Summary - InAs/GaAs quantum dots growth and characterization

Chapter 3: InAs/GaAs quantum dots growth and characterization

3.4 Summary

In summary, the growth and characterization of InAs QDs have been investigated extensively. The results of AFM and PL of grown QDs are used to study the growth condition dependence. The growth parameters, such as GaAs capped growth rate, As2/As4 ratio, and arsenic beam flux, have been studied and discussed extensively. We found that the diffusion of group III atoms is dominating factor in the QDs growth.

Arsenic molecular species would influence group III adatoms’ migration behavior on the surface. Besides, the capped growth rate has no or little effect on the QDs growth if the alloying effect dominates. We also study the growth and decline of QDs after S-K transition by a sequence of identical experiments.

References

[3.1] D. C. Liu and C.P. Lee, Appl. Phys. Lett. 63, 3503(1993)

[3.2] L. Goldstein, F. Glas, J. Y. Marzin, M. N. Charasse and G. Le Roux, Appl. Phys.

Lett. 47, 1099 (1985).

[3.3] J. Y. Tsao, Materals fundamentals of molecular beam epitaxy, Academic press, Inc. (1993)

[3.4]. W.K. Liu and M.B. Santos, Thin films: heteroepitaxial systems, World Scientific, (1999)

[3.5] Mitsuru Sugawara, Self-assembled InGaAs/GaAs quantum dots, Academic press, Tokyo, Japan (1999).

[3.6].W. Seifert, N. Carlsson, M. Miller, M. Pistol, L. Samuelson, and L. R. Wallenberg, Prog. Crystal Growth and Charact. 33, 423 (1996)

[3.7].S. D Lin, Doctoral dissertation, NCTU, Tawain(2002)

[3.8] G. S. Solomon, J. A. Trezza, and J. S. Harris, Jr., Appl. Phys. Lett. 66, 991 (1995).

[3.9] G. S. Solomon, J. A. Trezza, and J. S. Harris, Jr., Appl. Phys. Lett. 66, 3161 (1995).

[3.10] L. Chu, A. Arzberger, G. Böhm and G. Abstreiter, J. Appl. Phys. 85, 2355 (1999).

[3.11] M. G. Alessi, M. Capizzi, A. S. Bhatti, A. Frova, F. Martelli, P. Frigeri, A.

Bosacchi, and S. Franchi, Phys. Rev. B 59, 7620 (1999).

[3.12]J. M. Garcia, G. Medeiros-Ribeiro, K. Schmidt, T. Ngo, J. L. Feng, A. Lorke, J.

Kotthaus, and P. M. Petroff, Appl. Phys. Lett. 71, 2014 (1997).

[3.13] E. H. C. Parker, The technology and physics of molecular beam epitaxy, London, England (1985).

[3.14] J. M. Garcia, J. P. Silveira, and F. Briones, Appl. Phys. Lett. 77, 409 (1997) [3.15]Ferdos, Fariba; Wang, Shumin; Wei, Yongqiang; Sadeghi, Mahdad; Zhao,

Qingxiang; Larsson, Anders, J. Crystal. Growth, 251,145(2003)

[3.16] Heyn, Ch.and Hansen, W, J. Crystal. Growth, 251,140(2003)

Table 3.1 InAs QDs growth parameters

Fig.3.1 Fabrication procedure of ‘thermal etching’ quantum dots

GaAs capping layer deposition InAs 3-D island formation InAs 2-D layer deposition GaAs buffer layer growth Start from (100) GaAs substrate

Fig.3.2 Growth procedure for InAs/GaAs QDs in S-K mode.

A B

FM V W SK

Fig.3.3 Three growth modes in heterostructure epitaxy, Frank-van der Merwe (F-M), Volmer-Weber (VW), and Stranski-Krastanow (S-K).

InAs QDs

20nm GaAs 30nm Al

_0.3

Ga

_0.7

As

150nm GaAs InAs QDs 150nm GaAs 30nm Al

_0.3

Ga

_0.7

As

~250nm buffer GaAs (100) GaAs substrate

Fig.3.4 A typical sample structure of InAs QDs growth, the top InAs QDs are used in AFM measurement.

Fig.3.5 AFM images of uncapped InAs QDs.

1.00 1.05 1.10 1.15 1.20 1.25 0

5 10 15 20 25

PL intensity(a.u.)

