Allothermally Heated Reactors for Solar-Powered Implementation of Sulphur-Based Thermochemical Cycles

.


Introduction
Thermochemical cycles are a repeating set of two or more consecutive chemical reactions targeted to the synthesis of a specific chemical product or the storage and on-demand release of energy, involving intermediate compounds that are fully recycled at the end of the process.The cycles of the sulphur family recycle sulphur as the central element that appears in different compounds at changing oxidation state.Such cycles like the Hybrid Sulphur (HyS) and the Sulphur-Iodine (SI) cycle were originally conceived to produce hydrogen via water-splitting [1], but can be also modified to produce solid sulphur (SoSu) that can be combusted in air to produce high-temperature heat and hence can be used both as a renewable fuel as well as a seasonal solar energy storage medium [2].These three Sulphur-based cycles share the common step of the decomposition reactions of sulphuric acid (Table 1, [3]): first thermally to steam and SO 3 (1a) and subsequently the catalytic SO 3 dissociation (splitting) to SO 2 and oxygen (1b), their highest-temperature (650-950 o C) endothermic step.In the HyS cycle these two reactions are followed by the electrochemical oxidation of SO 2 with water to yield sulphuric acid and hydrogen (2a, 2b) in a so-called sulphur dioxide-depolarized electrolyzer (SDE).In the SI cycle the SO 2 produced in the first step reacts with water and iodine via the Bunsen reaction (3) to produce two acids, HI and H 2 SO 4 , that are subsequently respectively decomposed according to reaction (4) that produces Hydrogen and (1a, 1b) that re-generate SO 2 .In the SoSu cycle, reactions (1a), (1b) are followed by the disproportionation of SO 2 into elemental sulphur (S) and H 2 SO 4 (5).Elemental sulphur can then be on-demand combusted in air to produce high-temperature heat and SO 2 in (6).The SO 2 from combustion can be converted back into S and H 2 SO 4 to close the cycle.That is exactly the reason why the stoichiometry of reactions (1a), (1b) is written differently ion this specific cycle and the industrially practiced reactions for sulphuric acid production (7) and (8), the (catalytic) oxidation of the combustion product SO 2 to SO 3 (reaction 7, the so-called contact process, reverse of 1b) and, finally, absorption of SO 3 in water and its reaction to sulphuric acid (reaction 6, reverse of 1a) are added in Table 1, to "close" the mass balance so that the "net" sum of reactants and products is zero.The valuable outcome of this cycle is not a chemical product, but the highquality sulphur-combustion heat at temperatures in excess of 1200°C, suitable for combustion in gas turbines and allowing hence efficient combined cycle power generation.The SoSu cycle is flexible enough to be also operated on demand as an "open cycle" (i.e.without "closing" the mass balance in reactions 1a-7) employing/providing material feedstocks (H 2 SO 4 , S, SO 2 ) in combination with sulphuric acid production or with desulphurization of flue-gas or natural gas.The two endothermic steps (1a), (1b) are performed together in sequence and can be driven by heat supplied via concentrated solar energy (CSE).In this context, various CSEheated reactor concepts have been explored, including tubular reactors absorbing solar radiation on their external surface and transferring the heat to their internal one containing the catalyst wherein the reactant gas mixture flows [4] or with the SO 3 splitting catalysts coated on stationary, structured, direct absorbing, volumetric solar receivers from silicon carbide (SiC) foams and honeycombs [1].Even though the latter concept was in principle operational, drawbacks included high re-radiation losses from the front surface of structured bodies and capability to maintain the high temperatures achieved only within a limited length along the porous receiver body, necessitating implementation of complex heat recovery architectures.
Alternatively, the solar receiver and the sulphuric acid dissociation reactor can be spatially de-coupled; the receiver can be employed to heat a heat transfer fluid (HTF) that can in turn be transported via insulated tubes to the reactor where the endothermic chemical reactions take place, in an approach known as "allothermal" heating, first coined by the nuclear energy research community to exploit the high temperature (in the range of 900 o C) of the Helium gas stream from nuclear very high temperature reactors (VHTRs).Obviously, for analogous such solar operation, the receiver should be capable of heating the HTF at a similar level exceeding the temperature required for SO 3 splitting, i.e. ~ 850 -900 o C. Two recently developed such solar receivers employ as HTFs solid ceramic particles and gases, respectively.The first is a developed and patented by DLR centrifugal particle solar receiver (Centrec ® ) that has demonstrated capability of heating flowing sintered bauxite particle streams at temperature levels of the order of 900-950 o C [5].The other, developed by company SUNHELION, is based on direct absorption of thermal radiation by a gaseous HTF similarly to the well-known greenhouse gas effect [6] that can reach temperatures in the range of 1500 o C.These developments can render the implementation of the two sulphuric acid decomposition steps feasible "allothermally" via the enthalpy of solar receiver-heated solid particles in a catalytic reactor/heat exchanger placed on ground level; in such a case structural materials and reactor design issues relevant to the corrosive characteristics of sulphuric acid chemistry at high temperatures can be much easier addressed.

