Enhancing the Faradaic efficiency of solid oxide electrolysis cells: progress and perspective
Summary¶
Solid oxide electrolysis cells promise efficient hydrogen production at high temperature, but Faradaic efficiency losses from parasitic reactions, microstructural degradation, and transport limitations remain practical barriers. Gaikwad, Shin, and van Duin review how electrode, electrolyte, and interconnect materials, together with operating conditions, influence Faradaic efficiency in SOEC stacks. The article synthesizes experimental and modeling literature, highlighting multiscale simulation opportunities—including atomistic reactive modeling along ReaxFF-related lineages where chemistry at oxide surfaces matters—for diagnosing loss channels and guiding materials design. The review is positioned for computational materials audiences who must couple device metrics with atomistic mechanisms. It also stresses that Faradaic efficiency is an integrative metric: losses can originate in any stack layer, so localized fixes require systems thinking.
The review’s Methods stance is explicitly literature synthesis: it compares how experimental gas analytics, electrochemical impedance, microscopy, and multiscale models each constrain where Faradaic losses arise (electrode kinetics, electrolyte leakage, interconnect oxidation, steam starvation, etc.), rather than presenting one new in-house benchmark cell.
Methods¶
Review organization (D)¶
- Axes: Stack components (anode, cathode, electrolyte, interconnect), operating conditions (T, current density, steam fraction), and degradation modes tied to Faradaic efficiency losses.
- Literature types synthesized: Experiments, continuum transport/reactor models, kinetic Monte Carlo, electronic-structure studies, and atomistic surface models (including reactive approaches where cited). Comparative protocol (implicit): the review asks how local loss channels and interfacial chemistry in cited works map onto cell-level Faradaic metrics—not a new kinetic benchmark in one simulation code. N/A — the article does not introduce a single DFT/MD/continuum training dataset; it surveys primary studies. N/A — the Methods of cited experiments (gas analytics, EIS, T, p) vary by source; the review ties them to phenomenological categories of inefficiency.
Modeling chapter structure¶
Contrasts continuum mass/charge transport with atomistic elementary-step resolution; no single new benchmark simulation is central to the review.
Corpus PDF note¶
The ingested file may be a galley; layout can differ from final npj Computational Materials typesetting. N/A — this page is not a laboratory Methods log; VOR should be used for final section pagination.
Findings¶
Problem framing¶
Low Faradaic efficiency and high $/kg H\(_2\) persist as practical barriers despite favorable thermodynamic efficiency vs some low-T electrolyzers.
Mechanisms and modeling gaps¶
Parasitic chemistry, microstructural evolution, and transport limits siphon current from the oxygen evolution / hydrogen product channels; the review maps these to multiscale modeling opportunities (DFT, kMC, continuum, microstructure).
Outlook¶
Emphasizes integrated workflows rather than isolated DFT snapshots; Faradaic efficiency should be tracked alongside microstructure evolution, not current density alone.
How to use this page safely¶
Quantitative efficiency values are second-hand—verify gas compositions, seals, and operating windows in each cited primary study before reuse in MAS claims. Comparisons between cited SOEC and protonic/other oxide systems follow the primary references, not new fits in this review. Open directions the authors highlight (integrated continuum+atomistic loss accounting) are opinion-level outlook—treat as review synthesis with caveats.
Limitations¶
As a review, quantitative values are second-hand; consult primary references for uncertainties and experimental conditions.
Relevance to group¶
Positions Shin and van Duin within hydrogen and solid-oxide modeling discourse; complements application-focused articles such as paper:2023gaikwad-journal-of-t-modeling-dynamic where present in the corpus.