The effect of 3D carbon nanoadditives on lithium hydroxide monohydrate based composite materials for highly efficient low temperature thermochemical heat storage

Lithium hydroxide monohydrate based thermochemical heat storage materials were modified with in situ formed 3D-nickel-carbon nanotubes (Ni-CNTs). The nanoscale (5–15 nm) LiOH·H2O particles were well dispersed in the composite formed with Ni-CNTs. These composite materials exhibited improved heat storage capacity, thermal conductivity, and hydration rate owing to hydrogen bonding between H2O and hydrophilic groups on the surface of Ni-CNTs, as concluded from combined results of in situ DRIFT spectroscopy and heat storage performance test. The introduction of 3D-carbon nanomaterials leads to a considerable decrease in the activation energy for the thermochemical reaction process. This phenomenon is probably due to Ni-CNTs providing an efficient hydrophilic reaction interface and exhibiting a surface effect on the hydration reaction. Among the thermochemical materials, Ni-CNTs–LiOH·H2O-1 showed the lowest activation energy (23.3 kJ mol−1), the highest thermal conductivity (3.78 W m−1 K−1) and the highest heat storage density (3935 kJ kg−1), which is 5.9 times higher than that of pure lithium hydroxide after the same hydration time. The heat storage density and the thermal conductivity of Ni-CNTs–LiOH·H2O are much higher than 1D MWCNTs and 2D graphene oxide modified LiOH·H2O. The selection of 3D carbon nanoadditives that formed part of the chemical heat storage materials is a very efficient way to enhance comprehensive performance of heat storage activity components.


Introduction
As an important part of clean and efficient utilization of alternative energy, the development of thermal energy storage technology has become increasingly important in recent years owing to the incremental consumption of fossil energy and the emission of greenhouse gases. 1,2 These technologies are divided into three main types: sensible heat storage, 3,4 latent heat storage, 5,6 and thermochemical heat storage. 7,8 All of these technologies participate in solving the supply and demand mismatch of thermal energy, and improve energy efficiency. 9 Among these technologies, thermochemical heat storage, which uses reversible chemical reactions to store and release thermal energy, facilitates the efficient utilization of thermal energy due to its high heat storage density. 10 Based on heat storage working temperature, the thermochemical heat storage technology could be divided into two parts: high temperature heat storage and low temperature heat storage. 9,10 As core components of these technologies, numerous thermochemical materials (TCMs) have been selected. For instance, metal hydroxides, metal hydrides, and metal carbonates are typically used as TCMs for high-temperature thermochemical heat storage, whereas inorganic salt hydrates and salt ammoniate are considered as promising candidates for low-temperature thermochemical heat storage due to their different decomposition temperature. [11][12][13][14][15][16] For efficiently storing low-temperature thermal energy, the inorganic hydrate lithium hydroxide monohydrate (LiOH$H 2 O), which possesses high energy density (1440 kJ kg À1 ) and mild reaction process, was selected as a potential material. 17 However, similar to other inorganic hydrates 18,19 pure LiOH$H 2 O exhibit low hydration rate and thermal conductivity; 15,17 moreover, its heat storage density decreases aer hydration, thereby seriously limiting the application of the material. Hence, composite TCMs with high heat storage density and high thermal conductivity are of considerable synthetic value.
