Review of Reactors with Potential Use in Thermochemical Energy Storage in Concentrated Solar Power Plants

Gabriel Zsembinszki 1 , Aran Solé 2 , Camila Barreneche 3,4 , Cristina Prieto 5 , A. Inés Fernández 3 and Luisa F. Cabeza 1,* 1 GREiA Research Group, INSPIRES Research Centre, University of Lleida, Pere de Cabrera s/n, 25001 Lleida, Spain; gabriel.zsembinszki@udl.cat 2 Department of Mechanical Engineering and Construction, Universitat Jaume I, Campus del Riu Sec s/n, 12071 Castelló de la Plana, Spain; sole@uji.es 3 Department of Materials Science and Physical-Chemistry, Universitat de Barcelona, Martí i Franqués 1, 08028 Barcelona, Spain; c.barreneche@ub.edu (C.B.); ana.inesfernandez@ub.edu (A.I.F.) 4 BCES (Birmingham Centre of Energy Storage), School of Chemical Engineering, University of Birmingham, Birmingham B15 2TT, UK 5 Abengoa, C/Energía Solar 1, 41012-Seville, Spain; cristina.prieto@abengoa.com * Correspondence: lcabeza@diei.udl.cat; Tel.: +34-973-003576


Introduction
The production of electricity from concentrated solar power plants, which are known as CSP plants, is still a maturing technology, although it has been in operation commercially at the utility scale since 1985 [1].Nevertheless, it has experienced a worldwide large increase in deployment in recent years [2], and there are plans to expand its application in industrial processes and the generation of solar fuels.According to the Renewable Energy Policy Network for the 21st Century (REN21) [3], total installed capacity in CSP in 2016 was 4.81 GW, growing from 600 MW in 2009.The deployment was mostly in Spain (with 2300 MW) and the United States (USA) (with 1738 MW).Other countries with plants are India, Morocco, South Africa, the United Arab Emirates, Algeria, Egypt, Australia, China, and Thailand.Today, parabolic trough technology dominates the CSP market, both in number of projects and total installed capacity (around 85% of capacity) [4].
Energies 2018, 11,2358 3 of 23 to the use of phase change materials (PCM, the materials used in latent heat storage) in CSP plants, this technology is still not installed in commercial plants.
Finally, thermochemical energy storage is receiving increasing attention among researchers.Prieto et al. [15] presented a review of chemical reaction processes for CSP from the technological point of view.The authors presented the different cycles that have been studied by other researchers, such as sulfur base cycles and metal oxide redox cycles (calcium carbonate cycle, calcium hydroxide cycle, and manganese oxide cycle), and perovskite-form structures.Previously, reaction candidates for medium or high-temperature applications (250-800 • C) were listed by Felderhoff et al. [26].The listed candidates included metal hydrides, metal hydroxides, and metal carbonates.In 2014, Pardo et al. [27] published a state-of-the-art of the thermochemical heat storage solutions, focusing on temperatures comprised between 573-1273 K, and not focused only on CSP applications, but more generally on high temperature.The materials listed are metal hydrides, carbonate systems, hydroxide systems, redox systems, ammonia systems, and organic systems.In their conclusions, these authors claimed that thermochemical energy storage appears to be the most promising TES system for solar plants during long periods, because both the storage period and transport distance are theoretically unlimited, since there is no loss of heat during storage (storage happens at ambient temperature).
Although some of these reviews present and comment on the different configurations and prototypes used for each material or reaction, none of them has provided a detailed assessment of the reactor itself.Therefore, the aim of this new review is to fill this gap found in the literature, evaluating the reactor concepts, and highlighting their advantages and disadvantages.

