Hydrogen Production via Off-Sun Solar-Thermal Supercritical Water Gasification and Membrane Reforming of Piggery Waste

. Supercritical water gasification (SCWG) represents an emerging technology for liquid fuel synthesis, offering large potential in significantly improving the efficiency and environmental impact of clean fuel production. Compared to conventional gasification, SCWG proceeds at much lower temperatures, allowing char and tar-free gasification of biomass and for low-quality and high-moisture content biomass to be used. As the thermochemical processes involved in SCWG and steam methane reforming (SMR) are highly endothermic, combining them with concentrated solar power (CSP) and thermal energy storage (TES) could allow the process to be driven solely by renewable energy. As such, this work models an off-sun SCWG/SMR system using novel molten salt, proposed to reach up to 600°C [1], which overcomes the previously limiting molten salt temperature of 550°C. Using an off-sun configuration, it deals with the on-sun configuration’s issues of degraded lifetime and creep-fatigue, allowing for uniform heating and a reduced load on the reactors. The novel technology of an integrated Pd-based membrane reactor, highly selective to H ₂ , was chosen given its ability to carry out SMR at temperatures in the viable range of new molten salts . The system uses a waste feed-stock, piggery waste, to provide the dual benefit of reducing the environmental cost associated with piggery waste’s release of CH 4 . From the model created in Aspen Plus, a plant output of 7.2 kmolH ₂ /h (1,135 Nm³/h of H ₂ ), was obtained with a flow rate of 150 kg/h of dry piggery waste, a membrane area of 131.8 m², and a 73% CH 4 conversion from the feed.


Introduction
Bioenergy, produced from the conversion of biomass, has been identified by the IPCC as a major contributor in all scenarios to achieving the Paris Agreement and limiting global warming to 1.5°C [2].Specifically, bioenergy in the form of biofuels, including gaseous fuels like hydrogen and liquid fuels like biodiesel, biomethane, and bioethanol, has the potential to replace petroleum-based fuels and, when produced using renewable energy sources, can be carbon neutral [3].Biofuels can bridge gaps in key sectors like industry and transport that cannot be met by renewable electrification, including heavy transport and high-heat industrial processes, and can help overcome the intermittency of solar and wind energy by providing an alternate means of energy storage.Supercritical water gasification (SCWG) represents an emerging technology for such liquid fuel synthesis, offering significant potential for improvement to the efficiency and environmental impact of biofuel production [4].Unlike traditional gasification, SCWG allows char and tar-free gasification of biomass, and for low-quality and high-moisture content biomass to be used [5].Producing a methane-rich biogas, SCWG can be coupled with downstream reforming processes like steam methane reforming (SMR) to produce syngas, a gas mixture of H 2 and CO, which can then be used in Fischer Tropsch (FT) synthesis to produce hydrocarbons like biodiesel and biomethanol [5].As the thermochemical processes involved in SCWG are highly endothermic, often operated at temperatures above 550°C, the use of concentrated solar power (CSP) offers exciting potential for driving the SCWG reaction using renewable energy.Two main methods of CSP-driven SCWG have been identified in literature, referred to in this work as on-sun or off-sun.The on-sun method directly uses incoming concentrated solar radiation to run the SCWG reactors at temperatures around 750°C, while the off-sun method combines a CSP plant with thermal energy storage that acts as the heat transfer fluid (HTF) for the SCWG reaction at temperatures of around 550-600°C [1].
In previous work, Rahbari and Pye investigated a technology for converting farmed microalgae to solar fuels.In their initial work, they maximised the exergy efficiency of the SCWG process using a CSP thermal input [6], which then inspired the design and optimisation of a complete on-sun CSP-driven SCWG/SMR/FT system [7], [8].Their final works reviewed the microalgae SCWG system, discussing the potential benefits of CSP integration [9], [10].They found that the on-sun configuration of the system was limited due to the need for high-cost materials, thermal stresses, and mismatch between the fluctuating solar resource and steady thermal duty of the plant.The use of the off-sun configuration was therefore proposed by the authors to overcome these issues of degraded lifetime and creep-fatigue, allowing for uniform heating and a reduced load on the reactors.As conventional SMR requires temperatures of up to 850-950°C, the proposed off-sun configuration was however limited by the temperature bounds of thermal energy storage at 550-600°C, meaning it was incapable of meeting the high temperatures needed to drive the SMR reactions without additional thermal input.
The novel technology of an integrated Pd-based membrane reformer, highly selective to H 2 , can overcome this limitation, allowing for SMR to be performed at temperatures in the viable range of new molten salts due to its continual removal of permeated H₂.Past projects, specifically the CoMETHy project [1], [11], have developed promising integrated membrane reformers, incorporating a high degree of compactness and flexibility to the type of feedstock used for the conversion.This flexibility allows the use of alternate feedstocks, including waste feedstocks like piggery waste.In agriculture, manure management is the third highest source of emissions [12], accounting for over two thirds of total emissions in pork production.Being the largest source of emissions associated with pork production, alternative solutions to pork effluent management focused on the capture and reuse of methane, therefore have the potential to significantly reduce emissions associated with the pork industry [13].Using piggery waste as the feedstock in an off-sun SCWG/SMR system therefore can provide the additional benefit of reducing the environmental cost of piggery waste associated with its release of CH 4 and run-off into waterways.
The remainder of this paper is organised as follows: Section 2 provides a description of the model, including the CSP and molten salt storage system, SCWG reaction, membrane reformer, and selected piggery waste feedstock.Section 3 examines the results of the two reactor models, using sensitivity analysis to discuss the impact of key assumptions and design parameters.Section 4 concludes the paper.

