Investigation of high-enthalpy organic phase-change materials for heat storage and thermal management

The growing interest in phase-change materials (PCM) is related to their possible role in thermal energy storage and thermal management. The choice of materials depends strongly on the required temperature range, whereas the latent heat of solid–liquid phase transition has to be as high as possible. Among other organic PCM, sugar alcohols have gained some attention due to their availability and certain advantageous properties. However, the thermal processes in these materials still require investigation. In the present work, we focused on the materials with solid–liquid phase change within 80 °C–100 °C. A comprehensive literature survey was conducted to elucidate the available sugar alcohols relevant to this range. It was found that the use of pure materials of this type is not very practical, because of their scarcity in the required range and their specific features, like difficulties with crystallization and solidification. On the other hand, based on the literature, we have discerned three eutectic mixtures of erythritol with other organic materials, namely, erythritol–xylitol, erythritol–urea and erythritol– trimethylolethane (TME). In all those cases, it is remarkable that while the components commonly have rather high melting temperatures, the eutectic mixtures had the phase transitions in the required range. Still, each of these mixtures has its own peculiar features, especially at cooling and solidification. An extensive experimental study was performed to provide detailed visualization of these major processes. The results revealed the melting temperature and latent heat of the mixtures to be: 84 °C and 190 J g−1 for erythritol–xylitol, 82 °C and 227 J g−1 for erythritol–urea. Erythritol–TME has two phase transitions at 82 °C and 97 °C, with total latent heat of 198 J g−1. Based on the present findings, the erythritol–urea mixture is the best PCM candidate for the melting range within 80 °C–100 °C.


Introduction
Phase change materials (PCM) capture a large amount of thermal energy via solid-liquid phase transition, using their latent heat (either mass-or volume-specific). PCM can be integrated into latent heat thermal energy storage systems [1][2][3][4][5], as well as used in transient thermal management applications [6][7][8][9]. The current study focuses on certain applications related to heat storage and transient thermal management, where the physical properties needed are melting temperature in the range of 80 • C-100 • C, whereas the latent heat should be as high as possible. Low price is an additional consideration.
These requisites are set as the starting point of the present exploration, and may be related to applications such as high-end electronics transient cooling solutions: consider, e.g. Nvidia's GeForce series GPUs whose temperature rises to 93 • C at full load, while at normal load it should remain at 70 • C-75 • C [10]; another example is Intel's Core ™ i9-13900KS Processor, with maximum allowed temperature of 100 • C [11,12]. It is worth noting that the maximum allowed temperature is for performing demanding tasks and not for normal operation. Thus, it is vital to have proper transient cooling to prevent overheating and damage to the Shao et al [22] Höhlein et al [27] Mehling and Cabeza [2] Sugar alcohol Tm, • C L, J g −1 Tm, • C L, J g −1 Tm, •  components. Another example of possible application is mid-temperature thermal energy storage that is based on PCM with melting temperatures in the range of 80 • C-100 • C, which can be used to store thermal energy in a compact and efficient manner. These can be applied to excess heat generated by industrial processes, like chemical reactions and oil refining, or by equipment, e.g. engines and turbines [13,14]. Based on the application at hand, several PCM options are possible according to the temperature requirements, because the PCM should undergo phase transition in a specified range. PCM may be pure substances, but to manipulate the melting temperature where necessary, eutectic mixtures of various materials were suggested [2,8]. Classical PCM studies [1,6] outlined the desirable features of PCM which are deemed suitable for the purposes of heat storage and thermal management. Except for the appropriate melting temperature, which is crucial for a given application, the key desirable features include [8]: high latent heat, reversible solid-to-liquid transition, high thermal conductivity, non-toxicity and cycling stability. As mentioned frequently [8], there is no ideal PCM which 'checks every box' in the list of requirements. Thus, a PCM solution should be integrated with heat transfer enhancement means, e.g. metal foam structures [15][16][17], extended surfaces [18], or close-contact melting (either occurring naturally [19,20] or governed using external forces [9]).
