Droplet-Etched GaAs Quantum Dots

• Droplet-etched GaAs quantum dots are strain-free semiconductor nanostructures fabricated via droplet etching and nanoscale regrowth, yielding high symmetry and minimal fine-structure splitting.
• They offer high-purity single-photon and entangled photon emission with short radiative lifetimes, making them ideal for quantum communications and integrated photonics.
• Integration into devices such as waveguides, micropillars, and planar circuits—combined with effective surface passivation—ensures stable optical performance and scalability.

Searching arXiv for recent and foundational papers on droplet-etched GaAs quantum dots. Droplet-etched GaAs quantum dots are strain-free semiconductor nanostructures formed by using group-III droplets to etch nanoholes in an AlGaAs surface and subsequently filling those nanoholes with GaAs. In the GaAs/AlGaAs material system, this growth route combines a high degree of structural symmetry, short radiative lifetimes, small excitonic fine-structure splitting, and compatibility with nanophotonic integration. Across the literature, these attributes have been linked to high-purity single-photon emission, high photon indistinguishability, polarization-entangled photon-pair generation, spin-selective optical control, and device integration in waveguides, beamsplitters, micropillars, and shallow photonic structures (Silva et al., 2021).

1. Growth route and nanohole infilling

Droplet-etched GaAs quantum dots are produced by local droplet etching epitaxy in molecular beam epitaxy. A standard implementation begins from GaAs(001) with an AlGaAs barrier, after which a group-III flux is supplied under As-poor or As-closed conditions so that liquid droplets nucleate on the surface. Subsequent exposure to arsenic transforms the droplets into localized etchants that create nanoholes, and later GaAs deposition fills those holes to form strain-free GaAs islands embedded in AlGaAs (Gossink et al., 17 Apr 2026).

Several growth variants are reported. One overview describes deposition of $0.5$ ML of Al at $T \approx 600\,^\circ\mathrm{C}$ with $F_\mathrm{Al} = 0.5$ ML/s and $F_\mathrm{As} \approx 0$, yielding hemispherical droplets of density $N_d \approx 0.2\,\mu\mathrm{m}^{-2}$; arsenic exposure then produces shallow nanoholes of depth $\approx 5$ nm and diameter $\approx 30$ nm, followed by deposition of $1$–$4$ nm GaAs at $T \approx 580\,^\circ\mathrm{C}$ and capping with $T \approx 600\,^\circ\mathrm{C}$0 nm Al$T \approx 600\,^\circ\mathrm{C}$1Ga$T \approx 600\,^\circ\mathrm{C}$2As plus $T \approx 600\,^\circ\mathrm{C}$3 nm GaAs (Silva et al., 2021). Another review frames the process in three phases—droplet deposition, droplet etching, and nanohole regrowth—and emphasizes that density scales with deposition conditions through relations such as $T \approx 600\,^\circ\mathrm{C}$4 in the complete-condensation regime, with further coarsening described by $T \approx 600\,^\circ\mathrm{C}$5 (Gossink et al., 17 Apr 2026).

A Ga-based implementation used for near-surface studies employs a GaAs(001) substrate overgrown with an Al$T \approx 600\,^\circ\mathrm{C}$6Ga$T \approx 600\,^\circ\mathrm{C}$7As barrier, brief submonolayer Ga deposition at $T \approx 600\,^\circ\mathrm{C}$8, and As-triggered nanohole formation, followed by a thin GaAs cap and an Al$T \approx 600\,^\circ\mathrm{C}$9Ga$F_\mathrm{Al} = 0.5$0As sacrificial layer (Manna et al., 2022). In integrated waveguide structures, growth conditions include Ga-droplet formation at $F_\mathrm{Al} = 0.5$1 with $F_\mathrm{Al} = 0.5$2 ML of Ga at $F_\mathrm{Al} = 0.5$3 ML/s As background, nanohole etching for $F_\mathrm{Al} = 0.5$4 s under As overpressure $F_\mathrm{Al} = 0.5$5 Torr), GaAs infill of $F_\mathrm{Al} = 0.5$6 nm at $F_\mathrm{Al} = 0.5$7, and capping with $F_\mathrm{Al} = 0.5$8 nm AlGaAs followed by $F_\mathrm{Al} = 0.5$9 nm Al$F_\mathrm{As} \approx 0$0Ga$F_\mathrm{As} \approx 0$1As (Hornung et al., 2023).

