Removing 65Years of Approximation in Rotating RingDisk Electrode Theory with Physics-Informed Neural Networks

Polymer electrolyte fuel cells (PEFCs) are very promising as clean energy sources for future automotive applications. However, the performance and durability of PEFCs can be significantly decreased by the presence of contaminants. Such species may adsorb on the Pt catalyst, penetrate into the ionomer and membrane, or compete with the main reactions, resulting in a decrease of the active surface area, ionic conductivity of the catalyst layer and membrane, and cell performance. Also, H2O2 is a side product of the oxygen reduction reaction (ORR) and its associated yield often increases in the presence of organic contaminants, such as acetonitrile, propene, naphthalene, and methyl methacrylate [1-3]. A larger H2O2 generation rate exerts an undetermined influence on the performance and durability of PEFCs. The formation of H2O2 has been linked to the radical attack decomposition of the Nafion® ionomer, which is accelerated by the presence of Fe2+ and Cu2+ ions [4,5]. The Nafion® ionomer is not only used as an electrolyte membrane, but is also added to the catalyst layer to exploit its binding property and improve catalyst utilization. Consequently, the decomposition of the Nafion® ionomer is a significant risk to PEFC performance and durability. Therefore, it is necessary to investigate the ORR, hydrogen oxidation reaction (HOR), and H2O2 yield in the presence of contaminants to understand their effects on PEFCs. Ethylene glycol (EG) is of interest because it is widely used as a coolant, antifreeze agent, and de-icing solution. De-icing of airport runways and airplanes is the primary source of EG in the environment. EG is also dispersed in the environment by the disposal of products that contain it. Caprolactam is another potential contaminant released by PEFC system materials either as a result of degradation or leaching. It has been detected in many leachates including from polyphthalamide materials that are being considered for use as balance-of-plant structural materials [6,7]. Both EG and caprolactam can poison ORR Pt/C catalysts [8-10], but the H2O2 yield has not been measured. In contrast to the slow ORR kinetics, the HOR kinetics on Pt/C catalysts is so fast that the cell voltage losses are negligible even for very low Pt loadings. However, attention has not been given to the poisoning effects of EG and caprolactam on the HOR. These considerations are important for prevention because tolerance levels are currently missing. The rotating ring/disk electrode (RRDE) is the equipment of choice to reveal the H2O2 yield and reaction pathway. In this presentation, the contaminant impacts of EG and caprolactam on the kinetics of ORR and HOR in acid media using RRDE are reported. Results will emphasize the changes in the electrochemical surface area, ORR and HOR kinetic currents, Tafel slope, H2O2 yield,and reaction pathway as a function of EG and caprolactam concentration. Acknowledgments Authors are grateful to the Office of Naval Research (award N00014-13-1-0463) and the Hawaiian Electric Company for their ongoing support to the operations of the Hawaii Sustainable Energy Research Facility.

On flows induced by diffusion in a stably stratified fluid

Polymer-fluid interfaces are used widely in a variety of applications, including separations, which require exposure of the polymer to dynamic flow conditions. Despite the ubiquity of such interfaces, the importance of convective mass transport within the near-interface region of a polymer is a fundamental process that is still poorly defined. As a step toward better defining mass transport behavior within the near-interface portion of a polymer, in this work, a new application of a spectroscopic method based on the combination of Förster resonance energy transfer (FRET) and total internal reflectance fluorescence microscopy (TIRFM) is reported that allows quantification of the penetration depth of a laminar flow field (i.e., the slip length) in a densely grafted, thin poly(N-isopropylacrylamide) (pNIPAM) film as a model polymer system. Specifically, decay curves from FRET of an acceptor with a donor attached at the substrate surface are fit to a combined Taylor-Aris-Fickian mass transport model to extract apparent linear diffusion coefficients of acceptor molecules for different flow rates. Apparent diffusion coefficients range from 1.9 × 10(-12) to 9.1 × 10(-12) cm(2)/s for near-surface flow linear velocities ranging from 192 to 2952 μm/s. This increase in apparent diffusion coefficient with fluid flow rate suggests increasing contributions from convective mass transport that are indicative of flow field penetration into the polymer film. The depth of penetration of the flow field is estimated to range from ∼6% of the polymer film thickness in a good solvent at ∼192 μm/s to ∼60% of the film thickness at ∼2952 μm/s. Thus, flow field penetration into polymer thin films, with its concomitant contributions from convective mass transport within the near-interface region of the polymer, is demonstrated and quantified experimentally.

