New insights into Mn2O3 based metal oxide granulation technique with enhanced chemical and mechanical stability for thermochemical energy storage in packed bed reactors

DOI: https://doi.org/10.1016/j.solener.2022.06.010 © 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/(opens in new tab/window) This is the accept manuscript of the following article that appeared in final form in Solar Energy 241: 248-261 (2022), which has been published in final form at https://doi.org/10.1016/j.solener.2022.06.010. © 2022 International Solar Energy Society. Published by Elsevier Ltd. under CC BY-NC-ND licence (https://creativecommons.org/licenses/by-nc-nd/4.0/) New insights into Mn2O3 based metal oxide 1 granulation technique with enhanced chemical 2 and mechanical stability for thermochemical 3 energy storage in packed bed reactors 4 Daniel Bielsaa b *, Mikel Oreguib and Pedro L. Ariasb 5 a Centre for Cooperative Research on Alternative Energies (CIC energiGUNE), Basque Research and 6 Technology Alliance (BRTA), Albert Einstein 48, 01510, Miñano (Álava), Spain 7 b University of the Basque Country (UPV/EHU), calle Alameda Urquijo s/n, 48013 Bilbao, Spain 8 9 *Corresponding author at: [email protected] 10 Abstract 11 High temperature thermochemical energy storage still requires a significant research effort. 12 Most of the research has been carried out with materials at lab-scale and proper material 13 fabrication techniques need to be developed in order to make feasible the upscaling of the 14 technology. Agglomeration, abrasion or low volumetric energy density are some of the 15 negative consequences observed when trying to pass from the powder state to the material 16 shape and amount required for a thermochemical reactor. In this work, a granulation 17 technique is investigated, using a Si-doped manganese oxide as active material, determining 18 the critical parameters that provide the best chemical and mechanical stability of the 19 granules. The granulation process uses a polymeric binder to give consistency to the granules 20 and afterwards, it is removed to create a porous structure to facilitate the oxygen diffusion 21 in and out of the granule. We identified the importance of decreasing the solubility of the 22 binder to increase the volumetric energy density of the granules. Furthermore, it was 23 observed that increasing the mechanical stability through a high temperature treatment 24 does not decrease the chemical stability of the material. In order to provide the first insights 25 into the scalability of the solution, the chemical and mechanical stability of the granules has 26 been satisfactorily checked during 100 redox cycles, out of which 50 were carried out in a 27 home-made lab-scale packed bed reactor with an inner diameter of 15 mm and another 50 28 redox cycles in a simultaneous thermal analyzer. 29 30 31 Keywords: Thermochemical energy storage; Doped metal oxides; Redox reaction; 32 Concentrated solar power plant; Sintering inhibition; Packed bed reactor 33 1. Introduction 34 High temperature thermochemical energy storage (TcES) is at an early stage of development 35 and thus, little experimental research has been reported at reactor scale. For concentrated 36 solar power (CSP) plants, two existing possibilities have been principally considered. On the 37 one hand, direct use of solar radiation, where the thermochemical material absorbs the solar 38 thermal energy in a receiver, which acts at the same time as a reactor. On the other hand, 39 indirect use, where the thermochemical material is stored in a reactor and the heat exchange 40 is carried out by means of an intermediate heat transfer fluid (Zsembinszki et al., 2018). It has 41 to be noted that the selection of the adequate reactor technology cannot be detached from 42 the CSP configuration, provided that it should be integrated into the operation of the plant. 43 There is still little work concerning the configuration of new CSP generation at high 44 temperature. Nevertheless, one preliminary conclusion is that the TcES cannot simply replace 45 a conventional thermal energy storage system in a CSP such as a commercial molten salt 46 system (Schmidt and Linder, 2017) (Pelay et al., 2019) (Ströhle et al., 2016). Overall, taking 47 into consideration the advantages and disadvantages of the reported technologies, packed 48 bed reactors are the simplest in construction, which makes them an appropriate candidate 49 for a first technology upscaling and integration step in CSP plants. They are easy to build and 50 operate but their main drawbacks are the pressure drop inside the bed, that may induce 51 preferred gas channeling and the limited available contact surface of the reactants, which may 52 compromise the thermal power output (Wokon et al., 2017b). In this regard, the 53 thermochemical material properties, namely, shape, size and structural composition and 54 stability may have a significant impact on the performance of a thermochemical packed bed 55 reactor. Consequently, it may become one of the most important challenges of this 56 technology. 