Experimental Study of Thermo-Physical Characteristics of Molten Nitrate Salts Based Nano Fluids for Thermal Energy Storage

 GengQiao¹, Hui Cao², Feng Jiang³, XiaohuiShe^2,*, Lin Cong², Qing Liu³, Xianzhang Lei¹, Alessio Alexiadis² and Yulong Ding²

¹Global Energy Interconnection Research Institute Europe GmbH, Berlin, Germany

²BirminghamCentre for Energy Storage, University of Birmingham, Birmingham, UK

³University of Science and Technology Beijing, Beijing, China

*E-mail: X.She@bham.ac.uk

 

ABSTRACT:

Thermal energy storage (TES) is popular to shift peak-load, eliminate the intermittencyof renewable energy andrecoverindustrial waste heat. Molten nitrate salts arethe TESmaterials widely used in solar power plants. Pure nitrate saltsshow some unfavourable properties and hence composite salts are usually required. Adding nanoparticles into nitrate salts was reported as a good method to enhance their properties. However, it does not always show positive results andthe enhancements are sensitive to a number of parameters. Till now, there is also no a robust theoretical model to explain these phenomenadue to the lack of enough experimental data. Hence, more experimental studies are neededfor promoting the development of theoretical models. In this paper, three molten nitrate salts based nanofluids are synthesizedby doping SiO₂ nanoparticles into three popular single salts (NaNO₃, LiNO₃ and KNO₃). Effects of nanoparticle sizes (15 nm - 5 µm), mass fractions (0.5-4%) and temperature (200-380 ^oC) are considered. Thermo-physical properties of the composite materials are tested, includingmaterial structure, melting point,latent heat,specific heat capacity and thermal conductivity.Results show that the addition of SiO₂ nanoparticles has little effect on melting points of LiNO₃ and KNO₃, but it increases their latent heat by ~1.7%; for NaNO₃, itsmelting points are decreased by up to 2.5 ^oC whilelatent heat keeps stable with the reasonable SiO₂ nanoparticle size and mass fraction. The specific heat capacitiesof NaNO₃ and LiNO₃are enhanced by 27.6% and 12.3% with the addition of 60-70nm 2%and 0.5% SiO₂ nanoparticles, respectively; for KNO₃, the highest enhancement of 26% is achievedwith 15-20nm 1% SiO₂ nanoparticles. It is surprising to find that the addition of SiO₂ nanoparticles leads to adecreased thermal conductivity for all the three nitrate salts, which is probably due to the thermal resistance. This paper provides comprehensive and valuable experimental data, contributing to the development of robust models for predicting the effect of nanoparticles.

Keywords: Thermal energy storage; Nanoparticle; Molten salt; Nanofluid; Renewable energy

1. Introduction

Renewable energy is increasingly popular in recent years to combat climate changes and reduce carbon emissions. By 2017, renewable energy accounted for 10.3% of total global energy consumption for heat.¹ However, renewable energy is usually intermittent and hence there is a mismatch between energy supply and demand. Thermal energy storage (TES) provides an avenue to eliminate the uncertainty and inconsistency of renewable energy. This is accomplished by storing energy with TES materials at off-peak times when the supply exceeds demands and then releasingenergy at peak times or whenever it's needed.

Amongst different TES materials, molten salts attract much attention as they offer favourable specifications, such as wide melting temperatures of 250-1580 ^oC,² low vaporpressure, low viscosity and good chemical stability.³ Nitrate salts, as one kind of molten salts, have been widely used as energy storage medium and heat transfer fluid in solar power plants, such as the 19.9 MW TorresolGemasolar power tower⁴ and 49.9 MWAndasol power plant in Spain.⁵ In real applications, a composite nitrate salt is used rather than apure salt. The advantages include enhanced heat transfer, desired melting point, higher energy storage density, etc. A number of methods have been proposed to manufacture various composite nitrate salts to improve their different properties, including form-stable nitrate salts^{6, 7} for enhancing heat transfer, microencapsulation^{8, 9} to improve thermal stability and avoid corrosivity, multicomponent nitrate salts^10-14 for lowering melting points, molten salt based nanofluids^15-33 to improve energy storage density, etc.