The paraffins family (C n H n+2 ) is the most popular in current PCM investigations for relatively low temperatures [2,8]; however, pure paraffins are pretty expensive, which makes them unpractical for usage. The more affordable commercial grade paraffins, usually comprised of several different hydrocarbons, may have their own problems related, e.g. to extended melting range. In any case, paraffins are usually obtained by petroleum processing, i.e. they are related to fossil fuels production and thus are not 'green' [2].
Sugar alcohols have gained some attention as prospective PCM. One of the prominent ones, erythritol, was suggested for this purpose back in 1998 [10]. Sugar alcohols general chemical formula is HOCH 2 (CHOH) n CH 2 OH, while different forms are attained based on the location of the OH groups, see [2]. The materials of this class have melting points from 90 • C to about 200 • C, with relatively high latent heat (mass-specific). The latter feature, combined with their high density, results in very high volume-specific latent heat. These materials are considered non-toxic [21]. One of their known weaknesses is the fact that they show supercooling at solidification [2,21].
The renowned interest in the use of sugar alcohol as PCM for the medium range resulted in the recent studies [22][23][24], where an extensive in-depth analysis of sugar alcohols and their mixtures was performed. According to the results of [22], the entire temperature spectrum may be divided into two sub-ranges: the low-temperature one below 100 • C (which is of our interest) and the medium-temperature one between 100 • C and 250 • C.
In a preliminary screening reported in [22], a selection of six sugar alcohols was found, based on their affordable prices. These materials were: xylitol, d-sorbitol, erythritol, d-mannitol and d-dulcitol (all linear-chain ones) and inositol (cyclic chain). Of those materials, just two (xylitol and d-sorbitol) have melting temperatures below 100 • C, though those temperatures are well above 90 • C. Their properties, reported in [22], are given in table 1. It should be noted that the melting points and enthalpies of fusion of the six pre-screened sugar alcohols were measured in [22] by differential scanning calorimetry (DSC), at a representative ramping/cooling rate of 5 • C min -1 . Table 1 also includes data from [22] for erythritol-the latter, while obviously having a too high melting temperature for our purposes, has a very high latent heat and may be used in eutectic mixtures with tailored properties, as will be discussed below. The table also presents material data taken from two additional literature sources. One can note that, concerning the material properties, the agreement in the literature is not perfect, presumably due both to different manufacturers and deviations in the DSC measurements [25,26]. As mentioned in [27], erythritol and xylitol have remarkable volumetric heat storage densities of up to 450 MJ m −3 and comparatively high thermal diffusivity/conductivity. However, the research works in the literature stress that their high degree of supercooling must be overcome to make these materials applicable in practice.
DSC analysis, reported in [22], indicate that, when cooling is done, both xylitol and d-sorbitol apparently remain liquid instead of exhibiting solidification, i.e. they continue to exist as supercooled liquids well below their nominal phase-change temperature, even at low cooling rates applied during the DSC measurements. According to [22], such behavior may be due to the specific rotameric conformation of xylitol and d-sorbitol.
In any case, this behavior practically prevents the use of these materials in thermal energy storage systems, where the melting-solidification cycles are essential. Also, this would be a no-go in many thermal management cases, for the same reason, though there may be some limited applications (especially in the defense sector), where the cycles are not required. Of course, there might exist methods to enforce nucleation, e.g. by seeding particles or by some other external action, but their practicality beyond laboratory conditions is yet to be proved [28].
For the entire temperature range below 250 • C, 15 binary eutectic mixtures were prepared based on the chosen sugar alcohols and characterized in [22]. In table 2, we present the results of [22] for the temperature range below 100 • C. As stressed in [22], while generally a good agreement exists between the results for materials from various sources, the mixture of xylitol and erythritol was problematic in this sense.
All in all, based on [22][23][24], there are nine mixtures, all containing either xylitol or d-sorbitol, which have melting points below 100 • C, with the lowest melting point of ∼73 • C for the mixture of xylitol (52 mol%) + d-sorbitol. However, all mixtures containing d-sorbitol have relatively low melting enthalpies, apparently due to the low enthalpy of d-sorbitol itself, ∼179 J g −1 . The mixtures containing xylitol generally look more attractive in this sense, but as discussed above they have inherent solidification problems, suffering of non-crystallization. Also, the latent heats of the mixtures presented in table 2 are not especially high in any case, and thus other candidates should be considered. For instance, in [22], the eicosanoic acid (melting point of ∼74 • C, latent heat of ∼227 J g −1 ) is suggested as a replacement for the mixture of xylitol (52 mol%) and d-sorbitol.
