[Paper Review] Optical Analogue of the Dynamical Casimir Effect in a Dispersion-Oscillating Fibre
The paper demonstrates an optical analogue of the dynamical Casimir effect in dispersion-oscillating photonic crystal fiber, showing vacuum-origin photon pairs via Kerr nonlinearity and confirming non-classical correlations.
The dynamical Casimir effect is the generation of pairs of real particles or photons from the vacuum as a result of a non-adiabatic change of a system parameter or boundary condition. As opposed to standard parametric amplification where the modulation occurs both in space and in time, this fundamental process requires a pure modulation in time, which makes its detection particularly challenging at optical frequencies. In this paper we experimentally demonstrate a realisation of the optical analogue of the mechanical dynamical Casimir effect in the near-infrared optical regime in a dispersion-oscillating photonic crystal fibre. The experiments are based on the equivalence of the spatial modulation of the fibre core diameter to a pure temporal modulation when this is considered in the co-moving frame of the travelling pump pulse. We provide evidence of optical dynamical Casimir effect by measuring quantum correlations between the spectrally resolved photon pairs. The non-classical nature of the measured light is supported by evidence of anti-bunching photon statistics.
Motivation & Objective
- Motivate and realize an optical analogue of the dynamical Casimir effect (DCE) using a dispersion-oscillating fiber (DOF).
- Explain how spatial core-diameter modulation maps to temporal modulation in the co-moving frame of a pump pulse.
- Provide evidence of DCE by measuring quantum correlations between spectrally resolved photon pairs.
- Demonstrate non-classical light via anti-bunching statistics and heralded g(2)(0) measurements.
Proposed method
- Model the DCE condition in a dispersive medium with a temporally modulated refractive index proportional to χ(3)|E|^2 in the co-moving frame.
- Derive the DCE frequency condition in the lab frame as |Δω−v_g k| = mK/2 + v_g k_0, and simplify to β2Δω^2 + (1/12)β4Δω^4 = mK with energy-conservation implying symmetric ±Δω.
- Use a dispersion-oscillating fiber with periodic GVD modulation and a short pump pulse (600 ps) to induce the effect.
- Filter and spectrally resolve signal and idler photons around predicted wavelengths (e.g., 954 nm and 1173 nm) and measure coincidences with SPADs.
- Quantify non-classicality via coincidence-to-accidental ratio (CAR) and heralded g(2)(0) measurements, controlling Raman background.
Experimental results
Research questions
- RQ1Can DCE-like photon pair production be realized in an optical medium via a temporal modulation realized by a co-moving frame of a short pump pulse?
- RQ2Do spectrally resolved signal-idler photon pairs exhibit non-classical correlations consistent with vacuum-seeded DCE?
- RQ3How do dispersion properties (β2, β4) and fiber modulation shape determine the DCE emission frequencies?
- RQ4Is the observed light anti-bunched and does g(2)(0) fall below 1 for CAR>1, indicating non-classical emission?
Key findings
- Photon pairs are emitted at wavelengths predicted by the DCE condition, with clear quantum correlations evidenced by a large N_s,i(0) peak at zero delay.
- CAR remains significant for wavelength pairs satisfying the DCE-derived condition and diminishes otherwise.
- Best CAR (~5) and highest counts occur for λ_s=954 nm and λ_i=1173 nm under chosen pump parameters.
- g(2)(0) < 1 is observed for CAR > 1, indicating non-classical heralded emission.
- Finite Raman background is accounted for, with estimates showing DCE photon pairs are about 2×10^-3 per pump pulse at P_p=0.03 W (versus ~0.18 Raman photons).
- Results align with the DCE-derived spectral predictions and quasi-phase-matching intuition from parametric amplification.
Better researchstarts right now
From reading papers to final review, dramatically reduce your research time.
No credit card · Free plan available
This review was created by AI and reviewed by human editors.