57 Metal oxides are one of the most studied thermochemical materials for high temperature TcES 58 in CSP plants, since air can be used both as heat transfer fluid and as reactant, simplifying the 59 system integration. Nevertheless, most of the scientific studies that aimed to assess the 60 behaviour of metal oxides use only several mg in the form of micro-sized particle powder at 61 lab-scale(Neises et al., 2012) (Pestalozzi, 2013), whereas the technology upscaling up to a 62 packed bed thermochemical reactor for a real application might require kg or even tons of 63 material. This material amount, in the shape of fine powder, may cause extremely high 64 pressure drop and extra pumping power. In addition, fine powder is more prone to 65 agglomerate, which may contribute to enhance channelling and hinder re-oxidation kinetics 66 (Carrillo et al., 2014). Therefore, studies on technology up-scaling consider different material 67 preparation approaches (e. g. material pelletization or granulation, adding supports), in order 68 to produce particles in the scale of mm or cm, with sufficient mechanical and chemical stability 69 capable of withstanding a great number of thermal cycles. 70 In the case of a thermochemical packed bed reactor, the particles are subjected to different 71 stresses, namely, chemical, mechanical, and thermal stresses. The volume changes due to 72 both phase transitions and thermal expansion and shrinkage, together with the pressure 73 induced by the weight of the upper particle layers cause particle-wall or particle-particle 74 friction, known as ratcheting. In addition, pore reduction caused by particle sintering may 75 create overpressure during gas release in solid-gas thermochemical materials. As a result, 76 particles can crack, or their surface can be eroded, leading to a decrease on the void fraction 77 due to reallocation of particles. These negative effects can be significant on the lowest layers, 78 which are subjected to higher weight loads. Taking into account the cycling operation nature 79 of the TES systems and the long life-expectancy required for CSP plants, it can lead to a 80 significant increase of the pressure drop or to the collapse of the container wall. Therefore, 81 the metal oxide particles for a real application, together with fast kinetics, should demonstrate 82 sufficient mechanical strength to ensure reliable stability during the lifetime of these TES 83 systems, which may last several decades. 84 Particle strength can be increased by designing a proper synthesis route (e. g. sol-gel, spray-85 drying), applying a pre-treatment (e. g. high temperature sintering) or by adding binders or 86 support materials (e. g. Al2O3, TiO2 or ZrO2), provided that they do not react with the active 87 material leading to thermochemical deactivation. Another possibility is strengthening the 88 particle surface by encapsulation in a strong porous material or by rapid heating and cooling 89 of the particles. The latter would hinder changes in the core of the granules that might damage 90 the thermochemical material properties. 91 The addition of support materials in manganese-based oxides has been deeply studied both 92 for CO2 capture through chemical looping and TcES. The attrition resistance of manganese-93 iron oxide with the addition of Al2O3, MgAl2O4, CeO2, ZrO2 and Y2O3−ZrO2 was investigated by 94 G. Azimi in a fluidized bed reactor (Azimi et al., 2014)(Azimi et al., 2015). The materials were 95 prepared by spray-drying and calcinated up to 1200°C during 4 h. The research showed that 96 the crushing strength does not improve substantially with the addition of these supports, with 97 the exception of ZrO2, where a slight improvement was observed. Looking at chemical looping 98 applications, M. Abian observed that the addition of TiO2 may double the crushing strength of 99 manganese-iron oxides, reporting values of 3-5 N compared to the 1-2 N of the undoped 100 samples (Abián et al., 2017). 101 In regard of TcES research, N. C. Preisner studied the effect on the mechanical strength of 102 manganese-iron oxide by adding 20 wt% of different supporting materials, namely of ZrO2, 103 CeO2 and TiO2 for a moving bed reactor. The material was prepared using a build-up 104 granulation technique at 800°C during 10 h. Pure manganese-iron oxide particles showed a 105 clear tendency to agglomerate and to break into fine particles when subjected to thermal 106 cycling in air. Furthermore, the bed volume was increased by 17 % after 30 cycles in a packed 107 bed reactor containing 21 g of material, meaning that coarsening happened to some extent. 108 Nevertheless, both ZrO2 and CeO2 contributed to improve the attrition strength, with a slight 109 particle agglomeration and without coarsening. TiO2 reacts to form another stable phase and 110 thus, it cannot be considered for TcES (Neumann et al., 2018). 111 Other interesting approaches involved experiments with extruded metal oxide composites 112 (Pagkoura et al., 2015) and with a support structure, i. e. honeycomb, corderites or foams, 113 coated with the active redox material, through which the fluid can flow (Karagiannakis et al., 114 2016) (Singh et al., 2017) (Agrafiotis et al., 2015b) (Agrafiotis et al., 2015a) (Agrafiotis et al., 115 2016). The coated structures presented a better integrity over repetitive redox cycles, 116 although the volumetric energy storage is limited by the capacity of the active material 117 loading, leading to a substantial decrease in the energy storage density. 