Arbitrary control of the temporal waveform of photons during spontaneous emission

Control of the temporal waveform of photons produced during spontaneous emission from single quantum emitters provides a crucial tool in the establishment of hybrid quantum systems, optimization of quantum state transfer protocols and mitigation of effects due interferometric instability for network architectures based on flying qubits. We describe a method to generate photons of any temporal waveform from emitters of any excited state lifetime, limited only by the timing resolution of control hardware. We show how the temporal waveform of photons can be controlled by deterministically varying the population of an excited state which undergoes spontaneous emission. Our broadly applicable approach has only two requirements for a candidate quantum emitter: modulation of the (1) amplitude and (2) relative phase of a field coupling a ground state to the excited manifold. We detail how to identify optimal excitation pulses by employing variational algorithms to feed back on atomic populations. Additionally, we develop Quantum Monte Carlo based tools to determine photon-number statistics and establish techniques to identify optimal excitation strengths and post-selection thresholds for photon generation protocols. We situate our work in the context of other prior research on bespoke single photon sources and networking including post-emission pulse shaping, temporal gating and cavity-based methods. In comparison, our free-space process has greater flexibility in producing any waveform, requires less infrastructure, and can be readily applied across a wide range of quantum emitters. We discuss the applications and limits of this technique, including how increasing photon emission probabilities affects achievable temporal-mode overlap fidelities between emitted and target photon waveforms.

💡 Research Summary

The paper addresses a central challenge in quantum networking and hybrid quantum systems: the ability to tailor the temporal mode of single photons emitted via spontaneous decay from a quantum emitter. Existing approaches either reshape photons after emission using dispersive optics, electro‑optic modulators, or post‑selection, or they engineer the emission in‑situ by embedding the emitter in a cavity and employing STIRAP‑type Raman processes. The former suffer from wavelength‑dependent hardware limitations and insertion loss, while the latter require specialized cavity infrastructure and are constrained by cavity linewidths, limiting the achievable temporal features and compatibility with multi‑mode protocols.

The authors propose a fundamentally different, free‑space method that requires only two controllable degrees of freedom of the driving field that couples a ground state |0⟩ to an excited state |1⟩: (1) the time‑dependent amplitude Ω(t) of the Rabi frequency and (2) the instantaneous phase ϕ(t), which is restricted to binary values {0, π}. By applying a sequence of amplitude envelopes interleaved with π‑phase flips, the population ρ₁₁(t) of the excited state can be driven to follow any prescribed function. Since the instantaneous photon emission rate into a given polarization q (σ or π) is I_q(t)=Γ_q ρ₁₁(t), shaping ρ₁₁(t) directly sculpts the photon’s temporal waveform g_q(t).

The theoretical framework is built on a three‑level Λ system, exemplified by the ²S₁/₂ ↔ ²P₁/₂ transition of a ¹⁷⁴Yb⁺ ion. The Hamiltonian in the rotating frame includes the complex Rabi term ℏΩ(t)e^{iϕ(t)}|0⟩⟨1|+h.c., and spontaneous decay is modeled with Lindblad operators reflecting the 2:1 σ:π branching ratio. Numerical integration of the master equation is performed with QuTiP, yielding ρ₁₁(t) for any prescribed Ω(t),ϕ(t).

To find the optimal control pulse, the authors employ variational optimization algorithms that minimize the infidelity 1‑F, where F=|∫g_q^*(t)g_target(t)dt|² is the overlap between the generated and target temporal modes. The optimizer searches over a discretized set of amplitude values and the timing of π‑phase flips, constrained by the bandwidth of the available arbitrary waveform generator (typically ∼1 ns resolution). The resulting optimal pulses consist of short, high‑amplitude excitation bursts separated by π‑flips that coherently de‑excite the atom, allowing precise sculpting of the emission envelope.

A crucial practical issue is the probability of emitting more than one photon during a single excitation attempt, which becomes non‑negligible when the pulse duration approaches the excited‑state lifetime. Multi‑photon events degrade entanglement fidelity in remote‑entanglement protocols. To quantify and mitigate this, the authors develop a Quantum Monte Carlo (QMC) simulation that tracks individual quantum jumps for many stochastic trajectories. From the QMC data they extract the mean photon number ⟨N⟩, the multi‑photon probability P_multi, and the temporal distribution of each photon. Using these statistics they define optimal operating points (e.g., average emission probability 0.3–0.5) and design post‑selection criteria that discard trials where a second photon is likely (identified by early detection followed by a short‑time re‑excitation signature). This post‑selection reduces the overall success rate modestly (≈10 % loss) but raises the fidelity of the remaining entangled states above 0.95.

The authors compare their free‑space scheme to post‑emission shaping and cavity‑assisted methods. Their approach is wavelength‑agnostic (applicable from UV to telecom), requires only a fast phase modulator and an AWG, and can generate essentially arbitrary waveforms limited only by the control electronics’ timing resolution. In contrast, dispersive pulse shapers are unavailable in the UV, and cavity‑based STIRAP imposes adiabaticity constraints and a minimum temporal feature size set by the cavity linewidth. The free‑space method also avoids the need for mode‑matching into a resonator, simplifying scaling to many nodes.

Limitations are discussed: the need for high‑speed, low‑jitter phase switching; potential distortions for sub‑nanosecond pulses due to electronic bandwidth; and the inevitable trade‑off between emission probability and multi‑photon error, which still requires post‑selection. The paper suggests future extensions such as adaptive real‑time feedback to refine pulse shapes on the fly, generalization to multi‑level qudit systems, and integration with fiber‑based distribution channels for long‑distance quantum communication.

In summary, the work demonstrates that by controlling only the amplitude and binary phase of a resonant drive, one can deterministically engineer the temporal profile of spontaneously emitted photons across a broad class of emitters. The combination of variational pulse design, master‑equation modeling, and QMC‑based error analysis provides a practical toolkit for building high‑fidelity, waveform‑customized single‑photon sources that are compatible with existing quantum network architectures while minimizing hardware overhead.