The resulting morphology depends strongly on the specific recipe. Reported nanoholes span shallow geometries of depth $F_\mathrm{As} \approx 0$2 nm and diameter $F_\mathrm{As} \approx 0$3 nm (Silva et al., 2021), nanoholes $F_\mathrm{As} \approx 0$4 nm deep and $F_\mathrm{As} \approx 0$5 nm wide in quasi-resonant spin-control structures (Hopfmann et al., 2020), and much larger inverted-cone holes with openings of $F_\mathrm{As} \approx 0$6 nm along $F_\mathrm{As} \approx 0$7, $F_\mathrm{As} \approx 0$8 nm along $F_\mathrm{As} \approx 0$9, and average depth $N_d \approx 0.2\,\mu\mathrm{m}^{-2}$0 nm in high-resolution morphology studies (Zhang et al., 2024). After infilling, typical droplet-etched GaAs quantum dots are reported with base diameters of $N_d \approx 0.2\,\mu\mathrm{m}^{-2}$1–$N_d \approx 0.2\,\mu\mathrm{m}^{-2}$2 nm and heights of $N_d \approx 0.2\,\mu\mathrm{m}^{-2}$3–$N_d \approx 0.2\,\mu\mathrm{m}^{-2}$4 nm (Manna et al., 2022), or base diameters $N_d \approx 0.2\,\mu\mathrm{m}^{-2}$5–$N_d \approx 0.2\,\mu\mathrm{m}^{-2}$6 nm and heights $N_d \approx 0.2\,\mu\mathrm{m}^{-2}$7–$N_d \approx 0.2\,\mu\mathrm{m}^{-2}$8 nm (Silva et al., 2021). The 2026 review states that complete infilling and the use of As$N_d \approx 0.2\,\mu\mathrm{m}^{-2}$9 during regrowth promote symmetric quantum dots with minimal facet-induced anisotropy and fine-structure splitting $\approx 5$0 (Gossink et al., 17 Apr 2026).

2. Morphology, symmetry, and structural models

A central feature of droplet-etched GaAs quantum dots is their high in-plane symmetry. This has been repeatedly linked to reduced anisotropic electron-hole exchange and correspondingly small fine-structure splitting. One account describes the dots as exhibiting “almost perfect in-plane symmetry” with fine-structure splitting as low as $\approx 5$1 (Schöll et al., 2019), while another reports an ensemble-average $\approx 5$2 for $\approx 5$3 quantum dots (Silva et al., 2021). In the entangled-photon context, representative values of $\approx 5$4, $\approx 5$5, and $\approx 5$6 were measured for three individual quantum dots (Huber et al., 2016).

High-resolution structural work has refined the three-dimensional description of the nanohole and infilled dot. Cross-sectional STEM of uncapped DENI structures reveals an inverted conical nanohole with Al-rich sidewalls and defect-free interfaces, while selective chemical etching and AFM reveal asymmetries in element distribution (Zhang et al., 2024). In that study, the nanohole is modeled as an inverted cone with

$\approx 5$7

with base radii $\approx 5$8 nm and $\approx 5$9 nm, depth $\approx 30$0 nm, and $\approx 30$1, indicating a nearly linear sidewall profile (Zhang et al., 2024). The average sidewall angle is $\approx 30$2, characteristic of $\approx 30$3 planes, and small sidewall undulations of amplitude $\approx 30$4–$\approx 30$5 nm and period $\approx 30$6 nm are associated with facet mixtures and asymmetric Al incorporation (Zhang et al., 2024).

