Strain-Engineered Deterministic Quantum Dots for Telecom O-Band Emission Using Buried Stressors

Abstract

The deterministic realization of quantum light sources operating at telecom wavelengths is essential for long-distance fiber-based quantum communication and distributed quantum computing. In this work, we demonstrate that telecom O-band emission can be achieved from site-controlled InGaAs/GaAs quantum dots (QDs). Our concept utilizes a buried AlAs/Al$_2$O$_3$ stressor layer with the unique feature that induces a well-defined and controllable tensile strain field at the growth surface, enabling both a redshift of QD emission to the $\sim$1.3~μm range and site-selective nucleation at the mesa centers. This concept eliminates not only the need for strain-reducing layers (SRLs), which are known to degrade optical coherence, but also provides spatial control and spectral tunability. The grown telecom QDs show pure single-photon emission with $g^{(2)}(τ) = (5.0 \pm 1.0) \times 10^{-2}$ at 4 K and $(2.8 \pm 0.3) \times 10^{-1}$ at 77~K, demonstrating the quantum nature and thermal stability of the emitters. The emission characteristics of complex excitonic states are analyzed using 8-band $k \cdot p$ and configuration-interaction modeling, which quantitatively reproduces the experimental observations. Finally, we present a theory-supported strategy to further redshift the emission toward the center of the O-band and beyond by employing a multi-buried-stressor approach. This combined framework of experiment and theory establishes the buried stressor concept as a scalable route toward highly coherent, position-controlled O-band quantum emitters compatible with industrial photonic integration.

Figures (12)

• Figure 1: Optical characterization of O-band SCQDs using CL. (a) CL spectrum of the mesa's central region (black trace) and from an off-center region (red trace). The inset shows a magnified view of the QD emission in the telecom O-band. The spectral segment highlighted in yellow, spanning from 1260 to 1280 nm, indicates the emission from SCQDs, as seen in the CL intensity map in (b). (b) CL map superimposed on a SEM image of the mesa. (c) Waterfall spectra showing the emission from the center region of five different mesas numbered from 1 to 5 with nominal mesa sizes of 21.07 µ m, 21.00 µ m, 21.20 µ m, 21.34 µ m, and 21.07 µ m, respectively.
• Figure 2: (a) Waterfall plot displaying the µ PL emission from a selected SCQD under 1130 nm pulsed excitation, with pump powers ranging from 14 to 500 µ W at 4 K. Insets show Gaussian fits of the exciton ($X$) and negatively charged exciton ($X^-$) lines. (b) Polarization-resolved $\mu$PL spectra of the investigated SCQD at a pump power of 500 $\mu$W for polarization angles of 0$^\circ$ (black) and 90$^\circ$ (red). The inset shows a zoom of the exciton line, revealing an FSS of (60.0 $\pm$ 0.2) $\mu$eV. The blue curve represents the degree of linear polarization (DLP). (c) The left axis shows CI-calculated binding energies of $X^-$, $X^+$, and $XX$ relative to $X^{0}$ (set to 0 meV) as a function of indium content for a QD with 3 nm height and 34 nm width. The gray box marks the $\sim$70% indium range with the best agreement to experiment. The right axis (cyan) displays the simulated FSS of $X^{0}$ for the same geometry. The measured FSS is indicated by the dashed line and arrow, showing excellent agreement for 3 nm height, 34 nm base width, and 70% indium (gray region).
• Figure 3: (a) Waterfall plot showing µPL spectra of the investigated SCQD under pulsed laser excitation at 1130 nm and a constant pump power of 140 µW, measured at four different temperatures of 4 K, 20 K, 40 K, and 77 K. (b) µPL intensity (solid circles) and linewidth (open circles) for the four identified excitonic transitions — ($X$, black), ($X^+$, green), ($XX$, blue), ($X^-$, red)—as a function of temperature. (c) Second-order photon autocorrelation function for the investigated SCQD line at a pump power of 140 µ W. We observe clean single-photon emission up to 40 K, with a $g^{(2)}(\tau)$ value of $(8.0 \pm 1) \times 10^{-2}$. At 77 K, the $g^{(2)}(\tau)$ value increases to $(2.8 \pm 0.3) \times 10^{-1}$.
• Figure 4: (a) Calculated strain profiles at the growth surface for one (black), two (red), and three AlAs/Al$_2$O$_3$ stressor layers for an AlAs aperture of 500 nm. (b) Left: in-plane surface strain, $\varepsilon_{xx} + \varepsilon_{yy}$; right: corresponding strain-induced energy shift of a nominal QD (height: 3 nm, width: 34 nm, indium content: 70%), shown as a function of the number of AlAs/Al$_2$O$_3$ stressor layers (one, two, and three). (c) Calculated strain profiles for a two-layer AlAs/Al$_2$O$_3$ stressor configuration, illustrating the effect of varying the lower and upper aperture widths. Curves correspond to the following aperture combinations (lower/upper): 500/500 nm (black), 700/500 nm (red), 500/700 nm (green), 1000/500 nm (blue), and 500/1000 nm (magenta), highlighting how aperture geometry tunes the local tensile strain.
• Figure S1: (a) Epitaxial layer design consisting of a template structure that includes a 300 nm buffer layer, 10 pairs of GaAs/Al0.9Ga0.1As/GaAs DBR optimized for emission at 1300 nm, a 30 nm AlAs stressor layer, and an 80 nm GaAs cap. (b) After the UV lithography step, selective wet oxidation forms an unoxidized AlAs aperture at the center of the mesa. (c) The strain from this aperture facilitates the accumulation of indium atoms, leading to site-controlled nucleation of QDs above the aperture. Additionally, with a 90-second growth interruption, this strain redshifts the single QD emission. (d) The simulated structure of the truncated cone-shaped QD is shown, including its dimensions (height and width) and the indium concentration ($x$) within the dot.