The Receipt Address: A Bilayer Where the Fold Could Be Paid

What was measured

In a stack of two semiconducting sheets — molybdenum diselenide above, tungsten diselenide below, separated by a few atomic layers of insulating boron nitride — electrons can be held on one side and holes on the other. They bind across the barrier into excitons: neutral pairs that live long enough to be treated as a gas, and that can be pushed to condense into a single coherent state. The stack carries four such flavours, distinguished by spin and by which of two valleys in the crystal's band structure the carrier sits in.

A Berkeley group reports evidence, using magneto-optical spectroscopy in a dilution refrigerator, that at zero magnetic field the ground state is a coherent superposition of two intravalley flavours condensing at once — a two-component condensate rather than a single one. Raising the field switches that state to a two-component intervalley condensate through a first-order transition at a weak critical field, and at high field the state becomes a fully polarised single component. The two-component phases persist to about 1.8 kelvin. The paper says evidence for condensation, not settled proof, and this chapter repeats the word evidence for the same reason the paper chose it.

That is the whole of the established part, and it is worth pausing on how much it is. Two knobs — field and carrier density — tune a many-body state through a sequence of distinct phases, with the transitions located rather than inferred. Systems in which the controls are genuinely turnable and the state genuinely readable are rare, and every claim in the previous chapter needed exactly such a system.

Why it is a candidate address

The fold reading of Chapter 64 was left with a debt: it argued that a description which fails to close on itself may be reporting the shape of the parameter space rather than a missing ingredient, and it admitted that no measurement had yet been sent the bill. What it needed was a system with tunable controls, more than one mode in play, and a transition already located by someone else — so that a prediction could be made against a number nobody chose to accommodate it.

The bilayer offers all three. Two components means two modes. Field and density means a two-dimensional control plane in which loops can be drawn. And the switch at the weak critical field is a point in that plane whose location is already published. If the fold figure applies anywhere in condensed matter, this is the kind of place it would have to apply, and the address is specific enough to be wrong.

The reason it may be the wrong address

A first-order transition is not a coalescence. That sentence is the chapter's spine and it cuts against the reading. At a first-order quantum phase transition two competing states swap which one is lower; their descriptions remain distinct on either side, and the system jumps between them. At an exceptional point two states stop being distinct at all — the eigenvectors fuse and the basis becomes deficient. Those are different events, and the observable signature that distinguishes them is the one already named: square-root splitting near the point, and a state assignment that needs two circuits to return.

Worse, a first-order transition brings its own look-alike. It commonly shows hysteresis: sweep the field up and the switch happens at one value, sweep it down and it happens at another, so the state does not return on a single pass through the controls. That is failure-to-return without any fold at all. It is the cheapest possible imitation of the previous chapter's headline observation, and any honest use of the figure here must rule it out first rather than count it as support.

There is a further gap, and the conversation that prompted this chapter named it without being pressed. Nobody has written down the non-Hermitian operator whose coalescence would sit at the measured critical field. Excitons in this stack do have finite lifetimes, which is the physical ingredient non-Hermitian descriptions require, so such an operator is not forbidden. But until it exists on paper, with an exponent that reproduces a measured rate, the dictionary between the switch and the fold is a hope with an address attached, not a derivation.

What an address buys, and what it does not

It is worth being exact about what has improved between the two chapters, because the temptation is to feel that naming a system has done more work than it has. Nothing about the fold figure has become more likely. What has changed is that the figure can now be presented for payment somewhere specific, by someone with the apparatus, against numbers already in print. A figure with no address can be admired indefinitely; a figure with one gets a verdict.

This is the volume's standing discipline in its narrowest form. An abstraction that cannot fail did no work. The bilayer's value here is not that it supports the reading — as written, it mostly resists it — but that it converts a shape into a claim that can be settled, and settling includes losing.

And if it is lost, something is still gained, which is why the chapter is worth keeping either way. A demonstration that this switch is an ordinary competition between two distinct states, with hysteresis and linear splitting and no deficient basis, tells the volume that the fold family does not extend into equilibrium many-body phase transitions. That is a boundary on a borrowing, and boundaries on borrowings are the only thing this book has ever tried to accumulate.

What would settle it

Three measurements, in order of how much they cost. First, sweep the critical field in both directions and look for hysteresis; if the switch sits at two different field values depending on the direction of the sweep, the failure to return is thermodynamic and the fold is not needed. Second, measure how the separation between the two component energies grows as the field moves away from the critical value; a square root supports the reading, a straight line ends it. Third, and only if the first two survive, drive a slow closed loop in the field-and-density plane that encloses the critical point and ask whether the flavour assignment comes back after one circuit or requires two.

The reading is refuted for this system if the switch shows hysteresis, or if the splitting near the critical field grows linearly, or if a single slow loop enclosing the point returns the original assignment. Any one of the three is sufficient. If no non-Hermitian operator with a coalescence at the measured field can be written down at all, the dictionary is withdrawn for this system and the address is struck out rather than left standing as decoration.

The bill is now addressed. Whether it clears is not the author's to decide, and the chapter's usefulness does not depend on the answer.