Development of sodium acetate trihydrate-ethylene glycol composite phase change materials with enhanced thermophysical properties for thermal comfort and therapeutic applications

The heat packs using phase change materials (PCMs) are designed for possible applications such as body comfort and medical applications under adverse situations. The development and performance of such heat packs rely on thermophysical properties of PCMs such as latent heat, suitable heat releasing temperature, degree of supercooling, effective heat releasing time, crystallite size, stability against spontaneous nucleation in metastable supercooled liquid state and thermal stability during heating and cooling cycles. Such PCMs are rare and the available PCMs do not exhibit such properties simultaneously to meet the desired requirements. The present work reports a facile approach for the design and development of ethylene glycol (EG) and aqueous sodium acetate trihydrate (SAT) based composite phase change materials, showing these properties simultaneously. The addition of 2–3 wt% EG in aqueous SAT enhances the softness of SAT crystallites, its degree of supercooling and most importantly the effective heat releasing time by ~10% with respect to aqueous SAT material. In addition, the maximum heat releasing temperature of aqueous SAT has been tailored from 56.5 °C to 55 °C, 54.9 °C, 53.5 °C, 51.8 °C and 43.2 °C using 2%, 3%, 5%, 7% and 10 wt% EG respectively, making the aqueous SAT-EG composite PCMs suitable for desired thermal applications.

composite materials. We observed that ethylene glycol (EG) helps in reducing aqueous SAT crystallite size and increases the effective heat retention time about 10% with respect to aqueous SAT phase change material. We will discuss the observed thermophysical properties of investigated materials using structural, microstructural, thermal characterizations and the microscopic origin of enhanced thermal properties of aqueous SAT-EG composite phase change materials.

Results
X-ray Diffraction Analysis. The crystal structure of aqueous SAT (94 wt% sodium acetate trihydrate + 6 wt% deionized water) named as sample A hereafter, and EG modified aqueous SAT samples by 2, 3, 5, and 7 wt% EG are named sample B, C, D and E, respectively hereafter in text. The crystallographic information for these samples are investigated using powder X-ray diffraction (pXRD) measurements in 10°-40° 2Θ range and results are summarized in Fig. 2(a).
The XRD pattern of aqueous SAT is in good agreement with the reported monoclinic crystallographic structure with space group C2/c 32 . The XRD pattern is used to calculate the lattice structural parameters and agree with the reported literature values 32 . The theoretical XRD pattern has been calculated for monoclinic structure (space group C2/c) using the estimated lattice parameters. The generated pattern is summarized in Fig. 2(b) and observed that generated diffraction pattern exhibits only those hkl diffraction patterns for which h + k = 2n; (h0l) with h = 2n, l = 2n and (0k0) with k = 2n (n is an integer). These diffraction patterns substantiate the observed C2/c space group for aqueous SAT. The crystallographic structure of aqueous SAT has been generated using these structural parameters and shown in Fig. 2(c) with polyhedron around sodium atoms. These edge shared polyhedrons form long chains, may be responsible for long needle shaped crystals as observed in microscopic studies and discussed later. The XRD pattern for EG modified aqueous SAT samples (sample B, C, D and E) are identical to that of the sample A. However, the relative intensity of these diffraction patterns has reduced with increase in EG wt% ratio. The reduced intensity is attributed to the geometrical effects, where smaller crystals have resulted in the out of phase diffracted radiation, and thus relatively lower intensity. This has been substantiated with our microscopic observations, discussed later. The XRD measurement was not possible for sample F (Sample A with 10 wt% EG) due to its semi-liquidous characteristics at room temperature (~25 °C). Fourier Transform Infrared Spectroscopic Analysis. The room temperature Fourier Transform Infrared (FTIR) spectroscopic measurements are carried out on these samples. The respective spectrographs are shown in Fig. 2(d). FTIR spectrographs for -OH bending vibrations (1500-1800 cm −1 wavenumber) and -OH stretching vibrations (3000-3800 cm −1 wavenumber) of these samples are shown in insets of Fig. 2(d). These insets suggest that -OH bending and stretching vibrational frequencies shift towards higher and lower frequencies with increasing wt% of EG in aqueous SAT-EG composite materials respectively. The shifting of vibrational frequencies in FTIR vibrational spectra suggests that there are hydrogen bond interactions between H atoms of water and hydroxyl oxygen atoms of ethylene glycol 33 . The observed interaction between H (aqueous SAT) -OH (EG) may assist in reducing the melting temperature and increasing the supercooling temperature of aqueous SAT-EG composites against the aqueous SAT phase change material.
