Open Access
Add to BibSonomy
Abstract
Single layer self-assembled InGaAs quantum dots (QDs) are manipulated by using different arsenic species on GaAs (100) surface. The As4 molecules are experimentally observed to be more promising than As2 to promote the formation of one-dimensionally-aligned QD-chain arrays. The lateral alignment of QDs and the corresponding formation of dot chains are explained by the anisotropic surface kinetics in combination with the different reactivities of the two molecules with bonding sites on the GaAs (100) surface. Photoluminescence (PL) measurements demonstrate that the spectra of the QD-chains broaden to higher energy and increases in intensity with increasing excitation laser power. The PL band of the QD-chains also exhibits a 9 meV reduction in linewidth as temperature increases starting from 8 K. These observations confirm an efficient lateral coupling between neighboring QDs and thereafter polarized QD emission, whereas the randomly distributed QDs grown with As2 show no preferential polarization. Such QD-chains exhibiting anisotropic properties have the potential for nanophotonics applications like electro-optic modulators with very low drive voltage and ultra-wide bandwidth operation.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Over the recent twenty-five years semiconductor In(Ga)As/GaAs quantum dots (QDs) have been the subject of increasing interest based on their promising potential in a wide variety of optoelectronic device applications. The development of In(Ga)As/GaAs QD-based transistors, lasers, detectors, sensors, photovoltaic, and modulators is anticipated to produce exceptional devices that are more powerful, sensitive, and efficient than any existing today [1–4]. Currently the most common technique for fabrication of In(Ga)As QDs involves strain-driven self-assembly, known as the Stranski-Krastanov (SK) growth mode. However, SK grown QDs are typically characterized by large size fluctuations and random distributions of spatial position. These inhomogeneous characteristics limit the possibilities of utilizing the zero-dimensional density of states for the QDs and the corresponding performances for exceptional optoelectronic devices [5–7].
While In(Ga)As/GaAs QDs offer great potential for applications they remain a great technological challenge to the crystal growth community in the achievement of homogeneous and spatially ordered arrays of QDs. Approaches to control QD formation using strain engineering and selective area epitaxy have been studied in attempts to alter the SK growth and achieve uniform, position-controlled QDs [8–12]. Encouragingly, great achievements have been obtained with respect to the inhomogeneous broadening for site controlled InGaAs/GaAs QDs [9]. Also, experiments looking at stacking QD layers have exposed the possibility of both improved size uniformity and enhanced in-plane ordering of the QDs. In particular, distinct three-dimensional (3D) ordering of QDs has been demonstrated by stacking of InGaAs QDs in successive layers on GaAs (100) surfaces to form spatially-organized QD-chains [13–15]. The spatial organization of QDs to form QD-chains is attributed to enhanced surface diffusion and the vertical transfer of anisotropic strain [16,17]. In spite of the achievement of 3D spatial organization for InGaAs QD-chains, however, the dedicated study of optical properties related with the spatial ordering of QDs, in particular, for a single layer of InGaAs QD-chains on planar GaAs (100) surfaces, is still lacking. In this paper, we report that the arsenic species used have a significant influence on the morphology and optical properties of a single layer of InGaAs QDs grown on planar GaAs (100) by molecular beam epitaxy (MBE) as demonstrated by atomic force microscopy (AFM) and photoluminescence (PL). The As4 molecules are observed to be more promising than As2 molecules to form close-spaced one-dimensionally organized InGaAs QD-chain arrays. As results, the lateral carrier transfer is prominent between neighboring QDs for the InGaAs QD-chains grown with As4 flux, leading to a strong correlation between the PL performance and the self-alignment of the QDs in this single layer of InGaAs QDs.