118 Concerning lab-scale thermochemical reactor testing, spherical shape particles present lower 119 void fraction and better homogeneity within the reactor bed and thus, it has been one of the 120 most targeted material geometries. Spray drying is a commonly used method to make high-121 performance, fluidizable particles for chemical looping combustion applications. Ibraheam et 122 al. studied the effect on the redox behaviour and sintering of spray-dried particles of 123 manganese oxide with Al2O3, ZrO2, and Fe2O3 (Ibraheam et al., 2019). All the materials showed 124 very poor mass change during redox cycling, since they underwent significant sintering during 125 calcination at 1200°C. Nevertheless, ZrO2 and Al2O3 samples demonstrated enough structural 126 stability under the temperatures tested, although Al2O3 addition tends to form a MnAl2O4 127 phase, which is stable at the tested temperatures. Wokon et al. studied the kinetic 128 performance of manganese-iron oxide granules of 1-3 mm prepared by a build-up granulation 129 technique without any support or binder in a thermobalance (Wokon et al., 2017a). The 130 corresponding amounts of Mn3O4 and Fe3O4 powders were mixed in an Eirich mixer, based on 131 the principle of intensive mixing by an inclined arranged rotating mixing pan, providing mixing 132 effect in vertical and horizontal directions through the application of a rotating micro-133 granulator mixing tool. It was observed a significant particle volume increase after 100 redox 134 cycles, which led to lower density and more fragile particles. This fact might compromise the 135 stability within continuous redox cycling at a real scale and therefore, their mechanical 136 stability shall be enhanced. 137 Hamidi et al. synthesized granules of manganese-iron oxide by intensive mixing in an Einrich 138 mixer, introducing a maltodextrin solution organic binder at the end of the mixing process 139 (Hamidi et al., 2019). Particle sizes of 0.5 - 1 mm were used to study the reduction reaction of 140 manganese-iron oxide in a small packed bed reactor containing approximately 35 g of the 141 thermochemical material. No data regarding the mechanical properties and stability of the 142 particles under cycling was provided. Another approach was studied by Gigantino et al., who 143 modified a lab-scale granulation process, known as drop technique, to obtain CuO particles of 144 1-2 mm of diameter (Gigantino et al., 2020). The process consisted in dissolving an organic 145 polymer with the active material to create a paste that is added dropwise to a bath where the 146 immiscibility leads to particle sphericity. Different combinations of polymers, solvents and 147 surfactants were studied, selecting the best combination which led to the highest particle 148 strength and sphericity, even though no data regarding the influence on the reaction kinetics 149 and storage density was reported. In addition, in order to reduce the sintering effect, the CuO 150 powder was mixed with a Y2O3/ZrO2 stabilized powder, turning into less agglomeration after 151 100 redox cycles, for amounts of Y2O3/ZrO2 above 50 wt%. 152 Considering that adding a support material decreases the volumetric energy storage density 153 and may deactivate the thermochemical material by undesired chemical reactions, we present 154 a comprehensive study of the granulation technique proposed by Gigantino et al., applied to 155 Si-doped manganese oxide. In a previous work we demonstrated the improved kinetics and 156 chemical stability of Si-doped manganese oxide in powder state, representing a promising 157 material candidate for technology upscaling (Bielsa et al., 2020), which stores and releases 158 heat according to the following equation: 159 6(𝑀𝑀𝑀𝑀0.99𝑆𝑆𝑆𝑆0.01)2𝑂𝑂3(𝑠𝑠)+149 𝐽𝐽/𝑔𝑔 ↔ 4(𝑀𝑀𝑀𝑀0.99𝑆𝑆𝑆𝑆0.01)3𝑂𝑂4 (𝑠𝑠) + 𝑂𝑂2 (𝑔𝑔) Eq. 1 160 161 The investigation carried out in the present work aims to identify all the effects of the different 162 granulation synthesis parameters on the material behavior, paying special attention to the 163 chemical and mechanical stability and to increase the active material content of the granules, 164 providing new insights into a potential TcES material preparation route for large scale packed 165 bed thermochemical reactors. In the first part of the work, we evaluate the influence of the 166 different synthesis route parameters on the material behavior, from the kinetics to the 167 chemical stability and mechanical behavior of the granules. Subsequently, granules exhibiting 168 the most promising properties were selected for the upscaling study where 8 g of the material 169 was subjected first to 50th redox cycling study in a lab-scale packed bed reactor. Finally, a 170 sample of the tested granules were subjected to additional 50 redox cycles in a STA to analyze 171 in detail their chemical stability after a total of 100 redox cycles. The satisfactory results 172 observed concerning granules mechanical and chemical stability represent a step forward in 173 high temperature thermochemical energy storage upscaling, that can be applicable not only 174 to the material under study but to other metal oxides. 