Solar Thermal Chemical Looping Reforming (SoCLR)
As stated above, the conventional processes of SMR and ATR processes for syngas and hydrogen generation are very energy intensive. The parasitic energy consumption associated with ASU operation for oxygen preparation can account for up to 40% of the operating cost of a syngas production plant (Rostrup- Nielsen. 2002). Using the CLR process can eliminate the need for ASU. However, the endothermic reforming reaction of methane consumes a large amount of heat and creates a significant challenge for the CLR reactor design. On the other hand, solar radiation is the most abundant source of energy that can be utilized for industrial applications. Combining the solar thermal process with CLR eliminates the need for ASU, provides a much better heat management and effectively utilizes solar energy for chemical production and electricity generation. From a thermal energy storage (TES) perspective, the solar reforming combined cycle can be viewed as a heat conversion process that stores intermittent solar thermal energy as latent chemical energy. Under this design, the steam reforming catalyst can serve as the heat transfer fluid to absorb and transport the thermal energy throughout the process.
SoCLR for syngas and hydrogen generation combines the reaction engineering of metal oxide lattice oxygen transfer with the process design of concentrating solar power systems. It has the potential to reduce the carbonaceous fuel consumption associated with conventional syngas/liydrogen production. The key technical challenge of this technology lies in the selection of a suitable metal oxide whose required properties are similar to other chemical looping processes of combustion, gasification and reforming.
The solar thermochemical process for methane reforming to produce syngas replaces fossil fuels with solar energy to supply the energy for the endothermic heat of reaction. This design can potentially reduce the fossil fuel consumption in a reforming process by 20-35% (He and Li, 2014). In addition, using a metal oxide-based reforming catalyst as the heat transfer fluid has the potential to simplify the process flow and improve the economics. Since reforming processes and solar thermal processes are separately available on a commercial scale, the key engineering challenge exists in the process integration and scale- up of this combined cycle design.
In the SoCLR of methane process, methane and steam are the feedstock for the two redox reactors. In the reducer, methane is oxidized by the metal oxides to generate syngas, with the reaction
The reduced metal oxides are then transported to the combustor (oxidizer) to react with steam for H, generation.
Compared to the SMR process, the SoCLR process produces syngas ready for production of methanol or liquid fuel, as well as pure H„ with decreased CO, emissions and simpler syngas processing steps. CO, can also be used to replace steam in the combustor for CO generation with the reaction
He et al. (He and Li, 2014) evaluated the performance of the SoCLR process for syngas production to produce liquid fuel while co-producing H,. The simulation results were compared with conventional SMR and solar SMR processes. The process diagram of the SoCLR for liquid fuel and H, production is shown in Figure 25. Iron-based oxide with perovskite ciystal structure support material of lanthanum strontium ferrite (La0SSr0,FeO,_.) cycled between the oxidation states of Fe.Cfy Fe/FeO was used as the oxygen carrier in the process. The heat required in the reducer was compensated by solar energy. The syngas produced in the reducer was used to produce a hydrocarbon mixture with a Fischer-Tropsch (F-T) reactor and was then upgraded to naphtha and diesel in refining units. A pressure swing adsorption (PSA) system was used to purify hydrogen produced in the combustor after being compressed. Part of the hydrogen was used in the upgrader to upgrade fuels. Byproducts from the different units of F-T reactor, upgrader, and PSA were combusted to generate steam. Process heat was recovered by the heat recovery steam generator (HRSG) and steam turbines in order to produce enough electric power to satisfy the parasitic energy requirements of the system. As a large amount of solar energy is integrated into this system, the life cycle CO, emission from the SoCLR process is expected to be much smaller than traditional methane-reforming processes.
The methane processing capacity of the process was assumed to be 8 t/li, which required a solar input of approximately 60 MW[h and allowed for the integr ation of existing concentrated solar thermal systems into the process. Three different cases of SoCLR, whose parameters were listed in Table 10, were simulated. Case 1 assumed the system was operated under the conditions of the reducer temperature being at 900 °C and system pressure at 1 atm. Case 2 had a system pressure of 10 atm and reducer temperature of 950 °C. Both case 1 and case 2 assumed the reducer and the combustor reached thermodynamic equilibrium. Case 3 had the same operational conditions as Case 1, while the products used the experimentally obtained results from a fixed-bed reducer whose data was kinetically limited by the reactor design and
Figure 25. Sunplified schematic of the SoCLR process.
