From 8fed82d95ddf46f78a6cfba5eacf2b4954ea5b42 Mon Sep 17 00:00:00 2001 From: Claude Date: Sun, 18 Oct 2026 11:31:43 +0000 Subject: [PATCH 1/3] Quantum paths: where dimension 2 starts and stops, and who draws the path (issue #16) quantum-paths/paths.js reproduces the Abbott-Wise length of a path recorded to within dx and its crossover from dimension 2 to 1 near hbar/p (D = 1.5 at 0.81 hbar/p in 3D), the return to 1 below hbar/mc, and the heating from repeated records: (1/N) sum sqrt(k) times longer, the rough stretch widening as sqrt(N) hbar/p. It contrasts a Brownian particle, smooth below its ballistic length (Li et al. 2010: 47 and 154 us against 48 and 147 us). The Witness Map link: the surroundings record an object about once per wave clock of the gap where W = 1, to a resolution min(dx*, hbar/q), q the momentum the main witness carries. Every object on the map but the electron is recorded more coarsely than its thermal hbar/p; the electron, lighter than the gas, at its own hbar/|p|; a rubidium atom cooled to 1 uK about 4,500 times more finely. Fluctuation-dissipation with Epstein drag reproduces the map's corrected gas rate for the virus (44.8 against 44.9). tests/quantum-paths.mjs: 17 checks. README bullet and CORRECTIONS 16-17. Reviewed by independent agents (physics, claims, code); their findings are folded in. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01C2bpiHyHSkC8DAL5CcyciK --- CORRECTIONS.md | 14 +++ README.md | 7 +- quantum-paths/README.md | 231 ++++++++++++++++++++++++++++++++++++++++ quantum-paths/paths.js | 149 ++++++++++++++++++++++++++ quantum-paths/table.mjs | 38 +++++++ tests/quantum-paths.mjs | 103 ++++++++++++++++++ 6 files changed, 540 insertions(+), 2 deletions(-) create mode 100644 quantum-paths/README.md create mode 100644 quantum-paths/paths.js create mode 100644 quantum-paths/table.mjs create mode 100644 tests/quantum-paths.mjs diff --git a/CORRECTIONS.md b/CORRECTIONS.md index c2798e7..cdf4367 100644 --- a/CORRECTIONS.md +++ b/CORRECTIONS.md @@ -104,3 +104,17 @@ Claims made during the exploration that did not survive it, kept so the record s to grow, and at its radius the galaxy, not the black hole, holds the particles. Any cloud these black holes can grow within the age of the universe, at any particle mass, peaks within about 0.1 pc. See [notes/bh-lensing.md](notes/bh-lensing.md). + +## 18 October 2026: quantum paths (issue #16) + +16. **"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. The records also make the roughness: each kicks the momentum, the kicks add up, + and over N records the rough stretch widens to roughly √N ħ/p, so a long record is not one fixed fractal. + See [quantum-paths](quantum-paths/). +17. **"The same roughness as a pollen grain jostled by molecules."** Two slips. Brown watched particles about a + micron across released from inside pollen grains, which are about 100 µm long, 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 a 3 µm 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. diff --git a/README.md b/README.md index 05eb0b7..9f8f08b 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 but coarser than ħ/mc, has fractal dimension 2, against 1 for a classical + trajectory (Abbott and Wise, 1981). 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..25d80ed --- /dev/null +++ b/quantum-paths/README.md @@ -0,0 +1,231 @@ +# 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). If the records are far enough apart for the wave +to outgrow its starting width (Lake 2022 restates the condition as Δt ≳ 4mΔx²/ħ, four of the Witness Map's wave +clocks), 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: + +| 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. +- D passes 1.5 at Δx = 0.81 ħ/p (0.44 ħ/p for a path along one direction). The crossover takes about a decade. +- Records only one wave clock apart give the same two limits. The crossover then moves to 1.8 to 2.4 ħ/p, + 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. + +## 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. 