Nanoencapsulated Crystallohydrate Mixtures for Advanced Thermal Energy StorageMichael Graham*, José A. Coca-Clemente, Elena Shchukina and Dmitry Shchukin

Abstract

Controlled storage of thermal energy, especially in the ‘low temperature’ region <230 ºC, is very important for domestic applications, where effective heating and cooling of buildings is required in off-peak and seasonal periods. Additionally, domestic thermal energy storage systems have to provide energy mobility to expand their prospective application from buildings into other areas. The use of phase change materials (PCMs) can provide necessary storage and mobility levels. Inorganic crystallohydrate PCMs are non-flammable, have much higher latent heat capacity than organic PCMs and flexible transition temperatures, which depend on crystallohydrate composition. In this study, we demonstrated a universal method for encapsulation of crystallohydrate PCMs (Mg(NO 3)2·6H2O and Na2SO4·10H2O) and their mixtures into 100-300nm capsules with a poly(ethyl-2-cyanoacrylate) shell, employing in situ inverse miniemulsion polymerisation combined with ultrasonic treatment. Encapsulated crystallohydrates and their mixtures demonstrate high stability during energy uptake/release (>100 cycles as monitored by differential scanning calorimetry), due to the functional properties of the capsule shell and spatial confinement preventing water loss and incongruent melting during phase transitions. Mixtures of encapsulated crystallohydrates have differing energy uptake/release depending on their inherent design. Additive mixtures (mixtures of energy capsules containing single crystallohydrates) possess separate heat uptake/release transitions at temperatures corresponding to each crystallohydrate type. Latent heat of each transition is proportional to the content of the corresponding energy capsules in the mixture. Nanocapsules containing mixed crystallohydrate core have synergetic effects from Mg(NO3)2·6H2O and Na2SO4·10H2O components at their eutectic point: low-temperature phase transition and high latent heat capacity. Chemical composition, integrity and morphology of the capsules were analysed by scanning electron microscopy, Fourier transform infrared spectroscopy and thermogravimetric analysis.

IntroductionLatent heat storage (LHS) using phase change materials (PCMs) is an efficient way of storing thermal energy. Major concerns about pollution and climate change mean greener energy sources are vital (Felix De Castro, Ahmed, & Shchukin, 2016; Zhao & Wu, 2011). Energy storage has the potential to greatly improve the reliability of solar power(Wang, Tang, & Zhang, 2013), as well as reducing overall electricity demand. Energy storage is highly beneficial for environmental protection and energy security. Several important applications could benefit from LHS. Storage for concentrated solar power (CSP) plants 4 gives improved reliability and energy output due to peak shifting. When production is greater than demand, energy is stored. The energy can then be released at periods where demand outweighs production5–7. PCMs can provide ‘smart’ thermal management for buildings8,9, electronics10, photovoltaics11,12 etc., reacting to changes in local temperature. A large amount of heat is stored as the PCM melts with much higher efficiency than sensible heat storage materials13–15, which store energy as they heat up. PCMs release the stored energy during freezing. The narrow temperature range in which PCMs operate reduces heat loss to the environment, especially useful for low temperature heat (<230°C) 16, which is often produced as waste in industrial applications and can be transported for other applications such as greenhouses17. There are two main types of PCMs which can be used for thermal energy storage at temperatures below 200°C: organic paraffin waxes and inorganic salt hydrates (also known as crystallohydrates). Paraffins, despite their availability as commercial materials for heat storage, have latent heat capacity around 100-150 kJ/kg18 and are flammable19 which significantly reduce their potential applications. On the contrary, salt hydrates have advantages such as almost doubled energy storage density, non-flammability and low cost20,21. However, salt hydrates are chemically unstable due to phase separation (non-congruent melting) and suffer from supercooling during heat uptake and release cycles. Encapsulation in a polymer shell improves PCM properties by extending lifespan, preventing decomposition and increasing heat transfer area14. These effects are due to increased surface-area-to-volume ratio, and have been proven with much research on encapsulation of paraffins22–28. Encapsulation removes the need for additives to improve thermal performance (such as carbon nanotubes29 or graphite30) which decrease the amount of PCM in

the mixture, therefore reducing latent heat15. Capsule shell confines crystallohydrate composition to a limited volume, preventing water loss in the melted state and enhancing nucleation during cooling. Resulting micro and nanosized energy capsules can be used as active fillers for thermal energy storage in various environments: from textile materials31 to bulky heat exchangers32.As LHS is most effective for applications requiring a small temperature range, the melting temperature (T M) is crucial. It is important therefore to be able to tailor the TM to meet requirements15. For paraffin waxes, varying the amount of carbon atoms in the chain alters the TM