2. Experiments and methods
Two QD samples were grown on semi-insulating GaAs (100) substrates by a solid source MBE reactor. A cracked arsenic source with a base zone temperature of 350°C was used to generate As4 (As2) by setting the cracking zone temperature to 550°C (900°C) for sample A(B). For both samples, after the native oxide was removed, a 500 nm thick GaAs buffer layer was grown with a growth rate of 1.0 ML/sec using a beam equivalent pressure (BEP) ratio of As to Ga of 15 at a substrate temperature of 580°C. The substrate temperature was then lowered to 540°C and 8 ML of In0.4Ga0.6As was deposited to form the QDs, following standard growth conditions for InGaAs QD-chain structures [13,14]. After a 30 second growth interruption, which plays an important role in the formation of close-spaced one-dimensionally-aligned QDs, the QDs were capped by 50 nm of GaAs at 540°C and then 8 ML of In0.4Ga0.6As was deposited to form a top layer of QDs for morphological investigation. The samples were then cooled and removed from the growth chamber for AFM and PL analysis. Both samples follow exactly the same growth procedures except for the arsenic species used (As4 or As2) where the flux was modified such that the resulting BEP was the same. For PL and polarization PL investigations, the samples were mounted on the cold finger of a close-cycled cryostat (8-300K). A continuous-wave laser, operated at 532 nm, was focused on the sample surface with a diameter of ~20 μm to excite the QDs. The PL signal was dispersed by an Acton-2500i spectrometer and then detected by a CCD detector array.
3. Results and discussion
The QD morphology for both samples is studied by AFM. Figure 1(a) shows the AFM of sample A with large arrays of long QD-chains formed along the [01-1] direction. Here, it is found that the longest chains extend beyond one micrometer in length. The InGaAs QDs are measured to have an areal density of 3.9x1010/cm2, an average height of 9.3 ± 1.3 nm (see the histogram in Fig. 1(c)), and an average lateral diameter of 52.7 ± 5.5 nm. An interesting observation is that the InGaAs QDs in each chain preserve their round shape, and are close-grouped in together with almost no separation along [01-1] direction but a larger separation along [011] direction. From the auto-correlation image in the inset and the corresponding profile analysis in Fig. 1(b), the average tip-to-tip spacing between QD chains is measured to be about 66 nm, while the average tip-to-tip separation between QDs in each chain is about 55 nm, which is close to the lateral size of the QDs. As a result, we expect that lateral electronic coupling will occur between QDs along the [01-1] direction and change their optical properties.
Fig. 1 Morphology study of the QDs for sample A grown with As4 and sample B grown with As2. (a) and (d) are 1μm x 1μm AFM images to show the InGaAs QDs for sample A grown with As4 and sample B grown with As2. The insets show the corresponding three-dimensional plot for the auto-correlation image of the AFM. (b) and (e) are the profiles along [011] and [01-1] directions from the auto-correlation image in the inset for sample A and sample B, respectively. (c) and (f) are the histogram of QD height distribution for sample A and sample B.
For sample B, as shown by Figs. 1(d)-(f), the InGaAs QDs are measured to have an areal density of 1.9x1010/cm2, an average height of 9.7 ± 2.1 nm (see the histogram in Fig. 1(f)), and an average lateral diameter of 64.7 ± 9.2 nm. Although short range ordering of the QDs is observed in a few places here, the QD separation is in general much larger, and the coherent lengths of the grouped QDs are much shorter. Indeed, most of the InGaAs QDs are randomly distributed on the sample surface as common SK QDs. We find that the separation between the neighboring QDs is quite large, as measured with an average tip-to-tip distance of 71 nm along [01-1] direction.
The AFM analysis indicates that with only a single layer of QD growth, we have achieved well-aligned, QD-chain arrays which usually require stacking more than ten layers of InGaAs QDs. For stacking QD-chain structures, which normally show lateral alignment after 5 to 6 layers of QD deposition, a combination of anisotropic strain and enhanced adatom diffusion has been proposed as the major factor controlling the alignment [18–20]. However, in this experiment the dots are already lined up in the [01-1] direction after only a single layer of InGaAs QD deposition. This indicates that the formation mechanism for these single layer QD-chains is different from that of the previously reported multi-layer QD-chains. It is likely due to the anisotropic lateral adatom diffusion, but not the transferred strain field that is responsible for the formation of single layer QD-chains [21].