Microstructural Analysis. The microstructural properties of samples A, B, C, D and E are investigated using scanning electron microscope (SEM) and respective micrographs are summarized in Fig. 3(a-j) at 500× and 15 k × magnifications respectively. Figure 3(a) explains the sharp and big SAT crystallites, which are substantiated by the pXRD observations with relatively large diffraction intensity ( Fig. 2(a)), as explained above. The sharp and big crystals reduce the shape adaptability and flexibility of heat pack as per requirements. The SEM micrographs shown in Fig. 3(b-e) explain the reduction in size and edge sharpness of aqueous SAT crystallites with increasing wt% of EG. These results are in good agreement with pXRD results, Fig. 2(a), where the reduced intensity and enhanced FWHM are observed for samples B, C, D, and E with respect to sample A. This reduction in the diffraction intensity suggests that long range ordered crystallite phase ( Fig. 3(a)), has converted into short range ordered microstructures ( Fig. 3(b-e)). The SEM micrograph of sample A (Fig. 3(f)), explains the tightly packed crystallites, which are strong and difficult to break into smaller crystallites. Figure 3(g-j) explains the insertion of liquid EG into aqueous SAT crystallites during crystal growth, thus, weakening the crystals and assisting the reduction of crystallite size. The increasing EG wt% in aqueous SAT has led to the enhanced insertion of EG and thus, reducing the crystallite size into micro/Nano crystallites, as explained in Fig. 3(g-j).
Thermal Properties Analysis. Temperature-history measurements. We carried out T-history measurements to measure the thermal response of samples under investigation using an in-house built temperature history (T-history) set-up 34 . The schematic of experimental T-history set-up is shown in Fig. 4(a). The system  consists of a controlled heating/cooling insulated chamber of size 60 cm (L) × 32 cm (W) × 30 cm (H), 500 W capacity electric heater for heating samples, 600 W capacity refrigeration system for cooling samples, variable air circulating fan to regulate the temperature uniformity inside the chamber, k-type thermocouples integrated with data logger and a computer system for recording the temperature versus time data. The temperature inside T-history chamber can be varied in the range of 0-100 °C as per requirement.
The heating and cooling measurements on samples can be carried out in this T-history set-up and repeated sets of measurements can be carried out with better accuracy as the measurements are carried out without disturbing experimental setup 35 . There are different approaches to generate enthalpy vs. temperature curves for phase change materials using T-history measurements [36][37][38][39][40][41][42][43][44][45] . We have used Sandanes approach to analyze the T-history data and evaluate the thermophysical properties of investigated PCMs 36 . The deionized (DI) water is used as a reference sample. All the PCM samples and deionized water are collected in glass test tubes of size 15 mm diameter and 200 mm height. The weight of each sample has been kept constant (50 g) for T-history measurements. The physical properties of glass test tubes, all PCMs and reference samples, used for these measurements, are listed in Table 1. These samples are heated up to 80 °C inside the chamber as explained in Fig. 4(a) for carrying out temperature vs. time (T-history) measurements and data is recorded at every 10 second time interval. The sample containers are kept inside 150 mm diameter and 280 mm height cylindrical expanded polypropylene (EPP) thermal insulations. This effective thermal insulation is important to realize the low heating and cooling rates with minimal temperature gradients within the sample. This is important to achieve a low Biot number (<0.1) to use lump capacitance method for calculating the thermophysical properties of these materials with better accuracy 46,47 .