33, but in a limited range (5-70ºC)34. Mixing salt hydrates can give a similar effect, causing a flexible and controlled decrease in TM up to the eutectic point. It has been reported this effect is due to the PCMs inhibiting the crystallisation of each other35. For macroscale CSP heat exchangers, a single PCM for energy storage in applications has been shown to be inefficient. As the heat transfer fluid (HTF) flows along a PCM tank, its temperature decreases, creating a large temperature difference. Heat flow between PCM and HTF is consequently decreased. Using cascaded latent heat storage (CLHS), the temperature difference, and therefore heat flow, can be kept relatively constant36,37. CLHS consists of several PCMs arranged in order of decreasing TM with respect to the direction of HTF during charging. This was shown by Michels and Pitz-Paal 38 to have great benefits over single PCM systems. During charging of the CLHS system 92% of the PCM melted, and during discharge 67% froze. This is compared with 100% and 2% using a single PCM. Thus, energy transfer is greatly improved during discharge, due to constant heat flow. Current CSP exchangers would use heat capacitors with TMs of 300+ºC, so high temperature PCMs are required for encapsulation39. It has not been previously demonstrated that the implementation of CLHS improves efficiency of lower temperature applications such as passive air conditioning, where low-temperature PCMs can be used. It is possible that mixing salt hydrates together can improve their overall latent heat capacity at certain salt ratios40,41. In previous research, we showed magnesium nitrate hexahydrate (Mg(NO3)2·6H2O) can be made thermally stable over >100 cycles by nanoencapsulating it in a poly(ethyl-2-cyanoacrylate) (PECA) shell 42. Although other attempts at salt hydrate encapsulation have been successful, they did not display prolonged thermal cycling stability43–46. Our best results were obtained when lower amounts of surfactants were used as stabilisers for the initial inverse miniemulsion. Shearing by sonication facilitates the use of low amounts of surfactant 47. It was observed that the nanoencapsulation considerably reduced supercooling and phase separation during heat uptake and release without damaging the capsule shell. Here, we extended our approach to other crystallohydrates and their mixtures to demonstrate its universality as well as performance for thermal energy storage and controlled release. For this study, we encapsulated both single Mg(NO3)2·6H2O (TM = 89°C) and sodium sulphate decahydrate (Na2SO4·10H2O, TM = 32°C) and several mixtures of Mg(NO3)2·6H2O with Na2SO4·10H2O. Especially of interest was the effect on TM with different ratios of salt hydrates, and how encapsulation affects this. Two types of mixtures were investigated: (i) additive mixtures where single encapsulated crystallohydrates were mixed in different ratios, and (ii) energy nanocapsules containing pre-mixed crystallohydrates as core. Resulting energy capsules were characterised for their morphology, chemical stability and composition by scanning electron microscopy and Fourier transform infrared spectroscopy (FTIR), along with heat storage properties and stability by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA).

MethodsFormation of PECA nanocapsules with different compositions of crystallohydrate coreThe goal of the presented work is to investigate heat storage ability of the two crystallohydrate salts Mg(NO3)2.6H2O and Na2SO4·10H2O as well as their mixtures encapsulated into nanocapsules with poly(ethyl-2-cyanoacrylate) (PECA) shells. We have found in our previous studies 42 that the use of ultrasonic treatment on the first stages of the crystallohydrate encapsulation to form stable inverse miniemulsions is extremely important, in order to gain a high encapsulation yield and maintain salt hydrate composition.

Fig. 1 Formation of salt hydrate loaded capsules. (a) macroemulsion sheared by sonication to form inverse miniemulsion. (b) ECA dissolved in chloroform dropped in to form the PECA shell around aqueous phase. Depending on aqueous phase, nanocapsules are fabricated with (c) single salt hydrate core or (d) salt hydrate mixture core.

Fig. 1 schematically represents the overall encapsulation process. The aqueous phase containing crystallohydrates (1g) was added to the oil phase (9g) and stirred for an hour to create an initial macroemulsion (see Table 1 for preparation conditions for each sample). This mixture was then sonicated for 10 min (70% amplitude; 30s on, 20s off pulse regime) in an ice bath to prevent evaporation of solvent, using a Q700-220 sonicator (700W) (QSonica, USA) equipped with a 1/8” micro-tip. After sonication, the previously milky solution became transparent and stable against agglomeration. For the second stage, ECA monomer (200µL) in chloroform (600µL) was added dropwise to the miniemulsion and stirred magnetically for 4 hours at 400rpm in an open vial, to allow the chloroform to evaporate. A PECA shell formed around the aqueous miniemulsion droplets. The resulting capsule suspension was washed with ethanol and centrifuged (11000rpm, 10 min) to purify the capsules. The supernatant was removed, and the capsules allowed to dry overnight in a fume hood at ambient conditions.

CharacterisationCapsule morphology and size measurements were made using a JSM-7001F (JEOL, Japan) scanning electron microscope (SEM). Nanocapsules were diluted to 0.1 wt.% solid content in ethanol by bath sonication for 20 minutes. 50µL of the resulting dispersion was added to a glass cover slide attached to SEM stub, and dried at room temperature overnight. The samples were coated by chromium sputtering (100 Amp, 60s coating time) before measurement. Chemical analysis of the samples was made using a TENSOR II (Bruker, Germany) FTIR spectrometer equipped with an all-reflective diamond ATR. Transmittance measurements were taken on powder samples with 64 scans from 400 to 4000cm-1. Thermogravimetric analysis was taken using a SDT Q600 (TA Instruments, USA), providing data on thermal degradation behaviour of both bulk and encapsulated salt hydrates. Measurements were taken from room temperature up to 600ºC, with a ramp of 10ºC/min under a nitrogen atmosphere. DSC measurements were taken using a DSC6 (Perkin Elmer, USA) to study cycling stability of the energy nanocapsules during heat uptake and release. 100 cycles were undertaken between -20 to 120ºC with a ramp of 5ºC under a nitrogen atmosphere. Measurements of the salt hydrate mixture T Ms were taken using a DSC 214 (NETZSCH, Germany) with a ramp of 10ºC under a nitrogen atmosphere. TMs were averaged over the 2nd and 3rd cycles between -20 to 120ºC. X-ray Photoelectron Spectroscopy (XPS) measurements were made for NanoPCM1 and 2, using an Electron Energy Analyser (PSP, UK) with Al Kα X-ray source (hν =1486.6 eV, 12 kV, 144 W, pass energy = 20eV). The base pressure of the system was <2x10-10 mbar with H2 as the main residual gas. The NanoPCM samples were crushed to form a fine powder before measuring. XPS results can be found in the supporting information (SI) section.