Energy(eV)

200uW 1mW 5mW 14mW

Fig.3.6 The PL spectrum of grown InAs/GaAs QDs measured under different excitation powers

Cracker zone=840℃ Cracker zone=730℃

Cracker zone=680℃ Cracker zone=570℃

Fig.3.7 The surface morphology of uncapped InAs QDs using a mixture of As2/As4

vapor (on condition that a high V/III ratio is used during InAs deposition)

Cracker zone=840℃ Cracker zone=730℃

Cracker zone=680℃ Cracker zone=570℃

Fig.3.8 The surface morphology of uncapped InAs QDs using a mixture of As2/As4

vapor (on condition that a low V/III ratio is used during InAs deposition)

1.05 1.10 1.15 1.20 1.25 1.30 1.35 1.40

0 (a) with a high V/III ratio , (b) with a low V/III ratio during InAs deposition

1.05 1.10 1.15 1.20 1.25 1.30 1.35 1.40

Fig. 3.10 The PL spectrum of grown InAs QDs under different capped growth rate (using a mixture of As2/As4 vapor). (a) 1µm/hr, (b)0.3µm/hr

0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35 0.0

0.2 0.4 0.6 0.8 1.0

PL intensity(a.u.)

Energy (eV)

3^rd round 2^nd round

1^st round

Fig. 3.11 The PL spectrum of grown InAs QDs (without growth interruption)

1.00 1.05 1.10 1.15 1.20 1.25 0.0

0.2 0.4 0.6 0.8 1.0

PL intensity (a.u.)

Energy (eV)

Fig. 3.12 The PL spectrum of grown InAs QDs (with growth interruption)

Part A.

Ordered Quantum Dots Lattice Growth

Chapter 4 Selective growth of InAs quantum dots array on patterned GaAs(001) substrates

In this chapter, we study the MBE selective growth of InAs QDs on GaAs(100) substrates patterned by e-beam lithography and wet chemical etching. The purpose of QDs selective growth and a brief review are given. Then, the experiments about substrate patterning and MBE growth are presented. Various samples characterization and discussion are summarized finally.

4.1 Introduction

4.1.1Purpose to use selective growth on self-assembled QDs

InAs self-assembled QDs grown by S-K mode have been intensively studied in the last several years due to their fabrication-easy and defect-free properties. The high crystalline quality found in QDs formed by these processes enables incorporation of the dots in optical and electronic devices. However, the stochastic nature of the formation of these QDs driving by the elastic relaxation of the strain leads to undesirable nonuniformity in their sizes as well as their spatial distribution. The size fluctuation results in large inhomogeneous broadening in their energy spectrum. Besides, it is almost impossible to place the QDs in a long-range-ordering manner or on the designated locations. These seriously limit the potential device applications of such QDs. As for size fluctuation, recent studies concentrated on the formation of high

density and uniform sized QDs by optimizing the growth conditions. On the other hand, there has been increasing interest in the study of self-assembled QD selective formation on the grown surfaces, designed to improve position control.

4.1.2 Overview of selective growth methods

The common selective growth methods of self-assembled QDs can be classified into two categories: (a) patterned substrate methods and (b) strain engineering methods.

• Patterned substrate methods [4.1-4.9]

The method is based on the differences in migration behaviors or sticking coefficient of group III adatoms on the faceted surface: D. S. L. Mui et al.made use of differences in atom diffusion on faceted surfaces to control QDs formation. M.

Kitamura et al. grew QDs on the step edges of multi-atomic steps that result from the step bunching. R. Zhang et al. utilized selective area epitaxy and oxide-patterned substrate to position QDs on GaAs facet. C.Y. Hyon et al. selectively positioned QDs on GaAs patterning by AFM nanolithography. T. Fukui et al. demonstrated position- and number-controlled QDs via selective area epitaxy and SiNx (or SiO2)-patterned substrate. B.H. Choi et al. demonstrated almost one-dimensional alignment of QDs on Si substrate. S.C. Lee et al. fabricated a one-dimensional row of QDs using Orientation-dependent migration and incorporation of In atoms from (111)A to (100) facets. X. Mei et al. utilized anodic Al2O3 nanohole array template mask to fabricate QDs array.

• Strain engineering methods [4.10-4.13]

This technique makes use of introducing a build-in or artificial strain anisotropy

on the grown surface to control QDs formation: H. Lee et al.combined lithography with in situ lateral strain engineering to control QDs formed on mesoscopic surface. T. Mano et al. demonstrated a one-dimensional alignment of QDs by lateral surface strain field modulation generated by the underlying InGaAs superlattice template

4.2 Experiments

4.2.1 Electron-beam patterning

Two chessboard-like patterns with sizes of ~ 80µm × 80 µm were first defined on a flat (001) GaAs substrate via e-beam nanolithography and the wet chemical etching.