Experimental
In the perspective of scaling up the H 2 SO 4 dissociation/SO 3 splitting process and coupling it to a centrifugal particle solar receiver [5], extensive analysis described in previous works [7,8] concluded to the selection of a shell-and-tube type reactor/heat exchanger to perform the two reactions downstream of the receiver with solar-receiver-heated bauxite particles as HTF flowing downwards in the shell side and providing the necessary heat for the SO 3 splitting reaction at the upper part of the reactor and for the thermal dissociation of sulphuric acid at the lower.In the tube side, rising sulphuric acid vapours come into contact with a non-moving catalyst bed that can contain the catalyst in structured forms like honeycombs or foams.A proof-of-concept, 2 kW shell-and-tube reactor/heat exchanger designed to be heated through bauxite particles flowing in the shell through an electrically-heated inclined belt feeder (Figure 1a) and comprised of six catalytic tubes (Figure 1b) was constructed.Details on the prior catalytic systems tested and on the reactor's design and construction were provided in previous publications ( [3], [9], respectively).Thermocouples were placed on the surface of every tube at different positions to measure temperature distribution along them as shown in Figure 1c.One of the 6 tubes (tube 2) served as a reference tube being equipped with more thermocouples whereas the remaining 5 tubes had the same number of thermocouples placed at the same position.Besides, 12 thermocouples are placed on the parallelepiped shell side of the reactor, co-planar to the thermocouples on the tubes.Non-coated SiC foams of Ø24mm x 40mm were employed in the lower, sulphuric acid thermal dissociation zone as flow diffusers and heat transfer media upon which the H 2 SO 4 vapors decompose according to reaction (1a).Similar, yet Fe 2 O 3 -coated SiC foams comprised the catalytic systems of choice for the shelland-tube reactor since they have demonstrated reproducible near-equilibrium conversion at 850 o C, under a broad range of sulphuric acid flow rates as well as minute pressure drop even under high catalyst loadings (35-45 wt.%) [3].Three non-coated and five coated foams were placed in each one of the six tubes (Figure 1b).Aspects addressed in the design and construction included materials solutions for the manifolds for the injection of the sulphuric acid solution, evaporator tubes, sealants, special ceramic adhesives to join anti-corrosive metallicto-ceramic tubing e.g.(stainless steel to SiC) and materials and reactor solutions to compensate for thermal expansion/contraction [9].During the shutdown or emergency, the hot particles inside the reactor are vented out via an emergency exit (on the reactor door) into a water-cooled emergency container as per the schematic in Figure 1d.The unit was also equipped with an in-house developed customized UV-Vis spectrometry system with a specifically designed corrosion and high-temperature resistant gas cell, for on-line analysis of produced SO 2 .The produced gas is analysed and neutralised at the outlet of the reactor.Hereafter, the gas stream flows completely into the scrubber.The particle heating system has been described in detail previously [9]; briefly it consists of a 100-liters-volume particle container and a 1-meter-long, 35 o -inclined heater tube wrapped with resistively heated elements and encapsulated with insulation.The overall test rig is shown in Figure 1e.

