FIG 1 - uploaded by J. Martinez-Pastor
Content may be subject to copyright.
AFM images of a 2.5-ML-thick InAs layer deposited on a 200 nm buffer layer grown by ALMBE ͑ a ͒ and MBE ͑ b ͒ on an InP ͑ 001 ͒ substrate.
Source publication
We have studied the influence of InP buffer-layer morphology in the formation of InAs nanostructures grown on InP(001) substrates by solid-source molecular-beam epitaxy. Our results demonstrate that when InP buffer layers are grown by atomic-layer molecular-beam epitaxy, InAs quantum dot-like structures are formed, whereas InP buffer layers grown b...
Contexts in source publication
Context 1
... nanostructures grown on InP 001 are promising candidates for light emitting devices in the wavelength range 1.3–1.55 m. 1–3 A widely investigated technological ap- proach is to use self-organized nanostructures that appear spontaneously as an efficient way to relax elastic strain in lattice-mismatched heteroepitaxy. In order to obtain a well- defined emission wavelength, nanostructures should have a good uniformity in size and shape. Ordering in the growth plane and a precise control of the nanostructure morphology is a key issue for self-organized systems. Due to a 3.2% lattice mismatch of InAs on InP ͑ 001 ͒ , elastic strain relaxation takes place above a certain critical thickness via a change of morphology from two-dimensional ͑ 2D ͒ to three- dimensional ͑ 3D ͒ . Much of the research developed on self- organized growth of InAs on InP substrates is devoted to quantum-dot ͑ QD ͒ formation. 1–4 However, very recent results 5 have shown that the chemical composition of the buffer layer ͑ InP, InGaAs, or InAlAs ͒ is determinant in con- figuring the final shape of the InAs self-organized nanostructures. In this letter, we demonstrate that the growth conditions of the InP buffer layer also controls the surface rearrange- ment of the strained InAs layer grown on top. Therefore, it is possible to obtain either QD or quantum-wire ͑ QWr ͒ structures for identical InAs coverage and growth conditions. There are two sets of samples grown for this work. In the first one, InAs was deposited on a 200-nm-thick InP buffer layer grown either by molecular-beam epitaxy ͑ MBE ͒ or by atomic-layer molecular-beam epitaxy ͑ ALMBE ͒ ͑ Ref. 6 ͒ using solid sources. MBE buffer layers were grown in the 2 ϫ 4 surface reconstruction at T s ϭ 460 °C and at a beam- equivalent pressure ͑ BEP ͒ (P 2 ) ϭ 5 ϫ 10 Ϫ 6 Torr. For the ALMBE InP buffer layers, a substrate temperature T s ϭ 400 °C is used. The P 2 pulsed flux reaching the surface sample is controlled by means of the reflectivity difference RD technique in order to optimize surface stoichiometry for growing InP planar surfaces. 6 The InAs layers, 2.5 ͑ ML ͒ thick, were deposited at a BEP (As 4 ) ϭ 1.5– 2 ϫ 10 Ϫ 6 Torr, a growth rate of 0.5 ML/s, and T s ϭ 400 °C, which is chosen to minimize the P/As exchange. After InAs growth, an annealing at 480 °C under arsenic pressure during 10–20 s was performed. We observe, in agreement with other authors, 3,4 that InAs growth takes place in a 2D mode. The 2D–3D transition occurs during the annealing process at T s ϭ 480 °C, as shown by a change into a spotty reflection high-energy electron diffraction ͑ RHEED ͒ pattern. The sec- ond set of samples consist of 5-, 8-, and 12-ML-thick In x Ga 1 Ϫ x As layers with In contents x ϭ 0.95, 0.9, and 0.85, respectively, deposited on InP buffer layers grown by MBE. We have increased the layer thickness for decreasing x according to the expected dependence of critical thickness with misfit. 7 As-grown samples were either immediately cooled down and taken out of the growth chamber for morphology studies by atomic-force microscopy ͑ AFM ͒ , or covered with a 50- nm-thick InP cap layer for optical and structural character- ization by high-resolution transmission electron microscopy ͑ HRTEM ͒ . Photoluminescence ͑ PL ͒ experiments were performed as a function of temperature from 18 to 300 K. The signal was analyzed with a linear polarizer set in front of a 0.22 m SPEX single monochromator and synchronously de- tected by a 77 K cooled Ge photodetector. Figure 1 ͑ a ͒ shows the AFM image of an InAs/InP sample in which the InP buffer layer has been grown by ALMBE. The self-organized structures have a QD shape, whose height ranges between 1.6 and 6.8 nm, a full width at half maximum ͑ FWHM ͒ W Ϸ 26 nm, and a typical density Ϸ 2 ϫ 10 10 cm Ϫ 2 . On the other hand, the same amount of InAs deposited on a InP buffer layer grown by MBE gives rise to periodic undulations ͓ Fig. 1 ͑ b ͔͒ . These QWr-like structures are oriented along the ͓ 11 ̄ 0 ͔ direction. The height of the wires, h ranges between 0.6 and 2 nm and the length exceeds 1 m. Their FWHM W is around 18 nm and the