Composition mapping further shows a Ga-rich central region, an Al-depleted core, a $\approx 30$7–$\approx 30$8 nm thick AlAs-rich shell coating the cone walls, and a lower cone region that is “almost pure AlAs”; high-resolution HAADF-STEM and LAADF-STEM show atomically coherent GaAs/Al$\approx 30$9Ga$1$0As (or AlAs) interfaces without misfit dislocations or stacking faults, and the absence of strain contrast indicates strain-free embedding (Zhang et al., 2024). This structural picture is consistent with earlier spin studies that detected only small residual biaxial strain, with $1$1 inferred from NMR spectroscopy (Ulhaq et al., 2015).

A design-oriented treatment uses AFM-reconstructed morphology to model filled nanoholes as truncated-cone quantum dots with base $1$2 nm, top diameter $1$3 nm, and height $1$4 nm, imposing an approximate $1$5 point-group symmetry with only weak $1$6 perturbations from atomistic interfaces (Schimpf et al., 3 Apr 2025). This suggests that morphology-driven symmetry can be treated as an experimentally constrained design variable rather than merely an empirical consequence of growth.

3. Optical spectroscopy and quantum-light performance

Droplet-etched GaAs quantum dots are widely studied as sources of single photons and entangled photon pairs. Their optical response reflects the combination of short radiative lifetimes, narrow linewidths, reduced spectral diffusion in optimized conditions, and small fine-structure splitting (Silva et al., 2021).

For neutral excitons, reported transition energies lie near the $1$7 nm spectral region. One overview gives an emission-wavelength distribution of $1$8 nm $1$9 with $4$0 nm over $4$1 dots (Silva et al., 2021). In a resonant-fluorescence study, the neutral exciton wavelength was $4$2 nm and the measured radiative lifetime was

$4$3

implying $4$4 ps and a lifetime-limited homogeneous linewidth

$4$5

in the ideal limit (Schöll et al., 2019). A broader review gives $4$6 ps and $4$7 ps, together with a Fourier-limited linewidth $4$8 and a measured homogeneous linewidth of $4$9 from Michelson interferometry (Silva et al., 2021).

The single-photon purity reported for these emitters is very high under resonant or quasi-resonant excitation. Under two-photon excitation with a $T \approx 580\,^\circ\mathrm{C}$0 pulse of $T \approx 580\,^\circ\mathrm{C}$1 ps width, one study reports

$T \approx 580\,^\circ\mathrm{C}$2

with no background subtraction in the full-brightness regime (Silva et al., 2021). Under two-photon resonant excitation of the biexciton, measured values are $T \approx 580\,^\circ\mathrm{C}$3 and $T \approx 580\,^\circ\mathrm{C}$4 (Huber et al., 2016). In integrated waveguides under resonant $T \approx 580\,^\circ\mathrm{C}$5-pulse excitation, the raw zero-delay autocorrelation $T \approx 580\,^\circ\mathrm{C}$6 corresponds to a single-photon purity of

$T \approx 580\,^\circ\mathrm{C}$7

(Hornung et al., 2023).

Photon indistinguishability is another defining metric. For pulsed resonance fluorescence from a neutral exciton, the raw Hong–Ou–Mandel visibility reached

$T \approx 580\,^\circ\mathrm{C}$8

without Purcell enhancement (Schöll et al., 2019). The same overview of the field reports $T \approx 580\,^\circ\mathrm{C}$9 for a trion under resonant fluorescence with $T \approx 600\,^\circ\mathrm{C}$00 ns and a remote two-photon interference visibility $T \approx 600\,^\circ\mathrm{C}$01 for remote quantum dots in a p-i-n diode under cw resonance (Silva et al., 2021). On-chip waveguide devices yield fully corrected two-photon interference visibilities up to

$T \approx 600\,^\circ\mathrm{C}$02

for two consecutively emitted photons with $T \approx 600\,^\circ\mathrm{C}$03 ns (Hornung et al., 2023).

Entangled-photon generation proceeds via the biexciton–exciton cascade. Under two-photon resonant excitation, one study reports entanglement fidelity $T \approx 600\,^\circ\mathrm{C}$04, indistinguishability $T \approx 600\,^\circ\mathrm{C}$05, and a Bell-parameter violation with $T \approx 600\,^\circ\mathrm{C}$06 without temporal or spectral post-selection (Huber et al., 2016). A later overview gives a maximum measured fidelity

$T \approx 600\,^\circ\mathrm{C}$07

with concurrence $T \approx 600\,^\circ\mathrm{C}$08 and $T \approx 600\,^\circ\mathrm{C}$09 by more than $T \approx 600\,^\circ\mathrm{C}$10 when $T \approx 600\,^\circ\mathrm{C}$11 is tuned near zero and no post-selection is used (Silva et al., 2021).

