Directly Irradiated Fluidized Bed Reactor for Thermochemical Energy Storage and CO2/H2O Splitting

Authors: 
Tregambi, C., Università degli Studi di Napoli Federico II
Bevilacqua, C., Istituto di Ricerche sulla Combustione, Consiglio Nazionale delle Ricerche
Di Benedetto, A., Università di Napoli Federico II
Solimene, R., Consiglio Nazionale delle Ricerche
Salatino, P., Università degli Studi di Napoli Federico II
Landi, G., Istituto di Ricerche sulla Combustione, Consiglio Nazionale delle Ricerche
Montagnaro, F., Università degli Studi di Napoli Federico II
Extensive research and developments activities are in progress to exploit the huge amount of solar energy falling on Earth. Concentrating Solar Power (CSP) is a fast-growing renewable technology in which the solar energy, upon concentration, is focused onto a receiver whence it is converted to electricity or industrial process heat. One of the great advantages of CSP is the relatively easy integration with thermal energy storage (TES) systems. TES systems allow to decouple the two steps of solar energy absorption and exploitation, enabling dispatchability of solar energy. In this way, it is possible to overcome the intrinsic intermittent behaviour of the solar source, providing on‑demand power delivery even during the night or in not favourable weather conditions. TES systems can rely on sensible heat storage, latent heat storage and thermochemical energy storage (TCES). In TCES the concentrated solar energy is used to perform reversible chemical reactions characterized by high latent heat. During the endothermic step, the solar energy is collected, and the reaction products are separated and stored. Then, when energy is required, the products are put back into contact to release the stored energy. TCES is a quite complex technology, but rewards with high density of energy storage and improved stability over long time-scales, as the incident energy is stored in the noble form of chemical bonds. An ambitious strategy for TCES aims to convert the harmless and inexpensive H2O and CO2 molecules into H2 and CO in two‑step reaction cycles by using metal oxides systems, which act as catalysts at relatively high temperatures (1000–1500 °C). In a first endothermic step, the metal oxide is thermally reduced at high temperature, with the consequent release of O2. The reduced metal oxide is then oxidized back with H2O/CO2 to eventually produce H2/CO and close the cycle. In order to make this process economically feasible and compete with current available technologies, two challenges need to be addressed. A first one – material oriented – concerns the development of a metal oxide owning high reversibility and cyclability, fast reaction kinetics and a not too high temperature for the reduction step. The second one – reactor oriented – aims at developing a feasible and reliable reactor configuration to maximize the conversion of solar energy into solar fuels. Among the investigated materials, cerium oxide is one of the most interesting one, as it is characterized by a high reversibility and fast re‑oxidation kinetics. However, its conversion degree for each reaction cycle is low, and very high reaction temperatures (~ 1500 °C) are required to fulfil the endothermic reduction step. Since last years, mixed perovskites materials are widely investigated for H2/CO production as they feature significant redox properties, structure stability and a wide range of formulations.

