Interfacing Spins in an InGaAs Quantum Dot to a Semiconductor Waveguide Circuit Using Emitted Photons

An in-plane spin-photon interface is essential for the integration of quantum dot spins with optical circuits. The optical dipole of a quantum dot lies in the plane and the spin is optically accessed via circularly polarized selection rules. Hence, a single waveguide, which can transport only one in-plane linear polarization component, cannot communicate the spin state between two points on a chip. To overcome this issue, we introduce a spin-photon interface based on two orthogonal waveguides, where the polarization emitted by a quantum dot is mapped to a path-encoded photon. We demonstrate operation by deducing the spin using the interference of in-plane photons. A second device directly maps right and left circular polarizations to antiparallel waveguides, surprising for a nonchiral structure but consistent with an off-center dot.

Figure 1

Prototype spin-photon interface and experimental setup. (a) Scanning electron microscope image of the device. Two orthogonal nanowire waveguides are excited at their intersection by a cw laser, linearly polarized at 45° to the $x$ axis. QD emission into the two waveguides is measured via outcouplers, which scatter light into the $z$ direction. $O_T$, $O_R$, $O_B$, and $O_L$ signify the top, right, bottom, and left outcouplers, respectively. (b) Schematic diagram of the optical setup used in the spectroscopy and interference experiments. (c) PL intensity map, integrated over the QD distribution, recorded from the spin-photon interface, by scanning the collection fiber with the excitation laser fixed at the intersection. (d) PL spectra recorded from a single Zeeman split QD line at the intersection and two outcouplers of device A. At the intersection the transitions of the Zeeman doublet are found to be right and left circularly polarized, as expected. From the orthogonal outcouplers, $O_R$ and $O_T$, both ${\sigma }^{+}$ and ${\sigma }^{-}$ transitions are observed. The orange (left) and green (right) shaded regions indicate the spectral windows, corresponding to ${\sigma }^{+}$ and ${\sigma }^{-}$ emission, respectively, collected by the two APDs in the interference experiments shown in Fig. 2.

Figure 2

Interference measurements of device A. (a) Interference fringes, for ${\sigma }^{+}$ and ${\sigma }^{-}$ light detected from $O_R$ and $O_T$, recorded as the piezomirror position is varied at $t_d=0$. (b) Intensity plot of the interference data presented in (a) showing that the ${\sigma }^{+}$ and ${\sigma }^{-}$ transitions are $\sim \pi$ out of phase at $t_d=0$ and can therefore be identified using an interference measurement in an in-plane architecture. The solid red line shows a fit of an ellipse to the data, which is used to extract the relative phase between the ${\sigma }^{+}$ and ${\sigma }^{-}$ transitions, from $\Delta \varphi ={cos}^{-1}[(A/B)\tan (\theta )]$. (c) The left-hand axis shows a plot of $\Delta \varphi$ as a function of time delay. The solid blue line is a triangular waveform fit to the data. The right-hand axis plots the visibility of the white light interference of the QD distribution, which is fitted with the function, $V(t)=V_0+V_1{e}^{(t_d-t_0)/\tau }$ to determine $t_d=0$. (d) Intensity plots of the interference fringes recorded simultaneously from the ${\sigma }^{+}$ and ${\sigma }^{-}$ transitions at different delay times of the interferometer. The solid red lines show elliptical fits to the data.

Figure 3

Photoluminescence measurements for device B and a numerical investigation into the effects of QD position on device operation. (a) PL spectra recorded from outcouplers $O_R$ and $O_L$ showing pronounced asymmetry between the two antiparallel directions. (b) PL spectra recorded from $O_B$ and $O_T$, again showing the pronounced asymmetry of (a). (c) FDTD simulations showing the contrast of ${\sigma }^{\pm }$ polarized light in each waveguide for a source located at a distance $s$ along the diagonal from the center of the intersection. The dashed gray line indicates the location of the source for the simulations shown in (f) and (g), which reproduces well the emission properties of device B. (d)–(g) Electric field intensity $|E{|}^2$ at the center of the waveguides for different source locations and polarizations 0.65 ps after the cw source begins to emit. (d) ${\sigma }^{-}$ polarized source located at center. (e) ${\sigma }^{+}$ polarized source located at center. (f) ${\sigma }^{-}$ polarized source located off center at $s=90\mathrm{nm}$. (g) ${\sigma }^{+}$ polarized source located off-centre at $s=90\mathrm{nm}$. (h) Diagram illustrating the position of the source within the waveguide intersection as used in the simulations shown in (c). The dashed lines indicate the reflection planes of the diagonal and antidiagonal virtual beam splitters.