According to [24], the thermal endurance of pure and binary eutectic mixture sugar alcohols is of a significant concern, as both the melting point and latent heat of fusion of these sugar alcohols may degrade with increasing the heating duration, and a higher degree of superheat leads to faster degradation. Among the various candidates, erythritol exhibits the best thermal endurance [24], though its performance also may be affected at high superheats.
Another approach, found in the literature though somewhat limited in its scope, is to find mixtures that, while containing erythritol, notable for its very high latent heat of fusion of above 330 J g −1 , involve also organic materials other than sugar alcohols [26,29,30].
The investigation of Hidaka et al [29] was dedicated to erythritol-polyalcohol mixtures as PCM, which could be suitable for thermal management or heat storage in cases where melting temperature in the range of 80 • C-100 • C is desired. The phase-change temperature was adjusted in [29] by adding polyalcohols, such as trimethylolethane (TME), 2-ethyl-2-methyl-1, 3-propanediol, 2-amino-2-methyl-1,3-propanediol, 1,4-butanediol, 2-amino-1,3-propanediol and trimethylolpropane (TMP), to erythritol. It was found that 40% erythritol and 60% TME mixture melted at 86.1 and 97.8 • C with a latent heat of 246 J g −1 . Another suggested mixture, composed of 90% of the above-mentioned mixture and 10% TMP, melted at 80.0 and 95.0 • C with a latent heat of 231 J g −1 . The results of [29] illustrated that erythritol-polyalcohol mixtures could be potential candidates for PCM. The phase-change temperatures and latent heats of the erythritol-TME mixtures at various proportions, adapted from [29], are shown in table 3.
An additional suggested binary mixture, made of erythritol and urea for its application as a PCM, was explored in [26]. The reported properties for urea were a melting temperature of 133.3 • C and a latent enthalpy of 243.7 J g −1 . The eutectic composition formed by 54.9% (w/w) erythritol and 45.1% (w/w) urea  1. Melting points and latent heats of the sugar alcohols [22], their binary eutectic mixtures [22] and other erythritol-based eutectic mixtures with TME [29] and urea [26]:• pure sugar alcohols, ⃝ eutectic mixtures of sugar alcohols, erythritol-based eutectic mixtures.
was achieved, with a melting temperature of 81.13 • C and a latent heat of 248 J g −1 . Since urea undergoes thermal degradation upon heating, the suitability of the mixture as a PCM was evaluated. The results revealed that the enthalpy of the eutectic mixture was reduced by just 3.6%, vs a decrease of 75% of its initial storage ability for pure urea. It was deduced that the thermal degradation of liquid urea was greatly dependent on the temperature. This conclusion, similar to the results observed in 1829, reinforce the possibility for the use of urea as a component of eutectic PCM that work at maximum temperatures below 133 • C, i.e. below the melting temperature of pure urea. It is worth noting that these materials undergo density change due to solid-liquid phase transition. This might be advantageous in certain configurations [19,20], as it is expected that for normal materials the solid phase has a higher density. For the relevant PCM the density values are: xylitol has solid density of 1.5 g cm −3 and liquid density of 1.35 g cm −3 [27], erythritol has solid density of 1.48 g cm −3 and liquid density of 1.30 g cm −3 [29], erythritol-xylitol mixture has solid density of 1.31 g cm −3 and liquid density of 1.27 g cm −3 [31], erythritol-TME mixture has solid density of 1.26 g cm −3 and liquid density of 1.17 g cm −3 [29], erythritol-urea mixture has solid density of 1.33 g cm −3 and liquid density of 1.28 g cm −3 [26].