Table 1 Preparation conditions for PECA nanocapsules with different crystallohydrate core.

Sample Aqueous phase Oil phaseNanoPCM1 20 wt.%

Mg(NO3)2·6H2O, 80 wt.% water

5 wt.% 3:2 Tween80:Span

20,95 wt.%

cyclohexane

NanoPCM2 20 wt.% Na2SO4·10H2O, 80 wt.% water

NanoPCM320 wt.% 1:1

Mg(NO3)2·6H2O:Na2SO4·10H2O

, 80 wt.% water

NanoPCM420 wt.% 1:2

Mg(NO3)2·6H2O:Na2SO4·10H2O

, 80 wt.% water

ResultsMorphology and chemical composition of nanoencapsulated crystallohydrates and their mixturesWe found in our previous study42 that an excess of water is necessary for successful nanoencapsulation of the crystallohydrates into PECA shell. This helps to avoid phase separation upon peritectic melting43 during ultrasonic treatment of crystallohydrate emulsions in the cyclohexane phase accompanied by formation of the PECA shell. The same effect was observed for Na2SO4·10H2O and its mixtures with Mg(NO3)2·6H2O. SEM images are shown in Fig. 2. Initial crystallohydrates are large >1µm crystals combined into agglomerates up to 100µm in size. After encapsulation, the size of the crystallohydrate core is reduced to 100-300nm with the smooth PECA shell providing complete coverage of the core, with no pores present. This is important for ensuring the core material is protected from the environment and prevention of leakage - the shell is a barrier preventing water from evaporating, therefore keeping the salt hydrate in its correct hydration state. Several of the images (please also see SI, Fig. S1) show how the PECA nanocapsules tend to aggregate, however single capsules can be clearly seen confirming elasticity of the capsule shell and its robustness to the SEM high vacuum (please also see SI Fig. S2). The irregular (not spherical) shape of the capsules is caused by solid form of the crystallohydrate core appearing after cooling of initially liquid crystallohydrate core during preparation of energy capsules.

Fig. 2 SEM images of bulk Mg(NO3)2.6H2O (A), Na2SO4.10H2O (B) and 1:1 Mg(NO3)2.6H2O:Na2SO4.10H2O mixture (C). (D), (E) and (F) demonstrate energy nanocapsules with Mg(NO3)2.6H2O, Na2SO4.10H2O and 1:1 Mg(NO3)2.6H2O:Na2SO4.10H2O core, respectively. Scale bar for (A)-(C) images is 1µm; for (D)-(F) images is 100nm

FTIR spectra for bulk salt hydrates and NanoPCMs are shown in Fig. 3 (please also see Fig. S4 in SI section to demonstrate reproducibility of the synthetic methods). Mg(NO3)2·6H2O (Fig. 3a) has characteristic peaks for O-H at 3356cm-1, N=O bending at 1646cm-1, a mix of N-O stretching and bending and N=O at 1365cm-1, and NO3

- at

819cm-1. Na2SO4·10H2O (Fig. 3b) has peaks for O-H at 3334cm -1, a mix of S-O and S=O at 1078cm -1 and the SO42-

group at 613cm-1. All NanoPCMs (Figs. 3c-3f) have characteristic PECA peaks: C-H stretch at 2927cm -1, C=O at 1747cm-1 and C-O ester peaks at 1249 and 1013cm -1. Unusually, only NanoPCM3 has a peak at 2361cm -1 signifying the presence of a CN group, which can be explained by the presence of ECA residue in the capsule shell. NanoPCM1 contains Mg(NO3)2·6H2O as shown by peaks at 3387cm-1 for O-H, N=O bend at 1657cm-1, a mix of N-O stretching and bending and N=O at 1373cm-1, and NO3

- at 827 cm-1. NanoPCM2 contains Na2SO4·10H2O, it has a very small peak for O-H at 3539cm-1, S-O and S=O in the broad peak at around 1121cm -1 and SO4

2- at 620cm-1. NanoPCMs 3 and 4 contain both Mg(NO3)2·6H2O and Na2SO4·10H2O. They have peaks at 3259 and 3327cm-1

respectively for O-H, 1651 and 1657cm-1 respectively for N=O bend, 1370 and 1373cm-1 respectively for a mix of N-O stretching and bending and N=O, 1084cm-1 for S-O and S=O, and 613cm-1 for SO4

2-. NanoPCM3 has a small peak at 827cm-1 for NO3

-; this peak shifts to 830cm-1 for NanoPCM4 with the increase of Na2SO4·10H2O content in the capsule core. As one can see, all FTIR signals match the core/shell components of the corresponding nanocapsule sample. Small deviations between wavenumbers of the bulk crystallohydrates and nanoencapsulated ones are due to spatial confinement effect, which influences the H-bonding inside crystallohydrate core 42. XPS measurements (please see SI Fig. S3 and Table S1) also confirmed the presence of crystallohydrates in the capsule core.

Fig. 3 FTIR spectra for (a) Mg(NO3)2·6H2O, (b) Na2SO4·10H2O, (c) NanoPCM1, (d) NanoPCM2, (e) NanoPCM3 and (f) NanoPCM4.

