diff --git a/CORRECTIONS.md b/CORRECTIONS.md index 85e13fd..2fb953e 100644 --- a/CORRECTIONS.md +++ b/CORRECTIONS.md @@ -115,3 +115,23 @@ Claims made during the exploration that did not survive it, kept so the record s angular structure in the lens. Their mock data show they would have found a real 10⁹ M☉ lump (log Bayes factor 29). The black hole bound stands, so no lens in the note now has both a small-scale anomaly and a bound on its black hole. Fixed in [notes/bh-lensing.md](notes/bh-lensing.md) (issue #28). + +## 18 October 2026: quantum paths (issue #16) + +17. **"A free quantum particle's path has fractal dimension 2."** Too broad. Abbott and Wise's dimension 2 + belongs to the path recorded by repeated position measurements, and only at resolutions finer than ħ/p, the + de Broglie wavelength over 2π. Coarser, it has dimension 1, and in a single-particle model the speed limit + makes it 1 again below ħ/mc (that end is not in Abbott and Wise, who are non-relativistic). The records also + make the roughness: each kicks the momentum, and the kicks add up. With records a fixed time apart the rough + stretch widens to roughly √N ħ/p over N records; with records a fixed number of wave clocks apart, Abbott + and Wise's self-similar spacing, the exponent rises to about 3. A long record is not one fixed fractal. + See [quantum-paths](quantum-paths/). +18. **"The same roughness as a pollen grain jostled by molecules."** Two slips. Brown watched particles about 1 to + 5 µm across released from inside pollen grains, which are about 100 µm across, not the grains themselves + (Pearle et al. 2010). And a Brownian path has dimension 2 only above its ballistic length, about 20 nm from + Li et al.'s (2010) numbers for their bead in air; below it the path is straight. The dimension matches, but + the quantum path is rough at small scales and the Brownian one at large scales, and only if nothing holds + it: Li et al.'s bead was trapped within about 22 nm rms, so its measured path is never rough. With the + 2.79 µm diameter they used, slip-corrected Stokes drag gives 175 µs at 2.75 kPa against their 147.3 µs; + `tests/quantum-paths.mjs` records that as a disagreement. (The first draft on #25 used the nominal 3 µm, + which hid the gap, and called the trapped path smooth, then rough, then smooth; review there caught both.) diff --git a/README.md b/README.md index fc021a6..badc519 100644 --- a/README.md +++ b/README.md @@ -18,8 +18,11 @@ with Claude (Anthropic), for [Foundation0](https://github.com/Foundation-0/f0). with gravity in place of electric attraction (the "gravitational atom"). It is still a wave cloud, not planets on orbits. The same equation, a wave held by a pull that weakens with distance, gives the same solution at any scale. The nesting lives in the equations, not the objects. -- **Quantum paths are fractal.** A free quantum particle's path has fractal dimension 2, against 1 for a - classical trajectory (Abbott and Wise, 1981), the same roughness as a pollen grain jostled by molecules. +- **Quantum paths are fractal, between two lengths.** A free particle's path, recorded by position + measurements finer than ħ/p, has fractal dimension 2, against 1 for a classical trajectory (Abbott and Wise, + 1981); below ħ/mc, in a single-particle model, the speed of light makes it smooth again. Its surroundings draw + that roughness only if they kick it harder than its own momentum: a cold atom in a room-temperature vacuum, + yes; nothing as warm as, and heavier than, the gas around it ([quantum-paths](quantum-paths/)). - **Cosmic structure is fractal only up to a point.** Galaxies cluster in a fractal-like way on small scales, but beyond about 80 to 300 million parsecs (h⁻¹ Mpc, 80 to 300) they are spread almost evenly (Scrimgeour et al., 2012). The nesting stops, the way a coastline's roughness stops at the size of molecules. diff --git a/quantum-paths/README.md b/quantum-paths/README.md new file mode 100644 index 0000000..e12d71f --- /dev/null +++ b/quantum-paths/README.md @@ -0,0 +1,251 @@ +# Quantum paths: rough only between two lengths + +Status: **exploration, not peer reviewed, no new result claimed.** Issue +[#16](https://github.com/Foundation-0/fractals/issues/16). + +The repository's README said a free quantum particle's path has fractal dimension 2, against 1 for a classical +trajectory (Abbott and Wise, 1981), "the same roughness as a pollen grain jostled by molecules". This study asks +over what range of scales that is true, what ends it at each end, and whether a particle's own surroundings ever +draw its path finely enough to see it. + +## What Abbott and Wise showed + +Their path is an *observed* one: record the particle's position again and again, each time to within Δx, and +join the dots. Their abstract: "the observed path of a particle in quantum mechanics is a fractal curve with +Hausdorff dimension two". Their dimension is defined by how the joined path's length grows as Δx shrinks, +⟨l⟩ ∝ Δx^(1−D). + +Each record leaves the particle with a momentum spread ħ/(2Δx) (for their Gaussian records). Abbott and Wise +space the records in two ways. In one they keep the time between records, Δt, fixed as Δx shrinks, so the wave +soon outgrows its starting width: ħΔt/2mΔx² ≫ 1, or Δt much longer than two of the Witness Map's wave clocks +mΔx²/ħ. In the other, the one under which they find the path self-similar (their eq. 1.9), Δt ∝ mΔx²/ħ, a fixed +number of wave clocks. (Lake 2022 asks for at least four wave clocks, from an energy–time argument of his own.) +In the first, far-field limit the mean length over a time T is + +``` + = (T/m) E|p + δp|, δp ~ Normal(0, (ħ/2Δx)²) in each direction +``` + +and, in three dimensions, with this model's local slope D = 1 − d ln⟨l⟩/d ln Δx at each Δx: + +| Δx / (ħ/p) | 0.01 | 0.1 | 0.3 | 1 | 3 | 10 | +|---|---|---|---|---|---|---| +| D | 2.00 | 1.99 | 1.89 | 1.38 | 1.05 | 1.00 | + +- Below the reduced de Broglie wavelength ħ/p (λ/2π) the spread wins and D = 2: → √(2/π) ħT/(mΔx). +- Above it the drift