Influence of the atomic-scale structure on the exciton fine-structure splitting in InGaAs and GaAs quantum dots in a vertical electric field

Luo, Jun-Wei; Singh, Ranber; Zunger, Alex; Bester, Gabriel

doi:10.1103/physrevb.86.161302

Cited by 19 publications

(24 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This formula is expected from the classical "folded down" perturbation approach [1][2][3][4][5][6][7][11][12][13][14]16 . On the other hand, all data points of strained In(Ga)As/GaAs QDs exhibit a denominator ∆ HL that is effectively reduced by δ = 78.6 meV with respect to the class of unstrained GaAs QDs and fall then close to another curve:…”

Section: K · P Analysis Of Atomistic Microscopic Resultsmentioning

confidence: 99%

“…The perturbation potential δV C 2v lowers the QD symmetry to C 2v and hence introduces HH-LH mixing, which modifies the ground hole state HH0 [1][2][3][4][5][6][7][11][12][13][14]16,19,22 ,…”

Section: Discussion Of the Supercoupling Effectmentioning

confidence: 99%

“…We are thus free from the need to pre-judge at the outset which 3D bands will couple in nanostructrues, a judgment needed in the truncation of the k · p expansion. This method has been tested extensively over the past two decades for a broad range of spectroscopic quantities in self-assembled as well as colloidal nanostructures 4,10,[27][28][29][30][31][32][33] . The well known k · p approaches can also be in principle accurate at the limit of a large number (possibly infinite) of band basis functions 18 .…”

Section: Theoretical Methodologymentioning

confidence: 99%

“…Well established literature tests [31][32][33] have shown that even the state-of-the-art k · p using 8 basis functions and displacing atoms according to atomistic (valence force field) elasticity theory 26 still shows significant deviations from the corresponding all band pseudopotential description of the same self-assembled InGaAs QDs 33 as well as colloidal CdSe and InP dots 31,32 . Although we believe that a sufficiently complete basis set k · p approach might be able to convey realistically the HH-LH mixing, we wish to circumvent the uncertainties related to the convergence with number of basis states by using instead the well-tested 4,10,[27][28][29][30][31][32][33] all-bulk band direct diagonalization approach that treats a QD explicitly as a giant molecule without making the k · p expansion. By applying this approach to a broad range of QD shapes, symmetries and compositions we found an unexpected mechanism of supercoupling.…”

Section: Theoretical Methodologymentioning

confidence: 99%

“…The two components of the valence band have diametrically opposed fundamental features (including different Bloch functions, pseudospin, type of optical transition, magnitude of effective masses and deformation potentials), so the extent of their quantum mechanical mixing underlies our fundamental understanding of the duality of a broad range of physical properties associated with hole in low-dimensional nanostructures. Specifically, such HH-LH mixing plays important roles in e.g., QDs 1 , including (i) tuning of the excitonic fine-structure splitting [2][3][4] which controls the fidelity of entangled photon pairs,…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Supercoupling between heavy-hole and light-hole states in nanostructures

2015

Self Cite

View full text Add to dashboard Cite

The heavy-hole (HH) and light-hole (LH) components of the valence states in 3D bulk semiconductors can mix quantum mechanically as the dimensionality is reduced in forming 2D, 1D nanostructures and 0D quantum dots (QDs). This coupling controls the tuning of the excitonic fine structure splitting, provides an efficient channel for the spin coherence, and leads to polarization anisotropy of light emission, central to several quantum information schemes. Current understanding is that the mixing scales with the square of δV HL /∆ HL , where δV HL and ∆ HL are the coupling matrix elements of the crystal potential and the energy separation between the primary HH0 and LH0 states, respectively. We discuss two classes of HH-LH coupling mechanisms.First, coupling factors occurring through the numerator δV HL , referred to as "direct coupling", including the well-known (1) quantum confinement, (2) built-in strain, and (3) shape elongation, as well as three additional direct coupling mechanisms discussed here: (4) the intrinsic C 2v crystal field effect, (5) the local symmetry of the interface, and (6) the alloy disorder. We quantify these 6 direct HH-LH coupling effects by performing atomistic pseudopotential calculations on a range of strained and unstrained QDs of different morphologies. We find that in unstrained self-assembled QDs such as GaAs/AlGaAs effects (1)-(6) contribute 0%, 0%, 0%, 0%, 40%, and 60%, respectively, whereas in strained self-assembled QDs such as InGaAs/GaAs they contribute 0%, 0%, 78%, 0%, 8%, and 14%, respectively, to the direct HH-LH coupling δV HL . These relative contributions to direct HH-LH coupling differ significantly from what was previously believed.Second, we discover an unexpected HH-LH coupling that effectively reduces the denominator ∆ HL by the presence of a dense ladder of intermediate states between the HH0 and LH0 states (analogous to super-exchange in magnetism). Supercoupling amplifies and propagates the HH-LH interaction and is the dominant source of HH-LH mixing in strained nanostructures where ∆ HL is fairly large, so by the direct coupling mechanism alone δV HL /∆ HL would be expected to be rather small. Supercoupling explains a number of outstanding puzzles including the surprising fact that in strained (InAs/GaAs) QDs the mixing is very strong despite the fact that ∆ HL is large, and offers a new way to manipulate HH-LH mixing and hence associated properties in nanostructures.

show abstract