Thermal characterisationTGA analysis (Fig. 4) demonstrates fast degradation of both crystallohydrates in the pure state starting at 40°C for Na2SO4·10H2O and about 90°C for Mg(NO3)2·6H2O. Mg(NO3)2·6H2O showed 32% mass loss by 250°C related to water evaporation. Remaining mass was lost from 290 to 470°C due to Mg(NO 3)2 decomposition. 21% mass remains after 600°C, mainly consisting of magnesium oxide48. Na2SO4·10H2O has slightly different behaviour, losing all its water at 115°C (55% mass) without further decomposition of anhydrous Na 2SO4 within the investigated temperature range. The weight loss observed for NanoPCM4 up to 100°C may also be due to residual solvent (cyclohexane) along with any free water, which would explain why this sample starts weight loss earlier than NanoPCMs 2 and 3. All encapsulated PCM samples exhibit better thermal stability caused by PECA capsule shell protection against water evaporation from crystallohydrates at elevated temperatures. Depending on the quantity of free water remaining after capsule synthesis, the loss of water up to 150°C appears in all NanoPCM samples, which coincide with DSC results in Fig. 5 and 6 (see below). 16% water loss was observed for NanoPCM1. Increasing the Na2SO4·10H2O content in the capsule samples reduces the water loss for NanoPCM2-4 samples. There is an increase in the rate of mass loss after 150°C, corresponding with the decomposition of the PECA shell for all samples and, additionally, to the thermal decomposition of Mg(NO3)2 into magnesium oxide species. TGA data of the samples become stable after 450°C. The remaining mass at 600°C is 20% for NanoPCM1 and increases with more Na2SO4·10H2O content in the capsule core to 22%(NanoPCM3), 24% (NanoPCM4) and 29% (NanoPCM2).

Fig. 4 TGA curves for crystallohydrates and NanoPCMs.

Crystallohydrate nanocapsulesEncapsulated crystallohydrates and their mixtures were studied by DSC to find out their thermal stability during heat uptake and release cycles. Thermograms of pure crystallohydrate salts (Fig. 5a, b) demonstrate poor cycling stability and decompose after only 10 cycles. The thermogram for bulk Mg(NO3)2·6H2O (Fig. 5a) provides a latent heat of 160.4 J·g-1 and a melting point with Tonset, M = 83°C and TM = 93°C, which does not differ much from literature values of 162.8 J·g-1 for latent heat and 89°C for TM

21. The broad signal during melting can be associated with uneven size of the crystallohydrate particles. The main freezing peak has Tonset, F = 60°C and TF = 55°C. The difference between TM and TF of the bulk Mg(NO3)2·6H2O is 38°C, which is the supercooling (TM – TF). Much smaller peaks observed at 42°C and 68°C can be attributed to a solid-solid transition between different crystallohydrate phases. Bulk Na2SO4·10H2O has two endothermic peaks: the first one has Tonset, M = -2°C and TM = 3°C and the second one has Tonset, M = 26°C and TM = 32°C. First small peak can be assigned either to water melting or to Na2SO4·7H2O/Na2SO4·10H2O transition. Second peak is attributed to Na2SO4·10H2O solid/liquid transition (transition temperature 32.4°C, latent heat 254 J·g-1 from ref. 20). This is also confirmed by the latent heat (228.1 J·g-1) derived from DSC data for the bulk Na2SO4·10H2O sample. Exothermic peaks observed during crystallisation of Na2SO4·10H2O found the reverse transition processes: Tonset, F = 13°C and TF = 4°C for Na2SO4·10H2O solid/liquid transition and a small one at TF = -11°C for Na2SO4·7H2O and water crystallisation. The supercooling effect for bulk Na2SO4·10H2O is 28°C, similar to Mg(NO3)2·6H2O. Both pure crystallohydrates demonstrated rapid decay in their heat uptake/release properties during next DSC cycles. The melting and freezing peaks are strongly decreased (Figs 5a, b, 10th cycle), and signals are lost completely as the PCM dehydrates to Mg(NO3)2 and Na2SO4 salts together with less hydrated components (most likely Mg(NO3)2·2H2O and Na2SO4·7H2O). In general, the rapid reduction of heat uptake/release parameters together with large supercooling effect is the main inherent shortcoming for the application of crystallohydrates in their pure form for thermal energy storage. Encapsulation drastically increases stability of each crystallohydrate. Fig. 5c and d display DSC curves of the NanoPCM1 and NanoPCM2 samples containing encapsulated Mg(NO3)2·6H2O and Na2SO4·10H2O, respectively. NanoPCM1 does not undergo thermal degradation upon 100 heat uptake/release cycles, having a latent heat of 88.4 J·g-1 after 100 cycles, Tonset, M = 88°C, TM = 92°C and Tonset, F = 86°C, TF = 82°C and supercooling of only 10°C. The first DSC cycle also demonstrated broad peak at around 100 – 110°C, which is attributed to the evaporation of excess water appearing during synthesis of NanoPCM1. This water loss may be also caused by capsules damaged during formation of the capsule shell. This peak disappears during the next 2-3 cycles. Heat uptake/release properties of the NanoPCM1 remain almost constant during continuous DSC cycling between -20°C and 120°C with only 3% decrease in latent heat after the first cycle. NanoPCM2 also reveals stable characteristics during heat uptake and release. First DSC cycle showed Tonset, M = 26°C, TM = 32°C and Tonset, F = 21°C, TF = 17°C together with a small endothermic peak at 105°C related to free water remained after encapsulation. There were no other signals on DSC which can be associated to other Na2SO4 crystallohydrate phases, contrary to the DSC results for bulk Na2SO4·10H2O. Latent heat of 138.6 J·g-1 was observed after 100 cycles with 2% of reduction from the first heat uptake/release cycle. The supercooling effect was decreased to 15°C from 28°C found for bulk Na2SO4·10H2O. Both NanoPCM1 and NanoPCM2 samples with single crystallohydrates encapsulated into PECA shell demonstrated high cycling stability and reduced supercooling. Incongruent melting was therefore prevented by the PECA capsule shell terminating water exchange with the environment. Encapsulation efficiency, estimated

from the ratio between latent heat of the encapsulated salt hydrates and bulk salt hydrates, is 54% for Mg(NO3)2·6H2O (NanoPCM1) and 61% for Na2SO4·10H2O (NanoPCM2).

