Novel PEG6000–Silica-MWCNTs Shape-Stabilized Composite Phase-Change Materials (ssCPCMs) for Thermal-Energy Storage

This paper describes the preparation of new PEG6000–silica-MWCNTs composites as shape-stabilized phase change materials (ssPCMs) for application in latent heat storage. An innovative method was employed to obtain the new organic–inorganic hybrid materials, in which both a part of the PEG chains, used as the phase change material, and a part of the hydroxyl functionalized multiwall carbon nanotubes (MWCNTs-OH), used as thermo-conductive fillers, were covalently connected by newly formed urethane bonds to the in-situ-generated silica matrix. The study’s main aim was to investigate the optimal amount of PEG6000 that can be added to the fixed sol–gel reaction mixture so that no leakage of PEG occurs after repeated heating–cooling cycles. The findings show that the optimum PEG6000/NCOTEOS molar ratio was 2/1 (~91.5% PEG6000), because both the connected and free PEG chains interacted strongly with the in-situ-generated silica matrix to form a shape-stabilized material while preserving high phase-transition enthalpies (~153 J/G). Morphological and structural findings obtained by SEM, X-ray and Raman techniques indicated a distribution of the silica component in the amorphous phase (~27% for the optimum composition) located among the crystalline lamellae built by the folded chains of the PEG component. This composite maintained good chemical stability after a 450-cycle thermal test and had a good storage efficiency (~84%).


Introduction
In response to several international regulations regarding cost-optimal minimum energy performance requirements, thermal energy storage has emerged as a promising method for energy-efficient use. Thus, energy storage plays a crucial role in saving excess energy and resolving the imbalance between energy supply and demand [1,2]. For renewable energies such as solar and geothermal, energy storage helps to store a huge amount of energy that otherwise can be lost [3][4][5][6]. In solar thermal applications, collectors are employed to capture thermal sun radiations for home and water heating [7,8]. One popular option for the recovery of solar energy has been the use of PCMs as part of heat exchangers [9]. In addition, energy storage can help reduce energy consumption and CO 2 emissions in buildings and industries [10]. Due to their superior thermal storage capabilities, phase change materials (PCMs) have enormous promise for latent thermal energy storage, waste
A mechanical stirrer (STIRRER DLS, VELP Scientifica Srl, Usmate, Italy) was used to homogenize samples TP2-TP4 that were prepared for this study. The water bath was heated on an AREX 6 Heating Magnetic Stirrer (VELP Scientifica Srl, Usmate, Italy) equipped with a VTF Digital Thermoregulator (VELP Scientifica Srl, Usmate, Italy). The resulting material, after drying, was ground by hand, then pressed using a 15 ton manual press (Specac Ltd., Orpington, UK) in a Specac mold (13 mm Evacuable Die and a 13 mm Disk Holder), which is usually used to obtain KBr pellets for FT-IR. The synthesis of these new shape-stabilized phase-change materials, with thermalenergy storage properties, involved three reaction mixtures (A, B, and C, Table 2), whose addition corresponded to the following stages (A, B and C; Scheme 1): The flask containing all four of these ingredients was put into a water bath that was heated to 80 °C. NCOTEOS was added to the flask only after the PEG had completely melted and SDS had fully dissolved. Both PEG6000 and MWCNTs-OH reacted with NCO-TEOS in a coupling reaction that lasted for two hours at 80 °C under 250 rpm mechanical stirring in an inert nitrogen environment. Tao et al. previously reported the useful application of SDS as a surface active agent for a better dispersion of CNTs in PEG [45]. At this moment, we anticipated that PEG6000-Si silica building blocks could be found in the flask, Scheme 1. Bubbling and feeding funnels were introduced: (i) the required amount of PEG 6000 ( Table 2); (ii) the required amount of MWCNTs-OH; (iii) 4 g of surfactant (SDS) required for an improved dispersion of CNT in; and (iv) 1,4-diazabicyclo[2.2.2]octane (DABCO) in a proportion of 1% by weight relative to the amount of PEG 6000 . DABCO has a catalytic role in the coupling reaction between PEG 6000 and MWCNTs-OH with NCOTEOS.
Stage A. Covalent connection of PEG and MWCNTs with the silica co-precursor NCOTEOS (preparation of PEG 6000 -Si and MWCNT-Si). During this stage, a portion of the PEG 6000 chains formed covalent bonds with the silica co-precursor 3-(triethoxysilyl)propyl isocyanate (NCOTEOS). As a result, some PEG chains were able to create covalent connections with the silica network that subsequently formed in the presence of the silica co-precursors mixture (Stage C). This connection was aimed at stopping the PEG from leaking out of the final composite structure during numerous heating-cooling cycles.
At the same time as the formation of the PEG 6000 -Si silica building block, the connection of the OH-functionalized multi-walled carbon nanotubes (MWCNTs-OH) with NCOTEOS occurred in this step, with the goal of preventing their deposition on the bottom of the flask and favoring their homogeneous distribution in the finished PCM composite by covalently bonding the MWCNTs with the in-situ-formed silica network.
Following the reaction between the hydroxyl groups of the two components with the isocyanate functions of NCOTEOS, we obtained PEG 6000 and MWCNTs end-capped with Si(O-CH 2 -CH 3 ) 3 sol-gel reactive functions. The new building blocks, denoted PEG 6000 -Si and MWCNTs-Si, respectively, were to be used for the construction of the three-dimensional (3D) macromolecular network structure that would be produced in situ.
The flask containing all four of these ingredients was put into a water bath that was heated to 80 • C. NCOTEOS was added to the flask only after the PEG had completely melted and SDS had fully dissolved. Both PEG 6000 and MWCNTs-OH reacted with NCOTEOS in a coupling reaction that lasted for two hours at 80 • C under 250 rpm mechanical stirring in an inert nitrogen environment. Tao et al. previously reported the useful application of SDS as a surface active agent for a better dispersion of CNTs in PEG [45]. At this moment, we anticipated that PEG 6000 -Si silica building blocks could be found in the flask, arising from the coupling of PEG 6000 with NCOTEOS, free PEG 6000 , thermoconductive filler coupled with NCOTEOS (MWCNT-Si), and possibly free thermoconductive filler. Combining these components resulted in mixture A.
Stage B. Obtaining inverse emulsion of PEG in n-octane. The mixture A from the flask was combined with 80 mL of n-octane (mixture B) that was preheated to 70 • C. The newly formed mixture was mechanically agitated at 250 rpm while being kept at a temperature of 80 • C under an inert nitrogen environment. Shortly after n-octane was poured over mixture A, phase separation could be observed, revealing a thick black layer and a thin clear layer at the top.
Stage C. In situ synthesis by sol-gel process of the SiO 2 network. Using a feeding funnel, a mixture of 5 g TEOS, 0.567 g GPTMS, 0.327 g MeTMS, and 20 mL EtOH (molar ratio TEOS/GPTMS/MeTMS = 10/1/1) was slowly added to the flask. When the combination of sol-gel precursors dissolved in ethanol was added, the two phases existing in the flask began to homogenize. The presence of ethanol assisted in emulsifying the PEG in noctane. 1.5 mL of NH 4 OH (25%) was added after the mixture in the flask had completely homogenized (for about two minutes) and the pH was adjusted to 7.
The reaction continued under mechanical stirring at 250 rpm, at 80 • C and under nitrogen atmosphere until coagulation appeared or, in the case of sample TP5, for 3 h. Approximately 1 h after the NH 4 OH addition, two phases could be seen to be separated in samples TP2, TP3, and TP4, and the stirring was ended. A dark solid phase was formed at the bottom of the flask that had a gel-like appearance and a liquid clear phase at the top. After the clear phase was eliminated, the resulting gel was put into a polyethylene tray and allowed to air dry for 48 h at room temperature. The TP5 sample remained homogeneous even 3 h after the NH 4 OH addition, without the phase separation observed in the other samples. Because there was an excessive amount of PEG in this sample, the silica network was unable to form properly, which prevented PEG from coagulating. As all samples were prepared at 80 • C, the PEG from the flask existed in the melted state, so all the composites during preparation were soft and could flow, even after coagulation. Immediately after observing the occurrence of the coagulation, and while the obtained materials were still warm and soft, the samples were poured from the flask into polyethylene trays, where, as Polymers 2023, 15, 3022 6 of 23 they cooled, they hardened due to the solidification of the PEG. They were then allowed to dry at room temperature for at least 48 h before different specimens were removed for various characterizations.