The observed anisotropic lateral alignment and the corresponding formation of dot chains is explained by the anisotropic surface kinetics that occur during the InGaAs growth, as the indium and gallium migration on the surface can be altered by the arsenic molecules. The use of the As2/As4 background for effective manipulation of QD shape, positioning and deformations can be understood by considering surface diffusion. Due to the nature of the (2 × 4) GaAs (100) surface reconstruction at the growth temperature of 540°C, the adatom diffusion length along the [01-1] direction is estimated to be eight times larger than that along the [011] direction [22]. This anisotropic surface diffusion generally leads to elongation of the QDs or the formation of QD-chains along the direction of higher mobility.
The difference between QD alignment for growth using As2 or using As4 can be further understood as being due to the different reactivities of the two molecules with bonding sites on the GaAs (100) surface. The As2 will become chemisorbed to the surface very quickly whereas the As4 will exist in a physisorbed state for some time before becoming chemically bonded with the surface at a fixed deposition site. Based on this, several groups have claimed that As2 molecules fix the Ga at a stable site on the surface faster than As4, and thus reduce the In/Ga diffusion length on the GaAs (100) surface [23,24]. Therefore, for growth with As4, the In/Ga diffusion length is longer and thus more influenced by the surface anisotropy resulting in the aligned QD growth. Although, for our growth at the high temperature of 540°C, we believe that it is easier for As2 to bind with Gallium or Indium adatoms before bonding to the GaAs surface. This more complex molecule would no longer feel the anisotropy of the surface but still have a large diffusion length leading to the isotropic distribution of larger QD at lower density as seen in Fig. 1. Similar results have previously been presented that the surface diffusion of Indium atoms is enhanced under an As2 source [25]. So the QDs grown with the As2 source were larger and less dense than those grown with the As4. The use of As2 also reduced the anisotropic surface kinetics on the GaAs (100) surface.
These results as discussed are tied closely to the varying surface kinetics on GaAs (100) surface under As2 or As4. Therefore, any changes in the growth conditions (for example, the temperature, Indium mole fraction, growth interruption, et al) that can affect the surface diffusion could alter the QD-chain formation. It is because of this reasoning that we introduced a long growth interruption of 30 seconds after the InGaAs QD formations during the growth. This long growth interruption plays an important role in enabling the Indium and Gallium adatoms to have more migration to promote the formation of one-dimensionally-aligned QD-chains. Actually, previous research on stacking QD-chains have adopted the same approach to introduce multiple growth interruptions during the InGaAs growth, resulting in long chains of QDs with lengths well over 5 µm [26].
The distinct morphologies for samples A and B are expected to modify their optical performance accordingly. PL spectra for both samples were first measured at low temperature (8 K) with a low intensity excitation of 0.03 W/cm2 so that only the ground states of the QDs are excited. As shown in Fig. 2, sample A has a PL peak at 1.214 eV with a Full Width at Half Maximum (FWHM) of 37.8 meV. In comparison, the PL peak for sample B is broader, 58.1 meV FWHM, and shifted to the blue at 1.223 eV. Here we find that the PL from sample B is similar to that from normal randomly distributed QDs, well represented by a Gaussian distribution, whereas the PL from sample A is much narrower and asymmetric. As sample A has a symmetric QD height distribution in Fig. 1(c), its nonsymmetrical PL profile is likely a result of the efficient lateral coupling between closely spaced dots, where the smaller, higher energy dots transfer their carriers to the larger, lower energy dots. Even within this already narrow distribution, this skews the peak to the lower energy side. However, these effects do not completely explain sample A, with on average smaller QDs than sample B, having a lower energy PL peak than sample B. Here we believe that due to the subtle differences in the QD formation, there may be a real difference in Indium composition and strain within the InGaAs QD layer. However, more systematic study is needed to gain a reasonable explanation.
Fig. 2 Low temperature (T = 8 K) PL spectra. (a) and (b) are the low temperature PL spectra for sample A and sample B measured with an excitation intensity of 0.03 W/cm2.