Initially, a water sample is heated inside T-history chamber up to 80 °C, (above the melting temperature of sample A ~57 °C) and then cooled to the ambient temperature (23 ± 0.5 °C), (below the solidification temperature of sample A) to record T-history data for water and ambient temperature. The recorded T-history data are plotted in Fig. 4(a) for deionized water during cooling from 70 °C to the ambient temperature. Small periodic oscillations have been observed in ambient (chamber) temperature data because of the lag in compressor and controller feedback process. This small variation in temperature has no significant impact on PCM T-history measurements, as the temperature difference of successive temperatures measured at every 10 sec intervals has been considered for evaluating thermophysical properties. These recorded T-history data are used for calculating the rate of heat loss from deionized water sample in 70-25 °C cooling temperature range, using the following equation: Where  q w i , is the rate of heat loss from water test tube, m w mass of water, c p,w constant pressure specific heat capacity of water (4.18 kJ.kg −1 .K −1 ), m t mass of glass test tube, c p,t constant pressure specific heat capacity of test , temperature difference between two consecutive measurements of water and ∆ = − is time interval between two consecutive measurements, which is 10 second for these measurements 36 . The calculated rate of heat loss from deionized water sample versus temperature difference between deionized water and ambient (T w − T amb ) is plotted in Fig. 4(c). A second order polynomial has been fitted to these heat loss data, as shown by red solid line and used to estimate the heat loss coefficients k 0 , k 1 and k 2 , as described in the following equation: The calculated k 0 , k 1 and k 2 coefficient values are −0.0425 W, 0.03252 W/°C and 2.8363 × 10 − 4 W/°C 2 respectively. The similar T-history measurements are carried out for A, B, C, D, E and F samples. The sample weights, insulation, positioning of k-type thermocouples and ambient temperature conditions are kept identical during the measurements as explained above under T-history measurements section. The cooling T-history data has been recorded from 70 °C to ~30 °C with external nucleation in liquid samples at ~60 °C using solid SAT powder. The measured data are summarized in Fig. 5(a), with the respective ambient temperature.
The temperature of sample A is exhibiting undercooling up to 53.6 °C and after that nucleation started in this supercooled liquid, which led to instant rise in temperature up to 56.6 °C. Samples B, C, D, E and F have shown enhanced supercooling temperature up to 49.9 °C, 49.1 °C, 48.9 °C, 48.4 °C, and 45.4 °C respectively, as explained in the inset of Fig. 5(a). These observations suggest the enhancement in the degree of supercooling for aqueous SAT-EG composite samples with increasing wt% of EG. Thus, aqueous SAT-EG composite samples may exhibit better stability against spontaneous nucleation in metastable supercooled liquid state with respect to aqueous SAT sample A. In addition, the maximum temperature achieved during solidification (heat releasing temperature) of sample A shows decrease with increasing wt% of EG and the maximum reduction has been observed for 10 wt% EG in aqueous SAT composite sample, as explained in the inset of Fig. 5(a). These studies provide an avenue where variation of EG can be used as a parameter to tailor the heat releasing temperature and degree of supercooling of aqueous SAT PCM, required for different applications. The first order derivatives of T-history measurements, not shown here, are used to find out completion of liquid-solid phase transformation temperature more accurately 40 . We observed that liquid-solid phase transformation completion temperature decreases with increasing wt% fraction of EG i.e. the liquid to solid phase transformation temperature range has increased with increasing EG fraction. The samples B and C take ~10% excess time to cool from 70 °C to 40 °C as compared to the sample A for the same temperature range. This suggests that 2-3 wt% EG modified samples (sample B and C) may provide thermal energy for longer time durations, thus more suitable for heat pack applications. However, samples with higher wt% EG in sample A (samples D, E and F) showed nearly the same time as that sample A for cooling from 70 °C to 40 °C temperature range. This is attributed to the reduction in respective enthalpies with increase in wt% EG, as observed in Fig. 5(b). These temperature versus time measurements in conjunction with the calculated heat loss coefficients k 0 , k 1 and k 2 for deionized water are used to calculate the combined rate of heat loss from PCM and test tubes containing PCM samples, using following equation: and the rate of heat loss from PCM has been calculated using following equation: , temperature difference between two consecutive measurements of PCM and t i+1 − t i is time interval between two consecutive measurements, which is 10 seconds for these measurements. The calculated rate  (4), is used to estimate the change in enthalpy (ΔH PCM,i ) of PCM for i th time interval using following equation: The calculated ΔH PCM,i are summed to calculate the cumulative enthalpy of PCM samples with respect to the temperature. The calculated cumulative enthalpy versus temperature graphs, for Fig. 5(a), are plotted in Fig. 5(b).
