The Phonon Bound:
A 17% Efficiency Suppression Hidden in Cold-Atom Analog Gravity

This article is a companion to Paper 10 — “A Non-Equilibrium Efficiency Bound for Phonon Extraction in Bose–Einstein Condensate Analog Gravity Systems with Numerical Tests of the Underlying Thermodynamic Assumption” (2026). The full manuscript is at github.com/Windstorm-Institute/phonon-extraction-bound. This paper opens the Institute’s second research track — Entropic Bounds in Analog Systems — alongside the nine papers of the Throughput Basin arc.

In 2011 Erik Verlinde wrote a short, audacious paper arguing that gravity is not a fundamental force at all. It is, he said, an entropic force — the thing you feel when you try to push a system away from a state with more microscopic configurations toward one with fewer. Like the elastic snap of a stretched polymer chain, gravity in this view is the universe’s tendency to maximize entropy on holographic screens.

It was a beautiful idea. It recovered Newtonian dynamics from first principles. It connected to Jacobson’s 1995 derivation of Einstein’s field equations from the Clausius relation. And it had one fatal feature, never quite acknowledged: it was untestable.

Not because the math was wrong. Because every setting where gravity matters — planets, stars, galaxies, black holes — puts you in a thermodynamic regime where the prediction collapses into the trivial second law. The local Unruh temperature near Earth is something like 4×10⁻²⁰ kelvin. The combustion chamber of a Saturn V is 3500 kelvin. The ratio is 10⁻²³. Every entropic-gravity efficiency bound, in that regime, just says “efficiency ≤ 1”. Of course it is. So is every other thermodynamic process in the universe.

The trick is to find a setting where the ratio is order one

That’s what this paper is about.

Cold-atom physicists have spent the last twenty years building analog gravity systems — Bose–Einstein condensates with engineered density gradients that produce, for sound waves traveling through them, an effective curved spacetime. The acoustic horizon of such a system has its own analog Unruh temperature. And here is the key: in the lab, that analog Unruh temperature can be made comparable to the temperature of the condensate itself. Not 10⁻²³ apart. Not 10⁻¹⁰ apart. About 0.2 apart — right in the regime where any entropic-gravity efficiency bound becomes restrictive.

So we asked the question Verlinde never quite asked: given his screen-entropy expression and standard non-equilibrium thermodynamics applied to a two-temperature process, what is the maximum extraction efficiency, and where can it be tested?

The bound

Apply the Clausius inequality to a process that exchanges heat with a reservoir at temperature T_res while depositing heat on a holographic screen at the local Unruh temperature T. Use Verlinde’s screen-entropy expression. Make one substantive assumption — that the screen exchanges heat reversibly with its immediate environment, while the coupling to the reservoir carries the irreversibility. After two pages of straightforward bookkeeping, you get:

That’s the central result. In the astrophysical regime where T/T_res is essentially zero, the right-hand side collapses to 1 — the trivial second-law statement. In the BEC regime where T/T_res ≈ 0.2, the right-hand side is 1/1.2 ≈ 0.83 — a 17% suppression below what naive energy accounting would predict.

Seventeen percent is a lot. It’s well above the systematic-error floor of contemporary BEC phonon-manipulation experiments. It is, plausibly, measurable.

The honest part: there’s one assumption holding the whole thing up

The substitution above — that the screen exchanges heat reversibly while the reservoir coupling carries the irreversibility — is not a theorem. It is an assumption. We labeled it Assumption A and treated it as load-bearing throughout the paper. The empirical content is conditional on Assumption A. The BEC prediction follows under it. The proposed laboratory experiment tests the assumption and the construction simultaneously.

An honest paper, when it has a load-bearing assumption, doesn’t hide it. It tests it.

So we did. Five times.

Five tests of Assumption A

We built a series of QuTiP Lindblad master-equation simulations — standard open-quantum-system numerics — in which the substitution δQ = T · dS could be checked directly against numerically computed heat and entropy changes. Each test probed a different dimension along which the assumption might fail.

Test 1 — Timestep convergence. A single qubit coupled to a thermal bath. Initial states thermal at T = r · T_res for r ∈ {0.05, 0.10, 0.20, 0.30, 0.50}. Timesteps from 10⁻² down to 10⁻⁵. Result: for r ≥ 0.10 the ratio δQ / (T · dS) converges monotonically to 1.000 within 0.1%. Power-law convergence analysis gives clean O(dt) scaling for r ≥ 0.20. The first-order Trotter error in the integrated Lindblad evolution is exactly what theory predicts.

Test 2 — Bosonic mode (truncated harmonic oscillator). The qubit is replaced by a 10-level harmonic oscillator — the natural model for a phonon mode in a BEC. Result: indistinguishable from the qubit case in the same parameter regime. This is the most physically relevant generalization for the BEC application. The substitution holds with the same convergence structure.