Methods
SEM images were recorded with an ESEM-FEI Quanta 200 Scanning electron microscope (FEI Company, Eindhoven, The Netherlands). The working parameters used in the study of the materials through SEM were high vacuum, an LFD detector, and magnification between 50×-5000×. For their morpho-structural characterization, the materials were also examined by scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX), which analyzes the composition of the elements' distinctive X-rays. A field-emission scanning electron microscope, Hitachi TM4000plus II (Hitachi High-Technologies Corp Ltd., Tokyo, Japan), which was connected to an energy-dispersive X-ray spectrometer, was used. EDX was used to evaluate the distribution of C, O, and Si elements along a scan line in the analyzed samples. A BSE detector was employed at an accelerating voltage of 20 kV and a magnitude of 500 X.
Preparing the samples for this study involved taking a small piece of each sample and depositing it on a carbon strip and covering it with a 5 nm layer of Au (through sputter coating). For each sample, both SEM pictures and EDAX spectra were acquired. X-ray diffractograms were recorded within the range of 5 • to 70 • (2θ) with a Rigaku diffractometer, type Ultima IV (Rigaku Corporation, Tokyo, Japan), equipped with a Cu tube operating at 40 kV and 30 mA. The crystallinity degree of the PEG 6000 component was obtained by fitting the defined X-ray diffraction peaks over the range of 16 to 30 • (2θ), as described elsewhere [46].
UV-Raman spectra were collected by means of a LabRam HR800 spectrometer (Horiba France SAS, Palaiseau, France) equipped with a 325 nm laser line (Kimmon Koha Co., Ltd., Tokyo, Japan), a grating of 2400 lines, and a 40×/0.47 NUV objective (Olympus Corporation, Tokyo, Japan). The Edge filter prevented recording Raman signal at low wavenumbers. Silicon was used for peak position calibration. The laser power (P L ) was adjusted to avoid sample heating and degradation (P L = 5, 1.5 and 0.5 mW).
FTIR spectra were obtained with a Bruker Tensor 37 spectrometer (Bruker Instruments, Woodstock, NY, USA). The analysis parameters were 32 scans with a resolution of 4 cm −1 in the wavelength range of 4000 to 600 cm −1 . Samples were analyzed using the attenuated total reflection (ATR) module with a Golden Gate diamond unit (400-4000 cm −1 ).
The thermal stability of the composite material was studied in the temperature range of 0 to 700 • C. TGA analysis was performed using a Q5000 apparatus from TA Instruments (New Castle, DE, USA). The mass of the samples varied between 10 and 15 mg. These were placed in 100 µL platinum pans. All TGA experiments were conducted in a nitrogen environment with a purge flow rate of 40.0 mL/min. The TGA program consisted of heating up to 710 • C, with a 10 • C/min increment, followed by a 5 min. isotherm. DSC measurements were performed using a DSC Q2000 apparatus from TA Instruments (New Castle, DE, USA). DSC runs were carried out on powders of about 10 mg encapsulated in Tzero ® aluminum pans. The purge gas used was helium, at a flow rate of 5 mL/min. The DSC method used to analyze the thermal properties of the selected samples was a dynamic method, consisting of a cooling process with a rate of 10 • C/minute up to −10 • C, an isotherm of 3 min followed by a heating with 10 • C/min up to 80 • C, and another 3 min isotherm for sample stabilization. This run was repeated three times (3 heating-cooling cycles) for all samples and 450 times for sample TP4. Indium was used as the reference for calibration of melting temperature (extrapolated onset melting temperature of 156.49 ± 0.01 • C, obtained with onset values recorded at various heating rates of 2, 5, 10, and 20 • C) and enthalpy (28.27 ± 0.07 J/g).
The analysis of the thermo-physical properties (thermal diffusivity, specific heat capacity, effusivity, and thermal conductivity) were carried out by the "flash method", according to ASTM E 1461 [47], using an Netzsch LFA 447 Nanoflash device (Netzsch, Selb, Germany).
The LFA equipment allowed the direct measurement of the thermal diffusivity, while the specific heat of the materials was determined by a differential method, using a reference sample. Two mathematical models, the Adiabatic and Cowan models, both with linear baseline correction, were used for the analysis of the data. The analysis included the possibility of heat loss from both the front and back faces of the sample. It was also assumed that the energy pulse was rectangular in shape and of relatively short duration. For the specific heat measurement, the magnitude of the temperature rise of an unknown sample was compared to that of the pyroceram standard. The device could perform tests in the temperature range of 25 to 300 • C, for cylindrical samples with a diameter of 12.7 mm and a thickness of 2-3 mm. A high-performance xenon lamp was used as a source of radiation energy, and the irradiation time on the front face of the sample was 0.18 ms. A minimum of five measurements were performed for each sample at a temperature of 25 • C. The increase in temperature on the other surface of the sample was measured using an infrared (IR) InSb detector, cooled with liquid nitrogen.
Thermo-mechanical properties were analyzed with a DMA Q800 V20 instrument (TA Instruments, New Castle, DE, USA). The working conditions were a DMA controlled force mode, the stress-strain method, air-bearing gas, and air.

