Towards Stirling cooler operable single-photon sources based on low-noise GaAs quantum dots

AI summary

Overview

Research area: Quantum photonics — semiconductor quantum-dot single-photon sources, carrier–phonon dephasing, and two-photon interference (Hong–Ou–Mandel interference). Technical level: Intermediate (assumes familiarity with single-photon sources, indistinguishability, and basic phonon physics). Scope: A combined experiment–theory study of how temperature up to 55 K degrades photon indistinguishability from the negative trion in a low-noise GaAs quantum dot, and a demonstration that Purcell enhancement restores high indistinguishability at Stirling-cooler temperatures.

What This Paper Is About

Epitaxial quantum dots emit highly indistinguishable single photons, but so far only when cooled to about 4 K with bulky helium cryostats. Compact Stirling coolers reach roughly 30 K while drawing only about 100 W, so it matters whether quantum-dot sources can still perform well at those temperatures. The paper measures exactly how temperature destroys indistinguishability in a GaAs quantum dot, identifies the physical mechanism responsible, and tests a cavity-based fix.

Key Contributions

1. Temperature-resolved Hong–Ou–Mandel measurements (4.5 K to 55 K) on the negative trion of a low-noise GaAs/AlGaAs quantum dot, performed both with the full spectrum and with the phonon sideband spectrally filtered out.
2. A microscopic theory — multiband k·p plus configuration-interaction calculations of the trion and excited trion states, with explicit carrier–phonon coupling — that reproduces the measured temperature dependence quantitatively without ad-hoc fitting of the temperature trend.
3. Identification of two distinct phonon channels: diagonal coupling to longitudinal acoustic phonons (which produces the phonon sideband) and phonon-mediated cycling through energetically close excited trion states (which broadens the zero-phonon line).
4. An experimental verification of the proposed remedy: a different quantum dot in a circular Bragg grating resonator with an estimated Purcell factor of 7.7(1) reaches ℐ = 0.800(28) at 32 K, which the authors state is the highest reported value at that temperature.

Main Findings

• Near-Fourier-limited emission at 4.5 K: The zero-phonon line is only 6(2)% above the Fourier limit, and the filtered indistinguishability is ℐ = 0.966(6) for photons separated by 12.5 ns.
• Strong temperature degradation: Unfiltered indistinguishability falls from 0.857(7) at 4.5 K to 0.000(29) at 50 K. Filtered indistinguishability falls from 0.966(6) at 4.5 K to 0.048(35) at 55 K (quoted as 0.05(4) in the abstract). The conclusion states the filtered value reaches zero at about 55 K and the unfiltered at 45 K.
• Spectral filtering helps: At 4.5 K the filtered indistinguishability is 0.109(9) higher than the unfiltered value, and the filtered data stay substantially higher across the whole temperature range, confirming the phonon sideband as a major dephasing channel.
• Emergent excited trion lines: At 32 K, four weak emission lines appear at ΔE = 5.1 meV, 7.0 meV, 9.8 meV and 11.1 meV, attributed to thermally occupied excited trion states; a simulated spectrum reproduces the same pattern at similar energies.
• Two-level theory only works below ~10 K: A model with diagonal phonon coupling alone (two-level system) matches the data for temperatures below 10 K. Adding the first excited trion state |T*⟩ as an effective three-level model extends the agreement to the full temperature range.
• Broadening explains the loss: Measured zero-phonon-line linewidth stays near the Fourier-transform limit of 2.71(3) µeV below 16 K and grows to 42.7(1.3) µeV at 55 K. Indistinguishability tracks Γ₀/Γ approximately linearly, confirming ℐ′ = Γ₀/Γ; the line shape remains Lorentzian, arguing against thermally activated charge noise.
• Extracted dephasing rates: γ_ZPL = 5.5(2) µeV at 32 K for the bare dot; γ_ZPL = 6.0(7) µeV for the cavity dot, 0.5 µeV higher, attributed to its excited state lying only 3 meV above the trion (versus 5.1 meV).
• Purcell enhancement works: With a lifetime of τ = 31(1) ps (Γ₀ = 21.2(7) µeV), indistinguishability at 32 K improves from 0.314(25) to 0.80(3), consistent with the predicted 0.779(21).
• Brightness also drops: A larger fraction of light goes into the phonon sideband, and stronger phonon damping of Rabi rotations lowers the trion population probability. The sideband contributes roughly 5% of the intensity at 4.5 K and roughly 30% at 32 K.

Methodology in Plain English

The researchers grew GaAs quantum dots by droplet etching epitaxy and embedded them in a p-i-n diode so that a applied voltage of 1 V fixes the charge state to exactly one electron, making the negative trion the relevant emitter. The sample sits in a closed-cycle cryostat with a 4.5 K base temperature and a PID-controlled heater for tuning. A pulsed Ti:Sa laser at 80 MHz with about 3 ps pulses and a pulse area of π resonantly excites the trion, and a cross-polarization scheme removes stray laser light.