Molten salt based nanofluid wasprepared by adding a small amount of nanoparticleswith one dimension smaller than 100 nm into base salts. It was first studied by Shin and Banerjee¹⁵ in 2011. They proposed three independent competing inter-molecular interaction mechanisms. One of them suggested that the interfacial thermal resistance between the nanoparticles and base liquids provided extra thermal storage capability.Since then, numerous studies have been conducted.¹⁶ It was reported thatenhancements of specific heat, thermal conductivity and energy storage capacity were observed in molten nitrate salts due to the addition ofCuO,^17-19 Al₂O₃²⁰ and SiO₂^21-24 nanoparticles.Luo et al.¹⁷ synthesized composite nitrate salts composedof binary salt (60 wt.% NaNO₃-40 wt.% KNO₃) and CuO nanoparticles (~20 nm). The maximum enhancements of specific heat and total storage capacitywere 11.48% for the liquidphase and 4.71% over the base salt, respectively. Jr et al.¹⁸ focusedon three nitrate salts: KNO₃, NaNO₃and KNO₃-NaNO₃ (51.8:48.2 mole%) with CuO as nanoparticles (~40 nm). The improvements in thermal conductivity were observed for all the measurements. Hu et al.²⁰ investigated the effect of doping eutectic binary salt (NaNO₃-KNO₃) with Al₂O₃ nanoparticles (20 nm) on its specific heatcapacity.Results showed that the enhancement of the specific heatcapacity ranged from 1.9% to 8.3% with the mass fraction of nanoparticles varying from 0.5% to 2%. Seoand Shin²¹ doped different sizes of SiO₂ nanoparticles (5-60 nm) into ternary nitrate salt eutectic LiNO₃-NaNO₃-KNO₃ (38:15:47 mole%) at a 1% mass fraction.The specific heat of the mixture was enhanced by 13-16% and nanoparticle size had little effect on the specific heat. Chen et al.²³ studied molten salt based nanofluids which possessed low-melting point salt Ca(NO₃)₂·4H₂O-KNO₃-NaNO₃-LiNO₃ (2:6:1:2 wt.%) as base liquid and SiO₂ asnanoparticleswith a diameter of 20 nm. The specific heat was improved by 24.5% ata mass fraction of 0.5%. The mechanism of the enhancements was also disclosed. It was reported that the solid-like nanolayer around the surface of the nanoparticles^25-27 and fractal-like fluid nanostructures^27-29 contributed to the enhancement of specific heat capacity and thermal conductivity.

On the other hand, some research showed thatsometimes the addition of nanoparticles resulted inunfavourable properties ofmolten nitrate salts.^30-33 Chieruzzi et al.³⁰ dispersed SiO₂ (7 nm), Al₂O₃ (13 nm) and a mix of SiO₂/Al₂O₃ (2-200 nm) nanoparticles into NaNO₃-KNO₃ solar salt (60-40 wt%) at 300 °C, respectively. The results showed that significant enhancement of thespecific heat capacity only occurred for the solar saltwith SiO₂/Al₂O₃ nanoparticles. Nanofluids consisting of NaNO₃-KNO₃ (60-40 wt.%) and Al-Cu nanoparticles (80-20 wt.%, 20-300 nm) were prepared by Navarrete et al.³¹ The specific heat was decreased by up to 10% at all mass fractions of nanoparticles. What’s more, the thermal conductivity was decreased by ~24% at the 10 wt.% nanoparticles. Awad et al.³² used a binary nitrate salt NaNO₃-KNO₃(60-40 mole%) as the base salts dispersed withFe₂O₃ (20-40 nm), TiO₂ (~50 nm) and CuO (<50 nm)nanoparticles, respectively. The Fe₂O₃ and CuOnanoparticles showed positive effect on the properties of nitrate salts. However, theTiO₂nanoparticles decreased the specific heat of liquid phase andtotal energy storage capacity by 14.6% and 12%, respectively. Grosu et al.³³ tested the corrosivity of HitecXL salt (15 wt.% NaNO₃, 43 wt.% KNO₃ and 42 wt.% Ca(NO₃)₂) doped with two different nanoparticles Al₂O₃ (13 nm) and SiO₂ (12 nm). It was found that the corrosion rate ofcarbon steelwas increased by 2-3 times due do the fact that the nanoparticles entrapped additional air in the interparticle pores which accelerated the corrosion.

As shown above, it is quite controversial regarding the effect of adding nanoparticles into molten nitrate salts, andthere is not awell-acknowledged model or theorythat explains the anomalous behaviour of moltensalt based nanofluids. Therefore, more experimental and theoretical studies areneededconsidering various parameters, such as different base salts, nanoparticle types,nanoparticle sizes,nanoparticle concentration, materialtemperatures, material preparation methods, etc.In this paper, three molten nitrate salts based nanofluids are prepared with single nitrate salts LiNO₃, NaNO₃and KNO₃ as the base salts and SiO₂ as the nanoparticles. Experimental studies are conducted considering different nanoparticle sizes (15 nm - 5 µm), mass fractions (0.5 - 4%) and temperatures (200 - 380 ^oC). The effects of nanoparticles on thermo-physical properties of the molten nitrate salts are disclosed, including material structure, melting point,latent heat,specific heat capacity and thermal conductivity.Thisexperimental study will provide valuable data for the development of theoretical models.