pitch period is about 24 nm. Meanwhile, AFM profiles over a selected QWr along the ͓ 11 ̄ 0 ͔ direction show height fluctuations below 2 ML, h can vary up to 7 ML between different wires. The above given average dimensions for QD and QWr structures are consistent with the amount of deposited InAs. So, within the error introduced by the AFM estimate of the shape and even possible inaccuracies in the growth parameters, we can say that there is no significant amount of material from the buffer layer involved in the formation of the nanostructures. HRTEM images ͑ not shown ͒ from ͕ 110 ͖ cross sections of capped samples show that InAs QWr-like structures remain after capping with InP. However, no clear chemical contrast between InAs and InP could be obtained under any imaging conditions. The PL spectra of the different QWr-like samples ͑ Fig. 2 ͒ evidence a strong light emission centered around 1.55 m at room temperature ͑ RT ͒ , whose overall integrated intensity is about 20 times less intense than that observed at low temperatures. At lower temperatures, several lines can be re- solved from the PL spectra ͑ Fig. 2 ͒ . The measured linewidths of such individual lines are around 30 meV. The intensity of each individual line exhibits a different temperature evolu- tion. When increasing temperature, the excitons jump from the smaller wires ͑ probably by thermal scape ͒ to the higher ones, whose PL intensity begins to decrease above 200 K. 8 The PL peak energies agree approximately with those found in strained InAs/InP quantum wells 2,9 thicker than about 3–4 ML, and the energy separation is consistent with 1-ML- thickness fluctuation. The different PL lines of the QWr-like structures can be correlated with height variations from 3 to 9 ML, in good agreement with the h range found by AFM. No evidences of the existence of a wetting layer are found in our PL data, in accordance with the AFM estimate of the QWr volume and period. Another evidence for strong lateral confinement in our QWr-like structures lies in the polarization degree of the emitted light. We observe that PL intensity is stronger when the polarizer is set parallel to the wire direction ͓ 11 ̄ 0 ͔ , as shown in Fig. 2. The polarization degree, defined as P ϭ ( I ͓ 11 ̄ 0 ͔ Ϫ I ͓ 110 ͔ )/( I ͓ 11 ̄ 0 ͔ ϩ I ͓ 110 ͔ ), is independent of incident light polarization, and practically constant with temperature P Ϸ 30%. QWr-like structures are similar in size and shape to those obtained for the InAs/InGaAs/InP system reported in Ref. 5. In our case, QD or QWr structures appear using the same InAs growth process, so we can exclude the formation of an InGaAs intermediate layer as the cause of QWr formation. The difference in the surface morphology of the InAs layer, QD, or QWr, can be only ascribed to the different growth processes of the InP buffer layer. In strained heteroepitaxial systems, QD structures are commonly self-assembled. In the case of the QWr, we should ask about the role of the buffer layer on the QWr formation process. Nucleation of 3D features is usually as- sumed to be mainly controlled by elastic strain, surface energy, and surface diffusion kinetics. On one hand, strain and surface energy are characteristic magnitudes for a certain heteroepitaxial system ͓ InAs on InP ͑ 001 ͒ , in this case ͔ . On the other hand, surface diffusion kinetics can be dominated by extrinsic parameters. In particular, anisotropic buffer- layer morphology can make the surface diffusion different along the ͓ 110 ͔ and ͓ 11 ̄ 0 ͔ directions. In this case, the development of asymmetric morphologies during the 2D–3D transition seems a reasonable consequence. In fact, we have observed that InP buffer layers grown by MBE systematically show an anisotropic roughness with features preferentially aligned in the ͓ 11 ̄ 0 ͔ direction. Similar features have been observed in homoepitaxial growth of InP ͑ Ref. 10 ͒ and GaAs ͑ Ref. 11 ͒ on ͑ 001 ͒ -oriented substrates. Surface reconstruction during growth can play a decisive role in the observed growth-front morphology, as has been proposed. 11,12 A fast surface diffusion direction along the ͓ 11 ̄ 0 ͔ direction in the 2 ϫ 4 reconstructed surface will lead to a preferential island nucleation and to surfaces with a larger density of steps along that direction. Moreover, the higher reactivity of the step edges oriented along ͓ 110 ͔ enhances the anisotropy during growth. 13 Therefore, the development of an anisotropic morphology in the InP buffer layer could be responsible of the appearance of a periodic corrugation in the InAs layer as the best way for elastic strain relaxation. This idea is reinforced by the observed fact that QDs, instead of QWrs, are formed not only when InP buffer layers are grown by ...
