What You Are Building
You are designing an energy storage system that is meant to be charged, then held at state of charge for hours to days, then discharged, over and over, for the working life of a piece of grid infrastructure. This is the "multi-day duration" regime: the discharge duration your system must deliver is measured in tens of hours, and the gap between charge and use can be measured in days.
The people who have this problem are grid planners, microgrid integrators, and anyone sizing storage for renewable firming, multi-day resilience, or seasonal shifting. The hard part is rarely peak power. It is the combination of three things at once: a low cost per stored kilowatt-hour, a low rate of charge loss while sitting idle, and a life long enough that the capital cost amortizes over decades.
This guide teaches an architectural approach to that combination. It is grounded entirely in a filed patent disclosure. You will not download it; you will build it. The point here is to understand the design well enough to evaluate it and to start engineering it yourself.
Why the Obvious Approaches Fall Short
Almost every rechargeable cell in wide use, including lithium-ion, sodium-ion, lead-acid, and nickel-metal-hydride, shares one structural feature: an anode and a cathode held at different electrochemical potentials, separated by an ion-conducting, electron-insulating separator or membrane. That separator is the load-bearing component for holding charge. The two electrodes "want" to equalize, and the only thing stopping them is that the separator blocks electrons from crossing internally. This is an accurate and standard description of how these chemistries work; it is not a weakness invented for contrast. It is simply the design point.
Two consequences follow for multi-day duration. First, charge retention is a race against every pathway that lets the electrodes equalize anyway. Idle charge loss comes from real, well-characterized processes: electrolyte decomposition at the electrode interfaces, growth of solid-electrolyte-interphase layers, dissolution of cathode metals, and parasitic side reactions with trace contaminants. For lithium-ion cells at room temperature, these drive calendar capacity loss on the order of one to five percent per year, whether or not you cycle the cell. Over a multi-day hold that is small; over a multi-decade deployment it compounds into a real cost.
Second, cost and duration pull against each other. Cheap chemistries tend to have shorter life or higher self-loss; long-life chemistries tend to cost more per kilowatt-hour. Flow batteries answer duration by separating stored energy (tank size) from power (stack size), which is a genuinely good structural fit for long duration, but they add plumbing, pumps, and separated electrolyte management. The structural gap the guide addresses is different: it removes the separator as the retention mechanism altogether, so that idle retention no longer depends on insulating two mismatched electrodes from each other.
The Architecture
The disclosed cell is a single sealed enclosure with two carbon current collectors at opposite interior faces and, between them, a continuous bulk volume of a conductive carbon gel with metal nanoflakes dispersed through it. There is no internal separator, no membrane, and no internal barrier between the two collectors other than the gel itself. Every load-bearing idea below traces to that arrangement.
Charge is retained by saturation, not by insulation. In the charged state with no load connected, all the metal nanoflakes sit at substantially the same electrochemical potential. Because the gel is both electronically and ionically conductive, any local potential difference equalizes quickly, so no internal gradient exists to drive charge from one place to another. The cell holds its charge not because current is blocked but because there is no internal driving force for current to flow. Energy is stored in metal-hydrogen surface bonds distributed across the whole flake population, not in a separated electrode pair. This is the "bulk-equipotential" principle, and it is the reason the architecture is interesting for long holds: the retention mechanism is the absence of an internal potential difference, rather than the integrity of an insulating layer.
Discharge is created by the external circuit, not resisted internally. The equipotential state is broken only when you close an external circuit through a load. Closing it lets electrons exit one terminal, cross the load, and re-enter the other, which sets up a gradient inside the gel that drives stored charge toward the exit terminal. The cell discharges only as fast as the load allows; at open circuit, per the disclosure, it does not discharge at all. For a multi-day system this is the property you care about: idle time is genuinely idle.
Energy is stored as surface-bonded hydrogen on metal flakes. During charging, a proton from the gel and an electron from the charging circuit combine at a flake surface to form an atomic-hydrogen bond to a surface metal atom. Discharge reverses this: the load withdraws the bonding electron, the bond destabilizes, and the hydrogen returns to the gel as a proton. The disclosure frames these metal-hydrogen bonds as a kinetically trapped state that is stable at rest, requiring no continuous bias to maintain. The underlying surface chemistry, chemisorption of atomic hydrogen on metals such as aluminum, is pre-existing, characterized science; the novelty claimed is the architecture that puts it inside a separator-free equipotential cell, not the discovery of the effect.
Idle degradation pathways are engineered out rather than merely slowed. The disclosure attributes low idle charge loss to two things working together: the equipotential retention above, and the absence of internal degradation chemistries that operate at rest. The disclosure describes the gel as serving simultaneously as electrolyte, structural matrix, and reservoir for the mobile carbon that participates in flake healing and expansion during cycling, while the hydrogen storage bond itself lives on the flake surfaces rather than in the gel; because those gel roles are exercised by cycling and not by sitting charged, the storage state at rest does not consume the gel. On that basis the disclosure projects a "long-term storage mode" with calendar self-discharge well below one percent per year, and a long-storage example projecting roughly 0.1 to 1.0 percent per year, expressly noting these are projections to be confirmed by long-duration testing, not demonstrated results.
