High density polyethylene with phase change materials for thermal energy management

Materials

Three organic PCMs with melting points near 55 °C were investigated. The first one was an organic aliphatic compound, Rubitherm® RT54HC (RT), with a melting point of 54 °C, a melting enthalpy of 200 J/g, and a density of 0.80 g/cm³ at the liquid state, provided by Rubitherm GmbH (Berlin, Germany). The second was a bioderived and biodegradable PCM, Puretemp® PT58X (PT), with a melting point of 58 °C, a melting enthalpy of 225 J/g, a thermal conductivity of 0.2 W/m·K and a density of 0.81 g/cm³ at the liquid state, provided by PureTemp (Minneapolis, United States). The third was a bioderived mixture of fatty acids, specifically stearic and palmitic acids, commercially known as FATTY ACIDS C16-18 (SP), with a melting point of 55 °C, a melting enthalpy of 200 J/g and a density of 0.89 g/cm³ at the liquid state, provided by Carlo Erba Reagents (Val De Ruil, France). HDPE was used as a matrix for the PCM shape stabilization. The product PLASTENE® AD25 in the form of fine powder, with a density of 0.955 g/cm³ at 23 °C and a melt flow index (MFI) of 25 g/10 min (at 190 °C, 2.16 kg), was provided by Poliplast S.p.A. (Casnigo, Italy). A commercial expanded graphite SIGRAFLEX® EXPANDAT by SGL CARBON GmbH (Meitingen, Germany) with a tapped density of 25 g/L at 20 °C and a particle size of 4-40 µm was selected as a filler to improve the thermal conductivity.

Samples preparation

The first production step involved impregnating EG with PCM using a rotary evaporation apparatus, as extensively described in a previous work^[23]. PCM was manually stirred adding 12.7 wt.% of EG, which corresponds to 14 phr (parts per hundred ratio), optimum amount for both shape stabilization and thermal conductivity enhancement^[23]. 114 g of the dispersed mixture were inserted in a 1 L flask, which was connected to the rotary evaporator.A vacuum of 60 mbar was applied while continuing the mixing at a fixed rotation speed of 20 rpm at room temperature. After 5 min, the flask was placed in a water bath, heated to 90 °C, and the rotation was maintained for 30 min after the PCM melting to allow the complete absorption of the PCM within the EG matrix. After cooling outside the bath, the vacuum and rotation were removed, resulting in powdery stabilization of PCM within the EG. Preliminary tests for selecting the PCM to which the conductive filler would be added were performed on vacuum-impregnated PCM/EG samples, as listed in Table 1.

HDPE-stabilized PCM composites were prepared through melt compounding followed by hot compression molding: (i) using the PCMs without stabilization; and (ii) using one of the stabilized PCMs (SP).

The precursors were weighted with balance Kern KB3600 (sensitivity 0.01 g) and compounded using a Thermo Haake Rheomix 600 (Thermo Fisher Scientific, USA) equipped with counter-rotating rotors. The compounding process was performed at 135 °C for 8 min with only HDPE, followed by an additional 5 min after the addition of the PCMs. The rotor speed was 50 rpm and the total amount of material inserted into the mixer chamber was 50 g. The obtained blends were then pressed using a Carver hot press at 140 °C, with a pressure of 2.5 MPa for 5 min. The material was put between two stainless steel plates inside a mold having a frame of different dimensions: 40 × 40 × 10 mm³, 12 × 12 × 3 mm³ and 200 × 200 × 3 mm³. Water cooling to room temperature for 5 min was conducted before the remotion of the obtained sheets from the press.

Samples containing 50 wt.% of HDPE and complementary weight fraction of the three PCMs without EG stabilization were produced. In the case of EG-stabilized PCM (SP only), a composition with 30 wt.% of HDPE was prepared. In order to verify the effective necessity of SP vacuum impregnation, which permits the PCM intercalation between the graphitic lamellae, direct insertion of EG in the compounder (without rotavapor stabilization) was also performed, producing the sample PE-70SP/EG14 (no rot), which showed abundant PCM loss, as confirmed by differential scanning calorimetry (DSC). The four compositions selected for the full characterization are listed in Table 2.

A 90 μm thick polyethylene (PE) envelope for alimentary applications, provided by ORVET S.p.A. (Musile di Piave, Italy), was thermo-welded using an under-vacuum thermo-welding machine ORVET VM-16 in order to totally avoid PCM leakage.

Experimental methodologies

Full characterization

Thermogravimetric analysis (TGA) from 30 to 700 °C at a 10 °C/min heating rate under a 10 mL/min air flow was conducted using a Mettler TG50 thermobalance (Mettler-Toledo, Switzerland) equipped with an alumina crucible. The temperatures associated with a mass loss of 5 wt.% (T_5%), the temperature of maximum degradation rate (T_peak) as the peak of derivative thermogravimetric analysis (DTGA), and the residual mass at 350 °C (m₃₅₀), 400 °C (m₄₀₀) and 700 °C (m₇₀₀) were determined.