wins and D = 1: → pT/m, the classical length. +- These two limits are Abbott and Wise's. Between them, they write, D "is not well defined since it is rapidly + varying with Δx"; the middle of the table is this model's local slope, not their result. By that slope, D + passes 1.5 at Δx = 0.81 ħ/p (0.44 ħ/p for a path along one direction), and the crossover takes about a decade. +- Records closer together give the same two limits, with the crossover further out: 0.90 ħ/p at four wave + clocks apart, 1.1 ħ/p at two (Abbott and Wise's ħΔt/2mΔx² = 1), and 1.8 to 2.4 ħ/p at one, depending on + whether each record's own error is counted. + +So the README's sentence holds only for a recorded path, and only below ħ/p. The roughness is made by the +records: each one's kick ħ/(2Δx) is what makes the length grow as 1/Δx. Unlike a coastline, the path at +different Δx is not one curve seen at different magnifications. Abbott and Wise do find the path statistically +self-similar when Δt ∝ mΔx²/ħ, but only because every record restarts the particle with the same mean +momentum; "What repeated records do" below drops that. + +## The other end: the Compton wavelength + +Nothing moves faster than light. With relativistic speeds, v = pc²/E, the recorded length stops growing below the +reduced Compton wavelength ħ/mc and tends to cT, so the path is smooth again. This end is the model's own: +Abbott and Wise are non-relativistic throughout. In a single-particle model, then, a recorded path is rough only +between ħ/mc and ħ/p, a window that exists only when p ≪ mc: + +| Particle | ħ/mc | ħ/p | Window | +|---|---|---|---| +| Electron at 54 eV (Davisson and Germer) | 3.9×10⁻¹³ m | 2.7×10⁻¹¹ m | ×69 | +| Thermal neutron, 2,200 m/s | 2.1×10⁻¹⁶ m | 2.9×10⁻¹¹ m | ×1.4×10⁵ | +| Buckyball (C₆₀), 220 m/s | 2.9×10⁻¹⁹ m | 4.0×10⁻¹³ m | ×1.4×10⁶ | + +Localising a particle below its Compton wavelength is not a clean single-particle operation, so this end is a +bound, not a measurement. For the buckyball the whole window lies a thousand times below the molecule's own size +(about 1 nm), so its rough path is formal. Only light, slow particles have a ħ/p within reach. + +## What repeated records do + +The formula above treats each interval on its own, with the same mean momentum every time. But each record +kicks the momentum by about ħ/(2Δx), and the kicks add up: the particle heats. An exact simulation of repeated +Gaussian records (`sequential` in `paths.js`) shows three things. + +- For a particle starting at rest with records many wave clocks apart, the recorded path is (1/N) Σ√k ≈ (2/3)√N + times longer than the single-interval formula after N records. +- With records a fixed time apart (Abbott and Wise's fixed-Δt limit) the length still grows as 1/Δx, so D = 2 + survives. But the length grows as T^(3/2) rather than T. The rough stretch also widens: its upper end moves + out roughly as √N ħ/p, to about 11 ħ/p after 400 records. +- If the records are instead spaced a fixed number of wave clocks apart (Abbott and Wise's self-similar rule), + the length grows as Δx⁻²: D ≈ 3, not 2. + +So the dimension of a long record depends on how it is taken, and under Abbott and Wise's self-similar rule their +D = 2 does not survive the heating, which their reset at every record leaves out. Ding, Ong and Xu (2025) find +measurement shifting the dimension downward in a different (nonselective) model; the two are not compared here. + +## Brownian motion, and the pollen + +A Brownian particle is the reverse: smooth at small scales, rough at large ones. Below its ballistic length +v_rms·τ_p it moves in straight lines (D = 1). Above it, collisions have scrambled its velocity and the path is +rough (D = 2). + +- Li, Kheifets, Medellin and Raizen (2010) resolved the ballistic part for a silica bead, nominally 3 µm and + 2.79 µm in their analysis, held in an optical trap in air. Fits to its measured velocity autocorrelation gave + a momentum relaxation time of 48.5 µs at 99.8 kPa and 147.3 µs at 2.75 kPa, and its rms speed was 0.42 mm/s. +- Stokes drag with the Cunningham slip correction, for the 2.79 µm bead and the mass its rms speed implies, + gives 51 µs at 99.8 kPa, 5% high, and 175 µs at 2.75 kPa, 19% high. The ratio of the two times does not + depend on the mass: 3.43 here against 3.04 measured, so the slip correction used here overstates how much + freer the bead gets at low pressure. (The mass is not an independent check: Li et al. computed it from the + bead's size and density.) The test records the 2.75 kPa value as a disagreement. +- From their numbers the ballistic length is about 20 nm. A free bead would be rough above that, from about + 1 ms and 130 nm on. But the bead they measured was trapped (their eq. 1, trap frequency about 3.1 kHz), and the + trap holds it within about 22 nm rms, no more than its ballistic length at 99.8 kPa and a third of it at + 2.75 kPa. Its path goes from straight to confined with no rough stretch in between: the local dimension comes + within 0.1 of 2 only for brief moments, under 0.04 of a decade. + +Both kinds of path can be smooth, then rough, then smooth. The quantum one is rough between ħ/mc and ħ/p. A +free Brownian one is rough between v_rms·τ_p and whatever confines it or carries it along, and only if those +are far apart; for Li et al.'s trapped bead they are not. "The same roughness" is right about the dimension, 2, +but not about where it lives. + +Robert Brown did not watch pollen grains move. He watched particles about 1 to 5 µm across, lipid spherosomes and +starch amyloplasts, released from inside pollen grains about 100 µm across (Pearle et al. 2010). A whole grain does jiggle, but in water it moves +only about 0.1 µm in a second, too little for him to see. + +## Who draws the path: the Witness Map link + +An object's surroundings act as a measuring device that never stops. They record its position about once per +wave clock of the gap Δx\* where the Witness Map's W equals 1. That record can be fine or coarse in two ways. + +- **Weak records.