The electric double layer, and what it cannot do
A supercapacitor is not a better battery and it is not a bigger capacitor. It is a conventional capacitor in which the two plates have been brought absurdly close together and given an absurdly large area — and the same trick that makes it remarkable also caps what it can ever store. The limit is not an engineering problem awaiting a breakthrough. It is in the geometry.
What the layer is
Put an electrode in an electrolyte and apply a potential. Ions of opposite sign are drawn towards the electrode surface and arrange themselves against it, held at a standoff distance set by their own size and their shell of solvent molecules. The result is two sheets of opposite charge separated by a gap of the order of a nanometre: a capacitor, formed by nothing more than the interface itself. Helmholtz described the arrangement in the 1850s; the refinements that followed — Gouy and Chapman’s diffuse layer, Stern’s combination of the two — are about how sharply the charge is localised, not about whether the layer exists.
Nothing reacts. No ion enters the electrode lattice, no oxidation state changes, no phase transforms. Charging and discharging is ions arriving at a surface and leaving it again. That single fact explains everything that follows, good and bad.
Why the capacitance is so large
For a parallel-plate capacitor, C = εA/d — capacitance rises with the permittivity of the medium and the plate area, and falls with the separation. The double layer attacks both of the terms that matter:
- d is about a nanometre, rather than the micrometres of a film capacitor. Three orders of magnitude, for free, because the geometry is set by molecular dimensions rather than by manufacturing.
- A is enormous. The electrode is porous activated carbon whose internal surface area is commonly quoted in the region of 1,000 to 2,000 square metres per gram. A few grams of electrode carries the area of a tennis court.
Multiply a very large area by a very small separation and the capacitance of a small device lands in farads rather than microfarads. That is the whole trick, and it is why the name "supercapacitor" is accurate in a way that marketing names usually are not.
And now what it cannot do
The energy stored in a capacitor is E = ½CV². Capacitance is large, so the first term is generous. The problem is the second one.
1. The voltage is capped by chemistry, not by design. Raise the potential across the interface far enough and the electrolyte decomposes. That ceiling is a property of the solvent and salt, not something a better design can negotiate: organic-electrolyte devices are commonly rated around 2.7 volts per cell, aqueous ones far lower. Because energy goes as V², a ceiling on voltage is a hard ceiling on energy — and it is a more severe constraint than the capacitance is a generous one.
2. Only the surface stores anything. A battery electrode stores charge throughout its volume; the double layer stores it on an interface. However much area is folded into a gram of carbon, it is still area rather than volume. This is the structural reason specific energy lands roughly an order of magnitude below a lithium-ion cell — commonly quoted at around 5 to 10 watt-hours per kilogram against 150 to 250 — and it is why no amount of process improvement will close that gap while the mechanism stays electrostatic.
3. The voltage falls the moment you use it. Since V = Q/C, taking charge out lowers the terminal voltage in direct proportion. A battery holds a fairly flat plateau across most of its capacity; a capacitor slides steadily downhill from full. Recovering even three-quarters of the stored energy means discharging to half the rated voltage, which means the load must tolerate a two-to-one input swing — or, in practice, a converter must sit in between. That converter has mass, cost and losses of its own, and it is a recurring theme in why pairing the two devices is not simply adding them.
4. It does not hold charge for long. Self-discharge in these devices is measured in days rather than the months typical of a cell. For buffering a transient that is irrelevant; for storing energy overnight it is disqualifying.
5. Series strings need balancing. A single cell of 2.7 volts is rarely enough, so cells are stacked — and small differences in capacitance and leakage between them mean the voltage does not divide evenly. Left alone, one cell drifts above its rating and degrades. Balancing circuitry is not optional, and it is part of the true cost of the bank.
What it is therefore good for
Everything the list above rules out is a storage job. What is left is the delivery job, and at that the double layer is superb: because no material has to transform, charge can enter and leave as fast as the ions can move and the conductors allow, and it can do so hundreds of thousands of times without the electrode degrading. Cycle lives quoted at a million are not marketing — they follow from the fact that nothing is being worn out.
So the honest summary is narrow. A supercapacitor is a poor place to keep energy and an excellent place to put a transient. Pairing it with a cell is an attempt to use it only for the second of those, and the argument for doing so is the arithmetic in the note on I²R — that for a fixed quantity of charge, the heat wasted in a cell is proportional to the current used to move it.
The figures above, and their status
The numbers in this note — nanometre-scale separation, surface areas of 1,000 to 2,000 m²/g, cell ratings around 2.7 V, specific energies of 5 to 10 Wh/kg — are the values commonly published for commercial devices and in the standard literature on electrochemical capacitors. They vary by electrode material, electrolyte, format and measurement convention, and they are given here as orders of magnitude to support an argument about mechanism, not as specifications for any product. Nothing in this note describes a measurement made on the Hybrid Power Pack, and nothing in it should be read as a performance claim for one. What has and has not been measured on that architecture is set out in the open-questions list.