Table 10. Operating conditions of simulation cases for SoCLR process.
oxygen earner performance. As shown in Table 10, for case 3, even with a much higher Fe304:CH4 molar ratio, the purity of hydrogen is still less than the thermodynamic equilibrium, as shown in cases 1 and 2.
The SoCLR process has garnered significant research interest in recent years, with major researches on process thermodynamics and oxygen carrier materials. Steinfeld et al. conducted an appreciable amount of the pioneering work in this solar thermochemical scheme (Steinfeld et al., 1993; Steinfeld et al., 1995).
Steinfeld et al. proposed the iron/syngas solar co-production process based on an Fe304 and CH4 system in 1993 (Steinfeld et al., 1993). A thermodynamic analysis of the reaction system suggests that at 1 atm and temperatures above 1300 K, the equilibrium reaction components consist of solid metallic iron and a mixture of gaseous FI, and CO in a 2:1 ratio.
The direct contact of carbon with iron oxide can form intermediate products, such as cementite (Fe3C), which deactivates the metallic iron. Therefore, injection of CO, to induce the Boudouard reaction and minimize carbon deposition is critical to maintaining iron reactivity.
A two-step cyclic process based on iron oxide for hydrogen and syngas production from water and methane can be illustrated in a chemical looping diagram, as given in Figure 26. In the first step of this process, the highly energy intensive reduction of Fe304 by methane to form syngas is driven by solar energy. In the second, regeneration step, metallic iron is oxidized by water to form hydrogen and Fe304. This reaction is exothermic and occurs at a lower temperature.
To examine this process, a fluidized bed reactor with a solar concentrator was constructed (Steinfeld et al., 1993). The reactor was a quartz tube of diameter 2 cm operated in a fluidized bed mode. The solar receiver, a 10 cm ID steel cylinder, was installed perpendicular to the reactor on its outer surface. A layer of specular reflective gold was electroplated onto the inner wall of the solar receiver in order to reflect infrared diffuse radiation, thus minimizing energy loss. A solar concentrator collected solar energy, and the reactor was located at the focus of the solar concentrator. A water-cooled steel plate with a circular
Figure 26. Schematic diagram for syngas and hydrogen co-production using Fej04-Fe redox.
aperture of 6 cm diameter was also attached to the solar receiver to prer ent the radiation from spilling. The capacity of the solar heating was estimated to be 1.1 kW for the reactor. This is a typical fluidized bed experiment set-up that utilizes solar thermal energy to pror ide heat for the endothermic reaction. The experiment demonstrated that fluidized particles were effective in absorbing solar radiation. The reaction between Fe,04 and methane occurred in two stages. In the first stage, Fe,Oa was reduced to FeO, and more C0,/H,0 than СО/H, was generated. In the second stage, FeO was reduced to Fe, and more CO than CO, was generated. The first stage had a higher reactivity than the second stage. The sintering and recrystallizatiou of metallic iron contributed to the slower kinetics in the second stage. The conversion of the methane was ~ 20%.
Steiufeld et al. also proposed a SoCLR process for a zinc and syngas co-production process which used a ZnO/Zn redox cycle (Steiufeld et al.. 1995). The concept and advantages in the zinc process are the same as those in the iron process. The major difference is that zinc has a lower boiling point, 907 °C, at which metallic zinc in the product stream vaporizes. The zinc vapor is easily re-oxidized, lowering the overall process efficiency due to irreversible energy loss (Steiufeld et al., 1995). Even though the reactor system was not optimized for obtaining the best performance of the process, the feasibility of producing zinc and syngas simultaneously with solar power was demonstrated. Steinfeld et al.'s intention was to produce zinc, which was considered to be more valuable than syngas, and thus, the operational condition of the reactor was not intended to be used for high methane conversion and syngas generation. The reactor was operated with a CH4:ZuO molar ratio larger than 10. Under this condition, the yield of zinc can reach over 90%, however, methane conversion was very low.