3D carbon nanomaterials (3D-carbon nanotube sponge and arrays 20,21 ), which exhibit large surface area, low bulk density, and chemical stability, [22][23][24] are widely used in various elds, such as electronics 25,26 and catalysis 24,27 as typical carbon nanomaterials. In addition, these materials offer excellent hydrophilic property aer the introduction of surface oxygen groups. However, under normal conditions, traditional one-dimensional or two-dimensional carbon nanomaterials are selected and used for latent heat thermal energy storage. 2,28,29 3D carbon nanomaterials are rarely used for thermal energy storage, 30 especially thermochemical heat storage. In our previous work, 1D carbon nanomaterials, 31 MWCNTs were used to modify LiOH$H 2 O; heat storage performance was markedly enhanced, but the thermal conductivity needed improvement. Hence, 2D-carbon nanomaterial 15,32 graphene oxide was selected and, which markedly improved the energy storage performance of LiOH$H 2 O and Mg(OH) 2 . Graphene oxide also positively affected thermal conductivity to a certain extent. Hydrophilic substances, such as lithium chloride (LiCl), 13X-zeolite, and NaY-zeolite, were also used to enhance the heat storage performance and obtain expected results. 33 Kato et al. investigated the reaction behavior of metal-salt modied Mg(OH) 2based materials; during the heat storage process, LiCl and LiBr acted as catalysts and decreased the activation energy effectively and substantially improved heat storage performance. 5,11,34 Moreover, the functional groups on the surface of the materials markedly affected the reaction behavior of Mg(OH) 2 . However, the relationship between heat storage density and particle size, especially when materials reached nanoscale, was not discussed. In this work, in situ DRIFT spectroscopy technology and chemical reaction kinetics test were used for an in-depth analysis of the heat storage mechanism, which has not been involved in previous research. Moreover, to simultaneously improve the performance of heat storage density, hydration rate, and thermal conductivity, we synthesized a novel TCM composite of in situ formed 3D-nickelcarbon nanotubes (Ni-CNTs) and LiOH$H 2 O. Four kinds of TCMs were prepared, and the effect of 3D-carbon nanomaterials was extensively investigated. The highest heat storage density of Ni-CNTs-LiOH$H 2 O could reach 3935 kJ kg À1 , which is 2.2 times higher than that of 1D MWCNTs modied LiOH$H 2 O (1804 kJ kg À1 ) and the thermal conductivity (3.78 W m À1 K À1 ) is also much higher than MWCNTs modied LiOH$H 2 O (1.75 W m À1 K À1 ) 31 The heat storage density and the thermal conductivity of Ni-CNTs modied LiOH$H 2 O are also higher than 2D graphene oxide modied LiOH$H 2 O (1980 kJ kg À1 ; 1.70 W m À1 K À1 ), respectively. 15 It indicated that the selection of 3D nano carbon materials as composed part of the chemical heat storage materials is a very efficient way to enhance comprehensive performance of heat storage activity component.

Raw materials and synthesis method of LiOH$H 2 Obased TCMs
The 3D-carbon nanotubes were synthesized by catalytic chemical vapor deposition method with C 2 H 4 as carbon source and nickel foam as catalyst. 35 First, nickel foam was placed in a tubular furnace, which was heated to 700 C, and lled with Ar/H 2 (300 mL/100 mL) mixed reducing gas for 2 h. Aer reduction, Ar/H 2 mixed gas was replaced with Ar/C 2 H 4 (400 mL/ 100 mL), the temperature was increased to 550 C, and the system was allowed to react for 20 min. Aer the temperature decreased to 25 C, the 3D Ni skeleton CNTs (Ni-CNTs) was obtained. Then, aer being oxidized by 5% O 2 for 2 h at 250 C, the as-prepared 3D Ni-CNTs were composited with LiOH$H 2 O by impregnation method. First, 0.5 g LiOH$H 2 O was dissolved in 1 mL deionized water under vigorous stirring. Subsequently, 3D nanocarbon were added in the LiOH aqueous solution at room temperature, and the mixture was stirred continuously for 4 h. Aerward, the products were withdrawn and vacuum freeze-dried.