Fig. 5 DSC thermograms for (A) bulk Mg(NO3)2·6H2O, (B) bulk Na2SO4·10H2O , (C) NanoPCM1, (D) NanoPCM2.

Additive mixtures of crystallohydrate-based nanocapsulesFor the next stage, we investigated the potential of the encapsulated crystallohydrate mixtures for heat storage in multi-temperature cascade energy storage systems. Two types of mixtures were explored: (i) NanoPCM1 and NanoPCM2 single crystallohydrate capsules mechanically mixed in different ratios and (ii) NanoPCM3 and NanoPCM4 capsules containing crystallohydrate mixtures as core component. Fig. 6 demonstrates heat uptake/release behaviour of the first type of the mixtures, where NanoPCM1 and NanoPCM2 were mechanically mixed in 1:2, 1:1 and 2:1 weight ratios. 1:1 mixture of NanoPCM1 and NanoPCM2 exhibited additive effects on DSC curves attributed to both Mg(NO3)2·6H2O and Na2SO4·10H2O cores (Fig. 6a). First DSC cycle revealed peaks indicating Na2SO4·10H2O phase transition (Tonset, M = 24°C, TM = 32°C and Tonset, F = 19°C, TF = 12°C), Mg(NO3)2·6H2O phase transition (Tonset, M = 86°C, TM = 94°C and Tonset, F = 83°C, TF = 75°C) and peak between 100–110°C related to evaporation of free water remaining after encapsulation. The latter peak disappears during the second DSC cycle. No considerable changes were observed after second cycle and 1:1 NanoPCM1:NanoPCM2 mixture remains stable for >100 DSC cycles. Tonset, M = 24°C, TM= 32 °C and Tonset, F = 19°C, TF = 13°C peaks were recorded for Na2SO4·10H2O and Tonset, M = 87°C, TM = 94°C and Tonset, F = 84°C, TF = 75°C peaks were recorded for Mg(NO3)2·6H2O with latent heat of 67.2 J·g-1 for Na2SO4·10H2O and 44.1 J·g-1 for Mg(NO3)2·6H2O transitions. The supercooling effect for the mixed sample reached 19°C for both Mg(NO3)2·6H2O and Na2SO4·10H2O, which is considerably higher than supercooling temperatures for initial NanoPCM1 (10°C) and NanoPCM2 (15°C). Supercooling is affected when the nanocapsule ratio is not 1:1. Fig. 6b indicates that changing the ratio is accompanied by changes in latent heat for each phase transition in NanoPCM1 and NanoPCM2 mixtures. Decreasing the quantity of any capsule component in the mixture leads to the increase of the supercooling effect for the minor component and its reduction in the major one. Doubled NanoPCM1 content resulted in TM = 33°C and TF = 12°C for Na2SO4·10H2O and TM = 93°C and TF = 84°C for Mg(NO3)2·6H2O, reducing supercooling for encapsulated Mg(NO3)2·6H2O to 9°C and increasing for encapsulated Na2SO4·10H2O to 21°C. On the other hand, doubled NanoPCM2 quantity lead to TM = 32°C and TF = 15°C for Na2SO4·10H2O and TM = 97°C and TF = 74°C for Mg(NO3)2·6H2O with supercooling ΔT = 23°C for encapsulated Mg(NO3)2·6H2O and ΔT = 17°C for encapsulated Na2SO4·10H2O. Changes of the latent heat for each phase transition were observed in accordance to the proportion between Mg(NO3)2·6H2O and Na2SO4·10H2O in the mixtures: 59 J·g-1:46.1 J·g-1 for 2:1 NanoPCM1:NanoPCM2 and 29.1 J·g-1:93.1 J·g-1 for 1:2 NanoPCM1:NanoPCM2.

It is well known that supercooling effect for PCMs is caused by poor thermal conductivity of the material during heat uptake or release. The additive mixing of NanoPCM1 and NanoPCM2 enhances thermoinsulating properties at each phase transition process where one encapsulated component of the additive mixture can reduce heat transfer during phase transition of the second component. Capsules with Mg(NO 3)2·6H2O core remain intact and do not absorb heat during phase transition in the capsules with Na2SO4·10H2O core. Hence, NanoPCM1 capsules play a role of heat insulator for NanoPCM2 capsules at 10-30°C temperatures and the more NanoPCM1 capsules are in the mixture, the greater the supercooling effect for NanoPCM2 capsules. The same is observed during Mg(NO3)2·6H2O core transition in the capsule mixture. Na2SO4·10H2O core in NanoPCM2 capsules remains liquid and not active for heat transfer at phase transition temperature for Mg(NO3)2·6H2O core (70-100°C) and supercooling increases with increasing NanoPCM2 content. Contrary to the capsule mixtures, NanoPCM1 and NanoPCM2 samples with single crystallohydrate cores undergo phase transition within the same temperature range across the whole sample and remain thermally active for heat transfer. This is the most evident explanation for the supercooling effects observed in the additive energy capsule mixtures.

