Strain-Tunable GaAs Quantum Dot: A Nearly Dephasing-Free Source of Entangled Photon Pairs on Demand

We report on the observation of nearly maximally entangled photon pairs from semiconductor quantum dots, without resorting to postselection techniques. We use GaAs quantum dots integrated on a patterned piezoelectric actuator capable of suppressing the exciton fine structure splitting. By using a resonant two-photon excitation, we coherently drive the biexciton state and demonstrate experimentally that our device generates polarization-entangled photons with a fidelity of 0.978(5) and a concurrence of 0.97(1) taking into account the nonidealities stemming from the experimental setup. By combining fine-structure-dependent fidelity measurements and a theoretical model, we identify an exciton spin-scattering process as a possible residual decoherence mechanism. We suggest that this imperfection may be overcome using a modest Purcell enhancement so as to achieve fidelities $>0.99$, thus making quantum dots evenly matched with the best probabilistic entangled photon sources.

Figure 1

Erasure of the fine-structure splitting via a strain-tunable device. (a) Spectrum of a representative two-photon resonant excited GaAs quantum dot. The inset shows a sketch of the used six-leg device from the top. The sample (blue) with a solid immersion lens on top is bonded onto the piezoelectric actuator (golden part). The piezo is structured using three cuts (black areas) into six legs which are pairwise electrically connected on the backside (Legs 1–3). (b) Minimization of the FSS as described in the text by tuning the voltage on Leg 1 for $V2=0$ (red) and $V2=100\mathrm{V}$ (blue), respectively.

Figure 2

Two-photon density matrix of two representative GaAs quantum dots at zero fine-structure splitting. Real (a) and imaginary (b) parts of the measured two-photon density matrix for QD1 at zero fine structure splitting. (c), (d) Same measurement as in (a) and (b) but for a different quantum dot (QD2).

Figure 3

Near maximally entangled photons from quantum dots. (a) Autocorrelation measurement of the biexciton ($XX$) (red curve) and exciton ($X$) (blue curve) from QD1 ($g^{(2)}(0)$ at 102 ns). To improve readability, the $XX$ and $X$ curves are shifted by 5 ns. (b) Entanglement fidelity versus fine structure splitting ($S$) for QD1. The black data points represent the measurement data. The red curve is a fit according to Eq. (3), while the green curve is a fit taking into account a rotation of the state, as explained in the text. The blue curve takes into account background laser light (bg) but no spin dephasing $({\tau }_{ss}\to \infty )$. The inset shows the evolution of the fidelity versus the FSS for background-free entangled photons and no spin dephasing (orange line). The data points (black) are the measured data from the main figure, but corrected for the estimated background. Real (c) and imaginary (d) parts of the two-photon density matrix measured on QD2. The gray bars show the results calculated out of the raw data, while the red bars are related to the matrix after the correction for background photons, wave plate retardance, and dark counts.