InAs/InGaAs/InP Quantum Dots

• InAs/InGaAs/InP quantum dots are III–V semiconductor nanostructures grown on InP using methods like Stranski–Krastanov and droplet epitaxy, offering precise control over morphology and emission wavelengths.
• Advanced epitaxial techniques and barrier engineering enable tuning of optical operation across the telecom spectrum, including C-band emission and multimodal ensemble outputs.
• Electrical tuning, thermal activation control, and integrated photonic designs enhance excitonic properties and device performance for telecom applications.

to=arxiv_search.search 天天中彩票腾讯json content='{"query":"InAs InGaAs InP quantum dots telecom Stark tuning droplet epitaxy arXiv", "max_results": 10, "sort_by": "relevance", "sort_order": "descending"}' to=arxiv_search.search 大发快三走势图 nammineq content='{"query":"(Lebedev et al., 2017, Aghaeimeibodi et al., 2018, Holewa et al., 2019, Berdnikov et al., 2023, Holewa et al., 2020, Ha et al., 2016)", "max_results": 10, "sort_by": "relevance", "sort_order": "descending"}' InAs/InGaAs/InP quantum dots are III–V semiconductor nanostructures implemented on the InP materials platform, most commonly as InAs islands embedded either directly in InP or in lattice-matched ternary and quaternary barriers such as In(Al,Ga)As, InGaAlAs, and related alloys. In the literature grouped around this platform, the active nanostructures are realized by Stranski–Krastanov growth, droplet epitaxy, ripening-assisted molecular beam epitaxy, selective-area epitaxy, and electrically contacted capacitor geometries. Reported optical operation spans the telecom-relevant range from about $1.2$ to $2.0~\mu\mathrm{m}$, with isolated single-dot lines in the C band, multimodal ensemble emission over S/C/L/U bands, and strong sensitivity to wetting-layer states, barrier composition, and thermally activated carrier transfer (Holewa et al., 2019, Berdnikov et al., 2023, Ha et al., 2016, Holewa et al., 2020, Aghaeimeibodi et al., 2018).

1. Epitaxial realizations and growth control

On InP, several closely related growth routes define the platform. In low-density MOVPE-grown InAs/InP quantum dot-like structures, a $0.5~\mu\mathrm{m}$ InP buffer is grown on $(001)$ InP at $600^\circ\mathrm{C}$, the temperature is lowered to $480^\circ\mathrm{C}$, and $1.04$ ML of InAs is deposited with a very high V/III ratio of approximately $366$, followed by a $6$ s interruption, a $10$ nm InP cap, and then an additional $2.0~\mu\mathrm{m}$0 nm of InP. In this case the nanostructures nucleate on a $2.0~\mu\mathrm{m}$1-ML wetting layer that is ultimately identified as pure InAs (Holewa et al., 2019). A distinct MOVPE route uses near-critical Stranski–Krastanov growth on InP(001): As/P exchange first creates an $2.0~\mu\mathrm{m}$2 wetting layer, and then a near-threshold InAs coverage is added. In that regime, the density can be tuned between $2.0~\mu\mathrm{m}$3 and $2.0~\mu\mathrm{m}$4 by V/III ratio, InAs coverage, and post-deposition interruption under constant $2.0~\mu\mathrm{m}$5, while the dot size evolves largely independently of density (Berdnikov et al., 2023).

Molecular beam epitaxy on InP is used in two further important variants. In ripening-assisted growth on InP(001), a $2.0~\mu\mathrm{m}$6 nm InP buffer is followed by a $2.0~\mu\mathrm{m}$7 nm $2.0~\mu\mathrm{m}$8 barrier, $2.0~\mu\mathrm{m}$9 ML of InAs deposited at $0.5~\mu\mathrm{m}$0, and then a cooldown from $0.5~\mu\mathrm{m}$1 to $0.5~\mu\mathrm{m}$2 at $0.5~\mu\mathrm{m}$3 K/min under an $0.5~\mu\mathrm{m}$4 overpressure of $0.5~\mu\mathrm{m}$5 Torr; Ostwald-type ripening then converts the initial seed ensemble into a low-density multimodal dot distribution (Holewa et al., 2020). In droplet epitaxy on Fe-doped semi-insulating InP(111)A, $0.5~\mu\mathrm{m}$6 ML of In is deposited at $0.5~\mu\mathrm{m}$7, crystallized into InAs under $0.5~\mu\mathrm{m}$8 at $0.5~\mu\mathrm{m}$9, and capped by $(001)$0 nm of lattice-matched $(001)$1. This route uses the $(001)$2 symmetry of the (111)A surface to suppress in-plane elongation (Ha et al., 2016).