System Description
The proposed system comprises of two key reactors utilising 600°C solar salt supplied from a co-located CST plant as heat carriers: a SCWG reactor (MS-SCWG) that will convert the feedstock into biogas, and an integrated membrane reactor (MS-IMR) that will produce pure H₂ and CO₂ streams.The MS-SCWG reactor comprises of a shell and tube heat exchanger, with molten salt flowing in the shell and the wet feedstock flowing in the tubes.The design and performance for the MS-IMR was taken from that developed in the CoMETHy project [1], [11], but with molten salt flowing at 590°C rather than 540°C, and using the product biogas from the MS-SCWG reactor as the feed stream.The CST plant, storage, and reactors were sized at 2.3 MWth with 12-hour storage.To model the system, Aspen Plus V8.8 software was used (Figure 1), taking a Peng-Robinson Equation of State (PREoS) with van der Waals mixing rule for thermodynamic property calculations and a sequential modular simulation approach.Piggery manure is composed of organic matter, nutrients, including nitrogen and phosphorus, salts, and microorganisms.It has a high moisture content, typically >75%, and therefore low wet mass methane potential, meaning it is inappropriate for traditional gasification without drying and has low performance in anaerobic digestion systems but is well suited to SCWG.The piggery waste was modelled using its ultimate and proximate analysis, which determine the mass concentrations of its major organic chemical elements, carbon, hydrogen, nitrogen, sulfur, and oxygen, and its mass percentages of moisture, volatile matter, fixed carbon, and ash content, respectively.The ultimate and proximate analysis used in this work has been summarised in Table 1.

CSP and Molten Salt Storage
Concentrating solar power (CSP) technology has seen significant advancements in the last few decades, with large-scale storage options being developed to offer viable means of overcoming the intermittency of solar energy supply associated with on-sun operation, and allowing the 24-hour operation of industrial procedures.CSP systems use a combination of mirrors or lenses to concentrate direct-beam solar radiation and produce useful forms of energy like heat, electricity, or fuels through various downstream technologies.One of the most common CSP systems commercially deployed is the central receiver tower configuration, where solar radiation is concentrated by factors of 500 to several thousand by a field of heliostats and intercepted by a receiver that converts it to thermal energy.This thermal energy can either be directly used to drive chemical reactions, converted to electricity via thermal power cycles, or carried away via a HTF.
One of the most promising storage options that has been developed is two-tank molten salt storage, wherein molten salts act as both the HTF and heat storage fluid (HSF).In the twotank approach, the molten salt is cycled between a hot tank (∼565°C) and a cold tank (∼290°C) [15].Molten salts have the benefit of high boiling points, low viscosity, low vapour pressure, high volumetric heat capacities, and therefore low operating costs.Currently, hot tank molten salt storage has achieved temperatures of up to 565°C, with future technologies offering the potential to reach up to 600°C [1].Known as "solar salt", NaNO 3 /KNO 3 (60/40 w/w) has emerged as the most promising molten salt for solar applications.
For this model, a central receiver tower configuration with two-tank storage was selected.The off-sun system, consisting of a series of shell and tube heat exchangers where the molten salt flows in the shell at temperatures up to 600°C and feedstock flows in the tubes under high pressure, allows the following benefits: • fixes mismatch between fluctuating solar resource and steady thermal duty profile of chemical plant, and • reduces the need for high performance alloys and resolves issues of degraded lifetime and creep fatigue due to lower thermal stresses, lower peak temperatures, and more continuous operation.