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 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, the length grows as Δx⁻²: an exponent + of 3, not 2. + +So the dimension of a long record depends on how it is taken. 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 3 µm silica bead held in an optical + trap in air. Its momentum relaxation time was 48 µs at 99.8 kPa and 147 µs at 2.75 kPa, with an rms speed of + 0.42 mm/s. +- Stokes drag with the Cunningham slip correction gives 47 and 154 µs, using the mass that rms speed implies. + That mass corresponds to a density of about 1,600 kg/m³; bulk silica's ~2,000 would give about 58 µs. +- From their numbers the ballistic length is about 20 nm. A free bead would be rough above that. In the trap, + the roughness is cut off again at the trap's own confinement. + +Both kinds of path are smooth, then rough, then smooth. The quantum one is rough between ħ/mc and ħ/p. The +Brownian one is rough between v_rms·τ_p and whatever confines it or carries it along. "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 a micron across released from inside +pollen grains, which are about 100 µm long (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 dimension + is about 1.4, halfway between smooth and rough. +- **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 + π/8. 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 (D = 1.5) | 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, measured by Davisson and Germer | 1.65 Å | 1.67 Å | +| Bead in air, τ_p at 99.8 and 2.75 kPa (Li et al. 2010) | 48 and 147 µs | 47 and 154 µs | +| Free bead with those numbers: D over 0.04 nm and 1.3 µm | 1 and 2 | 1.000, 1.999 | +| 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 itself was not read: only its abstract and later restatements (Lake 2022; Granik + and Chapline 2003). Their exact equations and their rule for spacing the records are unchecked, and so is + Cannata and Ferrari's relativistic paper (Am. J. Phys. 56, 721 (1988)). +- 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 +- M. J. Lake, Eur. Phys. J. C 82, 928 (2022), which restates 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. Ding, Y. C. Ong and H. Xu, *Measurement-induced perturbations of Hausdorff dimension in quantum paths* (2025; Eur. Phys. J. Plus, 2026). 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..ee19b81 --- /dev/null +++ b/quantum-paths/paths.js @@ -0,0 +1,149 @@ +/* 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, and D = 1 - dln/dln dx is its length-scaling + * (divider) dimension; for a path in three dimensions it is the Hausdorff dimension of the trace. Each record leaves + * a momentum spread hbar/(2 dx), and with records far enough apart for the wave to outgrow its start, + * = (T/m) E|p + dp|, dp ~ Normal(0, (hbar/2dx)^2) in each direction (far field, dt >> m dx^2/hbar) + * D = 2 when dx << hbar/p (the spread wins) and D = 1 when dx >> hbar/p (the drift wins). + * Relativity caps the speed at c, so below hbar/mc the length stops growing and D falls back to 1. + * A real Brownian particle is the reverse: smooth (ballistic) below v_th tau_p, rough above 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); + // 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. + 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, 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..09dedd1 --- /dev/null +++ b/quantum-paths/table.mjs @@ -0,0 +1,38 @@ +// 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 = ${QP.localDimension(L, x * lb).toFixed(2)}`).join('\n')); + +console.log('\nThe rough window of real particles: from hbar/mc (speed limit) up to hbar/p (drift)\n'); +console.log('particle'.padEnd(30) + 'hbar/mc m'.padStart(11) + 'hbar/p m'.padStart(11) + 'width'.padStart(9)); +for (const [name, mass, v] of [['Electron at 54 eV', C.me, Math.sqrt(2 * 54 * C.e / C.me)], ['Thermal neutron, 2,200 m/s', C.mn, 2200], + ['Buckyball (C60), 220 m/s', 720 * C.u, 220]]) { + const