These results are commonly attributed to the coexistence of low fine-structure splitting and short exciton lifetimes. In one explicit formulation, droplet-etched GaAs quantum dots “intrinsically combine short lifetime with low noise, thus achieving near-unity indistinguishability without any cavity” (Schöll et al., 2019).

4. Near-surface operation, spectral diffusion, and passivation

Integration into shallow nanophotonic structures introduces a specific challenge: the optical properties of a droplet-etched GaAs quantum dot degrade when the emitter is brought close to a free surface. Near-surface operation is nevertheless required for coupling into planar Yagi–Uda antennas, nanopillars, Bragg gratings, and other structures whose mode overlap demands surface-to-dot distances below $T \approx 600\,^\circ\mathrm{C}$12 nm and, in practice for $T \approx 600\,^\circ\mathrm{C}$13 nm GaAs dots, often $T \approx 600\,^\circ\mathrm{C}$14 nm (Manna et al., 2022).

The underlying problem is the high density of GaAs surface states. One study gives

$T \approx 600\,^\circ\mathrm{C}$15

for midgap states, which pin the Fermi level and produce downward band bending. Under illumination, trap and detrap processes cause fluctuating electric fields and spectral diffusion, with the time-averaged line shape modeled through Kubo–Anderson theory as

$T \approx 600\,^\circ\mathrm{C}$16

(Manna et al., 2022).

Experimentally, low-temperature $T \approx 600\,^\circ\mathrm{C}$17-PL at $T \approx 600\,^\circ\mathrm{C}$18 K shows resolution-limited Gaussian FWHM $T \approx 600\,^\circ\mathrm{C}$19 for as-grown dots at $T \approx 600\,^\circ\mathrm{C}$20 nm. After etching to bring dots close to the surface, linewidths broaden to $T \approx 600\,^\circ\mathrm{C}$21 for one structure at $T \approx 600\,^\circ\mathrm{C}$22 nm and $T \approx 600\,^\circ\mathrm{C}$23 for another when the distance is reduced from $T \approx 600\,^\circ\mathrm{C}$24 to $T \approx 600\,^\circ\mathrm{C}$25 nm (Manna et al., 2022). Sulphur passivation followed by Al$T \approx 600\,^\circ\mathrm{C}$26O$T \approx 600\,^\circ\mathrm{C}$27 encapsulation partially recovers the linewidth to $T \approx 600\,^\circ\mathrm{C}$28–$T \approx 600\,^\circ\mathrm{C}$29 and increases photoluminescence intensity (Manna et al., 2022).

The chemical protocol consists of an HCl:H$T \approx 600\,^\circ\mathrm{C}$30O $T \approx 600\,^\circ\mathrm{C}$31 oxide strip for $T \approx 600\,^\circ\mathrm{C}$32 min after the final wet etch, followed by immersion in $T \approx 600\,^\circ\mathrm{C}$33 $T \approx 600\,^\circ\mathrm{C}$34 solution for $T \approx 600\,^\circ\mathrm{C}$35 min at room temperature, rinsing in DI water, and N$T \approx 600\,^\circ\mathrm{C}$36 drying; within $T \approx 600\,^\circ\mathrm{C}$37 min of drying, the sample is transferred to an ALD reactor at $T \approx 600\,^\circ\mathrm{C}$38, where Al$T \approx 600\,^\circ\mathrm{C}$39O$T \approx 600\,^\circ\mathrm{C}$40 is deposited by alternating TMA and H$T \approx 600\,^\circ\mathrm{C}$41O pulses at $T \approx 600\,^\circ\mathrm{C}$42 per cycle, giving thicknesses $T \approx 600\,^\circ\mathrm{C}$43 nm or $T \approx 600\,^\circ\mathrm{C}$44 nm (Manna et al., 2022). The passivation mechanism is described as S$T \approx 600\,^\circ\mathrm{C}$45 binding to Ga and As dangling bonds, reducing $T \approx 600\,^\circ\mathrm{C}$46 by more than one order of magnitude, while the ALD overlayer blocks O$T \approx 600\,^\circ\mathrm{C}$47/H$T \approx 600\,^\circ\mathrm{C}$48O ingress and prevents re-oxidation of Ga–S bonds (Manna et al., 2022).