DSC of neat PCMs and HDPE-based composites was performed using a Mettler DSC30 calorimeter, under a nitrogen flow (100 cm³/min) with a thermal cycle of heating/cooling/heating in the range of 30 to 60 °C into 40 μL aluminum crucibles. This scanning rate of 1 °C/min was selected to precisely determine the phase transition ranges^[23,42]. The melting temperatures during the first and the second heating scans (T_m1 and T_m2), the crystallization temperature (T_c) and the specific melting and crystallization enthalpy values (ΔH_m1, ΔH_c and ΔH_m2) were obtained. The effective PCM content in the second heating scan ($$ P C M_{m 2}^{e f f}$$) was determined as the ratio of the specific enthalpy of the specimens and the corresponding specific enthalpy of the neat PCM, as given in

(1) $$ \begin{equation} \begin{aligned} P C M_{m 2}^{e f f}=\frac{\Delta H_{m 2}}{\Delta H_{m 2 P C M}} \cdot 100 \end{aligned} \end{equation} $$

while the efficiency of melting during the second heating scan (η_m2) was evaluated using

(2) $$ \begin{equation} \begin{aligned} \eta_{m 2}=\frac{\Delta H_{m 2}}{\Delta H_{m 2, P C M} \cdot W_{P C M}} \cdot 100 \end{aligned} \end{equation} $$

where ΔH_m2,PCM is the specific enthalpy value associated with melting during the second heating scan of neat PCM, while W_PCM is the PCM weight fraction. The parameters expressed in Equations (1) and (2) represent the main indicators to verify the suitability for the PCM stabilization, affected by composition and production processes.

The ability of the produced composite sheets to retain PCM was investigated through leakage tests. These tests involved monitoring the weight of 120 × 120 × 3 mm³ sheets placed into an oven at 60 °C on an absorbent paper towel for 40 days. The residual PCM content was calculated as weight percentage with respect to the total specimen weight, considering the initial PCM percentage at the beginning of the test, i.e., the effective PCM content, as determined according to Equation (2) by the melting enthalpies in DSC tests; the entire weight loss over time was associated with the PCM leakage.

Scanning electron microscopy (SEM) observations of cryo-fractured samples were performed to observe the phase distribution, using a Zeiss Supra 40 field emission scanning electron microscope operating in vacuum at 10^-6 Torr, equipped with a secondary electron detector and working at an acceleration voltage of 2.5 kV. Before the observations, the fracture surfaces of the specimens were metalized through the deposition of a thin electrically conductive coating of Pt-Pd inside a vacuum chamber.

The apparent density (ρ_exp) of the samples was measured using an Ultrapyc 5000 helium pycnometer (Anton Paar, Austria) at 23 °C. For each sample 30 measurements were performed using a testing chamber of 4.5 cm³. The theoretical density (ρ_th) was evaluated by considering the weight percentage of the phases from DSC and using the measured densities of the PCMs and HDPE for the respective phases. For graphite, the theoretical density of 2.26 g/cm³ was used^[23].

The void fraction ($$\vartheta_{v}$$) was computed using

(3) $$ \begin{equation} \begin{aligned} \vartheta_{v}=\frac{\rho_{t h}-\rho_{exp}}{\rho_{th}} \cdot 100 \end{aligned} \end{equation} $$

The specific heat capacity at constant pressure (c_p) was determined at 30 °C following the ASTM E1269-11 standard^[43]. This method involves comparing DSC curves obtained from the sample with those from a sapphire reference of known specific heat capacity. Three specimens for each sample were analyzed using a Mettler DSC30 machine, performing 4 min of isotherm at 20 °C, followed by heating at 5 °C/min till 40 °C and other 4 min of isotherm at the final temperature.

Thermal diffusivity (α) of neat PCMs and blends at 30 °C was determined with two different techniques, using:

• a Neztsch Laser Flash Analyzer (LFA) 446 on disks with a 12.7 mm diameter and a 2.5 mm thickness, imposing laser signals of 250 V for a pulse width of 0.6 ms;

• a Hot Disk thermal analyzer, equipped with a 7577 sensor (diameter of 4.002 mm) on approximatively 40 × 40 × 10 mm³ specimens, according to ISO 22007 standard^[44] into a Memmert HCP climatic chamber at fixed 30% relative humidity applying for 1 to 3 s a power of 30 to 50 mW.

Three specimens for each sample were analyzed. Thermal conductivity (λ) at 30 °C was indirectly determined as product of thermal diffusivity α, bulk density of specimens ρ (measured as ratio between mass and geometrical volume) and the specific heat capacity c_p, as given in

(4) $$ \begin{equation} \begin{aligned} \lambda=\alpha \cdot \rho \cdot c_{p} \end{aligned} \end{equation} $$

Thermal conductivity values were distinguished as λ_HD and λ_LFA, depending on whether they were determined using the hot disk technique or LFA, respectively. The uncertainties were calculated considering the error propagation.

The mechanical behavior of the PE macro-encapsulated materials was evaluated through compression tests of 40 × 40 × 10 mm³ specimens using an Instron 5969 testing machine equipped with a load cell of 10 kN and a thermostatic chamber to perform the tests at 30 and 80 °C, above and below the melting temperature of the PCM, respectively. The tests were performed on neat HDPE and samples containing PCM at a rate of 1 mm/min, following the standard ISO 844^[45] (with different specimen geometries adapted to the specific case). From the obtained stress-strain curves, the compressive modulus of elasticity (E_C) in the linear part of the curve between 4% and 7% of strain and the stress at 10% of deformation (σ₁₀) were evaluated. The strain at 5 MPa (ε₅) was computed for the tests conducted at 30 °C, while for the tests at 80 °C, the strain at 0.5 MPa (ε_0.5) was selected due to the different responses of the specimens below or above the PCM melting point. Furthermore, the stress at 20% of deformation (σ₂₀) was evaluated at 80 °C. The results represent the average of three specimens.