175 176 2. Experimental 177 178 2.1 Characterization techniques 179 180 The particles size and morphology were determined by means of a Quanta 200 FEG scanning 181 electron microscope (SEM) operated in high vacuum mode at 20 kV and with a back scattered 182 electron detector (BSED). 183 Reaction kinetics and chemical stability were studied with the simultaneous thermal analyzer 184 STA 449 F3 Jupiter (Netzsch). In these studies, samples consisting in 3-4 granules accounting 185 for around 10 mg were placed into 85 µL open platinum/rhodium crucibles (Netzsch) and 186 subjected to charging and discharging cycles under an air stream of 100 mL/min. 187 The bulk density and true density of the different granules were measured using a helium 188 pycnometer AccuPyc II 1340. For the measurements a reference volume of 1 cm3 was 189 completely filled with the different composition granules, resulting in total mass varying from 190 0.1 to 0.3 g. 191 192 2.2 Granules synthesis route 193 194 The synthesis process applied, based on the methodology developed by Gigantino et al. 195 (Gigantino et al., 2020), consists of the following three steps: i) synthesis step, where the 196 precursors are transformed in a fine powder of the desired material, ii) granulation step, 197 where spherical granules are produced, and iii) hardening step, consisting in a sintering 198 process to enhance the mechanical stability of the granules. 199 First, Si-doped manganese oxide samples were synthesized by a sol-gel method, following the 200 procedure described in a previous work (Bielsa et al., 2021). The subsequent granulation 201 technique is based on preparing a mixture of this metal oxide powder and a solution of a 202 polymeric binder and a solvent. Subsequently the mixture is introduced dropwise through a 203 syringe into a precipitating bath, where the drops harden while the solvent leaves the granule. 204 After drying, the granules are calcined to remove the remaining organic matter and to 205 enhance their mechanical stability. As a consequence, the space occupied by the binder 206 becomes empty and then, porous solid granules are obtained. The following two conditions 207 shall be met: the solvent must be miscible with the precipitation bath and the polymeric 208 binder must be soluble in the solvent and insoluble in the precipitation bath. The preparation 209 steps are illustrated in Fig. 1. 210 211 Figure 1. Metal oxide spherical granules preparation route 212 Various granules of Si-doped Mn2O3 were prepared varying the proportions of the metal oxide 213 (MO), organic solvent (OS) and polymeric binder (PB) with the aim of finding the best 214 preparation route giving spherical shape granules with good mechanical and chemical 215 stability. The PB used was ethyl cellulose (Sigma Aldrich) and the OS 1-methyl 2-pyrrolidinone 216 (>99% Sigma Aldrich). The precipitation bath was made of deionized water, mixed with a 217 surfactant (Tween 80, Sigma Aldrich) to reduce its high surface tension and to promote the 218 solvent exit from the granule. 219 It was found that for the amounts of MO used (< 1 g), the ratio OS:PB shall be maintained in 220 the range 9:1. Lower and higher OS:PB ratio led to inconsistent granules in the precipitation 221 bath and after the hardening step, respectively. In principle, the amount of MO in the granules 222 is important because it might affect directly to the energy storage density, so high proportions 223 are preferred. Nevertheless, high MO proportions (MO/MO+PB > 0.7) increase the viscosity 224 of the mixture resulting in tile-like granules (Fig. 2a), being not practical to increase the MO 225 proportion above 0.8, since the viscosity imposes significant difficulties to the solution to flow 226 out of the syringe. In order to reduce the viscosity and maintain a high MO content, the OS+PB 227 mixture was heated at 40°C, contributing to improve the sphericity of the granules (Fig. 2b). 228 The distance between the precipitation bath and the syringe tip affected the granules in two 229 ways: distances > 5cm resulted in coin-shape granules due to high mechanical shock when the 230 granules reached the precipitation bath surface, and distances < 2 cm do not allow the 231 granules to penetrate the bath and remain floating in the surface for some time. Therefore, a 232 distance of 3 cm was selected for all the samples. The diameter of the syringe tip has a strong 233 influence on the final size of the granules, which should be maintained below 1:10-20 of the 234 reactor diameter in order to avoid gas channelling inside the thermochemical reactor 235 (Mederos et al., 2009). Therefore, a tip diameter of 2 mm was used, resulting in granules 236 within the average diameter range of 3-4 mm after step 2. Removing the water during the 237 drying step became critical, since it was observed that the water content after the synthesis 238 process may reach up to 80 wt% of the total granule mass and its extraction can destroy the 239 granule sphericity, unless it is carried out slowly. For that reason, the granules were dried at 240 room temperature until the water content was almost totally removed. Subsequently, the 241 