f value above 1 MWh/m 3 value was experimentally confirmed for the Fe-26Si-9B (Alloy #8) and Fe-46Si-5B (Alloy #7) alloys. Indeed, for both alloys, a very good agreement between predicted and experimentally measured heat of Fig. 4. A DSC curve recorded for pure Si (a reference material). Fig. 5. DSC curves recorded for newly developed Fe-Si-B alloys: Fe-49Si-4B (alloy #5) (a); Fe-46Si-5B (alloy #7) (b); Fe-26Si-9B (alloy #8) (c); Fe-43Si-5B (d). Table 2 Comparison of FactSage and DSC results on Heat of fusion and melting temperatures of selected Fe-Si-B alloys. Chemical composition, ICP-MS [wt%] Chemical composition, ICPMS [wt%] Melting Temperature range (ΔT) [◦C] Heat of fusion (ΔH f ) [J/g] Heat of fusion (ΔH f ) [MWh/ m 3 ] Alloy no. Fe B Si Fe B Si FactSage DSC FactSage DSC FactSage DSC T onset T max T end Pure Si − − 99.999 − − − 1414 (0) 1411 1427 1428 1788 1711 1.27 1.22 (4.0 %)* 5 46.000 4.447 49.553 45.472 4.243 50.285 1149–1167 (18) 1172 1199 1203 1090 580 1.12 0.623 (44.4 %)* 7 49.032 5.034 45.934 47.968 4.666 47.366 1149–1159 (10) 1173 1190 1196 1037 960 1.09 1.064 (2.4 %)* 8 65.104 10.848 24.048 65.120 8.973 25.908 1150 (0) 1196 1242 1264 888 834 1.09 1.02 (6.4 %)* 9 51.124 5.349 43.527 50.998 4.855 44.146 1149 (0) 1174 1191 1201 997 688 1.07 0.879 (17.9 %)*I *a difference between predicted vs. DSC established values. W. Polkowski et al. Materials & Design 260 (2025) 114958 6 fusion was obtained. To our best knowledge, this is so far the highest experimentally documented energy density for metallic PCM alloy at temperature range of 1100–1300 ◦C. In this regard, this result places the alloy beyond the already existed capabilities of PCMs reported in the literature (Fig. 6) [35]. As it has been shown by Datas et al. [36], the energy density of ~1 MWh/m 3 is 2–6 times higher than that of Li-ion batteries, 10–20 times higher than that of lead-acid batteries and 5–10 times than that of the current state of the art molten salt-based TES systems utilized in CSP applications. The underperformance of Alloy #5 and #9 might be due to their solidification paths: for instance, FactSage predicted Alloy #5 would melt in a narrow ΔT ~ 19 ◦C with phases Si + FeSi +SiB. If in reality an additional phase or segregation occurs (e.g. primary Si crystals or a ternary compound formation), part of the silicon may melt at a higher temperature, leaving a smaller latent heat peak around the main eutectic. From the FactSage calculation it is noted that all Si-rich alloys show a solid-state phase transition near 850–1050 ◦C. This pre-melting reaction would consume a bit of energy (sensible/ latent), but it is relatively minor. A more impactful factor could be the presence of the newly identified FeSiB ternary phase. If FeSiB (or a similar ternary borosilicide) forms in the Si-rich alloys, it might not be included in the FTLite database used for FactSage calculations. Omission of such a phase in the model could lead to an overestimation of latent heat −because the simulation might assume more Si segregates as pure Si or SiB (melting at lower temperature with high ΔH), whereas experimentally some Si is tied up in FeSiB which melts differently. Unfortunately, as neither FactSage nor EBSD databases include ternary FeSiB intermetallics, it is not possible to experimentally confirm existence of such phases. In summary, the thermodynamic calculations are a strong guide, but they might not cover all possible solutions. Nevertheless, in order to make a deeper insight into a possible impact of ternary