The cumulative enthalpy of sample A, B, C, D and E and F at temperature 70 °C, nucleated at 60 °C is plotted as an inset in Fig. 5(b), suggesting that the enthalpy of sample A decreases with increasing wt% of EG in SAT. These results are consistent with the expected reduction because of reduced fraction of aqueous SAT in aqueous SAT-EG composite phase change materials.
Additional T-history measurements are carried out for these samples, which are supercooled up to 30 °C and nucleated at this temperature using SAT fine powder. The measured T-history data are plotted in Fig. 5(c). A sharp increase in temperature has been observed for all these samples after external nucleation. This rise in temperature is a consequence of latent heat release while solidification. This temperature rise has been summarized in the inset of Fig. 5(c), suggesting decrease in temperature rise with increasing wt% of EG. This is consistent with that of observed, while nucleating at 60 °C (inset Fig. 5(a)). Inset in Fig. 5(c) suggests that the heat releasing temperature can be tailored from 57 to 43 °C by simply varying wt% of EG up to 10% in aqueous SAT. The respective enthalpy versus temperature data are calculated as explained above and the results are plotted in Fig. 5(d). The enthalpies of these samples at 30 °C against wt% EG are plotted as an inset in Fig. 5(d). The enthalpies of these samples (nucleated at 30 °C) are lower as comparted to that of samples, nucleated at 60 °C. The lower enthalpy values for samples, nucleated at 30 °C is mainly due to loss of thermal energy in the form of specific heat of liquid during PCM cooling from 70 to 30 °C. Differential Scanning Calorimetric Measurements. Further, thermophysical properties of these phase change materials are substantiated by Differential Scanning Calorimetric (DSC) measurements using DSC TA Q10 (TA Instruments USA make). The measuring sample is hermetically sealed in an aluminum pan and empty aluminum pan is used simultaneously as a reference sample. The instrument is calibrated using Indium reference material before carrying out DSC measurements on these samples. The weights of sample A, B, C, D and E used for DSC measurements are 5.19 mg, 5.12, 5.86, 6.27 and 6.08 mg respectively. The measurements are carried out at identical heating and cooling rate of 2 °C/min under inert environment using continuous nitrogen gas purging at 50 ml/min flow rate. The melting and solidification of PCM samples are identified as endothermic (down) and exothermic peaks (up) in DSC thermographs shown in Fig. 6(a,b).
Sample A exhibited sharp (narrow temperature window) endothermic and exothermic DSC peaks, as shown in Fig. 6(a). In contrast to sample A, EG modified composite PCM (samples B, C, D and E) showed wider endothermic and exothermic DSC peaks, Fig. 6(b). The narrow temperature range of exothermic peak for sample A suggests high rate of heat release in narrow temperature window as compared to EG modified composite PCM samples. The onset and endset temperatures and latent heat of fusion for these PCM samples are estimated using DSC thermographs, Fig. 6(a and b), and results are summarized in Table 2. The observations are consistent with T-history observations, suggesting that melting/solidification temperature decreases and melting/solidification temperature range increases with increasing wt% of EG in aqueous SAT. The self-nucleating temperature of aqueous SAT decreases with increasing wt% of EG, suggesting enhancement in degree of supercooling and thermal stability of metastable supercooled liquid against spontaneous nucleation.