Selective-area droplet epitaxy adds lithographic ordering. A PS-b-PDMS block-copolymer mask is converted into a silicon oxycarbide hard mask with $(001)$3–$(001)$4 nm openings, after which InP(001) is deoxidized under $(001)$5, etched in situ with $(001)$6, exposed to $(001)$7, and loaded with indium droplets using TMIn at $(001)$8 for $(001)$9 s before crystallization into InAs and capping with InP. This creates ordered InAs/InP quantum dots around $600^\circ\mathrm{C}$0, but also introduces barrier defects that later dominate thermal capture and quenching (Shikin et al., 2018).

2. Morphology, symmetry, and confinement

A recurring structural feature of InAs/InP-family dots is extreme flatness. In low-density MOVPE-grown InAs/InP structures, buried dots are flat-topped truncated pyramids protruding by small integer numbers of monolayers above a $600^\circ\mathrm{C}$1-ML wetting layer; AFM on surface dots gives widths $600^\circ\mathrm{C}$2–$600^\circ\mathrm{C}$3 nm, lengths $600^\circ\mathrm{C}$4–$600^\circ\mathrm{C}$5 nm, aspect ratio $600^\circ\mathrm{C}$6, and sidewall inclination near $600^\circ\mathrm{C}$7. A representative buried dot protrudes by $600^\circ\mathrm{C}$8 ML, and fitting of exciton energies to the observed photoluminescence bands yields a linear in-plane size relation $600^\circ\mathrm{C}$9, where $480^\circ\mathrm{C}$0 is the height in monolayers above the wetting layer (Holewa et al., 2019).

In the ripened $480^\circ\mathrm{C}$1 system, the dots are modeled as truncated pyramids on a $480^\circ\mathrm{C}$2-ML wetting layer with a $480^\circ\mathrm{C}$3 side-facet angle, heights from $480^\circ\mathrm{C}$4 to $480^\circ\mathrm{C}$5 ML above the wetting layer, and lateral size increasing from roughly $480^\circ\mathrm{C}$6 to $480^\circ\mathrm{C}$7 nm across the families. A defining result is that adjacent families differ by exactly $480^\circ\mathrm{C}$8 ML in height rather than $480^\circ\mathrm{C}$9 ML; the paper attributes this to long-range ordering in the quaternary barrier that stabilizes even-monolayer heights (Holewa et al., 2020). In droplet-epitaxy InAs/In(Al,Ga)As/InP(111)A, uncapped dots are flat disk-like truncated pyramids with heights $1.04$0–$1.04$1 nm, diameters $1.04$2–$1.04$3 nm, and densities $1.04$4 to $1.04$5, and cross sections along $1.04$6 and $1.04$7 are identical, confirming the absence of elongation (Ha et al., 2016).

The confinement landscape differs markedly from the better-known InAs/GaAs case. Atomistic pseudopotential calculations for lens-shaped InAs/InP dots with $1.04$8 nm and $1.04$9 nm give electron and hole confinement depths of about $366$0 meV and $366$1 meV, respectively, whereas comparable InAs/GaAs dots give about $366$2 meV and $366$3 meV. The same calculations find $366$4-orbital splittings of $366$5 meV and $366$6 meV in InAs/InP, compared with $366$7 meV and $366$8 meV in InAs/GaAs (Gong et al., 2010). Complementary atomistic tight-binding work shows that, for self-assembled InAs/InP, the overall confinement is set by the interplay of strain and valence-band offset rather than by valence-band offset alone. In that treatment, recommended InAs/InP valence-band offsets lie around $366$9–$6$0 meV, and strain is responsible for the shell-like organization of hole states in lens-type InAs/InP dots and the single-band-like hole structure in disc-type geometries (Zieliński, 2013).

3. Spectral range and telecom-band engineering

The main technological appeal of the InP platform is spectral access to telecom wavelengths without metamorphic buffers. In low-density InAs/InP quantum dot-like structures, ensemble emission extends across approximately $6$1–$6$2 eV, corresponding to roughly $6$3–$6$4, and the resolved D, E, and F bands coincide with the S, C, and L bands of the third optical window. Neutral exciton, biexciton, and trion lines were observed from spatially isolated emitters in those bands (Holewa et al., 2019). Near-critical Stranski–Krastanov growth on InP(001) produces well-isolated single-dot lines from $6$5 to $6$6 nm at $6$7 K, directly overlapping the $6$8–$6$9 nm telecom C band (Berdnikov et al., 2023).