Fe-Si-B intermetallics on latent heat of fusion of examined alloys, we reviewed reported literature data on the Fe-Si-B system. Specifically, we took into account ternary Fe-Si-B phases predicted in CALPHAD calculation by Poletti and Battezzati [37] and by Zaitsev et al. [38], as well as the results previously published by our team [29] (given in Table 3). By using predicted compositions of ternary phases we made another round of the FactSage calculations. It is clearly observed that ternary Fe-Si-B compounds show rather low specific ΔH f as compared to pure silicon (1800 J/g) [34] and binary eutectic Fe-Si alloys (930–1010 J/g) [39,40], so their possible presence should negatively affect a heat of fusion of Fe-Si-B alloys. However, a precise and more certain identification of both ternary phases and silicon borides, would require using transmission electron microscopy (TEM) based diffraction tools such as nanobeam precession TEM or atomic resolution STEM HAADF. The goal for a PCM is often a sharp phase change, but the new alloys do exhibit some melting ranges under DSC heating (ΔT of 23–68 ◦C for a 10 K/min ramp). The continuous heating tends to overestimate the apparent melting range (e.g. pure Si showed ΔT≈17 ◦C instead of 0 ◦C). When extrapolated to equilibrium, most of the proposed alloys likely have invariant or near-invariant behavior (predicted ΔT ≤20 ◦C). Notably, Alloy #7 (Fe-46Si-5B) melted over ~ 23 ◦C in DSC, which is still quite narrow in practical terms. Alloy #8 (Fe-26Si-9B) showed a larger ΔT≈68 ◦C in the DSC run, reflecting its off-eutectic composition. We believe that only Fe-46Si-5B and Fe-26Si-9B truly meet the “narrow ΔT +high ΔH” criterion among the tested alloys. This conclusion is supported by the fact that Alloy #9 and Alloy #5, despite high Si, did not fully deliver on the desired properties. A critical reader might wonder if further optimization (e.g. slight tweaks in Si/B ratio or using slower cooling to get closer to equilibrium microstructure) could improve those alloys. It could be also consider whether Alloy #9 predicted congruent melting at 1149 ◦C was adversely affected by boron loss (thus no longer congruent). Overall, the thermodynamic foundation is solid, and the experimental results mostly reinforce it, with the exception of the two alloys where the model over-predicted latent heat. This does not invalidate the approach, but it suggests that Fe-46Si-5B is the standout composition, while others might need further investigation or refinement. To summarize this part, it is worth underlining that the latent heat of Fe-Si-B alloys seems to be strongly affected by their phase compositions. From our theoretical and experimental studies supported by literature data, it appears that enhancement of ΔH f can be obtained by increasing content of Si, that in turn leads to a higher fraction of Si and silicide based phases at the expense of low ΔH f borides ans silicoborides. 4. Conclusions and future remarks In this work, the results of thermodynamic re-assessment of the Fe-SiB system were carried out, in terms of designing new PCMs for high temperature applications. Consequently, three new Si-rich alloys were proposed as potential candidates and then theoretically and experimentally compared to the already developed Fe-26Si-9B alloy. According to thermodynamic predictions, all the alloys should provide the high heat of fusion ΔH >1 MWh/m 3 in a narrow temperature range (ΔT up to 20 ◦C). However, the results of DSC studies confirmed these findings only for