The Skin That Only Sings Intact: A Signal at Seventy-One Trillion Hertz, and the Model It Is Asked to Carry
A preprint reports a spectroscopic signal that appears only while the mitochondrial membrane holds its structure, and a polariton model that fits the number without yet bearing the weight. Three columns are kept apart.
The signal that needs the structure
A September 2026 preprint posted to bioRxiv by Yu Yang, Zhenglong Gu and Bo Song reports Fourier-transform infrared spectroscopy on intact biological samples: cultured human HEK-293T cells, fresh mouse kidney, liver, heart and skeletal muscle, and mitochondria isolated with their membrane structure preserved. A spectral feature appears near seventy-one trillion hertz. It is not assignable to a molecule — no known vibration sits there — and what makes it unusual is not what makes it quantum. The feature disappears when the sample is dried or ground. Destroy the organised structure and the signal is gone.
That structural dependence is the load-bearing observation, and the volume can read it in its own register. Carrying, as this book uses the word, is not something seen through a lens; it is what you find out by trying to bear weight on it. Remove the form and the order fails to stand: the signal is carried by the intact organisation of the membrane and not by any of its chemical parts counted alone. This is the practical test, and it has been run, once, by one group, on one instrument, on biological material that is by definition hard to keep intact. The result is reported. It is not yet replicated.
Discipline runs with the observation. Seeing a structurally dependent spectroscopic line is not seeing a quantum state. A signal that requires higher-order structure tells you the organisation is implicated in the measurement; it does not tell you the organisation is coherent, polaritonic, or quantum in any specific technical sense. The gap between those two sentences is the gap the rest of the chapter refuses to close on credit.
The model the signal is asked to carry
The preprint proposes a polariton. A polariton is a hybrid made when a confined light field and a material vibration couple so strongly that the two lose their separate identities: the original levels split, and the system's excitations belong to neither partner alone. This is real condensed-matter physics, demonstrated in optical cavities for decades, and the splitting it predicts — a Rabi splitting, measurable as two peaks where one sat before — is its signature. The authors identify the material partner as the collective vibration of the methylene groups in the mitochondrial inner membrane's lipids, near eighty-seven trillion hertz, and the cavity as the membrane's own geometry; the model then predicts that a mode near eighty-seven trillion hertz should split into two, one branch near seventy-one and one near one hundred and three. The lower branch coincides with the observed signal. The upper sits where water and macromolecule absorption would hide it.
A number that matches is consistency, and consistency is not evidence. A fitted frequency confirms that the model can describe the observation; it does not show that the model caused it. What would carry the polariton reading is precisely what the volume calls carrying: a measurable consequence that survives only if the mechanism is present — a Rabi splitting resolved in the spectrum, a coherence lifetime measured directly, or a time-resolved record of energy cycling between the light field and the material. None is reported. Until one is, the polariton is a model the signal is asked to carry, not a result the signal has been shown to carry.
The fence has a second rail. Classical explanations for the same observation have not been excluded: a conventional electromagnetic resonance of the mitochondrial geometry, a collective vibration of lipid or protein without strong coupling, or even a multiple-reflection artefact of the measurement geometry would each predict structure-dependent spectral features without invoking a polariton at all. The preprint argues against several of these, but does not measure the discriminating quantity. The chapter records the model and keeps it in its column.
The frequency and the meter
Alongside the spectroscopy, the preprint reports a functional measurement. Ten minutes of illumination near seventy-one trillion hertz raised reported ATP in HEK-293T cells by roughly ten per cent; the same duration near eighty-seven trillion hertz did the same; a control near fifty-four trillion hertz produced no significant change. Eight samples per group, one human cell line, absolute ATP values not reported.
The frequency dependence is the one feature that argues against plain heating — different frequencies should heat similarly if the effect were purely thermal, and here two frequencies act and one does not. But the argument does not settle the question, because different infrared frequencies are absorbed by different molecular bands to different degrees, and the resulting local temperature differences could be small and frequency-specific without a direct probe in place to see them. No direct local temperature measurement is reported. The honest status of the functional claim is: a frequency-dependent correlation, in a small sample, with the thermal confound registered rather than excluded.
The volume reads this through its own instrument. Chapter 72 placed the membrane and its catalysts as one instrument with the meter in the middle — the rotary ATP synthase that converts the membrane's stored potential into the cell's carrier. Here a frequency is being tuned to the membrane and the meter reads a change. That is a striking image, and it is an image only. The meter moved at this frequency is a correlation; the mechanism by which an infrared photon at the membrane reaches the rotary enzyme that makes ATP is itself a chain of hypotheses — the preprint invokes TCA-cycle intermediates and metabolic photons near eighty-seven trillion hertz — none of it directly measured. The chapter does not convert a correlation into a mechanism.
The membrane, again
Braid the result back into the volume. The inner mitochondrial membrane already does three jobs in these chapters. Chapter 68 measured it as a field — roughly a hundred and fifty to two hundred millivolts across four nanometres, a real gradient with units, held by a bilayer. Chapter 70 read it as the low-entropy gate: the place where the Sun's high-quality radiation is banked, across four nanometres of lipid, before it is spent at living temperature. Chapter 76 heard it as a tuned skin, a held sheet whose warm modes are broad and overdamped. The preprint proposes a fourth job — the membrane as a resonant light-matter cavity, its lipid vibrations strongly coupled to a confined field — and the discipline of the braiding is to ask whether the membrane is being asked to carry more than the evidence loads.