Three-point bending tests were performed at 23 °C using an Instron 5969 dynamometer equipped with 1 kN load cell, following the ASTM D790-17 standard^[46], by using bars of 80 × 10 × 4 mm³ and the testing rate of 1.7 mm/min. From the obtained stress-strain curves, the flexural elastic modulus (E_f) was determined in the strain range of 0.1%-0.4%, while the flexural strength (S_f) and flexural strain at break (ε_fb) were determined at the maximum load as the average of three specimens.

Vicat softening temperature (VST) was determined following ASTM D1525-17e1 standard^[47] using a HDT-Vicat tester model MP/3 subjecting small bars of 4 mm thickness to a mass of 10 N and heating rate of 2 °C/min. The results represent the average of three specimens.

Shore A hardness was determined at 30 and 80 °C using a Hildebrand durometer with an applied load of 488 g for a duration of 5 s, performing ten measurements for each sample, following ASTM D2240^[48].

Finally, infrared (IR) thermography was conducted to observe the PCM homogeneity in the samples and observe the heating and cooling delay of sheets 120 × 120 × 3 mm³ subjected to a thermal cycle from 23 °C (room temperature) to 84 °C imposed into an oven, using an IR thermal imaging camera IRtech Fotric 348A. The sheet configurations to be compared in thermography are shown in Figure 1. During heating, the frames were acquired after at least 20 s from the paperboard removal from the oven.

Figure 1. Sheets of PCM-HDPE composites (120 × 120 × 3 mm³) placed on a paperboard and observed with IR thermography.

Preliminary comparison of PCMs and leaking tests on EG-stabilized PCMs

The three selected PCMs were compared by FTIR analysis, as shown in Figure 2. RT, a modified organic aliphatic compound (RT54HC), was identified by the aliphatic C-H stretching peaks at about 2,915 and 2,850 cm^-1, and bending and wagging peaks at about 1,470 and 725 cm^-1. Additionally, the intense peak at about 1,700 cm^-1, attributed to C=O stretching, and the broad band between 2,600 and 2,550 cm^-1 typical of -COOH, together with its close similarity to SP in the fingerprint zone (discussed later), suggest an attribution to aliphatic fatty acids.

Figure 2. FTIR spectra of RT, PT and SP phase change materials with melting temperature of about 55 °C.

The spectrum of PT is consistent with that of aliphatic alcohol, showing the -OH stretching in the range 3,400-3,050 cm^-1, the C-O stretching at 1,060 cm^-1, and the other three main double peaks of -CH₂-CH₂-, i.e., stretching at 2,920-2,850 cm^-1, bending at 1,470-1,455 cm^-1 and wagging at 720 cm^-1.

Fatty acid mixture SP is derived from stearic and palmitic acids. As expected, the FTIR spectrum exhibits the C=O stretching peak at about 1,700 cm^-1, and the typical band at 2,700-2,500 cm^-1 of -COOH, plus the signals of various -CH₂- and -CH₃. The near-equivalence of the SP and RT spectra suggests a qualitatively identical composition, with probably some minor quantitative differences. The EG ability to stabilize the different PCMs is investigated considering the three-hour long leakage test results on pressed disks of vacuum-impregnated EG with PCMs [Figure 3].

Figure 3. Leakage test results at 60 °C of vacuum-impregnated EG with PCMs.

A better interaction between SP and RT (both aliphatic fatty acids) and EG is evident compared to the behavior of PT, an aliphatic fatty alcohol. In the first hour, PT/EG14 shows a mass loss more than three times greater than that of the other two samples. Over the following 2 h, the trend becomes similar for all disks. After 3 h of testing, the total mass loss of RT/EG14 and SP/EG14 is 2.4 and 1.8 wt.%, respectively, indicating SP/EG14 is the best candidate for EG introduction into the HDPE-based composites; in addition, SP can be considered more sustainable as it is bio-derived. The better EG stabilization of the fatty acid mixture (SP and RT) compared to the aliphatic alcohol (PT) is likely due to the higher chain polarity and capillarity forces conferred by the carboxylic group, resulting in a quite stable PCM/EG interaction^[23].

Thermogravimetric analysis

Thermogravimetric analysis (TGA) curves of neat PCMs, neat HDPE and composite samples as residual mass and derivative of thermogravimetric analysis (DTGA) are reported [Figure 4]. Main results are summarized in Table 3: temperatures associated with a mass loss of 5 wt.% (T_5%), temperatures associated with the maximum rate of degradation (T_peak) with multiple peaks for HDPE composites, and the residual masses at 700 °C (m₇₀₀). The full thermal degradation of PCM is clearly evident in the range 200-300 °C, with a faster mass loss for aliphatic alcohol (PT) compared to aliphatic acids. The maximum degradation rates (DTGA peaks) center at 226 and 287 °C, respectively. On the other hand, the main degradation of HDPE occurs above 400 °C, and the produced composites show intermediate behavior, with a two-step degradation evidenced by two peaks in DTGA, attributable to PCM and the polymeric matrix, respectively.

