Quantum Laboratory - 2D

Momentum distribution |phi(k)|^2

Experiment result

Press "Tunnelling experiment": a fresh packet is run through the current barrier with the same solver and the transmitted probability is compared with the analytic transmission coefficient averaged over the packet's k-spread.

Live readouts

Independent numerical checks

The science

2D twin. This page runs exactly the same solver, qubit algebra and Bell/CHSH code as the 3D twin (one shared file, assets/core.js), so every number is identical. Here the state is drawn as plots: phase-coloured probability density with V(x) and <H> on one energy scale, a rotatable Bloch sphere with its three coordinate-plane projections, and the correlation curve E(a,b) with the CHSH bound chart.

Wave packet. Units are hbar = m = 1. The state obeys i d psi/dt = [ -1/2 d^2/dx^2 + V(x) ] psi. It is advanced with the Strang split-step Fourier scheme psi -> e^{-iV dt/2} F^-1 e^{-i k^2 dt/2} F e^{-iV dt/2} psi on a 1024-point grid (own radix-2 FFT). The scheme is unitary, second order in dt and exact for V = 0.

Initial state: psi = (2 pi sigma^2)^(-1/4) exp(-(x-x0)^2/(4 sigma^2) + i k0 x), so |psi|^2 has standard deviation sigma and mean momentum k0. Free spreading: sigma(t) = sigma sqrt(1 + (t/(2 sigma^2))^2). Harmonic trap: <x>(t) = x0 cos(wt) + (k0/w) sin(wt) and the packet revives after T = 2 pi/w.

Tunnelling. Walls are smoothed over 0.12 length units (tanh) so the grid resolves them; this makes the effective barrier very slightly narrower than the sharp-wall formula, a difference well below 1 percentage point. For a rectangular barrier and E < V0: T = 1/(1 + V0^2 sinh^2(kappa a) / (4 E (V0 - E))), kappa = sqrt(2(V0 - E)); for E > V0 sinh becomes sin(k2 a) with k2 = sqrt(2(E - V0)) and the denominator 4E(E - V0). A real packet has a k-spread sigma_k = 1/(2 sigma), so the experiment compares the simulated transmitted probability with T(E) weighted by the packet's |phi(k)|^2.

Measurement. "Measure position" draws x from |psi|^2 (Born rule) and multiplies psi by a Gaussian of width 0.5 around the result (finite-resolution collapse), then renormalises.

Qubit. |psi> = alpha|0> + beta|1>, Bloch vector (x,y,z) = (2 Re a*b, 2 Im a*b, |a|^2 - |b|^2). Gates are exact 2x2 unitaries (H, X, Y, Z, S, T, Rx/Ry/Rz). The Rabi drive integrates i d|psi>/dt = (Omega/2 (cos phi sx + sin phi sy) + Delta/2 sz)|psi> with RK4 and is checked against P1 = Omega^2/(Omega^2+Delta^2) sin^2(sqrt(Omega^2+Delta^2) t/2).

Bell test. For the state (|00>+|11>)/sqrt2 and observables A(theta) = cos theta Z + sin theta X, the correlation is E(a,b) = cos(a-b). The CHSH sum S = E(a,b) + E(a,b') + E(a',b) - E(a',b') reaches 2 sqrt2 = 2.828 for a=0, a'=pi/2, b=pi/4, b'=-pi/4, while any local hidden-variable model gives |S| <= 2.

Simplifications. 1D, single particle, no spin or decoherence in the packet mode; absorbing layers (a smooth mask on the outer 10 % of the box) remove outgoing probability, which is tracked as "absorbed". The qubit is ideal (no T1/T2). Shot noise in the Bell test is real sampling noise, not added by hand.