Context 2
... nanostructures grown on InP 001 are promising candidates for light emitting devices in the wavelength range 1.3–1.55 m. 1–3 A widely investigated technological ap- proach is to use self-organized nanostructures that appear spontaneously as an efficient way to relax elastic strain in lattice-mismatched heteroepitaxy. In order to obtain a well- defined emission wavelength, nanostructures should have a good uniformity in size and shape. Ordering in the growth plane and a precise control of the nanostructure morphology is a key issue for self-organized systems. Due to a 3.2% lattice mismatch of InAs on InP ͑ 001 ͒ , elastic strain relaxation takes place above a certain critical thickness via a change of morphology from two-dimensional ͑ 2D ͒ to three- dimensional ͑ 3D ͒ . Much of the research developed on self- organized growth of InAs on InP substrates is devoted to quantum-dot ͑ QD ͒ formation. 1–4 However, very recent results 5 have shown that the chemical composition of the buffer layer ͑ InP, InGaAs, or InAlAs ͒ is determinant in con- figuring the final shape of the InAs self-organized nanostructures. In this letter, we demonstrate that the growth conditions of the InP buffer layer also controls the surface rearrange- ment of the strained InAs layer grown on top. Therefore, it is possible to obtain either QD or quantum-wire ͑ QWr ͒ structures for identical InAs coverage and growth conditions. There are two sets of samples grown for this work. In the first one, InAs was deposited on a 200-nm-thick InP buffer layer grown either by molecular-beam epitaxy ͑ MBE ͒ or by atomic-layer molecular-beam epitaxy ͑ ALMBE ͒ ͑ Ref. 6 ͒ using solid sources. MBE buffer layers were grown in the 2 ϫ 4 surface reconstruction at T s ϭ 460 °C and at a beam- equivalent pressure ͑ BEP ͒ (P 2 ) ϭ 5 ϫ 10 Ϫ 6 Torr. For the ALMBE InP buffer layers, a substrate temperature T s ϭ 400 °C is used. The P 2 pulsed flux reaching the surface sample is controlled by means of the reflectivity difference RD technique in order to optimize surface stoichiometry for growing InP planar surfaces. 6 The InAs layers, 2.5 ͑ ML ͒ thick, were deposited at a BEP (As 4 ) ϭ 1.5– 2 ϫ 10 Ϫ 6 Torr, a growth rate of 0.5 ML/s, and T s ϭ 400 °C, which is chosen to minimize the P/As exchange. After InAs growth, an annealing at 480 °C under arsenic pressure during 10–20 s was performed. We observe, in agreement with other authors, 3,4 that InAs growth takes place in a 2D mode. The 2D–3D transition occurs during the annealing process at T s ϭ 480 °C, as shown by a change into a spotty reflection high-energy electron diffraction ͑ RHEED ͒ pattern. The sec- ond set of samples consist of 5-, 8-, and 12-ML-thick In x Ga 1 Ϫ x As layers with In contents x ϭ 0.95, 0.9, and 0.85, respectively, deposited on InP buffer layers grown by MBE. We have increased the layer thickness for decreasing x according to the expected dependence of critical thickness with misfit. 7 As-grown samples were either immediately cooled down and taken out of the growth chamber for morphology studies by atomic-force microscopy ͑ AFM ͒ , or covered with a 50- nm-thick InP cap layer for optical and structural character- ization by high-resolution transmission electron microscopy ͑ HRTEM ͒ . Photoluminescence ͑ PL ͒ experiments were performed as a function of temperature from 18 to 300 K. The signal was analyzed with a linear polarizer set in front of a 0.22 m SPEX single monochromator and synchronously de- tected by a 77 K cooled Ge photodetector. Figure 1 ͑ a ͒ shows the AFM image of an InAs/InP sample in which the InP buffer layer has been grown by ALMBE. The self-organized structures have a QD shape, whose height ranges between 1.6 and 6.8 nm, a full width at half maximum ͑ FWHM ͒ W Ϸ 26 nm, and a typical density Ϸ 2 ϫ 10 10 cm Ϫ 2 . On the other hand, the same amount of InAs deposited on a InP buffer layer grown by MBE gives rise to periodic undulations ͓ Fig. 1 ͑ b ͔͒ . These QWr-like structures are oriented along the ͓ 11 ̄ 0 ͔ direction. The height of the wires, h ranges between 0.6 and 2 nm and the length exceeds 1 m. Their FWHM W is around 18 nm and the pitch period is about 24 nm. Meanwhile, AFM profiles over a selected QWr along the ͓ 11 ̄ 0 ͔ direction show height fluctuations below 2 ML, h can vary up to 7 ML between different wires. The above given average dimensions for QD and QWr structures are consistent with the amount of deposited InAs. So, within the error introduced by the AFM estimate of the shape and even possible inaccuracies in the growth parameters, we can say that there is no significant amount of material from the buffer layer involved in the formation of the nanostructures. HRTEM images ͑ not shown ͒ from ͕ 110 ͖ cross sections of capped samples show that InAs QWr-like structures remain after capping with InP. However, no clear chemical contrast between InAs and InP could be obtained under any imaging conditions. The PL spectra of the different QWr-like samples ͑ Fig. 2 ͒ evidence a strong light emission centered around 1.55 m at room temperature ͑ RT ͒ , whose overall integrated intensity is about 20 times less intense than that observed at low temperatures. At lower temperatures, several lines can be re- solved from the PL spectra ͑ Fig. 2 ͒ . The measured linewidths of such individual lines are around 30 meV. The intensity of each individual line exhibits a different temperature evolu- tion. When increasing temperature, the excitons jump from the smaller wires ͑ probably by thermal scape ͒ to the higher ones, whose PL intensity begins to decrease above 200 K. 8 The PL peak energies agree approximately with those found in strained InAs/InP quantum wells 2,9 thicker than about 3–4 ML, and the energy separation is consistent with 1-ML- thickness fluctuation. The different PL lines of the QWr-like structures can be correlated with height variations from 3 to 9 ML, in good agreement with the h range found by AFM. No evidences of the existence of a wetting