The characterization and heat storage performance test method of LiOH$H 2 O-based TCMs
Surface topography was measured by eld-emission scanning electron microscopy (SEM, S-4800, Hitachi Limited). Transmission electron micrographs (TEM) were obtained with FEI Tecnai G212 operated at 100 kV and a JEOL JEM-2100F operated at 200 kV. X-ray diffraction (XRD) analysis was performed on a D8-advance X-ray diffractometer (Bruker, Germany) with Cu target (40 kV, 40 mA). The scan step size was 0.0167 , and counting time was 10.160 s. Nitrogen adsorption-desorption was measured at the boiling point of nitrogen (77 K) by a Quantachrome QDS-30 analyzer. BET (Brunauer-Emmett-Teller) surface area and pore structure were measured by nitrogen physisorption under normal relative pressure of 0.1-1.0. The thermal conductivity of the sample was measured by a DRL-II thermal conductivity tester (Xiangtan Xiangyi Instrument Co., Ltd.). LiOH$H 2 O, Ni-CNTs-LiOH$H 2 O-1, Ni-CNTs-LiOH$H 2 O-2, Ni-CNTs-LiOH$H 2 O-3 were used as raw substance then, LiOH Ni-CNTs-LiOH-1, Ni-CNTs-LiOH-2, and Ni-CNTs-LiOH-3 were synthesized by decomposing LiOH$H 2 O, Ni-CNTs-LiOH$H 2 O-1, Ni-CNTs-LiOH$H 2 O-2, and Ni-CNTs-LiOH$H 2 O-3 in a horizontal tubular quartz furnace with Ar gas at 150 C for 3 h. Dehydrated products were cooled to 30 C in an Ar atmosphere, and water vapor at a partial pressure of 2.97 kPa and carried by N 2 ow was introduced to the tube for 60 min for hydration operation at 30 C. Aer hydration, the endothermic heat and temperature of the samples were measured through Thermogravimetry and Differential Scanning Calorimetry (TG-DSC) (STA-200, Nanjingdazhan Co., Ltd.), which was also used for measuring weight change during dehydration step. Each TG-DSC measurement was repeated three times in order to ensure correctness. The activation energy of dehydration performance of all samples was calculated using Ozawa method, 36 which is applicable for calculating the activation energy of thermal decomposition and dehydration reaction. By using the Ozawa method, the following equation can be obtained based on the reaction rate expression and the Arrhenius's equation: In this equation, E is the activation energy (kJ mol À1 ), b is the heating rate (K s À1 ), T is temperature (K), R is the ideal gas constant (J (mol K) À1 ), A is a pre-exponential factor, a is the dehydration conversion, and f(a) is a function of dehydration conversion. During the calculation of activation energy, the dehydration conversion was maintained at 70%. The heating rates were 3 K min À1 , 7 K min À1 and 10 K min À1 , and the activation energy was obtained from the slope (ÀE/R) of this equation.  Fig. 2a, before the addition of 3Dcarbon nanoadditives, the bulk LiOH$H 2 O was aggregated with a large diameter (300 nm to 2 mm). Aer the modication of Ni-CNTs, no obvious change was observed in the regular structure for Ni-CNTs-LiOH$H 2 O-1 (Fig. 2b) and Ni-CNTs-LiOH$H 2 O-2 (Fig. 2c). This nding indicates that the intervention of LiOH$H 2 O did not lead to structural deterioration at a mass ratio < 23% and that LiOH$H 2 O was highly dispersed in the composite materials. This result agrees well with the XRD result. When the content of LiOH$H 2 O reached 39% (Ni-CNTs-LiOH$H 2 O-3), several LiOH$H 2 O particles could be seen on the surface of Ni-CNTs (Fig. 2d). Fig. 3 shows the TEM images of Ni-CNTs-LiOH$H 2 O with different contents of LiOH$H 2 O. In Fig. 3, LiOH$H 2 O nanoparticles with a diameter around 5 nm were successfully dispersed on Ni-CNTs ( Fig. 3b) with clear particle structure. LiOH$H 2 O particle size showed a growth trend with increasing LiOH$H 2 O content in Ni-CNTs-LiOH$H 2 O composite materials ( Fig. 3d and f). When LiOH$H 2 O content reached 39% (Fig. 3f), the LiOH$H 2 O nanoparticle size could increase to 15 nm, which is a bit bigger than the other two composite materials. Pure LiOH$H 2 O (Fig. 2a) existed in the form of stacked akes and showed the largest particle size (300 nm to 2 mm).   According to SEM and TEM characterization results, high specic surface area is an important cause of the nanoscale dispersion of LiOH$H 2 O particles.