We also measured the PL polarization for our samples at low temperature (8 K) with the same low excitation intensity of 0.03 W/cm2. To eliminate the influence of the spectrometer response on the different polarization components, a non-polarized white light source was measured in advance to calibrate the spectrometer response. The PL spectra measured along [011] direction and [01-1] direction for each sample are plotted in Fig. 3 along with polar plots showing the variation of PL intensity with polarization direction. These plots clearly demonstrate the two-fold symmetry for sample A. The QD-chains in sample A exhibited a significant polarization anisotropy of ~11%, while sample B showed only an insignificant amount of anisotropy. This anisotropy was calculated as normal by Eq. (1):
where the I[011] and I[01-1] are the intensities of the PL components along the [011] direction and the [01-1] direction of the sample surface, respectively. The polarized luminescence is a clear fingerprint of the anisotropic properties of the QD-chains. However, we did not observe clear anisotropic QD shape for either sample by AFM. Therefore, we speculate that in the QD-chain structures, the anisotropic polarization is induced through lateral electronic coupling and carrier transfer along the chain direction as it has been revealed recently [27]. This electronic coupling is related to the large penetration of electron wave function into the adjacent dots due to the close alignment of the QDs. This results in anisotropic excitonic states in the QD-chains, and subsequently in anisotropic polarized emission, even though the QDs in sample A are generally round in shape, i.e., not anisotropic in the plane [28]. In particular, we observe brighter PL emission in the direction perpendicular to the QD-chains instead of parallel to the chains, it is likely that the quantum confinement is reduced along the QD-chain direction due to the coupling between neighboring QDs along the chain. Therefore, the oscillator strength as well as the resulting emission intensity are smaller in the direction parallel to the chains.Fig. 3 The polarized PL spectra. (a) and (b) are the polarized PL spectra for sample A and sample B measured at 8 K with a laser excitation intensity of 0.03 W/cm2,the inset polar plots show the variation of PL intensity with polarization direction
The PL spectral profile, integrated PL intensity, and FWHM are then measured as functions of the excitation laser intensity at the low temperature of 8 K for both samples. For this measurement, the excitation laser intensity is kept lower than 30 W/cm2 to avoid higher energy states of the QDs being excited. As shown by the normalized PL spectra in Fig. 4(a), for sample A we observe a clear broadening towards the high energy side of the PL band, which corresponds to filling smaller QDs. This can be regarded as a fingerprint of carrier transfer between QDs inside the chains [29,30]. The normalized PL spectra for sample B are given in Fig. 4(b). Here, we observe no measureable broadening or change in the spectra as the excitation intensity increases. Therefore, for sample B there is no lateral carrier transfer between neighboring QDs, which is a result of their large separation.
Fig. 4 Excitation laser intensity-dependent PL measurements for both samples. (a) and (b) show the PL spectral profile measured at 10K as a function of the laser intensity. (c) the FWHM as a function of the temperature.
To verify our hypothesis of lateral carrier transfer between neighboring QDs in sample A, Fig. 4(c) shows that the two samples exhibit different dependencies of their FWHMs on the laser excitation intensity even starting from low pump power. For sample A, its FWHM slowly increases from 37 meV to 42 meV as the excitation intensity is increased from 0.003 W/cm2 to 30 W/cm2. This observed linewidth broadening is attributed to carrier transfer between the laterally coupled InGaAs QDs in sample A. In the QD-chain for sample A, the photon-excited carriers transfer from small QDs to the larger ones and therefore fill the larger, lower energy QDs first, leading to the asymmetry in the PL spectrum found in Fig. 2(a). However, as the excitation intensity increases, the possibility of the carrier recombination in smaller, higher energy QDs will increase as the larger ones become populated. Therefore, we observe the increase in the FWHM before the appearance of the first excited state of the QDs. While, for sample B, the QDs behave more like normal SK QDs where the low intensity PL is a true representation of the size distribution without lateral carrier transfer and where the FWHM of the ground state remains very stable with increasing intensity until the ground state emission of the QDs becomes saturated at ~30 W/cm2.