a = C.hbar / (mass * C.c), b = C.hbar / (mass * v); + console.log(name.padEnd(30) + f(a).padStart(11) + f(b).padStart(11) + f(b / a).padStart(9)); +} + +console.log('\nA Brownian path is the other way round on the inside: smooth below its ballistic length v_rms tau_p\n'); +const vrms = 0.4235e-3, mb = C.kB * 297 / (vrms * vrms); +for (const [P, label] of [[99.8e3, '3 um bead in air, 99.8 kPa'], [2.75e3, '3 um bead in air, 2.75 kPa']]) { + const tp = mb / QP.stokesCunningham({ r: 1.5e-6, eta: 1.81e-5, mfp: 68e-9 * 101.325e3 / P }); + console.log(`${label.padEnd(30)} tau_p ${(tp * 1e6).toFixed(0)} us, ballistic length ${(vrms * tp * 1e9).toFixed(0)} nm`); +} + +console.log('\nWho draws the path: how finely each Witness Map object\'s own surroundings record it, against its thermal hbar/p\n'); +console.log('object'.padEnd(32) + 'records'.padEnd(14) + 'main witness'.padEnd(32) + 'resolution m'.padStart(13) + 'hbar/p m'.padStart(10) + 'ratio'.padStart(9)); +const row = (name, d) => console.log(name.padEnd(32) + d.regime.padEnd(14) + d.witness.padEnd(32) + f(d.resolution).padStart(13) + f(d.lambdaBar).padStart(10) + f(d.ratio).padStart(9)); +for (const o of WM.OBJECTS) { + const d = QP.drawnPath(WM, o, WM.ENVS[o.home]); + if (d) row(o.name, d); else console.log(o.name.padEnd(32) + 'no witnesses'); +} +row('Rubidium atom cooled to 1 uK', QP.drawnPath(WM, WM.OBJECTS.find(o => o.id === 'rb'), WM.ENVS.uhv, 1e-6)); +console.log('\nResolution: dx* (the gap where W = 1) for weak records, hbar/q for full records. Ratio above 1: the surroundings'); +console.log('never record the object finely enough to see its quantum roughness.'); diff --git a/tests/quantum-paths.mjs b/tests/quantum-paths.mjs new file mode 100644 index 0000000..3f64c57 --- /dev/null +++ b/tests/quantum-paths.mjs @@ -0,0 +1,103 @@ +// Checks the quantum paths model against published numbers and its own closed forms. Run: npm test. +// Each check states the source, the published or expected value, the model's value and the tolerance. +import { createRequire } from 'node:module'; +const require = createRequire(import.meta.url); +const QP = require('../quantum-paths/paths.js'); +const WM = require('../witness-map/model.js'); +const { C, awLength, awLengthRelativistic, localDimension, dimensionFromS, brownianS } = QP; + +let fails = 0; +const fmt = x => (x === 0 ? '0' : x.toExponential(2)); +function within(name, got, want, factor, src) { + const ok = got > 0 && got / want <= factor && want / got <= factor; + console.log(`${ok ? 'PASS' : 'FAIL'} ${name}\n model ${fmt(got)} expected ${fmt(want)} (within x${factor}) ${src}`); + if (!ok) fails++; +} +function check(name, ok, detail, src) { + console.log(`${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, + `D = 1.5 at ${x15.toFixed(3)} hbar/p`, AW); + +// 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), measured 1.65 Angstrom'); + +// 5. A Brownian particle is smooth below its ballistic length. Li, Kheifets, Medellin & Raizen (Science 328, 1673 +// (2010)) held a 3 um silica bead in an optical trap in air: momentum relaxation 48 us at 99.8 kPa and 147 us at +// 2.75 kPa, rms speed 0.42 mm/s. Model: Stokes drag with the Cunningham slip correction, using the mass the +// measured speed implies (density ~1,600 kg/m^3; bulk silica's ~2,000 would give ~58 us). +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: 1.5e-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), 48e-6, 1.1, LI); +within('Bead in air at 2.75 kPa: momentum relaxation time (s)', taup(2.75e3), 147e-6, 1.1, LI); +const bead = { sigmaV: vrms, taup: 48e-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; the trapped bead is bounded above by its trap), tau_p from ' + 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'); +process.exit(fails ? 