Leakage Test
After 48 h of drying, both ground and cylinder-shaped samples were placed on circular filter papers with a diameter of 10 cm. Images of the samples taken prior to the thermal test are displayed in the upper part of Figure 1. In order to determine whether PEG 6000 flowed from the hybrid material in which it was integrated and whether it moistened the filter paper, the samples were put in an oven and heated for two hours at a temperature of 85 • C (far above the melting point of PEG 6000 , which is 61.4 • C, as shown in Table 3).

Leakage Test
After 48 h of drying, both ground and cylinder-shaped samples were placed on circular filter papers with a diameter of 10 cm. Images of the samples taken prior to the thermal test are displayed in the upper part of Figure 1. In order to determine whether PEG6000 flowed from the hybrid material in which it was integrated and whether it moistened the filter paper, the samples were put in an oven and heated for two hours at a temperature of 85 °C (far above the melting point of PEG6000, which is 61.4 °C, as shown in Table 3).
where ∆H 0 of 197 kJ/Kg corresponds to latent heat of fully crystallized PEG [48]. 2 XRD = percent of crystallized PEG in PCM, calculated from XRD analyses. 3 ∆T = subcooling range. The standard uncertainty for melting temperature and enthalpy were estimated to be 0.01 • C and 0.02·∆H m from the range of 25 to 75 • C.
As can be seen from Figure 1a (images collected after leakage test), PEG leakage occurred for none of the ground samples except for the TP5 and TP15 composites. In the case of the TP5 sample (which contained the highest percentage of PEG 6000 ), a slight wetting of the filter paper was visible. This was worsened in the case of PT15, with the highest content of MWCNTs-OH ( Figure 1b). Instead, the carbon nanotubes-free sample, PT0, easily passed the leakage test.
In addition to the leakage test conducted on ground samples, we also performed leakage tests on cylindrical specimens. These specimens were heated in an oven for 2 h at 80 • C. The time allocated to this test was sufficient to allow us to observe the leakage and shape deformation of the two samples with low shape-stabilized performances (TP5 and TP15). For the other samples (TP2, TP3, TP4, and TP0), no leakage or shape deformation occurred. In addition, for samples TP5 and TP15, a strong tendency to impregnate the filter paper on which they were placed during testing and to stick to it after cooling was observed. In contrast, samples TP2, TP3, TP4, and TP0 did not wet the paper and could be easily detached. However, the cylinder made from the TP2 composite was observed to be very brittle, especially after heating, thus limiting the usability of this composite. This friability was also confirmed by DMA analysis, as discussed below. Both the TP4 and TP3 samples exhibited very good mechanical and shape-stabilized performances, but after the evaluation of the thermophysical properties, we selected TP4 to check its long-term reliability (via a cycling test), because it exhibited the highest transition enthalpy (i.e., the highest latent-heat storage capacity).

Evaluation of the Thermal Energy Storage Capacity for the Obtained Materials
Differential scanning calorimetry (DSC) was used to assess thermal performances, e.g., the enthalpy and temperature of phase transitions, of pristine and PEG 6000 -based materials that had heat-storage properties.
The DSC curves, the melting and crystallization temperatures, and the melting and crystallization enthalpies for the PEG 6000 and TP(0/2/3/4/5/15) samples are displayed in Figure 2 and Table 3, respectively. PEG 6000 was selected as the phase-change material for this work because of its high value of melting and solidification enthalpies (Table 3)  terials that had heat-storage properties.
The DSC curves, the melting and crystallization temperatures, and the melting and crystallization enthalpies for the PEG6000 and TP(0/2/3/4/5/15) samples are displayed in Figure 2 and Table 3, respectively. PEG6000 was selected as the phase-change material for this work because of its high value of melting and solidification enthalpies (Table 3) and its phase-transition temperatures of Tm = 61.4 °C and Ts = 42.7 °C (Figure 2a), which were suitable for various applications in thermal-energy storage.  Single asymmetric melting and solidification effects were obtained for composite materials, in contrast with the pristine PEG 6000 with multiple-melting effects. Decreases in the melting temperature of the PEG 6000 component in the TP(0/2/3/4/15) composites might have been due to three factors: (i) covalently and weak interaction with SiO 2 , (ii) PEG confinement, and (iii) thermal-conductivity modification [49,50]. Even weak PEG-SiO 2 interaction through capillary forces, hydrogen bonding, and surface forces causes a lowering of melting points [49]. The subcooling (∆T values in Table 3) of the PEG component (calculated as the difference between the melting and solidification temperatures), which is detrimental for long-lasting heat storage purposes, was lowered for the composite materials obtained in this study, except for the MWCNT-free sample, TP0. The competition of these two processes (the "confining effect" [51] and covalent bonding), which restricted the growth to the stable phase (crystal) of the PEG, can explain the different behavior of the samples when the PEG content increased.
According to the high enthalpies of melting and solidification, exceeding 90 J/g, a reasonable amount of heat could be stored and released by these materials. However, these values were lower than the theoretical ones. Additionally, the PEG 6000 degree of coupling to the inorganic network, which is required to stabilize the shape of the material during the subsequent melting-solidification cycles, was partly to blame for the decrease in the enthalpy values recorded for the composite materials, in comparison to the pristine PEG. A relative enthalpy-efficiency coefficient was introduced to describe thermal energy storage capacity, according to the following equation [52]: where λ m = phase-transition capacity for the heating stage (the relative enthalpy efficiency); ∆H m = melting enthalpy of ssPCM; ∆H mPEG6000 = melting enthalpy of pure PEG 6000 ; w PEG6000 = percentage of PEG 6000 in the final material. The enthalpy-efficiency coefficient was greatly diminished (65.83% and 66.77%) for the TP2 sample, with the lowest PEG 6000 content. This difference was determined by the different interaction of PEG within the silica framework [40,52] and the crystallization behavior of the PEG chains, i.e., the crystallization degree and arrangement of the chains [46,[49][50][51][52][53][54]. The enthalpy-calculated crystallization degree, χ c (see Table 3) showed the smallest values for the TP2 (67.17%) and TP15 (78.87%) composites. In addition, the crystallization degree was calculated by integration of the X-ray patterns as presented in the following results.

