Thermochemical Storage of Middle Temperature Wasted Heat by Functionalized C / Mg ( OH ) 2 Hybrid Materials

For the thermochemical performance implementation of Mg(OH)2 as a heat storage medium, several hybrid materials have been investigated. For this study, high-performance hybrid materials have been developed by exploiting the authors’ previous findings. Expanded graphite (EG)/carbon nanotubes (CNTs)-Mg(OH)2 hybrid materials have been prepared through Mg(OH)2 deposition-precipitation over functionalized, i.e., oxidized, or un-functionalized EG or CNTs. The heat storage performances of the carbon-based hybrid materials have been investigated through a laboratory-scale experimental simulation of the heat storage/release cycles, carried out by a thermogravimetric apparatus. This study offers a critical evaluation of the thermochemical performances of developed materials through their comparison in terms of heat storage and output capacities per mass and volume unit. It was demonstrated that both EG and CNTs improves the thermochemical performances of the storage medium in terms of reaction rate and conversion with respect to pure Mg(OH)2. With functionalized EG/CNTs-Mg(OH)2, (i) the potential heat storage and output capacities per mass unit of Mg(OH)2 have been completely exploited; and (ii) higher heat storage and output capacities per volume unit were obtained. That means, for technological applications, as smaller volume at equal stored/released heat.


