Presets
Playback
Packet and barrier
Camera
Layers
Quality
Model check
The science
What you are looking at
An electron wave packet is fired along a track toward a rectangular barrier (red block). The glowing helix is the wavefunction itself: its height is Re ψ, its depth toward you is Im ψ, and the colour is the phase. The cyan profile on the back wall is the probability density |ψ|². When the packet reaches the barrier it splits: part reflects, part passes through even if E < V0. The green and orange columns are the running transmitted and reflected probabilities.
The model
iℏ ∂ψ/∂t = −(ℏ²/2m) ∂²ψ/∂x² + V(x)ψ for an electron (ℏ²/2m = 0.0381 eV·nm²), integrated with the split-step Fourier method (2048 points on 80 nm, Strang splitting, unitary). This is the 2D page's solver, copied verbatim. The initial state is a Gaussian packet, σx = 4 nm, mean energy E, starting at x = −20 nm. Plane-wave transmission for E < V0: T = [1 + V0² sinh²(κa) / (4E(V0−E))]−1, κ = √(2m(V0−E))/ℏ; for E > V0, sinh becomes sin and resonances with T = 1 appear. For κa ≫ 1, T ≈ 16(E/V0)(1−E/V0)e−2κa, the straight line in the last graph.
Honest simplifications
- The problem is one-dimensional. The helix is the complex ψ(x) plotted in 3D (it is not a particle moving in a spiral), and the back wall is a projection, not a second dimension of space.
- The barrier is only 0.1–2 nm thick while the packet is about 8 nm wide, so it is drawn at least 0.3 scene units wide (about 1.8 nm at this scale) when it would otherwise be a sliver; its true width is in the label (a star marks a barrier drawn wider than to scale) and in the numbers. The track length axis is 0.17 scene units per nm.
- A packet is not a plane wave: its energy spread smears T(E), so the simulated T differs slightly from the formula. Periodic boundaries are used, so each run stops before the packet wraps around. No spin, no many-body effects, rectangular barrier only.
Checks run in the page
- The 2D page's three checks: R + T + inside = 1, a wide packet reproduces the plane-wave formula within 0.03, and V0 → 0 gives T = 1.
- The positions of the helix beads and phase-sheet vertices, read back from the GPU buffers, equal (x, Re ψ, Im ψ) of the solver.
- The area under the density drawn on the wall is 1 (trapezoid on every third grid point, window ±38 nm).
Credits
Blender 5.2 test bench (rail with ruler, electron source, crystalline barrier block, detector columns, floor) with Cycles-baked textures; Three.js r160 drawing of the live split-step solution.