This work deals with the development of a directly irradiated lab‑scale fluidized bed reactor targeted at maximizing the collection of solar energy, withstanding the high‑concentrated flux typical of high‑temperature CSP applications (> 1 MW/m2) and ensuring a uniform temperature distribution of the reactive material. The reactor is made of two concentric circular columns: the internal one (OD=12 mm, ID=10 mm, length=80 mm) is referred as riser, the external one (OD=21 mm, ID=16 mm, length=70 mm) as annulus. The upper part of the annulus column is connected to a conical section (internal angle=30°, height=120 mm), which represents the freeboard of the reactor and hosts, at its extremity, a transparent quartz window to let the solar radiation enter keeping seal-tight operation of the reactor. During the reactor operation, a gas stream is fed into the internal column to induce the rise of a dense particle suspension, typically under fast fluidization conditions. At the outlet of the riser, the dense particle suspension directly interacts with the solar concentrated radiation (simulated by mean of a short‑arc Xe lamp) and increase its temperature. In the freeboard zone, due to the cross-section increase, the gas velocity is progressively reduced, and the solid particles fall down on the conical section and descend along it to eventually convey into the annulus column. Here, the particles move by gravity towards the bottom, where they re-enter the riser by mean of four small holes drilled in the riser column. During their motion along the annulus section, the “hot” particles transfer their heat to the riser column, thus preheating the rising dense suspension prior to its interaction with the concentrated solar radiation. This feature allows an autothermal operation of the reactor, enabling the possibility of reaching higher reaction temperatures. Gas exit is provided at the top of the conical section, just below the transparent window, by two small tubes (OD=6 mm, ID= 4 mm). The total inventory of the reactor is in the order of 15–25 g. As mentioned above, the solar concentrated radiation is simulated by a short‑arc Xe lamp of 7 kWel coupled with an elliptical reflector. The use of Xe lamps to simulate the solar spectrum is well established in the literature since the differences with the solar spectrum are quite small, especially in the visible spectral range. At full power, a peak flux of 2 MW/m2 is obtained in the focal point, while the total thermal power supplied to the system is of about 1.8 kWth. Temperature measurements are made by two K-type thermocouples: one at the bottom of the riser (Tdown), the other at the top of the annulus (Tup). The former measure the temperature of the “cold” material before its rise, the latter measure the temperature of the “hot” material before it descends along the annulus. It is worth noting that Tup does not represent the higher temperature reached by the material, as the highest temperature is probably experienced by the particles during their upward flight in the conical freeboard section. Unfortunately, this temperature cannot be currently measured in an affordable way. Exhaust gas are sent to a mass spectrometer to analyze the concentration of the relevant species, namely Ar, CO, CO2, N2, O2 for the performed experiments.

The experimental campaign on the fluidized bed reactor consisted in: i) thermal and hydrodynamic characterization under inert conditions; ii) characterization of a thermochemical energy storage process; iii) characterization of a CO2 splitting process.

The thermal and hydrodynamic characterization of the fluidized bed reactor was performed by means of different inert granular material, characterized by different density, colour and particle dimensions. Silica sand, silicon carbide and bauxite with a mean Sauter diameter in the 150–300 mm were used as bed materials. Steady state temperature of 900–1100 °C were reached in the system depending on material and fluidization conditions. Some of the materials exhibited a tendency to agglomerate and sinter during the reactor operation and were then rejected for the subsequent tests. This phenomenon was probably emphasized by the presence of small quantities of impurities which promoted the processes of agglomeration and sintering.

Concerning the thermochemical energy storage process, carbonation and calcination of a Ca‑based sorbent was achieved. More into detail, limestone particles were used as bed materials, and several iterated calcination and carbonation reaction cycles were successfully performed.

Finally, the CO2 splitting process was performed using solid particles of a perovskite material made of a mixture of La and Sr as “A” cation and Fe as “B” cation. To perform these tests, the system was initially operated using an inert material and heated up to 1000 °C using air as fluidizing gas. Then, the inlet gas was switched to pure N2 (99.999%v) and, through a pneumatic system, a small amount (1–2 g) of perovskite was fed to the system by one of the two outlet tubes. Exhaust gas was continuously analyzed to follow the O2 evolution. Once the material was completely reduced (no further O2 release), the inlet gas was switched to pure CO2 (99.999%v) so to re-oxidize the perovskite material and produce CO. During this step, the exhaust gas was continuously analysed to monitor the CO concentration. Finally, when no more CO was produced, the inlet gas stream was switched back to pure N2 to reduce the material again and start a new reaction cycle. The oxidation/reduction cycles were then performed isothermally at a temperature of about 1000 °C. Several iterated reaction cycles were performed and, for each cycle, the amount of O2 and CO produced were evaluated. A first estimate of thermal/chemical efficiency for the proposed reactor configuration was performed.