the Fe-26Si-9B reference and newly developed Fe-46Si-5B alloy. It is also worth mentioning that the new alloy has almost half the B content as compared to the Fe-26Si-9B counterpart, what is beneficial from the economic point of view. Although, this high storing energy density was experimentally confirmed for the first time for a metallic PCM and provided a very promising first indication for future development, there are still several functional and technical properties that ought to be evaluated before the final application in TES systems. Specifically, the other thermophysical Fig. 6. A possible energy storage capacity vs. melting temperature of PCMs: a comparison of presently obtained results with these reported in the literature (based on [17;35]). Table 3 The results of FactSage calculations on thermophysical properties of ternary FeSi-B intermetallics reported in the literature. Chemical composition (at. %) Phase formula Fe Si B Solidus T [◦C] Liquidus T [◦C] Heat of fusion [J/g] Fe4.7Si2B [23] 61.04 25.97 12.99 1130 1187 547.11 Fe2Si0.4B0.6 [23] 66.67 13.33 20.00 1156 1233 517.13 Fe5Si1B2 [23] 62.50 12.50 25.00 1156 1339 683.65 Fe6SiB [24] 75.00 12.50 12.50 1179 1221 373.5 FeSiB3 [15] 20.00 20.00 60.00 994 1147 820.98 W. Polkowski et al. Materials & Design 260 (2025) 114958 7 properties including thermal conductivity, volumetric expansion and enthalpy evolution over cyclic phase changes, must be examined in detail. Furthermore, the thermochemical compatibility of the new alloy (s) with refractories and a lack of properties degradation during longterm service is a particularly important practical aspect that must be included in the future research actions. CRediT authorship contribution statement Wojciech Polkowski: Writing – original draft, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Paolo Lai Zhong Lo Biundo: . Jianmeng Jiao: Methodology, Investigation. Maria Wallin: Project administration, Data curation. Adelajda Polkowska: Methodology, Investigation. Filip Kateusz: Methodology, Investigation. Merete Tangstad: Writing – review & editing, Project administration, Funding acquisition, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements SUNSON project has received funding from Horizon Europe Research and Innovation Action programme under Grant Agreement no 101083827, funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or CINEA. Neither the European Union nor the granting authority can be held responsible for them. Data availability Data will be made available on request. References [1] N. Zhang, Y. Yuan, X. Cao, Y. Du, Z. Zhang, Y. Gui, Latent heat thermal energy storage systems with solid–liquid phase change materials: a review, Adv. Eng. Mater. 20 (2018) 1700753. [2] Y. Shimizu, T. Nomura, Al–Si–Fe alloy-based phase change material for hightemperature thermal energy storage, High Temp. Mater. Processes 42 (2023) 20220280. [3] Y. Shimizu, T. Nomura, Die-cast Al–Si–Cu alloy (ADC12) as a phase change material for medium–high-temperature heat storage, J. Alloy. Compd. 1017 (2025) 179006. [4] Y. Shimizu, T. Kawaguchi, K. Tanahashi, T. Nakamura, M. Jeem, T. Nomura, Sensible/latent hybrid heat storage material using Solar Salt and Al–Cu–Si alloybased phase change material, Chem. Eng. J. 508 (2025) 160832. [5] L. Shan, T. Nakamura, T. Nomura, A. Martin, J.N. Chiu, Single and cascaded high temperature latent heat thermal energy storage in concentrated solar power, J. Storage Mater. 