Introduction
About 20%-50% of the energy consumed by industrial processes is lost as waste heat contained in streams of hot exhaust gases and liquids, as well as through heat conduction, convection, and radiation from hot equipment surfaces and from heated product streams [1].This waste heat is produced whenever the operation is running.The current restrictive global energetic policy, the conservation of fossil resources, CO 2 abatement, and cost savings are the driving forces for the development of more efficient processes [2].
Thermal energy storage (TES) systems comprise all of the technologies that store thermal energy by heating or cooling a storage medium so that the stored energy can be used at a later time [2,3].Indeed, through TES technologies, industrial users can put wasted energy back into the process that created it, route the power somewhere else in the facility, or sell it to the grid.Thus, reducing peak demand, energy consumption, CO 2 emissions and costs, while increasing overall efficiency of the processes.Since the energetic demand of industrial processes may vary during any given day and from one day to next, TES systems can help to balance energetic demand and supply on a daily, weekly and even seasonal basis.
There are three kinds of TES systems: (1) sensible heat storage, which is based on storing thermal energy by heating or cooling a liquid or solid storage medium (e.g., water, sand, molten salts, rocks) [4]; (2) latent heat storage through phase change materials-PCMs-(e.g., from a solid state into a liquid state) [5]; and (3) thermo-chemical storage (TCS), which uses reversible chemical reactions to store and release thermal energy on demand [6].
Currently, TES systems based on sensible heat are commercially available while TCS and PCM-based storage systems are mostly under development and demonstration.The commercial deployment of TCS and PCM technologies is dependent on challenges in energy density, power performance, lifetime, charging capabilities, and costs.
Sensible heat storage is relatively inexpensive compared to PCM and TCS technologies.However, it requires large volumes because of its low energy density (i.e., three and five times lower than that of PCM and TCS systems, respectively) [3].TCS can offer even higher storage capacities [7] and is especially suitable for long-term energy storage (e.g., seasonal storage), considering that heat is stored until the reactants remain separate for an infinitely long time.This means that heat can be released when required by the user.
For the TCS technology, materials are the key issues.The choice of a TCS material must be defined on the base of several parameters [8] Several TCS materials have been selected and investigated [9] on the basis of the desired waste heat temperatures.For instance, carbonate systems (e.g., PbCO 3 /PbO, CaCO 3 /CaO) [10,11] and Ca(OH) 2 /CaO [12] can be used to store heat at 450 • C and above.Metallic hydrates (e.g., LiH/Li, CaH 2 /Ca, MgH 2 /Mg) [13] and redox systems (e.g., Co 3 O 4 /CoO, BaO 2 /BaO) [14,15] have been studied for heat storage between 700 • C and 1100 • C.
This study focuses on the Mg(OH) 2 /MgO TCS system, for middle-temperature waste heat.Indeed, it operates in the temperature range 200-400 • C that is typical for exhaust gas of different processes (e.g., gas turbine exhaust, steam boiler exhaust, drying and baking ovens, reciprocating engine exhaust).The Mg(OH) 2 /MgO TCS system is based on the following reversible chemical reaction and is characterized by the following reaction enthalpy (∆H 0 ) [16]: The endothermic Mg(OH) 2 dehydration reaction corresponds to the heat storage phase.The heat is released when required through the exothermic hydration.The water vapour produced during the dehydration reaction is condensed and stored in a water tank and then reused for the subsequent hydration reaction.
The Mg(OH) 2 /MgO system, according to the Gibbs' phase rule, is mono-variant.The reaction temperature (T) and the vapor partial pressure (P H 2 O /P 0 ) are correlated by a univocal correspondence through the integral Van't Hoff equation: where ∆S 0 is the reaction (1) entropy (equal to 152.4 kJ/mol•K) and R is the ideal gas constant (equal to 8.314 J/mol•K).At equilibrium conditions, the reaction temperature (T eq ) is calculated as follows: The dehydration reaction requires an extra heat supply, in addition to the reaction heat, to raise the sample's temperature to the dehydration point.
The performance implementation of this storage medium is still an object of study.In order to increase its heat transfer property and to avoid the coalescence of the MgO product, which leads to grain growth and loss of bulk porosity diminishing the MgO rehydration kinetics, several strategies have been investigated.With vermiculite was used for the confinement of Mg(OH) 2 in its pores [17] and to avoid the coalescence of the MgO product.With vermiculite, it was not possible to improve the heat transfer property of Mg(OH) 2 considering its well-known low thermal conductivity.Zamengo et al. [18] experimentally observed the benefits of using expanded graphite (EG) as a heat-transfer enhancer for the dehydration/hydration reaction of Mg(OH) 2 /MgO system.However, the composite exhibited contact instability between two components.For improving the interaction between the organic (EG) and inorganic phase (Mg(OH) 2 ), the authors used a synthetic strategy through a deposition-precipitation method [19].The compatibility between the two components has been improved by the electrostatic interaction promoted by the different point of zero charge (pH PZC ) of the materials.This synthetic strategy has also been applied on carbon nanotubes (CNTs) [20], which have been chosen not only for their high thermal conductivity but also for acting as a Mg(OH) 2 confining matrix due to their porous structure.From this study [20], it was demonstrated that thanks to the presence of CNTs the maximum potential heat storage capacity of Mg(OH) 2 per mass unit (~1300 kJ/ kgMg(OH) 2 ) has been fully exploited.The introduction of oxygenated functionalities upon CNTs surface was fundamental to bettering the MgO hydration reaction kinetics [20].
A critical evaluation of the thermochemical performances of EG/CNTs-Mg(OH) 2 hybrid materials is proposed here.In this study, the synthetic approach by deposition-precipitation previously studied by the authors in [19,20] has been extended to another type of carbonaceous support: graphene oxide (GO).Carbonaceous materials have been used both as such and functionalized.In case of CNTs, functionalization, i.e., surface oxidation, has been carried out with HNO 3 vapours according to the procedure reported in literature [21].Functionalized CNTs for briefness are named F-CNTs.While, in the case of EG, a stronger oxidation treatment has been used to obtaining GO: the Hummers method [22].It is expected that the functionalization treatment over EG surface could improve the MgO hydration reaction kinetics as previously observed in case of F-CNTs.
The heat storage performances of the carbon-based hybrid materials have been investigated through an experimental simulation of the heat storage/release cycles.The comparison between the developed materials has been carried out in terms of heat storage and output capacities per mass and volume unit in order to understand (i) if the maximum potential heat storage capacity of Mg(OH) 2 could be exploited and (ii) the impact of the presence of the carbonaceous phase on the estimated volumetric energy density.

Carbonaceus Materials
Two carbon-type materials have been used for this study: exfoliated graphite (EG) and CNTs.EG has been supplied by TIMREX C-THERM 002 TIMCAL Ltd. (Bironico, Switzerland), while CNTs were prepared by catalytic chemical vapor deposition (CCVD) of iso-butane (i-C 4 H 10 ) as the C source in the presence of Fe dispersed on an Al 2 O 3 catalyst, then purified to eliminate the catalyst residue [23].
EG and CNTs have been applied to the hybrid materials synthesis as such and modified through the insertion of oxygenated functionalities on their surface.Specifically, EG was treated though the well-known Hummers' method [22] to obtain GO, while, CNTs were oxidized in nitric acid vapour (HNO 3 ) at T = 135 • C for 120 min, in agreement with the procedure reported in literature [21].The sample was named F-CNTs.With respect to the previous study [20], an implemented functionalization apparatus permitted to gain higher HNO 3 concentration in the vapour phase thus resulting in a stronger oxidative effect.The extent of the effects of the two functionalization treatments has been evaluated through thermogravimetric (TG) and derivative thermogravimetric analysis (DTG) in inert atmosphere (Ar) in the temperature range between 100 • C and 900 • C (20 • C/min).The results are shown in Figure A1 of Appendix A.1.