Heat storage performance test on LiOH$H 2 O-based TCMs
The heat storage performance test of pure LiOH$H 2 O, Ni-CNTs-LiOH$H 2 O-1, Ni-CNTs-LiOH$H 2 O-2 and Ni-CNTs-LiOH$H 2 O-3 were carried out and shown in Fig. 4. The Y axis is the amount of heating per unit time and mass (W g À1 ), and the X axis is temperature. The area of the curve is proportional to the change of enthalpy, as for heat storage materials it stands for the heat storage density.  Fig. 4c) and Ni-CNTs-LiOH$H 2 O-3 (Fig. 4d), respectively. Therefore, compared with pure LiOH and with the same duration of the hydration reaction LiOH, and H 2 O molecule can be fully converted to LiOH$H 2 O owing to the introduction of Ni-CNTs, the hydration reaction rate of Ni-CNTs-LiOH$H 2 O was substantially improved. On the one hand, owing to the formation of hydrophilic functional groups on the surface of Ni-CNTs during preparation, H 2 O adsorption became easier and provided a completely different reaction interface between LiOH and water molecules. On the other hand, Ni-CNTs-LiOH$H 2 O showed ultrahigh heat storage density exceeding that of pure LiOH$H 2 O owing to the existence of hydrophilic functional groups 41 and increased specic surface area, which markedly enhanced the dispersion of LiOH$H 2 O nanoparticles and the contact surface area with water molecules. The low specic surface area of pure LiOH$H 2 O may exert a negative effect on heat storage density. When the particle size reached nanoscale, the amount of surface atoms evidently increases; moreover, the crystalline eld and binding energy of internal atoms were notably different from those of surface atoms, which possessed numerous dangling bonds owing to the lack of adjacent atoms. Hence, the unsaturated bonds in atoms show that nanoparticles present enhanced thermodynamic property. 42,43 Meanwhile, due to increased number of surface atoms and the existence of hydrophilic functional groups, more H 2 O and LiOH can react, and therefore, heat storage performance can be improved. Furthermore, according to SEM and TEM characterization results, the heat storage density of Ni-CNTs-LiOH$H 2 O was higher than that of LiOH$H 2 O possible because of the smaller particle size of LiOH$H 2 O (5-15 nm) in Ni-CNTs-LiOH$H 2 O than that in pure LiOH$H 2 O (300 nm to 2 mm). Thus, small-sized nanoparticles can notably contribute to the enhancement of heat storage density for composite materials; as the particle size expands, the heat storage density of LiOH$H 2 O decreases. Furthermore, aer the addition of Ni-CNTs to LiOH$H 2 O, the thermal conductivity of composite materials evidently increased and exceeded that of pure LiOH$H 2 O (Fig. 5) owing to the high thermal conductivity of Ni-CNTs. The highest thermal conductivity of Ni-CNTs-LiOH$H 2 O can reach 3.78 W m À1 K À1 , which is 2.2 times higher than that of pure LiOH$H 2 O.