Materials Used in Thermochemical Reactions
Currently, there are different processes using solar energy, from CSP plant technology to performing a chemical reaction [28][29][30][31][32].All of them have in common the objective of reducing fossil fuel consumption, and therefore reducing CO 2 emissions.The different processes that require solar reactors and also working temperatures above 1000 • C are two-step thermochemical H 2 O/CO 2 splitting, the gasification/cracking of carbonaceous feedstock, thermochemical heat storage, and the recycling/procurement of raw materials.
The two-step thermochemical H 2 O/CO 2 splitting [33,34] is based on a two-redox reaction cycle for CO and H 2 production, which is considered as solar fuel.This gas is used as fuel for fuel cells or feedstock for in Fischer-Tropsch reactions.In this case, in order to split CO 2 and H 2 O molecules, a metal oxide is used to perform a thermal reduction of metal oxide, obtaining a metal (or metal oxide) (Equation ( 1)) that is capable of reducing (split) CO 2 or H 2 O (Equation ( 2)).Depending on the metal oxide, the working temperature or the system pressure must be adjusted [35].In addition, in Equation ( 1), the key is the solid product of the reaction, while in Equation (2), the important step is the gas product.These key issues will be the points affecting the reactor design.
The processes gasification/cracking of carbonaceous feedstock [29,36] are based on biogas cracking (Equation (3)) or carbonaceous feedstock gasification (Equation ( 4)) to obtain solar fuel, for fuel cell or for a Fischer-Tropsch reaction.When necessary, these reactions can be performed under catalytic reaction to accelerate kinetics.A system for collecting the gases will be significant for the reactor design in this type of process: Thermochemical energy storage [26,27] is designed to store heat through a reversible reaction (Equation ( 5)).A heat exchanger is required for this process in order to revert the solar heat that has been stored.However, although this technology/process has still not been neither extensively studied nor implemented, its definition can be found in the literature [37].
Finally, there is the recycling/procurement of raw materials.As it is well known, many processes for obtaining raw materials, such as Zn or CaO, require high-energy inputs to perform a carbothermal reduction.In this case, several materials can sustain thermal reductions by using concentrated solar power.
As mentioned above, all of the processes prioritize some design aspects (heat exchanger, product collection, versatility, etc.).However, all of them can be classified according the reaction type, heating system, or limiting step of process (reaction, diffusion, etc.).

Classification of Reactors
There are several ways to classify the solar reactors depending on the focus.Classifications based on the reactor or on the system can be found in the literature [28,38].Therefore, a new classification is suggested based on the process by itself, which is at the end what will give us the limiting step.The limiting step is essential to properly designing the reactor and the system containing the future TCM (thermochemical material) selected reaction, which will be implemented in solar power tower (SPT) technology.

Classification Based in the Reactor
Figure 1 shows the classification of the reactors that has been taken into account in Table 1.Within stacked beds (modular ones), fixed beds are recommended for solar catalytic reactions.For those reactions that require good thermal transfer properties, Villermaux [38] suggested the employment of fluidized beds.Cyclones are interesting if the further separation between solid and gas is desired, but their use is similar in terms of the ad/disadvantages of fluidized bed reactors (Table 1).been stored.However, although this technology/process has still not been neither extensively studied nor implemented, its definition can be found in the literature [37].
Finally, there is the recycling/procurement of raw materials.As it is well known, many processes for obtaining raw materials, such as Zn or CaO, require high-energy inputs to perform a carbothermal reduction.In this case, several materials can sustain thermal reductions by using concentrated solar power.
As mentioned above, all of the processes prioritize some design aspects (heat exchanger, product collection, versatility, etc.).However, all of them can be classified according the reaction type, heating system, or limiting step of process (reaction, diffusion, etc.).

Classification of Reactors
There are several ways to classify the solar reactors depending on the focus.Classifications based on the reactor or on the system can be found in the literature [28,38].Therefore, a new classification is suggested based on the process by itself, which is at the end what will give us the limiting step.The limiting step is essential to properly designing the reactor and the system containing the future TCM (thermochemical material) selected reaction, which will be implemented in solar power tower (SPT) technology.

