Cathodoluminescence Study of a Quantum Dot in a Nanowire for Single-Photon Emission

TL;DR

This work demonstrates that cathodoluminescence in a scanning electron microscope can be used not only to image and spectrally analyze a single CdSe quantum dot embedded in a ZnSe nanowire, but also to access quantum‑optical properties such as photon antibunching via a Hanbury Brown–Twiss interferometer. By combining CL imaging, hyperspectral mapping, and time‑resolved CL, the authors quantify diffusion‑limited carrier excitation, identify exciton complexes with energies around $2.40$–$2.38$ eV, and extract a phonon temperature of $T_{\text{ph}} = 12 \pm 2$ K under electron beam heating. The measured antibunching, with $\tilde{g}^{(2)}(0) \approx 0.25$, together with a fast exciton decay time of $τ_X \approx 0.9$ ns, confirms single‑photon emission from a single QD under CL excitation; the work also highlights the potential of e‑beam control to optimize emitter purity and enable high‑resolution, on‑chip single‑photon sources. Overall, CL provides a thermally comparable, spatially precise, and chemically versatile route to study and optimize novel single‑photon emitters for integrated photonics beyond traditional near‑IR systems.

Abstract

Cathodoluminescence in a scanning electron microscope was applied to a semiconductor quantum dot in a nanowire able to emit single photons. We show that cathodoluminescence can be used not only for imaging and spectroscopy, but also to measure the correlation function and characterize the purity of the single-photon emitter. The electron beam can be manipulated to minimize the collection of parasitic luminescence. At cryogenic temperatures, we observed that the thermal budget, as measured via the phonon sidebands, is close to that of non-resonant micro-photoluminescence. This makes cathodoluminescence an efficient tool in the quest of novel single-photon sources.

Figures (5)

• Figure 1: (a) CL setup with inset showing electron trajectories calculated with the CASINO software drouin_casino_2007 for 5 keV electrons impacting a thick ZnSe slab at 65°. (b) Panchromatic CL image taken with 65° tilt, recorded at room temperature with 5 kV acceleration voltage and 50 pA current. (c) SEM image of the same region, obtained with a Zeiss Ultra+ SEM, partially superimposed on the panchromatic image.
• Figure 2: (a) Preliminary SEM image of the NW taken with $80^\circ$ tilt (Zeiss Ultra+). The scale bar shows 1 µ m. (b) SEM image of the base of the emitter. The y-axis length is adjusted to account for the $65^\circ$ sample tilt. The corresponding CL map, recorded at 2.39 eV, is superimposed in green. The scale bar shows 200 nm. (c) CL profile (green symbols) obtained from (b) by integrating the CL signal over a 120-nm width across the NW, and fitted by an asymmetric normal-Laplace distribution. The vertical scale is the position along the NW, taking into account the 65° tilt. (d) Exciton (X) and charged exciton (CX) intensities integrated from the spectra shown in (e), and X/CX intensity ratio. (e) Spectrum measured at 5 K from the e-beam scan over the red area indicated in panel (b). We used 1800 grooves/mm, slits 0.05 mm, 5 s exposure time, 5 kV acceleration voltage and 21 pA current.
• Figure 3: (a) Intensity spectrum of a CdSe/ZnSe QD-NW at $T_\mathrm{c ryo}=5\,\mathrm{K}$ under electron-beam excitation (current of 21 pA). Stokes ($I_{\text{S}}$, orange) and anti-Stokes ($I_\text{AS}$, blue) intensities are separated by the ZPL (black line). (b) Stokes and anti-Stokes phonon sidebands. (c) Temperature extraction from the ratio $I_\text{S}/I_\text{AS}$.
• Figure 4: (a) Decay time of the neutral exciton acquired at 5 K using 5 kV acceleration voltage and a current of 960 pA with 1 MHz excitation rate. After subtraction of a constant baseline, the solid magenta line shows bi-exponential decay curve with fast decay time $\tau_X=0.9$ ns and a longer one of $14$ ns. The inset shows the exciton emission spectrum recorded with a tunable filter and a 7 mm slit width, under the same beam current and acceleration voltage conditions. (b) Exciton autocorrelation (acquisition time 600 s, 5 kV acceleration voltage, 960 pA beam current) obtained from an e-beam scan over a fixed area similar to the red rectangle in Figure \ref{['figure2']}(b). The magenta curve shows a fit based on the convolution of a Laplace distribution with the HBT instrumental response, with the adjustable parameters $T_\text{CL} = 0.6$ ns and $B/S\approx0.14$.
• Figure 5: Schematic of the SEM setup used for CL and TRCL measurements.