In situ measurements of As/P exchange during InAs/InP(0 0 1) quantum wires growth

The size and emission wavelength of self-assembled InAs/ InP͑001͒ quantum wires (QWrs) is affected by the P / As exchange process. In this work, we demonstrate by in situ stress measurements that P / As exchange at the InAs/ InP interface depends on the surface reconstruction of the InAs starting surface and its immediate evolution when the arsenic cell is closed. Accordingly, the amount of InP grown on InAs by P / As exchange increases with substrate temperature in a steplike way. These results allow us to engineer the size of the QWr for emission at 1. InAs/ InP quantum wires (QWrs) or quantum dots (QDs) are very promising for optoelectronic devices operating in the 1.30-1.55 m wavelength range employed in fiber optical telecommunications systems. 1-3 However, these nanostructures have not been as extensively exploited as InAs/ GaAs QD partly due to the intrinsic peculiarities of the interface between III-V A / III-V B compounds: On one hand, an asymmetric stress appears at the InAs/ InP interface; 4 on the other hand, complicated V A /V B exchange processes take place during epitaxial growth. Although P / As exchange has been widely studied, 5-15 a deeper understanding of this process is still needed since it determines the size of the nanostructures, governing their emission wavelength.In this work, we have carried out in situ and in real-time stress measurements on an InAs(001) surface exposed to a P 2 flux at different substrate temperatures T s in order to quantify the P / As exchange effect. We have observed a strong dependence on T s of the amount of InP formed (that is, grown without In delivering) during P 2 exposure in the 330-515°C range. We prove that this behavior is directly related to the InAs surface reconstruction, which depends on temperature for a fixed arsenic pressure. Finally, we show that such InAs/ InP surface dynamics can be used to engineer the size distribution of the nanostructures ensemble. To demonstrate this capability, we have studied the optical properties of the InAs QWr as a function of the growth temperature of the InP cap layer, T C . A redshift of the photoluminescence (PL) spectrum is observed as the T C decreases. The size reduction responsible for the observed change in emission energy has been directly demonstrated by transmission electron microscopy (TEM).The P / As exchange process has been assessed by in situ and in real-time measurements of the stress evolution during P 2 exposure of an InAs (001) surface through the optical monitorization of the substrate curvature. This in situ technique provides a direct measurement of the film accumulated stress, ⌺ (stress integrated along the layer thickness). 4 We have used thinned ͑175 m͒ InAs substrates, elongated along the [110] axis to detect stress variations in this direction, in which the whole amount of incorporated InP can be quantified. 5 The phosphorous and arsenic beam equivalent pressures (BEPs) used in these experiments were BEP͑P 2 ͒ = 2.37ϫ 10 −5 mbar and BEP͑As 4 ͒ = 0.53ϫ 10 −5 mbar. Simultaneous reflectance...

mentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Size control of InAs∕InP(001) quantum wires by tailoring P∕As exchange

Fuster¹,

González²,

González³

et al. 2004

“…So, stress driven processes must play an important role in the formation of excess of InAs. [13][14][15][16][17] Accordingly, we propose that InAs grows faster at the low strained areas of the surface, process that involves In migration and As/P exchange. In order to prove this assumption, we have measured ⌺ evolution in similar experiments but exposing the InP surface only to As 4 flux at 515°C.…”

mentioning

confidence: 99%

Stacking of InAs/InP(001) quantum wires studied by in situ stress measurements: Role of inhomogeneous stress fields

Fuster¹,

González²,

González³

et al. 2004

Size and spatial distribution homogeneity of nanostructures is greatly improved by making stacks of nanostructures separated by thin spacers. In this work, we present in situ and in real time stress measurements and reflection high-energy electron diffraction observations and ex situ transmission electron microscopy ͑TEM͒ characterization of stacked layers of InAs quantum wires ͑QWRs͒ separated by InP spacer layers, d(InP), of thickness between 3 and 20 nm. For d(InP)Ͻ20 nm, the amount of InAs involved in the created QWR from the second stack layer on, exceeds that provided by the In cell. Our results suggest that in those cases InAs three dimensional islands formation starts at the P/As switching and lasts during further InAs deposition. We propose an explanation for this process that is strongly supported on TEM observations. The results obtained in this work imply that concepts like the existence of a critical thickness for two-to three-dimensional growth mode transition should be revised in correlated QWR stacks of layers. © 2004 American Institute of Physics. ͓DOI: 10.1063/1.1759374͔The incorporation of nanostructures in electronic and optoelectronic devices provides properties improvement and design possibilities, yet requiring the control of their size and position. For self-assembled nanostructures, the size homogeneity and spatial distribution can be greatly improved by stacking several layers. [1][2][3][4] In this context, the vertical stacks of self-assembled InAs quantum wires ͑QWRs͒ grown on InP͑001͒ are of particular interest for lasers, as they emit light at 1.30 and 1.55 m. [5][6][7] In stacks of nanostructures, the buried ones produce inhomogeneous strain fields that propagate toward the capping layer surface where the next nanostructures will be formed. 1,2 This leads to a vertical correlation between the nanostructures layers depending both on the size of the buried nanostructures and the spacer layer thickness.In this work, we have studied the growth of multilayers of InAs QWRs with different spacer layer thicknesses, d(InP), by in situ and in real time stress measurements and reflection high-energy electron diffraction ͑RHEED͒. A strong influence of the spacer layer thickness on QWR formation process in the second and successive layers of the stack has been observed: QWRs are correlated for d(InP)Ͻ20 nm, where a reduction of the amount of deposited InAs necessary for QWR formation ͑observed by RHEED͒ together with an increase in InAs growth rate ͑ob-tained from stress measurements͒ take place. In order to explain these results, we propose a model that is strongly supported by transmission electron microscopy ͑TEM͒ images. Our experiments provide quantitative values of the extra amount of material involved in the formation of vertically correlated self-assembled nanostructures, as well as the evidence of the absence of a two-dimensional ͑2D͒-threedimensional ͑3D͒ growth mode transition as the relevant process in nanostructures self-assembling during stacking.The samples under study consist of...