One feature of the proposal keeps it honest, and one feature threatens it. What keeps it honest is that the infrared in the experiment is delivered from outside: the result shows that exogenous mid-infrared light can perturb the intact membrane and move the meter. That is a measured perturbation. What threatens it is the further step, never measured, that the same coupling is driven in vivo by an endogenous field — metabolic photons that the cell itself emits near eighty-seven trillion hertz. That endogenous field is, at present, an unmeasured hypothesis; without it, the experiment demonstrates a susceptibility, not a function. The chapter records the distinction rather than smoothing it over.
There is a scale question the chapter records rather than resolves. A polariton's strong coupling depends on the field being confined to a cavity whose dimensions relate to the wavelength, and the mid-infrared wavelengths here are several microns while a mitochondrion is smaller than one. The authors propose the cristae geometry rather than the whole organelle as the relevant cavity, which is plausible but under-specified, and the chapter does not endorse a geometry it cannot check. The carrying test, applied again, asks whether removing the cristae structure removes the coupling — and that is the same unperformed measurement the polariton column is waiting on.
What would carry, and what would not
Two claims leave this chapter at different strengths. The structural dependence of the signal is carried by the intact form: destroy the organisation and the signal goes, which is the practical test run and reported, once. The polariton reading is not carried, because the test that would carry it — a resolved Rabi splitting, a direct coherence measurement, a time-resolved energy exchange, or the failure of every classical alternative — has not been run. These are not the same claim held at different levels of confidence. They are different claims, and the chapter refuses to let the stronger-sounding one inherit the standing of the more modest one by proximity.
The volume's rule on things in common applies to the membrane chapters themselves. The mitochondrial inner membrane and the membrane-in-the-middle of Chapter 75 are not the same instrument. They hold in common a held sheet coupled to a readout, a grammar of permitted modes and boundary conditions; they do not share a mechanism, a temperature, a scale, or a coupling strength. Borrowing the placement is permitted. Borrowing the standing is not.
The chapter's own claim is narrow and falsifiable. It claims only that the structural dependence of the seventy-one-trillion-hertz signal is a genuine carrying observation — order that fails when the form is removed — and that the polariton is, at the date of this entry, a consistent model rather than a carried result. If an independent group recovers the signal and the same group or another then measures a Rabi splitting with a coherence lifetime that no classical model fits, the polariton moves from model to result and this chapter's standing line changes with it. If the signal fails to replicate, or is shown to arise from a classical resonance without strong coupling, or if the ATP effect vanishes under direct local temperature control, the columns fall in the order the falsifier names. The chapter is written to be revised by exactly those outcomes, and by none other.
Equations borrowed
- Yang, Gu and Song, “A quantum state of mitochondria in the living cell,” bioRxiv preprint, September 2026 — a single-group, un-peer-reviewed result, borrowed at that standing and no higher.
- The polariton concept and Rabi splitting in strong light-matter coupling — established condensed-matter and cavity-QED physics, borrowed as the model under test.
- Peng et al., “Mid-infrared photons enhance mitochondrial ATP synthesis,” Fundamental Research (2025), at 8.3 micrometres — a peer-reviewed precedent that infrared can modulate ATP, borrowed for the real prior at a different proposed mechanism.
- Chapter 68's bioenergetic bilayer, Chapter 70's low-entropy gate, Chapter 72's meter-in-the-middle, and Chapter 76's tuned skin, borrowed back as the volume's own membrane line.
Validity band
The spectroscopic signal is reported in a single bioRxiv preprint by one group, un-peer-reviewed and unreplicated; the intact-sample dependence is the authors' own observation and the destroyed-sample control is theirs. The polariton is a model fitted to the frequency, with no Rabi splitting, coherence lifetime or time-resolved energy-exchange measurement presented; classical alternatives (electromagnetic resonance, collective lipid or protein vibration, multiple reflection) are not excluded. The ATP effect is one human cell line with eight samples per group, absolute ATP unreported, and no direct local temperature probe; heating is argued against by frequency dependence but not excluded by measurement. The endogenous metabolic field the in-vivo reading would require is itself an unmeasured hypothesis. The membrane-chapter braiding is the volume's own, and changes nothing about the preprint's standing.
Falsifier
The structural-dependence observation fails if independent replication does not recover a seventy-one-trillion-hertz signal that requires intact structure — a single-group artefact would retire the entry's central observation. The polariton reading retires if the signal is shown to arise from a classical mechanism — electromagnetic membrane resonance, collective lipid or protein vibration, or multiple reflection — without strong light-matter coupling, or if a Rabi splitting is sought and not found at the predicted coupling strength. The functional claim collapses if the ATP effect vanishes under direct local temperature control, or proves frequency-independent once differential absorption is equalised. The carrying observation itself — order that fails when the form is removed — would fail only if the signal were recovered in dried or ground structure, which no present account predicts.
Where this chapter is weakest
The material is days old at the time of writing, from one group, unreviewed, and the chapter risks lending a preprint more weight than a preprint can carry; the three-column apparatus is the defence against that and is also the reason the chapter is longer than its evidence strictly warrants. The polariton's cavity argument is under-specified on scale — a mitochondrion is smaller than the mid-infrared wavelength, and the cristae-geometry proposal is recorded rather than assessed. The ATP effect's two active frequencies both coincide with model features, which is suggestive precisely in the way that invites fitting after the fact. And the chapter leans on the volume's own membrane chapters for its braiding; if those are restructured, this one must follow, and its internal reference points travel with them.
The volume-wide audit of these weak points is collected in Where This Volume Is Weak.