2. Experiment procedures

1. Materials selection

Inorganic salts are used in this study for high temperature thermal energy storage applications. These salts are among the most favourable candidates due to their virtues of thermophysical properties and stability. Rather than using eutectic salt mixtures, single salt is chosen with the aim of simplifying the system and eliminating interference factors that may affect the observation of the underlying physics. The salts used in the experiment are LiNO₃, NaNO₃ and KNO₃ with the purity of 99.0%, purchased from Sigma-Aldrich.

SiO₂ is chosen as the nanoparticles to enhance the thermal physical properties of the salts. SiO₂ nanoparticles with different average particle sizes (15 - 20 nm, 20-30 nm, 60 - 70 nm, 1 - 5 µm) are used to investigate the mechanisms of the enhancement. Those particles are purchased from US Research Nanomaterials with the purity of 99.5%. The density of pure materials is shown in Table 1, and the density of composite materials is calculated based on mixing rules.

2. Sample preparation 

Fig. 1 Ultrasonic mixing method for making salt-based nanomaterials

Fig. 1 illustrates the process of sample preparation for the salts with nanoparticles well dispersed. Salt and nanoparticles are weighed first with an analytical balance with ±0.1 µg precision (Mettler Toledo, type XP6U). The nanoparticles are then dispersed in distilled water (typically 20 ml) in a beaker and mixed by a high power ultrasonicator (Fisher scientific, CL 334) for 5 minutes. The saltsare then added to the aqueous based suspensions, and the resulting mixture is subjected to sonication for a further 5 minutes. The beaker is subsequently placed on a hot plate set at 130 ^oC to evaporate the water. This process gives a well dispersed salt-particle mixture. This method is similar to that used by Shin and Banerjee.¹⁵ However, instead of employing an ultrasonic bath and mixing the sample for 2 and 3 hours, a 500 W ultrasonicatoris used in this work and the mixing time is 5 minutes. This is because, after subjecting the aqueous suspension to the sonication for 5 minutes, the size of nanoparticles shows little change with longer sonicationas indicated by the results of the Zetasizer. In this work, salts with different mass fractions of nanoparticles (0.5%, 1.0%, 2.0%, and 4.0%) are prepared. Pure salts without nanoparticles are also prepared using the method to serve as the benchmark sample for results comparison.In addition, each test is made three times to ensure repeatability and minimize experimental errors. After each measurement, we cleaned the crucible with an ultra-sonic bath to ensure there is no contamination left in the crucible after each test. The procedure is as follows: we immerse crucible in a beaker of distilled water. We employed an ultra-sonic bath to clean the crucible in the beaker for 30 min and replace the water inside the beaker. After repeating the procedure twice, the crucible was placed in a drying oven at 120 °C.

3. Characterization

After the samples are prepared, their thermo-physical properties are tested, including decomposing temperature, crystal structure, melting point, latent heat capacity, specific heat capacity and thermal conductivity. The testing instruments are shown in Table 2.

Differential Scanning Calorimetry (DSC) is used to measure the melting point, latent heat capacity and the specific heat capacity. When the sample is undergoing a phase transition, more heat will flow to the sample than that to the reference to maintain both at the same temperature. By detecting this heat flow difference, the device will be able to work out the amount of heat absorbed or released during the phase change and hence the latent heat of the phase change by integrating the heat flow difference with respect to time. The onset of the heat flow variation is defined as the melting point.The specific heat capacity is measured withthe ASTM E1269-05 standard.A synthetic sapphire disk (α-aluminium oxide; alumina) was used as a heat flow calibration standard. The specific heat capacity of a sample is calculated by the following formula:

                                      (1)

whereC_p(sample) and C_p(standard) are specific heat capacities of the sample and sapphire standard, respectively; HF_(sample) and HF_(standard) are corrected heat flows of the sample and standard sapphire, respectively; m_(sample) and m_(standard) are mass of the sample and standard sapphire, respectively.

Laser flash analysis (LFA) is employed to measure the thermal diffusivity and thermal conductivity of the samples. A laser pulse is used to momentarily heat the bottom side of a sample with 12.5-14 mm diameter and 0.5-4 mm thickness. The temperature of the sample on the top side is measured with a radiation thermometer as a function of time.The thermal diffusivity can then be calculated by:

                                                        (2)

Where l is the thickness of the sample; t_1/2 is the time difference between the initiation of the pulse and the top side temperature when it reaches one-half of the maximum value. The thermal conductivity then can be related to the thermal diffusivity by:

                                                                 (3)

Where λ is the thermal conductivity of the sample; C_p is the specific heat capacity of the sample;  ρ is the density of the sample.