Structure and Morphology of the Semicrystalline PEG 6000 -Based Composites
The n-folded (n stands for 0, 1, and 2) chains of semicrystalline PEG 6000 form lamellae and, further, build spherulites [49,51]. The diffractograms of the TP(0/2/3/4/5/15) composites shown in Figure 3 are dominated by sharp peaks that are assignable to the crystalline phase of the PEG 6000 component, with the two main peaks at about 19.13 and 23.36 • and with an amorphous halo, due to amorphous counterpart of PEG and silica. The calculated d-inter-reticular spacing values of the TP0, TP4, and TP15 samples prepared with the same PEG6000/NCOTEOS ratio and variable amounts of MWCNTs-OH were comparable to those reported in the previous study for the PEG6000/epoxy composite [46] of 0.46 (d120), 0.38 (d032), 0.34 (d024), and 0.33 nm (d131) corresponding to X-ray peaks at ~19, 23, 26, and 27°, respectively.
Analogous to the PEG6000/epoxy composite (70/30), the more intense peak of PEG6000 at ~27°, corresponding to a larger crystallite size (24.41 nm) in a direction perpendicular to the (131) plane [46], was diminished for the TP(0/4/15) composites, very likely due to restrained thickening of the PEG lamellae under silica and/or filler action.
Wider and less-intense peaks were recorded for TP2 and TP15 and, hence, their crystallinity degree (see Table 3) was the lowest among the composites presented here. Good agreement between the DSC and XRD-derived crystallinity was obtained for all composites in this work.
The SEM micrographs in Figure 4 show different morphology, in accordance with the PEG6000/NCOTEOS ratio (see Table 1) and the MWCNTs-OH amount used in preparation of the PT(0/2/3/4/5/15) composites. Lamellae morphology is easier to depict in the Figures 4a and S1 from the TP2 to TP5. TP5 showed richer folds for pristine PEG6000 that form spherulites [51]. Slightly different chain folding and/or lamellae wrapping was noticeable for the TP3 and TP4 samples.
The MWCNTs-OH are densely entangled in Figure 4b. Increasing MWCNTs-OH content at the same PEG6000/NCOTEOS ratio used in preparation of the PT(0/4/15) samples triggered the formation of globular structures in the case of PT15 (Figure 4b).
According to the EDX data for the in-line scan shown in Figure S1, Si distributions into interlamellar space, where there is usually an amorphous phase of PEG, are present The calculated d-inter-reticular spacing values of the TP0, TP4, and TP15 samples prepared with the same PEG 6000 /NCOTEOS ratio and variable amounts of MWCNTs-OH were comparable to those reported in the previous study for the PEG 6000 /epoxy composite [46] of 0.46 (d 120 ), 0.38 (d 032 ), 0.34 (d 024 ), and 0.33 nm (d 131 ) corresponding to X-ray peaks at~19, 23, 26, and 27 • , respectively.
Analogous to the PEG 6000 /epoxy composite (70/30), the more intense peak of PEG 6000 at~27 • , corresponding to a larger crystallite size (24.41 nm) in a direction perpendicular to the (131) plane [46], was diminished for the TP(0/4/15) composites, very likely due to restrained thickening of the PEG lamellae under silica and/or filler action.
Wider and less-intense peaks were recorded for TP2 and TP15 and, hence, their crystallinity degree (see Table 3) was the lowest among the composites presented here. Good agreement between the DSC and XRD-derived crystallinity was obtained for all composites in this work.
The SEM micrographs in Figure 4 show different morphology, in accordance with the PEG 6000 /NCOTEOS ratio (see Table 1) and the MWCNTs-OH amount used in preparation of the PT(0/2/3/4/5/15) composites. Lamellae morphology is easier to depict in the Figures 4a and S1 from the TP2 to TP5. TP5 showed richer folds for pristine PEG 6000 that form spherulites [53]. Slightly different chain folding and/or lamellae wrapping was noticeable for the TP3 and TP4 samples.

FT-IR and UV-Raman Analysis
Since the share of the inorganic silica network in the final composites is small (below 25%), the FTIR spectra of all samples shown in Figure 5 were dominated by the signals coming from PEG. Small differences between the spectra were observed only in the range of 1500 to 1800 cm −1 (see the inset of Figures 5 and S2 and Table S1).
Thus, the peaks at ~1719 cm −1 (C=O vibration) and ~1530 cm −1 (N-H vibration), which are attributed to the newly created urethane bonds (-OCONH-) from isocyanate group (O=C=N-) of the silane coupling agent (NCOTEOS) and the hydroxyl group of PEG, HO-(CH2-CH2-O)n- [53] and/or MWCNTs-OH [54] (radical,R, in the reaction bellow), were clearly visible, especially for the samples in which the PEG weight was lower. These signals appeared for all TP(2/3/4/5) samples. The functionalization of the PEG chains and/or the MWCNTs with triethoxyxilane groups [53] represented a first step in the further solgel synthesis of the shape-stabilized organic-inorganic hybrids. The ATR FTIR was a very suitable technique in monitoring this step [53]. No isocyanate vibrations at about 2270 cm −1 [53] are depictable in Figure 5. The MWCNTs-OH are densely entangled in Figure 4b. Increasing MWCNTs-OH content at the same PEG 6000 /NCOTEOS ratio used in preparation of the PT(0/4/15) samples triggered the formation of globular structures in the case of PT15 (Figure 4b).
According to the EDX data for the in-line scan shown in Figure S1, Si distributions into interlamellar space, where there is usually an amorphous phase of PEG, are present in all composites, except for TP15. A smoother surface, without obvious lamellar morphology, was observed for TP15, as shown in Figure S1b. The melting of the PEG component due to the SEM electron gun was not ruled out in the latter case. An uneven pore structure disrupted the lamellar structure of the PEG component.