136 (2025) 118431. [6] N. Sheng, C. Zhu, G. Saito, T. Hiraki, M. Haka, Y. Hasegawa, H. Sakai, T. Akiyama, T. Nomura, Development of a microencapsulated Al–Si phase change material with high-temperature thermal stability and durability over 3000 cycle, J. Mater. Chem. A 6 (2018) 18143–18153. [7] Y. Wang, S. Wang, T. Nomura, A. Kurniawan, R.B. Cahyono, Z. Rao, N. Sheng, C. Zhu, Macroencapsulation and in-situ formed Al-Cu-Si ternary alloy phase change materials for high-temperature heat storage, Sol. Energy Mater. Sol. Cells 294 (2026) 113941. [8] T. Nomura, C. Zhu, N. Sheng, G. Saito and T. Akiyama, “Microencapsulation of metal-based phase change material for high-temperature thermal energy storage,” Scientific Reports 5 (2015), p. 9117, vol. 5, p. 9117, 2015. [9] H. Fan, Z. Zhang, J. Dong, W. Xu, China’s R&D of advanced ultra-supercritical coalfired power generation for addressing climate change, Therm. Sci. Eng. Prog. 5 (2018) 364–371. [10] X. Zhou, S. Yamashita, M. Kubota, H. Kita, Encapsulated copper-based phasechange materials for high-temperature heat storage, ACS Omega 6 (2022) 5442–5452. [11] A. Datas, A. L´ opez-Ceballos, E. Lopez, A. Ramos, C.D. Canizo, Latent heat thermophotovoltaic batteries, Joule 6 (2022) 418–443. [12] A. Datas, A. Marti, Thermophotovoltaic energy in space applications: review and future potential, Sol. Energy Mater. Sol. Cells 161 (2017) 285–296. [13] N. Maruoka, K. Sato, J.-I. Yagi, T. Akiyama, Development of PCM for recovering high temperature waste heat and utilization for producing hydrogen by reforming reaction of methane, ISIJ Int. 42 (2002) 215–219. [14] H. Sugo, E. Kisi, D. Cuskelly, Miscibility gap alloys with inverse microstructures and high thermal conductivity for high energy density thermal storage applications, Appl. Therm. Eng. 51 (2013) 1345–1350. [15] A. Datas, M. Francoeur, M. Shimizu, R. Vaillon, Advances in thermophotovoltaics: materials, devices, and systems, Sol. Energy Mater. Sol. Cells 240 (2022) 111711. [16] A. Ramos, E. Lopez, C.D. Canizo, A. Datas, Cost-effective ultra-high temperature latent heat thermal energy storage systems, J. Storage Mater. 49 (2022) 104131. [17] A. Datas, A. B. Cristobal, C. d. Ca˜ nizo, E. Antolín, M. Beaughon, N. Nikolopoulos, A. Nikolopoulos, M. Zeneli, N. Sobczak, W. Polkowski, M. Tangstad, J. Safarian, D. M. Trucchi, A. Bellucci, M. Girolami and e. al, “AMADEUS: Next generation materials and solid state devices for ultra high temperature energy storage and conversion,” AIP Conference Proceedings, vol. 2033, p. 170004, 2018. [18] J. Jiao, J. Safarian, M. Tangstad, The use of Fe-26Si-9B alloy as phase change material in Si3N4 container, Crystals (2022) 376. [19] http://www.chemistry-software.com/, “HSC Chemistry 9,” Chemistry Software Ltd, Kings Cross Lane, South Nutfield, Surrey, RH1 5NS, UK, 2023. [20] A. Datas, E. L´ opez, A. Ramos, N. Nikolopoulos, A. Nikolopoulos, M. Zeneli, N. Sobczak, W. Polkowski, M. Tangstad, J. Safarian, D.M. Trucchi, A. Bellucci, M. Girolami, D. Bestenlehner, S. Lang, A. Vitulano, G. Sabbatella, C. del Ca˜ nizo, A. B. Cristobal, A. Martí, “Ultra-high temperature energy storage and conversion: A review of the AMADEUS project results, Aip Conference Proceedings 2303 (2020) 0028552. [21] Y. Sato, T. Nishizuka, K. Hara, T. Yamamura, Y. Waseda, Density measurement of molten silicon by a pycnometric method, Int. J. Thermophys. 