Classification Based in the Reactor
Figure 1 shows the classification of the reactors that has been taken into account in Table 1.Within stacked beds (modular ones), fixed beds are recommended for solar catalytic reactions.For those reactions that require good thermal transfer properties, Villermaux [38] suggested the employment of fluidized beds.Cyclones are interesting if the further separation between solid and gas is desired, but their use is similar in terms of the ad/disadvantages of fluidized bed reactors (Table 1).Following a flow pattern classification, three main gas-solid technologies can be found, including fixed, moving and fluidized bed [39].Their main advantages and disadvantages are shown in Table 1.In fixed bed reactors (also called packed bed reactors), the solid particles are located inside a vessel with the flux of reactants flowing through the stationary bed.When the diameter of the packed bed increases, heat transfer rates are poor; therefore, high heat transfer rates should be considered when fluidized beds are required [40].In mobile/moving bed reactors, the bed can be moved continuously or periodically with a fluid circulation similar to the prior one.In addition, in fluidized bed reactors, the solid has a very small diameter and is maintained in suspension.Following a flow pattern classification, three main gas-solid technologies can be found, including fixed, moving and fluidized bed [39].Their main advantages and disadvantages are shown in Table 1.In fixed bed reactors (also called packed bed reactors), the solid particles are located inside a vessel with the flux of reactants flowing through the stationary bed.When the diameter of the packed bed increases, heat transfer rates are poor; therefore, high heat transfer rates should be considered when fluidized beds are required [40].In mobile/moving bed reactors, the bed can be moved continuously or periodically with a fluid circulation similar to the prior one.In addition, in fluidized bed reactors, the solid has a very small diameter and is maintained in suspension.

Reactors
Table 1.Comparison between fixed, moving, and fluidized bed reactors for solar applications [41].

Classification Based in the System
As mentioned in Table 1, the implementation of the reactors in the solar plant is the main problem for the system configuration.A possible classification based on the irradiation and the solar receiver can be established according to the heat integration mode into the reaction chamber.This is related with the maximum system efficiency (solar-to-chemical).Furthermore, solar reactors are indirect or direct, as shown in Figure 2.
Table 1.Comparison between fixed, moving, and fluidized bed reactors for solar applications [41].

Reactor Advantages Disadvantages
Fixed/packed bed Low cost Non-parasitic Easier for modeling Low heat and mass transfer High-pressure drop Difficulties to be implemented in solar receiver cavities at continuous commercial process Stacked bed needed, with higher complexity in solar-focused strategies Non-uniform irradiance distributions on the particle volume

Classification Based in the System
As mentioned in Table 1, the implementation of the reactors in the solar plant is the main problem for the system configuration.A possible classification based on the irradiation and the solar receiver can be established according to the heat integration mode into the reaction chamber.This is related with the maximum system efficiency (solar-to-chemical).Furthermore, solar reactors are indirect or direct, as shown in Figure 2. Indirect reactors are those where a container material is directly irradiated to heat the solid or the heat transfer fluid.This configuration is mostly used for fluidized beds, where the solid can be forced upward through irradiated tubes by an airflow, which fluidizes the particles and increases the heat transfer from the tube walls to the flowing particles [28,42].
On the other hand, in direct reactors, reactants are directly irradiated and heated by a solar radiation input.Some of them have a receiver aperture that can be opened (open reactor) or closed by a transparent window (mainly quartz-closed reactors).The quartz window is mandatory in a direct system at high temperature in order to reduce the heat and mass losses.The design of a suitable window for high-temperature processes (T > 900 • C) is still a bottleneck in the design of solar receivers.
Therefore, for direct reactors (both configurations: open and closed), it is important to consider the absorptivity/emissivity of the material, since the more radiation that is absorbed, the higher the reactor efficiency and conversion, because the materials are basically heated by radiation, although they are also heated by convection and conduction.Higher temperatures are expected within direct reactors when compared to indirect reactors, but it is difficult to get homogeneous temperatures inside the reaction chamber.
On the other hand, an indirect system is defined as a system that uses concentrated solar power to heat a black body, and then transfers the heat to the reaction cavity by conduction and convection.Therefore, an indirect reactor requires a thermal conductor between the receiver and reactor to heat the reaction cavity [43].However, such systems minimize the thermal shock (the temperature gradients are lower than in direct reactors) by maintaining a uniform temperature inside the reaction chamber.
Energy efficiency is directly affected by this choice, since when transferring sun energy to a material, there is always some losses, which at the end is translated to less system efficiency.
The indirect system implies one more resistance than the direct system.