4. Uncertainty 

The uncertainty of results comes from equipmentuncertainty, propagation of uncertainty and artificial uncertainty. In our test, each property of the materials is measured three times to minimize the artificial uncertainty.Suppose independent parameters X₁, X₂, …, X_n are measured with uncertaintiesdX₁, dX₂, …, dX_n, and the measured parameters are used to calculate the function R(X₁, X₂, …, X_n). Then, the uncertainty of R is calculated with the following equation:   

                                                                         (4)

3. Results and discussion

   1. Decomposing temperature

Fig. 2 Thermogravimetric analysis of three nitrate salts

All three nitrate salts decompose at a high temperature. As is shown in Fig. 2, the salts are heated to 800 ^oC in the atmosphere of helium. Significant mass loss occurs in LiNO₃ at 557 ^oC, which indicates the beginning of the decomposition. Considerable mass loss of NaNO₃ and KNO₃ is observed at 607 and 649 ^oC, respectively. In order to avoid the decomposition and insure the repeatability of our experiments, the temperatures in the corresponding tests in this work are set much lower than the decomposition temperature.

1. 2. Materials structure

The structural analysis is performed on the pure salts and their nanoparticle mixtures using an X-Ray Diffractometer. Before the analysis, the pure salts and nanoparticle mixtures are cycled 50 times (200 - 350 ^oC for NaNO₃, 200 – 310 ^oC for LiNO₃ and 270 – 380 ^oC for KNO₃) at a heating/cooling rate of 10 ◦C/min for stabilizing the physical structure. Fig. 3 shows the XRD graphs of the pure salts and their mixtures with 0.5%, 15-20 nm SiO₂. One can see that the addition of the nanoparticles does not change crystallinity of the nitrate salts and there is no new phase created. The SiO₂ particles used are amorphous and hence are reflected in the XRD graphs.

Fig. 3 XRD analysis of (a) NaNO₃,(b) LiNO₃ and (c) KNO₃with 0.5%, 15-20nm

SiO₂ nanoparticles

1. 3. Melting point and latent heat

The melting point and latent heat of the pure salts (NaNO₃, KNO₃ and LiNO₃) are shown in Fig. 4.

One can see that the pure NaNO₃ has a melting point of 305.9 ^oC and latent heat of ∼179 kJ/kg.The pure LiNO₃ salt has a melting point of 255.2 ^oC and latent heat of 370.6 kJ/kg. The pure KNO₃ has a melting point of 333.9 ^oC and latent heat of  99.2 kJ/kg.

Fig. 4 Melting point and latent heat of pure NaNO₃, LiNO₃and KNO₃

Fig. 5 shows how the addition of SiO₂ nanoparticles affects the melting point of the pure salts (NaNO₃, KNO₃ and LiNO₃). As shown in Fig. 5(a), the addition of smaller nanoparticles (15-20 nm and 20-30 nm) appears to give a decrease in the melting point of NaNO₃, which leads to up to ∼2.5 ^oC in the melting point with 4% SiO₂nanoparticles. However, particles with larger diameters (60-70 nm and 1-5 mm) do not cause a significant impact on the melting point.As known, the melting point of a solution falls along with the mass fraction of the solute. The quantitative dependency is related to the variation of the system free energy after adding the solute. The nanoparticles are small, which is analogous to the solute in the molten salt. The smaller nanoparticles have considerable specific surface area and act as nucleation sites, which decreases the free energy required for phase change and hence decreases the melting point. The measured melting point for LiNO₃ with SiO₂ nanoparticlesis shown in Fig.5(b). The addition of all sizes of SiO₂nanoparticles gives a change of the melting point within ∼1 ◦C and mass fraction of SiO₂ nanoparticles does not seem to have a much effect. The impact of the introduction of SiO₂ nanoparticles on the melting point of KNO₃ isshown in Fig. 5(c). Overall, the addition of SiO₂ nanoparticles to the salt decreases the melting point, and the extent of decrease appears to increase with nanoparticle concentration. This is similar to the results shown in Fig. 5(a) for NaNO₃ salt. However, the decrease in the melting point is small, within ∼1 ^oC, and hence is insignificant.

Fig. 5 Melting point of (a) NaNO₃, (b) LiNO₃and (c) KNO₃with SiO₂ nanoparticles.

The latent heat of NaNO₃ with SiO₂ nanoparticles is shown in Fig. 6. By adding 15-20 and 20-30 nm nanoparticles, the latent heat shows a decreasing trend with the increasing mass fraction of SiO₂ nanoparticles, as shown in Fig. 6(a) and (b), respectively. The NaNO₃ with SiO₂ nanoparticles exhibit a lower latent heat than the pure salt. As the mass fraction of SiO₂ nanoparticles is 4%, the latent heat of NaNO₃ with15-20and 20-30nm nanoparticles is 167.2 and 162.1kJ/kg, respectively. For the 60-70 nm SiO₂ nanoparticles, as shown in Fig. 6(c), little changes are seen on the latent heat due to the introduction of 0.5%, 1% and 2% SiO₂ nanoparticles. However, the latent heat decreases to 166.7 kJ/kg by introducing 4% SiO₂ nanoparticles. The addition of 1-5µm SiO₂ particles does not show any significant impact on the latent heat, as shown in Fig. 6(d).