Thermal-only tests
To study the temperature distribution within the reactor and to verify the target temperatures in the evaporation and catalyst zones, several thermal tests were conducted with air as process fluid within the tubes prior to any chemical tests.The particle heater and the reactor performance were thermally tested and the critical temperatures like particle inlet temperature During the start-up of every experiment, the reactor shell is filled completely with cold particles.Later when the particle heater is turned on, the screw feeder (at the bottom of the reactor, Figure 1a) enables the hot particles into the reactor by replacing the cold particles.The speed of the screw feeder defines the residence time of the hot particles falling into the reactor.The test plan has been evolved from test plan 1 to test plan 3 (TP1 -TP3) as shown in Table 2 and explained in detail below, according to the results obtained and the lessons learned at each stage.As per the very initial test plan TP1, a stepwise particle heated bed was planned by regulating the particle flow with respect to time and distance covered along the reactor (height) in such a way that the particles to be sufficiently hot when reaching the evaporation zone as shown in Figure 2. By adjusting the particle flow rate, the target PIT temperature can be achieved and by holding the particle flow for certain duration, the particles can fill required heights within the reactor.In reality, the measured target temperatures might deviate from the below calculated target temperatures because of the process fluid (in case of these tests this fluid was air rather than sulphuric acid) and heat losses.

Figure 2. Test plan 1 (TP1) -Stepwise particle heated bed.
The temperature distribution on reactor tube 2 (used as reference as already mentioned) during TP1 is shown in Figure 3.The particle flow rate was reduced from 43 kg/h to 10 kg/h in a timespan of approx.2:30 hours and the PIT at the outlet of the particle heater has recorded a corresponding progressive increase reaching a maximum temperature of 890°C with a particle flow rate of approximately 10 kg/h.The temperature difference between PIT and RT2-pos.11 (tip of catalyst) was approx.150 o C by the end of the test run due to the heat loss at the interface between the heater and the reactor.After approximately 4 operational hours, the lowermost thermocouple on the reactor tube in the evaporator zone, RT-Pos.0 has recorded a temperature of 100°C, still exhibiting a rising trend.Although the temperatures at RT-Pos.11 are higher than 700°C, the temperature in the evaporation section needs to be higher than the boiling point of sulphuric acid (approx.337°C for 98 wt% concentrated H 2 SO 4 ).It was clear that under these conditions the temperature reached was insufficient for the sulphuric acid splitting and moreover the heat of the particles is essentially used to preheat the reactor.Therefore, the test plan was modified to the one denoted with TP2.Test plan 2 (TP2) is a combination of TP1 with so-called "overnight" heating.The latter refers to running the test setup overnight with a constant, low flow rate of particles heated at a relatively low-to-intermediate temperature, namely 400 o C, followed by TP1 (stepwise particle heated bed) next day morning.The operation parameters are tabulated in Table 3.The temperature distributions on reactor tube 2 (RT2) during the overnight heating and the subsequent day operation according to TP2 are shown in Figure 4 and Figure 5 respectively.As the speed of the particle flow is low, the time required for the hot particles from heater to reach pos.0 is also longer.As shown in Figure 4, the evaporator section was not exposed to high enough temperatures even after overnight operation of the test rig.Soon after initiating the day operation according to the test plan, the temperatures on the surface of the reactor tube 2 started to rise as shown in Figure 5.The PIT has raised from 400°C to 890°C due to change in flow rate and it remained constant from and below 10 kg/h particle flow rate.The temperatures of pos.11 are suitable for sulphuric acid splitting but those at pos.6 are still not favourable for evaporation.Overall, the temperature measurements from TP2 concluded that the overnight heating had a comparatively positive effect on the day operation enabling to attain higher temperatures along the overall length of the tube but the repeated change in flow rates (particularly particle flow rate) is affecting the temperature distribution within the test rig.Therefore, TP2 was further modified to a final test plan (TP3) in which the optimal parameters are kept constant throughout the experimental test.
Test plan 3 (TP3) is actually a fine-tuned version of TP2.The particle flow rate in TP3 is always kept constant to approx.10 kg/hr unlike in TP1 and TP2 because all the set of tests and calculations determined that with this rate of 10 kg/h, the heater could deliver constant PIT of 890°C, higher than the targeted 850 o C for the SO 3 splitting step where the employed catalysts have demonstrated near-equilibrium conversion.The operation parameters of TP3 are shown in Table 4; they are identical during overnight heating and day operation.The temperature distribution on reference tube RT2 is shown in Figure 6.The air flow was interrupted soon after 2 hours of test start due to the malfunction of the air supply system.There was no air flow into the tubes during the overnight heating.This was an unfortunate harsh test on the reactor which however, resulted in a positive result in the sense that no physical damage of tubes occurred, as confirmed with a leak test.The air flow was manually resumed again during the day operation which resulted in the small drop of temperatures recorded throughout the tube.The decomposer zone had already reached more than 750°C during overnight operation.The temperature of pos.11 (tip of catalyst) was close to 800°C during the day operation, whereas the pos.0 of the evaporator zone was above 400 °C during the start of the day operation.In conclusion, the temperature distribution of the evaporation zone was improved by a factor of 2 using TP3 compared to TP2. 14:00 16:00 18:00 20:00 22:00 00:00 02:00 04:00 06:00 08:00 10:00 However, despite the fact that the reference reactor tube reached the targeted temperature, the temperatures of the six tubes at the same "vertical" position were not equal as shown on the relevant graphs of Figures 7a and 7b, where the temperatures at the top of the catalytic zone (TC at position 11) and at the tip of evaporation zone (TC at position 5) are compared.At position 11, tube 1 has reached the highest temperature (approx.800°C), but these of tubes 3 and 4 are almost 100°C lower than that of tube 1.The particle flow pattern affects the temperatures of different reactor tubes at a single vertical position of plane.The flow pattern of the reactor setup is shown in Figure 1d and the temperature differences of the tubes at the same horizontal position can be minimised by altering the flow pattern (which can be achieved by modifying the heat shield in Figure 1d.