Rotating ring-disk electrode method to evaluate performance of electrocatalysts in hydrogen peroxide activation via rapid detection of hydroxyl radicals

Templated bi-metallic non-PGM catalysts for oxygen reduction

Carbon dioxide is a major contributor to global warming and, beyond the mandatory reduction of our emissions, the possibility of recycling this greenhouse gas is becoming increasingly attractive. The electrochemical reduction of CO2 is an interesting pathway, since a broad range of useful products can be formed such as methanol and formic acid (fuels for PEFC), methane and ethylene (reactants for synthesis or combustion process), CO and H2 (syngas), etc. [1]. Nevertheless, on top of the high overpotential required to drive this reaction, the electrochemical reduction of CO2 suffers from a poor yield/selectivity of valuable products [2, 3].Copper-based electrocatalysts seem to be the most efficient to overcome these kinetic and barriers, [3], but it has been shown that the size of the Cu (nano)particles envisaged for practical applications can strongly influence the reaction selectivity [4]. In order to clarify this particle size effect, one needs to determine the electrochemically active surface area (ECSA) of the catalysts of interest that, for Cu, can be derived in-situ by lead underpotential deposition (Pb-UPD) [5]. Additionally, the selectivity of the CO2-reduction reaction on these well-characterized materials needs to be determined with an appropriate technique. With this motivation, we have performed Pb-UPD measurement on polycrystalline Cu and on carbon-supported Cu nanoparticles (NPs) of various particle sizes in a 3-electrode glass cell filled with 0.01 M of HClO4 with 50nM of Pb(ClO4)2 and 100 nM of KCl. In these measurements, we have used rotating ring disk electrode (RRDE) voltammetry to quantify the amount of of Cl- (typically ≈ 10%) co-adsorbed with the Pb2+ [5]. An example of the results obtained with this approach, Figure 1 displays the disc and ring currents recorded in a shielding experiment recording a cyclic voltammogram on the polycrystalline Cu disk while applying a constant potential of -0.25V vs. RHE at the ring, at which the continuous Pb-plating ring current is affected by the adsorption/desorption of Pb2+ species at the disk. Having estimated the ECSA for various catalysts, we then used differential electrochemical mass spectrometry (DEMS) to quantify the yield of the reaction’s gaseous products, using an in-house constructed DEMS setup combined with a dual thin layer flow cell adapted from previous works [7, 8] and separated from the quadrupole mass spectrometer chambers by a PTFE membrane. In summary, this contribution will introduce an RRDE-based methodology for the determination of the ECSA of Cu-based electrocatalysts That we have applied to the characterization of polycrystalline Cu and carbon-supported Cu nanoparticles of different sizes,. These results will be complemented with the corresponding CO2-reduction yields, as quantified with our DEMS setup.

Recent theoretical studies of the bromate electroreduction from strongly acidic solution have been overviewed in view of very high redox-charge and energy densities of this process making it attractive for electric energy sources. Keeping in mind non-electroactivity of the bromate ion the possibility to ensure its rapid transformation via a redox-mediator cycle (EC′ mechanism) is analyzed. Alternative route via the bromine/bromide redox couple and the comproportionation reaction inside the solution phase is considered within the framework of several theoretical approaches based on the conventional Nernst layer model, or on its recently proposed advanced version (Generalized Nernst layer model), on the convective diffusion transport equations. This analysis has revealed that this process corresponds to a novel (EC″) electrochemical mechanism since the transformation of the principal oxidant (bromate) is carried out via autocatalytic redox cycle where the bromate consumption leads to progressive accumulation of the bromine/bromide redox couple catalyzing the process. As a result, even a tracer amount of its component, bromine, in the bulk solution leads under certain conditions to extremely high current densities which may even overcome the diffusion-limited one for bromate, i.e. be well over 1 A/cm2for concentrated bromate solutions. This analysis allows one to expect that the hydrogen–bromate flow battery may generate very high values of both the current density and specific electric power, over 1 A/cm2and 1 W/cm2.

A series of numerical investigations was carried out to study the behavior of cavitating turbulent flows in an orifice. In the present work, two different cavitation models were used for the simulation. In the first model, flow was modeled as two interpenetrating fluids (liquid and vapor), and in the second model, the working fluid was assumed to be a mixture of three fluids (liquid, vapor and non-condensable gas). In both cases, we used a finite volume method to discretize the equations and SIMPLEC algorithm to link the pressure and velocity fields. An upwind scheme was used to model convective fluxes and other transport equations. Turbulence effects were considered using the k-ε model. Computations were performed at various inlet pressures and a fixed outlet pressure. The values of discharge coefficient obtained from the simulations were compared with published experimental data. Better agreement was found with the second model. This revealed the importance of non-condensable gases on cavitation. Furthermore, the distributions of vapor volume fraction and velocity magnitude were investigated with using both models. The results showed considerable differences between two models in description of inception of cavitation, distributions of vapor volume fraction and velocity magnitude.