Fig. 5 compares the temperature dependent PL results for both samples with the excitation laser intensity set at 3W/cm2 so as to be in the low excitation intensity regime. The PL peak shifts are plotted in Fig. 5(a) and 5(b), where the dash curve is the calculated results from InGaAs band gap according to the Varshni law. For sample B, the red-shift of the PL peak agrees well with the calculation for T<110K, which means that for sample B in this range, the PL peak shift is mainly caused by the variation of the InGaAs band gap. However, for sample A, the PL peak shifts to lower energy at a slightly slower rate starting from the temperature as low as 20K. Again, the same scenario as in the intensity dependent analysis, we assume that the barrier for carrier transfer between QDs in the QD-chains is very low such that increasing the temperature can provide sufficient energy for carriers to transfer from lower energy dots to higher energy dots along the QD-chains. This only happens between the neighboring QDs in the QD-chains because we are in the lower temperature regime, but the transfer is sufficient to observe this departure from the Varshni law.
Fig. 5 Temperature-dependent PL measurements for both samples. (a) and (b) show the peak energy shift as a function of the temperature where the dash curve is calculated results from the InGaAs band gap according to the Varshni law, (c) and (d) are the FWHM as a function of the temperature, (e) and (f) are the integrated PL intensity as a function of the temperature.
This carrier transfer can also be seen in the dependence of the FWHM on temperature in Fig. 5(c) and 5(d). For sample A at 8 K, we see that the FWHM is a maximum of 41 meV at 8 K. With increasing temperature, it immediately decreases to a plateau at ~70K, then to a minimum of 32 meV at ~150K, about 9 meV smaller than its maximum value at 8 K. This is a typical characteristic of InGaAs QD ensembles and the reduction of PL FWHM is generally attributed to the carrier thermal redistribution among the QDs, i.e., carriers are thermally excited from higher energy QDs, with low confinement energy, and relax into a smaller (narrower) ensemble of lower energy QDs. We believe the local minimum at ~70K represents the concept that the energy for transfer between QDs is anisotropic for this sample [31,32]. In comparison with sample A, sample B only has an insignificant FWHM variation of ~0.5 meV between 8 K and ~70 K, likely because the QDs in sample B do not have the lower energy transfer barrier that sample A has. But, similar to sample A, sample B reaches its FWHM minimum at 150K. This indicates that the high energy barrier for carrier activation in sample A is the same as the isotropic energy barrier between QDs in sample B.
Figures 5(e) and 5(f) present the integrated PL intensities of the QDs as functions of the temperature. The experimental data are fitted by using an Arrhenius Eq. involving two processes given by:
where I(T) is the integrated PL intensity at temperature of T, I0 is the intensity at temperature of T = 0K, k is Boltzmann's constant, C1 and C2 are fitting parameters, E1 and E2 are the thermal activation energies for the two processes governing the carrier thermal quenching in the low and high temperature regimes, respectively [33]. As shown in Fig. 5(e) and 5(f), sample A has small activation energies for both E1 and E2. In general, E1 corresponds to the binding energy or detrapping energy of carriers, while E2 corresponds to the energy that is required by carriers to jump out of the quantum confined energy state. It is clear that the activation energy E2 = 14.75 meV is smaller than the energy difference between the QD discrete energy levels or the energy separation from the highest QD excited state energy level to the wetting layer energy state. In comparison, sample B has larger activation energies, E1 = 4.75 meV and E2 = 21.63 meV, for both the low and high temperature regimes, indicating that the barrier for carrier jumping out from QDs for sample B is in fact higher than that for sample A. The low activation energies for sample A is likely the result of the low barrier for carrier transfer among QDs within the QD-chains, along the chain direction.The carrier transfer is further investigated through measuring the temporal decay of the PL for both samples at 8 K. For time-resolved PL (TRPL) experiments the samples are excited by a Ti:Sapphire mode-locked laser (780 nm, 78 MHz, 2.7 picosecond), while a Hamamatsu C5680 streak camera with an infrared-enhanced photocathode is used as the detection system. The overall system resolution for TRPL measurements is about 35 ps. First, the decay curves are measured at the QD peak wavelength with excitation intensity of ~4x107photons/pulse for both samples. As shown by Fig. 6(a), for both samples, the QD emission exhibits mono-exponential decay behavior, and the decay curves can be fitted by:
where I(t) is the PL intensity at time, t, and τ is the carrier lifetime. Obviously, sample A has a slightly longer lifetime (τ = 1.06 ns) than sample B (τ = 0.92 ns). The difference in lifetimes for the QDs is clearly not due to their size difference. We speculate that the lateral carrier transfer in sample A provides additional carriers for the large QDs and prolongs their measured carrier lifetime.Fig. 6 TRPL measurements. (a) PL decay curves measured at QD peak wavelength at T = 8 K. (b) PL spectrum and lifetime measured at different wavelengths for sample A. (c) PL spectrum and lifetime measured at different wavelengths for sample B.