Figure 4. TGA of neat PCMs, composite samples and neat HDPE. (A) Residual mass; (B) Derivative thermogravimetric analysis (DTGA).

The PCMs begin degradation at around 200 °C, with PT showing lower resistance to temperature (T_5% 204 °C), whereas RT and SP exhibit similar and higher values (T_5% 227 °C). This confirms PCM stability during processing at 140 °C. Above 400 °C, all the PCMs completely degrade with no residual masses at the end of the tests. HDPE starts thermal degradation above 300 °C, with 5 wt.% of degradation at 357 °C and a peak of weight loss at 434 °C, with the complete mass loss at about 550 °C. Two degradation peaks appear for the samples PE-50RT, PE-50SP and PE-50PT. The first peak corresponds to the degradation of the PCM (T_peak respectively at 258, 301 and 254 °C), followed by the degradation of HDPE (442, 443 and 426 °C). The mass loss at 350-400 °C could be considered a semiquantitative evaluation of PCM content, i.e., about 45%-50% in PE-50PCM, and about 60% in PE-70SP/EG14, partially agreeing with the nominal content.

No residues for PCMs, HDPE and their compositions at 50 wt.% are found at 700 °C due to the complete thermal degradation. On the other hand, PE-70SP/EG14 shows a residual mass of 8.1 wt.%, almost corresponding to the nominal content of EG, which is stable at 700 °C^[23]. This double-stabilized sample presents a more gradual mass loss profile, due to the presence of three distinct phases degrading at three different temperatures.

Differential scanning calorimetry

Phase transition temperatures and enthalpies of neat PCMs can be determined from the Differential Scanning Calorimetry (DSC) thermograms as shown in Figure 5, and the main results are summarized in Table 4.

Figure 5. DSC thermograms of the neat PCMs at 1 °C/min. (A) First heating; (B) Cooling; (C) Second heating. Fatty acids (RT and SP) and fatty alcohol (PT).

As shown in Figure 5, the entire melting transition occurs in the range 49-58 °C, while the crystallization peaks during cooling are observed in the range of 54-43 °C. According to the declared melting temperatures from the datasheets, there is a shift of a few degrees to higher temperatures as the transition occurs from fatty acids (RT and SP) to fatty alcohol (PT). PT shows the highest enthalpic content (250 J/g), making it preliminarily the best among the three PCMs, with its melting enthalpy much higher than that of RT (206 J/g) and SP (196 J/g). PT, whose composition is not specified by the producer, presents a double crystallization peak at 53 and 47 °C, respectively, indicating the probable presence of two distinct phases of fatty alcohol. It also shows among the three PCMs, the broadest temperature interval for crystallization, ranging from 54 to 43 °C.

Figure 6 displays the DSC thermograms of the HDPE composite samples, including melting, crystallization and melting during the second scan. Selected data of enthalpy and temperature transitions are provided in Table 5.

Figure 6. DSC thermograms of the HDPE-based samples at 1 °C/min. (A) First heating; (B) Cooling; (C) Second heating.

The melting temperature of the main peaks is in the range 52-54 °C, quite similar to the corresponding neat PCMs. However, it is worth noting that a relatively larger interval of PCM melting in HDPE compositions with the presence of shoulder peak, in both heating scans [Figure 6A and C], suggests the formation of different crystal domains.

Concerning crystallization, a broadening of the phase transition is observed [Figure 6C], with multiple peaks for the various PCMs in HDPE. It is also evident that the main peaks exhibit a lower crystallization temperature compared to those of pure PCMs. For instance, the main peak of crystallization of SP in HDPE is at 45 °C, instead of 51 °C [Table 4], showing a shoulder at about 50 °C. Moreover, three peaks were observed for PE-50RT (50, 43 and 38 °C) and PE-50PT (53, 48 and 33 °C) indicating HDPE inhibits crystallization, as demonstrated in other studies^[49,50]. Similar multiple melting peaks were documented for co-crystallization of syndiotactic polystyrene (sPS) in PA66^[51] and were attributed to the dispersed particles. The lower the size domain in the surrounding polymeric matrix, the lower the perfection of crystals and the lower the melting temperature. These findings were interpreted by Michell and Muller as fractionated crystallization upon cooling from the melt due to confinement effect, reduction in lamellar thickness, and decrease in melting temperature^[52]. Noteworthy in this context is the main crystallization peak of SP after impregnation in EG and dispersion in HDPE, which shifts from 53 to 46 °C, with a minor peak remaining at about 51 °C [Figure 6B and Table 5].

The energy involved in the thermal process was evaluated in the second heating scan by the melting enthalpy and the efficiency of melting [Table 5 and Equation (2)]. The first three compounds with a nominal content of PCM at 50 wt.% resulted in effective PCM content in the range 25-44 wt.% according to Equation (1). In particular, the lowest melting enthalpy (49 J/g for PE-50SP), with an efficiency η_m2 of 50%, is unacceptable in terms of material application. These results could be attributed to PCM loss during production, especially during pressing, and the reduced dimensions and imperfections of crystalline domains. A similar enthalpy reduction was observed in PCM dispersed in wood-based composites as a result of spatial confinement effects on PCM^[53].