1 : 0); From 114cf2d78e91419b3eda004de5874980393ceee4 Mon Sep 17 00:00:00 2001 From: Claude Date: Sun, 18 Oct 2026 11:47:50 +0000 Subject: [PATCH 2/3] Quantum paths: source corrections from a verification pass - Li et al.'s fitted momentum relaxation times are 48.5 and 147.3 us. - Brown's particles were about 1 to 5 um (spherosomes and amyloplasts), not about a micron (Pearle et al. 2010). - Lake's restatement is dt >> 4 m dx^2 / hbar, with 4 m dx^2 / hbar the shortest spacing he allows. - Epstein's diffuse-reflection factor is 1 + pi/8 for a conducting sphere and 1 + 9 pi/64 for an insulating one. - Davisson and Germer's 1.65 Angstrom is labelled as the usual quotation; Ding, Ong and Xu cited with their journal DOI. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01C2bpiHyHSkC8DAL5CcyciK --- CORRECTIONS.md | 4 ++-- quantum-paths/README.md | 21 +++++++++++---------- quantum-paths/paths.js | 2 +- tests/quantum-paths.mjs | 12 ++++++------ 4 files changed, 20 insertions(+), 19 deletions(-) diff --git a/CORRECTIONS.md b/CORRECTIONS.md index cdf4367..379f5c3 100644 --- a/CORRECTIONS.md +++ b/CORRECTIONS.md @@ -113,8 +113,8 @@ Claims made during the exploration that did not survive it, kept so the record s makes it 1 again below ħ/mc. The records also make the roughness: each kicks the momentum, the kicks add up, and over N records the rough stretch widens to roughly √N ħ/p, so a long record is not one fixed fractal. See [quantum-paths](quantum-paths/). -17. **"The same roughness as a pollen grain jostled by molecules."** Two slips. Brown watched particles about a - micron across released from inside pollen grains, which are about 100 µm long, not the grains themselves +17. **"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 a 3 µm 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. diff --git a/quantum-paths/README.md b/quantum-paths/README.md index 25d80ed..8eee16d 100644 --- a/quantum-paths/README.md +++ b/quantum-paths/README.md @@ -16,8 +16,8 @@ Hausdorff dimension two". Their dimension is defined by how the joined path's le ⟨l⟩ ∝ Δx^(1−D). Each record leaves the particle with a momentum spread ħ/(2Δx). If the records are far enough apart for the wave -to outgrow its starting width (Lake 2022 restates the condition as Δt ≳ 4mΔx²/ħ, four of the Witness Map's wave -clocks), the mean length over a time T is +to outgrow its starting width (Lake 2022 restates the condition as Δt ≫ 4mΔx²/ħ; 4mΔx²/ħ, four of the Witness +Map's wave clocks, is the shortest spacing he allows), the mean length over a time T is ``` = (T/m) E|p + δp|, δp ~ Normal(0, (ħ/2Δx)²) in each direction @@ -79,8 +79,8 @@ v_rms·τ_p it moves in straight lines (D = 1). Above it, collisions have scramb rough (D = 2). - Li, Kheifets, Medellin and Raizen (2010) resolved the ballistic part for a 3 µm silica bead held in an optical - trap in air. Its momentum relaxation time was 48 µs at 99.8 kPa and 147 µs at 2.75 kPa, with an rms speed of - 0.42 mm/s. + 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 gives 47 and 154 µs, using the mass that rms speed implies. That mass corresponds to a density of about 1,600 kg/m³; bulk silica's ~2,000 would give about 58 µs. - From their numbers the ballistic length is about 20 nm. A free bead would be rough above that. In the trap, @@ -90,8 +90,8 @@ Both kinds of path are smooth, then rough, then smooth. The quantum one is rough Brownian one is rough between v_rms·τ_p and whatever confines it or carries it along. "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 a micron across released from inside -pollen grains, which are about 100 µm long (Pearle et al. 2010). A whole grain does jiggle, but in water it moves +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 @@ -155,7 +155,8 @@ outside the high-temperature regime this holds in. For the virus in air (Kn ≈ 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 + π/8. For sand and the cat, much larger than air's +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. @@ -188,8 +189,8 @@ npm test # includes tests/quantum-paths.mjs | 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, measured by Davisson and Germer | 1.65 Å | 1.67 Å | -| Bead in air, τ_p at 99.8 and 2.75 kPa (Li et al. 2010) | 