The measured degree of supercooling for samples A, B, C, D and E using both DSC and T-history methods are plotted in Fig. 6(c). This suggests that the degree of supercooling has enhanced for EG modified composite PCM samples. The values of self-nucleating temperatures for all samples measured with DSC are lower, as compared to that from T-history measurements. The difference between these results are due to the variation in the mass of samples, used for these measurements. In DSC, a very small amount (few milligrams) of the sample is being used for the measurements, which reduces the probability of nucleation as compared to the voluminous sample (few tens of grams), used in T-history measurements. Hence, T-history measurement values may be more reliable as compared to that of DSC measurements, for any real application, where usually large quantities of samples are used.
Fabrication and Evaluation of PCM Heat Packs. Sample C (aqueous SAT with 3 wt% EG) is used for fabrication of PCM heat packs because of its superior thermophysical properties such as heat releasing temperature, latent heat of fusion, degree of supercooling, effective heat releasing time, small crystallites. Three heat packs, containing 300 g water, sample A and sample C, are fabricated to compare the thermal performance. Polyvinyl chloride (PVC) has been used as a packaging material for storing PCM and triggering device. The stainless-steel triggering device, consisting surface imperfections, is kept inside PCM packs to release heat at the time of requirement. The actual photograph and schematic of such fabricated heat pack (110 mm × 180 mm) and solidification mechanism in PCM heat packs is explained in Fig. 7(a-c). The triggering device is flexed to start nucleation for solidification of PCM in heat packs, Fig. 7(a). This solidification propagates in all the three directions, Fig. 7(b) and takes around one minute for complete solidification of PCM, Fig. 7(c). The thermal energy equivalent to the PCM latent heat is released during this process.
The heat packs are kept inside an insulated box (size 150 mm × 220 mm × 10 mm) made of 20 mm thick EPP material, to ensure thermal equilibrium across the heat pack during T-history measurements. This insulated box is kept inside T-history heating-cooling chamber and heat packs are heated up to 80 °C, followed by cooling up to ambient temperature ~30 °C. The thermal response of these heat packs is recorded from 75 to 30 °C at every 10 second time interval. The PCM inside the heat packs is nucleated at 60 °C above its melting temperature. The measured T-history responses of these three packs are plotted in Fig. 7(d). Water, sample A and C based heat packs took about 2.24, 6.4 and 7.03 hours respectively to cool from 75 to 40 °C temperature. Thus, extended heating time for heat packs containing sample A and C is about 186% and 214% more as compared to the water heat pack. Further, an enhancement of ~10% additional heat release time has been observed for sample C based heat pack with respect to sample A based heat pack.

Discussion
The proposed interactions between sodium acetate trihydrate and ethylene glycol at atomic level is shown schematically in Fig. 8, based on the molecular interactions, inferred from FTIR measurements.
The schematic molecular structure of SAT and EG are shown in Fig. 8(a and b) respectively. The metastable supercooled liquid SAT solidifies in large and lumped continuous crystallites, as shown schematically in Fig. 8(c). Such large lumped crystallites are not suitable for thermal heat pack applications. This can be mitigated by using the physical mixing of EG in SAT matrix. Here, EG is dispersed homogeneously in SAT matrix because of weak hydrogen bond interaction between SAT water hydrogen atoms and EG hydroxyl oxygen atoms, as observed and inferred from FTIR spectra of SAT-EG composite samples. The interaction mechanism is explained schematically in Fig. 8(d). The uniform dispersion of EG assists in the adsorption of liquid EG layers on the growing SAT crystal faces. Thus, these liquid EG layers inhibits the growth of large and lumped SAT crystallites. This liquid EG forms the laminar vesicles, as observed in microscopic studies, Fig. 3. The process of laminar vesicles formation in SAT is explained schematically in Fig. 8(d). These laminar vesicles of liquid EG in SAT crystals are responsible for weakening the SAT crystallites and thus, reducing the continuous growth of SAT crystallites. Further, this inculcated liquid EG will assist in softening, by breaking crystallites with external stimuli causing shear forces between EG separated SAT planes, as shown schematically in Fig. 8(e). The hydrogen bonding between ethylene glycol and H 2 O molecules of SAT may disturbs the hydrogen bonding between SAT and H 2 O molecules. It may be the responsible for enhancement of activation energy required to form nuclei with critical radius and enhances the degree of supercooling and thermal stability of SAT-EG metastable supercooled liquid PCM against spontaneous nucleation compared to SAT.