Although enthalpies are significantly reduced compared to neat PCMs due to the presence of the matrix, values of 92 and 95 J/g can be obtained for PE-50RT and PE-50PT, corresponding to melting efficiencies of 89% and 76%, respectively. This evidence a better ability of RT than PT to be stabilized in HDPE through the considered production process.

A significant improvement in the efficiency of SP in HDPE compounds was achieved through EG impregnation. In fact, graphite can provide a strong stabilizing effect due to capillary forces and surface tensions^[23], notwithstanding its low fraction (8.6 wt.%) in the composite. Effective PCM content increased up to 44% (86 J/g) after the direct insertion of the graphitic filler into the compounder, and to 57% (112 J/g) after rotavapor process, which is very close to the nominal value of 61%. For these reasons, PE-70SP/EG14, produced after vacuum impregnation of the EG, exhibited the highest efficiency (93%) and was selected for the subsequent characterizations. These enthalpy values are very promising and higher than those reported in many works exploring PCM shape stabilization using HDPE (10-66 J/g), as in the case of wood fiber composites 60 J/g^[54], or in combination with EG^[14,36,55] or other graphitic fillers^[20]. The DSC analysis results quantify the ability of PCM systems to subtract heat from a device in thermal management applications and store heat in TES.

Leakage test

The ability of HDPE composites to retain the PCM above its melting temperature over time was assessed through a leakage test performed at 60 °C for 40 days, as shown in Figure 7, which illustrates the PCM content in wt.% of the four samples as a function of time. The initial value corresponds to the effective PCM content from Table 5, and the final PCM loss is reported in the caption. A comparison of the leakage of PE-70SP/EG14 and PE-50PT sheets at the end of the test is shown in the pictures.

Figure 7. Leakage results at 60 °C of HDPE composites with different PCM contents.

The PCM loss is concentrated in the first 100 h, with a more pronounced loss observed for PE-50PT and PE-50RT samples, followed by a progressive reduction in the leakage rate until a threshold is reached after around 500 h, evidencing trends similar to those reported by Valentini et al.^[56]. PE-50SP is the sample suffering from the lowest PCM loss (-0.6 wt.%), probably due to the highest loss during the production process, with a quite low residual effective SP content of 24.6 wt.%. The second sample in terms of resistance to PCM loss is PE-70SP/EG14 (-1.2 wt.%), with a residual effective PCM content of 56.1 wt.%, confirming that the combination of HDPE and EG for the shape stabilization of SP is very efficient, while also achieving the highest PCM content. The leakage of PE-50PT and PE-50RT is more abundant, with losses of 3.9 and 2.5 wt.%, respectively, resulting in residual effective PT and RT contents of 35 and 42 wt.% respectively. In all cases, the values are much lower than in other studies that have used HDPE to stabilize non-microencapsulated PCMs^{[49,54,57-59]}. Thus, HDPE presents very good properties as a matrix to stabilize the considered PCMs, with the majority of PCM loss occurring during melt compounding and hot compression molding. An explanation could be the formation of polymeric layers of HDPE into the structure behaving as a barrier for the PCM penetration; hence, the PCM remains entrapped in the structure after the loss of the peripheral PCM dispersed in the specimens. Despite the 89% efficiency evaluated by DSC for PE-50RT, which can be considered quite lower than, for example, those close to 100% for only EG impregnation^[23], the total leakage in the 40 days in the presence of HDPE as supporting material is very well contained. The EG-stabilized samples^[23] showed continuous PCM loss (8 to 14 wt.% in total); thus, the high impregnation efficiency was contrasted by the abundant leakage in service above melting.

By encapsulating with the PE-envelope film, leakage can be totally avoided; this solution was adopted for thermal management applications to prevent material leakage into the environment^[23] and can also be used without any supporting matrix^[60,61]. Moreover, with this approach, PCM-based systems are suitable for adhesively attaching to surfaces, thus avoiding the loss of adhesion that a leaked PCM layer can cause onto the surface of the plate. Examples of macro-encapsulated panels with the thermo-welded PE envelope are shown in Figure 8.

Figure 8. PE enveloped panels having dimensions 120 × 120 × 3 mm³ (top) and 200 × 200 × 3 mm³ (bottom).

Scanning electron microscopy

Fracture surfaces of HDPE composites are investigated to gain an overview of the phase distribution. Scanning electron microscopy (SEM) micrographs at low magnifications are shown in Figure 9. From the low magnification microstructures, a homogeneous distribution of the HDPE matrix and very small PCM flakes can be observed, demonstrating the suitability of the parameters used in melt compounding to disperse the phases in the nominal composition at 50 wt.% of single PCM. The fracture surfaces show a decrease of RT, PT and SP particles in HDPE, with their effective percentages being 44%, 38% and 25%, respectively, as determined by DSC. In contrast, for PE-70SP/EG14 samples, the graphite lamellae containing entrapped SP are clearly distinguishable^[27], and some heterogeneity in the phase distribution can be observed, with even some cavities between different EG and HDPE domains, showing internal separation lines of a few hundred microns.

Figure 9. Overview of SEM micrographs of HDPE composites. (A) PE-50RT; (B) PE-50PT; (C) PE-50SP; (D) PE-70SP/EG14.