Classification Based in the Process Limiting Step
The classification shown in Figure 3 is suggested by the authors.From the point of view of our knowledge, it is very important to distinguish between a reaction with two phases and a decomposition of a solid due to the effect of heat.This classification enables us to make a relation with the limiting step of the process inside the reactor.Moreover, the limiting step (chemical or diffusive controlled) must be identified in order to start the reactor design.
Indirect reactors are those where a container material is directly irradiated to heat the solid or the heat transfer fluid.This configuration is mostly used for fluidized beds, where the solid can be forced upward through irradiated tubes by an airflow, which fluidizes the particles and increases the heat transfer from the tube walls to the flowing particles [28,42].
On the other hand, in direct reactors, reactants are directly irradiated and heated by a solar radiation input.Some of them have a receiver aperture that can be opened (open reactor) or closed by a transparent window (mainly quartz-closed reactors).The quartz window is mandatory in a direct system at high temperature in order to reduce the heat and mass losses.The design of a suitable window for high-temperature processes (T > 900 °C) is still a bottleneck in the design of solar receivers.
Therefore, for direct reactors (both configurations: open and closed), it is important to consider the absorptivity/emissivity of the material, since the more radiation that is absorbed, the higher the reactor efficiency and conversion, because the materials are basically heated by radiation, although they are also heated by convection and conduction.Higher temperatures are expected within direct reactors when compared to indirect reactors, but it is difficult to get homogeneous temperatures inside the reaction chamber.
On the other hand, an indirect system is defined as a system that uses concentrated solar power to heat a black body, and then transfers the heat to the reaction cavity by conduction and convection.Therefore, an indirect reactor requires a thermal conductor between the receiver and reactor to heat the reaction cavity [43].However, such systems minimize the thermal shock (the temperature gradients are lower than in direct reactors) by maintaining a uniform temperature inside the reaction chamber.
Energy efficiency is directly affected by this choice, since when transferring sun energy to a material, there is always some losses, which at the end is translated to less system efficiency.The indirect system implies one more resistance than the direct system.

Classification Based in the Process Limiting Step
The classification shown in Figure 3 is suggested by the authors.From the point of view of our knowledge, it is very important to distinguish between a reaction with two phases and a decomposition of a solid due to the effect of heat.This classification enables us to make a relation with the limiting step of the process inside the reactor.Moreover, the limiting step (chemical or diffusive controlled) must be identified in order to start the reactor design.The diffusion control step can be separated in two diffusion processes, which occur in solid-gas reactions: one from the gas to the solid surface (external diffusion), and a second from the outer surface of the solid particle to the inner part of the particle (internal diffusion) [37,38].
Another limiting step could be the chemical reaction, and thus kinetics itself.There are several criteria to identify the influence of diffusion and temperature difference within the particle on the reaction rate.This should be studied for each specific reaction in order to differentiate the processes controlled by chemical reaction (thermal decomposition) and processes to be studied as its limiting step (solid-gas reaction).

Solid-gas reaction
Thermal decomposition The diffusion control step can be separated in two diffusion processes, which occur in solid-gas reactions: one from the gas to the solid surface (external diffusion), and a second from the outer surface of the solid particle to the inner part of the particle (internal diffusion) [37,38].
Another limiting step could be the chemical reaction, and thus kinetics itself.There are several criteria to identify the influence of diffusion and temperature difference within the particle on the reaction rate.This should be studied for each specific reaction in order to differentiate the processes controlled by chemical reaction (thermal decomposition) and processes to be studied as its limiting step (solid-gas reaction).

