- Nano Express
- Open Access
Can misfit dislocations be located above the interface of InAs/GaAs (001) epitaxial quantum dots?
Nanoscale Research Lettersvolume 7, Article number: 486 (2012)
InAs/GaAs(001) quantum dots grown by droplet epitaxy were investigated using electron microscopy. Misfit dislocations in relaxed InAs/GaAs(001) islands were found to be located approximately 2 nm above the crystalline sample surface, which provides an impression that the misfit dislocations did not form at the island/substrate interface. However, detailed microscopy data analysis indicates that the observation is in fact an artefact caused by the surface oxidation of the material that resulted in substrate surface moving down about 2 nm. As such, caution is needed in explaining the observed interfacial structure.
Semiconductor quantum dots (QDs) are attracting increasing research interest because of their important potential applications in electronic and optoelectronic devices [1–4]. Among the many techniques to produce QDs, coherent island formation through heteroepitaxial growth of semiconductor materials has been the most important technique because of the possible combination of the QD growth and semiconductor integration techniques. Two different methods have been used to grow epitaxial QDs. The first one is the classical Stranski-Krastanow (S-K) growth, [5, 6] in which a semiconductor material with larger lattice parameters is first deposited on the substrate with smaller lattice parameters layer-by-layer, forming a wetting layer, followed by island formation to partially release the strain energy caused by the lattice mismatch between the epilayer and the substrate. The other method is droplet epitaxy [7–9] that has been used for the growth of III-V semiconductor QDs [10, 11] by firstly introducing liquid nanodroplets of the group III element on the substrate and then exposing the droplets to a gas-phase flow of the group V element. Different from the S-K growth mode, droplet epitaxy does not depend on lattice mismatch and therefore can be applied to more materials systems.
For both the S-K growth and the droplet epitaxy, misfit dislocations will be introduced to further release the strain caused by the lattice mismatch when the size of a QD reaches a critical value [12, 13]. Because misfit dislocations produce deleterious effects on QD properties, the mechanisms on how misfit dislocations are generated have been extensively investigated both experimentally and theoretically [11–20]. Based on the fact that the highest stress in a QD occurs at the corner where the island meets the substrate, it has been suggested that strain-relieving perfect misfit dislocations are generated at the island edge, when the island reaches its critical size [14–16]. The generation and morphology of perfect misfit dislocations have been considered theoretically [17–19]. Partial misfit dislocations have also been identified in relaxed islands [9, 20] and are believed to be energetically more favourable than perfect misfit dislocations in some regions of QDs . In some situations, e.g. Ge/Si(001) QDs grown at high temperature, partial misfit dislocations can originate from the surface of QD islands and then glide to the island/substrate interface . While most misfit dislocations are located at the island/substrate interface, it is interesting to note an exception in a recent report that presented evidence of misfit dislocations located above the island/substrate interface in an S-K-grown GaSb/GaAs(001) system . The phenomenon was explained based on the compressive stress induced on the GaSb islands by the GaAs substrate due to lattice mismatch between the epilayer and the substrate . In this letter, we conducted electron microscopy characterization of InAs/GaAs(001) QDs grown by droplet epitaxy. Misfit dislocations in relaxed InAs/GaAs(001) islands were found to be located about 2 nm above the island/substrate interface. However, detailed analysis of the transmission electron microscopy (TEM) data indicates that the observation is in fact an artefact caused by surface oxidation of the material that resulted in substrate surface moving down about 2 nm. We therefore conclude that caution is needed when explaining the interfacial structure of the QDs.
A double-layer InAs/GaAs(001) QD sample was grown by droplet epitaxy using metal organic vapour phase epitaxy. The QD sample was grown on a semi-insulating GaAs(001) substrate in a horizontal flow reactor (AIX200/4, AIXTRON SE, Herzogenrath, Germany) at a pressure of 100 mbar. Trimethylindium, trimethylgallium and AsH3 were used as the precursors and ultrahigh-purity H2 as the carrier gas. Firstly, a 200-nm GaAs buffer layer was deposited at 650°C, then the temperature was reduced to 500°C and the growth was interrupted for 10 s with all sources removed from the reactor to eliminate the influence of AsH3 source on the later deposition of indium droplets. After that, two monolayers of indium droplets were deposited, which were then immediately exposed to the AsH3 flow (3.0 × 10−4 mol/min) for 15 s. The InAs QDs were capped immediately by a 10-nm GaAs capping layer at 500°C. Then, the growth temperature was ramped up to 650°C and a 100-nm GaAs layer was deposited. The InAs QDs were capped immediately by a 100-nm GaAs capping layer whilst the temperature was ramped up to 650°C. Finally, a surface layer of InAs QDs was grown in the same fashion as the buried InAs QDs.