Carbon-Based Hybrid Materials Preparation and Characterization
For the preparation of the Carbon/Mg(OH) 2 hybrid materials a deposition-precipitation route has been employed [20].In detail, a specified amount of carbonaceous material (EG, GO, CNTs, or F-CNTs) was sonicated for 30 min in 50 mL of Mg(NO 3 ) 2 (purchased by Sigma Aldrich, Saint Louis, USA, Assay: 99%) solution.Then, 150 mL of NH 4 OH 1 M (purchased by Carlo Erba, Milan, Italy, Assay: 30 wt %) solution was gradually added (2.5 mL/min) under magnetic stirring through a peristaltic pump.The final solution (pH ≈ 11.5) was aged at ambient conditions for 24 h, after which it was vacuum filtered (0.22 µm) and the collected solid was washed with deionized water and dried in a vacuum oven at 50 • C for 16 h.
The Mg 2+ ions concentration in solution was chosen on the basis of the desired Mg(OH) 2 deposited amount.The initial moles of Mg 2+ in the solution are those that theoretically should be precipitated in case of complete precipitation according to the stoichiometric reaction: The Mg(OH) 2 load (wt %) of the hybrid materials was fixed at ~30% accordingly to previous studies [20], having obtained with this amount the better dispersion of Mg(OH) 2 among the CNTs.
Pure Mg(OH) 2 was synthesized according to the precipitation reaction reported above.
The prepared samples are listed in Table 1, reporting the relative carbon type of which the hybrid materials are made of, the Mg(OH) 2 load (wt %) and the bulk (or apparent) density.
The as-prepared samples were analysed by means of scanning electron microscopy (SEM, Quanta 450, FEI, Hillsboro, OR, USA) and X-Ray Diffraction (XRD, Bruker D8 Advance, Bruker, Billerica, MA, USA) to determining their morphology and crystal structure.The SEM analysis were performed on Cr-metallized samples and operating with an accelerating voltage of 10 kV under high vacuum conditions (6.92 × 10 −5 Pa).The XRD apparatus has a Bragg-Brentano θ-2θ configuration, with Cu Kα radiation (40 V, 40 mA).XRD patterns were collected in the range 10 • -80 • with a step of 0.1 • /s.

Experimental Simulation of a Heat Storage/Release Cycle
A heat storage/release cycle basically consists of the succession of a dehydration and a hydration reaction.It was experimentally simultated through a TG unit (STA 449 F3 Jupiter, Netzsch, Selb, Germany) coupled with an evaporation system for the water vapour supply.The TG apparatus is equipped with a special furnace suitable for analysis in water vapour atmosphere.A dehydration/hydration cycle representative of the chemical heat pump (CHP) was carried out on ~20 mg as follows: the sample was first dried at 110 • C in inert atmosphere (N 2 flow rate: 100 mL/min) for 60 min to remove the physically-adsorbed water.Then, the temperature was increased at a constant rate (10 • C/min) to the desired dehydration temperature (T d =350 • C) and Mg(OH) 2 dehydration reaction proceeded over 120 min.After the complete dehydration reaction, the temperature was decreased at constant rate (−10 • C/min) to the hydration temperature (T h = 125 • C).The water vapour necessary for the MgO hydration reaction was supplied by the evaporation unit at 2.22 g/h and mixed with 35 mL/min N 2 as carrier gas.After the hydration reaction proceeded over 120 min, the water vapour supply was stopped and the sample was kept at 110 • C for 10 min under a constant N 2 flow (100 mL/min) to physically remove the adsorbed water from the sample.This process was repeated for each cycle.