granules were subjected to a temperature program consisting in heating up to 450°C at 242 1°C/min, followed by an isothermal step for 4 h to remove the organic matter (Fig. 2c). The 243 final hardening step consisted in another heating step at 2°C/min up to the corresponding 244 calcination temperature, where different isothermal steps were used to promote particle 245 sintering. A noticeable granule shrinkage was observed during the water and organic matter 246 removal steps, being greater the fewer the content of MO, resulting in granules with a final 247 size of around 1-3 mm of diameter. Table 1 shows the different materials prepared in this 248 work. 249 a) b) c) 250 251 Figure 2. Granules of Mn2O3 prepared by the drop technique: a) pronounced tile-like granules, b) 252 spherical granules and c) granules after the calcination step at 450°C 253 Table 1. Description of materials prepared in this work 254 Description MO Proportion (wt %)* Diameter (mm) Powder 1 N/A MO 50 50 1,83 MO 70 70 2,34 MO 80 80 2,46 *MO proportion (wt %) = MO/(MO+PB) 255 256 2.3 Experimental setup 257 258 The experimental test rig consists of two main parts: the heating setup and the 259 thermochemical reactor (Fig.3a). The generation of a controlled gas flow rate above 800°C 260 required to carry out the reduction and oxidation of the material may become a complex issue 261 and therefore, a furnace placed around the thermochemical reactor carries out the 262 heating/cooling process. The furnace setup temperature was varied between 550°C and 800°C 263 at a heating/cooling rate of 20°C/min and the reactions were conducted switching the gas 264 flow between N2 and O2. The thermochemical reactor consists on an AISI 304 metal tube of 265 15 mm of inner diameter and 400 mm length. Around 8 g of MO were placed in the middle of 266 the tube length. Silicon carbide particles were placed below and above the MO, filling 267 completely the reactor, preventing any movement or fluidization of the MO. Once the gas 268 stream crosses the reactor, it is cooled down in a water bath and subsequently, the oxygen 269 content is measured by means of a zirconia gas analyzer KCD-ON320 (Sensorstecnics). The 270 reactor includes five temperature measuring points at different heights: T2, T3 and T4 placed 271 inside the reactor at the positions indicated in Fig. 3a and T2out and T4out placed on the exterior 272 surface of the reactor at the same height of T2 and T4, respectively. The signals were collected 273 with an acquisition system type ABSD-MD832-81-23-HLP (Yunrunyn) and registered by a 274 homemade Labview app. The complete set-up is illustrated on Fig. 3b. 275 a) b) 276 277 Figure 3. Thermochemical experimental setup: a) components diagram and b) setup photography 278 279 280 281 3. Results and discussion 282 283 3.1 Effects of the synthesis parameters on the granules behaviour 284 285 A comprehensive study was made to assess the influence of the granule conformation, the 286 MO content and the hardening step on the chemical and mechanical behavior of Si-doped 287 In order to assess the mechanical capability of the different calcined granules to withstand the 448 mechanical stresses imposed by the lab-scale packed bed reactor set-up, a layer composed of 449 3 of the different set of granules was submitted to an increased weight through a flat plate 450 until failure. The crushing strength results are shown in Table 2 and shall be taken only as a 451 reference value for the particular configuration tested in the present work. To obtain an 452 accurate value for the crushing strength specific lab device should be used (e. g. standard 453 compression testing machine). 454 Table 2. Crushing strength of the granules calcined at different temperature programs 455 Sample id Hardening program Maximum weight Crushing strength MO 50 1050°C 4h 38 g 0,126 N MO 50 1050°C 8h 44 g 0,146 N MO 50 1100°C 4h 50 g 0,166 N 456 457 3.2.2 Selection of the granules for the packed bed reactor experiments 458 459 The reactor consisted on a stainless-steel tube of 13 mm of internal diameter and 90 mm 460 height, turning into an effective volume of 11,9 cm3. In order to determine the energy density 461 contained in this volume it is necessary to measure the actual metal oxide content in the 462 granules and the packed bed void fraction. The metal oxide content in the granules was 463 determined using a reference volume of 1 cm3. This volume was filled with granules of the 464 different MO:PB ratios used and the bulk density and true density were measured using a 465 helium pycnometer AccuPyc II 1340. It was observed that the maximum MO in the granules 466 reached only 16.94%, likely caused by a deficient dissolution of the OS in the precipitation 467 bath. In order to get an increase on the MO content another batch of granules was prepared 468 decreasing the temperature of the bath by using ice in order to reduce the solubility of the PB 469 (MO 50-c). The resulting granules got a significant MO content increase. The results are 470 showed in Table 3. 