** Each photon or molecule carries a little information (the Witness Map's rate grows as Δx²). + The records add up, and the best estimate of the object's position stays uncertain by about 0.6 Δx\*: for + a rate kΔx², the steady conditioned state is (ħ/8km)^(1/4) wide. The resolution is about Δx\*. +- **Full records.** Each hit tells the two places apart outright. It pins the object to about ħ/q, where q is + the momentum the hit carries: 2.70 k_BT/c for a thermal photon, and μ·v_rel for a molecule, with μ the reduced + mass. Δx\* is then only how far the wave spreads between hits. + +Either way, the resolution is min(Δx\*, ħ/q). The quantum roughness is drawn only where that is below ħ/p, which +is the same as asking whether the records kick the object harder than its own momentum. For each object on the +map, in its usual surroundings, with p its thermal momentum per direction at the surroundings' temperature +(`node quantum-paths/table.mjs`): + +| Object | Records | Main witness | Resolution | ħ/p | Ratio | +|---|---|---|---|---|---| +| Electron (lab vacuum) | full | leftover gas | 9.9×10⁻¹⁰ m | 1.7×10⁻⁹ m | 0.58 | +| Rubidium atom | full | leftover gas | 1.7×10⁻¹¹ m | 4.3×10⁻¹² m | 3.8 | +| Buckyball | full | its own glow | 9.4×10⁻⁷ m | 1.5×10⁻¹² m | 6.3×10⁵ | +| Giant molecule | weak | its own glow | 6.3×10⁻⁷ m | 2.5×10⁻¹³ m | 2.5×10⁶ | +| Virus (air) | weak | air | 9.5×10⁻¹⁴ m | 2.1×10⁻¹⁵ m | 45 | +| Interstellar dust grain | full | hydrogen atoms | 4.5×10⁻¹² m | 2.0×10⁻¹⁶ m | 2.2×10⁴ | +| Grain of sand | weak | air | 7.7×10⁻¹⁹ m | 1.4×10⁻²¹ m | 550 | +| Cat | weak | air | 1.1×10⁻²¹ m | 8.2×10⁻²⁵ m | 1,300 | +| Moon | weak | solar wind | 1.5×10⁻²⁷ m | 6.3×10⁻³⁶ m | 2.5×10⁸ | +| Earth | weak | solar wind | 2.7×10⁻²⁸ m | 6.9×10⁻³⁷ m | 3.9×10⁸ | +| Sun | weak | its own glow | 2.3×10⁻³¹ m | 4.5×10⁻³⁹ m | 5.1×10⁷ | +| Milky Way | weak | its own starlight | 5.0×10⁻³⁷ m | 9.5×10⁻⁴⁵ m | 5.3×10⁷ | +| **Rubidium atom cooled to 1 µK** | full | leftover gas | 1.7×10⁻¹¹ m | 7.5×10⁻⁸ m | **2.2×10⁻⁴** | + +Three cases: + +- **Heavier than its witnesses and as warm as them: drawn smooth.** Every object on the map except the electron + is recorded more coarsely than its own ħ/p, the rubidium atom by the narrowest margin (3.8). A gas molecule + lighter than the object, at the same temperature, carries less momentum than the object does. +- **Lighter than its witnesses: drawn at the crossover.** A collision can transfer at most about the lighter + body's own momentum. So the electron is recorded at about ħ/|p|, whatever its temperature, where the model's local + slope is about 1.4, between Abbott and Wise's two limits. +- **Colder than its witnesses: drawn rough.** A rubidium atom laser-cooled to 1 µK in the same room-temperature + vacuum is hit by leftover gas molecules carrying far more momentum than it has. Each hit records it 4,500 times + more finely than its ħ/p, deep in the rough regime, and also knocks it out of its trap. + +So the surroundings draw the quantum roughness only of something colder than they are. + +**Weak records and fluctuation-dissipation.** For an object warmed and slowed by the same gas that records it, +in the weak-record regime, the records and the drag come from the same collisions: Λ = Mγk_BT/ħ², with γ the drag +per unit mass. Then W at the thermal ħ/p is τ_th/τ_p, where τ_th = ħ/k_BT (2.5×10⁻¹⁴ s at 300 K) and τ_p is the +momentum relaxation time, and + +``` +Δx* / (ħ/p_th) = (τ_p / τ_th)^(1/4) +``` + +The surroundings could draw quantum roughness this way only if momentum relaxed faster than ħ/k_BT, which is +outside the high-temperature regime this holds in. For the virus in air (Kn ≈ 1.4, Epstein drag for specular +reflection) the relation predicts 44.9 and the map gives 44.8. Before the map's gas rate was corrected (#9), it +gave 33.7, low by π^(1/4), the same factor. That makes fluctuation-dissipation plus Epstein's drag a route to +#9's coefficient independent of the closed forms #9 compared against. Both routes assume specular hard-sphere +scattering; diffuse reflection would raise the rate by 1.39 to 1.44 (Epstein's factors 1 + π/8 and 1 + 9π/64 for +a conducting and an insulating sphere). For sand and the cat, much larger than air's +mean free path, the real drag is viscous and much weaker than the collision count suggests. There the map's rate +exceeds what fluctuation-dissipation allows, so their ratios are lower bounds. + +## What "the path" means + +There is no protocol-free answer. The dimension 2 belongs to a record of repeated position measurements, and the +records make it. Average trajectories reconstructed from weak measurements (Kocsis et al. 2011, single photons) +and Bohmian trajectories are smooth, dimension 1. The paths summed in Feynman's path integral are Brownian, but +they are integration variables, not something recorded. + +## Run it + +``` +node quantum-paths/table.mjs # the dimension, the windows, the bead and every Witness Map object +npm test # includes tests/quantum-paths.mjs +``` + +`paths.js` is the whole model. It loads in Node (`require`) and in a browser (`window.QP`). + +## Checks + +`tests/quantum-paths.mjs`: + +| Check | Expected | Model | +|---|---|---| +| Fine records: Δx m/(ħT) (3D) | √(2/π) = 0.798 | 0.798 | +| Coarse records: | pT/m | pT/m | +| D at 0.01 and 100 ħ/p (Abbott and Wise) | 2 and 1 | 2.000, 1.000 | +| Crossover (local slope D = 1.5, the model's; scale ħ/p from Abbott and Wise) | ħ/p, to a factor of 2 | 0.81 ħ/p | +| D below ħ/mc, inside the window, above ħ/p | 1, 2, 1 | 1.001, 1.995, 1.000 | +| Repeated records, N = 400: length over the single-interval formula | (1/N)Σ√k = 13.36 | 13.4 | +| C₆₀ at 220 m/s: h/mv (the same arithmetic as Arndt et al.) | 2.5 pm | 2.52 pm | +| Electron at 54 eV, Davisson and Germer's measurement as usually quoted | 1.65 Å | 1.67 Å | +| Bead in air, τ_p at 99.8 kPa (Li et al. 2010) | 48.5 µs | 51 µs | +| Bead in air, τ_p at 2.75 kPa (Li et al. 2010): recorded disagreement | 147.3 µs | 175 µs (+19%) | +| Free bead with those