Based on the practical applications of our interest, and corresponding requirements for PCM melting range and enthalpy, three binary eutectic mixtures were found in the literature. Each of them contains erythritol, which has a very high latent heat but a too high melting temperature. Figure 1 summarizes graphically the relevant sugar-alcohol-containing PCM found in the literature, both pure and eutectic. The current study compares the behavior of three eutectic mixtures: erythritol-xylitol, erythritol-urea, and erythritol-TME, under the same conditions. Additional aspect of the investigation is to provide visual results which indicate the behavior of these materials at cooling and solidification. In our opinion, such results, scarcely found in the literature, are of interest for the extensive and expanding PCM research community.

Materials and methods
According to the requirements and the literature survey, three different erythritol based eutectic mixtures were identified: erythritol-xylitol, erythritol-urea and erythritol-TME. To achieve reliable results, first each of the pure materials was characterized. Then, the eutectic mixtures were prepared according to predefined recipe. The behavior and properties of the mixtures were studied to determine the best PCM candidate.
Three 10 g samples of eutectic mixtures were prepared. All samples were mixed as powders inside a glass vial, according to the following eutectic ratios (weighed ratios corresponding to erythritol): 25% for erythritol-xylitol, 54.9% for erythritol-urea and 40% for erythritol-TME. The powder mixtures were then melted inside a programmable oven (Memmert UFE) at the temperature of 130 • C for one hour. After melting, the samples were placed on a heated plate with fixed temperature of 140 • C, and a magnetic stirrer was used for about two hours to achieve homogeneous mixing. The samples were cooled at room temperature of about 25 • C. While the erythritol-urea and erythritol-TME mixtures solidified at room temperature to a firm solid bulk, the erythritol-xylitol mixture remained liquid, in accordance with the descriptions in the literature [1]. To enforce solidification on the erythritol-xylitol mixture, erythritol particle seeding was used, as discussed in section 3 below. From the bulk solid, PCM particles were scratched using a lab spatula to create DSC samples.

DSC characterization
DSC was used to measure the latent heat capacity of the pure materials and the erythritol-based eutectic mixtures. Samples of 7-9 mg were placed in a 40 µl aluminum crucible and heated under air flow at a 10 • C min -1 rate. The latent heat was determined in the Mettler-Toledo DSC 823 Star system by numerical integration of area under the peaks that represent phase transition. According to the manufacturer, the uncertainties of the measured temperature and latent heat of fusion using this DSC are ±0.2 • C and ±1%, respectively.

Visual observations of the phase change process
All seven samples (pure materials and mixtures) were heated in the oven (15 centigrade above the DSC melting point) to achieve full melting. The melted PCM were cooled at room temperature of 25 • C and the solidification process was captured using photos.

Thermal properties of the PCM candidates
The DSC scan results for heating of the pure materials can be seen in figure 3. The plots reveal the heat flow to the sample (units of W g −1 ), as a function of the sample temperature. It is expected that a sharp endothermic peak represents the melting temperature of the sample, which is the temperature at which the solid transitions to a liquid. Figure 3(a) reveals that erythritol undergoes melting at 119 • C with latent heat of 335 J g −1 . Xylitol ( figure 3(b)) exhibits melting at 93 • C with latent heat of 241 J g −1 . Urea melted at 134 • C with latent heat of 236 J g −1 , as seen in figure 3(c). Lastly, the DSC scan results for the TME, figure 3(d), present a different trend than the previous scans, as we encounter a solid-solid phase transition at 85 • C. This result is expected for TME, since below 85 • C the material has an ordered and monoclinic structure, whereas above 85 • C the TME transitions to disordered fcc phase-this phase change requires large transition enthalpy and is probably related to the breaking of the hydrogen bonds, see Suenaga et al [32] for in-depth discussion. The solid-liquid phase change occurs at 201 • C with corresponding latent heat of 42 J g −1 . Thus, two peaks appear in figure 3(d). The total latent heat for the two discussed phase transitions is 218 J g −1 . The results of figure 3 are in accordance with the literature presented in the Introduction above.