471 Table 3. Samples properties measurement over a reference volume of 1 cm3 472 Description Mass granule g Diameter mm Bulk density g/cm 3 True density g/cm 3 MO content % Powder N/A N/A 0,3345 5,3236 N/A MO 50 0,0023 1,835 0,735 5,116 14,38 MO 70 0,0055 2,342 0,828 5,0047 16,56 MO 80 0,0066 2,466 0,846 5,253 16,94 MO 50-c 0,002 1,396 1,402 5,653 24,81 473 Considering the size of the lab-scale reactor and the granules with the highest bulk 474 density, the maximum amount of material that can be introduced in the reactor without 475 applying pressure was a maximum of 8 g. Therefore, according to the data provided in Table 476 2, the three hardening programs provide to the granules with enough mechanical stability to 477 withstand the mechanical stress imposed by the weight of the packed bed. Consequently, the 478 granules with the highest MO content (MO 50-c) calcined at 1050°C for 4h, which in addition 479 present a more homogeneous sphericity, were selected and synthesized for further cycling 480 tests in the lab-scale reactor (Fig. 9). 481 482 483 Figure 9. Mn2O3 Si-doped granules 50:50 (MO:PB) for reactor testing: a) After step 2 of the synthesis 484 route and b) After hardening at 1050°C for 4 h. 485 486 3.2.3 Packed bed reactor experimental results 487 488 A summary of the main reactor parameters is presented in Table 4. 489 Table 4. Lab-scale reactor parameters 490 Description Value Reactor radius 13 mm Reactor length 90 mm Particle diameter (dp) 1,39 mm Reactor mass 7,96 g Bulk density 1,402 g/cm3 Void fraction (ɛ) 0,59 Material storage density 149 J/g (Bielsa et al., 2021) Storage capacity 0,33 Wh 491 In order to evaluate the chemical and mechanical stability of the granules in a packed bed 492 arrangement, the reactor was subjected to 50 redox cycles consisting in of: i) a heating step 493 up to 800°C at 20°C/min under 100 mL/min of N2, ii) an isothermal step at 800°C for 20 min to 494 ensure complete reduction of the material, since in a previous study, the reduction onset 495 temperature in N2 was identified at approximately 775°C (Bielsa et al., 2021) , and iii) a cooling 496 step down to 550°C with a cooling rate of 20°C/min under a gas flow of 10%/90% (N2/O2). In 497 the same work it was observed that in such conditions, oxidation takes place in less than 5 498 min, so complete material oxidation is expected during the cooling step, before starting the 499 following redox cycle. The temperature evolution at different points of the reactor and the 500 oxygen content in the gas stream flowing out of the reactor were recorded by the 501 instrumentation during the 50 redox cycles and are plotted in Fig. 10a. For better 502 understanding, the same parameters comprising only the 3rd and 4th cycles are plotted in Fig. 503 10b. 504 505 In regard of the reduction reaction, it is difficult to obtain useful information from the 506 temperature profiles of Fig. 10b, since it seems that reduction takes place when the furnace 507 reaches the temperature set point and reduces the power input, which is followed by an 508 abrupt temperature slowdown that makes it difficult to distinguish the effect of the 509 endothermic reduction reaction in the temperature profiles of the internal thermocouples. 510 The reaction occurrence can be ascertained from the oxygen variation measurement, since an 511 increase of the oxygen content in the gas stream reaching a peak of 5% was detected. 512 Nevertheless, the extent of the reaction cannot be evaluated, since integrating the area under 513 the curve, it results in a value reaching only 6% of the theoretical oxygen release expected, 514 considering a theoretical mass loss of 3%. In the case of the oxidation reaction, the internal 515 recorded temperatures (T2, T3 and T4) present a slightly lower cooling rate than the 516 temperatures recorded outside the reactor (T2out and T4out) at the beginning of the cooling 517 step. This fact can be caused by the exothermic nature of the oxidation reaction, even though 518 the starting and finishing point of the reaction cannot be distinguished. Regarding the oxygen 519 concentration, a slight variation during the oxidation reaction can be observed. The exact 520 oxygen decrease cannot be accurately determined as a consequence of the poor sensor 521 resolution. However, both evidences suggests that during the heating step reduction of the 522 material is taking place, while in the cooling step the material is being oxidized. 523 524 The morphological evolution and physical integrity of the granules after the 50 redox cycles in 525 the reactor was analyzed both by visual inspection (Fig. 10c) and SEM (Fig. 10d). None of the 526 granules presented the reddish color typical of the reduction phase and thus, showing 527 incomplete oxidation, maintaining their spherical shape after removing them from the 528 reactor. As can be observed from the SEM image, the particles have not suffered from 529 noticeable sintering, since their size remain similar to their initial sate after the hardening step 530 (Fig. 8a). Therefore, no reversibility loss would be expected. To confirm this statement three 531 granules of the already cycled material in the reactor were subjected to an additional 50 redox 532 cycles in the STA, using the same temperature program, completing a total program of 100 533 redox cycles. The mass loss/gain during the STA program together with a SEM image of the 534 particles after the 100 redox cycles are shown in Fig. 11a and Fig. 11b, confirming no loss of 535 the chemical stability of the granules with no noticeable change on the particle morphology. 