Light’s Internal Timing Errors Corrected for Secure Quantum Networks

Ismail Nassar and colleagues from the Technion–Israel Institute of Technology have restored polarization entanglement from solid-state photon sources. A photonic-compensation protocol directly addresses phase evolution imprinted on emitted light by quantum emitters. The approach overcomes the suppression of observable entanglement caused by stochastic emission times and finite detector timing resolution. Applying synchronized, time-dependent coherent operations to photon reverses accumulated phase, recovering polarization entanglement without temporal post-selection. This provides a scalable pathway towards strong entangled-photon sources for quantum communication and photonic information processing in integrated platforms.

Mitigating deterministic phase evolution in integrated photonic entanglement sources

Entangled-photon sources are under investigation, and a strategy has been developed for removing the imprint of coherent emitter dynamics on photonic entanglement in integrated platforms. Quantum states of light are vital resources for quantum technologies, spanning single-photon states, entangled photon pairs, and multipartite photonic entanglement, underpinning secure communication, quantum-state teleportation, and photonic quantum computation. Realising sources that are simultaneously deterministic, high-fidelity, and scalable in solid-state platforms presents a key challenge.

A common obstacle in many quantum emitters is that coherent internal evolution during photon emission imprints a deterministic, time-dependent phase on the emitted light. Level splittings and related dynamics map onto the emitted photonic state as a time-dependent relative phase between radiative pathways. When experiments average over stochastic emission times, or when detector timing resolution is finite, this phase becomes effectively unresolved, resulting in a reconstructed photonic state appearing mixed and observable entanglement being suppressed.

Researchers are pursuing a conceptually new route to overcome this obstacle by reversing the phase imprint after emission using synchronized, time-dependent photonic operations. These operations undo the deterministic phase evolution mapped onto the photons and convert a non-stationary photonic quantum state into a stationary one on an event-by-event basis, without temporal post-selection. Semiconductor quantum dots provide a paradigmatic setting for this problem and an attractive platform for on-demand entangled-photon generation via the biexciton-exciton radiative cascade.

Ideally, this cascade emits a maximally entangled polarization Bell state. In practice, exciton fine-structure splitting lifts the degeneracy of the intermediate exciton states and induces spin precession, imprinting a well-defined phase evolution on the two decay paths. Averaged over the random biexciton-to-exciton delay, this phase encodes which-path information and washes out the observable polarization entanglement. Considerable effort has focused on suppressing the fine-structure splitting below the radiative linewidth using strain tuning, applied electric and magnetic fields, and engineered cavity architectures.

These approaches, while powerful, require careful device optimisation, typically operate over limited ranges, and are difficult to implement uniformly across large ensembles. Residual coherent dynamics often remain and become detrimental whenever detector timing cannot fully resolve the evolution. Photonic compensation is implemented for a single quantum dot driven by deterministic two-photon excitation of the biexciton, applying dynamic phase modulation synchronized to the excitation clock and matched to the exciton precession frequency set by the fine-structure splitting.

Time-resolved, full two-photon polarization-state tomography shows that dynamic phase modulation restores a stationary two-photon polarization state and recovers high-fidelity polarization entanglement without temporal filtering and independently of detector timing resolution. This approach applies whenever coherent emitter evolution maps onto the emitted photons as a deterministic phase referenced to a clock, providing a scalable route to strong entangled-photon sources in solid-state and integrated photonic platforms. In the biexciton-exciton cascade of a semiconductor quantum dot, coherent emitter dynamics, specifically the exciton fine-structure splitting ∆FSS, map onto the emitted photons as a deterministic, time-dependent relative phase between the two radiative pathways.

Because the biexciton-to-exciton delay varies stochastically from event to event, and because detector timing resolution is finite, experiments effectively average over this phase evolution, washing out polarization coherences and suppressing the observable entanglement. The resulting two-photon state can be written as |of(t)⟩= 1/√2 |HXXHX⟩+ e−iωX(tX−tXX)|VXXVX⟩, where ωX = ∆FSS/ħ is the exciton precession frequency and tXX (tX) denotes the biexciton (exciton) emission time. Instead of suppressing the fine-structure splitting at the emitter, the phase imprint is compensated directly in the photonic domain.

Specifically, an electro-optic modulator implements a phase φ(t) between the H and V polarization components, synchronized to the excitation clock. A linear phase ramp with slope φ = ωX cancels the exciton-induced phase evolution on an event-by-event basis, yielding a time-independent two-photon state, |of⟩= 1/√2 |HXXHX⟩+ |VXXVX⟩, which is the Bell state |Φ+⟩. This photonic-compensation concept restores polarization entanglement without modifying the quantum dot. Polarization-resolved photoluminescence from a single quantum dot under resonant two-photon excitation of the biexciton was observed.