21 (6) (2000) 1463–1471. [22] W. Polkowski, N. Sobczak, R. Nowak, G. Bruzda, A. Kudyba, A. Polkowska, I. Krzak and A. Tch´ orz, “Solid/liquid interaction between Si-based phase change materials and refractories,” in Ultra-High Temperature Thermal Energy Storage, Transfer and Conversion, Cambridge, UK, Woodhead Publishing Series in Energy, 2021, pp. 113–137. [23] W. Polkowski, N. Sobczak, A. Polkowska, Silicon as a phase change material: performance of h-BN ceramic during multi-cycle melting/solidification of silicon, JOM 71 (2019) 1492–1498. [24] J. Jiao, S. Jayakumari, M. Wallin, M. Tangstad, Graphite crucible interaction with Fe-Si-B phase change material in pilot-scale experiments, High Temp. Mater. Processes 42 (2023) 20220288. [25] S. Betharia, “Boron,” in Encyclopedia of Toxicology (Third Edition), The Netherlands, Academic Press, 2014, pp. 536–539. [26] B. T. LIMITED, “Introduction to Boron Mining – Locations and Major Players,” 2021. [Online]. Available: https://borates.today/boron-mining-locations/. [Accessed 07 11 2023]. [27] W. Polkowski, N. Sobczak, G. Bruzda, R. Nowak, D. Giuranno, A. Kudyba, A. Polkowska, K. Pajor, T. Kozieł, I. Kaban, The effect of boron content on wetting kinetics in Si-B Alloy/h-BN system, J. Mater. Eng. Perform. 28 (2019) 3819–3825. [28] A. Guruprasad, L. Lingkun, X. Huang, F. Guiyin, Thermal energy storage materials and systems for solar energy applications, Renew. Sustain. Energy Rev. 68 (2017) 693–706. [29] J. Jiao, B. Grorud, C. Sindland, J. Safarian, K. Tang, K. Sellevoll, M. Tangstad, The use of eutectic Fe-Si-B alloy as a phase change material in thermal energy storage systems, Materials 12 (2019) 2312. [30] C. Brosset, B. Magnusson, The silicon – boron system, Nature 187 (1960) 54–55. [31] R. Olesinski, G. Abbaschian, The B-Si (Boron-Silicon) system, Bull. Alloys Phase Diagr. 5 (1984) 479–484. [32] W. Polkowski, G. Bruzda, A. Polkowska, I. Krzak, A. Tch´ orz, R. Nowak, A. Kudyba, M. Ksią˙ zek, K. Karczewski, Microstructural characteristics and mechanical properties of Si-B alloys for functional and structural ultra-high temperature applications, J. Alloy. Compd. 935 (2023) 167672. [33] M. Homa, N. Sobczak, Measurements of temperature and heat of phase transformation of pure silicon by using differential scanning calorimetry, J. Therm. Anal. Calorim. 138 (2019) 4215–4221. [34] Chemistry Software Ltd., “HSC Chemistry,” Chemistry Software Ltd, Surrey, 2023. [35] G. Alva, L. Liu, X. Huang, G. Fang, Thermal energy storage materials and systems for solar energy applications, Renew. Sustain. Energy Rev. 68 (2017) 693–706. [36] A. Datas, A. Ramos, A. Marti, C.D. Canizo, A. Luque, Ultra high temperature latent heat energy storage and thermophotovoltaic energy conversion, Energy 107 (2016) 542–549. [37] M.G. Poletti, L. Battezzati, Assessment of the ternary Fe–Si–B phase diagram, CALPHAD: Comput. Coupl. Phase Diagrams 43 (2013) 40–47. [38] A. Zaitsev, N.E. Zaitsev, A.A. Kodentsov, A thermodynamic study of liquid Fe-Si-B alloys; an influence of ternary associates on a liquid : glass transition, Metall. Mater. Trans. B 34B (2002) 887–898. [39] W. Polkowski, P.L.Z.L. Biundo, J. Jiao, M. Wallin, B. Kalicki, J. Ciftci, L. Zrodowski, A. Polkowska, F.K. Aleksandra Betkowska, M. Tangstad, Micro-encapsulation of SiFe ultra-high temperature phase change material: fabrication and basic energy storage properties, Results Eng. 25 (2025) 104191. [40] P.L.Z.L. Biundo, J. Jiao, M. Wallin, A. Polkowska, R. Nowak, G. Bruzda, A. Kudyba, F. Kateusz, A. Bętkowska, M. Tangstad, W. Polkowski, Thermophysical properties of eutectic Fe-57Si ultra-high-temperature metallic phase change material, J. Therm. Anal. Calorim. 150 (2025) 9903–9911. W. Polkowski et al. Materials & Design 260 (2025) 114958 8