Efficient bidirectional quantum frequency conversion between telecom and visible bands using programmable III-V nanophotonic waveguides

Abstract

Quantum frequency conversion (QFC) is essential for interfacing quantum systems operating at different wavelengths and for realizing scalable quantum networks. Despite extensive progress, achieving QFC with simultaneous high efficiency, low pump power, minimal noise, broad bandwidth, and pump-wavelength flexibility remains challenging. Here, we demonstrate efficient, low-noise, and bidirectional QFC between the telecom (1550-nm) and visible (780-nm) bands using unpoled indium gallium phosphide (InGaP) $χ^{(2)}$ nanophotonic waveguides, eliminating the need for a long-wavelength pump. Leveraging the large nonlinear susceptibility of InGaP together with programmable modal-phase-matching control, we obtain record-low pump power (20 mW) -- an order of magnitude lower than that in previous demonstrations using integrated thin-film waveguides -- and record-high loss-inclusive normalized conversion efficiency among non-resonant QFC implementations. The platform maintains quantum coherence and entanglement of input photons with noise well below the single-photon level. These results mark a significant advance in integrated nonlinear photonics for high-performance QFC, facilitating the development of versatile and scalable quantum networks.

Figures (5)

• Figure 1: Programmable InGaP nanophotonic waveguide for QFC.a. Schematic illustration of QFC between the 1550-nm and 780-nm bands with an ancillary pump via SFG or DFG. b. False-colored scanning electron microscope (SEM) images of a phase-matching programmable InGaP nanophotonic waveguide integrated with a nanoheater array. Blue: InGaP waveguide. Yellow: gold electrode. c. Simulated mode profile overlaid on the SEM of the cross section of the waveguide.
• Figure 2: Phase-matching optimized SFG and DFG.a. Initial and optimized SFG spectra of a 6-mm waveguide. Inset shows the applied voltages of nanoheaters to optimize the phase matching. b. Internal SFG conversion efficiency versus the pump power after optimization. Solid line is model fitting. c. Initial and optimized DFG spectra. d. Internal DFG conversion efficiency versus the pump power after optimization. e. Peak conversion efficiency and corresponding pump power of the 2.5-mm and 6-mm waveguides. Error bars are due to uncertainty in the fiber-optic coupling efficiency. Solid lines are theoretical modeling. f. On-chip noise flux in the 1550-nm band at the wavelength separated from the pump by 20 nm ($-2.5$ THz) versus the pump power of the 6-mm waveguide. The line is a linear fitting. Inset illustrates the pump-induced Raman noise and QFC wavelengths.
• Figure 3: Time--energy entanglement of telecom and visible photons via QFC.a. Experimental schematic and setup. ES: entanglement source. TA: tapered amplifier. PPLN-WG: periodically-poled lithium niobate waveguide. WDM: wavelength-division multiplexer. FFP filter: fiber Fabry-Perot filter. PD: photodetector. TF: tunable filter. WM: wavelength meter. VOA: variable optical attenuator. SNSPD: superconducting nanowire single-photon detector. TCSPC: time-correlated single-photon counting module. SPAD: single-photon avalanche diode detector. b. Coincidence counts of the signal and idler in 150-ps binwidth integrated in 60 s for various glass MZI temperatures. c. Coincidence counts of the converted signal and idler. d. Time--difference histograms of the converted signal and idler for two glass MZI temperatures. Time-bin width is 50 ps.
• Figure 4: Down-conversion of time-bin qubit.a. Experimental schematic and setup. IM: intensity modulator. ASE: amplified spontaneous emission. For other acronyms see Fig. \ref{['fig::3']}. b. Density matrices of $\left|+\right\rangle$-like state before (top) and after (bottom) QFC. c. Density matrices of $\left|R\right\rangle$-like state before (top) and after (bottom) QFC. d. Fidelities of the six basis states after QFC.
• Figure 5: Comparison of non-resonant $\chi^{(2)}$-based QFC performance. Loss-inclusive normalized conversion efficiency is defined as $\eta_\mathrm{L,norm}\equiv\eta_{\mathrm{max}}/P_p/L^2$, where $\eta_{\mathrm{max}}$ is the peak internal conversion efficiency, $P_p$ is the corresponding pump power, and $L$ is the device length. Thin-film poled waveguide: fan2021photonwang2023quantum. Bulk poled waveguide: morrison2021brightniizeki2020twobock2018highrakher2010quantumates2012twozaske2012visiblevonChamier2025lowdreau2018quantumikuta2018polarization. Bulk crystal: mann2023lowkerdoncuff2021quantumbrevoord2025quantum.