Figure 4.1(a) shows a schematic diagram of the grid pattern defined on the GaAs substrate. The two chessboard-like patterns were composed of cross stripes oriented at angles of 0◦, 90◦ (pattern A) and 45◦, 135◦ (pattern B) with respect to the [110] direction.

On both pattern A and pattern B, those have a pitch size of 0.1 µm. The GaAs substrate was first covered with 2% PMMA, and then exposed to an e-beam to define the patterns.

After PMMA development and wet chemical etching in a H2SO4:H2O2:H2O=1:8:80 solution, the e-beam-defined patterns were transferred onto the GaAs substrate. Figure 4.1(b) shows the scanning electron microscope (SEM) images of the patterns after the wet chemical etching. The anisotropic etching rate for different faceted surface results in pits with area of about 150 nm ×100 nm and depth about 15 nm [4.14]. The pits on pattern A have their edges parallel to [110], [1-10]. For pits on pattern B, their edges are parallel to [100], [010]. The pits were rectangular in shape since the etching processes were anisotropic on the two perpendicular facets inside the pits. This resulted in different thicknesses of the walls surrounding the pits. For example, on pattern A, the

4.2.2 MBE regrowth

Before introducing the patterned substrate into the MBE system, the wafer was cleaned in solvents and a thin layer of GaAs was removed from the top using wet chemical etching using H3PO4:H2O2:H2O = 3:1:50. The purpose of this procedure is to remove the contaminations that introduced from PMMA development and successive process. The MBE growth process started with oxide desorption under an As2 flux at 610^oC.After the oxide layer was removed, a 100Å GaAs buffer layer was deposited at 610^oC to recover the etched surface. Then the substrate temperature was lowered to 520^oC to deposit 2MLs of InAs at a growth rate of about 0.056µm/hr. Finally, 500Å GaAs capped layer was grown at the same temperature and then cooled down under As2

flux. In order to investigate the effect of patterned substrate on QDs growth, we deposited 2ML of InAs, which is above the critical thickness of S-K transition. The QDs were formed on both patterned (A and B) and non-patterned regions, under the same growth conditions. The same procedure was repeated for AFM sample, except that after the QDs growth, the growth was stopped and then cooled down under As2 flux immediately.

4.3 Results and discussion

4.3.1 TEM and AFM characterization

The transmission electron microscope (TEM) image of the QDs formed on the non-patterned area is shown in the upper part of Fig.4.2. We have also shown the AFM images taken from the centre of pattern A in Fig.4.3. The AFM image in Fig.4.3 looks slightly different from the SEM image of the e-beam defined pattern shown in Fig.4.1.

The reason is that the wet chemical etching process operative before introducing the patterned sample into the MBE chamber has effected further thinning of the stripes parallel to [110] on pattern A. Therefore, the image of the stripes parallel to [110] is not evident in Fig.4.3. Nevertheless, the inset in Fig.4.3 shows that only a single row of dots was formed on stripes parallel to [1-10]. Those stripes have top widths of about 50 nm.

The dots that appeared on the stripes have an average base width of about 30 nm and are more uniform in size than dots formed on the non-patterned area. It is worth to note that no dot was found on a stripe parallel to [110] due to its much narrower top width. The AFM image taken from the inside of the pits on pattern A is also given in Fig.4.4. The inset in Fig.4.4 indicates that there is also a single row of dots formed inside the pits.

The position of those dots that appeared inside the pits looks asymmetric with respect to the centre of the pits. The TEM image of QDs on pattern A as shown in Fig.4.5 is also consistent with results from the AFM images. In contrast to the results on pattern A, the TEM image of the dots on pattern B (as shown in Fig.4.6) shows that the dots were only grown inside the pits and no dots can be found on the top of mesas. This is due to the undercut edges of the etched sidewalls of the pits, which have prevented the dots from being grown on the top of the stripes. The estimated density of the QDs formed on pattern A is about 3 × 10¹⁰ cm−² and is much higher than the densities of those on pattern B (about 4 × 10⁹ cm−²) and the non-patterned area (about the same as for pattern B).

