Formation of Nanopits in Si Capping Layers on SiGe Quantum Dots
© Cui et al. 2010
Received: 10 August 2010
Accepted: 15 September 2010
Published: 2 October 2010
In-situ annealing at a high temperature of 640°C was performed for a low temperature grown Si capping layer, which was grown at 300°C on SiGe self-assembled quantum dots with a thickness of 50 nm. Square nanopits, with a depth of about 8 nm and boundaries along 〈110〉, are formed in the Si capping layer after annealing. Cross-sectional transmission electron microscopy observation shows that each nanopit is located right over one dot with one to one correspondence. The detailed migration of Si atoms for the nanopit formation is revealed by in-situ annealing at a low temperature of 540°C. The final well-defined profiles of the nanopits indicate that both strain energy and surface energy play roles during the nanopit formation, and the nanopits are stable at 640°C. A subsequent growth of Ge on the nanopit-patterned surface results in the formation of SiGe quantum dot molecules around the nanopits.
Heteroepitaxy has been a powerful method to fabricate functional quantum structures, e. g. quantum wells , quantum dots (QDs)  and quantum rings (QRs) [3, 4]. On the one hand, strain is the most important factor affecting the formation of nanostructures [1–3] and even their capping layers in heteroepitaxy . The evolution of the strain can result in a variety of nanostructures, such as QRs . On the other hand, strain induced by heteroepitaxy has been a very prominent improvement in technology to increase carrier mobility . Recently, strained Si channel induced by SiGe QDs has been proposed to enhance hole mobility in field effect transistors . Thus strain, together with its distribution and evolution is a key to understand the growth mechanism of the quantum structures and realize the desired structures.
Nanopits are interesting for their use as template to achieve positioning growth of QDs, [8–10] QRs  and QD molecules . In 1998, Deng and Krishnamurthy  fabricated nanopits by depositing carbon impurity in Si matrix, in which SiGe QD molecules were grown around each nanopit. In recent years, nanopits fabricated by electron beam lithography, [8, 13] holography lithography  and nanosphere lithography [10, 11] have been used to fabricated ordered SiGe QDs or QRs. Recently, nanopits were also observed in III–V material system by self-assembling based on droplet epitaxy growth .
QD molecules are promising candidates as building blocks for the quantum computing [15–17]. It is highly desired to grow QD molecules on semiconductors for the possible applications in quantum computing. By self-assembly, in III–V system, GaAs , and InAs [19, 20] QD pairs were grown by droplet epitaxy. In SiGe system, though the QD molecules has been observed by introducing carbon impurities , the growth of defect-free QD molecules needs to be further explored.
In this paper, a strain governed process, as well as related nanostructures in Si/Ge system are reported. Square nanopits, with a depth of about 8 nm and boundaries along 〈110〉, were formed in the low temperature grown Si capping layer by thermal annealing at 640°C. The formation mechanism is proposed based on strain energy relaxation and surface energy minimization. On the nanopit-patterned surface, by depositing Ge SiGe QD molecules were fabricated.
The sample growth was carried out in a molecular beam epitaxy (Riber Eva-32) system with two electron beam evaporators of Ge and Si sources. P-type Si(001) substrates with a resistivity of 1–10 Ω cm were used to grow QDs. The substrates were chemically cleaned by Shiraki  method before put into load-lock chamber. After transferred into growth chamber, the substrates were heated to 980°C for 10 min to remove the protecting oxide, then clear 2 × 1 reconstruction pattern could be observed by reflection high energy electron diffraction. The chamber pressure was below 1 × 10-9 Torr. Then a 50 nm thick Si buffer layer was deposited at 650°C with a growth rate of 0.36 Å/s. At 640°C, by two-step growth method , two layers of Ge with a total thickness of 1.0 nm were deposited in sequence to form uniform QDs. The deposition rate was 0.11 Å/s. Then the sample was cooled down to 300°C to grow Si capping layers at a growth rate of 0.54 Å/s. The in-situ annealing process was performed at a high temperature of 640°C. After growth or annealing processes, the temperature was decreased to room temperature immediately. In order to unveil the detailed kinetics for the nanopit formation, in-situ annealing at a low temperature of 540°C was also carried out. Atomic force microscopy (AFM, Solver P47-SPM-MDT) was used to measure surface morphology. Cross-sectional transmission electron microscopy (TEM) was used to characterize microstructures.