Quantitatively, Gaussian fits imply modulation amplitudes reducing from $T \approx 600\,^\circ\mathrm{C}$49–$T \approx 600\,^\circ\mathrm{C}$50 to $T \approx 600\,^\circ\mathrm{C}$51, with inferred correlation times $T \approx 600\,^\circ\mathrm{C}$52 ps (Manna et al., 2022). The same work reports that only-etched samples broaden by an extra $T \approx 600\,^\circ\mathrm{C}$53–$T \approx 600\,^\circ\mathrm{C}$54 over two months of ambient storage, whereas passivated dots degrade by less than $T \approx 600\,^\circ\mathrm{C}$55, with $T \approx 600\,^\circ\mathrm{C}$56 nm Al$T \approx 600\,^\circ\mathrm{C}$57O$T \approx 600\,^\circ\mathrm{C}$58 giving the slowest aging (Manna et al., 2022).

A rate-equation description formalizes the role of traps. With excited and trap populations obeying

$T \approx 600\,^\circ\mathrm{C}$59

and

$T \approx 600\,^\circ\mathrm{C}$60

the optical coherence decays as

$T \approx 600\,^\circ\mathrm{C}$61

with pure dephasing $T \approx 600\,^\circ\mathrm{C}$62 (Manna et al., 2022). Post-passivation values of $T \approx 600\,^\circ\mathrm{C}$63 and $T \approx 600\,^\circ\mathrm{C}$64 are reported as consistent with the observed linewidth reduction from $T \approx 600\,^\circ\mathrm{C}$65 to $T \approx 600\,^\circ\mathrm{C}$66 (Manna et al., 2022).

5. Spin, magnetic response, and coherence environment

Droplet-etched GaAs quantum dots are also studied as spin-photon interfaces. Their spin physics reflects the combination of quasi-strain-free confinement, high symmetry, and predictable electronic structure.

A foundational spin study reports nearly vanishing electron $T \approx 600\,^\circ\mathrm{C}$67-factor, with $T \approx 600\,^\circ\mathrm{C}$68 in some nanohole-filled GaAs/AlGaAs dots, together with optical manipulation of the nuclear spin environment up to $T \approx 600\,^\circ\mathrm{C}$69 polarization and nuclear spin lifetimes exceeding $T \approx 600\,^\circ\mathrm{C}$70 s (Ulhaq et al., 2015). Magneto-PL yields $T \approx 600\,^\circ\mathrm{C}$71 and $T \approx 600\,^\circ\mathrm{C}$72 for type B dots, and $T \approx 600\,^\circ\mathrm{C}$73 with $T \approx 600\,^\circ\mathrm{C}$74 for type A dots (Ulhaq et al., 2015). NMR spectroscopy of $T \approx 600\,^\circ\mathrm{C}$75As satellite transitions gives $T \approx 600\,^\circ\mathrm{C}$76 kHz for type A dots, corresponding to $T \approx 600\,^\circ\mathrm{C}$77, and $T \approx 600\,^\circ\mathrm{C}$78 kHz for type B dots, corresponding to $T \approx 600\,^\circ\mathrm{C}$79 (Ulhaq et al., 2015). This residual strain is small, yet sufficient to suppress nuclear spin diffusion and stabilize the nuclear bath (Ulhaq et al., 2015).