4.3.2 Photoluminescence studies

We have studied and compared the PL spectra from the QDs grown on pattern A, pattern B and the non-patterned area after the samples were capped with 50 nm GaAs.

region of the sample: pattern A, the non-patterned area and pattern B. The QDs grown on pattern A emitted the strongest PL intensity among the three (about four times stronger than that from the other areas). We attribute this to the improvement of the dot quality and the higher density of QDs formed in this area. It also indicated that our patterned surface has recovered during the regrowth from the processing damage. The peak of the luminescence also showed a significant blue shifting (about 25 meV) compared to the luminescence signal from the dots on the non-patterned area. This indicates that quantum dots formed on pattern A are different in size and composition compared to the dots on pattern B and on the non-patterned area. For QDs grown on pattern B, the luminescence intensity and peak position are not very different from those for dots on the non-patterned area.

4.4 Summary

In summary, we have grown QDs via molecular beam epitaxy on patterned GaAs(001) substrates prepared by e-beam lithography and wet chemical etching. TEM and AFM images show ordering of QDs formed on the stripes and inside the pits. From PL studies, we found that the QDs grown on pattern A gave the strongest luminescence intensity among the three areas that we have investigated. We attribute this to the higher density, ordering and improvement in quality of the dots.

References

[4.1] D. S. L. Mui, D. Leonard, L.A. Coldren, and P. M. Petroff, Appl. Phys. Lett. 66, 1620 (1995).

[4.2] M. Kitamura, M. Nishioka, J. Oshinowo, and Y. Arakawa, Appl. Phys. Lett. 66, 3663 (1995).

[4.3] R. Zhang, R. Tsui, K. Shiralagi, D. Convey, and H. Goronkin, Appl. Phys. Lett. 73,

505 (1998).

[4.4] C. K. Hyon, S. C. Choi, S.-H. Song S. W. Hwang, M. H. Son, D. Ahn, Y. J. Park and E. K. Kim, Appl. Phys. Lett. 77, 2607 (2000).

[4.5] Cheol-Koo Hahn, Junichi Motohisa, and Takashi Fukui, Appl. Phys. Lett. 76, 3947 (2000).

[4.6] Hyo Jin Kim, Junichi Motohisa, and Takashi Fukui Appl. Phys. Lett. 81, 5147 (2002).

[4.7] B. H. Choi, C. M. Park, S.-H. Song, M. H. Son, S. W. Hwang, D. Ahn, and E. K.

Kim, Appl. Phys. Lett,78, 1403 (2001).

[4.8] S. C. Lee, L. R. Dawson, K. J. Malloy, and S. R. J. Brueck, Appl. Phys. Lett. 79,

Phys. Lett, 81, 1705 (2002).

[4.12] Kwang Moo Kim, Young Ju Park, Young Min Park, Chan Kyeong Hyon, Eun

Kyu Kim, and Jung Ho Park, J. Appl. Phys. 92, 5453 (2002).

[4.13] T. Mano, R. No¨ tzel, G. J. Hamhuis, T. J. Eijkemans, and J. H. Wolter, J. Appl.

Phys. 95, 109 (2004).

[4.14] R. E. Williams, Gallium arsenide processing techniques (1984)

[1-1 0]

[1 1 0]

Pattern A Pattern B

Fig.4.1 (a) Two types of square grids patterns with pitch of 0.1µm, (b) SEM images of the patterns after the wet chemical etching

Fig.4.2 The TEM image of the uncapped QDs on the non-patterned area (in the upper half of the figure) and pattern B (in the lower half part).

Fig.4.3 The AFM images of QDs grown on the pattern A. The inset shows a single row of dots that formed on the mesa stripes

Fig.4.4 The AFM images of QDs inside the pits of pattern A. The inset also shows one single row of dots sited against one of the sidewalls.

Fig.4.5. The TEM image shows that dots formed on the ridge and valley, which is consistent with the AFM image

Fig.4.6 The TEM image shows that dots formed inside the pits only and no dot landed on the mesa top

Fig.4.7 The PL spectra of the QDs sample are taken from three different areas: pattern A, non-patterned area, and pattern B.

Chapter 5 Selective growth of single InAs quantum dots

In this chapter, a method to achieve long-range ordering and selective positioning of single InAs self-assembled QDs has been developed. The selective growth was achieved by manipulating the strain distribution on the sample surface. We have demonstrated single QD formation on small e-beam defined mesas. Controlled single dot arrays have also been achieved by using this technique.