In situ DRIFT test was conducted to investigate the effect of hydrophilic functional groups of 3D-carbon nanotubes on the hydration/dehydration reaction of LiOH$H 2 O. Raw materials LiOH$H 2 O and the Ni-CNTs-LiOH$H 2 O composite were exposed to a ow of N 2 (300 mL min À1 ) at 150 C to decompose for 24 h. Then, the as-prepared samples LiOH and Ni-CNTs-LiOH were placed in the in situ reactor of an FT-IR spectrometer. The reactor was vacuumed and purged using a He ow. The in situ DRIFT test started with the hydration reaction of LiOH and H 2 O and nished with dehydration reaction of LiOH$H 2 O.    44 In the spectra, the peak at around 3679 cm À1 and 1573 cm À1 could be attributed to the stretching vibrations (n OH ) and bending vibrations (b OH ) of the structure water in LiOH$H 2 O, respectively. 45 Besides, broad peaks in the 2842-3423 cm À1 range were centered at around 3235 cm À1 . This peak (3235 cm À1 ) and another peak at the lower band (1644 cm À1 ) were the n OH and b OH of OH, respectively, in the physical adsorbed H 2 O. 44,46 In Fig. 6a it could be also observed that when the hydration reaction was ready for start, the weak peaks, which stand for OH in structure H 2 O and physical adsorbed H 2 O molecules have been existed. This nding may be due to the reaction of LiOH and residual physical adsorbed H 2 O on the surface of LiOH. With prolonged hydration reaction, the peak intensities of structural OH (3679 cm À1 and 1573 cm À1 ) markedly increased at 15 min hydration. These peaks slowly increased during 15 min to 120 min hydration reaction, indicating the continuous reaction of LiOH and water steam and the decrease in hydration reaction rate. During this reaction, no obvious change could be observed for the peak intensities of OH (3235 cm À1 and 1644 cm À1 ) in physical adsorbed H 2 O because of the steady water steam ow in the in situ reactor. Aer the hydration reaction, the reactor was vacuumed and purged using a dry He ow, then LiOH$H 2 O was heated at a rate of 0.5 C s À1 under the control of temperature-programmed technology. Fig. 6b shows the in situ DRIFT spectroscopy of dehydration reaction of LiOH$H 2 O obtained from 1 h hydration of LiOH. When LiOH$H 2 O was heated to 60 C, the peak intensities of OH in structural H 2 O (3679 cm À1 and 1573 cm À1 ) started to decrease. When the temperature exceeded 70 C, the trend became more evident. Meanwhile, the peaks of OH (3235 cm À1 and 1644 cm À1 ) in physical adsorbed H 2 O appeared and increased from 60 C to 80 C, indicating the form of free water. The intensity of the peaks gradually decreased owing to easy  This journal is © The Royal Society of Chemistry 2018 desorption and blowing away of physical adsorbed water at elevated temperature. Fig. 6c shows the in situ DRIFT spectroscopy of hydration reaction of Ni-CNTs-LiOH and H 2 O. During hydration reaction, the peak intensities of OH in structure H 2 O (3679 cm À1 and 1573 cm À1 ) showed a marked increase with 5 min hydration because of the existing hydrophilic groups, such as C-OH and C]O, on the surface of 3D-carbon nanotubes. Aer 20 min, the peaks were virtually unchanged, which indicating that the hydration reaction rate of LiOH and water steam was enhanced, exceeding that of pure LiOH. The peak intensities of OH (3181 cm À1 and 1625 cm À1 ) in physical adsorbed H 2 O did not notably change. The band at around 3520 cm À1 can be assigned to stretching vibrations, whereas the band at around 1400 cm À1 was attributed to the bending vibrations of C-OH. 44 The band at 1720 cm À1 was assigned to C]O groups. When the water steam owed into the reactor, the wavenumber of C]O groups shied from 1720 cm À1 to 1710 cm À1 because of the formation of hydrogen bonding 2 between C]O groups and adsorbed H 2 O.
