High-purity frequency-degenerate photon pair generation via cascaded SFG/SPDC in thin film lithium niobate

TL;DR

The paper tackles the problem of generating frequency-degenerate photon pairs in integrated photonics without contamination from parasitic single-pump processes. It introduces a dual-pump cascaded SFG/SPDC mechanism implemented in a layer-poled thin-film lithium niobate waveguide, leveraging phase matching and modal engineering to funnel energy into degenerate SPDC while suppressing unwanted SP processes. The authors demonstrate on-chip brightness around $1.0\times 10^5$ Hz nm$^{-1}$ mW$^{-2}$ and achieve SP parasitic suppression exceeding 40 dB, with Raman scattering characterized and mitigated by pump detuning; acceleration of degenerate photon-pair production is shown in the dual-pump configuration, and comparison to DP-SFWM indicates competitive performance with a simpler, single-pass architecture. Overall, this work provides a scalable, telecom-compatible pathway to high-purity frequency-degenerate photon sources suitable for continuous-variable quantum information processing and precision metrology, potentially enabling broadband squeezed states with high purity.

Abstract

Frequency-degenerate photon pairs generated using nonlinear photonic integrated devices are a crucial resource for scalable quantum information processing and metrology. However, their realization is hindered by unwanted parametric processes occurring within the same phase matching band, which degrade the signal-to-noise ratio and reduce the purity of the associated quantum states. Here, we propose a dual-pump scheme to produce frequency-degenerate photon pairs, based on cascaded sum-frequency generation and spontaneous parametric down-conversion occurring within a single waveguide, while strongly suppressing parasitic photon pair generation from single-pump processes. This approach significantly simplifies the design compared to microresonator-based methods and enables both pumping and collection of photon pairs entirely in the telecom band. We experimentally validate the concept in a layer-poled thin film lithium niobate waveguide, achieving frequency-degenerate photon pair generation with a brightness of \SI{1.0(3)e5}{\hertz \per \nm \per \square \milli \watt } and a 40 dB suppression of unwanted single-pump processes.

Figures (3)

• Figure 1: (a) Cascaded SFG/SPDC process: dual-pump scheme where two input pumps at $\omega_1$ and $\omega_2$ create SFG at $\omega_\mathrm{SF} =\omega_1 + \omega_2$ which acts as the pump for frequency-degenerate SPDC. $\omega_\mathrm{SF}$ is centered on the SHG phase matching frequency. The single-pump SHG processes are suppressed as they fall outside of the phase matching bandwidth of SHG, $\Omega_{\mathrm{SHG}}$. (b) Ideal layer-poling for the MPM between the (c) fundamental TE$_{00}$ mode in the telecom band and the TE$_{01}$ mode in the near-visible. The LN ferroelectric domains at the bottom of the waveguide (dark blue) are reversed compared to the ones on top (light blue). The boundary sit at half the waveguide height. (c) Shows the real part of the transverse (TE) component of the electrical fields, normalized such that the maximum over the cross-section is equal to 1. (d) Theoretical and experimental SHG spectrum, from which we infer the SHG suppression away from phase matching.
• Figure 2: (a) Setup used to measure the emission spectrum from the chip. EDFA: erbium doped fiber amplifier, ASE filter: amplified spontaneous emission filter, WDM filters: wavelength division multiplexer, FBG: fiber bragg grating. (b) Emission spectrum from the TFLN chip centered around the pump wavelength (solid) and residual contribution from the fibers, obtained bypassing the chip (dashed). The highest peaks indicate the Raman shifts associated with the mode 251 1A1. The spectrometer intensity at the wavelengths indicated with vertical lines are recorded while increasing the pump power and shown in (c). The experimental data are fitted on a linear scale with the model $ax^2 + bx$, which includes a quadratic term (from cSHG/SPDC) and a linear term (from Raman scattering), and shown in panel c). The two dashed lines are a guide to the eye representing pure linear and pure quadratic scaling, illustrating the transition from a Raman-dominated to an SPDC-dominated regime. The Raman coefficient is 7 times higher at the 251 1A1 mode position than at 1502 nm.
• Figure 3: (a) Setup to measure the frequency-degenerate photon pairs from cSFG/SPDC in DP experiment. WS: tunable waveshaper in a notch filter configuration, CWDM: coarse-wavelength division multiplexer, BS: beam splitter, PC: polarization controller, SNSPD: superconducting nanowire single photon detector. (b) Estimated on-chip (detected) pair generation rate data on the left (right) axis, fitted with a linear regression in a log-log scale (solid line), scaling the power of one pump power and keeping the other constant at 0.17 mW. (c) CAR for the same power scaling. (d) CAR (on-chip PGR) as a function of the detection bandwidth on the left (right) axis. The CAR is fitted using $ax^{-1}$ with $a=388$, while the PGR is fit with a linear model $ax$. (e) Time-correlation histogram for frequency-degenerate photon pairs at $\omega_\mathrm{D}$ from cSFG/SPDC, measured as shown in the inset. TCSPC: time-correlated single photon counting (f) Time-correlation histogram for non-degenerate photon pairs from cSHG/SPDC, with one photon from each pair at $\omega_\mathrm{D}$, measured as shown in the inset. (g) PGR of the DP processes (linear fit: $1.95x -2$) and PGR of the SP processes (linear fit: $1.8x -43$), showing a relative suppression of 40 dB consistently for various pump power.