Quasi-resonant excitation has enabled deterministic spin preparation using excited-state resonances. In one comprehensive study, the single-particle spectrum forms shells with s–p shell onset at $T \approx 600\,^\circ\mathrm{C}$80 meV above the s-shell and the p-shell manifold around $T \approx 600\,^\circ\mathrm{C}$81 meV (Hopfmann et al., 2020). For the neutral exciton, quasi-resonant PLE identifies resonances at $T \approx 600\,^\circ\mathrm{C}$82 meV, $T \approx 600\,^\circ\mathrm{C}$83 meV, $T \approx 600\,^\circ\mathrm{C}$84 meV, and $T \approx 600\,^\circ\mathrm{C}$85 meV above $T \approx 600\,^\circ\mathrm{C}$86 (Hopfmann et al., 2020). Spin-preparation fidelity is defined through excitation-induced polarization degrees over the three orthogonal polarization bases, leading to a global norm

$T \approx 600\,^\circ\mathrm{C}$87

which is independent of the relative orientation of lab and quantum-dot polarization eigenbases (Hopfmann et al., 2020). Experimentally, $T \approx 600\,^\circ\mathrm{C}$88 reaches up to $T \approx 600\,^\circ\mathrm{C}$89 for $T \approx 600\,^\circ\mathrm{C}$90 and $T \approx 600\,^\circ\mathrm{C}$91 resonances of the s–p shell and approximately $T \approx 600\,^\circ\mathrm{C}$92 for $T \approx 600\,^\circ\mathrm{C}$93 (Hopfmann et al., 2020).

The same study finds non-radiative relaxation times as low as

$T \approx 600\,^\circ\mathrm{C}$94

for $T \approx 600\,^\circ\mathrm{C}$95 excited via an s–p shell$T \approx 600\,^\circ\mathrm{C}$96 resonance, with similar values of $T \approx 600\,^\circ\mathrm{C}$97 ps for $T \approx 600\,^\circ\mathrm{C}$98 and $T \approx 600\,^\circ\mathrm{C}$99 ps for $F_\mathrm{Al} = 0.5$00 under other conditions (Hopfmann et al., 2020). Time-resolved correlation spectroscopy reveals that the excitation scheme significantly impacts the electronic environment: under above-band pumping, $F_\mathrm{Al} = 0.5$01 ns and $F_\mathrm{Al} = 0.5$02, while under quasi-resonant pumping $F_\mathrm{Al} = 0.5$03 ns and $F_\mathrm{Al} = 0.5$04 (Hopfmann et al., 2020). This suggests that excitation conditions tune not only preparation fidelity but also charge-noise dynamics and coherence.

A more recent design study treats optical and magnetic response as predictable from high-symmetry morphology. Using an $F_\mathrm{Al} = 0.5$05 envelope-function Hamiltonian in the Burt–Foreman formalism, constrained by AFM morphology, simulations and measurements track the in-plane electron $F_\mathrm{Al} = 0.5$06-factor through a zero-crossing from $F_\mathrm{Al} = 0.5$07 to $F_\mathrm{Al} = 0.5$08 over the $F_\mathrm{Al} = 0.5$09–$F_\mathrm{Al} = 0.5$10 nm range (Schimpf et al., 3 Apr 2025). In that work, $F_\mathrm{Al} = 0.5$11 crosses zero at $F_\mathrm{Al} = 0.5$12 nm, equivalently $F_\mathrm{Al} = 0.5$13 nm for barrier fraction $F_\mathrm{Al} = 0.5$14, while $F_\mathrm{Al} = 0.5$15 with anisotropy $F_\mathrm{Al} = 0.5$16 (Schimpf et al., 3 Apr 2025). The charged-exciton transition dipole moments in Voigt geometry are predicted to rotate rigidly with the in-plane magnetic field,

$F_\mathrm{Al} = 0.5$17

and experiments show measured alignment deviation $F_\mathrm{Al} = 0.5$18 for dots with fine-structure splitting $F_\mathrm{Al} = 0.5$19 (Schimpf et al., 3 Apr 2025).

6. Device integration and quantum-photonic architectures

Because droplet-etched GaAs quantum dots are strain-free and compatible with GaAs/AlGaAs processing, they have been incorporated into several photonic architectures. A recurring motivation is the combination of intrinsically short lifetimes and high indistinguishability without mandatory high-$F_\mathrm{Al} = 0.5$20 microcavities (Schöll et al., 2019).