48 and 147 µs | 47 and 154 µs | +| 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 and 2.75 kPa (Li et al. 2010) | 48.5 and 147.3 µs | 47 and 154 µs | | Free bead with those numbers: D over 0.04 nm and 1.3 µm | 1 and 2 | 1.000, 1.999 | | Continuously recorded particle: purity, and W at its width | ħ²/4, 1/8 | exact | | Gas rate against drag (Schlosshauer 2019, Epstein 1924) | equal | equal | @@ -217,7 +218,7 @@ npm test # includes tests/quantum-paths.mjs - 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 - M. J. Lake, Eur. Phys. J. C 82, 928 (2022), which restates 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. Ding, Y. C. Ong and H. Xu, *Measurement-induced perturbations of Hausdorff dimension in quantum paths* (2025; Eur. Phys. J. Plus, 2026). https://arxiv.org/abs/2512.13046 +- 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 diff --git a/quantum-paths/paths.js b/quantum-paths/paths.js index ee19b81..f6f0e8a 100644 --- a/quantum-paths/paths.js +++ b/quantum-paths/paths.js @@ -77,7 +77,7 @@ 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. + // 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); diff --git a/tests/quantum-paths.mjs b/tests/quantum-paths.mjs index 3f64c57..9e34832 100644 --- a/tests/quantum-paths.mjs +++ b/tests/quantum-paths.mjs @@ -52,18 +52,18 @@ within('Repeated records: recorded length over the single-interval length, N = 4 // 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), measured 1.65 Angstrom'); +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 3 um silica bead in an optical trap in air: momentum relaxation 48 us at 99.8 kPa and 147 us at -// 2.75 kPa, rms speed 0.42 mm/s. Model: Stokes drag with the Cunningham slip correction, using the mass the +// (2010)) held a 3 um silica bead 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 0.42 mm/s. Model: Stokes drag with the Cunningham slip correction, using the mass the // measured speed implies (density ~1,600 kg/m^3; bulk silica's ~2,000 would give ~58 us). 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: 1.5e-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), 48e-6, 1.1, LI); -within('Bead in air at 2.75 kPa: momentum relaxation time (s)', taup(2.75e3), 147e-6, 1.1, LI); -const bead = { sigmaV: vrms, taup: 48e-6 }, Sb = t => brownianS(t, bead); +within('Bead in air at 99.8 kPa: momentum relaxation time (s)', taup(99.8e3), 48.5e-6, 1.1, LI); +within('Bead in air at 2.75 kPa: momentum relaxation time (s)', taup(2.75e3), 147.3e-6, 1.1, 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`, From 5b392ca57d141243088e9e585063a072c83df401 Mon Sep 17 00:00:00 2001 From: Claude Date: Sun, 18 Oct 2026 17:15:30 +0000 Subject: [PATCH 3/3] Quantum paths: Li et al.'s 2.79 um trapped bead, Abbott-Wise attributions (#25 review) From the primary-source check on #25: - Li et al. used a 2.79 um diameter, not the nominal 3 um. With it the slip-corrected Stokes drag gives 51 us at 99.8 kPa (48.5 measured) and 175 us at 2.75 kPa (147.3), 19% high; the mass-free ratio is 3.43 against 3.04. The 2.75 kPa check is now a recorded disagreement (DIFF, as in tests/checks.mjs), and the 1,600 kg/m^3 density remark is gone. - The bead Li et al. measured was trapped (eq. 1, w0 = 2 pi x 3064 and 3168 Hz): its trap size, ~22 nm rms, is no more than its ballistic length, so its path is never rough. New trappedS() and a check: D comes within 0.1 of 2 for 0.037 and 0.025 decade only; a free bead needs ~970 us and ~127 nm. README and CORRECTIONS #18 no longer call the trapped path smooth-rough-smooth. - The D table, the 0.81 hbar/p crossover and "the dimension" at hbar/p are the model's local slope; Abbott and Wise give only the two limits and call D "not well defined" between them. The header, README, test label and check table now say so. - Abbott and Wise's own spacing rules are stated: fixed dt as dx -> 0 (hbar dt/2m dx^2 >> 1), and the self-similar dt ~ m dx^2/hbar, under which the heating gives D ~ 3. Lake's four wave clocks are his own. - The hbar/mc end is the model's (AW are non-relativistic); the root README cites AW only for the hbar/p half, and Cannata & Ferrari are not cited for the cutoff. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01C2bpiHyHSkC8DAL5CcyciK --- CORRECTIONS.md | 14 +++++-- README.md | 8 ++-- quantum-paths/README.md | 93 +++++++++++++++++++++++++---------------- quantum-paths/paths.js | 25 +++++++---- quantum-paths/table.mjs | 9 ++-- tests/quantum-paths.mjs | 44 +++++++++++++++---- 6 files changed, 129 insertions(+), 64 deletions(-) diff --git a/CORRECTIONS.md b/CORRECTIONS.md index 4544409..2fb953e 100644 --- a/CORRECTIONS.md +++ b/CORRECTIONS.md @@ -121,11 +121,17 @@ Claims made during the exploration that did not survive it, kept so the record s 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. The records also make the roughness: each kicks the momentum, the kicks add up, - and over N records the rough stretch widens to roughly √N ħ/p, so a long record is not one fixed fractal. + 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 a 3 µm 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. + 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 aa3fdec..badc519 100644 --- a/README.md +++ b/README.md @@ -19,10 +19,10 @@ with Claude (Anthropic), for [Foundation0](https://github.com/Foundation-0/f0). 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, between two lengths.** A free particle's path, recorded by position - measurements finer than ħ/p but coarser than ħ/mc, has fractal dimension 2, against 1 for a classical - trajectory (Abbott and Wise, 1981). 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/)). + 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 index 8eee16d..e12d71f 100644 --- a/quantum-paths/README.md +++ b/quantum-paths/README.md @@ -15,15 +15,18 @@ join the dots. Their abstract: "the observed path of a particle in quantum mecha 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). If the records are far enough apart for the wave -to outgrow its starting width (Lake 2022 restates the condition as Δt ≫ 4mΔx²/ħ; 4mΔx²/ħ, four of the Witness -Map's wave clocks, is the shortest spacing he allows), the mean length over a time T is +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: +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 | |---|---|---|---|---|---|---| @@ -31,19 +34,25 @@ and, in three dimensions: - 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. -- D passes 1.5 at Δx = 0.81 ħ/p (0.44 ħ/p for a path along one direction). The crossover takes about a decade. -- Records only one wave clock apart give the same two limits. The crossover then moves to 1.8 to 2.4 ħ/p, - depending on whether each record's own error is counted. +- 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. +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. 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: +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 | |---|---|---|---| @@ -63,14 +72,15 @@ 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 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, the length grows as Δx⁻²: an exponent - of 3, not 2. +- 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. Ding, Ong and Xu (2025) find measurement shifting -the dimension downward in a different (nonselective) model; the two are not compared here. +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 @@ -78,17 +88,24 @@ A Brownian particle is the reverse: smooth at small scales, rough at large ones. 