Barrier engineering broadens the accessible spectral range. In droplet-epitaxy InAs dots on InP(111)A, changing the lattice-matched barrier from $10$0 to $10$1 and then to $10$2 systematically red-shifts the ensemble from $10$3–$10$4 through robust C-band coverage and beyond $10$5. The stated mechanism is reduction of the conduction- and valence-band offsets, which lowers the ground-state exciton energy $10$6 and thus increases $10$7 (Ha et al., 2016). In the ripened $10$8 system, multimodal families B–G span $10$9–$2.0~\mu\mathrm{m}$00, with adjacent family spacings decreasing from about $2.0~\mu\mathrm{m}$01 meV to $2.0~\mu\mathrm{m}$02 meV and family linewidths of $2.0~\mu\mathrm{m}$03–$2.0~\mu\mathrm{m}$04 meV (Holewa et al., 2020).

A separate high-density InP-substrate literature on InAs dots and dashes embedded in different lattice-matched barriers shows the same tuning logic. In nearly rotationally symmetric dots with $2.0~\mu\mathrm{m}$05 barriers, $2.0~\mu\mathrm{m}$06 ML structures emit at $2.0~\mu\mathrm{m}$07 eV $2.0~\mu\mathrm{m}$08 and $2.0~\mu\mathrm{m}$09 ML structures at about $2.0~\mu\mathrm{m}$10 eV $2.0~\mu\mathrm{m}$11. In elongated dashes with $2.0~\mu\mathrm{m}$12 barriers, $2.0~\mu\mathrm{m}$13, $2.0~\mu\mathrm{m}$14, and $2.0~\mu\mathrm{m}$15 ML structures emit near $2.0~\mu\mathrm{m}$16, $2.0~\mu\mathrm{m}$17, and $2.0~\mu\mathrm{m}$18, and single-QDash emission was demonstrated near $2.0~\mu\mathrm{m}$19 (Jahan et al., 2012).

Structure class	Matrix or barrier	Reported emission
Low-density InAs/InP QD-like structures	InP	$2.0~\mu\mathrm{m}$20–$2.0~\mu\mathrm{m}$21
Near-critical SK InAs/InP	InP	$2.0~\mu\mathrm{m}$22–$2.0~\mu\mathrm{m}$23 nm
Droplet epitaxy on InP(111)A	$2.0~\mu\mathrm{m}$24	$2.0~\mu\mathrm{m}$25 to beyond $2.0~\mu\mathrm{m}$26
Ripened InAs/InAlGaAs/InP	$2.0~\mu\mathrm{m}$27	$2.0~\mu\mathrm{m}$28–$2.0~\mu\mathrm{m}$29
InAs dots and dashes on InP	$2.0~\mu\mathrm{m}$30, $2.0~\mu\mathrm{m}$31	$2.0~\mu\mathrm{m}$32–$2.0~\mu\mathrm{m}$33

4. Excitons, charged states, and radiative properties

Excitonic spectroscopy on the InP platform shows a combination of long telecom wavelengths, nontrivial Coulomb physics, and comparatively long lifetimes. In the low-density InAs/InP system, $2.0~\mu\mathrm{m}$34PL from isolated mesas resolves trions across the L, C, and S bands, as well as neutral exciton and biexciton lines. For an L-band emitter, the biexciton binding energy is $2.0~\mu\mathrm{m}$35 meV and the exciton fine-structure splitting is $2.0~\mu\mathrm{m}$36; representative power laws were $2.0~\mu\mathrm{m}$37 and $2.0~\mu\mathrm{m}$38 (Holewa et al., 2019). The same work found exciton lifetimes on the order of $2.0~\mu\mathrm{m}$39 ns and nearly independent of emission energy across the S/C/L bands. The interpretation was that strongly height-dependent Coulomb correlations flatten the lifetime dispersion that had been expected from weaker electron confinement in InAs/InP.

Atomistic many-body calculations establish that the hierarchy of charged-exciton transitions in InAs/InP differs qualitatively from InAs/GaAs. For the lowest shell, the Hartree–Fock emission energies are

$2.0~\mu\mathrm{m}$40

$2.0~\mu\mathrm{m}$41

In InAs/InP, the ordering $2.0~\mu\mathrm{m}$42 makes $2.0~\mu\mathrm{m}$43 typically blue-shifted and $2.0~\mu\mathrm{m}$44 red-shifted relative to $2.0~\mu\mathrm{m}$45, while $2.0~\mu\mathrm{m}$46 is typically blue-shifted or close to zero binding for taller dots. The hidden correlation,

$2.0~\mu\mathrm{m}$47

lies in the range $2.0~\mu\mathrm{m}$48–$2.0~\mu\mathrm{m}$49 meV, smaller than in InAs/GaAs. The same calculations give radiative lifetimes of about $2.0~\mu\mathrm{m}$50–$2.0~\mu\mathrm{m}$51 ns for neutral excitons in flat InAs/InP dots and the approximate relation $2.0~\mu\mathrm{m}$52 (Gong et al., 2010).