Kodama et al. experimentally compared the performance of different metal oxides, including Fe304, ZnO, ln,03, SnO,, V,05, MoO,, and W03, for syngas and H, or CO co-production with SoCLR of methane a reactor temperature of 1000 °C (Kodama, 2003). W03 and V,Os were found to be reactive and selective for the process. The reduced metallic tungsten from the reducer can split water for hydrogen generation at a temperature as low as 800 °C. Supported tungsten oxides were examined in the temperature range of 800-1000 °C in a bid to improve the reactivity of W03. Experimental results showed that the ZrO,- supported WO, impror ed methane conversion from 40% to 70% and steam conversion from 7% to 30% with a H, selectivity fr om 69% to 97%. However, the CO selectivity dropped from 97% to 86% when CO, was used in the combustor. The improvements were partially attributed to the interaction between WO, and the support material. Shimizu et al. suggested that tungsten carbide (WC) was formed when ZiO,- supported WO, was reduced above 850 °C (Shimizu et al., 2001). In a subsequent study at a temperature of 1077 °C, the solar simulator with 50 wt.% ZrO,-supported W03 reached to a methane conversion of up to 93% with the selectivities of H, and CO of 46% and 71%, respectively (Kodama et al., 2002). However, the solar-to-chemical efficiency was low because H,0 conversion was low (less than 30%).
Kodama et al. also investigated the iron-based oxygen carriers with mixing of other bivalent metal oxides, including nickel, cobalt, and zinc oxide (Kodama et al., 2002). The particles were synthesized with a constant dopant to non molar ratio of 0.15, giving a general chemical formula of M039Fe,61O4 (M = Ni. Co, Zn). A fixed bed reactor of SoCLR was used to examine the particles at an operating temperature of 900 °C. The nickel ferrite (Ni039Fe,61O4) particles provided a higher CO yield and selectivity (22% and 72%, respectively) than pure Fe304 (8% and 63%, respectively). The addition of highly reactive Ni attributed to the enhanced performance. However, severe sintering under the operating temperature significantly hindered the following water splitting reaction. To suppress the sintering effect,
ZrO, was then added to support the nickel ferrite particle (Kodama et al., 2008). Five redox cycles were tested with the ZrO, supported nickel ferrite with good recyclability reported. Methane conversions of around 46-58% with a CO selectivity of around 47% and a H,:CO mole ratio of around 2.5 were obtained from the reduction reaction, while a steam conversion of around 20% was obtained with the oxidation reaction. Cu-fenites with ZrO, support material were investigated by Cha et al. (Cha et ah, 2009; Cha et ah. 2010). It was reported that, in the syngas production step, increasing Cu content in the Cu Fe, л04/ ZrO, medium suppressed carbon deposition and enhanced the reaction rate, which reached its highest point when x = 0.7. As a result of less carbon deposition as compared to the particle without Cu, CO selectivity was higher. In the oxidation step, the addition of Cu promoted the gasification of the deposited carbon. In addition, adding Ce as a binder also improved reactivity in the syngas production step and yielded the highest reactivity when the Ce:Zr molar ratio was 3:1. The Cu0,Fe,,04/Ce-Zr0, oxygen carrier showed high durability and recyclability in ten repeated cycles.
Ceria has been studied widely as a redox material for SoCLR processes, mainly due to its stable fluorite structure up to a nonstoichiometry of 0.25 and the extremely fast kinetics of oxygen diffusion (William andSossina, 2009; Chueh and Haile, 2010). Otsuka etal. tested a series ofCeO,-ZrO, composite oxides (Ce,_AZrvO,) with Zr content below 50% (Otsuka et al., 1998; Otsuka et al.. 2007; Otsuka et al., 1998; Otsuka et al., 1999). It was reported that the syngas formation rates were increased and the activation energy was remarkably decreased due to the addition of ZrO, into CeO, (Otsuka et al., 1998a, 2007,1998b; Otsuka et al.. 1999). By adding Pt catalyst, the reaction was further accelerated. When using CeQ sZr0,0, in the presence of Pt, the operating temperature for converting CH4 to H, and CO can be as low as 500 °C. Krenzke et al. explored the performance of fibrous ceria oxide particles with a low surface area of 0.143 m-/g under different temperature and methane flow rate conditions (Krenzke et al., 2016). It was reported that methane and steam conversion increased when temperature increased fr om 900 °C to 1000 °C, with modest improvements in syngas selectivity. When the temperature was increased to 1100 °C, carbon deposition was observed. There was a tradeoff between high methane conversion, high syngas selectivity and high steam conversion. The methane conversion increased when a slower methane flow rate was used, with reduced syngas selectivity and steam conversion observed.