Supercritical Water Gasification (SCWG)
Traditional gasification technologies encounter several difficulties that limit their development and application.Firstly, the quality of the product gas is often low due to contamination by impurities like char and tar [16].Secondly, standard gasification technologies require dry biomass, as a high percentage of moisture in the biomass will decrease the temperature inside the gasifier, reducing its efficiency.As biomass usually contain a high percentage of moisture this then requires that biomass be dried before it is fed into the gasifier, contributing substantial energy and cost additions to the process.
The use of supercritical water in the process of supercritical water gasification (SCWG) has therefore emerged as a potential solution to these limitations of typical gasification.For pure water, the critical point is 374°C and 22.1 MPa, meaning above this temperature and pressure, water is in its supercritical phase.In this supercritical phase, the physical properties of water change drastically, with water behaving as a homogeneous fluid phase exhibiting both liquid-like density and gas-like viscosity [10].In the supercritical state, water's density is reduced by an order of magnitude, drastically reducing its hydrogen bonding and the dipoledipole interaction and allowing water to act as a non-polar rather than polar solvent with a high solubility for organics and low solubility for salts.This unique solvent property, opposite to that experienced in normal conditions, allows char and tar-free gasification of biomass, and the potential for low-quality and high-moisture content biomass to be converted into gas [17].

Membrane Reformer
Palladium has emerged as the leading dense metal membrane due to its high selectivity to hydrogen.Pd and Pd-based membranes can dissociate molecular hydrogen into monatomic form ready for its fast diffusion through its lattice.Pd alone however is expensive and so it is often alloyed with inexpensive elements like Cu, with alloys like Pd-Ag having almost double the hydrogen permeability of pure Pd [18].
The performance of Pd-based membrane reactors can be described by the reaction's conversion, permeate flow rate, hydrogen yield, and hydrogen recovery.The H 2 permeation through a Pd-based membrane is usually described by Sieverts' law, where   2 ,   , and   refer to the membrane's permeance, hydrogen partial pressure on the retentate side, and hydrogen partial pressure on the permeate side, respectively.Pd-based membrane reformers benefit from a high selectivity, but, as with other membranes, face challenges regarding degradation and stability over time, as well as potentially higher costs.
Of the membrane reactor configurations, including packed bed and fluidised bed configurations, few have been designed to operate with renewable energy vectors like off-sun CSP.The CoMETHy project therefore stands as the key reference point for an off-sun membrane reactor.Running from December 2011 to December 2015, the CoMETHy (Compact Multifuel-Energy to Hydrogen converter) project led to the development of an innovative, compact membrane reformer, in which the membrane is integrated with the catalyst bed and the molten salts heat exchanger.The integrated membrane reformer (IMR) achieved an experimental production rate of up to 3 Nm 3 /h of pure, permeated hydrogen (>99.8%H 2 with < 100 ppm CO content) and methane conversions of over 60%, which are over double that attainable in a conventional reformer under the same low temperature conditions.From characterisation of the membrane, a H 2 permeance of 10-15 (Nm 3 /h)/m 2 /bar 0.5 was obtained.In a TEA performed for the membrane reactor, the pilot membrane reactor was then scaled up to a 1500 Nm 3 /h capacity, taking the hydrogen permeance of 15 (Nm 3 /h)/m 2 /bar 0.5 , a methane conversion of 73%, and hydrogen recovery of 90%.A sweep steam, designed to achieve a steam-tocarbon ratio of 3.2 (on molar base), was injected in the inner tube of the membrane reactor to increase the difference of the hydrogen partial pressures between the feed and permeate sides of the reactor by removing permeated hydrogen from the reactor.The results of this IMR design were used in this work's model, with Figure 2 showing the 2D representation for the IMR that was implemented.