Reactors Used in Solar Plants
Table 2 is a summary of all of the available experimental reactors that have been published until September 2017, of which most work at temperatures above 1000 • C (1273 K).
Sixty-eight experimental reactors have been found in the literature working at high temperatures, including 48 of them working above 1000 • C. Both thermal decomposition and solid-gas chemical reaction processes are carried out in these reactors, with the chemical reaction type the being the dominant one.Most of the reactors (81%) are part of a direct and closed system, followed by indirect systems (15%); direct and open systems only make up 3%.Then, when looking at the highest-level group of system configurations (direct and closed), the main observations are:

•
All of the available and described reactor types are presented: entrained, stacked, and fluidized.

•
The main available reactors that have been tested at the lab scale are stacked/fixed bed, 50%.

•
Fluidized beds are the second most reported option, with around 21%.

•
Stacked/rotary reactors are in third position, with 15% of the implementation within this group.

•
Cyclone reactors have been implemented in all of the possible system configurations: direct/open, direct/closed, and indirect.The most available ones are in active/direct/closed system configuration.
In general, and as a summary of all of the reactor configurations listed in Table 2, the statistics of different combinations of system, reactors, and processes are presented in Table 3.
The rotary reactors, which are within the stacked bed reactors, are mainly found in the direct and closed system group, with half of them using decomposition processes, and the other half using chemical reactions.Those reactors account for around 24% of the total decomposition processes.The solar efficiencies of these decomposition processes are between 12-88%, and chemical conversion varies between 30-95% in these reactors.Rotary reactors reduce radiation losses with constant temperature in the internal wall.Moreover, the movement of the particles increases the heat transfer [44].
Fluidized beds are used in reactors in direct/closed systems, with most of them using solid-gas reactions (13 out of 14).They represent 28% of the total number of reactors used for solid-gas reactions.Chemical conversions vary from less than 10% to 100%.Data regarding solar efficiency is only available for three of the reactors, with values ranging between 10-15%.
Fixed bed reactors, which are the most frequent type of reactors, have been used for both processes; almost 32% for thermal decomposition and 68% for solid-gas reactions.They represent 49% of the reactors used for solid-gas reactions.Chemical conversions vary from 25% to 100% for thermal decomposition and from less than 15% to 85% for solid-gas reactions (note that there is a reactor presenting a 2% chemical conversion, which uses a foam matrix, and has been discarded in this comparison [28]).Furthermore, very little data is available concerning solar efficiency, especially for direct, closed fixed bed reactors, due to the difficulties of scaling up this technology in commercial solar applications.