Conclusions and outlook
The development of solar receivers capable of delivering high-temperature (> 850 o C) heat transfer fluids, provides for implementation of endothermic reactions via concentrated solar energy in suitably designed allothermally-heated reactors/heat exchangers placed away from the solar receiver.Such a 2-kW laboratory-scale reactor to perform thermal sulphuric acid decomposition and catalytic sulphur trioxide splitting was in-house designed, built and tested with electrically heated bauxite particles, demonstrating the in-principle feasibility of the proposed concept.
The reactor successfully underwent multiple thermal test runs demonstrating the inprinciple feasibility of both sulphuric acid decomposition as well as of SO 3 splitting.The temperatures reached in the lower sulphuric acid evaporation zone (~ 400 o C) were sufficient to ensure complete sulphuric acid evaporation.However, the ones reached in the upper SO 3 splitting zone were of the order of 750°C, high enough to demonstrate SO 3 splitting but not reaching the levels required for satisfactory conversion (~ 850 o C).An improved version of the reactor is under construction incorporating design modifications based on lessons learned from the test campaigns, in the perspective of scaling up the process and coupling it to a centrifugal particle solar receiver operating on a solar tower.

Figure 1 .
Figure 1.(a) CAD of lab-scale particles-heated reactor/heat exchanger test setup (from [9]); (b) schematic of six reactor tubes included in the actual reactor within two 3-tube bundles and photographs of typical non-coated and Fe2O3-coated SiC foams used in the evaporation and catalytic sections respectively; (c) thermocouple positions for temperature measurements along the tubes; (d)operation schematic of side safety exit gate for hot particles stream drain (from[9]); (e) overall labscale insulated particles-heated reactor/heat exchanger test setup unit with particles heater and feeder and all necessary peripherals.
, temperatures at position 1 (beginning of the tubes), position 5 (end of the evaporator zone) and position 11 (top tip of the catalyst) were reviewed during multiple thermal tests.

Figure 3 .
Figure 3. Temperature distribution on Tube 2 according to TP1

Figure 4 .Figure 5 .Table 4 .
Figure 4. Temperature distribution during overnight heating on reactor tube 2 at every position to TP2 with particles flow rate of 4 kg/h and air flow rate of 22.5 slm.

Figure 6 .
Figure 6.Temperature distribution on reactor tube 2 at every position according to TP3

Table 2 .
Test plan development