Electrochemical reaction kinetics at electrode–electrolyte interfaces can constrain how quickly batteries charge and discharge. For engineers modeling these systems, Marcus-Hush-Chidsey kinetics offer a way to account for electron-transfer behavior beyond the commonly used Butler-Volmer approach. But a key modeling assumption can limit their usefulness: treating the electrode’s electronic density of states as constant across energy. Incorporating the electrode density of states (DOS) can produce similar predictions for some materials and substantially different ones for others.
The difference matters when evaluating interfaces for lithium metal batteries, including lithium deposition on a copper current collector and reactions through a solid-electrolyte interphase (SEI). The material’s electronic states, the location of its Fermi level, and the reaction’s reorganization energy all influence the calculated rates.
Why include electrode density of states in Marcus-Hush-Chidsey kinetics?
Marcus-Hush-Chidsey (MHC) theory describes electron-transfer rates by integrating contributions from electronic states, accounting for their occupation and the energy associated with molecular reorganization. In conventional applications, the electrode-state contribution is often treated as weakly dependent on energy and factored out of the rate calculation. That simplification can work when the DOS is relatively flat over the relevant energy range.
However, electrode materials do not all have the same electronic structure. Transition metals can have pronounced features associated with localized d-bands, while semiconductors have bandgaps where the DOS is absent. If these variations are ignored, a model may miss how available electronic states affect the reaction rate as overpotential changes.
A DOS-aware MHC calculation incorporates the electrode DOS directly into the rate integrals. In the reported analysis, the DOS was calculated using density functional theory (DFT) and evaluated numerically. This approach provides a way to represent material-specific electronic structure rather than assuming a single, energy-independent contribution.
Material-specific effects on lithium reaction rates
Lithium metal: a relatively small difference
For lithium deposition and stripping on a lithium surface, the calculated DOS is relatively flat across the range relevant to the comparison. As a result, the traditional MHC treatment and the DOS-aware version produce similar rate predictions over the experimental overpotential range examined. Differences become more noticeable over a wider range, but remain modest compared with the other cases.
This example shows why a constant-DOS approximation may appear adequate in some systems: when the electronic structure changes little across the energies that contribute to the reaction, adding that variation has limited effect on the result.
Copper current collectors: d-bands influence predicted rates
The picture changes for lithium deposition on copper, a configuration relevant to anode-free lithium metal batteries. Copper’s DOS includes a high density of states associated with d-bands, creating a substantial energy dependence that a constant-DOS model does not capture.
In the modeled Cu(111) case, the DOS-aware calculation predicts a pronounced change in reaction rate as the overpotential reaches the energy range associated with those states. The paper reports an approximately order-of-magnitude rate-constant change near the onset of the high-DOS region. The practical modeling implication is that copper’s electronic structure can influence predicted deposition kinetics, particularly when examining larger overpotentials. A simplified model may therefore miss a material-specific feature relevant to high-rate operation.
Semiconducting SEI interfaces and reaction asymmetry
An SEI adds another level of complexity. The analysis considered a model interface formed from lithium fluoride (LiF) and lithium carbonate (Li2CO3), representative of inorganic components that can occur in an SEI. These materials are semiconducting or insulating, so the DOS includes a bandgap rather than a continuous supply of electronic states near the Fermi level.
For the modeled interface, the pristine equilibrium Fermi level lies near the middle of a gap of roughly 4 eV. Because the DOS is zero through the gap, accounting for it leads to rate predictions that differ dramatically from the constant-DOS model—even at low overpotentials. This illustrates why a flat-DOS assumption can be particularly unsuitable for semiconductor–electrolyte interfaces.
The position of the Fermi level also matters. In an electrochemical system, the redox reaction at the interface sets that level; it should not always be treated as fixed at the position calculated for a pristine material. When the modeled Fermi level shifts toward the valence-band edge, the predicted rate becomes asymmetric with overpotential. Moving farther into the valence band increases the rate sharply, while moving back into the gap suppresses it before rates rise again as the conduction-band edge becomes relevant.
Model parameters and practical considerations
Two additional considerations help interpret DOS-aware predictions:
- Fermi-level position: Changing the equilibrium Fermi level alters which electronic states contribute to the reaction. This can change both the rate and its dependence on charge or discharge direction.
- Reorganization energy: In the rate calculation, reorganization energy acts in part like a Gaussian smoothing function applied to the DOS. Increasing it smooths the variation in the predicted rate with overpotential and shifts the relationship between the rate and the energy states being sampled.
These effects mean that a rate curve cannot be interpreted from the material’s DOS alone. The relevant energy alignment and kinetic parameters must also be considered. The analysis further notes that the coupling between electrolyte states and electrode states can have energy dependence, although the main focus is explicitly incorporating the electrode DOS.
Implications for battery modeling
For electrochemical modelers, the central lesson is to assess whether the constant-DOS approximation is appropriate for the interface being studied. A relatively flat DOS, as in the lithium-surface example, may yield only minor changes. Strongly varying metallic states or a semiconductor bandgap can make the approximation consequential.
Including material-specific DOS can help distinguish predicted behavior across electrode and interphase materials, and can reveal charge–discharge asymmetry tied to Fermi-level alignment. The reported work provides open-source analysis code and an interactive visualization for exploring how parameters such as DOS, Fermi level, reorganization energy, and temperature affect predictions.
Conclusion
Marcus-Hush-Chidsey kinetics become more representative of a particular electrode interface when the electrode’s energy-dependent density of states is included. The effect may be limited for lithium metal, substantial for copper with its d-band features, and especially pronounced for a semiconducting SEI with a bandgap. For battery engineers and model developers, the practical step is to treat electronic structure and Fermi-level alignment as potential contributors to interfacial reaction rates—not automatically as constant background factors.