Numerical Evaluation of the Hybrid Materials' Performances
Congruent with previous studies [19,20], the performances of the samples have been expressed in terms of reacted fraction (β (%)), calculated by the Equation (5): where ∆m real (%) is the instantaneous real mass change, and ∆m th (%) is the theoretical mass change due to the dehydration of 1 mol Mg(OH) 2 normalized to the total amount present in the sample, respectively expressed by Equations ( 4) and ( 5): where m in (g) and m inst (g) are respectively the initial sample mass and the instantaneous mass during the TG analysis.The dehydration and hydration conversions (∆β d/h (%)) are respectively expressed by Equations ( 8) and ( 9): where β i d (%) and β f d (%) are, respectively, the reacted fraction at the beginning and at the end of the dehydration treatment.While β h (%) is the final reacted fraction of MgO at the point of water supply termination.
The heat storage and output capacities per mass (Q M s/r (kJ/kg Mg(OH) 2 )) and volume unit (Q V s/r (MJ/m 3 )) were calculated respectively by the following Equations ( 10) and ( 11): where ∆H 0 (kJ/mol) is the enthalpy of reaction, and c (kg/m 3 ) is the density of the sample.

Samples' Morphological and Structural Characterizations
As reported in Table 1, the hybrid materials containing the F-CNTs (FNM sample: functionalized nanotubes and magnesium hydroxide) and GO (GOM sample: graphene oxide and magnesium hydroxide) are characterized by higher bulk density values than the ones made of EG (GM sample: graphite and magnesium hydroxide) and CNTs (NM sample: nanotubes and magnesium hydroxide), and pure Mg(OH) 2 .Specifically, the bulk density of the materials ranks in the order: FNM > GOM > Mg(OH) 2 > NM > GM.
The morphological aspect of the as prepared samples is shown by SEM micrographs in Figure 1.Pure Mg(OH) 2 (Figure 1a) has a plate-like morphology typical of brucite [24] and the platelets naturally tend to agglomerate.Upon Mg(OH) 2 deposition-precipitation over non-oxidized EG and CNTs a scarce interaction between the two components can be observed (Figure 1b,c).Indeed, the deposition-precipitation of Mg(OH) 2 upon EG (GM sample) leads to a poor carbonaceous surface covering (Figure 1b).In the case of the NM sample, few Mg(OH) 2 clusters (Figure 1c white circles) are sparse among the CNT bundles (Figure 1c black arrow).Both in GM and NM samples it cannot be excluded that bulk Mg(OH) 2 precipitation occurs without carbon (EG/CNTs) interaction.The lower bulk density values of GM and NM samples could be clarified at the light of their morphology.Indeed, due to the scarce interaction between Mg(OH) 2 and the carbonaceous phase, and to the larger amount of carbon (ca.65-70 wt %) with respect to Mg(OH) 2 (ca.30%-35%), the main contribution to the bulk density of hybrid samples could be determined by the carbonaceous phase that is characterized by low bulk density values (respectively, 40 and 170 g/cm 3 for EG and CNTs).
A stronger interaction is observed between Mg(OH) 2 and oxidized carbonaceous materials: GO and F-CNTs.Specifically, in the case of the GOM sample, Mg(OH) 2 densely packs the GO phase (Figure 1d), and, in the FNM sample, Mg(OH) 2 particles are homogeneously dispersed among the F-CNTs (Figure 1e).The more packed morphology of these last samples is totally in agreement with their bulk density values.
For the sake of completeness, SEM analysis of the carbonaceous materials are reported in Figure A2 of Appendix A.2. and F-CNTs.Specifically, in the case of the GOM sample, Mg(OH)2 densely packs the GO phase (Figure 1d), and, in the FNM sample, Mg(OH)2 particles are homogeneously dispersed among the F-CNTs (Figure 1e).The more packed morphology of these last samples is totally in agreement with their bulk density values.