Test 3 — Strong coupling. The same qubit setup with the system–bath coupling γ/ω pushed from 0.05 (textbook weak coupling) up to 2.0 (where standard Markovian assumptions become questionable). Result: for moderate temperature ratios r ≥ 0.30 the substitution remains exact across the entire coupling sweep. The BEC application sits comfortably in the weak-coupling regime, so this constraint does not affect the paper’s claim within its stated scope.

Test 4 — Two-bath driven non-equilibrium steady state. The hardest test conceptually: a qubit coupled simultaneously to a hot bath and a cold bath, driven into a non-equilibrium steady state. Heat currents computed directly from the dissipator structure rather than inferred from energy differences. Result: energy conservation closes to one part in 10¹⁶. The ratio Q_in / (T_eff · dS_hot) equals 1.000 exactly across all coupling configurations tested. This addresses the most important conceptual concern in the original derivation — that Assumption A was being tested only in relaxation, not in genuine driven non-equilibrium settings.

Test 5 — Non-thermal initial states (the negative result). The qubit is prepared in a coherent superposition |+⟩ with zero von Neumann entropy and maximum off-diagonal coherence. Result: the ratio takes values 1.37, 0.43, 0.21 at r = 0.10, 0.30, 0.50 — far from 1, strongly r-dependent, and inconsistent with Assumption A. Assumption A fails for non-thermal initial states with macroscopic quantum coherences.

This is a meaningful constraint and we do not minimize it. The framework presupposes that the system entering the non-equilibrium process is in or near a thermal state. Phonon modes in a thermalized BEC condensate naturally satisfy this condition; coherently prepared atomic superpositions or squeezed acoustic states would not. The proposed BEC experiment must therefore be conducted with thermally populated phonon modes, and the paper identifies this as an explicit experimental requirement.

The five tests, summarized

Within the validated regime — thermal initial states, weak-to-moderate coupling, T/T_res in [0.20, 0.50], bosonic or qubit systems — Assumption A holds within numerical error. The proposed BEC experiment sits squarely inside this regime.

The prediction

For laboratory BEC parameters representative of current experimental capability (Steinhauer 2016; Muñoz de Nova et al. 2019):

Condensate temperature T_condensate ∼ 50 nK. Achievable analog Unruh temperature T_analog ∼ 10 nK. Ratio T/T_res = 0.2. Bound:

Naive energetic accounting would predict η ≈ 1, with deviations only from mundane experimental losses. The bound’s prediction differs from naive accounting by approximately 17 percentage points.

How to falsify it

Three failure modes, all clean:

(1) A controlled BEC experiment measures phonon extraction efficiencies that exceed the bound by more than experimental uncertainty — a direct violation of the inequality.

(2) Efficiencies show no scaling with T/T_res — the holographic-screen substitution does not transpose to acoustic analog systems.

(3) Efficiencies do scale with T/T_res but with a functional form inconsistent with 1/(1 + T/T_res) at the level of experimental error bars.

A constant-offset suppression at any single value of T/T_res can be reproduced by mundane loss channels — thermal phonon populations, mode-coupling losses, density inhomogeneities. The decisive test is the specific functional form of η vs. T/T_res across the accessible range. That cannot be reproduced by mundane losses without conspiracy.

What this paper does not claim

This is the part that matters most.

We do not modify Newtonian or general-relativistic gravity. We do not derive a new equation of motion. We do not unify entropy and gravity. We claim only that the framework, within its stated scope, predicts a falsifiable laboratory signature in cold-atom analog gravity that distinguishes it from naive energetic accounting.

Verlinde’s 2011 paper proposed a deep reframing of gravity itself. This paper is much narrower. It says: if you take the entropic-gravity construction seriously and apply standard non-equilibrium thermodynamics to it, you get a specific quantitative prediction in a specific laboratory regime, and that prediction is testable in the next several years. That’s the entire claim. It is small. It is also discriminating.

A successful BEC test would establish that holographic-screen entropy of the Verlinde type has thermodynamic content beyond the equilibrium force law — a small but non-trivial empirical result, the first of its kind. A failed test would falsify the construction or its acoustic transposition. Either outcome is informative.

Why this is Track 2

The Windstorm Institute’s first nine papers traced a single research arc: the throughput basin in serial decoding systems, from ribosomes to transformers. That arc is complete. Paper 10 opens a different line of inquiry — non-equilibrium thermodynamic bounds in analog physical systems — using the same mathematical lens (the Clausius inequality and entropy production) applied to a different substrate.

The two tracks share a meta-thesis: thermodynamic and information-theoretic constraints set the limits of what physical systems can do, across substrates. The throughput basin is one face of that thesis. The phonon bound is another.

We open Track 2 with one paper and no roadmap. Where it leads will depend on what the laboratories do, and on what the next analytical question turns out to be.

The Phonon Bound is Paper 10 of the Windstorm Institute — the first paper in the Entropic Bounds in Analog Systems track.
Zenodo: 10.5281/zenodo.20014391 · Code & data: github.com/Windstorm-Institute/phonon-extraction-bound
Download the full paper (PDF) · Markdown source