Experimental Proposal: Testing Frame Width of Coupling in Entangled Photon Correlations

Quantum entanglement correlations, while predicted probabilistically by quantum mechanics, are usually characterized in terms of the joint spectral and temporal amplitudes (JSA/JTA) of the source rather than by any explicit compact-support parameter. We propose an experimentally accessible, purely phenomenological ansatz: a finite Frame Width of Coupling (FWC) that enters as a compact-support kernel in the two-photon interference envelope. Operationally, the FWC is defined as a bounded temporal window in the JTA over which coincidence amplitudes remain phase-correlated and thus contribute to Hong–Ou–Mandel (HOM) visibility. The central question is empirical rather than ontological: does the HOM coincidence-versus-delay profile admit a compact-support component that is invariant under standard broadening mechanisms (filter bandwidth, pump chirp, crystal length), or is it fully accounted for by the usual JSA physics? We test this by comparing a standard model class (Gaussian and sinc-derived HOM envelopes with chirp and measured JSA) against an FWC class where the envelope is convolved with a finite rectangular kernel. The latter generically produces a flat-bottomed dip with steeper edges; critically, its plateau width is predicted to remain approximately invariant when the coherence time is broadened or narrowed, in contrast to standard models. We consider a type-II SPDC source (1 mm ppKTP, pumped at 405 nm via frequency-doubled ∼14 fs Ti:Sapphire pulses) generating ∼20 fs pump pulses at 405 nm. Polarization-entangled 810 nm pairs are analyzed in a HOM interferometer with a stabilized free-space delay line scanned over ±2 ps in 10 fs steps. With ∼5×10³ coincidences per point (total ∼2×10⁶ counts), the design is primarily powered to detect FWC plateaus with TFWC≳50 fs (high power in AIC/BIC discrimination) and acts mainly to set upper bounds on any significantly smaller compact-support scale (e.g., TFWC≲30 fs), which remains resolution-limited at the chosen 10 fs sampling. Model comparison using Poisson-likelihood AIC/BIC, an edge-steepness metric, and invariance tests under filter, power, chirp, and optional spatial-displacement sweeps either reveals an FWC-like plateau or strengthens the standard JSA-based description.