FT-IR and UV-Raman Analysis
Since the share of the inorganic silica network in the final composites is small (below 25%), the FTIR spectra of all samples shown in Figure 5 were dominated by the signals coming from PEG. Small differences between the spectra were observed only in the range of 1500 to 1800 cm −1 (see the inset of Figures 5 and S2 and Table S1). Two signals at ~1760 cm −1 and 1664 cm −1 (both carbonyl vibrations), observed for samples with richer PEG content (Figure 5a), might have indicated partial oxidation of the organic components. It was surprising that neither samples TP2 and TP3 nor pure PEG6000 exhibited this signal. Their presence also indicated formation beside the urethane bonds of amide (1664 cm −1 ) and anhydride bonds, (RC(O))2O (1760 cm −1 and 1814 cm −1 , respectively). In addition, the band at ~1590 cm −1 could be assigned to stretching vibrations of the ester groups (O-C=O). The signal at 3431 cm −1 , which denoted the presence of unreacted hydroxyl groups, was readily discernible for PEG-richer samples, while the peak at 2882 cm −1 was present for all samples and was assigned to C−H stretching from the PEG chains. The strongest ether vibrations, at about 1100 cm −1 [55], originated from PEG6000 ( Figure 5). Three main bands were expected for CNTs at ~3430, 1700, and 1180 cm −1 (OH, C=O in carboxyl, and C-O bonds, respectively) [56]. The Si-O-Si symmetric stretching vibrations [34] were noticeable as a shoulder of the TP15 spectrum at about 800 cm −1 , as shown in Figure 5b.
UV-Raman spectroscopy is a very sensitive technique for the investigation of the sample surface, especially for functionalized sp 2 carbon materials. The thermal and electrical behavior of the CNTs is highly influenced by their interaction with the embedding matrix. The UV-Raman spectra of all composites prepared in this work, as shown in Thus, the peaks at~1719 cm −1 (C=O vibration) and~1530 cm −1 (N-H vibration), which are attributed to the newly created urethane bonds (-OCONH-) from isocyanate group (O=C=N-) of the silane coupling agent (NCOTEOS) and the hydroxyl group of PEG, HO-(CH 2 -CH 2 -O) n - [55] and/or MWCNTs-OH [56] (radical, R, in the reaction bellow), were clearly visible, especially for the samples in which the PEG weight was lower. These signals appeared for all TP(2/3/4/5) samples. The functionalization of the PEG chains and/or the MWCNTs with triethoxyxilane groups [55] represented a first step in the further sol-gel synthesis of the shape-stabilized organic-inorganic hybrids. The ATR FTIR was a very suitable technique in monitoring this step [55]. No isocyanate vibrations at about 2270 cm −1 [55] are depictable in Figure 5.
Two signals at~1760 cm −1 and 1664 cm −1 (both carbonyl vibrations), observed for samples with richer PEG content (Figure 5a), might have indicated partial oxidation of the organic components. It was surprising that neither samples TP2 and TP3 nor pure PEG 6000 exhibited this signal. Their presence also indicated formation beside the urethane bonds of amide (1664 cm −1 ) and anhydride bonds, (RC(O)) 2 O (1760 cm −1 and 1814 cm −1 , respectively). In addition, the band at~1590 cm −1 could be assigned to stretching vibrations of the ester groups (O-C=O). The signal at 3431 cm −1 , which denoted the presence of unreacted hydroxyl groups, was readily discernible for PEG-richer samples, while the peak at 2882 cm −1 was present for all samples and was assigned to C−H stretching from the PEG chains. The strongest ether vibrations, at about 1100 cm −1 [57], originated from PEG 6000 ( Figure 5). Three main bands were expected for CNTs at~3430, 1700, and 1180 cm −1 (OH, C=O in carboxyl, and C-O bonds, respectively) [58]. The Si-O-Si symmetric stretching vibrations [34] were noticeable as a shoulder of the TP15 spectrum at about 800 cm −1 , as shown in Figure 5b.
UV-Raman spectroscopy is a very sensitive technique for the investigation of the sample surface, especially for functionalized sp 2 carbon materials. The thermal and electrical behavior of the CNTs is highly influenced by their interaction with the embedding matrix. The UV-Raman spectra of all composites prepared in this work, as shown in Figure 6, were dominated by the spectral features of the PEG 6000 [46,59]. No bands for MWCNTs [60] are shown in Figure 6b except for the G band at~1580 cm −1 (in-plane vibrations of the sp 2 carbon atoms) and the 708 cm −1 band with unknown origin [60] in the richer MWCNT containing sample TP15. The defect band, D, of the MWCNTs, located at about 1417 cm −1 , was masked by the stronger bands of PEG 6000 . Unlike the G band of the sp 2 hybridized carbon materials, the position and intensity of the D band (the breathing mode of the sp 2 -carbon atoms in the rings) are dependent on the wavelength of the exciting laser [60,61]. In addition, the defect band, D 1 (4-SiO 4 rings), was noticeable in the Raman spectrum of the MWCNT-rich sample TP15 [62]. Unfortunately, the band at about 530 cm −1 had multiple assignments, e.g., C-C-O bending vibrations of the PEG chains and -C-Si-O-bonds [55]. Due to the distinct spectral features of the TP15 composite, Raman spectra (Figure 6b) of the samples containing NCOTEOS and MWCNTs-OH in a PEG 6000 environment (see Scheme 1 part A) were collected at various time spans (5, 30, 60, 120, and 180 min). In addition, this sample had the same PEG 6000 /NCOTEOS ratio as the TP4 sample (Table 1) Intriguingly, during the first 60 min, the G band of the MWCNT-OH underwent broadening. Thus, this band accommodated three bands (1562, 1609, and 1660 cm −1 bands in Figure S2 with full width at the half maximum, FWHM, of 79, 80, and 125 cm −1 , respectively). It seems that the silane coupling agent reacted first with hydroxyl groups of the MWCNTs-OH. This explains the poor PEG leakage behavior in TP15. Conversely, the NCOTEOS for the TP5 with 0.02% MWCNTs-OH load was insufficient to covalently bind a large quantity of PEG 6000 (PEG 6000 /NCOTEOS = 3/1) for a shape-stabilized composite. The position of the stretching and bending modes of the hydroxyl groups at 2691 cm −1 and 1692 cm −1 (PEG 6000 ) [46,63] were not affected during the NCOTEOS coupling (Figure 6b). The covalent and non-covalent interaction of CNTs with the matrix could be identified by some Raman spectral features, such as the position of the D and G bands and their intensity ratio, I D /I G . This ratio increased from 0.14 in the case of pristine MWCNTs-OH to 0.99 for PT15-60 min. ( Figure S3). In the latter sample, the intensity of the 840 and 860 cm −1 bands did not correspond to the solid PEG 6000 .
Polymers 2023, 15, x FOR PEER REVIEW 14 of 24 Figure 6, were dominated by the spectral features of the PEG6000 [46,57]. No bands for MWCNTs [58] are shown in Figure 6b except for the G band at ~1580 cm −1 (in-plane vibrations of the sp 2 carbon atoms) and the 708 cm −1 band with unknown origin [58] in the richer MWCNT containing sample TP15. The defect band, D, of the MWCNTs, located at about 1417 cm −1 , was masked by the stronger bands of PEG6000. Unlike the G band of the sp 2 hybridized carbon materials, the position and intensity of the D band (the breathing mode of the sp 2 -carbon atoms in the rings) are dependent on the wavelength of the exciting laser [58,59]. In addition, the defect band, D1 (4-SiO4 rings), was noticeable in the Raman spectrum of the MWCNT-rich sample TP15 [60]. Unfortunately, the band at about 530 cm −1 had multiple assignments, e.g., C-C-O bending vibrations of the PEG chains and -C-Si-O-bonds [54]. Due to the distinct spectral features of the TP15 composite, Raman spectra (Figure 6b) of the samples containing NCOTEOS and MWCNTs-OH in a PEG6000 environment (see Scheme 1 part A) were collected at various time spans (5, 30, 60, 120, and 180 min). In addition, this sample had the same PEG6000/NCOTEOS ratio as the TP4 sample (Table 1) Intriguingly, during the first 60 min, the G band of the MWCNT-OH underwent broadening. Thus, this band accommodated three bands (1562, 1609, and 1660 cm −1 bands in Figure S2 with full width at the half maximum, FWHM, of 79, 80, and 125 cm −1 , respectively). It seems that the silane coupling agent reacted first with hydroxyl groups of the MWCNTs-OH. This explains the poor PEG leakage behavior in TP15. Conversely, the NCOTEOS for the TP5 with 0.02% MWCNTs-OH load was insufficient to covalently bind a large quantity of PEG6000 (PEG6000/NCOTEOS = 3/1) for a shape-stabilized composite. The position of the stretching and bending modes of the hydroxyl groups at 2691 cm −1 and 1692 cm −1 (PEG6000) [46,61] were not affected during the NCOTEOS coupling (Figure 6b). The covalent and non-covalent interaction of CNTs with the matrix could be identified by some Raman spectral features, such as the position of the D and G bands and their intensity ratio, ID/IG. This ratio increased from 0.14 in the case of pristine MWCNTs-OH to 0.99 for PT15-60 min. ( Figure S3). In the latter sample, the intensity of the 840 and 860 cm −1 bands did not correspond to the solid PEG6000. The shape and intensity recovery for the crystallinity band at 844 cm −1 of PEG6000, attributable to the stretching vibrations of the C-O bonds and the rocking vibrations of the CH2 [58], took place in the sample collected after 120 min of reaction. Heating and lowering of the molecular weight of PEGs were two causes of the diminishing intensity of the  The shape and intensity recovery for the crystallinity band at 844 cm −1 of PEG 6000 , attributable to the stretching vibrations of the C-O bonds and the rocking vibrations of the CH 2 [59], took place in the sample collected after 120 min of reaction. Heating and lowering of the molecular weight of PEGs were two causes of the diminishing intensity of the 844 and 863 cm −1 bands. The spectra for the PT15-120 min and 180 min samples were alike, meaning that the process was complete in two hours.