For the sake of completeness, SEM analysis of the carbonaceous materials are reported in Figure A2 of Appendix A.2. XRD analysis results are reported in Figure 2. As a common feature, the samples patterns show peaks relative to crystalline Mg(OH)2 (2θ: 18.5°, 32.5°, 38°, 50.6°, 58.5°, 62°, 68°, 72°) and to the carbonaceous phase (2θ: 26.5°, 42.5°, 45°, 51°, 55°, 78°) in agreement with standard data (respectively JCPDS 7-0239 and JCPDS 25-0284), and no peaks arising from impurities.In the case of the GOM sample, it has to be pointed out that the most intense peak typical of GO at 2θ = 11.4°(shown in Figure A3 of Appendix A.3) has not been detected (see Figure 2c).The reason can be attributed in the preparation method of the hybrid material.Indeed, the Mg(OH)2 deposition-precipitation over GO, as explained in the Materials and Methods section (Section 2), is driven by Equation ( 2), which gives as a by-product ammonium nitrate (NH4NO3) and residual ammonia (NH4OH) being that the reaction is carried out with a large stoichiometric excess of the latter reagent.Ammonium hydroxide can be easily intercalated in the GO layers [25] and ammonium hydroxide acts as reducing agent [26].It can be argued that the synergistic action of these two reagents leads to obtaining graphene, which explain why the GO diffraction peak cannot be detected, given that long-range order of the layer is too poor [27].This is also in agreement with the small-angle XRD (1°-10°) pattern of the GOM sample (shown in Figure A3b of Appendix A.3), in which no peaks can be detected., 78 • ) in agreement with standard data (respectively JCPDS 7-0239 and JCPDS 25-0284), and no peaks arising from impurities.In the case of the GOM sample, it has to be pointed out that the most intense peak typical of GO at 2θ = 11.4 • (shown in Figure A3 of Appendix A.3) has not been detected (see Figure 2c).The reason can be attributed in the preparation method of the hybrid material.Indeed, the Mg(OH) 2 deposition-precipitation over GO, as explained in the Materials and Methods section (Section 2), is driven by Equation ( 2), which gives as a by-product ammonium nitrate (NH 4 NO 3 ) and residual ammonia (NH 4 OH) being that the reaction is carried out with a large stoichiometric excess of the latter reagent.Ammonium hydroxide can be easily intercalated in the GO layers [25] and ammonium hydroxide acts as reducing agent [26].It can be argued that the synergistic action of these two reagents leads to obtaining graphene, which explain why the GO diffraction peak cannot be detected, given that long-range order of the layer is too poor [27].This is also in agreement with the small-angle XRD (1 • -10 • ) pattern of the GOM sample (shown in Figure A3b of Appendix A.3), in which no peaks can be detected.
reaction is carried out with a large stoichiometric excess of the latter reagent.Ammonium hydroxide can be easily intercalated in the GO layers [25] and ammonium hydroxide acts as reducing agent [26].It can be argued that the synergistic action of these two reagents leads to obtaining graphene, which explain why the GO diffraction peak cannot be detected, given that long-range order of the layer is too poor [27].This is also in agreement with the small-angle XRD (1°-10°) pattern of the GOM sample (shown in Figure A3b of Appendix A.3), in which no peaks can be detected.For a deeper comprehension of the Mg(OH) 2 crystal structure of the different hybrid materials, Figure 3 shows the comparison between the peaks relative only to the hydroxide phase.The Mg(OH) 2 intensity ratios of reflections [001] to [101] and [110] (I 001 /I 101 , I 001 /I 110 ), which are labeled in Figure 3, are listed in Table 2.For a deeper comprehension of the Mg(OH)2 crystal structure of the different hybrid materials, Figure 3 shows the comparison between the peaks relative only to the hydroxide phase.The Mg(OH)2 intensity ratios of reflections [001] to [101] and [110] (I001/I101, I001/I110), which are labeled in Figure 3, are listed in Table 2.In pure Mg(OH)2, peaks are narrower and more intense than those of the hybrid materials.GOM and NM samples show similar features.The FNM sample is characterized by the broadest and less In pure Mg(OH) 2 , peaks are narrower and more intense than those of the hybrid materials.GOM and NM samples show similar features.The FNM sample is characterized by the broadest and less intense Mg(OH) 2 peaks, indicative of smaller particle sizes which is also in agreement with SEM analysis (Figure 1e).Mg(OH) 2 particles, when pure, or combined with EG, are preferentially oriented to the (001) direction, while, in case of GOM, NM, and FNM samples, Mg(OH) 2 particles are preferentially oriented to the (101) direction.

