Zero-added-loss entanglement multiplexing using time-bin spectral shearing

High-quality quantum communications that enable important capabilities, such as distributed quantum computing and sensing, will require quantum repeaters for providing high-quality entanglement. To realize high-rate heralded entanglement for quantum repeaters, Chen et al. [Phys. Rev. Appl. 19, 054209 (2023)] proposed a scheme for heralded-multiplexed generation of quasi-deterministic entangled photon pairs, called zero-added-loss multiplexing (ZALM). Here, we propose a design of ZALM source using time-bin entanglement and spectral shearing. Additionally, we provide an analysis of experimentally relevant spectral-shearing parameters to optimize the spectral multiplexing. Moreover, we experimentally verify the compatibility of time-bin pulses and spectral shearing, as supported by observation of no phase shift when the same shearing is applied to both time bins. These results expand the benefits of applying a ZALM source to time-bin entanglement use cases. Moreover, more fully demonstrating time-bin and spectral shearing compatibility clears a path towards a broader use of spectral shearing that provides a deterministic frequency shift of high utility.

💡 Research Summary

The paper presents a concrete implementation of a zero‑added‑loss multiplexing (ZALM) entangled photon source that combines time‑bin entanglement with spectral shearing (serrodyne modulation). Building on the ZALM concept introduced by Chen et al., which relied on broadband SPDC and nonlinear up‑conversion to align the frequencies of multiplexed photon pairs, the authors replace the nonlinear frequency shifter with an electro‑optic phase modulator driven by carefully shaped RF waveforms. This substitution eliminates the need for additional lasers or nonlinear crystals, thereby reducing system complexity and loss.

The experimental architecture starts with a 460 MHz repetition‑rate, ~100 ps optical pulse train generated by two cascaded DC‑bias‑controlled amplitude modulators. The pulses are sent through an unbalanced Mach‑Zehnder interferometer (UMZI) to create two well‑defined time bins. A second wavelength serves as a phase‑reference for stabilizing the UMZI. The pump light is frequency‑doubled and filtered before pumping two spontaneous parametric down‑conversion (SPDC) crystals. The SPDC spectrum is engineered to straddle the C‑ and L‑bands, yielding one photon of each pair in each band. C‑band photons are interfered on a 50/50 beamsplitter, then sorted into discrete frequency bins using cascaded fiber‑Bragg gratings (FBGs) and circulators. L‑band photons are delayed and routed to an electro‑optic phase modulator driven by a ~1.4 GHz sine wave; the feed‑forward controller (FPGA) dynamically adjusts the slope and sign of the RF drive based on detection events, thereby applying a deterministic frequency shift to each time‑bin photon. After the modulator, a narrowband filter selects the zero‑th bin, producing heralded photon pairs that are subsequently converted from time‑bin to polarization qubits for state tomography.

A substantial portion of the manuscript is devoted to a quantitative model for spectral‑multiplexing optimization. The authors define the frequency‑bin width (δfb), transition band (δs), and the time‑bandwidth product (ΔfΔt) of the filter. They derive the required frequency‑bin spacing (ΔFb) and time‑bin spacing (Δtb) to avoid inter‑bin interference, imposing a conservative rule of thumb of >10 FWHM (≈24 σ) separation. Constraints on the RF drive frequency are derived for three waveform shapes (sawtooth, triangle, sine), ensuring that the entire optical pulse fits within the linear region of the phase‑modulation ramp. The maximum allowable drive frequencies are expressed as DA = 3 Δtb (sawtooth), DT = Δtb (triangle), and DS = Δtb (sine).

Using the drive power P, modulator π‑voltage Vπ, and the slope‑to‑frequency conversion Δf = A·Vπ, the authors calculate the achievable frequency shift for each waveform and, consequently, the number of multiplexed bins n. The resulting expressions (Eqs. 16‑18) show that, for realistic parameters (P ≈ 10 W, Vπ ≈ 1 V, δfb ≈ 12.5 GHz, ΔFb ≈ 2 GHz, ΔfΔt ≈ 0.89), the sawtooth waveform can support roughly 28 δfb √P σ/(Vπ ΔfΔt(δfb ΔfΔt + δs)) + 1 bins, i.e., on the order of dozens of frequency channels.

Voltage‑noise analysis reveals that a phase error of 5° corresponds to a tolerable RMS voltage fluctuation of ~28 mV for Vπ = 1 V, a level comfortably below the noise floor of state‑of‑the‑art RF drivers.

Experimentally, the authors verify the compatibility of spectral shearing with time‑bin qubits by applying identical shearing to both bins and measuring no observable phase shift in the interferometric output. This confirms that the shearing operation preserves the relative phase between time bins, a crucial requirement for maintaining high‑fidelity entanglement after frequency multiplexing.

Overall, the paper delivers a complete, experimentally validated blueprint for a ZALM source that leverages time‑bin entanglement and deterministic electro‑optic frequency shifting. By eliminating nonlinear up‑conversion, the scheme reduces loss and hardware overhead, while the detailed optimization framework provides clear guidance for scaling the number of multiplexed channels. The work therefore represents a significant step toward practical, high‑rate quantum repeaters and long‑distance quantum networks.