Only the structures of QDs on the top layer were investigated. Cross-sectional TEM specimens were prepared using a Gatan precision ion polishing system (Gatan, Inc., Pleasanton, CA, USA) with Ar+ energy of 2.5 keV. Structural characterization of the QDs was conducted using scanning electron microscopy (SEM; Zeiss Ultra+, Carl Zeiss, Inc., Oberkochen, Germany) operated at 2 kV and TEM (JEM-3000 F, JEOL Ltd., Akishima-shi, Japan) operated at 300 kV. Quantitative compositional analysis was conducted using X-ray energy-dispersive spectroscopy (XEDS) in JEM-2200 TEM (JEOL Ltd.) operated at 200 kV and the ESPRIT software. The electron probe size used for the XEDS was 1 nm. XEDS data collection time was controlled to make sure that high counts (larger than 10 thousands) were obtained for datum points so that the statistical errors were less than 1%.
Results and discussion
Figure 1a shows a typical plan-view SEM image of the QDs on the surface. Large islands with widths (measured along the  direction) larger than approximately 30 nm are usually elongated and faceted, showing typical strain-relaxed morphologies . Most of small islands with widths smaller than approximately 15 nm have an elliptical shape, although a small number of them are circular. The large islands are all relaxed through the formation of misfit dislocations. Figure 1b shows a typical cross-sectional high-resolution TEM image of a relaxed large island on the sample surface. Three misfit dislocations, which are indicated with three white arrows, are seen at the same atomic layer in the island, which is approximately 2 nm above the crystalline surface of the substrate. A white line marked as ‘1’ indicates the crystalline surface of the substrate, and the other white line marked with ‘2’ specifies the position of the atomic layer on which the misfit dislocations lie.
The phenomenon of misfit dislocations positioned above the crystalline substrate surface seen in Figure 1 is very similar to the phenomenon reported in . However, another interesting phenomenon is also seen in Figure 1. There are two amorphous structures with distinctly different contrasts above the crystalline structure: one is a thin amorphous layer with a thickness of about 2 nm immediately above the crystalline substrate and island surface and with contrast clearly darker than the other part of the amorphous area above the sample. The top surface of the 2-nm thin amorphous layer is located at exactly the same level as the atomic plane where misfit dislocations lie, i.e. the layer indicated by line 2. While the amorphous structure with bright contrast is the epoxy used for cross-sectional sample preparation, it is not clear what the dark amorphous structure is.
XEDS analysis was conducted to explore the composition of the dark-contrast amorphous layer and surrounding areas. Figure 2 shows the results obtained from an island and its nearby area. Line scans were conducted along lines ABCD and EFG shown in Figure 2a with the interval of the datum points at approximately 2 nm. O, Ga, As and In were detected in the material. The quantitative concentrations of these elements as a function of positions along lines ABCD and EFG are shown in Figure 2b,c, respectively. From Figure 2b, a small amount of In content is detected in the substrate; this is caused by the combined effect of the contamination induced during the sample preparation process and the electron beam spreading to the island through the scatter of the electron beam in the TEM sample. The latter effect can be removed when XEDS was conducted in the substrate far away from the island, as shown in Figure 2c. Another possible reason that In is detected in the substrate is the mutual diffusion of In into the substrate and Ga into islands during the high-temperature sample deposition process. This explains the phenomenon that over 10% of Ga is detected in the island. The inter-diffusion process of In and Ga is in fact an effective way to reduce the strain energy of the system [16, 25, 26].
The most interesting phenomena revealed by the XEDS data in Figure 2 are (1) although O is detected everywhere in the sample, which is caused by surface oxidation after TEM sample preparation, the O content increases significantly at the dark amorphous layer; (2) the contents of Ga at the position immediately below misfit dislocations and at the dark-contrast amorphous layer immediately above the crystalline substrate are very high; and (3) the In content at the dark-contrast amorphous layer covering the island is very high. All these evidence point to the conclusion that the dark-contrast amorphous layer is actually an oxidised layer that was originally the surface of the crystalline GaAs substrate and the surface of InAs islands. The oxidised amorphous layer formed after the epitaxial sample surface was exposed to the air and therefore was only seen on the sample surface, not in the buried QD layer. Therefore, misfit dislocations in large relaxed QD islands formed exactly at the interface of the epilayer and the substrate, not at a level approximately 2 nm above the interface as it looks in Figure 1. Because oxidation of the surface of semiconductor nanostructures has been widely reported in literatures [27–29], not just in the InAs/GaAs QD structure reported here, caution is needed when investigating the interfacial structures of epitaxial materials.