Monolithic photonic integrated circuits have been demonstrated using single-mode waveguides and on-chip beamsplitters (Hornung et al., 2023). In that platform, the layer stack includes Al$F_\mathrm{Al} = 0.5$21Ga$F_\mathrm{Al} = 0.5$22As claddings, a $F_\mathrm{Al} = 0.5$23 nm GaAs core, and Al$F_\mathrm{Al} = 0.5$24Ga$F_\mathrm{Al} = 0.5$25As such that the waveguide cross section is $F_\mathrm{Al} = 0.5$26 nm $F_\mathrm{Al} = 0.5$27 $F_\mathrm{Al} = 0.5$28 nm, supporting a single TE mode at $F_\mathrm{Al} = 0.5$29 nm; the degree of polarization is $F_\mathrm{Al} = 0.5$30, and propagation loss is $F_\mathrm{Al} = 0.5$31 dB/mm over $F_\mathrm{Al} = 0.5$32–$F_\mathrm{Al} = 0.5$33 nm (Hornung et al., 2023). A $F_\mathrm{Al} = 0.5$34 MMI beamsplitter with $F_\mathrm{Al} = 0.5$35m and $F_\mathrm{Al} = 0.5$36m yields a measured splitting ratio of $F_\mathrm{Al} = 0.5$37 over $F_\mathrm{Al} = 0.5$38–$F_\mathrm{Al} = 0.5$39 nm and transmission $F_\mathrm{Al} = 0.5$40 (Hornung et al., 2023). The radiative coupling factor is estimated as $F_\mathrm{Al} = 0.5$41 from decay-time reduction relative to bulk (Hornung et al., 2023).

Micropillar implementations provide a complementary route. In a deterministic fabrication study, droplet-etched GaAs quantum dots embedded in a low-$F_\mathrm{Al} = 0.5$42 cavity were localized relative to alignment markers with better than $F_\mathrm{Al} = 0.5$43 nm RMS and then integrated into circular pillars of diameter $F_\mathrm{Al} = 0.5$44–$F_\mathrm{Al} = 0.5$45m (Madigawa et al., 13 Feb 2025). All $F_\mathrm{Al} = 0.5$46 pre-selected quantum dots were found after fabrication at pillar centers, giving a spatial placement yield of $F_\mathrm{Al} = 0.5$47 (Madigawa et al., 13 Feb 2025). Under pulsed p-shell excitation, the planar dots exhibit monoexponential decay $F_\mathrm{Al} = 0.5$48 ns, whereas the pillars show biexponential decay with fast component $F_\mathrm{Al} = 0.5$49 ns and slow component $F_\mathrm{Al} = 0.5$50 ns (Madigawa et al., 13 Feb 2025). Count-rate fluctuations of order $F_\mathrm{Al} = 0.5$51–$F_\mathrm{Al} = 0.5$52 around the mean are attributed to random charging of traps, but low-power above-band LED excitation reduces the standard deviation of the fluctuations by nearly $F_\mathrm{Al} = 0.5$53 and increases the peak detected count rate from $F_\mathrm{Al} = 0.5$54 counts/s to $F_\mathrm{Al} = 0.5$55 counts/s, corresponding to efficiency increasing from $F_\mathrm{Al} = 0.5$56 to $F_\mathrm{Al} = 0.5$57 (Madigawa et al., 13 Feb 2025). At saturation, $F_\mathrm{Al} = 0.5$58 confirms near-ideal antibunching (Madigawa et al., 13 Feb 2025).

Other nanophotonic implementations summarized in the literature include planar DBR cavities with zirconia solid immersion lenses, metal-semiconductor-metal Yagi–Uda antennas, GaP hemispherical lenses, and circular Bragg resonator plus broadband reflector structures (Silva et al., 2021). Reported extraction efficiencies are $F_\mathrm{Al} = 0.5$59 for a planar DBR cavity with zirconia lens, $F_\mathrm{Al} = 0.5$60 for a planar Yagi–Uda antenna, $F_\mathrm{Al} = 0.5$61 for a frustrated-TIR GaP hemispherical lens with PMMA gap, and for a CBR-HBR structure $F_\mathrm{Al} = 0.5$62, $F_\mathrm{Al} = 0.5$63, with Purcell factor $F_\mathrm{Al} = 0.5$64 and $F_\mathrm{Al} = 0.5$65 reduced from $F_\mathrm{Al} = 0.5$66 to $F_\mathrm{Al} = 0.5$67 ps (Silva et al., 2021).