Morphologies with higher specific surface area can improve syngas production rates if the surface area is stable over many redox cycles. Gao et al. demonstrated that nanostructured ceria with a high surface area and porosity can significantly enhance the initial and long-term syngas production performance (Gao et al.. 2016). Among the three types of ceria morphologies they synthesized, flame-made CeO, nanopowders with a surface area of 77 ny'g had up to 191%, 167% and 99% higher initial average production rates than the flower-like, commercial (8 nr/g of surface area) and sol-gel ceria powders, respectively, given a reaction temperature of 900 °C. High H, and CO production rates as well as nonstoichiometry up to 0.25 were reported. However, after 10 redox cycles, 89% of the specific surface area and 96% of the pore volume of the particle were lost, resulting in syngas production dropping to 57%.
Tests of ceria oxide in lab-scale reactors were conducted. Warren et al. conducted fixed bed experiments in a solar simulator (Warren et al., 2017). The fixed bed held 1130 g of ceria which was then loaded into the reactor, corresponding to a 1/14th capacity of the reactor. SoCLR with carbon dioxide was operated at a temperature of 1120 °C. The reduction gas was argon diluted methane with a CH4:Ar ratio of 1:9. The reported methane conversion was 52% with H, and CO selectivities of 83% and 59%, respectively. Significant carbon deposition was observed. Low efficiency of the system was obtained, partially due to diluted methane supply, high thermal losses in a prototype, and probably the selection of reactant flow rates and cycling tunes. Welte et al. investigated the reduction of ceria with diluted CH4 in a solar particle-transport reactor (Welte et al., 2017). Ceria particles with an mean diameter of 40 nun were dosed downward through a vertically oriented alumina tube enclosed in a solar cavity- receiver, with a flow rate of 0.1-0.6 g/s. Experiments with both co-current and counter-current operations were conducted. Argon diluted CH4 with a concentration of 2.5-10% and with total flow rates of 0.67-2.69 mmol/s was used in the test, which resulted in gas residence times less than 1 s. Apeak thermal efficiency of 12% was projected for operation at 1300 °C with co-current flows of 0.13 g/s of ceria and
2.02 mmol/s of 10% CH4 in Ar. The projected efficiencies are lower than thermodynamic predictions, partially due to high sensible heating requirements incurred by feeding the ceria particles and heavily diluted CH4 into the reactor at near ambient temperature. Chuayboon et al. conducted SoCLR with reticulated porous ceria foam that was directly irradiated in a solar concentrator (Chuayboon et al., 2019). Ceria material with a porosity of 91.8% and a surface area of less than 1 m2/g was used. The nominal temperature of the ceria foam was varied between 950 and 1050 °C. Reduction was performed with 33-67% СН4 in Ar at total flow rates from 11.0 to 20.9 mmol/s per gram of C’eO,. Oxidation was carried out with 55% H.O in Ar at 16.3 mmol/s per gram of CeO,. The energetic upgrade factors were 0.97-1.10 and thermal efficiencies were 2.73-5.22%. Severe carbon deposition was observed with a H,:CO ratio of up to 3.5 in the syngas product. The efficiency was low because of the combined effects of high sensible heating requirements, long reduction and oxidation durations, and the loss of carbon particles. Fosheim et al. operated of a prototype high-flux fixed bed solar reactor with ceria at thermal steady state in a solar simulator for more than 20 redox cycles with CH4 and CO, (Fosheim et al., 2019). First, incremental changes were made in the operating conditions in order to eliminate carbon accumulation and maximize efficiency (Fosheim et al., 2019). Then, the reactor was operated with a CH4 concentration of 75% at 955 °C and 1000 °C for ten cycles at thermal steady state. Fligher temperature promotes better performance of the reactor. At the operating temperature of 1000 °C, CF14 conversion is 0.36, with H, and CO selectivities of 0.90 and 0.82, respectively. The oxidization step had a CO, conversion of 0.69. The energetic upgrade factor under this condition was 1.10, with a heat recovery effectiveness of over 95%. Reported solar-to-fiiel efficiency and thermal efficiency were 7% and 25%, respectively, with projected 31% and 67% for the full-scale reactor, and 56% and 85% for a commercial reactor with lower thermal losses. The projected scaled-up efficiencies suggest SoCLR could be a competitive approach for the production of solar fuels.