Fig. 6 Latent heat capacity of NaNO₃with various SiO₂ particle sizes: (a)15-20nm; (b) 20-30nm; (c) 60-70nm; (d)1-5 µm

The latent heat of LiNO₃ with SiO₂ nanoparticles is shown in Fig. 7. The addition of 15-20 nm SiO₂ nanoparticles decreases the latent heat of LiNO₃, and the extent of the decrease depends on the mass fraction of SiO₂ nanoparticles, as shown in Fig. 7(a). With 0.5% nanoparticles, the latent heat decreases to 368.8 kJ/kg, andit decreases to 351.6 kJ/kg with 4% nanoparticles. For the 20-30nm SiO₂ nanoparticles, as shown in Fig. 7(b), the addition of 0.5% or 1% nanoparticles increases the latent heat, whereas further increasing the nanoparticles to 2% and 4% shows a decrease in the latent heat, compared with the pure salt. The largest latent heat of ~377 kJ/kg is obtained as the mass fraction of SiO₂ nanoparticles is 1%. Fig. 7(c) shows similar observations for the 60-70 nm SiO₂ nanoparticles. With the mass fraction of SiO₂ nanoparticles increasing from 0.5% to 1%, the latent heat increases gradually, while it increases from 1% to 4%, the latent heat decreases significantly. The largest latent heat of ~378 kJ/kgis achieved as the mass fraction of SiO₂ nanoparticles is 1%. As shown in Fig. 7(d), for 1-5mm SiO₂ nanoparticles, the addition of 0.5%, 1% and 2% SiO₂ nanoparticles gives an increase in the latent heat, whereas the addition of 4% SiO₂ nanoparticles leads to a lower latent heat than the pure salt. The suggested mass fraction of SiO₂ nanoparticles is 0.5% and corresponding latent heat is ~379 kJ/kg.

Fig. 7 Latent heat capacity of LiNO₃with various SiO₂ particle sizes: (a)15-20nm; (b) 20-30nm; (c) 60-70nm; (d)1-5 µm

The latent heat of KNO₃ with SiO₂ nanoparticles is shown in Fig. 8. The addition of 15-20nm SiO₂ nanoparticles leads to changes to the latent heat and the extent of changes depends on the mass fraction of SiO₂ nanoparticles, as shown in Fig. 8(a). The largest latent heat of ~99.9 kJ/kg is achieved as the mass fraction of SiO₂ nanoparticles is at 0.5%. Introducing 20-30 nm SiO₂ nanoparticles into KNO₃ gradually decreases the latent heat as the mass fraction of SiO₂ nanoparticles increases from 0.5% to 4%, as shown in Fig. 8(b). The largest latent heat of ~100.3 kJ/kg is obtained with 0.5% SiO₂ nanoparticles. For the 60-70 nm SiO₂ nanoparticles, as shown in Fig. 8(c), the latent heat increases gradually to ~100.8 kJ/kgas the mass fraction of SiO₂ nanoparticles increases to 1% and then decreases generally with further increasing the mass fraction of SiO₂ nanoparticles. Fig. 8(d) shows that the addition of 1-5 mm SiO₂nanoparticles has little effect on the latent heat as the mass fraction of SiO₂ nanoparticles is 1% or 2%. However,with 0.5% and 4% SiO₂ nanoparticles, a lower latent heat is observed than the pure salt.

Fig. 8 Latent heat capacity of KNO₃with various SiO₂ particle sizes: (a)15-20nm; (b) 20-30nm; (c) 60-70nm; (d)1-5 µm

1. 4. Specific heat capacity

Specific heat capacity is an important parameter which determines the energy storage densityof sensible heat. Fig. 9 shows the specific heat capacity of NaNO₃ with various SiO₂ particles. For the 15-20nm SiO₂, enhancement of specific heat capacity occurs at all mass fraction of SiO₂ nanoparticles, as shown in Fig. 9(a). The specific heat capacity increases with increasing the mass fraction of SiO₂ nanoparticles and peaks at ∼1% SiO₂, beyond which a decrease is seen with a further increase in mass fraction of SiO₂ nanoparticles. Temperature has been found to give a relatively weak effect on the enhancement ofspecific heat capacity. Fig. 9(b) demonstrates the specific heat capacities with 20-30 nm SiO₂ nanoparticles. With 0.5% SiO₂ in the salt, the specific heat capacity reaches 1.73 kJ/kg/K, giving an enhancement of 6% with respect to the pure salt. A decreasing tendency occurs with further increase in the mass fraction of SiO₂ nanoparticles. The peak specific heat capacity occurs at 0.5% SiO₂. The temperature appears to have little effect on the enhancement.The specific heat capacity of NaNO₃ with 60-70 nm SiO₂ nanoparticles is shown in Fig. 9(c). A significant enhancement is observed under all mass fraction of SiO₂ nanoparticles. The peak enhancement of 27.6% occurs at 2% SiO₂.In general, a higher temeprature leads to a larger specific heat capacity. The results for the 1-5 µm SiO₂ particles are shown in Fig. 9(d). One can see that the specific heat capacity is only enhanced at 0.5% and 1% SiO₂. A further increase in the mass fraction of SiO₂ particlesleads to a decreasing of specific heat capacity, even lower than that of the pure salt. The highest enhancement of 2.4% occurs at 1% SiO₂.