SCWG Reactor
In the SCWG reactor, it was assumed all ash and 15% of the char do not interact in the MS-SCWG equilibrium reactions and can be removed, and that all nitrogen (N 2 ) and sulfur (S) react to form NH 3 and H 2 S that is then removed from the MS-SCWG products stream.The piggery waste enters the system at standard conditions of 25°C and 1 bar pressure.The piggery waste is first sent to a pump where it is pressurised to 240 bar pressure.It is then sent through a series of heaters where it is heated to 600°C.The heated piggery waste is then sent to a RYIELD reactor where all hydrogen was converted to H 2 , all oxygen to O 2 , all nitrogen to N 2 , all sulfur to elementary S, and all carbon to elementary C. The product gas is then removed of all N 2 and S in the form of NH 3 and H 2 S using a separator.This cleaned gas is then sent to a final RGIBBS reactor, potential product species are set by those inputted to the model and calculated according to Gibbs energy minimisation.The species considered as products were H 2 O, H 2 , N 2 , O 2 , S, C, CO, CO 2 , CH 4 , H 2 S, and NH 3 .
All reactors within the MS-SCWG model were assumed to operate under isothermal and isobaric conditions without the use of catalysts.The heat streams for the RYIELD reactor and separator were fed to the RGIBBS reactor to determine the total heat duty of the SCWG reaction.From the MS-SCWG reactor, biogas with a 0.259 molar fraction CH₄, 0.212 molar fraction CO 2 , and 0.068 molar fraction H 2 was produced.

Integrated Membrane Reactor
To perform the sequential modular simulation in Aspen Plus, it was assumed the reactor was operating under steady state, with counter current flow, constant pressure across the reactor, local equilibrium achieved in each partial area of the reactor, membrane is only selective to H 2 , the CO content in the feed stream doesn't hinder the permeability of the membrane, all sweep steam entering the inlet of the reactor exits at the outlet, and there are no losses to reactor performance over time.
In the simulation, the separation of hydrogen gas through the membrane was split between n partial areas using a series of ( + 1) RGIBBS sub-reactors and  sub-separators.The RGIBBS sub-reactors model the local equilibrium reactions occurring across the MS-IMR as H 2 is permeated through the membrane and removed from the reactor by the sweep steam.After each sub-separator, which separates a permeated hydrogen stream from the feed stream, the equilibrium for the WGS reaction in the feed stream is shifted to the left, the reactants side, as per Le Chatelier's principle, allowing the reaction to proceed further in favour of the products, H 2 and CO 2 , until equilibrium is reached again between the reactants and products.Alternating the RGIBBS sub-reactors and the sub-separators, this H 2 removal and redriving of the WGS reaction to equilibrium is modelled for each of the n partial areas of the MS-IMR and then summed to determine the final mole flow for the permeate and retentate streams.Each separated H 2 stream is combined in a mixer with the sweep steam stream to produce the final permeate stream product.After the th sub-separator, the feed stream is sent to a final RGIBBS sub-reactor, from which the retentate CO 2 , unreacted gases, and non-permeated H 2 exit the MS-IMR.Governing the membrane's permeance was the calculation of the hydrogen flux based off the hydrogen partial pressure difference between the feed side and permeate side of the membrane according to Sievert's law.
To calculate the hydrogen partial pressure difference, a 9.5 bar pressure on the feed side of the membrane, 0.4 bar pressure on the permeate side of the membrane, and 0.5 fraction of H 2 in the permeate stream (with the other 0.5 fraction H 2 O from the permeated sweep steam) were used.The best case permeance of 15 (Nm 3 /h)/m 2 /bar 0.5 obtained from the membrane's characterisation was used.Providing the hydrogen flux in (Nm 3 /h)/m 2 , this calculation was then converted to flux on a mole basis, (kmol/h)/m 2 , and multiplied by the area of the membrane to determine the total permeation of hydrogen.The area was calculated as that needed to remove 90% of H2 in the feed stream to the sub-separator, based off the 90% recovery factor that was achieved by Giaconia et al. [1].A steam to carbon ratio (/) (moles of H 2 O to moles of atomic carbon) of 3.2mol/mol of sweep steam to feed stream was used, with the sweep steam entering the MS-IMR at 441°C.In their experimental work, Giaconia et al. found that methane conversion steadily increased with increasing /, achieving up to 99% conversion with a 13 mol/mol / with inlet molten salts at 544°C, but that only up to approximately 7 mol/mol did the / have a significant effect on the CH 4 conversion.Use of higher / were not recommended due to the higher heat input required for steam generation, leading to the optimal compromise of a 3.2 mol/mol /.The outlet retentate stream was assumed to exit at 8.5 bar and 590°C.
The ratio of H 2 O to atomic carbon in the feed stream was also found to influence the H 2 permeation and CH 4 conversion in the MS-IMR.In the flowsheet, after exiting the MS-SCWG reactor and being cooled to 90°C, the liquefied water is removed from the product stream using a separator to be recycled or re-heated for use as sweep steam for the MS-IMR.The amount of H 2 O in the feed stream to the MS-IMR can therefore be determined by the fraction of H 2 O that is removed at this separator (H2OSEP).With no H 2 O in the initial feed stream, the steam reforming reaction dominates over the WGS reaction, resulting in a higher yield of CO in the final retentate stream and reducing the mole fraction of CO 2 .At higher mole flow rates of H 2 O, the mole fraction of hydrogen in the feed stream is reduced, decreasing the hydrogen partial pressure difference between the feed and permeate side and thus the potential hydrogen that is permeated through the membrane.For the 10 partial areas, a H 2 O mole flow in the feed stream was set for achieving a CH 4 conversion of 73% overall, as was achieved by Giaconia et al.The results of the sequential MS-IMR model was a plant output of 7.2 kmolH₂/h (1,135 Nm³/h of H₂) with a flow rate of 150 kg/h of dry piggery waste, a total membrane area of 131.8 m², 104.1 kW heat duty, and a 73% CH 4 conversion from the feed.