In summary, novel ethylene glycol and aqueous sodium acetate trihydrate composite phase change materials with enhanced thermophysical properties have been designed and developed for different thermal applications. Thermophysical properties of aqueous SAT PCM, such as melting/solidification temperatures, degree of supercooling, heat retention/release time, have been tailored and optimized by varying EG weight fraction in aqueous SAT-EG composite phase change materials for specific applications. The optimal ~10% enhancement in heat retention time has been observed for 3 wt% EG modified aqueous SAT PCM, without affecting other thermophysical properties significantly. The thermal stability of metastable supercooled liquid SAT against spontaneous nucleation has enhanced significantly for aqueous SAT-EG composite PCMs. In addition, the insertion of liquid EG into SAT crystallites helps in controlling the SAT crystallites size by hindering the growth of large and lumped SAT crystallites. The thermal response of developed aqueous SAT-EG composite PCMs has shown promising Sample Name

Materials and Methods
Sodium acetate trihydrate and ethylene glycol (EG) grade "excel R" are purchased from Qualigens Fine Chemicals Pvt. Ltd. Mumbai, India and used without any further purification.  Preparation of aqueous SAT-EG composites. Aqueous SAT-EG composite samples are prepared by simple physical mixing process. The materials are weighed by a digital micro balance model CX65S (the accuracy of balance is ±0.01 mg, make Citizen, USA). The different aqueous SAT-EG composite samples are prepared by physical mixing of different wt% of EG in aqueous sodium acetate trihydrate (94% SAT + 6% DI water) using magnetic stirrer for 20 minutes at 70 °C. The six samples, sample A (94 wt% SAT + 6 wt% deionized H 2 O), sample B (sample A + 2 wt% EG), sample C (sample A + 3 wt% EG), sample D (sample A + 5 wt% EG), sample E (sample A + 7 wt% EG) and sample F (sample A + 10 wt% EG) are prepared. Additional deionized water in pristine SAT is added to enhance its thermal stability during heating/cooling cycles. The effect of EG on solidification temperature, maximum temperature during solidification, degree of supercooling, heat release time, stability of metastable supercooled liquid and softness of aqueous SAT have been investigated to develop aqueous SAT-EG composite systems as a phase change material for possible heat pack applications.
Thermophysical characterization. The structural properties of aqueous SAT and aqueous SAT-EG composite samples are carried out using Bruker D8 Advance X-ray diffractometer using copper K α radiation ((λ = 1.5406Å). The XRD measurements are recorded in the range of 10 to 40° with step size 0.02°. The microstructural properties of these samples are investigated using Carl Zeiss EVO 18 Scanning Electron Microscope (SEM). The vibrational modes of these materials are identified using Bruker vertex 70 v Fourier Transform Infrared (FTIR) spectroscopy system in the range of 4000-400 cm −1 . An in-house fabricated semi-automated temperature-history (temperature versus time) system, as explained in Fig. 4(a), is used for thermal response measurements of these composite samples. K-type thermocouples (0.5 mm diameter) connected with National Instruments (NI-USA) data logger are used for collecting temperature versus time (T-history) measurements at every 10 second time interval. Differential scanning calorimeter (DSC) Q10 is used to measure latent heat of fusion, melting temperature of samples under investigation and validation of T-history results.