Measurement of solute diffusivities. Part I. Analysis of coupled solute buoyancy-driven convection and mass transport

A solution phase-based nanocapsule method was successfully developed to synthesize non-platinum metal catalyst—carbon supported Ag nanoparticles (Ag/C). XRD patterns and TEM image show Ag nanoparticles with a small average size (5.4 nm) and narrow size distribution (2–9 nm) are uniformly dispersed on the carbon black Vulcan XC-72 support. The intrinsic activity and pathway of oxygen reduction reaction (ORR) on the Ag/C and commercial Pt/C were investigated using rotating ring disk electrode (RRDE) tests at room temperature. The results confirmed that the 4-electron pathway of ORR proceeds on small Ag nanoparticles, and showed comparable ORR activities on the self-prepared Ag/C and a commercial Pt/C. A single H2-O2 anion exchange membrane fuel cell (AEMFC) with the Ag/C cathode catalyst exhibited an open circuit potential of 0.98 V and a peak power density of 190 mW/cm2 at 80°C.

Large uncertainties in natural geomaterials hinder the practical use of theoretical predictive models in geotechnical engineering. While geophysical techniques have been employed to reduce the epistemic components of these uncertainties, most existing methods rely on sparse surface or borehole data and often use only partial seismic information, such as first-arrival times. Full-waveform inversion (FWI) offers a more comprehensive solution by utilizing the entire seismic record, but its high computational cost and complex implementation have limited its adoption in practice. This study introduces a novel inversion framework that combines FWI with Physics-Informed Neural Networks (PINNs) to provide a robust and computationally efficient tool for geotechnical site characterization. By embedding governing physical laws directly into the loss function, PINNs enable accurate subsurface property estimation using fewer data points and can generalize predictions beyond the training range. In our approach, unknown wave velocities are treated as trainable parameters within the network. We demonstrate the method using a 1D site response problem, where geometric and material properties are expressed through shear wave velocities of layers and their thicknesses. Synthetic training data are generated using a finite volume solver for the Navier–Cauchy equation, and the proposed FWI-PINNs framework is used to recover wavefield components and velocity profiles. With ongoing advances in GPU-accelerated machine learning, this method offers a promising path toward fast, accurate, and practical seismic imaging for geotechnical applications.

Abstract. The Earth's magnetosphere is populated by particles originating from the solar wind and the terrestrial ionosphere. A substantial fraction of the plasma from these sources are convected through the magnetotail lobes. In this paper, we present a statistical study of convective plasma transport through the Earth's magnetotail lobes for various geomagnetic conditions. The results are based on a combination of density measurements from the Electric Field and Waves Experiment (EFW) and convection velocities from the Electron Drift Instrument (EDI) on board the Cluster spacecraft. The results show that variations in the plasma flow is primarily attributed to changes in the convection velocity, whereas the plasma density remains fairly constant and shows little correlation with geomagnetic activity. During disturbed conditions there is also an increased abundance of heavier ions, which combined with enhanced convection, cause an accentuation of the mass flow. The convective transport is much slower than the field aligned transport. A substantial amount of plasma therefore escape downtail without ever reaching the central plasma sheet.

Application of SECM in tracing of hydrogen peroxide at multicomponent non-noble electrocatalyst films for the oxygen reduction reaction

A method for estimating the convective radon transport velocity in soils

The design and experimental characterization of a laboratory‐scale uniform injection cell (UIC) are presented. The experimental cell contains a disk electrode of radius embedded in the center of an insulating shroud. Parallel and coaxial with the disk electrode is an electrolyte injector consisting of a porous disk of radius that is also embedded in an insulating shroud. The counterelectrode is placed upstream of the porous disk injector. The thin gap between the parallel disks has a length between 0.01 and 0.15 cm. Electrolyte is injected with a (nominally) uniform velocity into the gap between the disks. Convective mass transport in the UIC was investigated for Reynolds numbers between 0.03 and 108 and was compared to theoretical predictions. Experimental results for low and intermediate flow rates agree with theory, but the highest‐flow‐rate data deviate somewhat. Measurements of the collection efficiency for a ring‐disk electrode placed in the UIC also agree with theory at low and intermediate flow rates, but deviate as the injection velocity exceeds . We suggest that the high‐flow‐rate deviations between theory and experiments may be due to nonuniform electrolyte injection.