High-performance fuel cell designed for coking-resistance and efficient conversion of waste methane to electrical energy

A novel micro-monolithic fuel cell converts low-grade methane efficiently to clean electrical energy.

PHD22/200 HPsi).A weight ratio of 70/30 was selected for the water-ethanol internal coagulant to ensure the optimized morphology in the resultant anode precursor, which was subsequently dried and straightened at room temperature.YSZ-GDC dual-layer electrolyte was dip-coated onto the anode precursor, while preventing it accessing the channels in the anode by sealing using PTFE film.After both layers of electrolyte coating were dried, they were co-sintered at 1400 °C in ambient air for 6 hours.Finally, the LSCF-GDC cathode with a length of 10 mm was dip-coated onto the sintered half-cell that had been cut to a length of 50 mm and subsequently sintered at 1000 °C for 1 h.The thickness of the cathode layer was approximately 25 µm.

X-ray computed tomography
Non-destructive X-ray computed tomographic imaging of the fabricated micro-monolithic anode was conducted on a Zeiss Xradia Versa 520 X-ray microscope at the Electrochemical Innovation Lab (EIL), University College London.2001 X-ray projections with a voxel size of 1.55 m were collected with an exposure time of 18 seconds during the angular rotation of 360 degrees.These radiographic projections were then reconstructed by a proprietary Feldkamp-Davis-Kress (FDK) algorithm in the Zeiss Xradia XMReconstructor software.Post-processing of the reconstructed CT data was conducted using Avizo 9.4 (Thermo Fisher Scientific, UK).

Computational fluid dynamics (CFD) simulation
The reconstructed 3D volume of the monolithic anode was firstly segmented into solid and pore phases using standard thresholding method based on the grayscale difference.A 260 m thick disc with half of the geometry was then meshed and imported into the commercial CFD software Star CCM+ (CD-adapco Inc., London) for image-based simulation.Two types of inlet boundary conditions were set: pure H 2 and 20 % H 2 + 80 % inert diluent to compare the mass transport performance at different operating conditions.At the outlet, a negative source flux corresponding to the current density of 1, 2 and 4 A cm -2 was set respectively to mimic the consumption of reactants at the EAR near the electrode/electrolyte interface.In a porous electrode, the molar flux density can be expressed by the modified Fick's law, exemplified in 1D as: where J H2 represents the hydrogen molar flux density (mol m -2 s -1 ), ε and τ the porosity and tortuosity factor, D the diffusion coefficient, C the local concentration (mol m -3 ), and x the coordinate along permeate direction (m).The CT-resolved micro-channels were modelled as the fluid domain and the solid phase was modelled as the porous medium, with the porosity (29.2 %) and tortuosity factor (8.9) taken from previous measurements 21 .

Morphological and electrochemical characterizations
The morphology of the micro-monolithic cell was determined using field-emission scanning electron microscopy (FE-SEM, Gemini LEO 1525) under secondary electrons imaging (SEI) mode.
Prior to the electrochemical experiments, current collectors were applied by wrapping silver wires

Techno-economic evaluation
Outline: The techno-economic analysis was conducted using a model developed previously [1].In general, various types of costs, including the fixed cost (capital expenditure (CAPEX), auxiliary equipment, maintenance) and operating expenditure (OPEX) (reactant cost, heating and electrical energy input) were estimated together with the profit of generated electricity.To achieve a profitable operation, an electricity price threshold should be first determined, at which the revenue generated from electricity sales will exceed the costs of running the plant.
Plant overview: There are various sources of LCVGs.For example, the Natuna field in Indonesia is the largest gas field in south Asia, with approximately 1.2 trillion cubic metres of reserves and CO 2 content up to 87 % [Journal of Natural Gas Chemistry 21(2012)7-10].The daily production rate of an exemplar natural gas well is set at 2 million metric cubic feet (MMCF) in this model.
Supplementary Figure S2 shows the proposed plant layout for using LCVGs as the fuel in a ceramic fuel cell (without external reformer).Plant specifications, operating conditions and assumptions are outlined in Supplementary Table S2.

Capital equipment costing:
The ceramic fuel cell stack was sized based on its 'reacted' LCVG molar throughput of 6.32 kmol hr -1 resulting in a required reactor area (A reactor , m 2 ) of 135.5 m 2 using equation (S2), (S2)

𝑗
where F is the Faraday's constant (96485 s A mol -1 ), dn/dt is the molar flow rates and j the current density of the cell.
The purchase cost of equipment that is industrially available (i.e.blowers, heat exchangers and furnace) was determined using cost correlations [3,4].The blowers (B A and B C , where A and C denote the anode and cathode side, respectively) were sized according to the volumetric flow rates of fuel and oxidants.The cost of ceramic fuel cell stack was estimated by doubling the material cost.
The mass of individual component, including yttria-stabilised zirconia (YSZ), nickel oxide, gadoliniumdoped ceria (GDC) and lanthanum strontium cobalt ferrite (LSCF), was calculated based on the dimensions shown in Figure 1.Silver wires were selected to estimate the cost of current collector (0.25 mm diameter).Using raw material prices of 71.6 $ kg -1 , 1500 $ kg -1 , 1000 $ kg -1 , 74 $ kg -1 and 0.245 $ m -1 for YSZ, GDC, LSCF, NiO and silver wire respectively [5], the costs for SOFC stack manufacture were estimated to be ca.$ 4100 m -2 .CAPEX for all units installed was estimated using Hand Factors (multiplying factor applied to the purchase costs), which are 2.5, 3.5, 2.5 and 2.0 for fuel cell, heat exchangers, blowers and furnaces, respectively [3].Thus, operating for 7500 years over 10 years, the capital expenditure (CAPEX) per tonne of LCVG was ca.$ 8.

( 0 .
25 mm diameter, 99.99 % purity, Advent Materials Ltd, UK) along the cathode and on the exposed part of the anode.Silver paste was employed to enhance the electric contact between wires and electrode surfaces.A complete single cell with a typical length of 50 mm was then sealed into alumina support tubes with ceramic sealant (Aremco Ceramic Sealant 552, USA) to construct a complete reactor.Wires from both electrodes were connected to a PGSTAT302 N potentiostat with with Nova 1.10.2software (Metrohm Autolab B.V., Netherlands).The base case study was conducted between 650-800 °C, with 50 cm 3 min -1 of dry H 2 fuel fed to the anode and ambient air to the cathode.For measurements with simulated LCVGs, dry hydrogen was replaced by a mixture of CH 4 and CO 2 (total flow rate 50 cm 3 min -1 , CH 4 concentration of 10, 15 and 20 %).Electrochemical impedance spectroscopy (EIS, 10 mV p-p amplitude) was conducted under open-circuit conditions over the frequency range of 10 5 -10 -2 Hz.

Table 2 .
The fitted results of EIS in Fig.2(b) using LR estimated based on the change in specific heat of the streams per hour.Assumption was made that unreacted effluent from anode (CO, H 2 and CO 2 ) to be burnt to provide energy that maintains furnace temperature.The price of oxidant (compressed air) is $ 0.02 per cubic meters.