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 3 µm silica bead 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 gives 47 and 154 µs, using the mass that rms speed implies. - That mass corresponds to a density of about 1,600 kg/m³; bulk silica's ~2,000 would give about 58 µs. -- From their numbers the ballistic length is about 20 nm. A free bead would be rough above that. In the trap, - the roughness is cut off again at the trap's own confinement. - -Both kinds of path are smooth, then rough, then smooth. The quantum one is rough between ħ/mc and ħ/p. The -Brownian one is rough between v_rms·τ_p and whatever confines it or carries it along. "The same roughness" is -right about the dimension, 2, but not about where it lives. +- 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 @@ -133,8 +150,8 @@ Three cases: 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 dimension - is about 1.4, halfway between smooth and rough. + 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. @@ -185,13 +202,15 @@ npm test # includes tests/quantum-paths.mjs | 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 (D = 1.5) | ħ/p, to a factor of 2 | 0.81 ħ/p | +| 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 and 2.75 kPa (Li et al. 2010) | 48.5 and 147.3 µs | 47 and 154 µs | +| 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 | @@ -201,9 +220,9 @@ npm test # includes tests/quantum-paths.mjs ## What the model leaves out -- Abbott and Wise's paper itself was not read: only its abstract and later restatements (Lake 2022; Granik - and Chapline 2003). Their exact equations and their rule for spacing the records are unchecked, and so is - Cannata and Ferrari's relativistic paper (Am. J. Phys. 56, 721 (1988)). +- 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 @@ -215,8 +234,8 @@ npm test # includes tests/quantum-paths.mjs ## 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 -- M. J. Lake, Eur. Phys. J. C 82, 928 (2022), which restates the Abbott–Wise argument. https://arxiv.org/abs/2206.00609 +- 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). diff --git a/quantum-paths/paths.js b/quantum-paths/paths.js index f6f0e8a..cab6b2d 100644 --- a/quantum-paths/paths.js +++ b/quantum-paths/paths.js @@ -1,13 +1,15 @@ /* 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, and D = 1 - dln/dln dx is its length-scaling - * (divider) dimension; for a path in three dimensions it is the Hausdorff dimension of the trace. Each record leaves - * a momentum spread hbar/(2 dx), and with records far enough apart for the wave to outgrow its start, + * 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) - * D = 2 when dx << hbar/p (the spread wins) and D = 1 when dx >> hbar/p (the drift wins). - * Relativity caps the speed at c, so below hbar/mc the length stops growing and D falls back to 1. - * A real Brownian particle is the reverse: smooth (ballistic) below v_th tau_p, rough above it. + * 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 @@ -71,6 +73,15 @@ // 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)); @@ -142,7 +153,7 @@ } const api = { - C, erf, meanNorm, awLength, awLengthRelativistic, localDimension, dimensionFromS, brownianS, stokesCunningham, + 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; diff --git a/quantum-paths/table.mjs b/quantum-paths/table.mjs index 09dedd1..d756de4 100644 --- a/quantum-paths/table.mjs +++ b/quantum-paths/table.mjs @@ -20,11 +20,14 @@ for (const [name, mass, v] of [['Electron at 54 eV', C.me, Math.sqrt(2 * 54 * C. } 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, label] of [[99.8e3, '3 um bead in air, 99.8 kPa'], [2.75e3, '3 um bead in air, 2.75 kPa']]) { - const tp = mb / QP.stokesCunningham({ r: 1.5e-6, eta: 1.81e-5, mfp: 68e-9 * 101.325e3 / P }); - console.log(`${label.padEnd(30)} tau_p ${(tp * 1e6).toFixed(0)} us, ballistic length ${(vrms * tp * 1e9).toFixed(0)} nm`); +for (const [P, meas, f] of [[99.8e3, 48.5e-6, 3064], [2.75e3, 147.3e-6, 3168]]) { + const tp = mb / QP.stokesCunningham({ r: 1.395e-6, eta: 1.81e-5, mfp: 68e-9 * 101.325e3 / P }); + console.log(`${(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 ${(vrms / (2 * Math.PI * f) * 1e9).toFixed(0)} nm`); } +console.log('The trap holds the bead within its ballistic length, so the measured path is never rough; a free bead would be.'); console.log('\nWho draws the path: how finely each Witness Map object\'s own surroundings record it, against its thermal hbar/p\n'); console.log('object'.padEnd(32) + 