Then Fig. 6(b) plots the carrier lifetime as a function of monitored emission energies for both samples. It can be seen that for sample A the carrier lifetime depends strongly on the photon energy being detected within the QD PL band. As the detection energy moves off the QD peak, the lifetime drops rapidly, from 1060 ps at the QD peak energy of 1.203 eV, to 165 ps at 1.31 eV. The shorter carrier lifetime at the high energy side is attributed to the carrier relaxation to lower-lying energy states in addition to the radiative recombination at that energy, as the decay of excitons depends not only on the radiative recombination but also on carrier transfer process to the low-energy lying states in sample A [34,35]. As before, this carrier transfer and thereafter the electronic coupling is related to the large penetration of the electron wave-function into adjacent dots due to the close alignment of QDs along the [01-1] direction inside the QD-chains. As a comparison, the results of the more SK-like QDs in sample B are shown in Fig. 6(c). The measured PL decay time shows much less energy dependence from 900 ps at the QD peak energy of 1.218 eV to 200 ps at 1.347 eV, indicating a negligible lateral electronic coupling among the QDs in this sample.
In summary of the photoluminescence measurements, we confirm that the QD-chains. i.e., the 1D lateral-ordering of the InGaAs QDs, can modify the quantum confinement and enable anisotropic emission properties for the QDs. Such spatially-organized QD-chain arrays are different from either normal 0-Dimensional SK QDs or 1D nanowire quantum structures. These QD-chains preserve the 0-dimensional nature for QDs, but enable 1D anisotropic optical performance. It is an ideal candidate for the development of certain photonics elements, like electro-optic modulators and electro-absorption modulators with very low drive voltage and ultra-wide bandwidth operation [28]. We also should mention that, by careful control of the growth conditions not only the As flux but also the indium composition, temperature, and growth interruption, it is possible to achieve the transition from 0D SK QDs to laterally-organized QD-chains, and then to 1D nanowires [36]. Therefore, it is a flexible approach to manipulate the InGaAs QD self-assembled growth to obtain nanostructures of different quantum confinement nature for photonics device applications.
4. Conclusions
In conclusion, single layers of InGaAs QDs grown with As2 and As4 flux are carefully investigated. It is shown that the As4 molecules form close-spaced one-dimensionally-aligned QD arrays, i.e., QD-chains, whereas the QDs formed using As2 are much less organized and less dense resembling normal SK grown QDs. This is understood as being the result of changes in surface mobility of the metal atoms when growing with the two different arsenic species in combination with the different reactivities of the two arsenic molecules with bonding sites on GaAs (100) surface. PL measurements show that the QD-chains grown with As4 exhibit higher degrees of QD uniformity as seen in the narrower FWHM, and they demonstrate prominent carrier transfer between neighboring QDs due to the close-spaced one-dimensional-ordering as can be generally seen in the increase in the ground state PL linewidths with increasing excitation intensity. In addition, the ground state PL linewidths of the QD-chains exhibit a 9 meV reduction resulting from increasing the temperature from 8 K to 150 K. These observations indicate a strong lateral carrier transfer between neighboring QDs for the QD-chains, so that we build a clear correlation between the PL performance and the spatially-ordering properties for a single layer of InGaAs QDs. This research provides a flexible approach to manipulate single layer InGaAs QDs by using the arsenic species to obtain spatially-organized QD-chain arrays for nanophotonics applications, like electro-optic modulators and electro-absorption modulators with very low drive voltage and ultra-wide bandwidth operation.
Funding
National Natural Science Foundation of China (NSFC) (61774053); Advanced Talents Incubation Program of the Hebei University (8012605); National Science Foundation of the United States (NSF) (EPSCoR OIA-1457888).