layer are found in our PL data, in accordance with the AFM estimate of the QWr volume and period. Another evidence for strong lateral confinement in our QWr-like structures lies in the polarization degree of the emitted light. We observe that PL intensity is stronger when the polarizer is set parallel to the wire direction ͓ 11 ̄ 0 ͔ , as shown in Fig. 2. The polarization degree, defined as P ϭ ( I ͓ 11 ̄ 0 ͔ Ϫ I ͓ 110 ͔ )/( I ͓ 11 ̄ 0 ͔ ϩ I ͓ 110 ͔ ), is independent of incident light polarization, and practically constant with temperature P Ϸ 30%. QWr-like structures are similar in size and shape to those obtained for the InAs/InGaAs/InP system reported in Ref. 5. In our case, QD or QWr structures appear using the same InAs growth process, so we can exclude the formation of an InGaAs intermediate layer as the cause of QWr formation. The difference in the surface morphology of the InAs layer, QD, or QWr, can be only ascribed to the different growth processes of the InP buffer layer. In strained heteroepitaxial systems, QD structures are commonly self-assembled. In the case of the QWr, we should ask about the role of the buffer layer on the QWr formation process. Nucleation of 3D features is usually as- sumed to be mainly controlled by elastic strain, surface energy, and surface diffusion kinetics. On one hand, strain and surface energy are characteristic magnitudes for a certain heteroepitaxial system ͓ InAs on InP ͑ 001 ͒ , in this case ͔ . On the other hand, surface diffusion kinetics can be dominated by extrinsic parameters. In particular, anisotropic buffer- layer morphology can make the surface diffusion different along the ͓ 110 ͔ and ͓ 11 ̄ 0 ͔ directions. In this case, the development of asymmetric morphologies during the 2D–3D transition seems a reasonable consequence. In fact, we have observed that InP buffer layers grown by MBE systematically show an anisotropic roughness with features preferentially aligned in the ͓ 11 ̄ 0 ͔ direction. Similar features have been observed in homoepitaxial growth of InP ͑ Ref. 10 ͒ and GaAs ͑ Ref. 11 ͒ on ͑ 001 ͒ -oriented substrates. Surface reconstruction during growth can play a decisive role in the observed growth-front morphology, as has been proposed. 11,12 A fast surface diffusion direction along the ͓ 11 ̄ 0 ͔ direction in the 2 ϫ 4 reconstructed surface will lead to a preferential island nucleation and to surfaces with a larger density of steps along that direction. Moreover, the higher reactivity of ...
Context 3
... 1 ͑ b ͔͒ . These QWr-like structures are oriented along the ͓ 11 ̄ 0 ͔ direction. The height of the wires, h ranges between 0.6 and 2 nm and the length exceeds 1 m. Their FWHM W is around 18 nm and the pitch period is about 24 nm. Meanwhile, AFM profiles over a selected QWr along the ͓ 11 ̄ 0 ͔ direction show height fluctuations below 2 ML, h can vary up to 7 ML between different wires. The above given average dimensions for QD and QWr structures are consistent with the amount of deposited InAs. So, within the error introduced by the AFM estimate of the shape and even possible inaccuracies in the growth parameters, we can say that there is no significant amount of material from the buffer layer involved in the formation of the nanostructures. HRTEM images ͑ not shown ͒ from ͕ 110 ͖ cross sections of capped samples show that InAs QWr-like structures remain after capping with InP. However, no clear chemical contrast between InAs and InP could be obtained under any imaging conditions. The PL spectra of the different QWr-like samples ͑ Fig. 2 ͒ evidence a strong light emission centered around 1.55 m at room temperature ͑ RT ͒ , whose overall integrated intensity is about 20 times less intense than that observed at low temperatures. At lower temperatures, several lines can be re- solved from the PL spectra ͑ Fig. 2 ͒ . The measured linewidths of such individual lines are around 30 meV. The intensity of each individual line exhibits a different temperature evolu- tion. When increasing temperature, the excitons jump from the smaller wires ͑ probably by thermal scape ͒ to the higher ones, whose PL intensity begins to decrease above 200 K. 8 The PL peak energies agree approximately with those found in strained InAs/InP quantum wells 2,9 thicker than about 3–4 ML, and the energy separation is consistent with 1-ML- thickness fluctuation. The different PL lines of the QWr-like structures can be correlated with height variations from 3 to 9 ML, in good agreement with the h range found by AFM. No evidences of the existence of a wetting layer are found in our PL data, in accordance with the AFM estimate of the QWr volume and period. Another evidence for strong lateral confinement in our QWr-like structures lies in the polarization degree of the emitted light. We observe that PL intensity is stronger when the polarizer is set parallel to the wire direction ͓ 11 ̄ 0 ͔ , as shown in Fig. 2. The polarization degree, defined as P ϭ ( I ͓ 11 ̄ 0 ͔ Ϫ I ͓ 110 ͔ )/( I ͓ 11 ̄ 0 ͔ ϩ I ͓ 110 ͔ ), is independent of incident light polarization, and practically constant with temperature P Ϸ 30%. QWr-like structures are similar in size and shape to those obtained for the InAs/InGaAs/InP system reported in Ref. 5. In our case, QD or QWr structures appear using the same InAs growth process, so we can exclude the formation of an InGaAs intermediate layer as the cause of QWr formation. The difference in the surface morphology of the InAs layer, QD, or QWr, can be only ascribed to the different growth processes of the InP buffer layer. In strained heteroepitaxial systems, QD structures are commonly self-assembled. In the case of the QWr, we should ask about the role of the buffer layer on the QWr formation process. Nucleation of 3D features is usually as- sumed to be mainly controlled by elastic strain, surface energy, and surface diffusion kinetics. On one hand, strain and surface energy are characteristic magnitudes for a certain heteroepitaxial system ͓ InAs on InP ͑ 001 ͒ , in this case ͔ . On the other hand, surface diffusion kinetics can be dominated by extrinsic parameters. In particular, anisotropic buffer- layer morphology can make the surface diffusion different along the ͓ 110 ͔ and ͓ 11 ̄ 0 ͔ directions. In this case, the development of asymmetric morphologies during the 2D–3D transition seems a reasonable consequence. In fact, we have observed that InP buffer layers grown by MBE systematically show an anisotropic roughness with features preferentially aligned in the ͓ 11 ̄ 0 ͔ direction. Similar features have been observed in homoepitaxial growth of InP ͑ Ref. 10 ͒ and GaAs ͑ Ref. 11 ͒ on ͑ 001 ͒ -oriented substrates. Surface reconstruction during growth can play a decisive role in the observed growth-front morphology, as has been proposed. 