The excitation laser is tuned to the two-photon resonance midway between the XX0 and X0 transitions, enabling deterministic preparation of the biexciton state. The measured photoluminescence spectrum exhibits the expected linearly polarized XX0 and X0 doublets, separated by a fine-structure splitting of ∆FSS = 8.80 ±0.04 μeV. This splitting sets the exciton precession frequency ωX = ∆FSS/ħ and corresponds to a precession period of Tp = h/∆FSS = 470 ±2ps. To implement this compensation in practice, the applied phase must remove the stochastic relative phase ωX(tX −tXX). This is achieved using two complementary phase ramps referenced to the same excitation clock: one ramp cancels the −ωXtX term during the X0 wavepacket, and the other cancels the +ωXtXX term during the delayed XX0 wavepacket. The XX0 and X0 photons are spectrally separated and routed through a delay line that imposes a well-defined differential delay of ∆t = 1.9ns. The two photons are then recombined into a common spatial mode and directed into a fibre-based Mach-Zehnder interferometer.

A polarizing beam splitter separates the H and V components into different arms, and an electro-optic modulator placed in the V arm applies the programmed dynamic phase modulation synchronized to the excitation pulses. After the interferometer, a narrow band-pass filter transmits the biexciton photon and reflects the exciton photon, enabling independent polarization analysis using liquid-crystal variable retarders and a polarizing beam splitter. Photon arrival times are recorded using superconducting nanowire single-photon detectors and a time-tagging module.

Active stabilisation of the Mach-Zehnder interferometer maintains phase stability throughout the long integrations required for tomography. The temporal waveform driving the electro-optic modulator was also examined. Each modulation cycle contains eight excitation pulses: the first four serve as an unmodulated reference, while the subsequent four implement the programmed dynamic phase modulation. Prior to the arrival of the X0 photon, and throughout its radiative decay, the electro-optic modulator voltage is ramped down linearly at a slope matched to ωX, applying a phase +ωXt across the X0 wavepacket, thereby cancelling the factor e−iωXtX. Before the delayed XX0 photon arrives, the voltage is ramped up with the opposite slope, applying a phase −ωXt that cancels the factor e+iωXtXX. The combined waveform therefore removes the full stochastic phase ωX(tX −tXX), stabilising the two-photon polarization state for every emission event.

The induced phase evolution was verified by transmitting a continuous-wave, linearly polarised laser through a Mach-Zehnder interferometer and projecting the output onto horizontal, vertical, diagonal, and anti-diagonal polarisation bases. The inferred phase evolution accurately tracks the programmed waveform, apart from a constant offset resulting from the interferometer’s bias point. Dynamic phase modulation was further verified using exciton fine-structure splitting in a semiconductor quantum dot.

Time-resolved, polarisation-sensitive photoluminescence reveals clear oscillations in the exciton emission under fine-structure splitting, which are removed when a specific portion of the waveform is applied. Thus, dynamic phase modulation dynamically suppresses the exciton precession and yields a time-independent exciton polarization during its radiative lifetime. The temporal separation of the XX0 and X0 photons by ∆t = 1.9ns using a pair of dichroic mirrors and a schematic of the Mach-Zehnder interferometer with an integrated phase modulator in one arm were shown.

A histogram of the induced phase measured using a D-polarized continuous-wave laser tuned to the exciton wavelength was also presented. A histogram of time-resolved photoluminescence, showing eight consecutive emission pulses projected into four polarizations, was displayed. Insets show the sum over the first four pulses and the last four pulses. Coincidence maps as a function of detection times for representative polarization bases were also shown: collinear (HH) without dynamic phase modulation, co-circular (RR) without dynamic phase modulation, and co-circular (RR) with dynamic phase modulation.

The capital letters denote the biexciton and exciton polarization projections, and the colour scale indicates the number of coincidences per 16ps × 16ps bin. Coincidence rate as a function of the arrival-time difference between X0 and XX0 photons for various polarization bases was also examined. A photonic-compensation protocol is demonstrated that removes emitter-induced phase evolution directly in the photonic domain. By actively correcting for the deterministic phase shifts imprinted on photons by internal emitter dynamics, a scalable method for generating strong entanglement without needing to carefully optimise device characteristics or rely on high-resolution timing has been created.

Active phase error correction rescues entanglement from semiconductor quantum dots

The foundations for scalable quantum networks are being built, demanding reliable sources of entangled photons. While this method offers a clever way to rescue entanglement from solid-state emitters, it currently relies on precise synchronisation and control of photonic elements. The question remains: how easily will this photonic-compensation protocol translate to other materials and quantum systems lacking such well-defined internal dynamics. Acknowledging that translating this photonic-compensation protocol to diverse materials presents a challenge, the demonstrated principle remains significant. The team successfully countered internal dynamics within a semiconductor quantum dot, exhibiting unique light-emitting properties that typically degrade entanglement quality. This work establishes a method for actively correcting phase errors in emitted photons, key for building strong quantum communication links and photonic technologies.

The researchers demonstrated a photonic-compensation protocol that successfully removed phase evolution caused by internal dynamics in a semiconductor quantum dot. This is important because such dynamics typically suppress entanglement, a crucial resource for quantum communication and photonic information processing. By applying time-dependent coherent operations to emitted photons, they restored a stationary two-photon polarization state and recovered polarization entanglement independently of detector timing resolution. The team suggest this approach offers a scalable route to robust entangled-photon sources and a strategy for removing the impact of emitter dynamics on photonic entanglement.