Monitoring the Thermal Stability of the ssCPCMs
TGA and DTG analysis of the TP(0/2/3/4/5/15) composites, as shown in Figure 7, enabled the obtaining of the thermal decomposition range and the temperatures of the maximum decomposition. A slight improvement in the weight was noticeable (Figure 7) for the TP(2/3/4/5/15) composites' different content of PEG integrated in the silica matrix, in contrast with pure PEG 6000 . Different thermal degradation intervals are visible on the TGA curves. According to the DTG curves shown in Figure 7, the composites with the lowest PEG 6000 content, e.g., TP2 and TP3, started the degradation process at 170.4 and 179 • C, respectively. For the other materials, the starting degradation temperature increased in succession: TP15 (217.8 • C) < PEG 6000 (223.5 • C) < TP4 (226.3 • C) < TP0 (230.9 • C) < TP5 (217.8 • C). The first DTG peak was below 200 • C ( Figure 7a) and was due to removal of physically adsorbed water and alcohols [39,64] generated by completion of the hydrolysis and by condensation reactions to obtaining the silica.  Figure 7, enabled the obtaining of the thermal decomposition range and the temperatures of the maximum decomposition. A slight improvement in the weight was noticeable (Figure 7) for the TP(2/3/4/5/15) composites' different content of PEG integrated in the silica matrix, in contrast with pure PEG6000. Different thermal degradation intervals are visible on the TGA curves. According to the DTG curves shown in Figure 7, the composites with the lowest PEG6000 content, e.g., TP2 and TP3, started the degradation process at 170.4 and 179 °C, respectively. For the other materials, the starting degradation temperature increased in succession: TP15 (217.8 °C) < PEG6000 (223.5 °C) < TP4 (226.3 °C) < TP0 (230.9 °C) < TP5 (217.8 °C). The first DTG peak was below 200 °C ( Figure 7a) and was due to removal of physically adsorbed water and alcohols [39,62] generated by completion of the hydrolysis and by condensation reactions to obtaining the silica. Therefore, the mass loss in the first interval for samples TP2 and TP3, in which PEG6000 was present in only a percentage of 75% and 85%, respectively, was 16-18% (Table 4). A relatively small amount of mass was lost (0.4-0.7%) in samples where the polyethylene glycol component predominated (TP4-91.5% and TP5-94%). In the range of 215 to 260 °C, hydrolysis continued with the formation and evapora- Therefore, the mass loss in the first interval for samples TP2 and TP3, in which PEG 6000 was present in only a percentage of 75% and 85%, respectively, was 16-18% (Table 4). A relatively small amount of mass was lost (0.4-0.7%) in samples where the polyethylene glycol component predominated (TP4-91.5% and TP5-94%).
In the range of 215 to 260 • C, hydrolysis continued with the formation and evaporation of ethanol and methanol and condensation occurred with removal of water. However, the dissociation of the PEG-inorganic network urethane bonds, which was accompanied by the release of carbon dioxide and isocyanate, was the primary cause of the mass loss during this thermal interval [64,65].  Between 260 and 450 • C, the largest mass loss (58-97%) took place. Pielichowski et al. [66] reported a single step for thermal degradation of polyethylene oxide, at about 400 • C, as was noticed for the DTG curves of the rich-PEG 6000 samples, e.g., TP(0/4/5/15). Chemisorbed water from the Si-OH groups can be lost in the same stage [39]. For samples TP2 and TP3, the contribution of the organic functions of the silica co-precursors (having shorter molecular chains than PEG) to the mass loss in this range was significant. In addition, the lower crystallinity degree of the TP2 and TP3 samples underscored that the amorphous phase amount of semicrystalline PEG was bigger. Hence, they were prone to lower thermal stability, when compared to samples TP4 and TP5.
Above 450 • C, partial oxidation of MWCNTs and a possible degradation of residual Si-OH occurred. The differences that were observed between the theoretical and experimental values of the inorganic residues (i.e., silica, SiO 2 ) can be explained by the inhomogeneity of the samples taken for the thermogravimetric analysis. They could be extracted from the PEG-rich or PEG-poor regions, which caused a small difference between the experimentally determined weight of the inorganic component and the theoretically predicted one. The sample with the largest discrepancy between the value of the inorganic residue as found experimentally and the value as calculated theoretically was the TP2 sample.
3.6. Evaluation of Thermal Diffusivity, Specific Heat, Thermal Conductivity, and Effusivity for PEG 6000 -Silica-MWCNTs PCM Composites The thermophysical properties of the PEG 6000 -containing materials were dependent on composition, microstructure (crystalline/amorphous phases, chain flexibility/rigidity, etc. [63]), homogeneity and density of the material, and, last but not least, the processing route.
It was already known that silica is a poor thermal conductor. Therefore, with its incorporation into the in-situ-generated silica matrix, PEG 6000 s thermophysical characteristics were found to diminish, as shown in Table 5 and Figure 8.
Moreover, temperature, CNT loading, and sizes (area/volume ratio) were few major parameters influencing thermal conductivity and, hence, heat discharging in functionalized CNTs/PEG materials [67]. However, the TP(2/3/15) encountered lower thermal conductivity values than TP0-i.e., the MWCNTs-free sample. In addition, the compacted MWCNTs-OH had a rather small value in comparison with the value for longitudinal direction [30]. The misoriented MWCNTs, partially covalently linked, did not seem to cause an increase in thermal conductivity. Among the shape-stabilized composites presented here, the TP4 sample had the biggest thermal conductivity, which was a rather low value. Enhancement of the thermal conductivity of the shape stabilized PEG 6000 -SiO 2 -MWCNTs-OH composites remained as a problem to solve in future work.