Fig. 6 DSC thermograms for (A) 1:1 NanoPCM1:NanoPCM2 ratio at different cycles, (B) different NanoPCM1:NanoPCM2 ratio on 100th heat uptake cycle.

Nanoencapsulated crystallohydrate eutecticsOne of the main advantages of the crystallohydrates as PCMs for thermal energy storage is their possibility to make mixtures with controlled phase transition temperatures defined by the ratio between mixed crystallohydrate salts49. This phenomenon has been known for many years; however, its direct application in thermal energy storage systems faced the shortcomings similar to the single crystallohydrates, but to a higher extent. The melting of the first component of the mixture initiates dissolution of the components with higher transition temperatures by formed excess of free, non-bound water molecules 40. On the other hand, the crystallisation process occurs in the mixture of different salts with an excess of water for the higher TM

crystallohydrate component and a possible water deficit for phases with lower transition temperatures. This can lead to the formation of several new crystallohydrate phases during heat uptake/release cycles, which can be in non-equilibrated conditions to each other and their weight balance may differ in each energy uptake/release cycle. Crystallohydrate mixtures can reach equilibrated conditions at a certain ratio between crystallohydrate components (eutectic point)34. At this point, crystallohydrates form a single phase with a defined T M. Eutectic mixtures have transition temperatures lower than that of the single components 50 and, at the same time, phases of the single components disappear. The influence of the Mg(NO3)2·6H2O:Na2SO4·10H2O ratio on the melting temperature of their pure, non-encapsulated mixture is shown in Fig. 7 (T M has been derived from DSC data of the corresponding salt hydrate mixtures except TM for pure Mg(NO3)2·6H2O and Na2SO4·10H2O taken from Ref. 20).Mixing the two salt hydrates alters their TM, with all mixtures giving a lower TM than pure Mg(NO3)2·6H2O. The TM

decreases at 10, 20 and 25 wt.% of Mg(NO3)2·6H2O, until the eutectic point at around 33.3 wt.% (TM = 15.0ºC) is achieved. The TM increases greatly from 40 to 66.6 wt.% of Mg(NO3)2·6H2O (16.7ºC to 67.4ºC). Curiously, the TM

decreases at 75 and 80 wt.% Mg(NO3)2·6H2O (57.5ºC and 57.7ºC), exhibiting a second eutectic ratio before increasing at 90 wt.% (79.1ºC). A similar ‘double bell’ shaped curve has been previously observed with mixtures of Mg(NO3)2·6H2O and aluminium nitrate nonahydrate51.

From the data depicted in Fig. 7, we selected two Mg(NO 3)2·6H2O:Na2SO4·10H2O mixtures for further investigation of their heat uptake/release behaviour in the PECA-encapsulated form: (i) 1:1 wt.% ratio, which is far above the eutectic point (TM = 45.5°C, NanoPCM3), and (ii) 1:2 wt. % eutectic ratio (T M= 15.0ºC, NanoPCM4). Both pure bulk mixtures are not stable during heat uptake/release cycles (Figure 8a, b) and phase transition peaks on DSC thermogram are almost negligible during 10th cycle. However, mixtures demonstrate different thermal properties during first cycles. 1:1 wt.% Mg(NO3)2·6H2O:Na2SO4·10H2O sample indicates three peaks on heating. The first peak around 0°C can be assigned to the melting of free water, but the other at T M = 45.2°C and a broad one at TM = 67.1°C clearly show two separate phase transitions in the 1:1 bulk salt mixture. On the cooling step, only a broad peak at ca. 4.3°C is seen proving incongruent crystallisation and water loss during DSC cycle within -20°C - +120°C temperature range. The latent heat capacity is 103.3 J·g-1, which is lower than the additive sum of the crystallohydrates (194.3 J·g-1, please see above). 1:2 wt.% Mg(NO3)2·6H2O:Na2SO4·10H2O mixture has a melting peak at TM = 15.9°C and the latent heat of 189.3 J·g-1, almost equal to the additive sum of the crystallohydrates in the mixture (205.5 J·g-1). The high latent heat can further be explained by additional thermochemical hydration heat40. Taking into account data from Fig. 7, 1:2 wt.% Mg(NO3)2·6H2O:Na2SO4·10H2O mixture forms a salt hydrate eutectic with phase transition at 15.7°C, the shoulder until ca. 45°C is most likely due to excess water. The freezing transition peak at the eutectic point is more pronounced with TF = -11.4ºC, supercooling is 27.1°C. Encapsulation of the crystallohydrate mixtures prevents the loss of water during prolonged thermal cycling and reduces supercooling for both NanoPCM3 and NanoPCM4 samples (Fig. 8c, d). Two peaks can be found for NanoPCM3 during heating and one broad peak is recognised during cooling (Fig. 8c). Transition temperatures are not stable with any heat uptake/release cycle and difficult to predict. These inconclusive thermal properties of the capsule core are caused by the presence of several crystallohydrate phases inside the PECA capsule shell. Melting and crystallisation processes in the mixed phases have an uncertain nature and the capsule core has several phases in which the ratio is dependent on stochastic nucleation processes. Encapsulated 1:2 wt.% Mg(NO3)2·6H2O:Na2SO4·10H2O mixture (NanoPCM4 sample) has, on the contrary, one well defined phase transition peak with TM = 15.4°C and TF = -1.1°C and reduced supercooling of ΔT=16.5°C. The transition is stable over >100 heat uptake/release cycles and has latent heat capacity of 126.8 J ·g-1, which results in 67% encapsulation efficiency. The effect of nanoencapsulation on salt hydrate mixtures is similar to its effect on single crystallohydrates, producing thermally stable energy capsules. Please also see the summary of thermal parameters in Table S2 of SI.