Conclusions
The analysis of the available literature shows that there will not be a single reactor configuration to be applied for solar thermochemical process.The diversity of the detailed problems shows that the technology is still in low technology readiness level (TRL)-less than five-and significant research and development are still demanded.The design and optimization of the reactor to be used must be defined by the chemistry of the reaction, the dominant transfer mechanisms, and the compatibility of the materials.Moreover, all of this must be designed with ambitious solar efficiency objectives in order to minimize the cost of the solar field, which continues to be one of the limitations in the cost savings of CSP.
Reactors implemented in direct systems are by far the most studied ones.A very high percentage of both solid-gas reactions and thermal decompositions, around 81% of the total systems analyzed, consists of a direct and closed system configuration.Within this configuration, the majority of all of the processes-around 69%-take place in stacked reactors, and within this reactor type, fixed bed reactors are preferred (68%).The direct configuration has a significant lower cost than an indirect system.Additionally, the closed configuration is mandatory at high temperature to increase the solar efficiency in the reactor receiver.
Fixed beds are being used to undergo both processes; they are more used for chemical reactions and show better chemical conversion in thermal decomposition processes.However, their use in solar applications is at a very preliminary stage of application.The control of the temperature in the irradiated solid, the uniformity of the flux map, the effect of the diffusion mechanism, and the scalability of the processes makes their applicability difficult.
Fluidized beds are preferred for solid-gas reactions, and are known to increase two phases (gas and solid) contact and thus heat and mass transfer.Therefore, they are recommended for solid-gas reactions, whose limiting step is diffusion, which is either external or internal.Nevertheless, complexity in design, increase in the self-consumption, and the erosion/wear of internal components are big drawbacks to overcome.
Solar efficiencies up to now have been very low.Those values cannot be compared, since very little data is found in the published papers.The state-of-the-art showed 50% as the maximum solid receiver efficiency achieved so far; however, a value higher than 80% is required to make this solar system economically feasible.
Chemical conversions are acceptable in some specific cases; these are mainly fluidized bed reactors for solid-gas chemical reactions, and fixed bed reactors and rotary reactors for thermal decompositions.
Most of the problems found when working experimentally with these reactors are due to material resistance and low chemical conversions due to an insufficient radiation power or bad use of it.Important operational and design parameters when working with a CSP reactor are the geometry of the reactor, which needs to be optimized, the preheating of the inlet gas, the reactor configuration (continuous or batch), and re-radiation losses.

Figure 1 .
Figure 1.Classification according to the reactor type.

Figure 1 .
Figure 1.Classification according to the reactor type.
mass transfer High-pressure drop Difficulties to be implemented in solar receiver cavities at continuous commercial process Stacked bed needed, with higher complexity in solar-focused strategies Non-uniform irradiance distributions on the particle volume Mobile/moving bed Direct heat transfer between solids and the gas Increase of the heat transfer coefficient Difficulties to be implemented in solar reactors Non-uniform irradiance distributions on the particle flow Need to control the residence time to increase the heat transference in particle receiver Complex hydrodynamics Rotary High chemical conversion due to high heat and mass transfer Versatility Long life components Difficulties of scalability to lager commercial solar systems Increase of parasitic energy consumption due to the movement of the reactor Higher risk of mechanical maintenance cost due to the use of a rotatory element at high temperature Fluidized bed Minimization of the risk of hotspots and thermal instability Heat transfer coefficients are high Difficulties to be implemented in solar reactors Need of gas for fluidization Increase of parasitic energy consumption due to the need of fluidized gas Erosion of internal components Complex reactor hydrodynamics and modeling between solids and the gas Increase of the heat transfer coefficient Difficulties to be implemented in solar reactors Non-uniform irradiance distributions on the particle flow Need to control the residence time to increase the heat transference in particle receiver Complex hydrodynamics Rotary High chemical conversion due to high heat and mass transfer Versatility Long life components Difficulties of scalability to lager commercial solar systems Increase of parasitic energy consumption due to the movement of the reactor Higher risk of mechanical maintenance cost due to the use of a rotatory element at high temperature Fluidized bed Minimization of the risk of hotspots and thermal instability Heat transfer coefficients are high Difficulties to be implemented in solar reactors Need of gas for fluidization Increase of parasitic energy consumption due to the need of fluidized gas Erosion of internal components Complex reactor hydrodynamics and modeling

Figure 3 .
Figure 3. Classification according to the thermochemical material (TCM) process.

Figure 3 .
Figure 3. Classification according to the thermochemical material (TCM) process.

Table 2 .
Summary of reactor configurations.

Table 3 .
Statistics of systems, reactors, and processes undergone.

Table 3 .
Statistics of systems, reactors, and processes undergone.

Table 3 .
Statistics of systems, reactors, and processes undergone.

Table 3 .
Statistics of systems, reactors, and processes undergone.

Table 3 .
Statistics of systems, reactors, and processes undergone.

Table 3 .
Statistics of systems, reactors, and processes undergone.

Table 3 .
Statistics of systems, reactors, and processes undergone.