Results and discussion
Strain energy relaxation and surface energy minimization are considered as the two dominating factors for the nanopit formation. Firstly, atomic intermixing or interdiffusion between Si and Ge, and Ge surface segregation could be excluded during the process for the nanopit formation because the mounds consist purely of Si atoms and the capping layer is thick enough for ruling out Ge surface segregation (50 nm). The diffusion of Ge atoms in Si at the temperature lower than 650°C can be neglected. The whole process for the nanopit formation is only related to Si atomic migration.
In the growth of ordered SiGe QDs, lithography technologies [8–10] are frequently used to fabricated nanopits on Si substrate. The grown QDs located in the nanopits, rather than around them, which is contrary to our results. The nucleation modes on the nanopit-patterned substrates have been observed and discussed based on the growth temperature . They found that at a low growth temperature of 550°C, nucleation at the nanopits is in metastable phase. At high growth temperature of 750°C, QDs are stable at the terrace between nanopits. At a intermediate temperature of 650°C, random nucleation is observed. They attribute the variation of nucleation modes to the kinetic limitation at high growth temperature. In another theoretical paper , only random and ordered (in the nanopits) nucleation modes are predicted. In our experiments, the growth temperature is 650°C. Nearly all the QDs located around nanopits. However, by growing buffer layers with increasing growth temperature from 500°C to 640°C on nanopit-patterned substrates, QDs nucleate in the nanopits by growing at 640°C . It can be deduced that different pre-deposited surface morphologies in the two cases result in different nucleation modes. It may be the key point that gradually increasing the growth temperature during the growth of Si buffer layers can effectively decreases the depth of the nanopits as well as the aspect ratio. Nevertheless, the morphological detail of nanopits plays important roles in determining the nucleation whether in or out of nanopits along with growth temperature.
In summary, nanopits are obtained by in-situ thermal annealing of low temperature Si capping layer on QDs. The formation of the nanopits is discussed based on the strain energy relaxation and surface energy minimization. The strained Si mounds formed over the QDs become instable under thermal annealing, the Si atoms in the mounds migrate to the surrounding area and subsequently the nanopits are formed. The strain distribution in the Si capping layer defines the lateral size of the nanopits, which is close to the lateral size of QDs. QD molecules are grown by a subsequent deposition of Ge on the nanopit-patterned surface.
The work was supported by the special funds for the Major State Basic Research Project No. 2006CB92155 and partial supported by NSFC under No. 10875144.