The enclosure participates in retention. In the preferred embodiment the innermost enclosure layer is aluminum in ohmic contact with the gel and electrically isolated from both terminals, so it floats at the gel's bulk potential. The disclosure describes this as extending the equipotential volume to the case, which removes a parasitic capacitive-leakage path at the enclosure wall, while also acting as a Faraday shield and, via its native oxide, an oxygen barrier holding interior oxygen below about ten parts per million. For a multi-day design these are all retention-relevant: fewer leakage paths and a tightly sealed, low-oxygen interior.
It is built to scale up and to be rebuilt. The disclosure explicitly admits thicker aluminum inner layers for large stationary cells and names grid-scale and stationary storage as target applications. It also describes end-of-life remanufacturing, recovering and regenerating the internal materials into a new cell, as the path to multi-decade cumulative life. Note carefully: this cell is fully sealed with no field-service ports; the disclosure states the internal contents are not designed for in-place replacement, and end-of-life recovery is done by full remanufacturing, not by refilling a deployed cell.
How to Approach the Build
You are implementing an architecture, not installing a package. A realistic path:
Fix your duration and self-discharge budget first. Decide the discharge duration (for example, a system that must supply rated power for many tens of hours) and the maximum tolerable charge loss across your longest expected idle hold. The architecture's appeal is idle retention, so make the idle-loss target the number you design against, and treat the disclosure's sub-one-percent-per-year figure as a projection you must verify, not a spec you can assume.
Choose the cell chemistry parameters from the disclosed ranges. The active metal is selected for hydrogen surface chemisorption; aluminum is the preferred embodiment for cost, low mass, and abundance. Flake loading, fractal flake morphology, and the optional dynamic-expansion mechanism trade energy density against complexity. Use the disclosed ranges as your design envelope, and frame every number as "as disclosed / projected." Do not invent densities, cycle counts, or efficiencies the disclosure does not state.
Treat the gel synthesis as the hard engineering. The disclosure is explicit that the principal manufacturing complexity lives in the dual-domain proton-conducting carbon gel: a boron-doped, statically-reduced turbostratic graphene matrix, functionalized into a hydrophilic proton-conducting domain and a hydrophobic hydrogen-rejecting domain, with the nanoflakes self-distributing into the hydrophobic domain by acquired surface charge. Everything else (enclosure, collectors, terminals) is described as conventional. Your program plan should reflect that: the gel is where the risk and the research sit.
Design the sealed, low-oxygen, floating-case enclosure. Plan a hermetic enclosure with an aluminum inner layer in ohmic contact with the gel and dielectrically isolated from the terminals so it floats at bulk potential. Size the inner-layer thickness toward the upper end for large stationary cells. Assemble and cast the gel under inert atmosphere, seal, then inject hydrogen through a port that is permanently closed afterward, since the disclosure notes hydrogen will not stay in an unsealed gel.
Push retention into the pack and system layers too. Because a single cell holds charge by equipotential saturation, your gains survive only if pack wiring, balancing electronics, and monitoring do not add their own idle draw. A parasitic milliamp in the battery-management system will dominate a cell whose intrinsic idle loss is projected below one percent per year. Budget standby power at the system level with the same discipline you apply to the cell.
Plan for remanufacturing, not field refill. Because the sealed cell is not serviceable in place, your multi-decade economics depend on an end-of-life recovery and remanufacture loop. Design the enclosure and collectors for disassembly and reuse from the start.
A minimal, illustrative interface sketch (faithful to the disclosure's operating modes, not runnable code):
# ILLUSTRATIVE ONLY - not a library, not benchmarked
class LongDurationCellModel:
def hold(self, soc, days):
# idle retention is projected, not measured;
# equipotential saturation => no internal driving force at open circuit
return soc # minus a projected calendar loss you must verify empirically
def discharge(self, load):
# cell delivers only as fast as the external load draws
...
Treat that as a specification stub you fill in against your own measurements, not as behavior the architecture guarantees.
What This Does Not Give You
This is an architecture, not a drop-in product. There is no SDK, no reference cell you can order, and nothing here "just works." You are the one who has to synthesize the gel, build the enclosure, and validate the result.
Nothing in the disclosure has been built, benchmarked, or production-proven; it is a provisional filing of a design. Every performance figure it states, energy density, cycle life, round-trip efficiency, self-discharge, is a projection that the disclosure itself flags as requiring empirical confirmation through long-duration and prototype testing. Do not carry those numbers into a bid or a datasheet as if they were measured.
The materials science underneath, hydrogen chemisorption on metals, proton-conducting carbon gels, turbostratic graphene, mechanochemical effects, is pre-existing and prior art. The claimed novelty is the combination and the resulting new category of separator-free equipotential cell, not any newly discovered physics. Treat the basic effects as known and the integration as the thing to engineer.
Finally, this architecture is aimed at stationary, long-hold, multi-decade deployments where idle retention and rebuildability matter most. The disclosure also describes short-duration high-power and cycling modes, but if your problem is, say, a small portable device or a purely power-limited application, a different design may serve you better. Match the architecture to the multi-day duration problem it was described for.
Disclosure Scope
The approach described in this guide is disclosed in U.S. Provisional Application No. 64/055,649, the Hydrogen-Aluminum Energy Cell inventive step. This guide is educational: it explains the disclosed architecture so a skilled developer can understand and attempt to build it. It is not a warranty, not a performance guarantee, and not an offer of software or of a finished product. All quantitative characteristics referenced here are drawn from that filing and are projections stated therein, to be confirmed by the reader's own testing.