In summary, misfit dislocations in InAs/GaAs(001) QDs grown by droplet epitaxy are observed to be located approximately 2 nm above the crystalline substrate surface. However, detailed compositional analysis suggests that this is an artefact caused by surface oxidation. The oxidised surface is of an amorphous structure with a thickness of approximately 2 nm.
ZBC is a postgraduate student. Dr WL is a research associate professor. Dr BC is a postdoctoral research associate. Dr YBW is an Australian Research Council (ARC) Australian Postdoctoral Fellow. Dr XZL is an associate professor and ARC Future Fellow. Dr HHT is a senior fellow and ARC Future Fellow. Dr JZ is a professor and ARC Future Fellow. SPR is the Director of the Australian Centre for Microscopy and Microanalysis, the Executive Director and CEO of the Australian Microscopy and Microanalysis Research Facility, and the Director of the Bandwidth Foundry International Pty Ltd. Dr CJ is a distinguished professor and ARC Laureate Fellow.
Amlani I, Orlov AO, Toth G, Bernstein GH, Lent CS, Snider GL: Digital logic gate using quantum-dot cellular automata. Science 1999, 284: 289–291. 10.1126/science.284.5412.289
Salter CL, Stevenson RM, Farrer I, Nicoll CA, Ritchie DA, Shields AJ: An entangled-light-emitting diode. Nature 2010, 465: 594–597. 10.1038/nature09078
Imamoglu A, Awschalom DD, Burkard G, DiVincenzo DP, Loss D, Sherwin M, Small A: Quantum information processing using electron spins and cavity-QED. Phys Rev Lett 1999, 83: 4204–4207. 10.1103/PhysRevLett.83.4204
Burkard G, Loss D, DiVincenzo DP: Coupled quantum dots as quantum gates. Phys Rev B 1999, 59: 2070–2078. 10.1103/PhysRevB.59.2070
Stranski IN, Krastanow L: Zur Theorie der orientierten Ausscheidung von lonenkristallen aufeinander, Sitzungsberichte d. Akad. d. Wissenschaften in Wien. Math Nat KI IIb 1937, 146: 797–810.
Lee JH, Wang ZM, Strom NW, Mazur YI, Salamo GJ: InGaAs quantum dot molecules around self-assembled GaAs nanomound templates. Appl Phys Lett 2006, 89: 202101(1)-202101(3).
Koguchi N, Ishige K, Takahashi S: New selective molecular-beam epitaxial growth method for direct formation of GaAs quantum dots. J Vac Sci Technol B 1993, 11: 787–790. 10.1116/1.586789
Wu J, Shao DL, Dorogan VG, Li AZ, Li SB, DeCuir EA Jr, Manasreh MO, Wang ZM, Mazur YI, Salamo GJ: Intersubband infrared photodetector with strain-free GaAs quantum dot pairs grown by high temperature droplet epitaxy. Nano Lett 2010, 10: 1512–1516. 10.1021/nl100217k
Jo M, Mano T, Abbarchi M, Kuroda T, Sakuma Y, Sakoda K: Self-limiting growth of hexagonal and triangular quantum dots on (111)A. Cryst Growth Des 2012, 12: 1411–1415. 10.1021/cg201513m
Mano T, Watanabe K, Tsukamoto S, Fujioka H, Oshima M, Koguchi N: Fabrication of InGaAs quantum dots on GaAs(001) by droplet epitaxy. Jpn J Appl Phys 2000, 39: 504–508.
Wang ZM, Holmes K, Mazur YI, Ramsey KA, Salamo GJ: Self-organization of quantum-dot pairs by high-temperature droplet epitaxy. Nanoscale Res Lett 2006, 1: 57–61. 10.1007/s11671-006-9002-z
Chen Y, Lin XW, Liliental-Weber Z, Washburn J, Klem JF, Tsao JY: Dislocation formation mechanism in strained InxGa1-xAs islands grown on GaAs(001) substrates. Appl Phys Lett 1996, 68: 111–113. 10.1063/1.116773
Liao XZ, Zou J, Cockayne DJH, Qin J, Jiang ZM, Wang X, Leon R: Strain relaxation by alloying effects in Ge islands grown on Si(001). Phys Rev B 1999, 60: 15608.