Higher magnification SEM micrographs, shown in Figure 10, allow for better recognition of morphologic features. The PCMs appear as micro-flakes of varying dimensions dispersed in a continuous phase of the HDPE matrix, whose filamentous fracture is evident in all compositions. Figure 10A shows RT micrometric particles with an equivalent diameter mainly between 4 and 10 microns, and a thickness in the range of 1-2 microns; some smaller particles, about 1-2 microns, could also be recognized in the HDPE domain. These different particle sizes can be related to the various melting DSC peaks. A similar aspect is also observed for PT with different groups of PCM particles dispersed in the HDPE matrix [Figure 10B]. Particles of about 1-2 microns are evident in the larger fracture domains with an almost planar morphology, whereas wider PT particles of about 3-5 microns, partially squeezed out, are visible between the elongated HDPE fibrils evidencing a ductile fracture^[62].

Figure 10. High magnification SEM micrographs of HDPE composites. (A) PE-50RT; (B) PE-50PT; (C) PE-50SP; (D) PE-70SP/EG14.

From Figure 10C, the distinctive planar SP particles can be noticed^[63], with an average width in the range of 10-25 microns and very thin lamellae, less than 0.5 microns, confirming the correct selection of the fatty acid mixture to be intercalated among the graphitic lamellae. PE-70SP/EG14 [Figure 10D] presents a more complex morphology due to the coexistence of at least three phases of three components (HDPE, PCM and EG). The nominal content of HDPE is only 30 wt.%, but the filament morphology of HDPE remains visible in the fracture zone. Moreover, the higher porosity introduced by the EG and the typical lamellar structure of EG can be observed^[21,27,55].

The various morphologies observed in SEM microscopy confirm the presence of different PCM structures, as previously discussed in DSC analysis (melting and crystallization peaks of PCM phases or domains), evidencing some porosity for PE-70SP/EG14.

Density measurements

Density analysis makes it possible to compare different morphologies and to highlight possible defects, and is reflected on the thermal conductivity both directly [see Equation (4)] and indirectly (dependence of diffusivity on void fraction). Table 6 reports the apparent densities, measured using a helium pycnometer, of neat PCMs, HDPE, EG and composite samples, along with the theoretical density values (considering the evaluated densities of the single phases and the effective contents from DSC) and the porosity evaluated with Equation (3).

The two aliphatic fatty acids SP and RT possessed a similar density of about 0.985 g/cm³, slightly higher than that of HDPE, while the value for PT is lower than 0.9 g/cm³. Datasheets report PCM densities in the liquid state, whose values (in the order of 0.8 g/cm³) are coherently lower than those evaluated in the solid state due to the large volumetric expansion after melting, a characteristic typical of PCMs^[1]. The density measurements of HDPE are only 0.2% lower than those provided by the producers, and this slight difference could be attributed to residual voids and/or to a lower crystallinity content of the processed polymer. EG density, measured by pressing it into pellet form, is 0.3% higher than the theoretical value of 2.26 g/cm³ present in literature^[64]. PE-50SP and PE-50PT samples present very low void fractions, while PE-50PT has a density higher than the theoretical one with a corresponding negative void fraction which has no physical sense; this can be explained by the probable presence of voids in the neat PT granules tested in the pycnometer due also to the lower measured density than the other two PCMs and the similar values of the HDPE-stabilized samples. On the other hand, a considerable void fraction (13.4%) is confirmed in PE-70SP/EG14, in agreement with the SEM observations.

Specific heat capacity and thermal conductivity

In order to evaluate the best thermal performance in heat transmission, the specific heat capacity determined through DSC, bulk density and thermal diffusivity and conductivity values at 30 °C are listed in Table 7.

All the c_p values range from 1.7-2.1 J/(g·K), with the highest values observed for the PCMs and the lowest for HDPE, agreeing with datasheets and literature^[65]. In the case of the PCM-HDPE blends, the c_p values are slightly lower compared to the neat PCMs, while the presence of graphite reduces the value of the PE-70SP/EG14 sample. Thermal diffusivities, presented in Table 7 and used for the thermal conductivity evaluation, exhibit the same trend as described for thermal conductivity below.

The thermal conductivity of the PCMs is in the order of 0.2 W/(m·K) at 30 °C as expected^[9], while the measured thermal conductivity of HDPE is 0.5 W/(m·K), consistent with the value reported by Yang et al.^[66]. The PCM-HDPE compounds exhibit around double thermal conductivity of the neat PCMs. To be noted, the presence of EG as a conductive filler leads to an increase of the thermal conductivity of PE-70SP/EG14 with a measured LFA value up to 1.65 ± 0.12 W/m·K, confirming that this addition is beneficial to boost the heat transfer, fundamental requirement for the PCMs usage. The corresponding enhancement relative to neat SP is about nine times greater when comparing LFA and hot disk values, respectively. This result shows a much higher improvement than those reported in most works where thermal conductivity of PCM and HDPE composites is enhanced using similar content of EG^{[49,54,57,59]} or other conductive fillers^{[20,50,58,67]}, and is comparable only when higher EG weight fractions are used^[14,66].

Limited differences between the values found using hot disk and LFA techniques can be observed. This is quite common when thermal conductivity tests are performed, as noticed and extensively treated by Weingrill et al.^[68]; these discrepancies are not so marked and acceptable in the present study.

Measurements of the thermal conductivity above the PCM melting temperature will be possible with specific equipment due to the leakage problems of the materials, and could be the subject of future research.