Dehydration and Hydration Reactions
Mg(OH) 2 and hybrid materials have undergone three dehydration/hydration cycles experiments under conditions reported in Section 2.3.
Figure 4a,b shows the dehydration reaction curves, expressed in terms of reacted fraction (β (%)) as a function of reaction time (t (min)).The curves are relative to the third cycle, when the thermochemical behavior of the samples has been observed to be stable [19,20].The reaction conversion rate (dβ/dt (%/min)) curves, relative to the first 60 min of the dehydration reaction, are shown in Figure 4c,d

Sample Code
T peak ( From the dehydration profiles it is evident that, in the case of pure Mg(OH) 2 , GM and NM samples show a single decomposition step occurring between 200 • C and 350 • C (Figure 4a,b) that can be clearly addressed to the Mg(OH) 2 dehydration reaction.GOM and FNM samples present a change in the curve slope which allows two overlapped dehydration processes to take place according to different reaction rates (Figure 4a,b).The fastest, which occurs between 200 • C and 350 • C, is clearly addressed to the Mg(OH) 2 dehydration reaction (as already seen for Mg(OH) 2 , GM, and NM samples).The slowest, which occurs at lower temperatures (110-200 • C), can presumably be indicative of the structural water removal which can be retained by the hydrophilic oxygenated groups present on the carbonaceous surface [20].For the sake of completeness, the TG profiles of GO and F-CNTs from 100 • C up to 1000 • C in an inert ambient environment (Ar 100 mL/min) are reported in Figure A1 of Appendix A.1.They show that upon heating up to 350 • C to drive the Mg(OH) 2 dehydration, oxygenated groups present on GO and F-CNTs surfaces are only partially decomposed.The remaining groups are able to physically retain the adsorbed water, thus explaining the first dehydration step.
The T peak of pure Mg(OH) 2 and FNM sample is 307 • C (Figure 4d).This moves to lower values in the case of GM and NM samples, respectively, 292 • C and 282 • C (Figure 4c,d); while, in the case of the GOM sample, it is shifted towards a higher value: 318 • C (Figure 4c).The dehydration rate ranks in the order: Mg(OH) 2 ≈ GOM > GM > NM > FNM (the exact values are listed in Table 3).
The hydration reaction proceeds faster for hybrid materials than pure Mg(OH) 2 , in particular for GOM and FNM samples (see Figure 5a,b).The hydration rate ranks in the order: GOM ≈ FNM > GM > NM > Mg(OH) 2 .
Pure Mg(OH) 2 exhibits the lowest reaction conversion (Figure 6).Hybrid materials show even higher hydration conversion than pure Mg(OH) 2 .The highest ∆β d/h values are achieved for GOM and FNM samples.In this case reaction conversions slightly exceed the 100% value due to the contribution of the water physisorbed on the functionalized carbonaceous surface.The presence of physisorbed water has been previously observed in the dehydration curves plotted in Figure 4a,b.
Pure Mg(OH)2 exhibits the lowest reaction conversion (Figure 6).Hybrid materials show even higher hydration conversion than pure Mg(OH)2.The highest Δβd/h values are achieved for GOM and FNM samples.In this case reaction conversions slightly exceed the 100% value due to the contribution of the water physisorbed on the functionalized carbonaceous surface.The presence of physisorbed water has been previously observed in the dehydration curves plotted in Figure 4a,b.As it can be observed from the curves of the stored and released heat per mass unit (Figure 7c,d), congruent to previous studies [20] with pure Mg(OH)2, only half of its capacity is used.Hybrid materials show higher stored and released heat per mass unit than Mg(OH)2, with GOM and FNM samples fully exploiting Mg(OH)2 heat storage/output capacity.
The stored/released heat per volume unit capacity (Q s/r V ), i.e., the volumetric energy density, is a benchmark for comparing the different storage materials.This gives an indication of the global efficiency of the hybrid material.Indeed, despite its fundamental role in enhancing the performances  of the storage medium, the carbonaceous phase could be intended as an "inert" component since it does not directly participate to the dehydration/reaction.The stored/released heat per volume unit curves are plotted in Figure 7c,d.
The stored/released heat capacity per volume unit (i.e., volumetric energy density) of a material, as can be derived from its expression in Equation ( 11), directly depends from the bulk material density and its dehydration/hydration conversion.Hence, at equal Mg(OH)2 content the higher these two parameters are, the higher the volumetric energy density.In case of hybrid materials, it also depends from the Mg(OH)2 load: the lower the content, the lower the volumetric energy density.Due to its high density and dehydration/hydration conversion, the FNM sample exhibits the highest volumetric energy density, ~340 MJ/m 3 .Pure Mg(OH)2 volume energy density is comparable to that of the GOM sample (~220 MJ/m 3 ).NM and GM samples exhibit the lowest volumetric energy density, respectively, at ~80 and ~52 MJ/m 3 .

