IIT Roorkee, 2 September 2026: A Fluorescent-Protein Spin Qubit Six to Eight Orders Short of Neural Nitric Oxide
Quentir Medicine Monitor
Evidence-based insights for quantum medicine. Published by Quentir Systems LLC · September 3, 2026.
An individual nitric oxide molecule diffuses about two nanometers in roughly 0.2 nanoseconds, which makes the signal close to its point of production exceptionally transient. The methods available today each miss that measurement in a different way: a microelectrode cannot be placed with nanometer precision, a small-molecule indicator disturbs the redox chemistry it is there to report, and neither reads the radical's own magnetism. On 2 September 2026 five researchers at three Indian institutes published the calculation that says how far one new class of protein sensor sits from doing that job.
The preprint, posted to arXiv as 2609.02792 by Parul Raghuvanshi, Sagnik Ganguly, Sharika E, Mohana Priya T. and Vishvendra S. Poonia of IIT Roorkee, IIEST Shibpur and IIT (ISM) Dhanbad, works out the detection limit for spin relaxometry with the fluorescent-protein spin qubit. Their verdict on the sensor as it exists today is that it misses physiological concentrations of nitric oxide by six to eight orders of magnitude. The bottleneck they identify is the qubit's own relaxation time, which is set by the vibrations of the protein that houses it.
Practical takeaway. The measurement gap has been converted into two specifications that must be met together. Micromolar sensing of neural radicals needs an intrinsic relaxation time of 10 to 100 microseconds, and nanomolar sensing needs roughly 1 millisecond, against a native room-temperature value of 0.03 to 0.1 microseconds; the authors name a roughly twofold stiffening of the chromophore's surroundings as the route to the first. Alongside it they name photon yield per molecule as a co-equal bottleneck, so neither the spin engineering nor the optics is sufficient on its own.
What Feder and colleagues published in Nature on 20 August 2025: a spin-1 qubit inside enhanced yellow fluorescent protein
The object this calculation is about was demonstrated last year. Jacob Feder, Benjamin Soloway and their co-authors reported in Nature 645, 73 (2025) that enhanced yellow fluorescent protein hosts an optically addressable spin-1 qubit in the metastable triplet state of its chromophore. The qubit is initialized by 488 nanometer light through spin-selective intersystem crossing, driven by microwaves, and read out through optically activated delayed fluorescence, in which a 912 nanometer pulse lifts the triplet through a higher triplet state back into the singlet manifold. The delayed light that comes back is separated in time from prompt autofluorescence, so the emission carries the spin state.
Its Hamiltonian is a spin-1 with zero-field splitting, measured at D of 2.356 gigahertz and E of 0.458 gigahertz, which puts three zero-field magnetic resonance lines at roughly 0.92, 1.90 and 2.81 gigahertz. The property that matters for medicine is not spectroscopic. Because the sensor is a protein, a cell can be instructed to build it, and it can be fused to a chosen partner protein. That places the sensor spin within the three-nanometer barrel of an enzyme that makes the very radical one wants to measure. This Monitor read the wider case for cells that grow their own quantum sensors on 26 August 2026; the paper now under discussion is the first hard test of whether the idea survives contact with numbers.
How the Roorkee group built the detection limit: a Redfield relaxation matrix checked against Lindblad simulation and nitrogen-vacancy benchmarks
Spin relaxometry is a simple idea carried out carefully. A free radical carries an unpaired electron, and the magnetic field of that electron flickers as the radical diffuses past. A nearby sensor spin feels the flicker and loses its polarization faster, so the longitudinal relaxation time shortens in proportion to how many unpaired electrons are in the neighborhood. Measuring the relaxation rate therefore measures a local radical concentration, and does so through the radical's magnetism instead of through a binding event or a redox current.
The authors derive the transition-resolved relaxation matrix for the zero-field-split triplet coupled to a bath of diffusing radicals, establish the conditions under which that matrix collapses to a single exponential decay, and check the result twice: against direct Lindblad simulations, and against the established nitrogen-vacancy relaxometry literature. They then propagate three practical limits into the answer. Photon shot noise sets how finely a relaxation rate can be resolved. Photobleaching caps the total number of photons any one molecule will ever emit. Finite measurement bandwidth converts a static sensitivity into a time-resolved one. The paper is theoretical and computational throughout, built on QuTiP, NumPy, SciPy and Matplotlib, and the authors state that no experimental facility was used.
Quantum pillar: sensing. Technology readiness: TRL 2 of 9. Rung two means the idea has been worked out on paper and checked by simulation, with no measurement yet made on a real radical; the underlying protein qubit itself has only been demonstrated on a laboratory bench, and this paper concludes that in its present form it would not detect the signal at all.
Why room-temperature relaxation is the whole problem: a two-phonon Raman channel that dominates by roughly 720 to one
The central figure of the paper maps the minimum detectable nitric oxide concentration against two axes: the intrinsic relaxation time of the qubit and the number of photons collected. The native room-temperature sensor lands at a detection limit of order a million micromolar. Nitric oxide bursts near an active neuronal nitric oxide synthase enzyme reach the micromolar range, and resting levels sit near 10 nanomolar. The sensor as built is therefore about six orders above the loudest signal it might be asked to hear and about eight above the quiet one.