The quantum-communication relevance of these emitters has been explicitly demonstrated in teleportation, entanglement swapping, and QKD experiments summarized in the field overview. Reported values include a teleportation fidelity $F_\mathrm{Al} = 0.5$68, entanglement-swapping fidelity $F_\mathrm{Al} = 0.5$69, and BBM92 QKD over $F_\mathrm{Al} = 0.5$70 m fiber with raw key rate $F_\mathrm{Al} = 0.5$71 bit/s and QBER $F_\mathrm{Al} = 0.5$72 (Silva et al., 2021). These figures are system-level results rather than intrinsic material metrics, but they show that droplet-etched GaAs quantum dots have progressed from growth studies to deployed quantum-light nodes.

7. Interpretive themes, misconceptions, and research directions

A common misconception is that droplet-etched GaAs quantum dots are defined primarily by a single morphology. The literature instead shows multiple geometrical regimes, from shallow $F_\mathrm{Al} = 0.5$73 nm nanoholes and compact $F_\mathrm{Al} = 0.5$74–$F_\mathrm{Al} = 0.5$75 nm-high dots (Silva et al., 2021) to deeper DENI inverted cones of $F_\mathrm{Al} = 0.5$76 nm depth with elliptical asymmetry and Al-rich shells (Zhang et al., 2024). This suggests that “droplet-etched GaAs quantum dot” denotes a growth family with shared physical principles rather than a unique geometric archetype.

Another misconception is that the optical quality of these emitters is intrinsically immune to surfaces. Near-surface studies show the opposite: when the dot-to-surface distance is reduced to $F_\mathrm{Al} = 0.5$77 nm, linewidth broadening and intensity degradation occur unless the free surface is chemically passivated and encapsulated (Manna et al., 2022). A plausible implication is that the intrinsic quality of the quantum dot and the extrinsic quality of the processed photonic environment must be treated jointly in device engineering.

The relation between morphology and quantum-optical performance remains a central research axis. Structural modeling connects nanohole ellipticity, Al-rich sidewall shells, and infill amount to binding energies and fine-structure splitting (Zhang et al., 2024). Growth reviews similarly relate regrowth mode, arsenic species, ripening, and facet formation to final dot symmetry and emission energy, including the scaling $F_\mathrm{Al} = 0.5$78 with hole depth (Gossink et al., 17 Apr 2026). Design-oriented spin studies extend this logic by predicting $F_\mathrm{Al} = 0.5$79-tensors and optical dipole orientations from AFM-derived morphology (Schimpf et al., 3 Apr 2025). Together, these results support the view that droplet etching is evolving from a recipe-driven technique into a quantitatively modeled platform.

The platform has also expanded beyond single-dot spectroscopy toward scalable integration. The first monolithic integration into single-mode waveguides and beamsplitters (Hornung et al., 2023), deterministic placement into micropillars (Madigawa et al., 13 Feb 2025), and systematic treatment of shallow-surface passivation (Manna et al., 2022) indicate that fabrication constraints now play a role comparable to emitter physics. At the same time, reviews continue to emphasize best-practice growth windows near $F_\mathrm{Al} = 0.5$80, moderate arsenic overpressure, and sufficient GaAs regrowth to obtain symmetric, strain-free dots with small fine-structure splitting (Gossink et al., 17 Apr 2026).

In that broader context, droplet-etched GaAs quantum dots occupy a distinct niche among semiconductor quantum emitters: a quasi-strain-free, symmetry-favored, morphologically engineerable GaAs/AlGaAs platform whose experimentally demonstrated performance spans high-purity single photons, near-unity indistinguishability, high-fidelity entanglement, deterministic spin preparation, and direct photonic integration (Schöll et al., 2019).