Discussion
Despite pure Mg(OH)2 exhibiting the highest dehydration reaction rate (Figure 4), it shows the lowest reaction conversion and the worst rehydration kinetics, in terms of reaction rate.
Demonstrating that, in agreement with previous studies [19,20], the carbonaceous phase is fundamental in order to improve the thermochemical performances of storage medium in terms of reaction conversion and kinetics, it was evaluated with this study if and how the functionalization treatment influences the thermochemical performances of either CNT-and EG-based hybrid materials.
Both functionalized CNTs and EG allowed obtaining the highest reaction conversion and the As it can be observed from the curves of the stored and released heat per mass unit (Figure 7c,d), congruent to previous studies [20] with pure Mg(OH) 2 , only half of its capacity is used.Hybrid materials show higher stored and released heat per mass unit than Mg(OH) 2 , with GOM and FNM samples fully exploiting Mg(OH) 2 heat storage/output capacity.
The stored/released heat per volume unit capacity (Q V s/r ), i.e., the volumetric energy density, is a benchmark for comparing the different storage materials.This gives an indication of the global efficiency of the hybrid material.Indeed, despite its fundamental role in enhancing the performances of the storage medium, the carbonaceous phase could be intended as an "inert" component since it does not directly participate to the dehydration/reaction.
The stored/released heat per volume unit curves are plotted in Figure 7c,d.The stored/released heat capacity per volume unit (i.e., volumetric energy density) of a material, as can be derived from its expression in Equation (11), directly depends from the bulk material density and its dehydration/hydration conversion.Hence, at equal Mg(OH) 2 content the higher these two parameters are, the higher the volumetric energy density.In case of hybrid materials, it also depends from the Mg(OH) 2 load: the lower the content, the lower the volumetric energy density.Due to its high density and dehydration/hydration conversion, the FNM sample exhibits the highest volumetric energy density, ~340 MJ/m 3 .Pure Mg(OH) 2 volume energy density is comparable to that of the GOM sample (~220 MJ/m 3 ).NM and GM samples exhibit the lowest volumetric energy density, respectively, at ~80 and ~52 MJ/m 3 .

Discussion
Despite pure Mg(OH) 2 exhibiting the highest dehydration reaction rate (Figure 4), it shows the lowest reaction conversion and the worst rehydration kinetics, in terms of reaction rate.
Demonstrating that, in agreement with previous studies [19,20], the carbonaceous phase is fundamental in order to improve the thermochemical performances of storage medium in terms of reaction conversion and kinetics, it was evaluated with this study if and how the functionalization treatment influences the thermochemical performances of either CNT-and EG-based hybrid materials.
Both functionalized CNTs and EG allowed obtaining the highest reaction conversion and the fastest rehydration, which means that through FNM and GOM, in the case of their application as heat storage materials, it is possible to fully exploit the Mg(OH) 2 potential heat storage capacity, and that they can deliver to the user almost immediately (~10 min) all of the stored heat in the case of energetic demand, as evinced from Figure 7b,d.
According to the obtained results, the introduction of oxygenated functionalities over the carbonaceous surface has the dual role of: (i) improving the affinity of carbon with Mg(OH) 2 ; and (ii) enhancing the hybrid material affinity with water vapour.
The oxygenated groups on the carbonaceous surface improve the dispersibility of carbon in the aqueous solution during the Mg(OH) 2 synthesis reaction.Hence, more carbonaceous surface is available for Mg(OH) 2 deposition on it.Thus, through the improvement of carbon's affinity with Mg(OH) 2 , it was possible to deposit Mg(OH) 2 more compactly into the carbonaceous phase and hence to increase the density of the hybrid material (GOM and FNM samples).This allows for a higher volumetric energy density (see Figure 7c,d).The intrinsic porous nature of GO and F-CNTs permits the water vapour flowing through the bulk material from/to the deeper Mg(OH) 2 particles despite its compactness and hence allowing dehydration/hydration reaction.
The enhancement of the hybrid material affinity with water vapour allows for a faster rehydration kinetics (see Figure 5a,b).
It is worth to point out that the Mg(OH) 2 load of the hybrid materials is only ~30%.However, despite such low Mg(OH) 2 content, it was observed in the case of the FNM sample that a volumetric energy density increase with respect to pure Mg(OH) 2 .GOM and pure Mg(OH) 2 have a comparable volumetric energy density.Nevertheless, the rehydration kinetics of GOM are superior with respect to Mg(OH) 2 .
As a future goal, higher Mg(OH) 2 loads will be taken into account in order to further increase the volumetric energy density.The stability of the developed materials would be evaluated over more than 10 dehydration/hydration cycles.A study on the effect of heating rates on the heat storage step (dehydration reaction) will be considered.