numbers: D over 0.04 nm and 1.3 µm | 1 and 2 | 1.000, 1.999 | +| Li et al.'s trapped bead: longest stretch with D within 0.1 of 2 (model, their eq. 1) | under 0.05 decade | 0.037 and 0.025 decade | +| Continuously recorded particle: purity, and W at its width | ħ²/4, 1/8 | exact | +| Gas rate against drag (Schlosshauer 2019, Epstein 1924) | equal | equal | +| Virus: Δx\*/(ħ/p) against fluctuation-dissipation | 44.9 | 44.8 | +| Every object but the electron recorded more coarsely than ħ/p | ratio > 1 | smallest 3.8 (rubidium) | +| Electron recorded at ħ/\|p\| | 1/√3 of ħ/p per direction | 0.577 at 300 K and 10⁴ K | +| Rubidium at 1 µK recorded finer than ħ/p | ratio < 10⁻³ | 2.2×10⁻⁴ | + +## What the model leaves out + +- Abbott and Wise's paper was checked in review on #25 against its preprint (SLAC-PUB-2427, 1979), not the + journal pages. Cannata and Ferrari's relativistic paper (Am. J. Phys. 56, 721 (1988)) could not be read; the + ħ/mc end here is this model's own (spinless, speeds capped at c), not theirs. +- p is the thermal momentum at the surroundings' temperature. Objects in beams move faster (smaller ħ/p, a wider + gap). Cooled objects move slower (larger ħ/p), which is where the gap closes, as the rubidium row shows. +- The resolution uses the main witness at Δx\*. Rarer, sharper witnesses can record finer now and then: gas + hits on the buckyball, at about ħ/q = 1.7×10⁻¹¹ m, are still 11 times coarser than its ħ/p. +- The Witness Map's own information per gas hit uses the molecule's thermal momentum. For the electron, lighter + than the molecules, the real transfer is set by the electron's momentum, as `witnessMomentum` assumes. That + only matters at gaps below about a nanometre. +- Strong damping at low temperature, where momentum relaxes in less than ħ/k_BT, is not covered. + +## Sources + +- L. F. Abbott and M. B. Wise, *Dimension of a quantum-mechanical path*, Am. J. Phys. 49, 37 (1981). https://doi.org/10.1119/1.12657 (preprint SLAC-PUB-2427, 1979) +- M. J. Lake, Eur. Phys. J. C 82, 928 (2022), which revisits the Abbott–Wise argument. https://arxiv.org/abs/2206.00609 +- A. Granik and G. Chapline, Phys. Lett. A 310 (2003), which applies the Abbott–Wise analysis. https://arxiv.org/abs/quant-ph/0302013 +- Y.-W. Ding, Y. C. Ong and H. Xu, *Measurement-induced perturbations of Hausdorff dimension in quantum paths*, Eur. Phys. J. Plus (2026). https://doi.org/10.1140/epjp/s13360-026-07554-3 (https://arxiv.org/abs/2512.13046) +- H. Kröger, *Fractal geometry in quantum mechanics, field theory and spin systems*, Phys. Rep. 323, 81 (2000). +- T. Li, S. Kheifets, D. Medellin and M. G. Raizen, *Measurement of the instantaneous velocity of a Brownian particle*, Science 328, 1673 (2010). https://doi.org/10.1126/science.1189403 +- T. Li and M. G. Raizen, *Brownian motion at short time scales*, Ann. Phys. (Berlin) 525, 281 (2013). https://arxiv.org/abs/1211.1458 +- P. Pearle et al., *What Brown saw and you can too*, Am. J. Phys. 78, 1278 (2010). https://arxiv.org/abs/1008.0039 +- S. Kocsis et al., *Observing the average trajectories of single photons in a two-slit interferometer*, Science 332, 1170 (2011). +- K. Jacobs and D. A. Steck, *A straightforward introduction to continuous quantum measurement*, Contemp. Phys. 47, 279 (2006). https://arxiv.org/abs/quant-ph/0611067 +- A. O. Caldeira and A. J. Leggett, *Path integral approach to quantum Brownian motion*, Physica A 121, 587 (1983). +- P. S. Epstein, *On the resistance experienced by spheres in their motion through gases*, Phys. Rev. 23, 710 (1924). +- M. Schlosshauer, *Quantum decoherence*, Phys. Rep. 831, 1 (2019). https://arxiv.org/abs/1911.06282 +- M. Arndt et al., *Wave-particle duality of C₆₀ molecules*, Nature 401, 680 (1999). +- C. Davisson and L. H. Germer, *Diffraction of electrons by a crystal of nickel*, Phys. Rev. 30, 705 (1927). diff --git a/quantum-paths/paths.js b/quantum-paths/paths.js new file mode 100644 index 0000000..cab6b2d --- /dev/null +++ b/quantum-paths/paths.js @@ -0,0 +1,160 @@ +/* Quantum paths model: how rough is a path, at what scale, and who draws it. No dependencies. + * + * Abbott & Wise (1981): record a free particle's position again and again, each time to within dx, and join the + * dots. The joined path's mean length grows as dx shrinks; where ~ dx^(1-D), D is the path's Hausdorff + * dimension. Each record leaves a momentum spread hbar/(2 dx) (their Gaussian records), and with the time between + * records held fixed as dx shrinks (so hbar dt / 2m dx^2 >> 1, their eq. 2.10), the mean length over a time T is + * = (T/m) E|p + dp|, dp ~ Normal(0, (hbar/2dx)^2) in each direction (far field, dt >> m dx^2/hbar) + * They find D = 2 when dx << hbar/p (the spread wins) and D = 1 when dx >> hbar/p (the drift wins), and call D not well + * defined in between. This model uses the local slope D = 1 - dln/dln dx there; the crossover values are its own. + * Relativity caps the speed at c, so below hbar/mc the length stops growing and D falls back to 1 (this model; Abbott + * & Wise are non-relativistic). + * A real Brownian particle is the reverse: smooth (ballistic) below v_th tau_p, rough above it, if nothing holds it. + * + * Witness Map link: a surroundings records an object about once per wave clock of the gap dx* where W = 1. Weak + * records (each carrying a little information) add up to a resolution of about dx*; a full record pins the object to + * about hbar/q, q the momentum the record carries. The path is drawn rough only where that resolution is below hbar/p. + * All SI units. + */ +(function (root) { + 'use strict'; + const C = { + h: 6.62607015e-34, hbar: 1.054571817e-34, c: 2.99792458e8, kB: 1.380649e-23, e: 1.602176634e-19, + u: 1.66053906660e-27, me: 9.1093837015e-31, mn: 1.67492749804e-27, + }; + + function erf(x) { // series below 3, continued fraction for erfc above; good to ~1e-15 + const s = Math.sign(x); x = Math.abs(x); + if (x < 3) { let sum = x, term = x; for (let n = 1; n < 200; n++) { term *= -x * x / n; sum += term / (2 * n + 1); if (Math.abs(term) < 1e-17) break; } return s * 2 / Math.sqrt(Math.PI) * sum; } + let f = 0; for (let n = 60; n >= 1; n--) f = n / 2 / (x + f); + return s * (1 - Math.exp(-x * x) / Math.sqrt(Math.PI) / (x + f)); + } + // E|X| for X ~ Normal(mean of length mu, variance s^2 in each of dim = 1 or 3 directions). + function meanNorm(mu, s, dim) { + if (dim !