5.1 Introduction

5.1.1 Single QDs fabrication

The self-assembled quantum dots (QDs) grown by S-K mode are promising candidates for quantum devices because of ease of fabrication, and their high-quality, defect-free properties. However, as mentioned in the previous chapter, the ensemble of QDs grown using the S-K growth mode are usually randomly distributed on the grown surface and suffer from strain fluctuations in a random manner. It is almost impossible to place the QDs in a long-range ordered way or on the predesigned locations. These seriously limit the potential device applications of such QDs. Especially for single QDs;

they should find applications in many new generation quantum devices, such as single electron transistors, single photon source, and single-photon photodetectors

Recently, in order to solve these problems, there have been a great amount of efforts in the study of self-assembled QDs formation on patterned substrates to improve size uniformity and position controllability. However, as for single QDs fabrication,

fewer studieshave been reported, K. Asakawa et al. have demonstrated site control of individual QDs by in-situ nanolithography combined with MBE using an ultrahigh-vacuum multi-chamber system [5.1-5.2]. But the main problems they suffer from are direct QD-growth on etched surface, unknown materials introduced during the patterning process, and complex multi-chamber system.

For the sake of avoiding the above-mentioned problems, we develop a selective growth method, which is capable of placing a single InAs QD on a given location on GaAs substrate based on strain accumulation (or strain engineering).

5.1.2 Technique foundation

The basic principle behind this technique originates from: (1) self-assembled QDs are formed because of strain relaxation during the dot formation resulting in a decrease in the total strain energy of the epilayer/substrate system, and (2) if a local strain is artificially introduced in a given region, the 2D-to-3D morphology change will take place before the critical thickness (θc:1.4 ~ 1.7 ML) is reached for InAs QD formation.

In this method, the selective growth of QDs is achieved by artificially introducing additional strain energy in certain pre-designed locations defined by e-beam lithography. QDs are therefore grown only on the selected locations while in other regions only 2D growth takes place. Using this technique, the QDs are formed on pre-designed mesas with added strain and two-dimensional single QD arrays would be realized.

5.2 Experiments

Prior to selective growth on patterned substrate, several QDs growth tests should be performed on the flat GaAs substrate in advance. As shown in Fig.5.1, we need to

find out the minimum amount of InAs deposited on these structures. The experimental results have shown that QDs are formed by depositing 1.35MLs of InAs on structure A and B, while there is no QD found on structure C and D. Based on these results, we can control the formation of QDs below the critical thickness at any given region with added strain.

5.2.1 Pregrowth and e-beam patterning

Fig.5.2 illustrates the sequence of our selective growth technique. First, in order to obtain a local strain region, we deposited a 100 Å In0.2Ga0.8As layer and 200 Å GaAs capping layer on an epi-ready GaAs (001) substrate by MBE. The substrate was then coated with 2 % PMMA, and exposed to an e-beam to define a square lattice of mesas.

The linear dimension of each mesa was 200 nm × 200 nm, and mesas were separated by 500 nm. After the PMMA development and wet chemical-etching in an H2SO4: H2O2: H2O = 1: 8: 80 solution, e-beam defined patterns were then transferred onto the GaAs substrate. The etched mesas have a height of about 500 Å.

5.2.2 MBE regrowth

Before introduced into the MBE system, the patterned GaAs (001) substrate was cleaned in solvent and then about 100 Å GaAs was removed by wet chemical-etching using a H3PO4 : H2O2 : H2O solution. The MBE regrowth started with oxide desorption under an As2 flux at 610^oC. After the oxide layer was desorbed, the substrate temperature was lowered to 510^oC to deposit an 80 Å In0.1Ga0.9As buffer layer, which was used as a strain-fine-tuning layer on the surface. This layer increases the strain energy to ensure the subsequent QD formation on the mesas. Finally, 1.35 ML of InAs was deposited at a growth rate of about 0.056µm/hr. The sample was cooled down

under an As2 flux immediately afterwards. While 1.35 ML of InAs is not enough to cause QD formation on regular GaAs surface, the added strain from this InAs deposition on the pre-strained mesas is adequate to cause QD growth.

5.3 Sample characterization and results

Fig. 5.3(a) shows the SEM image of the mesa lattice after MBE regrowth. It shows that the mesas in the lattice are elongated along the [1⎯1 0] direction resulting in a rectangular shape and giving a new base dimension on the order of about 250 nm × 150 nm. This anisotropic growth is caused by different diffusion rates of adatoms along the two different <110> directions. The sample was also studied by a Digital Instruments D5000 AFM system using the tapping mode. The AFM image is shown in Fig. 5.3(b). It shows that QDs were formed on the top of the mesas. No dots were found on the non-patterned region; only surface roughness that resulted from etching process was observed. Fig.5.3(c) shows the surface profile across a mesa and a single QD. The dots landed on the mesas have an average base width of ~ 500 Å and an average height

在文檔中規則化量子點晶格及半導體量子環之研究 (頁 47-121)