Conclusions
In this study functionalized and un-functionalized EG/CNTs-Mg(OH) 2 hybrid materials have been analyzed for middle temperature thermochemical heat storage application through a laboratory-scale experimental simulation of a heat storage/release cycle.
It has been demonstrated that the carbonaceous materials improves the thermochemical performances of the storage medium in terms of hydration reaction rate and dehydration/hydration conversion with respect to pure Mg(OH) 2 .
The hybrid materials containing functionalized EG and CNTs (GOM and FNM samples) are able to fully exploit the potential heat storage and output capacities per mass unit of Mg(OH) 2 (~1300 kJ/kg Mg(OH) 2 ).
A higher density characterizes GOM and FNM samples.The functionalization treatment increases the compactness (and hence the density) of the hybrid materials thus allowing for higher volumetric energy density.That means, for technological application, smaller volume at equal stored/released heat.The GOM sample heat storage capacity is ~220 MJ/m 3 , which is comparable to that of 100% Mg(OH) 2 .The FNM sample exhibits an even higher heat storage capacity, about 340 MJ/m 3 .
For practical uses, further development of such C/Mg(OH) 2 hybrid materials would need to be evaluated over more than 10 dehydration/hydration cycles.In addition, higher Mg(OH) 2 loads will be taken into account in order to further increase their volumetric energy density.

Figure 3 .
Figure 3.Comparison between all the samples of the peaks relative only to the Mg(OH)2 phase.

Figure 3 .
Figure 3.Comparison between all the samples of the peaks relative only to the Mg(OH) 2 phase.
. Hydration curves are plotted in Figure 5a,b.The following reaction parameters are reported in Table 3: the temperatures of the derivative reacted fraction maximum peak (T peak ( • C)), the reaction conversion rate (dβ/dt (%/min)), and the dehydration/hydration conversions (∆β d/h (%)).The dehydration/hydration conversions are expressed considering a percentage error of 2% determined considering the measurement uncertainty derived from six repeated tests.Energies 2017, 10, 70 9 of 16 according to different reaction rates (Figure 4a,b).The fastest, which occurs between 200 °C and 350 °C, is clearly addressed to the Mg(OH)2 dehydration reaction (as already seen for Mg(OH)2, GM, and NM samples).The slowest, which occurs at lower temperatures (110-200 °C), can presumably be indicative of the structural water removal which can be retained by the hydrophilic oxygenated groups present on the carbonaceous surface [20].For the sake of completeness, the TG profiles of GO and F-CNTs from 100 °C up to 1000 °C in an inert ambient environment (Ar 100 mL/min) are reported in Figure A1 of Appendix A.1.They show that upon heating up to 350 °C to drive the Mg(OH)2 dehydration, oxygenated groups present on GO and F-CNTs surfaces are only partially decomposed.The remaining groups are able to physically retain the adsorbed water, thus explaining the first dehydration step.

3. 3 .
Heat Storage and Output Capacities Stored/released heat per mass unit (Q s/r M ) curves (Figure 7a,b) give an indication about how much of the Mg(OH)2 potential heat storage and output capacities have been exploited.The theoretical maximum heat storage capacity of pure Mg(OH)2 can be easily calculated from the dehydration reaction enthalpy change (ΔH 0 = ±81 kJ/mol) and its molecular weight (MMg(OH) 2 = 58.3g/mol).The value is 1389 kJ/kgMg(OH) 2 .

3. 3 .
Heat Storage and Output Capacities Stored/released heat per mass unit (Q M s/r ) curves (Figure 7a,b) give an indication about how much of the Mg(OH) 2 potential heat storage and output capacities have been exploited.The theoretical maximum heat storage capacity of pure Mg(OH) 2 can be easily calculated from the dehydration reaction enthalpy change (∆H 0 = ±81 kJ/mol) and its molecular weight (M Mg(OH) 2 = 58.3g/mol).The value is 1389 kJ/kg Mg(OH) 2 .Energies 2017, 10, 70 11 of 16

Figure 7 .
Figure 7. (a) Stored and (b) released heat per mass unit; (c) stored and (d) released heat per volume unit.

Figure 7 .
Figure 7. (a) Stored and (b) released heat per mass unit; (c) stored and (d) released heat per volume unit. :