Since the results of the pure PCM were found to be reliable, DSC measurements were carried out for the eutectic mixtures, as illustrated by figure 4. Figure 4(a) presents the erythritol-xylitol mixture, where melting is obtained at 84 • C with latent heat of 190 J g −1 . While the melting temperature matches the reported data in the literature [22], in the current study latent heat is found to be just 77% of the reported data. A possible reason for this discrepancy is incomplete preceding solidification of the erythritol-xylitol mixture, see section 2 above, although the material seemed like a firm solid. The erythritol-urea mixture shows a melting point of 82 • C, vs 81 • C reported in [26], and the latent heat is 227 J g −1 , which is 92% of the literature value. The third mixture of erythritol-TME (see figure 4(c)) has solid-solid phase transition at 82 • C, followed by melting at 97 • C, and the corresponding latent heat is 83 J g −1 and 115 J g −1 , respectively (total of 198 J g −1 ). These peaks match the phase change occurring for pure TME, as discussed for the results of figure 3(d).
Comparison with the literature shows that the latent heat is 91% of the reported one [29].
The solidification behavior of the eutectic mixtures used in this study was explored first via solidification tests using DSC. Figure 5 presents an example of the solidification results for erythritol-TME eutectic mixture. It is worth noting that the solidification curve presented in figure 5 is the original graph generated by the program of the DSC used in this study. The test was repeated twice, so in the figure one can see two solidification sequences, each of which has two peaks, due to the two phase transitions in TME, as discussed earlier. Two additional curves, shown in figure 5, are part of the heating sequence, like that depicted in figure 4(c). The results reveal that the mixture suffers from supercooling, probably due to the inclusion of erythritol. In the next sub-section we investigate visualization of the solidification process for the explored PCM.

Visualization of the phase change process
Visual observations of the phase change process of the pure erythritol, xylitol, urea and TME were conducted. The results are depicted in figures 6-9. The results are structured in a systematic way, so that part (a) of each figure presents the melted material freshly out of the oven; part (b) illustrates the initial observed solidification; part (c) shows an intermediate stage of the solidification; finally, part (d) presents the completely solidified PCM. It is worth noting that this behavior excludes xylitol, figure 7, as discussion below. Previous studies [22][23][24]26, 29] obtained significant characterization, but lacked such detailed visualization of the phase change process as depicted below. The current results are summarized graphically in figure 10.  Figure 6 demonstrates the solidification process of erythritol. The melted material is perfectly transparent (see figure 6(a)). As the erythritol is cooled below the melting temperature, a first sign of solidification can be observed as a white solid which 'rises' from bottom wall. It takes about 2 min for solidification to begin. Figure 6(d) shows a completely solidified material which has a non-transparent white color appearance.
The cooling of xyliton is unusual in regards to conventional materials, as one can see in figures 7(a)-(c). The melted material is perfectly transparent (a little bit yellowish), and behaves like a 'regular' liquid. When allowed to cool down to the room temperature, after two hours the material did not exhibit any changes, see figure 7(b). However, when the vial was tilted on the side for a substantial time, first no change was apparent and then the material flowed. As a result, its upper interface became non-parallel to the vial bottom. Once the vial was restored to its upright position, this interface remained skewed, as shown in figure 7(c). Therefore, one can conclude that the material exhibited properties of a high-viscosity liquid and did not solidify.
The solidification mechanism of urea is described in figure 8. The solidification is initiated almost immediately, see figure 8(a), and then the material is solidified rapidly next to the lateral walls of the vial, see figures 8(b) and (c). The last stage of the solidification is achieved after 4 min, as shown in figure 8(d). The solid obtained is non-transparent with white color. Figure 9 displays the solidification of TME. As the sample was out, first solid precipitates appeared at the bottom of the vial, with a 'needle'-like shape ( figure 9(a)). After one minute, most of the sample is solidified, see figure 9(b). A complete solid is achieved after 3 min. The solid has a visible texture which might be attributed to non-uniform solidification, see figures 9(c) and (d).