Fig. 9 Specific heat capacity of NaNO₃with various SiO₂ particle sizes: (a)15-20nm; (b) 20-30nm; (c) 60-70nm; (d)1-5 µm

Fig. 10 shows the specific heat capacity of LiNO₃ with various SiO₂ particles. In terms of 15-20 nm SiO₂ nanoparticles, the addition of 0.5% SiO₂ gives the highest specific heat capacity of~2.10 kJ/kg/K, as shown in Fig. 10(a). This is equivalent to an enhancement of ~8%. With an increase in the mass fraction of SiO₂ nanoparticles, the specific heat capacity decreases. What’s more, a higher temperature results in a larger specific heat capacity. The specific heat capacity of LiNO₃ with 20-30 nm SiO₂ nanoparticlesis shown in Fig. 10(b). A significant enhancement can be observed for 0.5%, 1% and 2% SiO₂ nanoparticles, with the maximum of ~10.3%. The highest specific heat capacity is ∼2.14 kJ/kg/K at 0.5% SiO₂ nanoparticles. For the 60-70 nm SiO₂ nanoparticles, an addition of 0.5% SiO₂ gives the maximum enhancement of specific heat capacity by 12.3%, as shown in Fig. 10(c).A further increase in the mass fraction of SiO₂ nanoparticles leads to the decrease of the specific heat capacity. At 4% SiO₂,it is even slightly lower thanthat of the pure salt. Fig. 10(d) presents that the addition of 1-5 µm SiO₂particles only gives an enhancement when the mass fraction of SiO₂ particles is at 0.5%, whereas the additon of 1-4% SiO₂ particles shows no or even negative enhancement.

Fig. 10 Specific heat capacity of LiNO₃with various SiO₂ particle sizes: (a)15-20nm; (b) 20-30nm; (c) 60-70nm; (d)1-5 µm

Fig. 11 shows the specific heat capacity of KNO₃ with various SiO₂ particles. The specific heat capacity of KNO₃ pure salt is ~1.31 kJ/kg/K. With the addition of 15-20 nm SiO₂ nanoparticles, the maximum specific heat capacity of ~1.65 kJ/kg/K is achieved as the mass fraction of SiO₂ nanoparticlesis at 1%, and the corresponding enhancement is ~26%, as shown in Fig. 11(a). The increase of salt temperature leads to a slight increase of the specific heat capacity for 1% and 2% SiO₂ nanoparticles and a decrease for 0.5% and 4% SiO₂ nanoparticles. The specific heat capacity of KNO₃ with 20-30 nm SiO₂ nanoparticlesis presented in Fig. 11(b). The maximum specific heat capacity is obtained at 2% SiO₂ nanoparticles with the enhancement of ~6.8%. A higher temperature results in a larger specific heat capacity. For the addition of 60-70 nm SiO₂ nanoparticles, as shown in Fig. 11(c), the maximum specific heat capacity is 1.54 kJ/kg/Kwith mass fraction of SiO₂ nanoparticles at 1%; a further increase of the mass fraction leads to a decrease in the specific heat capacity. What’s more, the specifc heat capacity slightly increases with the increase of salt temperature. In terms of KNO₃ containing 1-5 µm SiO₂ particles, the specific heat capacity is only slightly enhanced as the mass fraction of  SiO₂ particles is above 2%, as shown in Fig. 11(d). What’s more, a slight increasing trend of the specific heat capacity is seen with the increase of salt temperature.