Sensitivity Analysis
Figure 3 shows the results of scenarios analysed to interpret the behaviour and sensitivity of the MS-SCWG reactor.Figure 3 (left) shows the effect of temperature of the MS-SCWG reactor, whereby the production of H 2 from the piggery waste is the most dependent on the reactor temperature, significantly decreasing at temperatures below 550°C.Figure 3 (right) models the MS-SCWG reactor using five different ultimate and proximate analyses for the piggery waste [5], [19], [20], [21], showing how the product H 2 remains steady across all ultimate analyses but the product CH 4 and CO 2 experience significant variation, which could impact the results of the subsequent MS-IMR.
Figure 4 shows simulation results for the performance of the sequential model for the MS-IMR under different temperatures for the hot MS in the reactor's shell and for varying water mole fractions in the feed stream, which have been identified as two key parameters influencing the performance of the membrane.Figure 4 (left) shows that at higher MS temperature and with a higher water mole fraction in the feed stream, the flow rate of H 2 in the permeate increases.The results for the area of the membrane (Figure 4 (right)) are much more variable, specifically in the case of the 500°C MS where at low (below 0.3) and high (above 0.5) water mole fractions a very large area is needed to allow just a small permeation of H 2 or achieve a H 2 permeation close to that achieved at 550°C 600°C.The two higher MS temperatures both follow similar trends for membrane area, with a minimum at 0.3 water mole fraction.For MS at 600°C, however, the area and H 2 permeation rate plateaus at water mole fractions >0.65 as the CH 4 in the stream reaches near to complete conversion, and the hydrogen partial pressure in the feed drops below the partial pressure in the permeate.Desirable operating ranges to reduce membrane area are therefore greater or equal to 550°C and at <0.5 water content.

Conclusion
The modelled technology of a system consisting of a 2.3 MWth CST plant with two tank molten salt (taken to be emerging 600°C solar salt) storage, an off-sun SCWG reactor, and a membrane reactor has achieved the production of a pure 7.2 kmol/h H 2 stream from 150 kg/h dry pig waste.From the MS-SCWG reactor, biogas with a 0.259 molar fraction of CH₄ was produced, which was then converted with a 73% efficiency in the MS-IMR using a membrane area of 131.8 m² and with a 63% pure CO₂ retentate stream.The proposed system has overcome barriers associated with on-sun SCWG/SMR, converting the piggery waste to H₂ within one integrated system and within the temperature bounds of molten salt, avoiding the high temperatures needed for conventional SMR.Going forward, optimisation of the system's exergybased model, scale, and efficiency is needed, including the design of a more efficient heat exchanger network to reduce exergy losses throughout the system via coolers and other equipment.Improved designs should then be followed by a pilot project to test the system's performance against the mathematical and equilibrium-based model created in this work.

Figure 2 .
Figure 2. Diagram of 2D MS-IMR model used for sequential modular simulation.

Figure 3 .
Figure 3.Effect of MS temperature and pig waste's ultimate analysis on MS-SCWG reactor.

Figure 4 .
Figure 4. Total H 2 permeate flow rate and membrane area at different water mole fractions for IMR with MS at 600°C, 550°C, and 500°C.