Acknowledgments
The authors acknowledge the useful discussion with Prof. Zhao Ding at Guizhou University.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. D. Bimberg, M. Grundmann, and N. N. Ledentsov, Quantum Dot Heterostructures, (Wiley: New York, 1998).
2. Z. Yuan, B. E. Kardynal, R. M. Stevenson, A. J. Shields, C. J. Lobo, K. Cooper, N. S. Beattie, D. A. Ritchie, and M. Pepper, “Electrically driven single-photon source,” Science 295(5552), 102–105 (2002). [CrossRef] [PubMed]
3. K. A. Sablon, J. W. Little, V. Mitin, A. Sergeev, N. Vagidov, and K. Reinhardt, “Strong enhancement of solar cell efficiency due to quantum dots with built-in charge,” Nano Lett. 11(6), 2311–2317 (2011). [CrossRef] [PubMed]
4. H. Y. Liu, T. Wang, Q. Jiang, R. Hogg, F. Tutu, F. Pozzi, and A. Seeds, “Long-Wavelength InAs/GaAs quantum-dot laser diode monolithically grown on Ge substrate,” Nat. Photonics 5(7), 416–419 (2011). [CrossRef]
5. Z. Mi and P. Bhattacharya, “Molecular-beam epitaxial growth and characteristics of highly uniform InAs/GaAs quantum dot layers,” J. Appl. Phys. 98(2), 023510 (2005). [CrossRef]
6. T. Yang, J. Tatebayashi, S. Tsukamoto, M. Nishioka, and Y. Arakawa, “Photoluminescence linewidth (meV) from highly uniform self-assembled InAs/GaAs quantum dots grown by low-pressure Metal Organic Chemical Vapor Deposition,” Appl. Phys. Lett. 84(15), 2817–2819 (2004). [CrossRef]
7. H. B. Lan and Y. C. Ding, “Ordering, positioning and uniformity of quantum dot arrays,” Nano Today 7(2), 94–123 (2012). [CrossRef]
8. P. Hove, B. Abbey, E. C. Le Ru, R. Murray, and T. S. Jones, “Strain-interactions between InAs/GaAs quantum dot layers,” Thin Solid Films 464–465, 225–228 (2004).
9. A. Mohan, P. Gallo, M. Felici, B. Dwir, A. Rudra, J. Faist, and E. Kapon, “Record-low inhomogeneous broadening of site-controlled quantum dots for nanophotonics,” Small 6(12), 1268–1272 (2010). [CrossRef] [PubMed]
10. P. J. Simmonds, C. D. Yerino, M. Sun, B. Liang, D. L. Huffaker, V. G. Dorogan, Y. Mazur, G. Salamo, and M. L. Lee, “Tuning quantum dot luminescence below the bulk band gap using tensile strain,” ACS Nano 7(6), 5017–5023 (2013). [CrossRef] [PubMed]
11. S. Maier, K. Berschneider, T. Steinl, A. Forchel, S. Höfling, C. Schneider, and M. Kamp, “Site-controlled InAs/GaAs quantum dots emitting at telecommunication wavelength,” Semicond. Sci. Technol. 29(5), 052001 (2014). [CrossRef]
12. P. Alonso-González, L. González, Y. González, D. Fuster, I. Fernández-Martínez, J. Martín-Sánchez, and L. Abelmann, “New process for high optical quality InAs quantum dots grown on patterned GaAs (001) substrates,” Nanotechnology 18(35), 355302 (2007). [CrossRef]
13. Yu. I. Mazur, W. Q. Ma, X. Y. Wang, Z. M. Wang, G. J. Salamo, M. Xiao, T. D. Mishima, and M. B. Johnson, “InGaAs/GaAs three-dimensionally-ordered array of quantum dots,” Appl. Phys. Lett. 83(5), 987–989 (2003). [CrossRef]
14. Z. M. Wang, K. Holmes, Yu. I. Mazur, and G. J. Salamo, “Fabrication of (In,Ga)As quantum-dot chains on GaAs(100),” Appl. Phys. Lett. 84(11), 1931–1933 (2004). [CrossRef]