Finally, the role of the PE film covering the final plates can be considered negligible due to the very low thickness (90 μm)^[23] and the thermal conductivity comparable with those of the composites due to the same material of the sample matrix.

Compression test

The thermo-mechanical behavior of the samples was investigated through compression tests conducted at 30 and 80 °C, temperatures around 25 °C below and above the PCM melting points. The stress-strain curves of a representative specimen for each sample are reported in Figure 11, while the main results of the compressive tests are summarized in Tables 8 and 9.

Figure 11. Stress-strain curve compression test. (A) 30 °C; (B) 80 °C.

Comparing Figure 11A and B, the decrease in the mechanical properties of all the samples above the PCM melting temperature is evident, showing large deformations and about half of the maximum stress due to the presence of fluid domains. The only exception is neat HDPE, for which the deterioration of the properties is much less pronounced due to the absence of molten phases. At 80 °C, a plateau can be noticed for PE-70SP/EG14 sample, from 10% to 30% strain at about 1.2 MPa.

From Table 8, the compressive elastic modulus at 30 °C of neat HDPE is the highest (72 ± 5 MPa), similar to that of PE-50PT (73 ± 0.1 MPa), and much higher than PE-50RT (49 ± 3 MPa). This demonstrates a better interaction of PT with HDPE in comparison with RT, even if more PT was lost during sample production than RT. Despite the low SP content of the PE-50SP sample, its modulus is quite lower than HDPE (62 ± 13 MPa), even if it is improved by the presence of EG (68 ± 18 MPa), also considering the highest PCM content of the PE-70SP/EG14 sample. For both SP-based samples, the values are scattered with a high standard deviation; thus, the fatty acid introduces uncertainty in the response of material, aggravated by the addition of EG. Besides the limited compressive test references available in the literature, the produced materials can be considered stiffer and more resistant than PCM/HDPE composites present in literature, such as those characterized by Nemati et al.^[69]. A comparison with the study of Cheng et al. is not possible due to the absence of elastic modulus determination and the reported deformation stresses at much higher strains^[67].

Neat HDPE does not reach 10% strain before the end of the test, while all other samples reach it under stress of 4 to 6 MPa; similar behavior among the specimens can be noticed by examining the stress-strain curve trends [Figure 11A]. The strain at 5 MPa of applied stress is similar for all samples, with values ranging from 8% to 9% for neat HDPE and PE-50PT, and 10% to 11.5% for the other samples.

As anticipated, at 80 °C, all compressive properties are drastically reduced, with a more pronounced effect in the presence of PCM. In fact, the mean compressive elastic modulus at 80 °C is reduced to 62.5% of its value at 30 °C for HDPE, compared to 26.5% for PE-50RT, 13.7% for PE-50PT, 25.8% for PE-50SP, and 23.5% for PE-70SP/EG14. Regarding the stresses required to achieve 10% strain, HDPE reaches it with an applied stress of 3.9 ± 0.1 MPa, while the mean stresses at 80 °C are 35.2%, 15.0%, 25.5% and 30.8% of the values measured at 30 °C. Once again, the beneficial effect of EG can be noticed, considering also that PE-70SP/EG14 presents the highest melting enthalpy [Table 5]. For the PCM/HDPE blends, the stresses double when the strain value doubles, while for PE-70SP/EG14, the stress increase is only 25%; however, HDPE never reached 20% strain. The strains at 80 °C evaluated at 0.5 MPa (one order of magnitude below the tests at 30 °C) range from 5.1% to 6.6%, except for neat HDPE (3.8% ± 0.2%).

In Figure 12, a specimen for each sample after the test is shown, underlying the spreading of the material tested at 80 °C, along with the evident permanent deformation of the specimens. None of the envelopes present any damage after the tests, confirming that the PE macro-encapsulation is a suitable technique for containing the PCM, also under unexpectedly high loads.

Figure 12. Compression tested samples at 30 °C (above) and 80 °C (below). From left to right: neat HDPE, PE-50SP, PE-50RT, PE-50PT and PE-70SA/EG14.

Three-point bending test

The flexural behavior of the materials is investigated as auxiliary information of mechanical properties for thin geometries applications. In Figure 13, a flexural stress-strain curve of one representative specimen for each sample is shown, and in Table 10 the main flexural properties are listed. HDPE remains unbroken throughout the test as shown in Figure 13, evidencing a very ductile behavior; the stress decrease after the maximum point is due to the slippage of the specimen from the supports. On the other hand the addition of PCM resulted in a lower maximum stress and embrittlement. A quite ductile behavior is also observed in PE-50SP, probably due to the low PCM fraction.

Figure 13. Flexural stress-strain curve at room temperature.

The flexural elastic modulus increases for PE-50RT (1.07 ± 0.05 GPa) and PE-50PT (1.21 ± 0.09 GPa) compared with neat HDPE (0.93 ± 0.06 GPa), while the addition of SP reduces it (0.57 ± 0.08 GPa), unless combined with EG (1.64 ± 0.04 GPa), which results in the highest reached value, again demonstrating the beneficial effect of EG in enhancing stiffness. The flexural strength is negatively affected by the addition of PCM, decreasing from 22.9 ± 0.3 MPa to values ranging between 12 and 18 MPa. EG improves this, increasing it from 12.0 ± 0.2 MPa in the PE-50SP sample to 15.1 ± 1.4 MPa in the PE-70SP/EG14 sample. The flexural strain at break follows an inverse trend relative to the PCM content in the sample: it decreases from 7.0% for neat HDPE to 1.5% for the PE-70SP/EG14 sample. Sun et al. obtained comparable properties for neat HDPE; however, the flexural strength and elastic modulus of the PCM composites are lower than those obtained in this study, despite their materials having a lower PCM content^[33]. In the open literature, to achieve comparable flexural properties, other phases are incorporated as supporting matrices in addition to HDPE, drastically reducing the PCM content^[35,37].