Sample Code
PEG 6000 and TP(2/3). A linear relationship between the amount of PEG in these PCM composite and effusivity existed. Thermal effusivity influences thermal stresses and strains during transient heat conduction and plays a key role in thermal fatigue and thermal shock. The dispersion and concentration of CNTs also influence the C p values of the PEG/CNTs composites [67]. However, the Cp value of TP4, shown in Table 5 In order to carry out the DMA analysis, all the samples were pressed in the form of disks in an FTIR mold under high pressure.
Sample TP4 was a homogeneous composite material, in which the presence of the inorganic silica network helped to preserve mechanical properties comparable to those of pure PEG. In addition, PEG was present in a sufficiently large amount to allow the formation of the crystalline component in a weight close to that of pure PEG. The best mechanical properties were presented by the TP3 sample, due to the slightly higher weight of the inorganic network in the final material, compared to that of the TP4 sample ( Figure 9). However, the presence of silica in a higher percentage affected the other properties of interest for thermal-energy storage (thermal conductivity, effusivity, and transition enthalpies). In contrast, sample TP5 exhibited weaker mechanical properties than TP4 because of an insufficient inorganic network to reinforce the material.

Evaluation of the Mechanical Properties of Materials with Stabilized-Shape and Thermal-Energy Storage Properties (Dynamic Mechanical Analysis (DMA))
In order to carry out the DMA analysis, all the samples were pressed in the form of disks in an FTIR mold under high pressure.
Sample TP4 was a homogeneous composite material, in which the presence of the inorganic silica network helped to preserve mechanical properties comparable to those of pure PEG. In addition, PEG was present in a sufficiently large amount to allow the formation of the crystalline component in a weight close to that of pure PEG. The best mechanical properties were presented by the TP3 sample, due to the slightly higher weight of the inorganic network in the final material, compared to that of the TP4 sample ( Figure  9). However, the presence of silica in a higher percentage affected the other properties of interest for thermal-energy storage (thermal conductivity, effusivity, and transition enthalpies). In contrast, sample TP5 exhibited weaker mechanical properties than TP4 because of an insufficient inorganic network to reinforce the material.
In turn, sample TP2 had too much inorganic component, which led to an inhomogeneous distribution of PEG and silica in the composite material and a lack of adhesion between these components, phenomena that made the sample crumbly after pressing. This sample, therefore, presented the weakest mechanical properties.

Thermal Reliability and Reusability
For sample TP4, with a PEG6000/NCOTEOS = 2/1 molar ratio, it was shown above that the material retains its shape after exposure to high temperature (~80 °C for 2 h). At the same time, the values of the melting and solidification enthalpies maintain high values (146.1 J/g for the melting stage and 145.1 J/g for the solidification stage, respectively). These thermal properties make the TP4 material a very good candidate for thermal-energy storage applications. As a result, this sample was selected for the cycling test, which provides information about the material's long-term behavior after repeated uses.
The DSC method was used to assess the long-term stability of TP4 ssPCM, and 450 repeated melting-solidification cycles of were performed. The third and 450th heatingcooling cycles ( Figure 10) were utilized to estimate the changes in the sample's thermal In turn, sample TP2 had too much inorganic component, which led to an inhomogeneous distribution of PEG and silica in the composite material and a lack of adhesion between these components, phenomena that made the sample crumbly after pressing. This sample, therefore, presented the weakest mechanical properties.

Thermal Reliability and Reusability
For sample TP4, with a PEG 6000 /NCOTEOS = 2/1 molar ratio, it was shown above that the material retains its shape after exposure to high temperature (~80 • C for 2 h). At the same time, the values of the melting and solidification enthalpies maintain high values (146.1 J/g for the melting stage and 145.1 J/g for the solidification stage, respectively).
These thermal properties make the TP4 material a very good candidate for thermal-energy storage applications. As a result, this sample was selected for the cycling test, which provides information about the material's long-term behavior after repeated uses.
The DSC method was used to assess the long-term stability of TP4 ssPCM, and 450 repeated melting-solidification cycles of were performed. The third and 450th heating-cooling cycles ( Figure 10) were utilized to estimate the changes in the sample's thermal characteristics between the test's initial moment and its ending. We aimed to eliminate the thermal history of the sample; therefore, the first two heating-cooling cycles were not considered. In addition, Figure S5 (Supplementary Materials) in the supporting information provides a visual representation of an overlay representing each of the 450 heating-cooling cycles. characteristics between the test's initial moment and its ending. We aimed to eliminate the thermal history of the sample; therefore, the first two heating-cooling cycles were not considered. In addition, Figure S5 (Supplementary Materials) in the supporting information provides a visual representation of an overlay representing each of the 450 heating-cooling cycles. The melting temperatures (Tmo and Tmp) and the crystallization temperatures (Tfo and Tfp) recorded for sample TP4, as shown in Table 6, were very close after 450 melting-crystallization cycles. As expected, after 450 consecutive heating-cooling cycles, the phasetransition enthalpies for the melting stage (Hm) and the cooling stage (Hf), respectively, showed slight decreases (8 J/g and 7 J/g, respectively). Coefficients ƞm and ƞf, which represent the magnitude of change in the phase-transition enthalpies as a result of the cycling test [66], accounted for 5.2% and 4.8% decrease, respectively in enthalpies values (Table 6). There were no noticeable visual changes in the sample after cycling.
Therefore, from the point of view of the thermal conductivity and the capacity of the material to transfer heat, and from the point of view of the stabilization of the shape and The melting temperatures (T mo and T mp ) and the crystallization temperatures (T fo and T fp ) recorded for sample TP4, as shown in Table 6, were very close after 450 meltingcrystallization cycles. As expected, after 450 consecutive heating-cooling cycles, the phasetransition enthalpies for the melting stage (H m ) and the cooling stage (H f ), respectively, showed slight decreases (8 J/g and 7 J/g, respectively). Coefficients η m and η f , which represent the magnitude of change in the phasetransition enthalpies as a result of the cycling test [68], accounted for 5.2% and 4.8% decrease, respectively in enthalpies values (Table 6). There were no noticeable visual changes in the sample after cycling.
Therefore, from the point of view of the thermal conductivity and the capacity of the material to transfer heat, and from the point of view of the stabilization of the shape and the homogeneity of the final composite, the TP4 sample presented the best characteristics. In addition, it exhibited a high phase-transition capacity (87%) compared to that of pure PEG, preserving high values of transition enthalpies (~150 J/g).