Owing to the effect of hydrogen bonding, the adsorption of H 2 O on the surface of Ni-CNTs-LiOH was considerably enhanced, facilitating an easier and more rapid reaction of H 2 O and LiOH than before. Fig. 6d presented the in situ DRIFT spectroscopy of dehydration reaction of Ni-CNTs-LiOH$H 2 O. When Ni-CNTs-LiOH$H 2 O was heated to 50 C, the peak intensities of OH in structural H 2 O (3679 cm À1 and 1573 cm À1 ) started to decrease; beyond 60 C, LiOH$H 2 O was continuously dehydrated until the structure of H 2 O was fully lost. Meanwhile, the peaks of OH (3181 cm À1 and 1625 cm À1 ) in physical adsorbed H 2 O appeared, increased from 50 C to 90 C, and eventually decreased; this trend was ascribed to the easy removal of H 2 O at elevated temperature. The formed water could also produce hydrogen bonding with C]O groups (1720 cm À1 ) on the surface of Ni-CNTs, thereby causing a shi of the C]O band to a lower position (1710 cm À1 ). During the dehydration reaction from 30 C to 120 C, the intensities of C]O and C-OH (3520 cm À1 and 1400 cm À1 ) peaks showed no obviously change, indicating that the physical-chemical property of Ni-CNTs is steady within the total dehydration temperature range. The dehydration temperature of Ni-CNTs-LiOH$H 2 O was approximately 10 C lower than that of pure LiOH$H 2 O; this nding is in good agreement with the results of the TG-DSC test.
As shown in Fig. 7, the activation energies of the dehydration reaction of (a) LiOH$H 2 36 The activation energies of composite TCMs were markedly lower than that of pure LiOH$H 2 O due to the surface effect of nano-LiOH$H 2 O composite TCMs. The specic surface area or surface-to-volume ratio, which changes with particle size, depends on activation energy. 47 By combining the SEM, TEM, and BET characterization results, it could be found that as the Ni-CNT content increased, the specic surface area of the composite TCMs also increased, whereas the particle size of LiOH$H 2 O decreased. The activation energy of LiOH$H 2 O dehydration reaction in composite TCMs showed similar trend to the particle size variation of LiOH$H 2 O. This trend can be attributed to diminished particle size, leading to increased surface-to-volume ratio and molar surface energy of nanoparticles, which mainly result in decreased activation energy. 48,49 In summary, the addition of 3D-carbon nanomaterial Ni-CNTs can not only enhance water absorption at the LiOH particle surface but also decrease activation energy.
Owing to the addition of Ni-CNTs, LiOH$H 2 O are dehydrated more easily, and the reaction mechanism of composited TCMs may deviate from that of pure LiOH$H 2 O because of the surface effect of 3D Ni-CNTs-LiOH$H 2 O nanocomposite materials during the thermochemical reaction process.

Conclusions
For investigating the effects of in situ formed 3D-carbon nanoadditives (Ni-CNTs) on the thermal performance of lowtemperature LiOH$H 2 O-based composites as thermochemical heat storage materials, four kinds of LiOH$H 2 O-based composite TCMs were successfully constructed and characterized. Owing to the addition of 3D-carbon nanoadditives, the nanoscale (5-15 nm) LiOH$H 2 O particles were well dispersed in the composite with Ni-CNTs. The heat storage capacity and thermal conductivity of the composite materials were markedly improved. Meanwhile, the hydration rate was enhanced due to the hydrogen bonding formed between H 2 O and hydrophilic groups on the surface of Ni-CNTs, as shown by the combined results of in situ DRIFT spectroscopy characterization and the heat storage performance test. The activation energy for the thermochemical reaction process notably decreased aer the addition of Ni-CNTs possibly because Ni-CNTs provide efficient hydrophilic reaction interface and exhibit surface effect in the hydration reaction. Among these TCMs, Ni-CNTs-LiOH$H 2 O-1 showed the lowest activation energy (23.3 kJ mol À1 ), highest thermal conductivity (3.78 W m À1 K À1 ), and highest heat storage density (3935 kJ kg À1 ), which is 5.9 times higher than pure lithium hydroxide aer the same hydration duration. The heat storage density and the thermal conductivity of Ni-CNTs-LiOH$H 2 O are great higher than 1D MWCNTs and 2D graphene oxide modied LiOH$H 2 O. The selection of 3D carbon nanoadditives as composed part of the chemical heat storage materials is a very efficient way to enhance comprehensive performance of heat storage activity component.

Conflicts of interest
There are no conicts to declare.