The results presented in figures 6-9 are summarized graphically in figure 10. The figure is structured as bars graph where the x-axis represents different pure PCM and the y-axis is time in minutes; the different colors correspond to solidification stages observed during the visualized solidification tests. Erythritol exhibits the longest solidification time, due to the combination of the supercooling effect with its lowest melting point and largest latent heat among the materials explored. As described above, xylitol does not solidify at all. Urea and TME have a common behavior: both start to solidify almost immediately after being carried out from the oven.       A unique observation is obtained for the erythritol-xylitol eutectic mixture, see figure 11. The sample stays liquid even after a long period of time at a room temperature, as no solidification occurs; this observation is in accordance with the results of [22]. Since the erythritol-xylitol mixture does not solidify under regular conditions, to initiate the solidification process, erythritol crystallite powder was added to the sample as nucleation sites ( figure 11(a)). A large number of particles were seeded manually, thus some of them were attached to the walls of the vial, as seen in the photos of figure 11.
At first, no changes were noted. However, the solidification process was slowly developed, and after 24 h some of the erythritol particles sank in the liquid and small solid spheres were created, see figures 11(b) and (c). On the following day, after 48 h, the solidification front has advanced a little bit, as shown in figure 11(d). The solidification of the mixture took about a week. The solid 'stick' observed in figure 11(c), was created by piercing the liquid with a metal rod.   Figure 12 illustrates the solidification of erythritol-urea eutectic mixture. As the sample is taken out of the oven, initial small spherical solids appear to be homogeneously distributed in the liquid medium ( figure 12(a)). After 15 min, 'lath'-like particles with rectangular shape are formed in the center of the volume ( figure 12(b)). It is visible in figure 12(c) that after 25 min at room temperature, the solidification product is not uniform. In the final observation, after 45 min a complete solid is obtained, but the result also indicates on non-uniform solidification process ( figure 12(d)).
The solidification sequence of erythritol-TME eutectic mixture is shown in figure 13. The sample has the fastest solidification process, when compared with the other eutectic mixtures studied here. The melted material, as it is out of the oven, has a transparent appearance (see figure 13(a)). A first solid is formed after 5 min mainly at the bottom of the vial, see figure 13(b). The solidification front advances from below to the top (figure 13(c)) and after 7 min the solidification process is completed.

Closure
In the present work, thermal processes related to solid-liquid phase-change were investigated for such potentially useful PCM as sugar alcohols, mixtures within this class of materials, and mixtures between sugar alcohols and some other organic materials, carefully chosen based on the existing literature. Specifically, we focused on the materials with solid-liquid phase change within 80 • C-100 • C, which are of considerable interest in some thermal energy storage and thermal management applications.
Consistent with the existing literature, it was found that the use of pure materials of this type is not very practical in the required range, because of their limited choice and their specific disadvantages, including their relatively low latent heat and problems with their crystallization and solidification. On the other hand, we have discerned from the literature three potentially useful binary mixtures of a sugar alcohol (erythritol), one of which was with another sugar alcohol (xylitol) and the two others contained organic materials beyond the sugar alcohols class (urea and TME). In all three cases, it is remarkable that while the components commonly have rather high melting temperatures, their eutectic mixtures underwent the phase transitions in the required range.
Via the measurements, melting temperatures and latent heats were obtained, generally in agreement with the existing literature. The results revealed the melting temperature and latent heat of the mixtures to be: 84 • C and 190 J g −1 for erythritol-xylitol, 82 • C and 227 J g −1 for erythritol-urea. Erythritol-TME has two phase transitions at 82 • C and 97 • C, with total latent heat of 198 J g −1 . Based on the present findings, the erythritol-urea mixture is the best PCM candidate for the melting range within 80 • C-100 • C. Also, it was discovered that each of these mixtures had its own peculiar traits, especially at cooling and solidification. A detailed visualization study of these major processes was undertaken, revealing some previously unreported features and details pertinent to the behavior of these PCM.
These results will be very helpful in the future modeling of melting and solidification of these materials in their role as PCM. Also, it is worth noting that the mixtures considered in this study undergo the solid-liquid phase change at the temperatures much lower than the melting temperatures of their components, like urea and TME. This feature should definitely contribute to the mixture endurance, because there is no need to heat the system to temperatures where their properties might be damaged. This issue will be addressed in a future study.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).