Conclusions
The new ssCPCM developed in the current work (sample TP4) was obtained via a chemical procedure (an in situ sol-gel process) and included PEG 6000 as the phase-changing material, in-situ-generated silica as the inorganic stabilizing structure, and MWCNTs as the thermoconductive fillers. Both a part of the PEG chains and a part of the MWCNTs-OH were covalently connected by newly formed urethane bonds to the in-situ-generated silica matrix in order to prevent the PEG from flowing out of the structure in which it was integrated. For this innovative PEG-silica-MWCNTs ssCPCM, various physico-chemical characterization techniques were evaluated. Thus, the novel material, which contains 91.5% PEG 6000 after drying, exhibited a solid-liquid phase transition, with melting occurring in the temperature range of 55 • C to 70 • C, with the maximum at~61 • C, and solidification occurring in the temperature range of 30 • C to 45 • C, with the maximum at~42 • C. Its determined relative enthalpy efficiency was 87%.
To verify its thermal reliability and reusability, TP4 composite PCM was subjected to 450 complete successive heating-cooling cycles, with no leakage of the melted PEG from the composite structure being observed. During this test, the melting (146.1 J/g) and solidification (145.1 J/g) enthalpy preserved high values, compared to those of the pure PEG (~187.63 J/g).
These properties demonstrated that the composite PCM we created has the required thermal behavior for use in solar energy storage units to shield some exposed construction elements from overheating, or to prevent electrical equipment, such as batteries, processors, engines, and other devices, from overheating. In future studies, the optimization of the amount and orientation of the multi-and single-wall CNT thermoconductive filler in order to improve the thermal conductivity of the composites, while maintaining the thermal energy storage performance, will be considered. In the next work, light-to-heat conversion will be tested, together with thermal-energy storage for the optimized compositions.

Patents
A patent application entitled "Process for the preparation of a product with thermal energy storage properties based on a PEG 6000 -silica-carbon nanotubes composite material with stabilized shape" (a 2022 00629/13.10.2022), resulting from the work reported in this manuscript, was submitted to O.S.I.M. Romania (the State Office for Inventions and Trademarks).

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/polym15143022/s1. Figure S1. SEM micrographs (left), EDX Scanning Line (middle) and EDX Line Sum Spectrum (right) for: (a) samples with increasing amount of PEG 6000 (Series TP2, TP3, TP4 and TP5) and (b) samples with increasing amount of MWCNTs-OH (series TP0, TP4 and TP15. Figure S2. FT-IR spectra of sample TP2 during the synthesis, at time t0 (time of NCOTEOS feeding) and at time t1 (after 2 h from NCOTEOS feeding. Table S1. Peak position and assignments of the FT-IR spectra within the 1500-1900 cm −1 range. Figure S3. FittedUV-Raman spectra within 750-1850 cm −1 range of the PEG 6000, MWCNTs-OH, and TP15 during the synthesis, after 30 min from NCOTEOS feeding. Table S2. Thermophysical properties of the obtained PEG-silica-MWCNTs composites, evaluated by the Cowan model. Figure S4. Modification of thermophysical properties depending on the content of PEG and MWCNTs-OH in the material, analyzed by the Cowan mathematical model. Figure S5 DSC diagrams corresponding to the 450 successive heatingcooling cycles (left) and modification of the thermal parameters during the 450 melting cycles (right). Table S3. The evolution of the values of the thermal parameters recorded for sample TP4 during the 450 successive heating-cooling cycles.

Conflicts of Interest:
The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper. Nomenclature ssCPCMs = shape-stabilized composite phase change materials; MWCNTs-OH = multi-wall carbon nanotubes functionalized with hydroxyl groups; NCOTEOS = 3-(Triethoxysilyl)propyl isocyanate; PEG 6000 -Si = PEG 6000 coupled with NCOTEOS (newly created building block for the synthesis of the 3D network); MWCNTs-Si = MWCNTs-OH coupled with NCOTEOS; TEOS = tetraethylorthosilicate; GPTMS = (3-Glycidyloxypropyl)trimethoxysilane; MeTMS = methyltrimethoxysilane; w PEG6000 = percentage of PEG 6000 in the final material; λ m = phase-transition capacity for the heating stage; ∆H m = experimental melting enthalpy (latent heat) of ssPCM; ∆H mPEG6000 = melting enthalpy of pure PEG 6000 ; λ f = phase-transition capacity for the cooling stage; ∆H f = experimental crystallization enthalpy (latent heat) of ssPCM; ∆H sPEG6000 = solidification enthalpy of pure PEG 6000 ; T m = temperature at which the melting rate is maximum; T f = temperature at which the rate of crystallization is maximum; α = thermal diffusivity; Cp = specific heat capacity; ρ = density of ssPCM composite; λ = thermal conductivity; e = thermal effusivity; T mo = onset melting temperature for cycling tests; T mp = temperature at which the melting rate is maximum for cycling tests; η m = melting efficiency = (∆H m (cycle 450) − ∆H m (cycle 3))/∆H m (cycle 3) × 100; T fo = onset crystallization temperature for cycling tests; T fp = temperature at which the crystallization rate is maximum for cycling tests; η f = crystallization efficiency = (∆H f (cycle 450) − ∆H f (cycle 3))/∆H f (cycle 3) × 100.