11,12 A fast surface diffusion direction along the ͓ 11 ̄ 0 ͔ direction in the 2 ϫ 4 reconstructed surface will lead to a preferential island nucleation and to surfaces with a larger density of steps along that direction. Moreover, the higher reactivity of the step edges oriented along ͓ 110 ͔ enhances the anisotropy during growth. 13 Therefore, the development of an anisotropic morphology in the InP buffer layer could be responsible of the appearance of a periodic corrugation in the InAs layer as the best way for elastic strain relaxation. This idea is reinforced by the observed fact that QDs, instead of QWrs, are formed not only when InP buffer layers are grown by ALMBE ͓ Fig. 1 ͑ a ͔͒ , but also when InAs layers are deposited directly on the InP substrate, after desorbing the oxide, or on top of a very thin InP buffer layer ͑ 8.5 nm ͒ grown by conventional MBE. In other words, nucleation of the QD would take place when the buffer-layer asymmetry has not been fully developed. We would expect the QWr dimensions will depend on the misfit strain of the epilayer, given that strain relaxation is the driving mechanism for the formation of 3D structures. The misfit strain can be easily tuned by growing In x Ga 1 Ϫ x As layers of different compositions, as was described above. Our results show that QWr-like structures are also formed along ͓ 11 ̄ 0 ͔ after deposition of an appropriate amount of In x Ga 1 Ϫ x As and a subsequent anneal under As 4 . The dependence of the pitch period on the lattice mismatch, ⑀ can be observed The authors in Fig. 3. wish These to data acknowledge can be fitted the to an Spanish expres- ‘‘CICYT’’ sion ⑀ 2 0 ϭ for K , which financial is obtained support under from energy Project balance No. TIC96- con- 1020-C02. siderations of the surface roughness developed for stabiliza- tion of stressed surfaces. 14,15 In conclusion, our results show that once there is an established anisotropy in the InP surface buffer layer, the 2D InAs layer transforms into periodic QWr-like structures to allow elastic strain relaxation. We have demonstrated that the period of the wires is determined by the misfit strain between the layer and the substrate. QWr-like structures capped with InP exhibit a strong PL emission at 1.55 m at RT. The emission is anisotropic with a degree of polarization around a 30%. The authors wish to acknowledge the Spanish ‘‘CICYT’’ for financial support under Project No. TIC96- ...
Similar publications
Uniform, compact, and thick InGaN layers are grown on Si(111) substrates by plasma-assisted molecular beam epitaxy without any buffer layers at low temperatures of around 320 �C. By adjusting the Ga/In flux ratio, InGaN layers with In compositions between 10 and 33% are obtained, providing emission covering the whole visible spectral range. The In...
Citations
... However, single emitters are needed for this, i.e., QDs spatially and spectrally (energetically) isolated from others. At this point, a problem arises, as quantum dots manufactured in the InP-based material systems naturally grow in dense ensembles [4][5][6][7] , resulting from the relatively small lattice mismatch between the QD and barrier materials. The latter, combined with in-plane anisotropic indium diffusion, which is one of the driving factors in the growth by molecular beam epitaxy, causes such QDs to usually grow in a highly elongated shape [7][8][9] , which results in a dense and atypical spectrum of exciton states 10,11 with significant fine structure splitting 12 and specific optical properties 13 . ...
We calculate the spectrum of excited exciton states in application-relevant self-assembled pyramidal quantum dots grown in InAs/InP and InAs/AlGaInAs material systems. These types of dots have been recently shown to combine the emission in the third optical fiber window with low surface density and a reasonable level of in-plane symmetry of emitters, which predestines them for studies on single- and entangled-photon emission and for corresponding applications. The spectrum of optically active excited states is crucial for successful resonant and quasi-resonant excitation of emitters, allowing for conservation of angular momentum and addressing individual selected quantum states. Here, we show that in both types of studied dots, due to their specific morphology of truncated pyramid, the density of excited-state ladder, especially the s–p shell splitting may follow an unconventional dependence on emission energy, opposite to the one typically met in regular quantum dots. We obtain this result via modeling based on available morphological data and calculation within the multi-band $${{\varvec{k}} {\cdot } {\varvec{p}}}$$ k · p envelope-function theory combined with the configuration-interaction method used to calculate exciton states. Then, we explain this observation in purely geometric terms, as a result of an increasing effective quantum confinement width in a pyramid that is progressively cut from the top. Additionally, we show that the inverted trend is also manifested in the amount of electron-hole correlation in the exciton ground state, which also shows an anomalous dependence on emission energy and quantum dot volume.