Fig. 7 Influence of the Mg(NO3)2·6H2O:Na2SO4·10H2O wt% ratio on melting temperature of the corresponding mixture

Fig. 8 DSC thermograms for (A) 1:1 wt% Mg(NO3)2·6H2O:Na2SO4·10H2O bulk mixture, (B) 1:2 wt% Mg(NO3)2·6H2O:Na2SO4·10H2O bulk mixture, encapsulated 1:1 wt% Mg(NO3)2·6H2O:Na2SO4·10H2O (C, NanoPCM3) and encapsulated 1:2 wt% Mg(NO3)2·6H2O:Na2SO4·10H2O (D, NanoPCM4).

Chemical stability of NanoPCMsFTIR spectra in Fig. 9 confirm the stability of the crystallohydrate cores for all NanoPCM samples after 100 DSC cycles. They are largely identical to the initial spectra in Fig. 3, with no peak shifting. The only minor differences are intensities of the peaks. This is displayed most prominently in the spectrum for NanoPCM3, with the peaks at 2361 and 1651cm-1 having reduced intensity. These results show that the NanoPCMs shell and core material are chemically stable during prolonged thermal cycling, which is of great benefit for energy storage capsules.

Fig. 9 FTIR spectra for (A) NanoPCM1, (B) NanoPCM2, (C) NanoPCM3 and (D) NanoPCM4 samples after 100 DSC thermal cycles.

ConclusionsWe demonstrated a universal approach for the encapsulation of crystallohydrate salts (Mg(NO3)2·6H2O and Na2SO4·10H2O) into poly(ethyl-2-cyanoacrylate) nanocapsules of 100-200nm in size by in situ inverse miniemulsion polymerisation employing ultrasound to create inverse miniemulsions. FTIR analysis confirmed the presence of the crystallohydrates in the nanocapsule cores both before and after 100 heat uptake/release cycles. DSC results demonstrated high thermal stability of nanoencapsulated single and mixed crystallohydrates, which remained unchanged after 100 thermal cycles. Two types of crystallohydrate mixtures were investigated: (i) additive mixtures of nanocapsules with single crystallohydrate cores and (ii) nanocapsules with mixed Mg(NO3)2·6H2O and Na2SO4·10H2O crystallohydrate cores. Additive mixtures of crystallohydrate-loaded nanocapsules have high potential to design multi-temperature heat storage systems containing energy capsules with different PCM cores sensitive to different transition temperatures. Encapsulation of mixtures of two or more crystallohydrates can reach lower transition temperatures at their eutectic points, with high latent heat. However, predictable congruent melting/freezing and, as a consequence, consistent heat uptake/release was not achieved at a crystallohydrate ratio different to the eutectic mixture due to the formation of several phases in the capsule core. Our future research is focused on improving thermal conductivity of the capsule shell by introducing graphite components, as well as on the encapsulation effect of reducing crystallohydrate corrosiveness towards container materials in applied heat storage systems.

Acknowledgements

We thank ERC Enercapsule project (project number 647969) for supporting this research.

References

1 P. Felix De Castro, A. Ahmed and D. G. Shchukin, Chem. - A Eur. J., 2016, 22, 4389–4394.2 C. Y. Zhao and Z. G. Wu, Sol. Energy Mater. Sol. Cells, 2011, 95, 3341–3346.3 Y. Wang, B. Tang and S. Zhang, Adv. Funct. Mater., 2013, 23, 4354–4360.4 A. Gil, M. Medrano, I. Martorell, A. Lázaro, P. Dolado, B. Zalba and L. F. Cabeza, Renew. Sustain. Energy Rev., 2010,

14, 31–55.5 D. W. Hawes, D. Feldman and D. Banu, Energy Build., 1993, 20, 77–86.6 L. F. Cabeza, J. Illa, J. Roca, F. Badia, H. Mehling, S. Hiebler and F. Ziegler, Mater. Corros., 2001, 52, 140–146.7 N. Zhu, Z. J. Ma and S. W. Wang, Energy Convers. Manag., 2009, 50, 3169–3181.8 C. Voelker, O. Kornadt and M. Ostry, Energy Build., 2008, 40, 937–944.9 L. F. Cabeza, A. Castell, C. Barreneche, A. De Gracia and A. I. Fernández, Renew. Sustain. Energy Rev., 2011, 15,