Electrical tuning has advanced this excitonic picture from spectroscopy to device-level control. In a vertical capacitor geometry using a microtransferred InP waveguide with InAs dots, a $2.0~\mu\mathrm{m}$53 nm $2.0~\mu\mathrm{m}$54 insulator suppresses carrier injection from the contacts and enables continuous quantum-confined Stark tuning by $2.0~\mu\mathrm{m}$55 meV, equivalent to an $2.0~\mu\mathrm{m}$56 nm red shift for a representative line. The Stark shift is fitted by

$2.0~\mu\mathrm{m}$57

with $2.0~\mu\mathrm{m}$58 nm and $2.0~\mu\mathrm{m}$59. The single-photon character is preserved across much of the tuning range, with $2.0~\mu\mathrm{m}$60 at $2.0~\mu\mathrm{m}$61 V and $2.0~\mu\mathrm{m}$62 at $2.0~\mu\mathrm{m}$63 V. Time-resolved photoluminescence shows $2.0~\mu\mathrm{m}$64 ns at $2.0~\mu\mathrm{m}$65 V, increasing to $2.0~\mu\mathrm{m}$66 ns at $2.0~\mu\mathrm{m}$67 V and then decreasing to $2.0~\mu\mathrm{m}$68 ns at higher field as tunneling sets in (Aghaeimeibodi et al., 2018).

5. Wetting layers, thermal activation, and carrier redistribution

The most persistent physical complication of the platform is the coupling between quantum dots and nearby reservoirs: wetting layers, barrier localized states, and defect states. In low-density InAs/InP structures, the wetting layer is central to the temperature dependence. High-energy bands A and B quench with increasing temperature, whereas lower-energy families C–F initially gain intensity before quenching, showing thermally assisted transfer from shallower to deeper dots through the wetting layer. Using the wetting-layer absorption edge $2.0~\mu\mathrm{m}$69 eV and the ratio $2.0~\mu\mathrm{m}$70, the fitted activation energies were assigned to electron escape for bands A–B, exciton escape for bands C–E, and mixed behavior for the lowest-energy families. For a single C-band trion line, two activation energies were resolved, $2.0~\mu\mathrm{m}$71 meV and $2.0~\mu\mathrm{m}$72 meV, the latter matching transfer to bright excited trion orbitals (Holewa et al., 2019).

In ripened $2.0~\mu\mathrm{m}$73, the wetting layer again acts as a reservoir, but the dominant quenching mechanism shifts. The WL-related A band has activation energies $2.0~\mu\mathrm{m}$74 meV and $2.0~\mu\mathrm{m}$75 meV. The QD families show low-temperature activation around $2.0~\mu\mathrm{m}$76 meV for B–D and quenching energies between $2.0~\mu\mathrm{m}$77 and $2.0~\mu\mathrm{m}$78 meV for B–F, with the G family showing $2.0~\mu\mathrm{m}$79 meV plus an additional $2.0~\mu\mathrm{m}$80 meV channel. Comparison with multiband $2.0~\mu\mathrm{m}$81 and configuration-interaction level diagrams led to the conclusion that hole escape to the InAlGaAs barrier is the primary photoluminescence quenching mechanism, particularly for larger dots (Holewa et al., 2020).

A different reservoir appears in selective-area droplet epitaxy. There, strong confinement is expected from strained InAs/InP offsets of $2.0~\mu\mathrm{m}$82 meV and $2.0~\mu\mathrm{m}$83 meV, yet the optical response is dominated by defect states inside the InP bandgap introduced during lithography and etching. Arrhenius fits of the form

$2.0~\mu\mathrm{m}$84

yield shallow activation energies of approximately $2.0~\mu\mathrm{m}$85–$2.0~\mu\mathrm{m}$86 meV and deeper ones around $2.0~\mu\mathrm{m}$87–$2.0~\mu\mathrm{m}$88 meV, with an additional $2.0~\mu\mathrm{m}$89 meV process at high power. Time-resolved PL shows a rise time increasing monotonically from about $2.0~\mu\mathrm{m}$90 ps at $2.0~\mu\mathrm{m}$91 K to about $2.0~\mu\mathrm{m}$92 ps at $2.0~\mu\mathrm{m}$93 K, while the dominant decay remains near $2.0~\mu\mathrm{m}$94 ns, indicating that carrier loss occurs mainly before capture into the dots (Shikin et al., 2018).