... As a result, considerable research efforts have gone into obtaining room-temperature light emission from InAs QDs in these materials systems. 149,[181][182][183][184][185][186][187] An unusual feature of InAs self-assembly in these InP-based systems is the ability to synthesize either discrete QDs or elongated quantum wires (QWrs) depending on the MBE conditions, specific choice of buffer materials, or substrate offcut. 149,184,[186][187][188][189][190] The InAs QWrs tend to line up parallel to the 110 ½ direction due to the anisotropic relaxation of strain (Fig. 11). ...
... 149,[181][182][183][184][185][186][187] An unusual feature of InAs self-assembly in these InP-based systems is the ability to synthesize either discrete QDs or elongated quantum wires (QWrs) depending on the MBE conditions, specific choice of buffer materials, or substrate offcut. 149,184,[186][187][188][189][190] The InAs QWrs tend to line up parallel to the 110 ½ direction due to the anisotropic relaxation of strain (Fig. 11). 188,191 García et al. show that the group V stabilized (2 Â 4) surface reconstruction gives rise to QWr formation in heteroepitaxial systems involving different group V elements. ...
... 180,192 In all cases, the QWrs form parallel to the dimer rows of the (2 Â 4) reconstruction, which are aligned along 110 ½ . 191 Differences in QWr (or QD) morphology and light emission characteristics are observed when InAs is deposited on InP, In x Al 1Àx As, or In x Ga 1Àx As. 149,[181][182][183][184][185][186][187]193 To grow self-assembled InAs QWr's or QDs, the InP(001) substrate oxide is desorbed at 500 C for 10 min. 194 Regardless of the buffer material used, a 200-400 nm smoothing layer suffices for QD growth. ...
Research into self-assembled semiconductor quantum dots (QDs) has helped advance numerous optoelectronic applications, ranging from solid-state lighting to photodetectors. By carefully controlling molecular beam epitaxy (MBE) growth parameters, we can readily tune QD light absorption and emission properties to access a broad portion of the electromagnetic spectrum. Although this field is now sufficiently mature that QDs are found in consumer electronics, research efforts continue to expand into new areas. By manipulating MBE growth conditions and exploring new combinations of materials, substrate orientations, and the sign of strain, a wealth of opportunities exist for synthesizing novel QD nanostructures with hitherto unavailable properties. As such, QDs are uniquely well positioned to make critical contributions to the development of future quantum technologies. In this tutorial, we summarize the history of self-assembled QDs, outline some examples of quantum optics applications based on QDs, discuss the science that explains the spontaneous formation of QDs, and provide recipes for successful QD growth by MBE for some of the most commonly used semiconductor materials systems. We hope that compiling this information in one place will be useful both for those new to QD self-assembly and for experienced researchers, ideally supporting the community’s efforts to continue pushing the boundaries of knowledge in this important field.
... Experimentally, the most commonly studied QDs based on S-K growth are InAs/GaAs [15][16][17][18][19], InAs/InP [20][21][22][23][24][25], InP/GaP [26][27][28][29], Ge/Si [30][31][32][33] and SiGe/Si [34][35][36]. To understand the epitaxial S-K growth, theoretical models have been developed from both thermodynamic [37][38][39][40][41][42][43][44][45][46][47][48][49] and kinetic [39,[50][51][52][53][54][55][56][57] point of view in last two decades. ...
... Towards saturation we can observe slightly increase in diameter of 46.2 nm and a height of 9.2 nm. The InAs QDs on InP(0 0 1) substrate mentioned in the literature also have height and diameter distribution 10-2 nm and 70-30 nm with density in a range of 3E9-3E10 cm −2 [20,21,[70][71][72][73][74]. ...
Quantum Dots (QDs) are considered as an efficient building block of many optoelectronic applications, such as semiconductor laser, photodetector, whereby their physical dimension is the key parameter to be controlled. In this work, we have studied experimentally the growth of InAs QDs on InP(0 0 1) substrate by MOVPE and established a theoretical model explaining the observed epitaxial behaviour. In variance with the classical Stranski-Krastanov growth, we show that during the growth there is an intermediate stage whereby although first large 3D islands are formed, an increased density of small QDs is formed concurrent with the shrinkage of the large 3D islands. As a result, the growth can be divided into three regimes: 2D layer growth, large 3D islands growth and QDs growth. To explain this evolution, a thermodynamic model has been developed accounting for the process driven by surface energy, elastic relaxation energy and inter-island interaction energy. It will be shown that the balance between the surface energy and the elastic relaxation energy provides the transition from 2D layer to large 3D islands (around at 3 ML of InAs growth). This model also supports the energetically favourable truncated pyramidal shape for large 3D islands and the spherical cap shape for QDs. We show in this paper that a balance between surface energy, elastic relaxation energy and inter-island elastic interaction explains the volume shrinkage of the early large 3D islands towards the formation of small QDs in high density.
... Despite these advantages, the control of properties from a dot to dot of an inhomogeneous ensemble of the InAs on InP system remains challenging. The growth of InAs nanostructures on the technologically important InP(001) substrate leads to the formation of strongly in-plane asymmetric and large nano-islands (called quantum dashes -QDashes) instead of more common in-plane symmetric QDs [21][22][23][24] . These kind of unique objects are preferentially elongated and aligned along the [1][2][3][4][5][6][7][8][9][10] crystallographic direction and are quite non-uniform in the length, size, internal strain and exhibit irregularities in their morphology 25 . ...