536 537 a) b) 538 539 c) d) 540 541 542 543 Figure 10. Si-doped Mn2O3 granules redox cycling in the thermochemical reactor: a) thermocouples 544 signals record during the whole 50 redox program, b) magnification of the thermocouple signals and 545 oxygen sensor during the 3rd and 4th redox cycle, c) Si-doped granules together with SiC particles 546 appearance after cycling and removing from the reactor, and d) SEM image showing the 547 morphological evolution after the 50th redox cycle 548 549 a) b) 550 551 Figure 11. Si-doped Mn2O3 granules subjected to 50 additional redox cycles in the STA: a) mass 552 loss/gain recorded in the STA, and b) SEM picture after the additional 50 redox cycles 553 554 4. Conclusions 555 In this work, a granule preparation route was studied for thermochemical energy 556 storage upscaling using a novel Si-doped manganese oxide for concentrated solar power 557 plants. The research work comprises material preparation and thermochemical performance 558 evaluation in a lab-scale packed bed reactor. The process succeeded in obtaining spherical 559 porous granules of 1-2 mm with different active material content. The results identify the 560 critical parameters of the synthesis process which provide the best mechanical and chemical 561 stabilities of the granules in order to be used in a thermochemical packed bed reactor. It was 562 observed that the hardening step aimed to increase the mechanical stability of the granules 563 does not affect significantly their chemical stability. This fact confirms that the severity of the 564 hardening process could be further increased, leading to more mechanically stable granules, 565 required for larger scale reactors. Furthermore, we observed that decreasing the solubility of 566  -  10 µm  ---  -  10 µm  --- the polymeric binder in the synthesis bath almost double the active material content in the 567 granules and consequently, enhances the volumetric energy storage capacity. We achieved 568 an active material content of 24.81%, and this result suggests that there is a way to keep 569 increasing the energy storage density of the granules, which need to be further developed. In 570 addition, 8 g of Si-doped granules were subjected to 50 redox cycles in a lab-scale packed bed 571 reactor, showing satisfactory mechanical and chemical stability, which was confirmed over 572 additional 50 redox cycles in a thermobalance, with complete re-oxidation over the whole 573 program. In summary, both the Si-doped manganese oxide and the granule preparation route 574 conducted have demonstrated enough reliability to be used on a larger scale and thus, 575 contribute to push the high temperature thermochemical storage technology for a next 576 generation of concentrated solar power plants. 577 578 579 Acknowledgements 580 This work has been supported by the Department of Economic Development and 581 Infrastructures of the Basque government, through the funding of the ELKARTEK CIC 582 Energigune-2017 research program. The authors express their sincere thanks to Cristina 583 Luengo, Mikel Intxaurtieta and Yagmur Polat for their technical support. 584 585 Bibliography 586 Abián, M., Abad Secades, A., Izquierdo Pantoja, M.T., Gayán Sanz, P., Diego Poza, L.F., García 587 Labiano, F., Adánez Elorza, J., 2017. Titanium substituted manganese-ferrite as an 588 oxygen carrier with permanent magnetic properties for chemical looping combustion 589 of solid fuels. Fuel 195, 38–48. 590 Agrafiotis, C., Becker, A., Roeb, M., Sattler, C., 2016. Exploitation of thermochemical cycles 591 based on solid oxide redox systems for thermochemical storage of solar heat. Part 5: 592 Testing of porous ceramic honeycomb and foam cascades based on cobalt and 593 manganese oxides for hybrid sensible/thermochemical heat storage. Solar Energy 139, 594 676–694. 595 Agrafiotis, C., Roeb, M., Sattler, C., 2015a. Hybrid sensible/thermochemical solar energy 596 storage concepts based on porous ceramic structures and redox pair oxides chemistry. 597 Energy Procedia 69, 706–715. 598 Agrafiotis, C., Roeb, M., Schmuecker, M., Sattler, C., 2015b. Exploitation of thermochemical 599 cycles based on solid oxide redox systems for thermochemical storage of solar heat. 600 Part 2: Redox oxide-coated porous ceramic structures as integrated thermochemical 601 reactors/heat exchangers. Solar Energy 114, 440–458. 602 Azimi, G., Leion, H., Matisson, T., Rydén, M., Snijkers, F.M.M., Lyngfelt, A., 2014. Mn–Fe 603 Oxides with Support of MgAl2O4, CeO2, ZrO2 and Y2O3–ZrO2 for Chemical-Looping 604 Combustion and Chemical-Looping with Oxygen Uncoupling. Ind. Eng. Chem. Res 53, 605 10358–10365. 606 Azimi, G., Matisson, T., Leion, H., Rydén, M., Lyngfelt, F.M.M., 2015. Comprehensive study of 607 Mn–Fe–Al oxygen-carriers forchemical-looping with oxygen uncoupling (CLOU). 608 International Journal of Greenhouse Gas Control 34, 12–24. 609 Bielsa, D., Zaki, A., Arias, P.L., Faik, A., 2021. Development of a kinetic reaction model for 610 reduction and oxidation of Si doped Mn2O3 for thermochemical energy storage in 611 concentrated solar power plants. Journal of Energy Storage 43, 103271. 