To measure indistinguishability, the team sends two photons emitted 12.5 ns apart into an unbalanced Mach–Zehnder interferometer and records the Hong–Ou–Mandel coincidence histogram; the visibility is corrected for the measured multiphoton probability and for the beam splitter's 0.495/0.505 splitting ratio. They repeat this with a ±20 meV band-pass filter (unfiltered case, sideband included) and with narrow notch filters that keep only the zero-phonon line (filtered case). Linewidths come from Michelson interferometry.

On the theory side, they compute electron, hole and trion states with multiband k·p plus configuration interaction, based on an atomic-force-microscopy scan of a representative etched nanohole, and derive phonon coupling matrix elements for deformation-potential and piezoelectric coupling. Dynamics use a laser-driven few-level model with Lindblad radiative decay, process-tensor treatment of non-Markovian diagonal coupling, and Markovian Lindblad rates for phonon-assisted transitions into and out of the excited trion, weighted by Bose–Einstein occupation. The only fitting parameters are the electron and hole confinement lengths and the phonon relaxation rate γ⁰_ph.

Why This Matters

The work shows that the dominant obstacle to warm operation is not charge noise — the diode structure keeps the line Lorentzian and near the Fourier limit to 16 K — but phonon-mediated cycling through excited trion states. That reframes the engineering problem as one of cavity design and band structure rather than materials purity, and it gives a quantitative design target for warm single-photon sources.

Real-world applications:

• Quantum key distribution, which the paper explicitly names as a target application for on-demand indistinguishable photons.
• Linear optical quantum computing, where two-photon interference is the core resource.
• Field-deployable or space-constrained quantum photonics, where kilowatt helium cryostats are impractical and a ~100 W Stirling cooler is viable.
• Remote-entanglement and quantum-repeater links, which depend on photons from separate emitters being mutually indistinguishable.

Industry relevance: The paper connects a device-level metric (Purcell factor) to an application-level metric (indistinguishability above 0.9) and shows the Purcell factor of at least 17 needed for the dots studied here, versus 7.7(1) demonstrated. It also notes that InGaAs cavities have reached Purcell factors above 25 in prior work (Ref. [40]), suggesting the target is within reach of existing cavity fabrication.

Future Directions

• Push the Purcell factor to 17 or beyond for these GaAs dots, for example with optimized circular Bragg grating resonators, to exceed ℐ′ = 0.9 at around 30 K.
• Reduce γ_ZPL by increasing the excited-state energy separation through quantum-dot geometry and size; the paper notes that ℏγ_ZPL would need to fall to 0.31 µeV for the same target at fixed Γ₀, and cites prior work suggesting reduced dephasing for dots larger than 5 nm under a simplified model, while cautioning that this approximation is crude.
• Extend the comparison across more quantum dots to separate dot-to-dot variation (only two dots are compared directly) and to test whether the 3 meV versus 5.1 meV excited-state splitting explanation holds systematically.
• Demonstrate operation of a cavity-enhanced source on an actual Stirling cooler platform, and extend the temperature study to other transitions; the paper reports that similar measurements on the neutral exciton appear in the supplementary information, but those numbers are not given here.

Target Audience

Quantum-photonics experimentalists and theorists working on semiconductor single-photon sources; engineers designing compact cryogenic or cavity-integrated photonic hardware; and graduate students who want a well-matched example of combining microscopic phonon theory with Hong–Ou–Mandel benchmarking. Beginners will find the phonon-channel decomposition conceptually clear but will need background in two-photon interference to follow the visibility corrections.

Authors’ abstract

For photonic quantum technology applications, sources capable of emitting photons with indistinguishability close to unity are essential. Ideally, these sources should not require demanding cooling systems. Here, we present temperature-dependent two-photon-interference measurements on photons produced by the radiative decay of the negative trion in a low-noise GaAs quantum dot, which are in quantitative agreement with theoretical calculations accounting for carrier-phonon interactions and coupling to excited states. While at at the lowest explored temperatures the emission linewidth reaches values only 6(2) % above the Fourier limit and the indistinguishability I between subsequently emitted photons reaches 0.966(6), the latter drops to 0.05(4) at 55 K. We show that this loss can be explained with the coupling with energetically close excited trion states and suggest that the photon indistinguishability at elevated temperatures can be increased by employing Purcell enhancement of the emission rate or by increasing the energy separation of the excited states. Using cavity-enhanced emission, we experimentally verify the first route and demonstrate an improvement in photon indistinguishability from 0.314(25) to 0.80(3) at 32 K, which - to our knowledge - is the highest reported value at such temperature.