The charge asymmetry: why the buffering direction matters

By Hemant K. Rohera — ORCID 0009-0005-3275-1743
Published 22 September 2026 · Last updated 22 September 2026

Hybrid storage is usually sold on the discharge side: the capacitor supplies the surge, the battery is spared. That is real, but it is the weaker half of the argument. The hardest failure boundary inside a lithium-ion cell sits on the charge side, and it is close — a couple of hundred millivolts at best, and far less when the cell is cold. A buffer that absorbs a charge transient is working against a boundary that actually exists. A buffer that supplies a discharge transient is not. They should never be presented as one benefit.

Where the boundary is

Charging a graphite anode means driving lithium ions into the graphite lattice. That reaction has a potential, and so does the reaction that competes with it — plating lithium as metal on the particle surface instead of inserting it.

Chen and colleagues state the two standard potentials explicitly for a graphite electrode against a lithium reference: Literature intercalation to form LiC₆ sits at 0.20 V and lithium plating at 0 V vs Li/Li⁺ [1]. The staging transitions that carry most of the capacity run lower still, with the stage 3 → 2 and stage 2 → 1 transitions appearing between roughly 0.05 V and 0.07 V [1]. The review of record on plating puts the boundary the same way: Literature plating occurs once the graphite electrode potential reaches 0 V vs Li/Li⁺ and below [2].

So the entire useful working window of the negative electrode lies in a band of roughly 50 to 200 millivolts above a boundary it must not cross. Derived — that range is the difference between the intercalation potentials in [1] and the 0 V plating potential in [1] and [2]. There is no large margin to spend.

The asymmetry itself

Anode overpotential has a sign. Under load — discharge — the anode potential moves up, away from 0 V. Under charge it moves down, toward it. The same magnitude of current produces opposite outcomes with respect to the one boundary that matters:

A capacitor branch that absorbs a charge transient — a regenerative-braking pulse, a cloud-recovery edge on a photovoltaic input, a fast-charge current step — is operating in the direction where the failure boundary lies. A branch that supplies a discharge transient is operating in the direction where it does not.

This is why the two cases cannot be collapsed into one claim. Discharge-side buffering earns its benefit through the thermal and resistive route set out in the companion page on resistive loss and ageing rate. Charge-side buffering earns that same benefit and acts on a distinct electrochemical failure mode that current reduction is the direct remedy for.

What narrows the margin

Three things push the anode toward 0 V: low temperature, high charge rate, and high state of charge. All three raise the overpotential required to insert lithium at a given current [2].

The temperature term is the sharp one, and it has been measured. In three-electrode cells, Literature the graphite potential crossed 0 V vs Li/Li⁺ at 412 mAh·g⁻¹ of charge passed at −20 °C, but at only 160 mAh·g⁻¹ at −30 °C and 106 mAh·g⁻¹ at −40 °C [1]. The cell reaches the plating boundary earlier and earlier in the charge as it gets colder.

The asymmetry has a second, quieter cause. Literature across −40 °C to 30 °C the lithium plating potential is largely insensitive to temperature, while the intercalation potential shifts by as much as 60 mV [1]. The boundary stays where it is; the working window moves toward it.

Why plating is worth designing against

Plating is damaging twice over. The deposited metal consumes cyclable lithium, which is capacity lost permanently. If it grows dendritically it creates an internal short pathway, which is a safety mechanism rather than a performance one [2].

It is also quiet. Plating is not signalled by terminal voltage or surface temperature at the moment it happens, which is what makes it a hard limit rather than a soft one — a cell can be plating while every quantity a pack controller measures looks normal. That is an argument for controlling the current waveform at the architecture level rather than detecting the consequence in firmware.

One further consequence, because it is routinely got wrong in accelerated testing: Literature post-mortem analysis of commercial cells shows the dominant ageing mechanism changes across the temperature range, plating-dominated below roughly room temperature and interphase- and cathode-dominated above it [3]. An Arrhenius acceleration factor is only valid if the same mechanism dominates at both temperatures, so a life figure extrapolated across that crossover is a hypothesis, not a result.

Where this argument stops

Three limits, stated plainly, because an argument that is not bounded is not an argument.

A discharge pulse is not a remedy for a charge-side risk. Cold cranking is a discharge event. Cold charging is what plates lithium. Offering the first as protection against the second does not connect, and any document that does so is wrong — including earlier documents of mine. Cold regenerative braking and DC charging below 0 °C are the cases that need answering, and they are answered by limiting the current into the cell.

Current reduction does not reach most other ageing paths. Interphase growth continues on the calendar, driven by temperature and mean state of charge. Particle cracking is driven principally by depth of discharge. Transition-metal dissolution and electrolyte oxidation are driven by temperature and electrode potential limits. A capacitor branch touches those only indirectly, through temperature. Anyone claiming a buffer addresses all degradation is overstating it.

The effect on this architecture is unmeasured. Design estimate — whether the Hybrid Power Pack measurably reduces plating on real charge transients has not been tested. The experiment that would settle it is a low-temperature charge-acceptance comparison against a battery-only control of the same chemistry and capacity, with dV/dQ or post-mortem confirmation that plating is what changed. Until that is run, the claim on this page is a claim about direction, resting on published electrochemistry, and not a performance figure.

What the architecture has to do to use this

Absorbing a charge transient is not the same capability as supplying a discharge one, and not every hybrid topology has it. A capacitor branch can only take current in if its coupling to the bus is bidirectional.

That is what the granted claim provides. In US 10,523,019 B2 the second storage component is coupled to the input/output port via a bidirectional switch, to provide a charging path as well as a discharging path, while the battery branch is coupled through two unidirectional switches that separate its charge and discharge paths, and a controller draws sustained power from either component or both [4]. Derived — a capacitor branch with a charging path of its own is, by construction, able to take the charge transient the cell would otherwise take. Whether it does so to a degree that matters is the measurement above, still outstanding.

Summary

The plating boundary is real, it is close, and it lies on one side only. That single fact reorders the usual argument for hybrid storage: the discharge-side case is thermal and gradual, while the charge-side case acts on a hard electrochemical limit that narrows with cold and with rate. Both are worth making. They are not the same claim, and the stronger one is the one most often left out.

References

[1] Chen, B., Hope-Glenn, N., Wright, A., Messinger, R. J. & Couzis, A. (2025). Mechanistic understanding of lithium-ion adsorption, intercalation, and plating during charging of graphite electrodes. ACS Electrochemistry, 1(5), 574–587. Open access. doi:10.1021/acselectrochem.4c00079
[2] Waldmann, T., Hogg, B.-I. & Wohlfahrt-Mehrens, M. (2018). Li plating as unwanted side reaction in commercial Li-ion cells — A review. Journal of Power Sources, 384, 107–124. doi:10.1016/j.jpowsour.2018.02.063
[3] Waldmann, T., Wilka, M., Kasper, M., Fleischhammer, M. & Wohlfahrt-Mehrens, M. (2014). Temperature dependent ageing mechanisms in lithium-ion batteries — A post-mortem study. Journal of Power Sources, 262, 129–135. doi:10.1016/j.jpowsour.2014.03.112
[4] US 10,523,019 B2, Hybrid power pack, granted 31 December 2019. Claim 1. patents.google.com