LeGoues FK, Reuter MC, Tersoff J, Hammar M, Tromp RM: Cyclic growth of strain-relaxed islands. Phys Rev Lett 1994, 73: 300–303. 10.1103/PhysRevLett.73.300
LeGoues FK, Tersoff J, Reuter MC, Hammar M, Tromp R: Relaxation mechanism of Ge islands/Si(001) at low temperature. Appl Phys Lett 1994, 67: 2317–2319.
Wedler G, Walz J, Hesjedal T, Chilla E, Koch R: Stress and relief of misfit strain of Ge/Si(001). Phys Rev Lett 1998, 80: 2382–2385. 10.1103/PhysRevLett.80.2382
Johnson HT, Freund LB: Mechanics of coherent and dislocated island morphologies in strained epitaxial material systems. J Appl Phys 1997, 81: 6081–6090. 10.1063/1.364357
Zhou S, Liu YM, Wang DL, Xin X, Cao G, Lu PF, Yu ZY: The preferential formation site of dislocations in InAs/GaAs quantum dots. Superlattice Microst 2012, 51: 53–61. 10.1016/j.spmi.2011.10.004
Spencer BJ, Tersoff J: Dislocation energetics in epitaxial strained islands. Appl Phys Lett 2000, 77: 2533–2535. 10.1063/1.1319186
Yamaguchi H, Belk JG, Zhang XM, Sudijiono JL, Fahy MR, Jones TS, Pashley DW, Joyce BA: Atomic-scale imaging of strain relaxation via misfit dislocations in highly mismatched semiconductor heteroepitaxy: InAs/GaAs(111)A. Phys Rev B 1997, 55: 1337–1340. 10.1103/PhysRevB.55.1337
Ovid’ko IA: Relaxation mechanisms in strained nanoislands. Phys Rev Lett 2002, 88: 046103–046106.
Zou J, Liao XZ, Cockayne DJH, Jiang ZM: Alternative mechanism for misfit dislocation generation during high-temperature Ge(Si)/Si(001) island growth. Appl Phys Lett 2002, 81: 1996–1998. 10.1063/1.1506414
Kim YH, Lee JY, Noh YG, Kim MD, Oh JE: High-resolution transmission electron microscopy study on the growth modes of GaSb islands grown on a semi-insulating GaAs(001) substrate. Appl Phys Lett 2007, 90: 241915–241917. 10.1063/1.2747674
Costantini G, Manzano C, Songmuang R, Schmidt OG, Kern K: InAs/GaAs(001) quantum dots close to thermodynamic equilibrium. Appl Phys Lett 2003, 82: 3194–3196. 10.1063/1.1572534
Liao XZ, Zou J, Cockayne DJH, Jiang ZM, Wang X, Leon R: Composition and its impact on shape evolution in dislocated Ge(Si)/Si islands. Appl Phys Lett 2000, 77: 1304–1306. 10.1063/1.1290384
Rosenauer A, Gerthsen D, Van Dyck D, Arzberger M, Bohm G, Abstreiter G: Quantification of segregation and mass transport in InxGa1-xAs/GaAs Stranski-Krastanow layers. Phys Rev B 2001, 64: 245334–245348.
Wang YB, Wang LF, Joyce HJ, Gao Q, Liao XZ, Mai YW, Tan HH, Zou J, Ringer SP, Gao HJ, Jagadish C: Super deformability and Young’s modulus of GaAs nanowires. Adv Mater 2011, 23: 1356–1360. 10.1002/adma.201004122
Bhat TN, Rajpalke MK, Roul B, Kumar M, Krupanidhi SB, Sinha N: Evidences for ambient oxidation of indium nitride quantum dots. Phys Status Solidi B 2011, 248: 2853–2856. 10.1002/pssb.201147112
Shiramine K, Muto S, Shibayama T, Sakaguchi N, Ichinose H, Kozaka T, Sato S, Nakata Y, Yokoyama N, Taniwaki M: Tip artefact in atomic force microscopy observations of InAs quantum dots grown in Stranski-Krastanow mode. J Appl Phys 2007, 101: 033527–033531. 10.1063/1.2434806
The authors are grateful for the scientific and technical support from the Australian Microscopy & Microanalysis Research Facility node at the University of Sydney. We thank the Australian National Fabrication Facility for providing access to growth facilities used in this work. This research was financially supported by the Australian Research Council.
The authors declare that they have no competing interests.
ZBC, WL and XZL designed the study. ZBC conducted the microscopy experiments. WL, HHT and CJ grew the QD specimens. ZBC and XZL wrote the paper. All authors discussed the results and contributed to the paper. All authors read and approved the final manuscript.