Fig. 11 Specific heat capacity of KNO₃with various SiO₂ particle sizes: (a)15-20nm; (b) 20-30nm; (c) 60-70nm; (d)1-5 µm

1. 5. Thermal conductivity

Thermal conductivity is a key parameter to evaluate heat transfer performance. Fig. 12 shows the thermal conductivity of NaNO₃, LiNO₃, and KNO₃ after melting, with the addition of 15-20 nm SiO₂ nanoparticles. As shown in Fig. 12(a), the pure NaNO₃ has a thermal conductivity of ∼0.54 W/m/Kin thetemperature range of 320-370 ^oC. With the introduction of 0.5% SiO₂ nanoparticles, the average thermal conductivity decreases to 0.49 W/m/K. As the mass fraction of SiO₂ nanoparticles increases from 0.5% to 4%, the thermal conductivity decreases first and increases later. In general, the addition of SiO₂ nanoparticles is unfavourable to the thermal conductivity of NaNO₃ in the studied range. A higher temperature usually leads to a slightlylower thermal conductivity. Fig. 12(b) shows the thermal conductivity of LiNO₃ with 15-20 nm SiO₂ nanoparticles in the temperature range of 260-300 ^oC. One can see that the introduction of SiO₂ nanoparticles all decreases the thermal conductivity compared with the pure LiNO₃. What’s more, increasing the temperature decreases the thermal conductivity. The pure KNO₃ has an average thermal conductivity of ∼0.46 W/m/K in the temperature range of 340-380 ^oC, as shown in Fig. 12(c). The addition of SiO₂nanoparticles from 0.5% to 4% gives a decrease in the thermal conductivity.

Fig. 12 Thermal conductivity of (a) NaNO₃(b) LiNO₃and (c) KNO₃with 15-20nm SiO₂ nanoparticles

SiO₂ is commonly added in nanofluids as the conductivity enhancer attributing to its high thermal conductivity of 1.3-1.5 W/m/K. With the addition of SiO₂ in nitratesalts, it is expected to exhibit an enhanced thermal conductivity since SiO₂is more conductive thanall nitratesalts. However, the experimentresults are against our expectations. A model is hence proposed to help explain this phenomenon, as shown in Fig. 13. Though the thermal conductivity of SiO₂ is higher than that of the molten salt, an interfacial layer, existed between the solid SiO₂ and the liquid molten salt, provides a much higher thermal resistance, which significantly hinders the heat transfer. The thermal conductivity is then determined by the SiO₂ nanoparticles, molten salt and the interfacial layer. This model will be validated from molecular scale in the future work.

Fig. 13 A nanoparticle with an interfacial layer in a molten salt fluid

4. Conclusions

Thermal energy storage (TES) is a powerful technology to shift peak loads of power gridsand smooth intermittency of renewable energy. Molten nitrate salts are popular TES materialsfor solar power plants. However, pure nitratesalts usually show unfavourable properties, which affects the energy storage performance. In this paper, molten nitrate salts based nanofluids are prepared with a novel method. Three popular single salts (NaNO₃, LiNO₃, and KNO₃) are selected as the base salts and SiO₂ nanoparticles as the additives. Variousparameters, such as SiO₂ sizes (15 nm - 5 µm), mass fractions (0.5-4%) and working temperatures (200-380 ^oC), are considered. The thermo-physical properties are experimentally studied, including material structure,melting point, latent heat capacity, specific heat capacity and thermal conductivity.The main conclusions are as follows:

• The addition of SiO₂ nanoparticles does not change crystallinity of the three nitrate salts and there is no new phase created, which shows agood chemical compatibility between SiO₂ nanoparticles and nitrate salts.
• The effect of adding SiO₂nanoparticles on melting points of LiNO₃ and KNO₃ is insignificant. However, for NaNO₃,its melting points are decreased by up to 2.5 ^oC, and a higher mass fraction of SiO₂nanoparticles leads to a lower melting point. The decrease of melting points is probably because the small nanoparticles have a considerable specific surface area and act as nucleation sites, which reduces the free energy required for phase change and hence decreases the melting point.
• The latent heat of NaNO₃ is decreased or keeps stable with the addition of SiO₂ nanoparticles.However,for LiNO₃ and KNO₃, their latent heat is increased to~377kJ/kgwith20-30 nm 1% SiO₂ and~100.8  kJ/kg with60-70 nm 1% SiO₂, respectively.
• The introduction of SiO₂ nanoparticles enhances the specific heat capacity of the three nitrate salts. ForNaNO₃ and LiNO₃, the highest enhancementsof 27.6% and 12.3%are observed with the addition of 60-70 nm 2% and 0.5% SiO₂ nanoparticles, respectively. For KNO₃, the highest enhancement of 26% is achievedwith 15-20 nm 1% SiO₂ nanoparticles. In general, a higher temperature leads to a higher specific heat capacity.
• The three nitrate salts exhibit adepressed thermal conductivity with the addition of SiO₂ nanoparticles. This is out of expectation since SiO₂ has a much higher thermal conductivity than the molten salt. It is assumed that there is an interfacial layer between the solid nanoparticle and the liquid molten salt,which provides a much higher thermal resistance and hence decreases the thermal conductivity.This assumption will be validated in the future work from molecular scales.