DBR-mediated excitation in a C-band InAs/InP structure demonstrates that even photonic mirrors can become electronic reservoirs. In a device with an InP/$2.0~\mu\mathrm{m}$95 bottom DBR, the InGaAlAs band edge appears at $2.0~\mu\mathrm{m}$96 eV and the QD ensemble at $2.0~\mu\mathrm{m}$97 eV $2.0~\mu\mathrm{m}$98 nm). Below saturation, the QD ensemble intensity increases with temperature before quenching: by a factor of about $2.0~\mu\mathrm{m}$99 for $0.5~\mu\mathrm{m}$00 excitation with a maximum near $0.5~\mu\mathrm{m}$01 K, $0.5~\mu\mathrm{m}$02 for $0.5~\mu\mathrm{m}$03 near $0.5~\mu\mathrm{m}$04 K, and $0.5~\mu\mathrm{m}$05 for $0.5~\mu\mathrm{m}$06 mW near $0.5~\mu\mathrm{m}$07 K. The increase is anticorrelated with a decrease in the InGaAlAs emission, and PLE shows a roughly $0.5~\mu\mathrm{m}$08 drop when the excitation energy is taken below the DBR edge, directly proving carrier transfer from the DBR reservoir to the dots. Because the valence-band alignment favors hole transfer, the work identifies this path as potentially detrimental to coherence (Musiał et al., 2024).

6. Symmetry, tuning strategies, and broader device context

Structural symmetry is one of the defining advantages of InP-platform droplet epitaxy. On InP(111)A, the $0.5~\mu\mathrm{m}$09 surface symmetry eliminates the elongation that typically drives anisotropic exchange, and prior work cited within that literature reported a $0.5~\mu\mathrm{m}$10 probability of zero fine-structure splitting for InAs/In(Al,Ga)As on InP(111)A. Combined with the barrier-tunable wavelength range from the O band through the C and L bands and beyond $0.5~\mu\mathrm{m}$11, this makes the (111)A droplet platform a realistic route to telecom entangled-photon sources (Ha et al., 2016). A plausible implication is that symmetry control and wavelength control are unusually decoupled on this surface, because the measured shape distribution is essentially independent of barrier choice.

For integrated photonics, the same platform supports several complementary strategies. Near-critical Stranski–Krastanov growth gives low areal densities down to approximately $0.5~\mu\mathrm{m}$12, so isolated C-band lines can be obtained without aggressive post-processing (Berdnikov et al., 2023). Stark tuning supplies a second degree of freedom by spectrally aligning nominally dissimilar emitters over $0.5~\mu\mathrm{m}$13 meV without charge-state jumps (Aghaeimeibodi et al., 2018). Conversely, the DBR-mediated excitation study shows that quaternary mirror layers and native InP defects can introduce long capture paths and imbalanced carrier occupation, so electronic decoupling of photonic mirrors from the active dots becomes part of the device-design problem rather than a purely optical consideration (Musiał et al., 2024).

The label “InAs/InGaAs/InP quantum dots” also borders a distinct electrostatic tradition in InGaAs/InP heterostructures. In gated InGaAs/InP quantum wells grown by chemical-beam epitaxy, few-electron quantum dots can form beneath narrow Schottky gates through potential fluctuations. In that setting, Coulomb-blockade diamonds give a charging energy $0.5~\mu\mathrm{m}$14 meV and lever arm $0.5~\mu\mathrm{m}$15, photon-assisted tunneling is observed under $0.5~\mu\mathrm{m}$16–$0.5~\mu\mathrm{m}$17 GHz microwave irradiation, and a singlet–triplet transition occurs near $0.5~\mu\mathrm{m}$18 T (Granger et al., 2010). This is a different architectural class from self-assembled InAs/InP emitters, but it underscores that the InP platform supports both optically active self-assembled dots and laterally confined few-electron devices.

Taken together, the InAs/InP family and its InGaAs-, InGaAlAs-, and In(Al,Ga)As-based variants define a broad InP-platform quantum-dot ecosystem rather than a single heterostructure recipe. The central materials trade-off is consistent across that ecosystem: stronger confinement and cleaner symmetry favor spectral stability, small anisotropy, and predictable few-particle structure, whereas reduced barriers, wetting-layer localization, and defect-assisted transfer improve wavelength reach or photonic integration at the cost of thermal robustness and carrier-relaxation simplicity (Gong et al., 2010, Zieliński, 2013, Jahan et al., 2012, Holewa et al., 2020).