Vicat test

In order to evaluate the maximum using temperature, the resistance of the materials to external penetration in temperature is quantified by the Vicat Softening Temperature (VST) values reported in Table 11.

The penetration of 1 mm for HDPE occurs at a temperature a few degrees lower than its melting point, while in the presence of PCM it decreases by about 40-50 °C, with the exception of PE-50SP, whose VST of 90 ± 5 °C could be attributed to a lower amount of PCM. The lower the HDPE content, the lower the VST. One of the main advantages of using HDPE is the possibility to subject the composite material to temperatures higher than 95-97 °C.

Shore A hardness

The hardness of HDPE and composites was evaluated below and above the melting temperature of PCM, in order to obtain further data on the panel performance. In Table 12 the results of shore A hardness measured at 30 and 80 °C are reported.

The shore A hardness (ShA) at 30 °C is similar for all the samples, in the range 95-97, and at 80 °C shows a ShA decrease between 87 and 94 ShA in the presence of PCM. It is worth noting that these experimental results of Shore A at 80 °C above the melting temperature of PCM are much higher than those obtained by Ehid et al. (Shore A of 48 at 50 °C for blends of HDPE and 50 wt.% of paraffin), while the hardness values at ambient temperature are consistent^[70].

Infrared (IR) thermography

A verification of the heat absorbance at constant temperature of the PCM-based composite materials, which is fundamental for the thermal management mechanism, is shown in Figure 14. The thermography of the sheets depicted in Figure 2 is presented during heating from 23 to 84 °C (left) and cooling with the inverse ramp (right) after 2 (A and B), 15 (C and D) and 45 (E and F) minutes.

Figure 14. IR thermography images during heating (left) and cooling (right) of the sheets 120 × 120 × 3 mm³ shown in Figure 2 after: 2 min (A and B), 15 min (C and D) and 45 min (E and F).

The maintenance of the constant temperature for all the samples can be clearly observed once the PCM has molten. During cooling, a slower decrease in temperature compared to the paperboard is noticed until the recrystallization temperature of PCM is reached; at this point, the temperature remains constant until the PCM recrystallization is completed. Figure 14D shows the highest enthalpic content of PE-70SP/EG14 (right-lower corner), as indicated by the fact that, unlike the other samples, all areas of this sample remain around 50 °C, suggesting that a fraction of PCM has not yet solidified. In addition, the temperature is well-homogeneous, confirming the effectiveness of EG in improving the thermal response. After 45 min of cooling, the system reaches equilibrium at room temperature. Inhomogeneities can be observed in the materials without EG, with temperature variations across different spots in the sheets; however, the phase distribution is much more uniform compared with other PCM distributions^[71]. The paperboard temperature serves as a reference, assuming the panels were composed of only HDPE.

In Figure 15, the recorded temperatures at the center of the sheet surface and on a paperboard spot during cooling are reported. The delay in reaching a certain temperature due to the presence of the PCM is evident, in the order of 30 min considering either 30 or 35 °C. The starting temperature of the supporting paperboard starts its sudden temperature decrease immediately after the remotion from the oven, while the PCM-based sheets present a much less pronounced cooling rate. At the beginning of cooling, PE-70SP/EG14 is at SP melting temperature; hence, a PCM fraction is still solid; the other samples instead are heated above 60 °C after the complete melting of the PCM, confirming the lower enthalpic content. Similarly, the lower slope of PE-70SP/EG14 temperature profile confirms a higher stored heat providing 10 min of additional delay to reach 45 °C with respect to the other samples. These delays will increase with the mass (total enthalpy)^[23], in either heating (thermal management) or cooling; unfortunately, during heating, a continuous representation of the temperature with time is not possible because the IR thermo-camera must be kept outside the oven.

Figure 15. IR thermography temperatures in cooling.

Comparative properties of HDPE-based panels

In view of a potential final application and considering panels 2 cm thick, the main thermo-mechanical properties of the systems are compared in Table 13 to allow a qualitative and quantitative evaluation of the performances in the material selection. The thermal management ability (TMA) for a panel of thickness 2 cm is expressed as an energy normalized for surface unit, as given in

where Energy is related to a panel of volume (V) 100 × 100 × 2 cm³.

Starting with a general comparison of the thermal management plates, considering the similar densities and elastic flexural moduli, as well as maximum melting enthalpy observed in the presence of EG, the applicable temperature is higher for the composites containing larger fractions of HDPE. Furthermore, PE-70SP/EG14 exhibits the highest flexural modulus (1.6 GPa) and the best properties in terms of thermal conductivity (1.65 W/m·K) and volumetric enthalpy (112.1 J/cm³). Additionally, it shows a narrower crystallization temperature range and the highest surface density, which contributes to the highest TMA value of 2.2 MJ/m². This value can be considered a direct indicator of the thermal performance of the panels.