612 Bielsa, D., Zaki, A., Arias, P.L., Faik, A., 2020. Improving the redox performance of 613 Mn2O3/Mn3O4 pair by Si doping to be used as thermochemical energy storage for 614 concentrated solar power plants. Solar Energy 204, 144–154. 615 Bielsa, D., Zaki, A., Faik, A., Arias, P.L., 2019. Efficiency improvement of Mn2O3/Mn3O4 616 redox reaction by means of different operation strategies. 617 https://doi.org/10.1063/1.5117750 618 Carrillo, A.J., Serrano, D., Pizarro, P., Coronado, J.M., 2014. Thermochemical heat storage 619 based on the Mn2O3/Mn3O4 redox couple: influence of the initial particle size on the 620 morphological evolution and cyclability. J. Mater. Chem. A 2, 19435–19443. 621 Carrillo, A.J., Serrano, D.P., Pizarro, P., Coronado, J.M., 2015. Improving the Thermochemical 622 Energy Storage Performance of the Mn2O3/Mn3O4 Redox Couple by the Incorporation 623 of Iron. ChemSusChem 8, 1947–1954. 624 Gigantino, M., Sas Brunser, S., Steinfeld, A., 2020. High-Temperature Thermochemical Heat 625 Storage via the CuO/Cu2O Redox Cycle: From Material Synthesis to Packed-Bed Reactor 626 Engineering and Cyclic Operation. Energy & Fuels 34, 16772–16782. 627 Hamidi, M., Bayon, A., Wheeler, V.M., Kreider, P., Wallace, M.A., Tsuzuki, T., Catchpole, K., 628 Weimer, A.W., 2019. Reduction kinetics for large spherical 2:1 iron–manganese oxide 629 redox materials for thermochemical energy storage. Chemical Engineering Science 201, 630 74–81. 631 Ibraheam, A.S., Ehrhart, B.D., Ward, B.J., Bayon, A., Wallace, M.A., Bader, R., Kreider, P., 632 Weimer, A.W., 2019. Particle design and oxidation kinetics of iron-manganese oxide 633 redox materials for thermochemical energy storage. Solar Energy 183, 17–29. 634 Karagiannakis, G., Pagkoura, C., Halevas, E., Baltzopoulou, P., Konstandopoulos, A.G., 2016. 635 Cobalt/cobaltous oxide based honeycombs for thermochemical heat storage in future 636 concentrated solar power installations: Multi-cyclic assessment and semi-quantitative 637 heat effects estimations. Solar Energy 133, 394–407. 638 Mederos, F.S., Ancheyta, J., Chen, J., 2009. Review on criteria to ensure ideal behaviors in 639 trickle-bed reactors. Applied Catalysis A: General 355, 1–19. 640 Neises, M., Tescari, S., de Oliveira, L., Roeb, M., Sattler, C., Wong, B.Y., 2012. Solar-heated 641 rotary kiln for thermochemical energy storage. Solar Energy 86, 3040–3048. 642 Neumann, N.C., Block, T., Linder, M., Leion, H., 2018. Stabilizing particles of manganese-iron 643 oxide with additives for thermochemical energy storage. Energy Technology 6. 644 Pagkoura, C., Karagiannakis, G., Zygogianni, A., Lorentzou, S., Konstandopoulos, A.G., 2015. 645 Cobalt oxide based honeycombs as reactors/heat exchangers for redox 646 Thermochemical Heat Storage in Future CSP Plants. Energy Procedia 69, 978–987. 647 Pelay, U., Luo, L., Fan, Y., Stitou, D., Castelain, C., 2019. Integration of a thermochemical 648 energy storage system in a Rankine cycle driven by concentrating solar power: Energy 649 and exergy analyses. Energy 167, 498–510. 650 Pestalozzi, F., 2013. Experimental Investigation of Mn-Based Redox Reaction Kinetics for 651 Thermochemical Energy Storage. 652 Schmidt, M., Linder, M., 2017. Power generation based on the Ca(OH)2/ CaO 653 thermochemical storage system – Experimental investigation of discharge operation 654 modes in lab scale and corresponding conceptual process design. Applied Energy 203, 655 594–607. 656 Schoenfelder, C.W., Swisher, J.H., 1973. Kinetics of Thermal Decomposition of TiH2. Journal 657 of Vacuum Science and Technology 10. https://doi.org/10.1116/1.1318443 658 Singh, A., Tescari, S., Lantin, G., Agrafiotis, C., Roeb, M., Sattler, C., 2017. Solar 659 thermochemical heat storage via the Co3O4/CoO looping cycle: Storage reactor 660 modelling and experimental validation. Solar Energy 144, 453–465. 661 Ströhle, S., Haselbacher, A., Jovanovic, Z.R., Steinfeld, A., 2016. The Effect of the Gas-Solid 662 Contacting Pattern in a High-Temperature Thermochemical Energy Storage on the 663 Performance of a Concentrated Solar Power Plant. Energy & Environmental Science 9, 664 1375–1389. 665 Varsano, F., Alvani, C., la Barbera, A., Masi, A., Padella, F., 2016. Lithium manganese oxides 666 as high-temperature thermal energy storage system. Thermochimica Acta 640. 667 https://doi.org/10.1016/j.tca.2016.07.018 668 Wokon, M., Block, T., Nicolai, S., Linder, M., Schmücker, M., 2017a. Thermodynamic and 669 kinetic investigation of a technical grade manganese-iron binary oxide for 670 thermochemical energy storage. Solar Energy 153, 471–485. 671 Wokon, M., Kohzer, A., Linder, M., 2017b. Investigations on thermochemical energy storage 672 based on technicalgrade manganese-iron oxide in a lab-scale packed bed reactor. Solar 673 Energy 153, 200–214. 674 Wong, B., 2011. THERMOCHEMICAL HEAT STORAGE FOR CONCENTRATED SOLAR POWER. 675 Golden, CO (United States). https://doi.org/10.2172/1039304 676 Zsembinszki, G., Sole, A., Barreneche, C., Prieto, C., Fernandez, A.I., Cabeza, L.F., 2018. 677 Review of Reactors with Potential Use in Thermochemical Energy Storage in 678 Concentrated Solar Power Plants. Energies 11, 2358. 679 680