'records'.padEnd(14) + 'main witness'.padEnd(32) + 'resolution m'.padStart(13) + 'hbar/p m'.padStart(10) + 'ratio'.padStart(9)); diff --git a/tests/quantum-paths.mjs b/tests/quantum-paths.mjs index 9e34832..ed58051 100644 --- a/tests/quantum-paths.mjs +++ b/tests/quantum-paths.mjs @@ -1,18 +1,25 @@ // Checks the quantum paths model against published numbers and its own closed forms. Run: npm test. // Each check states the source, the published or expected value, the model's value and the tolerance. +// PASS: agreement within the tolerance. DIFF: a disagreement recorded in CORRECTIONS.md. A DIFF still passes, but only +// while the model's ratio to the published value stays inside the recorded range, so any change shows up here. import { createRequire } from 'node:module'; const require = createRequire(import.meta.url); const QP = require('../quantum-paths/paths.js'); const WM = require('../witness-map/model.js'); const { C, awLength, awLengthRelativistic, localDimension, dimensionFromS, brownianS } = QP; -let fails = 0; +let fails = 0, diffs = 0; const fmt = x => (x === 0 ? '0' : x.toExponential(2)); function within(name, got, want, factor, src) { const ok = got > 0 && got / want <= factor && want / got <= factor; console.log(`${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++; @@ -32,7 +39,7 @@ check('Dimension 2 well below hbar/p, 1 well above', Math.abs(localDimension(L3, `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, - `D = 1.5 at ${x15.toFixed(3)} hbar/p`, AW); + `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); @@ -55,19 +62,38 @@ within('C60 at 220 m/s: h/mv (m)', QP.deBroglie(720 * C.u, 220), 2.5e-12, 1.02, 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 3 um silica bead 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 0.42 mm/s. Model: Stokes drag with the Cunningham slip correction, using the mass the -// measured speed implies (density ~1,600 kg/m^3; bulk silica's ~2,000 would give ~58 us). +// (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: 1.5e-6, eta: 1.81e-5, mfp: 68e-9 * 101.325e3 / P }); +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); -within('Bead in air at 2.75 kPa: momentum relaxation time (s)', taup(2.75e3), 147.3e-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 ${(Math.sqrt(Sb(1e-1)) * 1e6).toFixed(1)} um`, - 'model (free bead; the trapped bead is bounded above by its trap), tau_p from ' + LI); + 'model (free bead), tau_p from ' + LI); +// The bead they measured was trapped (eq. 1, w0 = 2 pi x 3064 Hz at 99.8 kPa and 3168 Hz at 2.75 kPa). Its trap holds +// it within about 22 nm rms, no more than its ballistic length, so its path never settles into D = 2. +const slopeD = (S, tau) => { const e = 1e-4, a = Math.log(S(tau * (1 + e)) / S(tau * (1 - e))) / Math.log((1 + e) / (1 - e)); return a > 0 ? 2 / a : NaN; }; +function roughDecades(S) { // longest stretch of log10 tau over which the local D = 2 / (dln S / dln tau) is within 0.1 of 2 + let best = 0, run = 0; const step = 1e-3; + for (let lt = -8; lt <= -2; lt += step) { if (Math.abs(slopeD(S, Math.pow(10, lt)) - 2) < 0.1) best = Math.max(best, run += step); else run = 0; } + return best; +} +const trap = [[99.8, 0.422e-3, 48.5e-6, 3064], [2.75, 0.425e-3, 147.3e-6, 3168]].map(([P, sigmaV, tp, f]) => { + const o = { sigmaV, taup: tp, w0: 2 * Math.PI * f }; + return { P, size: sigmaV / o.w0, ballistic: sigmaV * tp, rough: roughDecades(t => QP.trappedS(t, o)) }; +}); +const tFree = crossing(t => 2 - slopeD(Sb, t), 1e-6, 1, 0.1), xFree = Math.sqrt(Sb(tFree)); +check('The trapped bead Li et al. measured has no rough stretch: D is within 0.1 of 2 for under 0.05 of a decade', + trap.every(t => t.rough < 0.05) && xFree > 3 * Math.SQRT2 * trap[0].size, + trap.map(t => `${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. @@ -99,5 +125,5 @@ const rbCold = QP.drawnPath(WM, WM.OBJECTS.find(o => o.id === 'rb'), WM.ENVS.uhv 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'); +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);