What closes that distance is a single physical quantity. The measured relaxation of the protein triplet follows a law with a term linear in temperature and a term rising as the seventh power of temperature. The first is the direct process, in which one lattice vibration carries away exactly the spin transition energy. The second is the two-phonon Raman process, in which two vibrations exchange energy with the spin and only their difference matters. At room temperature the authors find the Raman channel dominating by about 720 to one. That is a useful thing to learn, because the Raman coefficient scales as the inverse tenth power of the sound velocity in the material around the spin. A stiffness that is hard to improve by a factor of ten is not needed. Doubling it recovers a relaxation time near 100 microseconds, which is enough for micromolar sensing.
Their arithmetic on leverage is equally plain. A decade of intrinsic relaxation time buys a decade of concentration sensitivity, while a decade of photon budget buys only half a decade, because the shot-noise floor improves with the square root of counts. That makes the protein scaffold the more efficient place to spend effort, and it does not make it the only one. The paper is explicit that photon yield per molecule is a co-equal bottleneck, and its own improvement waterfall reaches the physiological range only when scaffold engineering, photon recycling, better optics and ensemble averaging are stacked together. A group that fixes the relaxation time and leaves the photon budget alone will not arrive.
What the 79-microsecond direct-process ceiling separates: micromolar sensing from nanomolar sensing
Suppressing the Raman channel does not remove the floor underneath it. When the two-phonon term is engineered away, the direct process remains, and the authors compute the relaxation time it permits at about 79 microseconds. That ceiling falls between the two targets. Micromolar sensitivity, which needs 10 to 100 microseconds, sits at or just under it and can be reached by decoupling the spin from the vibrations it currently couples to. Nanomolar sensitivity needs roughly a millisecond, which is more than an order of magnitude past the ceiling, so it requires genuine stiffening of the chromophore pocket rather than better decoupling.
Two intermediate operating points make the gradient concrete. At 80 kelvin, with a relaxation time of 141 microseconds and a hundred million detected photons, the sensor reaches a few tens of micromolar. An engineered qubit at a millisecond with a hundred billion photons crosses into the physiological burst range. Cryogenic operation is of limited interest to a hospital, and the authors present it as a diagnostic of the physics rather than as a route to the clinic.
Bandwidth costs more than it first appears. An engineered millisecond sensor asked to resolve a 100 millisecond transient reaches the micromolar burst level only with about a hundred million sensor molecules, and resting nanomolar levels at that speed stay out of range for any plausible number. The static thresholds assume integration over seconds. A relaxometric protein sensor is therefore suited to sustained or slowly varying radical levels, and tracking millisecond spikes costs one to two orders of magnitude of concentration sensitivity.
What anchoring the sensor 2 to 5 nanometers from the enzyme buys against a nanodiamond sitting 10 to 20 nanometers away
The comparison that matters clinically is with nitrogen-vacancy centers in nanodiamonds, which are already used to watch free radicals in and around living cells. On the physics the diamond wins comfortably: its room-temperature relaxation time is 1 to 3 milliseconds, three to four orders longer than the protein's, which gives it a far lower noise floor.
The trade a buyer would be making is between sensitivity and address. A nanodiamond has to be delivered into the cell as a foreign crystal. Its surface can be functionalized to steer it toward a class of target, which is real targeting and a blunter instrument than genetic fusion, because the crystal and its passivation layers still hold the sensing spin 10 to 20 nanometers off any particular protein. A protein sensor is written into the genome of the cell that will use it and can be attached to a chosen partner, which puts it 2 to 5 nanometers from the source. Two separate effects compound over that shorter distance, since the relaxation coupling falls as the inverse sixth power of separation and the local radical concentration is itself falling away from the point of production. A factor of four in distance is therefore worth a great deal more than a factor of four in signal.
Proximity has a limit that is worth stating plainly, because it bounds the most attractive version of this experiment. Watching one enzyme work is the thing a molecular-scale sensor seems to promise, and the diffusion arithmetic in the paper says it is close to impossible. An enzyme turning over at a hundred to a thousand radicals per second holds only about 1 nanomolar around itself at 3 nanometers, because the product leaves almost as fast as it appears. That figure sits below the nanomolar target the authors already class as the harder of the two, so single-molecule work is not the first application even on the optimistic engineering path. What remains available is coarser and still useful: bursts where production is correlated in time, compartments where concentration is locally held up, and ensembles of anchored sensors read together, which recovers sensitivity by giving away the spatial resolution that motivated the sensor in the first place.
The protein sensor also offers something that fixed-frequency, single-transition nitrogen-vacancy relaxometry does not. Because the relaxation rate depends on the noise spectrum at the probe frequency, measuring at two frequencies and taking the ratio cancels the unknown concentration prefactor and leaves a quantity that depends only on the radical's correlation time. Nitric oxide at about 5 picoseconds, superoxide at about 50 picoseconds and a nitroxide at about 1 nanosecond map to ratios spanning nearly two decades. Doing this cleanly for nitric oxide against superoxide needs a strong applied field, around 1.4 tesla, where the drivable transition moves to roughly 42 gigahertz, along with a generous photon budget. Distinguishing a fast radical from a slow one is much easier.