1675–1695.10 E. M. Alawadhi and C. H. Amon, IEEE Trans. Components Packag. Technol., 2003, 26, 116–125.11 M. J. Huang, P. C. Eames and B. Norton, Int. J. Heat Mass Transf., 2004, 47, 2715–2733.12 A. Hasan, S. J. McCormack, M. J. Huang, J. Sarwar and B. Norton, Sol. Energy, 2015, 115, 264–276.13 D. C. Hyun, N. S. Levinson, U. Jeong and Y. A. Xia, Angew. Chemie-International Ed., 2014, 53, 3780–3795.14 P. B. Salunkhe and P. S. Shembekar, Renew. Sustain. Energy Rev., 2012, 16, 5603–5616.15 M. K. Rathod and J. Banerjee, Renew. Sustain. Energy Rev., 2013, 18, 246–258.16 S. Seddegh, X. Wang, A. D. Henderson and Z. Xing, Renew. Sustain. Energy Rev., 2015, 49, 517–533.17 M. T. Johansson and M. Söderström, Energy Effic., 2014, 7, 203–215.18 M. M. Kenisarin, Renew. Sustain. Energy Rev., 2010, 14, 955–970.19 M. M. Farid, A. M. Khudhair, S. A. K. Razack and S. Al-Hallaj, Energy Convers. Manag., 2004, 45, 1597–1615.20 A. Abhat, Sol. Energy, 1983, 30, 313–332.21 B. Zalba, J. M. Marin, L. F. Cabeza and H. Mehling, Appl. Therm. Eng., 2003, 23, 251–283.22 A. Sari, C. Alkan, A. Karaipekli and O. Uzun, Sol. Energy, 2009, 83, 1757–1763.23 A. Sari, C. Alkan, A. Biçer, A. Altuntaş and C. Bilgin, Energy Convers. Manag., 2014, 86, 614–621.24 P. Felix De Castro and D. G. Shchukin, Chem. - A Eur. J., 2015, 21, 11174–11179.25 K. Tumirah, M. Z. Hussein, Z. Zulkarnain and R. Rafeadah, Energy, 2014, 66, 881–890.26 G. H. Zhang, S. A. F. Bon and C. Y. Zhao, Sol. Energy, 2012, 86, 1149–1154.27 Z. Zheng, J. Jin, G.-K. Xu, J. Zou, U. Wais, A. Beckett, T. Heil, S. Higgins, L. Guan, Y. Wang and D. Shchukin, ACS Nano,

2016, 10, 4695–4703.

28 Z. Zheng, Z. Chang, G.-K. Xu, F. McBride, A. Ho, Z. Zhuola, M. Michailidis, W. Li, R. Raval, R. Akhtar and D. Shchukin, ACS Nano, 2017, 11, 721–729.

29 L. W. Fan, X. Fang, X. Wang, Y. Zeng, Y. Q. Xiao, Z. T. Yu, X. Xu, Y. C. Hu and K. F. Cen, Appl. Energy, 2013, 110, 163–172.

30 H. Ji, D. P. Sellan, M. T. Pettes, X. Kong, J. Ji, L. Shi and R. S. Ruoff, Energy Environ. Sci., 2014, 7, 1185–1192.31 S. Mondal, Appl. Therm. Eng., 2008, 28, 1536–1550.32 Y. Allouche, S. Varga, C. Bouden and A. C. Oliveira, Energy Convers. Manag., 2015, 94, 275–285.33 T. Kousksou, A. Jamil, T. El Rhafiki and Y. Zeraouli, Sol. Energy Mater. Sol. Cells, 2010, 94, 2158–2165.34 K. Pielichowska and K. Pielichowski, Prog. Mater. Sci., 2014, 65, 67–123.35 P. W. Stott, A. C. Williams and B. W. Barry, J. Control. Release, 1998, 50, 297–308.36 M. Liu, N. H. S. Tay, S. Bell, M. Belusko, R. Jacob, G. Will, W. Saman and F. Bruno, Renew. Sustain. Energy Rev.,

2016, 53, 1411–1432.37 M. Liu, W. Saman and F. Bruno, Renew. Sustain. Energy Rev., 2012, 16, 2118–2132.38 H. Michels and R. Pitz-Paal, Sol. Energy, 2007, 81, 829–837.39 S. Pendyala, P. Sridharan, S. Kuravi, C. K. Jotshi, M. K. Ram, M. Rahman, E. Stefanakos and D. Y. Goswami, Proc.

ASME 6th Int. Conf. Energy Sustain. - 2012, Pts A B, 2012, 593–597.40 K. Posern and C. Kaps, Thermochim. Acta, 2010, 502, 73–76.41 K. Nagano, K. Ogawa, T. Mochida, K. Hayashi and H. Ogoshi, Appl. Therm. Eng., 2004, 24, 221–232.42 M. J. Graham, E. Shchukina, P. Felix De Castro and D. G. Shchukin, J. Mater. Chem. A, 2016, 4, 16906–16912.43 D. Platte, U. Helbig, R. Houbertz and G. Sextl, Macromol. Mater. Eng., 2013, 298, 67–77.44 A. Schoth, K. Landfester and R. Muñoz-Espí, Langmuir, 2015, 31, 3784–3788.45 J. Huang, T. Y. Wang, P. P. Zhu and J. B. Xiao, Thermochim. Acta, 2013, 557, 1–6.46 F. Salaün, E. Devaux, S. Bourbigot and P. Rumeau, Carbohydr. Polym., 2008, 73, 231–240.47 L. C. Peng, C. H. Liu, C. C. Kwan and K. F. Huang, Colloids Surfaces A Physicochem. Eng. Asp., 2010, 370, 136–142.48 J. Madarasz, P. P. Varga and G. Pokol, J. Anal. Appl. Pyrolysis, 2007, 79, 475–478.49 B. Li, D. Zeng, X. Yin and Q. Chen, J. Therm. Anal. Calorim., 2010, 100, 685–693.50 S. S. Chandel and T. Agarwal, Renew. Sustain. Energy Rev., 2017, 67, 581–596.51 Y. Marcus, A. Minevich and L. Ben-Dor, Thermochim. Acta, 2004, 412, 163–170.