“…Before opening the In cell, the P and As cell are switched producing an InAs wetting layer ͓thicknessϷ 1 monolayer͑ML͔͒ due to As/P exchange on the surface at a substrate temperature T S = 515°C. 11 The InAs layer deposition was done at T S = 525°C through a pulsed mode where In and As cells are alternated. In each growth cycle, the In cell was opened during 1 s at an equivalent InAs growth rate of 0.1 ML/s, and the As cell during 0.6 s at a BEP͑As 4 ͒ of 6.4ϫ 10 −6 mbar.…”

mentioning

confidence: 99%

Direct formation of InAs quantum dots grown on InP (001) by solid-source molecular beam epitaxy

Fuster¹,

Rivera²,

Alén³

et al. 2009

We have developed a growth process that leads to the direct formation of self-assembled InAs quantum dots on InP͑001͒ by solid-source molecular beam epitaxy avoiding the previous formation of quantum wires usually obtained by this technique. The process consists of a periodically alternated deposition of In and As correlated with InAs͑4 ϫ 2͒ ↔ ͑2 ϫ 4͒ surface reconstruction changes. Based on the results obtained by in situ characterization techniques, we propose that the quantum dots formation is possible due to the nucleation of In droplets over the InAs͑4 ϫ 2͒ surface during the In deposition step and their subsequent crystallization under the As step. © 2009 American Institute of Physics. ͓DOI: 10.1063/1.3108087͔ Semiconductor self-assembled quantum dots ͑QDs͒ have received much attention because of their application in advanced optical devices. 1,2 In particular, the active elements based on the InAs/InP heteroepitaxial system are interesting for their use in laser devices emitting at 1.55 m, compatible with optical fiber communications. 3 The self-assembling of QDs is a strain-driven process that takes place in highly lattice-mismatched semiconductor materials. The strained material grows in a layer-by-layer mode up to a certain critical thickness and then the growth mode switches from two dimensional ͑2D͒ to three dimensional ͑3D͒ with the net result of elastic energy relaxation ͑Stranski-Krastanov process͒. The InAs/GaAs heteroepitaxial system is a well studied example of this kind of process. 4 However, in the InAs/ InP͑001͒ heteroepitaxial system growing under usual molecular beam epitaxy ͑MBE͒ conditions, quantum wires ͑QWRs͒ elongated along the ͓110͔ direction are formed instead of QDs, which are in principle more efficient for elastic energy relaxation. Other authors have recently recognized the key role of the InAs As-rich ͑2 ϫ 4͒ surface reconstruction in a growth model for QWR formation. 5 Other works report the InAs/InP͑001͒ QD formation using MBE but always through the evolution of previously formed QWRs: either by the ripening of the initial QWRs during substrate cooling down under arsenic overpressure 6 or after substrate annealing under no arsenic flux at the InAs͑4 ϫ 2͒ surface reconstruction. 7 In a previous model we proposed that the growth of InAs on InP under a ͑2 ϫ 4͒ V element ended surface reconstruction produces a stress asymmetry at the InAs/InP interface, inducing an asymmetric relaxation process that finally results in QWR formation. 8,9 According to our previous model, the built-in interface strain anisotropy could be inhibited if the surface would be terminated in III element. 9 Therefore, if those conditions could be experimentally achieved, for example, growing InAs on the In-rich ͑4 ϫ 2͒ surface reconstruction, QDs instead of QWRs would be directly obtained in the InAs/InP͑001͒ system.In this work we have developed a method to implement this idea. The method consists of depositing InAs in a pulsed mode under experimental conditions in which the surface shows most of the growt...