Finally, NaNO₃ with 60-70 nm 0.5% SiO₂ nanoparticles is suggested, which enhances the specific heat capacity by ~27% without affecting the latent heat and thermal conductivity much.For LiNO₃, 20-30 nm 1% SiO₂nanoparticlesare proposed, while the addition of 60-70 nm 1% SiO₂ nanoparticles is recommended for KNO₃ which increases the specific heat capacity by 17.6% and latent heat by 1.6%.

Acknowledgements

This paper was supported bykeytechnologies for high temperature sensible heat storage using molten salts program from State Grid Corporation of China and GlobalEnergy Interconnection Research Institute Europe GmbH No. SGRI-DL-71-16-018.The authors would like to acknowledge networking support by the COST Action CA15119.

References

1. Renewables 2018 Global Status Report, REN21, 2018
2. M. M. Kenisarin, Renew. Sust. Energ. Rev., 2010, 14(3), 955-970.
3. G. Wei, G. Wang and C. Xu, Renew. Sust. Energ. Rev., 2018, 81, 1771–1786.
4. R. I. Dunn, P. J. Hearps and M. N. Wright, Proc IEEE, 2012, 100(2), 504-515.
5. The parabolic trough power plants Andasol 1 to 3, Solar Millennium, (2008) 1-26
6. G. Xu, G. Leng and C. Yang, Sol. Energy, 2017, 146, 494-502.
7. S. Pincemin, X. Py and R. Olives, J. Sol. Energ., 2008, 130(1), 011005.
8. J. Li, W. Lu and Z. Luo, Sol. Energ. Mat. Sol. C., 2017, 171, 106-117.
9. J. Li, W. Lu, and Z. Luo, Sol. Energ. Mat. Sol. C., 2017, 159, 440–446.
10. P. Zhang, J. Cheng and Y. Jin, Sol. Energ. Mat. Sol. C., 2018, 176, 36-41.
11. L. L. Zou, X. Chen, Y. T. Wu, X. Wang and C. F. Ma, Sol. Energ. Mat. Sol. C., 2019, 190, 12-19.
12. X.  Li, Y.Wang, S. Wu and L. D. Xie, Energy, 2018, 160, 1021-1029.
13. Á. G. Fernández and J. C. Gomez-Vidal, Renewable Energy, 2017, 101, 120-125.
14. H.  Kim, H. S. Kim, S. N. Lee and J. K. Kim, Int. J. Heat Mass Tran., 2018, 127,465–472.
15. D. Shin and D. Banerjee, Int. J. Heat Mass Tran., 2011, 54, 1064-1070.
16. M. S. Belén, N. M. Javier and I. I. Torres, Renew. Sust. Energy Rev., 2018, 82, 3924–3945.
17. X. Luo, X. Du and A. Awad, Int. J. Heat Mass Tran., 2017, 104, 658-664.
18. P. D. M. Jr, T. E. Alam and R. Kamal, Appl. Energy, 2016, 165, 225-233.
19. A. Awad, A. Burns and M. Waleed, Sol. Energy, 2018, 172,191-197.
20. Y. Hu, Y. He and Z. Zhang, Energ. Convers. Manage., 2017, 142, 366-373.
21. J. Seo and D. Shin, Appl. Therm.  Eng., 2016, 102, 144-148.
22. W. Song, Y. Lu and Y. Wu, Sol.  Energ. Mat.  Sol. C., 2018, 179, 66-71.
23. X. Chen, Y. T. Wu, L. D. Zhang, X. Wang and C. F. Ma, Sol.  Energ. Mat.  Sol. C., 2019, 191, 209-217.
24. Z. Jiang, A. Palacios, X. Z. Lei, M. E. Navarro, G. Qiao, E. Mura, G. Z. Xu and Y. L. Ding, Appl. Energy, 2019, 235, 529–542
25. G. Qiao, A. Alexiadis and Y. Ding, Powder Technol., 2017, 314, 660–664.
26. G. Qiao, M. Lasfargues and A.Alexiadis, Appl. Therma. Eng., 2017, 111, 1517-1522.
27. K. Khanafer, F. Tavakkoli and K. Vafai, Int. Commun. Heat  Mass, 2015, 69, 51-58.
28. D. Shin, H. Tiznobaik and D. Banerjee, Appl. Phys. Lett., 2014, 104(12), 121914.
29. H. Tiznobaik and D. Shin, Int. J. Heat  Mass Tran., 2013,  57(2), 542-548.
30. M. Chieruzzi, G. F.Cerritelli  and A. Miliozzi, Sol. Energ. Mat. Sol. C., 2017, 167, 60-69.
31. N. Navarrete, R. Mondragon and D. Wen, Energy, 2019, 167, 912-920.
32. A. Awad, H. Navarro and Y. Ding , Renew. Energ., 2018, 120, 275-288.
33. Y. Grosu, N. Udayashankar and O. Bondarchuk, Sol. Energ. Mat.  Sol. C., 2018, 178, 91-97.