The exciton and biexciton confinement regimes in strongly anisotropic epitaxial InAs nanostructures called quantum dashes (QDashes) embedded in an In0.53Ga0.23Al0.24As matrix, which is lattice-matched to InP(001) substrate, have been investigated. For that purpose, we have performed low-temperature spatially and polarization-resolved photoluminescence and time-resolved photoluminescence measurements on a set of single QDashes. The main conclusions are drawn based on the experimentally obtained distribution of the ratio between the exciton and biexciton lifetimes. We have found that a majority of QDashes for which the abovementioned ratio falls into the range of 1.2 ± 0.1–1.6 ± 0.1 corresponds to the so called intermediate confinement regime, whereas for several cases, it is close to 1 or 2, suggesting reaching the conditions of weak and strong confinement, respectively. Eventually, we support this data with dependence of the lifetimes' ratio on the biexciton binding energy, implying importance of Coulomb correlations, which change significantly with the confinement regime.
... 7bshows that for InAs/InP , the luminescence can be tuned between 1.3 and 1.55 µm by adjusting the core or shell sizes. These wavelengths are very much sought after for telecommunications[38,39,40]. These diagramsare helpful to concept structures for various optoelectronic applications by tailoring their dimensions in the limits (a/b) 0 to a/b 0.5. ...
The control of charge carriers states in coated semiconductors quantum dots constitute a challenge to achieve a new range of tunable optoelectronic devices. The purpose of this work is to investigate the ground state energies and the behaviors of correlated and non correlated pair of electron (e) and hole (h) in and type I core/shell spherical quantum dots taking into account the discontinuities of the dielectric constants and effective masses. Analytical study of the uncorrelated e and h states allows to evaluate the critical values of the core size corresponding to the escape of the electron and hole from the core to the shell material. Then, the confined e and h wave functions and their fundamental energies are determined according to their localization in the core/shell structure. Knowing that the electron-hole interaction i.e “excitonic effect” turns out to be crucial for a comprehension of the optical spectra, we extend our concept to the investigation of the electron-hole correlations by a variational method.
... Usually these nanowalls are directly grown on bare substrates such as c-plane sapphire and Si(111) which is already known to incorporate defects from lattice mismatch considerations in normal epilayers. Buffer layers have already been used to obtain better crystallinity in nanostructures 12,13 earlier, and can be used to enhance the crystalline quality of nanowalls also. It would be useful to probe the effect of a thin AlN intermediate layer on the structural and optical properties of nanowalls grown by MBE. ...
In the present study, GaN nanowall network (NWN) structures are grown using rf-plasma assisted molecular beam epitaxy (PA-MBE) system under highly nitrogen rich conditions over nitrided and bare c-plane sapphire substrates. Nitridation is carried out prior to growth for two of the samples and its effect on the crystal quality of the resultant nanowall structure is studied and compared with the same structure grown on bare sapphire. It is found that nitriding the c-plane sapphire prior to growth results in an improvement of crystalline quality of the resulting nanowall as observed by the reduction in FWHM of (0002) XRC.
... One of the main technology hurdles for growing high quality InAs Qdots has been the inclination of elongated wire-like nanostructures formation along the [0-11] crystal directions known as Qdashes. The formation of Qdashes on InP substrate has been attributed to the small lattice mismatch of 3.2% between InAs and InP and the possible complex growth kinetics, growth conditions, buffer layer surface morphology and composition [22,23]. In addition, another major problem in the growth of InAs Qdots on InP substrate is the relatively large Qdots inhomogeneity compared to the InAs/GaAs Qdots system. ...
... MBE growth of InAs Qdashes instead of Qdots on InGaAlAs buffer layers has been explored to elucidate their growth kinetics and provide key insights for eventual realization of InAs Qdots on (100) InP substrate. For instance, the role of buffer layer surface morphology and alloy contents was examined in Ref. [22] as they affect the formation of the nanostructures; growth of high density InAs nanostructures resembling like elongated Qdots [36]; sparsely populated InAs nano-islands [37] on InGaAs, InAlAs, and InP buffer layers on (100) InP [23]. In this respect, Stinz et al. [38] suggested the utilization of vicinal (100) InP substrates for the growth of InAs Qdots. ...
... Early work of Qdash growth on this material system started from the investigations on InP buffer layer on nominal [22] and vicinal [201,202] (100) InP substrates on MBE system. Gonzalez et al. [22] demonstrated that the buffer layer surface morphology played an important role in the formation of InAs nanostructures by growing the InP buffer layer using atomic-layer MBE and MBE. ...
The advances in lasers, electronic and photonic integrated circuits (EPIC), optical interconnects as well as the modulation techniques allow the present day society to embrace the convenience of broadband, high speed internet and mobile network connectivity. However, the steep increase in energy demand and bandwidth requirement calls for further innovation in ultra-compact EPIC technologies. In the optical domain, advancement in the laser technologies beyond the current quantum well (Qwell) based laser technologies are already taking place and presenting very promising results. Homogeneously grown quantum dot (Qdot) lasers and optical amplifiers, can serve in the future energy saving information and communication technologies (ICT) as the work-horse for transmitting and amplifying information through optical fiber. The encouraging results in the zero-dimensional (0D) structures emitting at 980 nm, in the form of vertical cavity surface emitting laser (VCSEL), are already operational at low threshold current density and capable of 40 Gbp s error-free transmission at 108 fJ/bit. Subsequent achievements for lasers and amplifiers operating in the O–, C–, L–, U–bands, and beyond will eventually lay the foundation for green ICT. On the hand, the inhomogeneously grown quasi 0D quantum dash (Qdash) lasers are brilliant solutions for potential broadband connectivity in server farms or access network. A single broadband Qdash laser operating in the stimulated emission mode can replace tens of discrete narrow-band lasers in dense wavelength division multiplexing (DWDM) transmission thereby further saving energy, cost and footprint. We herein reviewed the progress of both Qdots and Qdash devices, based on the InAs/InGaAlAs/InP and InAs/InGaAsP/InP material systems, from the angles of growth and device performance. In particular, we discussed the progress in lasers, semiconductor optical amplifiers (SOA), mode locked lasers, and superluminescent diodes, which are the building blocks of EPIC and ICT. Alternatively, these optical sources are potential candidates for other multi-disciplinary field applications.
