Wave 28: Debye Shielding, Bayesian Updating & Fabry–Pérot

Three new physics & statistics simulations land in Wave 28: plasma Debye shielding with Yukawa screened potential, interactive Bayesian updating with Beta-Binomial conjugate model, and the Fabry–Pérot interferometer with Airy resonances. All ship with EN + UK translations.

Platform Stats

462
Simulations
75
Categories
48
Devlogs
28
Waves

Wave 28 at a Glance

🔆

Debye Shielding

Yukawa screened Coulomb potential φ ∝ e−r/λ_D/r, Debye length λ_D, plasma frequency ω_p, and coupling parameter Γ. Three view modes: potential profile, 2D shielding cloud, and Debye sphere.

Try it →
📊

Bayesian Updating

Beta(α, β) prior updated by binomial coin-flip observations. Click to flip, watch posterior sharpen, read 95% credible interval — all in real time. Tunable prior and hidden true p.

Try it →
🔬

Fabry–Pérot Interferometer

Airy function T = 1/[1 + F·sin²(δ/2)] for resonant cavities. Sharp transmission peaks, FSR = c/2nd, finesse ℱ = π√R/(1−R), sweepable cavity length and wavelength.

Try it →

Debye Shielding

In a plasma, free electrons cluster around positive test charges and exponentially suppress the electrostatic field beyond the Debye length:

λ_D = √(ε₀ k_B T / n e²)     φ(r) = Ze/(4πε₀r) · exp(−r/λ_D)

The simulation offers three complementary views:

Controls expose log₁₀ n (m⁻³), temperature T in eV, and charge Z. The info bar reports λ_D in SI units (pm → µm depending on density), N_D, ω_p, and the coupling parameter Γ = e²/(4πε₀ a k_BT) where a = n−1/3.

Debye lengthYukawa potential plasma physicsscreened Coulomb plasma frequencycoupling parameter

Bayesian Updating

The Beta-Binomial model is the textbook conjugate pair for proportion inference. Choose a prior Beta(α, β) — encoding α pseudo-heads and β pseudo-tails observed before the experiment — then flip a coin whose true bias p is set by a slider (hidden from the “Bayesian”). After k heads in n flips the posterior is analytic:

Posterior = Beta(α + k,  β + n − k)

The canvas plots both distributions continuously, with an amber line for the true p and a teal 95% credible interval computed via the regularised incomplete Beta function (Lentz continued-fraction, 80-iteration bisection for quantiles). Four flip speeds and a 500-entry coloured flip-log make it easy to see convergence at a glance.

Bayes’ theoremBeta distribution conjugate priorcredible interval binomial likelihoodposterior convergence

Fabry–Pérot Interferometer

A Fabry–Pérot etalon is a pair of parallel, partially reflective mirrors separated by distance d. Light reflected back and forth interferes constructively only when the round-trip phase δ = 4πnd·cos(θ)/λ is a multiple of 2π, producing sharp resonance peaks described by the Airy function:

T(λ) = 1 / (1 + F · sin²(δ/2))    F = 4R/(1−R)²

The simulation visualises the transmission spectrum, the Free Spectral Range FSR = λ²/(2nd), and the finesse ℱ = π√R/(1−R) that quantifies how sharp the peaks are. Sliders allow sweeping mirror reflectivity R, cavity length d, refractive index n, and incidence angle θ to directly observe how each affects the resonance structure.

Airy functionfinesse free spectral rangeresonant cavity etalonmulti-beam interference

Engineering Notes

Debye shielding: SI-aware value formatting

Plasma density spans from laboratory discharges (~1020 m−3) to stellar interiors (~1032 m−3), so the Debye length swings from micrometres to femtometres. Rather than printing raw scientific notation, a lightweight fmtSI(v, unit) routine maps the SI prefix (f, p, n, µ, m, k, M, …) automatically by computing floor(log10(v)/3). A seeded Mulberry32 pseudo-random generator ensures the electron dots in the Debye-sphere view are reproducible across slider changes.

Bayesian updating: incomplete Beta via Lentz continued fraction

Computing 95% credible intervals requires the inverse regularised incomplete Beta function Ix(a, b). The forward CDF uses the Lentz modified continued fraction (200 iterations, convergence threshold 10−12), with the standard symmetry identity Ix(a,b) = 1 − I1−x(b,a) to select the more numerically stable branch. The quantile is solved by bisection over Ix (80 iterations, converges to 8 significant figures on the unit interval). The logGamma evaluator uses the 7-term Lanczos approximation, supporting α, β down to 0.5 without overflow.

Fabry–Pérot: smooth finesse rendering across all R

For high reflectivity (R → 0.99) the Airy peaks become extremely narrow (ℱ → 300). Rendering them faithfully requires at least ~20 sample points per peak. The simulation adaptively increases the wavelength resolution near peaks by pre-sampling at coarse resolution, locating the minima of (1 − T), then inserting refined sub-samples between adjacent minima. This keeps total evaluations under 2000 even at maximum finesse, well within a single animation frame.

What’s Next — Wave 29 Preview

Planned topics for Wave 29 include:

All Wave 29 simulations will ship with EN + UK pages on launch day.