Which five predictions a laboratory can test now with purified protein and a relaxometry bench
What makes this paper fundable rather than merely interesting is that its central claims can be checked on a bench that already exists, using purified protein and no cells at all. Two of the checks are cheap and settle whether the physics is right. Incubating the protein with known amounts of a nitric oxide donor should trace a straight line between relaxation rate and concentration at a slope the theory states in advance, and a sensor fused to a radical-producing enzyme should show a specific enhancement over a sensor left loose in the cytosol, which is the anchoring claim standing on its own.
The test that would actually settle the program is the phonon one, because it is the one that can fail. If the account of the relaxation is correct, a mutation that stiffens the scaffold should lengthen the room-temperature relaxation time steeply, following the tenth-power dependence, and the improvement should then stop near 79 microseconds as the direct process takes over. A mutation series that raises stiffness and moves the relaxation time hardly at all would falsify the mechanism the whole design target rests on. A funder can ask for that curve before committing to cell work, and the same bench can check the two-frequency discrimination of nitric oxide against superoxide on plain radical solutions.
The authors are candid about what their framework assumes. The Debye model of a protein is order-of-magnitude guidance and not a first-principles calculation. The room-temperature relaxation time, which is the single most important input, has not been measured directly and carries a factor of about 1.6 of uncertainty, so they sweep it as a parameter and draw no conclusion that depends on its precise value. The correlation time of nitric oxide is uncertain across 1 to 10 picoseconds and enters the answer linearly. Every one of those choices was made in the direction that makes the sensor look worse, which is the right direction for a feasibility study.
How Quentir Reads It
This paper is worth attention because of what it does to a claim, and the claim it disciplines is one that has been circulating since the Nature demonstration. A quantum sensor that a cell can grow for itself is genuinely new, and the temptation is to describe it as a bedside instrument in waiting. The Roorkee calculation says it is not one yet, and says so with a number that anyone can check.
The move that gives the work its value is the translation. A biosensing question, which is hard to argue about, became a materials question with a right answer and a measurable target. That is the discipline this Monitor looks for in every quantum-medicine claim: which physical quantity has to change, by how much, before the application is possible at all. Here there are two such quantities. The relaxation time needs three to four orders, which the authors convert into a chromophore environment about twice as stiff, and the photon yield per molecule has to rise alongside it. A press release that reports progress on one of them has reported half the problem.
For a hospital buyer there is nothing here to procure and nothing to evaluate. For a research director choosing between platforms the reading is more useful, because the target is stated in a language protein engineers already work in. Directed evolution on fluorescent proteins has thirty years of history behind it, and asking that history for a stiffer chromophore pocket is an ordinary request. Two questions are worth putting to any group proposing this work, and they are this Monitor's own. Which scaffold mutations do they intend to try, and will they publish the measured room-temperature relaxation time, with its uncertainty, before they publish a picture of a cell. That number is currently inferred rather than measured, and measuring it is what would move this from a calculation to a candidate.
Sources
Primary source: Parul Raghuvanshi, Sagnik Ganguly, Sharika E, Mohana Priya T. and Vishvendra S. Poonia (IIT Roorkee, IIEST Shibpur and IIT (ISM) Dhanbad), "Phonon-limited detection thresholds for genetically encoded fluorescent-protein spin-qubit relaxometry of neural radicals," arXiv:2609.02792v1, submitted 2 September 2026, for the detection-limit theory and every quantitative claim here: the six-to-eight-order shortfall, the roughly 720-to-1 Raman dominance at room temperature, the inverse-tenth-power scaling with sound velocity, the 79-microsecond direct-process ceiling, the 10-to-100 microsecond micromolar and roughly 1 millisecond nanomolar thresholds, the native 0.03 to 0.1 microsecond relaxation time, the 141-microsecond 80 kelvin and 1 millisecond engineered operating points, the 2-to-5 against 10-to-20 nanometer standoff comparison, the roughly 1 nanomolar local concentration sustained by a single synthase at 3 nanometers, the 1.4 tesla and roughly 42 gigahertz discrimination conditions, the five testable predictions and the stated approximations. That preprint is a theoretical and computational study and reports no new measurement. The underlying qubit demonstration, its 488 nanometer initialization, its 912 nanometer delayed-fluorescence readout and its 2.356 and 0.458 gigahertz zero-field splitting parameters are from Feder, Soloway and colleagues in Nature 645, 73 (2025), published 20 August 2025, which this reading cites as the source of the platform rather than of the sensing analysis. An earlier Quentir Medicine Monitor reading of 26 August 2026, on genetically encoded quantum sensors, supplies the wider context drawn on here. The judgments are this Monitor’s own: that the paper’s value lies in converting a biosensing question into a spin-phonon materials target, that the assumptions were made in the direction unfavorable to the sensor, that there is nothing here for a hospital to procure, and that a directly measured room-temperature relaxation time is the milestone to watch.