... In the last decades the development of new techniques for growing quantum well wires (QWWs) has allowed obtaining novel one-dimensional structures of various compositions, sizes, and shapes [1][2][3][4][5][6][7][8][9][10]. The theoretical study of the external field effects on the electronic structure and optical properties of these systems is one of great relevance for potential technological applications. ...
The conduction subband structure of a triangular cross-section GaAs/AlGaAs quantum well wire under magnetic field is theoretically investigated by taking into account a finite confining potential and two orientations of the field relative to the wire axis. The calculation of the subband energy levels is based on a two-dimensional finite element method within the effective mass approximation. It is shown that the magnetic field could be used for tuning the intersubband transitions: in the transverse field there is an obvious augment of the energy levels, whereas an axial field induces blueshifts/redshifts on the subband energies, depending on the azimuthal quantum number. We found that an axial orientation of the field allows the third harmonic generation and this process is enhanced for a particular polarization of the incident light and a proper field strength. A third-order nonlinear susceptibility with a peak value of 10−13(m/V)2 is predicted when the triple resonance condition is achieved.
... In this case the wires are formed by the Stranski-Krastanow growth mode, in which the materials that are deposited on top of each other have a substantially different lattice parameter. Spontaneous formation of self-assembled InAs quantum wires on InP (001) substrate, having 3.2% lattice mismatch, has been recently demonstrated [6,7]. These nanostructures are promising candidates for light-emitting devices for wavelengths 1.30 µm and 1.55 µm [8,9]. ...
... The growth of vertical 1D-nanostructures is governed by different strategies, e.g., nanorods and nanowires that are perpendicular or almost perpendicular to the substrate, and that have been produced by metallic catalytic and self-catalytic vapor-liquid-solid mechanism [10,11], by vapor-solid condensation by Gibbs energy reduction [12] such as molecular beam epitaxy technique [13], by vertical evolution on hydrothermal deposition of seeds with the same materials [14], etc. In addition to vertical nanowires, surface quantum wires (formed parallel to the substrate) can be formed as a consequence of the anisotropic strain appearing during their formation by epitaxial growth [15]. ...
... Furthermore, they exhibit efficient optical emission at even 1.6 mm [12]. These QWRs are formed instead of quantum dots due to a strong stress anisotropy built up during the growth process, due to the distortion of As-In bonds along [110] direction and As-As dimerization along [110] [15], which leads to a stress relaxation that takes place mainly along the [110] direction, resulting in elongated nanostructures along the perpendicular one (i.e., [110]). ...
... Self-assembly, through the SK growth mode, is recognized as one of the most remarkable method to optimize the 3D growth of heteroepitaxial InAs/InP QWRs. In this case, nanostructures spontaneously undergo a transition to 3D growth when the accumulated biaxial stress exceeds a given value and after a critical thickness is attained [15]. The epitaxial growth in heterostructures, with reticular misfit of 3.2 %, led the coupled structure to be relaxed by providing the surface energy needed to create these nano-objects. ...
With the aim to improve the design and potential use of 1D-nanostructures as functional materials, structural and compositional changes of these nanostructures must be correlated with the growth conditions. The process temperature, deposition time, diffusivity of participant specimens, the buffer or intermediate layer characteristics grown previously to the nanomotives, etc., can influence strongly on the final behavior of these systems. In this sense, techniques such as conventional and high-resolution electron microscopy imaging and energy dispersive
X-ray and electron energy loss spectroscopies are valuable tools for the analysis of their morphology, crystalline structure, and composition.
In this work, we apply these techniques to examine a number of 1D-nanostructures formed by ZnO, AlGaxN1�x or InAsxP1-x. In the case of chemical bath pre-deposited ZnO/Si nanorods, we analyze the role of the buffer layer and how it influences the vapor phase transport deposition process subsequently carried out. We also etermine the compositional and morphological changes of ZnO nanowires’ manufacture by the thermal oxidation of Zn films on CdTe substrate.On the other hand, we study the defect-free core–shell structures spontaneously formed in AlxGa1�xN nanowires grown by plasma-assisted molecular beam epitaxy on GaN substrate and the complex chemical composition gradient along the wire axis for both core and shell blocks. Finally, we analyze the nucleation,the compositional and morphology evolution of single and stacked layers of
self-assembled InAs/InP quantum wires formed through the Stranski-Krastanov transition by molecular beam epitaxy. For this work, the determination of high-resolution compositional maps extracted by aberration – orrected Z-contrast imaging combined with strain field map obtained by finite element methods have been especially successful.