- Sturm JC, Manoharan H, Lenchyshyn LC, Thewalt MLW, Rowell NL, Noël J-P, Houghton DC: Phys Rev Lett. 1991, 66: 1362. 10.1103/PhysRevLett.66.1362View ArticleGoogle Scholar
- Wang X, Jiang ZM, Zhu HJ, Lu F, Huang DM, Liu XH, Hu CW, Chen YF, Zhu ZQ, Yao T: Appl Phys Lett. 1997, 71: 3543. 10.1063/1.120385View ArticleGoogle Scholar
- Cui J, He Q, Jiang XM, Fan YL, Yang XJ, Xue F, Jiang ZM: Appl Phys Lett. 2003, 83: 2907. 10.1063/1.1616992View ArticleGoogle Scholar
- Garcia JM, Medeiros-Riberio G, Schmidt K, Ngo T, Feng JL, Lorke A, Kotthaus J, Petroff PM: Appl Phys Lett. 1997, 71: 2014. 10.1063/1.119772View ArticleGoogle Scholar
- Lin JH, Wu YQ, Cui J, Fan YL, Yang XJ, Jiang ZM, Chen Y, Zou J: J Appl Phys. 2009, 105: 024307. 10.1063/1.3068192View ArticleGoogle Scholar
- Lee ML, Fitzgerald EA, Bulsara MT, Currie MT, Lochtefeld A: J Appl Phys. 2005, 97: 011101. 10.1063/1.1819976View ArticleGoogle Scholar
- Kar GS, Kiravittaya S, Denker U, Nguyen BY, Schmidt OG: Appl Phys Lett. 2006, 88: 253108. 10.1063/1.2214150View ArticleGoogle Scholar
- Bollani M, Bonera E, Chrastina D, Fedorov A, Montuori V, Picco A, Tagliaferri A, Vanacore G, Sordan R: Nanoscale Res Lett. 10.1007/s11671-010-9773-0Google Scholar
- Zhong Z, Bauer G: Appl Phys Lett. 2004, 84: 1922. 10.1063/1.1664014View ArticleGoogle Scholar
- Chen P, Fan Y, Zhong Z: Nanotechnology. 2009, 20: 095303. 10.1088/0957-4484/20/9/095303View ArticleGoogle Scholar
- Ma YJ, Cui J, Fan YL, Zhong Z, Jiang ZMunpublished unpublishedGoogle Scholar
- Deng X, Krishnamurthy M: Phys Rev Lett. 1998, 81: 1473. 10.1103/PhysRevLett.81.1473View ArticleGoogle Scholar
- Pezzoli F, Stoffel M, Merdzhanova T, Rastelli A, Schmidt O: Nanoscale Res Lett. 2009, 4: 1073. 10.1007/s11671-009-9360-4View ArticleGoogle Scholar
- Wang ZM, Liang BL, Sablon KA, Salamo GJ: Appl Phys Lett. 2007, 90: 113120. 10.1063/1.2713745View ArticleGoogle Scholar
- DiVincenzo DP: Science. 1995, 270: 255. 10.1126/science.270.5234.255View ArticleGoogle Scholar
- Porod W, Lent CS, Bernstein GH, Orlov AO, Amlani I, Snider GL, Merz JL: Int J Electron. 1999, 86: 549. 10.1080/002072199133265View ArticleGoogle Scholar
- Li SS, Long GL, Bai FS, Feng SL, Zheng HZ: Proc Natl Acad Sci USA. 2001, 98: 11847. 10.1073/pnas.191373698View ArticleGoogle Scholar
- Wang ZM, Holmes K, Mazur YI, Ramsey KA, Salamo GJ: Nanoscale Res Lett. 2006, 1: 57. 10.1007/s11671-006-9002-zView ArticleGoogle Scholar
- Lee JH, Wang ZM, Strom NW, Mazur YI, Salamo GJ: Appl Phys Lett. 2006, 89: 202101. 10.1063/1.2388049View ArticleGoogle Scholar
- Liang BL, Wang ZM, Wang XY, Lee JH, Mazur YI, Shih CK, Salamo GJ: ACS Nano. 2008, 2: 2219. 10.1021/nn800224pView ArticleGoogle Scholar
- Songmuang R, Shiraki YJ: Electrochem Soc. 1986, 133: 666. 10.1149/1.2108651View ArticleGoogle Scholar
- Jiang WR, Qin J, Hu DZ, Xiong H, Jiang ZM: J Cryst Growth. 2001, 227: 1106. 10.1016/S0022-0248(01)00997-6View ArticleGoogle Scholar
- Schmidt OG, Eberl K: Phys Rev B. 2000, 61: 13721. 10.1103/PhysRevB.61.13721View ArticleGoogle Scholar
- Pascale A, Berbezier I, Ronda A, Kelires PC: Phys Rev B. 2008, 77: 075311. 10.1103/PhysRevB.77.075311View ArticleGoogle Scholar
- Bergamaschini R, Montalenti F, Miglio L: Nanoscale Res Lett. 10.1007/s11671-010-9723-xGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.