== 1 && dim !== 3) throw new Error('meanNorm: dim must be 1 or 3'); + const a = Math.abs(mu) / s; + if (dim === 1) return s * (Math.sqrt(2 / Math.PI) * Math.exp(-a * a / 2) + a * erf(a / Math.SQRT2)); + if (a < 1e-6) return 2 * s * Math.sqrt(2 / Math.PI); + return s * (Math.sqrt(2 / Math.PI) * Math.exp(-a * a / 2) + (a + 1 / a) * erf(a / Math.SQRT2)); + } + + // Mean length of the recorded path over time T, records far apart (the default), or one record every `clocks` + // wave clocks (dt = clocks m dx^2 / hbar), the wave starting dx wide and ending dx sqrt(1 + clocks^2/4) wide. The + // second form leaves out each record's own error; adding it moves the crossover at clocks = 1 from 1.8 to 2.4 hbar/p. + function awLength(dx, { m, p, T, dim = 3, clocks = Infinity }) { + if (!isFinite(clocks)) return T / m * meanNorm(p, C.hbar / (2 * dx), dim); + const dt = clocks * m * dx * dx / C.hbar; + return T / dt * meanNorm(p * dt / m, dx * Math.sqrt(1 + clocks * clocks / 4), dim); + } + // The far-field length in one direction with relativistic speed v(q) = q c^2 / sqrt(q^2 c^2 + m^2 c^4). + function awLengthRelativistic(dx, { m, p, T }) { + const s = C.hbar / (2 * dx), n = 4000, a = p - 10 * s, w = 20 * s / n, mc = m * C.c; + let sum = 0; + for (let i = 0; i <= n; i++) { + const q = a + i * w, z = (q - p) / s; + sum += (i === 0 || i === n ? 1 : i % 2 ? 4 : 2) * Math.abs(q) / Math.hypot(q, mc) * Math.exp(-z * z / 2); + } + return T * C.c * sum * w / 3 / (s * Math.sqrt(2 * Math.PI)); + } + // Local dimension D = 1 - dln L / dln dx of a length function L(dx). + function localDimension(L, dx) { + const e = 1e-4; + return 1 - (Math.log(L(dx * (1 + e))) - Math.log(L(dx * (1 - e)))) / (Math.log(1 + e) - Math.log(1 - e)); + } + // Same, from a structure function S(tau) = <(x(t+tau) - x(t))^2>: sampled every tau, a path of steps sqrt(S) + // has length ~ T sqrt(S)/tau, so D = 2 / (dln S / dln tau). Ballistic S ~ tau^2 gives 1, diffusive S ~ tau gives 2. + // Only meaningful for log-slopes between 1 and 2 (a steady process); NaN outside. + function dimensionFromS(S, tau) { + const e = 1e-4, a = (Math.log(S(tau * (1 + e))) - Math.log(S(tau * (1 - e)))) / (Math.log(1 + e) - Math.log(1 - e)); + return a < 1 - 1e-6 || a > 2 + 1e-6 ? NaN : 2 / a; + } + + // Classical Brownian particle (Ornstein-Uhlenbeck velocity, 1D rms speed sigmaV, momentum relaxation time taup) + // with a steady drift vDrift: smooth below ~sigmaV taup, rough above, smooth again above ~2 sigmaV^2 taup / vDrift. + const brownianS = (tau, { sigmaV, taup, vDrift = 0 }) => + 2 * sigmaV * sigmaV * taup * taup * (tau / taup - 1 + Math.exp(-tau / taup)) + Math.pow(vDrift * tau, 2); + // The same particle held in a harmonic trap of angular frequency w0, in equilibrium (Li et al. 2010, eq. 1): + // S = 2 (sigmaV/w0)^2 [1 - e^(-tau/2taup) (cos w1 tau + sin(w1 tau) / (2 w1 taup))], w1^2 = w0^2 - 1/(4 taup^2); + // cosh and sinh when overdamped. It levels off at 2 (sigmaV/w0)^2. + function trappedS(tau, { sigmaV, taup, w0 }) { + const g = 1 / (2 * taup), d = w0 * w0 - g * g, w1 = Math.sqrt(Math.abs(d)), x = w1 * tau; + const decay = d > 0 ? Math.exp(-g * tau) * (Math.cos(x) + g * Math.sin(x) / w1) + : ((1 + g / w1) * Math.exp(-(g - w1) * tau) + (1 - g / w1) * Math.exp(-(g + w1) * tau)) / 2; + return 2 * Math.pow(sigmaV / w0, 2) * (1 - decay); + } + // Drag on a sphere of radius r: Stokes with the Cunningham slip correction (gas mean free path mfp, viscosity eta). + function stokesCunningham({ r, eta, mfp }) { + const Kn = mfp / r, Cc = 1 + Kn * (1.257 + 0.4 * Math.exp(-1.1 / Kn)); + return 6 * Math.PI * eta * r / Cc; // force per unit speed + } + // Drag on a sphere much smaller than the mean free path, specular reflection (Epstein 1924): (4 pi/3) a^2 n m_g vbar. + // Diffuse reflection with full accommodation multiplies it by 1 + pi/8 (conducting sphere) or 1 + 9 pi/64 (insulating). + const epsteinDrag = ({ a, n, mg, T }) => (4 * Math.PI / 3) * a * a * n * mg * Math.sqrt(8 * C.kB * T / (Math.PI * mg)); + // Long-wavelength localisation rate of a sphere in a gas, Gamma = Lambda dx^2 (Schlosshauer 2019, eqs. 67-68). + const gasLambda = ({ a, n, mg, T }) => 8 * n * Math.sqrt(2 * Math.PI * mg) * a * a * Math.pow(C.kB * T, 1.5) / (3 * C.hbar * C.hbar); + + // Continuous position measurement with localisation rate Gamma = k dx^2 (Jacobs & Steck 2006 convention). + // Steady conditioned Gaussian: Vx = sqrt(hbar/8km), Vp = hbar sqrt(2kmhbar), Cxp = hbar/2 (a pure state), so the + // best estimate of the position is uncertain by about 0.6 dx*, where dx* = (hbar/km)^(1/4) is the gap with W = 1. + function monitored(k, m) { + const Vx = Math.sqrt(C.hbar / (8 * k * m)), Vp = C.hbar * Math.sqrt(2 * k * m * C.hbar), Cxp = C.hbar / 2; + return { Vx, Vp, Cxp, sigma: Math.sqrt(Vx), dxStar: Math.pow(C.hbar / (k * m), 0.25), tStar: Math.sqrt(m / (k * C.hbar)) }; + } + + // Exact repeated Gaussian position measurements (resolution dx, one every `clocks` wave clocks) on a free 1D + // particle, keeping the momentum each record leaves behind. Returns the recorded length and the time it took. + function sequential({ m, dx, clocks, N, p = 0, rng }) { + const dt = clocks * m * dx * dx / C.hbar, s2 = dx * dx, H2 = C.hbar * C.hbar / 4; + let x = 0, q = p, Vx = s2, Vp = H2 / s2, Cxp = 0, prev = 0, L = 0; + const gauss = () => { let u = 0; while (u === 0) u = rng(); return Math.sqrt(-2 * Math.log(u)) * Math.cos(2 * Math.PI * rng()); }; + for (let i = 0; i < N; i++) { + x += q * dt / m; Vx += 2 * Cxp * dt / m + Vp * dt * dt / (m * m); Cxp += Vp * dt / m; + const den = Vx + s2, r = x + Math.sqrt(den) * gauss(), d = r - x; + x += Vx / den * d; q += Cxp / den * d; Vx = Vx * s2 / den; Cxp = Cxp * s2 / den; Vp = (H2 + Cxp * Cxp) / Vx; + L += Math.abs(r - prev); prev = r; + } + return { L, T: N * dt }; + } + // For a particle starting at rest with records far apart, the k-th step carries the summed kicks of k records, + // so the recorded length is (1/N) sum_k sqrt(k) ~ (2/3) sqrt(N) times the single-interval formula. + const heatingFactor = N => { let s = 0; for (let k = 1; k <= N; k++) s += Math.sqrt(k); return s / N; }; + function mulberry32(seed) { // small seeded generator so checks are repeatable + return function () { seed |= 0; seed = seed + 0x6D2B79F5 | 0; let t = Math.imul(seed ^ seed >>> 15, 1 | seed); t = t + Math.imul(t ^ t >>> 7, 61 | t) ^ t; return ((t ^ t >>> 14) >>> 0) / 4294967296; }; + } + + const deBroglie = (m, v) => C.h / (m * v); + + // Momentum carried by one record from the named witness: 2.70 kT/c for thermal photons, h/lambda for starlight, + // and mu v_rel for gas or a particle beam, with mu the reduced mass, so an object lighter than the molecules + // that hit it gets about its own momentum, whatever theirs. + function witnessMomentum(o, env, label, T) { + const photon = Tp => 2.701 * C.kB * Tp / C.c; + if (label === 'Its own heat glow') return photon(o.selfHeat ? o.T : env.Tamb); + if (label === 'Its own starlight') return C.h / o.glowLambda; + const pop = env.pops.find(x => x.label === label); + if (!pop) return NaN; + if (pop.type === 'photon') return photon(pop.T); + const mu = o.m * pop.m / (o.m + pop.m); + return mu * Math.sqrt(3 * C.kB * T / o.m + (pop.type === 'gas' ? 3 * C.kB * pop.T / pop.m : pop.v * pop.v)); + } + // Witness Map bridge: how finely an object's own surroundings record it, against its thermal hbar/p, with + // p = sqrt(m kB T) per direction and T its translational temperature (by default its surroundings'). dx* is the + // gap where W = 1; the dominant witness there carries q per record; the resolution is min(dx*, hbar/q). + // WM is the witness-map model (require('../witness-map/model.js') or window.WM). + function drawnPath(WM, o, env, T = env.Tamb) { + const W = dx => WM.rates(o, env, dx, null).W; + let lo = 1e-50, hi = 1e30; + if (!(W(hi) >= 1) || W(lo) >= 1) return null; + for (let i = 0; i < 200; i++) { const mid = Math.sqrt(lo * hi); if (W(mid) < 1) lo = mid; else hi = mid; } + const top = WM.rates(o, env, hi, null).parts.slice().sort((a, b) => b.rate - a.rate)[0]; + const q = witnessMomentum(o, env, top.label, T), lambdaBar = C.hbar / Math.sqrt(o.m * C.kB * T); + const resolution = Math.min(hi, C.hbar / q); + return { dxStar: hi, witness: top.label, q, regime: C.hbar / q < hi ? 'full records' : 'weak records', resolution, T, lambdaBar, ratio: resolution / lambdaBar }; + } + + const api = { + C, erf, meanNorm, awLength, awLengthRelativistic, localDimension, dimensionFromS, brownianS, trappedS, stokesCunningham, + epsteinDrag, gasLambda, monitored, sequential, heatingFactor, mulberry32, deBroglie, witnessMomentum, drawnPath, + }; + if (typeof module !== 'undefined' && module.exports) module.exports = api; else root.QP = api; +})(typeof window !== 'undefined' ? window : globalThis); diff --git a/quantum-paths/table.mjs b/quantum-paths/table.mjs new file mode 100644 index 0000000..d756de4 --- /dev/null +++ b/quantum-paths/table.mjs @@ -0,0 +1,41 @@ +// Prints where quantum and Brownian paths are rough, and how finely each Witness Map object's surroundings draw its path. +// Usage: node quantum-paths/table.mjs +import { createRequire } from 'node:module'; +const require = createRequire(import.meta.url); +const QP = require('./paths.js'); +const WM = require('../witness-map/model.js'); +const { C } = QP; +const f = x => x.toExponential(1); + +console.log('Abbott-Wise path, recorded to within dx (3D, records far apart): dimension D against dx / (hbar/p)\n'); +const m = C.me, p = 1e5 * C.me, lb = C.hbar / p, L = dx => QP.awLength(dx, { m, p, T: 1 }); +console.log([0.01, 0.1, 0.3, 1, 3, 10, 100].map(x => `${String(x).padStart(5)}: D = ${QP.localDimension(L, x * lb).toFixed(2)}`).join('\n')); + +console.log('\nThe rough window of real particles: from hbar/mc (speed limit) up to hbar/p (drift)\n'); +console.log('particle'.padEnd(30) + 'hbar/mc m'.padStart(11) + 'hbar/p m'.padStart(11) + 'width'.padStart(9)); +for (const [name, mass, v] of [['Electron at 54 eV', C.me, Math.sqrt(2 * 54 * C.e / C.me)], ['Thermal neutron, 2,200 m/s', C.mn, 2200], + ['Buckyball (C60), 220 m/s', 720 * C.u, 220]]) { + const a = C.hbar / (mass * C.c), b = C.hbar / (mass * v); + console.log(name.padEnd(30) + f(a).padStart(11) + f(b).padStart(11) + f(b / a).padStart(9)); +} + +console.log('\nA Brownian path is the other way round on the inside: smooth below its ballistic length v_rms tau_p\n'); +console.log('Li et al. (2010), 2.79 um silica bead in air: tau_p measured and from slip-corrected Stokes drag; trap size v_rms/w0\n'); +const vrms = 0.4235e-3, mb = C.kB * 297 / (vrms * vrms); +for (const [P, meas, f] of [[99.8e3, 48.5e-6, 3064], [2.75e3, 147.3e-6, 3168]]) { + const tp = mb / QP.stokesCunningham({ r: 1.395e-6, eta: 1.81e-5, mfp: 68e-9 * 101.325e3 / P }); + console.log(`${(P / 1e3 + ' kPa').padEnd(10)} tau_p ${(meas * 1e6).toFixed(1)} us measured, ${(tp * 1e6).toFixed(0)} us model; ballistic length ` + + `${(vrms * meas * 1e9).toFixed(0)} nm; trap ${(vrms / (2 * Math.PI * f) * 1e9).toFixed(0)} nm`); +} +console.log('The trap holds the bead within its ballistic length, so the measured path is never rough; a free bead would be.'); + +console.log('\nWho draws the path: how finely each Witness Map object\'s own surroundings record it, against its thermal hbar/p\n'); +console.log('object'.padEnd(32) + 'records'.padEnd(14) + 'main witness'.padEnd(32) + 'resolution m'.padStart(13) + 'hbar/p m'.padStart(10) + 'ratio'.padStart(9)); +const row = (name, d) => console.log(name.padEnd(32) + d.regime.padEnd(14) + d.witness.padEnd(32) + f(d.resolution).padStart(13) + f(d.lambdaBar).padStart(10) + f(d.ratio).padStart(9)); +for (const o of WM.OBJECTS) { + const d = QP.drawnPath(WM, o, WM.ENVS[o.home]); + if (d) row(o.name, d); else console.log(o.name.padEnd(32) + 'no witnesses'); +} +row('Rubidium atom cooled to 1 uK', QP.drawnPath(WM, WM.OBJECTS.find(o => o.id === 'rb'), WM.ENVS.uhv, 1e-6)); +console.log('\nResolution: dx* (the gap where W = 1) for weak records, hbar/q for full records. Ratio above 1: the surroundings'); +console.log('never record the object finely enough to see its quantum roughness.'); diff --git a/tests/quantum-paths.mjs b/tests/quantum-paths.mjs new file mode 100644 index 0000000..ed58051 --- /dev/null +++ b/tests/quantum-paths.mjs @@ -0,0 +1,129 @@ +// Checks the quantum paths model against published numbers and its own closed forms. Run: npm test. +// Each check states the source, the published or expected value, the model's value and the tolerance. +// PASS: agreement within the tolerance. DIFF: a disagreement recorded in CORRECTIONS.md. A DIFF still passes, but only +// while the model's ratio to the published value stays inside the recorded range, so any change shows up here. +import { createRequire } from 'node:module'; +const require = createRequire(import.meta.url); +const QP = require('../quantum-paths/paths.js'); +const WM = require('../witness-map/model.js'); +const { C, awLength, awLengthRelativistic, localDimension, dimensionFromS, brownianS } = QP; + +let fails = 0, diffs = 0; +const fmt = x => (x === 0 ? '0' : x.toExponential(2)); +function within(name, got, want, factor, src) { + const ok = got > 0 && got / want <= factor && want / got <= factor; + console.log(`${ok ? 'PASS' : 'FAIL'} ${name}\n model ${fmt(got)} expected ${fmt(want)} (within x${factor}) ${src}`); + if (!ok) fails++; +} +function differs(name, got, want, lo, hi, why, src) { + const r = got / want, ok = got > 0 && r >= lo && r <= hi; + console.log(`${ok ? 'DIFF' : 'FAIL'} ${name}\n model ${fmt(got)} published ${fmt(want)} (model/published ${r.toPrecision(3)}, recorded ${lo} to ${hi}) ${src}\n ${why}`); + if (ok) diffs++; else fails++; +} +function check(name, ok, detail, src) { + console.log(`${ok ? 'PASS' : 'FAIL'} ${name}\n ${detail} ${src}`); + if (!ok) fails++; +} +function crossing(f, lo, hi, level) { // where a falling f(x) crosses level, by bisection in log x + for (let i = 0; i < 100; i++) { const mid = Math.sqrt(lo * hi); if (f(mid) > level) lo = mid; else hi = mid; } + return Math.sqrt(lo * hi); +} + +// 1. Abbott & Wise (1981): the recorded path has dimension 2 below hbar/p and 1 above it (3D, records far apart). +const AW = 'Abbott & Wise, Am. J. Phys. 49, 37 (1981)'; +const m = C.me, p = 1e5 * C.me, T = 1, lb = C.hbar / p; +const L3 = dx => awLength(dx, { m, p, T }); +within('Fine records: dx m / (hbar T) tends to sqrt(2/pi)', L3(1e-4 * lb) * 1e-4 * lb * m / (C.hbar * T), Math.sqrt(2 / Math.PI), 1.001, 'closed form, 3D'); +within('Coarse records: tends to the classical length pT/m', L3(1e4 * lb), p * T / m, 1.001, 'closed form'); +check('Dimension 2 well below hbar/p, 1 well above', Math.abs(localDimension(L3, 0.01 * lb) - 2) < 0.01 && Math.abs(localDimension(L3, 100 * lb) - 1) < 0.01, + `D = ${localDimension(L3, 0.01 * lb).toFixed(3)} at 0.01 hbar/p, ${localDimension(L3, 100 * lb).toFixed(3)} at 100 hbar/p`, AW); +const x15 = crossing(dx => localDimension(L3, dx), 1e-3 * lb, 1e3 * lb, 1.5) / lb; +check('The crossover sits at the reduced de Broglie wavelength hbar/p, to a factor of 2', x15 > 0.5 && x15 < 2, + `local slope D = 1.5 at ${x15.toFixed(3)} hbar/p`, 'model (Abbott & Wise give the two limits and the scale hbar/p; D between them is not theirs)'); + +// 2. Relativity: below hbar/mc the speed limit caps the length, so the path is smooth again (one direction). +const pr = 1e-3 * C.me * C.c, lc = C.hbar / (C.me * C.c); +const Lr = dx => awLengthRelativistic(dx, { m: C.me, p: pr, T }); +check('Smooth below the Compton wavelength, rough between, smooth above hbar/p (p = 0.001 mc)', + localDimension(Lr, 1e-3 * lc) < 1.01 && localDimension(Lr, 10 * lc) > 1.98 && localDimension(Lr, 1e4 * lc) < 1.01, + `D = ${localDimension(Lr, 1e-3 * lc).toFixed(3)}, ${localDimension(Lr, 10 * lc).toFixed(3)}, ${localDimension(Lr, 1e4 * lc).toFixed(3)} at 0.001, 10, 10^4 hbar/mc`, 'model'); + +// 3. Repeated records heat the particle: each kicks the momentum by ~hbar/(2 dx), and the kicks add up. For a +// particle starting at rest with records many wave clocks apart, the recorded length is (1/N) sum sqrt(k) times +// the single-interval formula, about (2/3) sqrt(N). +const rng = QP.mulberry32(2026), N = 400, dx = 1e-9, clocks = 10, R = 300; +let Lseq = 0; for (let r = 0; r < R; r++) Lseq += QP.sequential({ m, dx, clocks, N, rng }).L; +const single = awLength(dx, { m, p: 0, T: N * clocks * m * dx * dx / C.hbar, dim: 1 }); // far field, 1D +within('Repeated records: recorded length over the single-interval length, N = 400', Lseq / R / single, QP.heatingFactor(N), 1.05, + 'model (exact Gaussian updates, 300 seeded runs) against (1/N) sum sqrt(k)'); + +// 4. de Broglie wavelengths that set hbar/p for real particles. +within('C60 at 220 m/s: h/mv (m)', QP.deBroglie(720 * C.u, 220), 2.5e-12, 1.02, 'arithmetic; Arndt et al., Nature 401, 680 (1999) quote 2.5 pm'); +within('Electron at 54 eV: de Broglie wavelength (m)', QP.deBroglie(C.me, Math.sqrt(2 * 54 * C.e / C.me)), 1.65e-10, 1.02, 'Davisson & Germer (1927), 1.65 Angstrom as usually quoted (2.15 Angstrom x sin 50 deg)'); + +// 5. A Brownian particle is smooth below its ballistic length. Li, Kheifets, Medellin & Raizen (Science 328, 1673 +// (2010)) held a silica bead (nominally 3 um; 2.79 um in their analysis) in an optical trap in air: momentum +// relaxation 48.5 us at 99.8 kPa and 147.3 us at 2.75 kPa (fits to the velocity autocorrelation), rms speed +// 0.422 and 0.425 mm/s. Model: Stokes drag with the Cunningham slip correction, for the 2.79 um bead and the mass +// the rms speed implies (Li et al. computed that mass from the bead's size and density, so it is not independent). +const LI = 'Li et al., Science 328, 1673 (2010); Li & Raizen, Ann. Phys. 525, 281 (2013)'; +const vrms = 0.4235e-3, Tb = 297, mb = C.kB * Tb / (vrms * vrms); +const taup = P => mb / QP.stokesCunningham({ r: 1.395e-6, eta: 1.81e-5, mfp: 68e-9 * 101.325e3 / P }); +within('Bead in air at 99.8 kPa: momentum relaxation time (s)', taup(99.8e3), 48.5e-6, 1.1, LI); +differs('Bead in air at 2.75 kPa: momentum relaxation time (s)', taup(2.75e3), 147.3e-6, 1.15, 1.22, + `the slip correction overestimates the low-pressure time: the ratio of the two times, which does not depend on the mass, is ${(taup(2.75e3) / taup(99.8e3)).toFixed(2)} against ${(147.3 / 48.5).toFixed(2)} measured`, LI); +const bead = { sigmaV: vrms, taup: 48.5e-6 }, Sb = t => brownianS(t, bead); +check('A free bead with these numbers is smooth below its ballistic length (~20 nm) and rough above it', + dimensionFromS(Sb, 1e-7) < 1.01 && dimensionFromS(Sb, 1e-1) > 1.99, + `D = ${dimensionFromS(Sb, 1e-7).toFixed(3)} over ${(Math.sqrt(Sb(1e-7)) * 1e9).toFixed(2)} nm, ${dimensionFromS(Sb, 1e-1).toFixed(3)} over ${(Math.sqrt(Sb(1e-1)) * 1e6).toFixed(1)} um`, + 'model (free bead), tau_p from ' + LI); +// The bead they measured was trapped (eq. 1, w0 = 2 pi x 3064 Hz at 99.8 kPa and 3168 Hz at 2.75 kPa). Its trap holds +// it within about 22 nm rms, no more than its ballistic length, so its path never settles into D = 2. +const slopeD = (S, tau) => { const e = 1e-4, a = Math.log(S(tau * (1 + e)) / S(tau * (1 - e))) / Math.log((1 + e) / (1 - e)); return a > 0 ? 2 / a : NaN; }; +function roughDecades(S) { // longest stretch of log10 tau over which the local D = 2 / (dln S / dln tau) is within 0.1 of 2 + let best = 0, run = 0; const step = 1e-3; + for (let lt = -8; lt <= -2; lt += step) { if (Math.abs(slopeD(S, Math.pow(10, lt)) - 2) < 0.1) best = Math.max(best, run += step); else run = 0; } + return best; +} +const trap = [[99.8, 0.422e-3, 48.5e-6, 3064], [2.75, 0.425e-3, 147.3e-6, 3168]].map(([P, sigmaV, tp, f]) => { + const o = { sigmaV, taup: tp, w0: 2 * Math.PI * f }; + return { P, size: sigmaV / o.w0, ballistic: sigmaV * tp, rough: roughDecades(t => QP.trappedS(t, o)) }; +}); +const tFree = crossing(t => 2 - slopeD(Sb, t), 1e-6, 1, 0.1), xFree = Math.sqrt(Sb(tFree)); +check('The trapped bead Li et al. measured has no rough stretch: D is within 0.1 of 2 for under 0.05 of a decade', + trap.every(t => t.rough < 0.05) && xFree > 3 * Math.SQRT2 * trap[0].size, + trap.map(t => `${t.P} kPa: trap ${(t.size * 1e9).toFixed(1)} nm rms, ballistic ${(t.ballistic * 1e9).toFixed(1)} nm, ${t.rough.toFixed(3)} decade`).join('; ') + + `; a free bead reaches D = 1.9 only after ${(tFree * 1e6).toFixed(0)} us, ${(xFree * 1e9).toFixed(0)} nm rms`, 'model, with eq. (1) and the trap frequencies of ' + LI); + +// 6. Witness Map link. A surroundings that records at rate k dx^2 keeps the best estimate of the position uncertain +// by about 0.6 dx*, dx* = (hbar/km)^(1/4) being the gap where W = 1. +const k = 1e20, mo = QP.monitored(k, C.me); +check('Continuously recorded particle: steady state is pure, and W = 1/8 at its width', + Math.abs((mo.Vx * mo.Vp - mo.Cxp * mo.Cxp) / (C.hbar * C.hbar / 4) - 1) < 1e-9 && Math.abs(k * mo.Vx * C.me * mo.Vx / C.hbar - 1 / 8) < 1e-9, + `Vx Vp - C^2 = ${((mo.Vx * mo.Vp - mo.Cxp * mo.Cxp) / (C.hbar * C.hbar / 4)).toFixed(6)} hbar^2/4; width ${(mo.sigma / mo.dxStar).toFixed(4)} dx*`, + 'model; convention of Jacobs & Steck, Contemp. Phys. 47, 279 (2006)'); +// Fluctuation-dissipation: the localisation rate in a gas and the drag on a small sphere come from the same +// collisions, Lambda = M gamma kT / hbar^2 (gamma = drag / M), for specular reflection. For the virus in air +// (Kn ~ 1.4) it predicts dx*/lambda_bar = (tau_p / tau_th)^(1/4), tau_th = hbar/kT. +const air = { a: 5e-8, n: 2.5e25, mg: 28 * C.u, T: 300 }; +within('Gas localisation rate equals drag x kT / hbar^2 (Schlosshauer 2019 eqs. 67-68 against Epstein 1924)', + QP.gasLambda(air), QP.epsteinDrag(air) * C.kB * air.T / (C.hbar * C.hbar), 1.000001, 'Epstein, Phys. Rev. 23, 710 (1924), specular reflection'); +const virus = WM.OBJECTS.find(o => o.id === 'virus'), vd = QP.drawnPath(WM, virus, WM.ENVS.air); +const fdt = Math.pow(virus.m / QP.epsteinDrag(air) / (C.hbar / (C.kB * 300)), 0.25); +within('Virus in air: Witness Map dx*/lambda_bar against fluctuation-dissipation', vd.ratio, fdt, 1.03, + 'Witness Map; agrees since the gas rate was corrected in #9 (before it, low by pi^(1/4))'); +// Who records finer than hbar/p? Only witnesses that carry more momentum than the object. +const rows = WM.OBJECTS.map(o => ({ o, d: QP.drawnPath(WM, o, WM.ENVS[o.home]) })).filter(x => x.d); +const heavy = rows.filter(x => x.o.id !== 'electron'), least = heavy.reduce((a, b) => (a.d.ratio < b.d.ratio ? a : b)); +check('Every other object on the map, at the temperature of its surroundings, is recorded more coarsely than hbar/p', + heavy.every(x => x.d.ratio > 1), `${heavy.length} objects; smallest ratio ${least.d.ratio.toFixed(1)} (${least.o.name.toLowerCase()}, ${least.d.regime})`, 'Witness Map model'); +const el = WM.OBJECTS.find(o => o.id === 'electron'), e300 = QP.drawnPath(WM, el, WM.ENVS.uhv), e1e4 = QP.drawnPath(WM, el, WM.ENVS.uhv, 1e4); +check('An object lighter than the molecules that hit it is recorded at its own hbar/|p|, whatever its temperature', + [e300, e1e4].every(d => Math.abs(d.ratio * Math.sqrt(3) - 1) < 0.01), + `electron in lab vacuum: resolution ${e300.ratio.toFixed(3)} and ${e1e4.ratio.toFixed(3)} of hbar/p per direction at 300 K and 10^4 K (1/sqrt 3 = 0.577)`, 'Witness Map model'); +const rbCold = QP.drawnPath(WM, WM.OBJECTS.find(o => o.id === 'rb'), WM.ENVS.uhv, 1e-6); +check('A rubidium atom cooled to 1 uK in a room-temperature vacuum is recorded far finer than its hbar/p', + rbCold.ratio < 1e-3, `resolution ${fmt(rbCold.resolution)} m against hbar/p ${fmt(rbCold.lambdaBar)} m (ratio ${fmt(rbCold.ratio)}), by leftover-gas hits`, 'Witness Map model'); + +console.log(fails ? `\n${fails} check(s) failed` : `\nall checks pass (${diffs} recorded disagreement${diffs === 1 ? '' : 's'}, see CORRECTIONS.md)`); +process.exit(fails ? 1 : 0);