15. X. Wang, Z. M. Wang, B. Liang, G. J. Salamo, and C. K. Shih, “Direct spectroscopic evidence for the formation of one-dimensional wetting wires during the growth of InGaAs/GaAs quantum dot chains,” Nano Lett. 6(9), 1847–1851 (2006). [CrossRef] [PubMed]
16. W. Q. Ma, M. L. Hussein, J. L. Shultz, G. J. Salamo, T. D. Mishima, and M. B. Johnson, “Enhancing the in-plane Spatial Ordering of Quantum Dots,” Phys. Rev. B 69(23), 233312 (2004). [CrossRef]
17. M. Schmidbauer, Sh. Seydmohamadi, D. Grigoriev, ZhM. Wang, Y. I. Mazur, P. Schäfer, M. Hanke, R. Köhler, and G. J. Salamo, “Controlling planar and vertical ordering in three-dimensional (In,Ga)As quantum dot lattices by GaAs surface orientation,” Phys. Rev. Lett. 96(6), 066108 (2006). [CrossRef] [PubMed]
18. E. Placidi, F. Arciprete, V. Latini, S. Latini, R. Magri, M. Scuderi, G. Nicotra, and F. Patella, “Manipulating surface diffusion and elastic Interactions to obtain quantum dot multilayer arrangements over different length scales,” Appl. Phys. Lett. 105(11), 111905 (2014). [CrossRef]
19. M. Ldpez and Y. Nomura, “Surface diffusion length of Ga adatoms in molecular-beam epitaxy on GaAs (100)-(110) facet structures,” J. Cryst. Growth 150, 68–72 (1995). [CrossRef]
20. K. Ohta, T. Kojima, and T. Nakagawa, “Anisotropic surface migration of Ga atoms on GaAs (001),” J. Cryst. Growth 95(1-4), 71–74 (1989). [CrossRef]
21. R. Nedzinskas, B. Cechavicius, J. Kavaliauskas, V. Karpus, G. Valusis, L. H. Li, S. P. Khanna, and E. H. Linfield, “Temperature driven three-dimensional ordering of InGaAs/GaAs quantum dot superlattices grown under As2 gas flux,” Nanoscale Res. Lett. 7, 609 (2012). [CrossRef] [PubMed]
22. M. Hata, T. Isu, A. Watanabe, and Y. Katayama, “Distributions of growth rates on patterned surfaces measured by scanning microprobe reflection high-energy electron diffraction,” J. Vac. Sci. Technol. B 8(4), 692–696 (1990). [CrossRef]
23. M. A. Salmi, M. Alatalo, T. Ala-Nissila, and R. M. Nieminen, “Energetics and diffusion paths of gallium and arsenic adatoms on flat and stepped GaAs(001) surfaces,” Surf. Sci. 425(1), 31–47 (1999). [CrossRef]
24. E. S. Tok, J. H. Neave, J. Zhang, B. A. Joyce, and T. S. Jones, “Arsenic incorporation kinetics in GaAs (001) homoepitaxy revisited,” Surf. Sci. 374(1-3), 397–405 (1997). [CrossRef]
25. T. Sugaya, T. Amano, and K. Komori, “Improved optical properties of InAs quantum dots grown with an As2source using molecular bean epitaxy,” J. Appl. Phys. 100(6), 063107 (2007). [CrossRef]
26. Zh. M. Wang, H. Churchill, C. E. George, and G. J. Salamo, “High anisotropy of lateral alignment in multilayered (In,Ga)As/GaAs(100) quantum dot structures,” J. Appl. Phys. 96(11), 2908 (2004). [CrossRef]
27. B. R. Wang, B. Q. Sun, Y. Ji, X. M. Dou, Z. Y. Xu, Zh. M. Wang, and G. J. Salamo, “Optical study of lateral carrier transfer in (In, Ga)As/GaAs quantum-dot chains,” Appl. Phys. Lett. 93(1), 011107 (2008). [CrossRef]
28. W. Liu, B. L. Liang, D. L. Huffaker, and H. Fetterman, “Anisotropic performance of high speed electro-optic modulators with InGaAs quantum dot chain active region,” Opt. Lett. 38, 4262–4264 (2013). [CrossRef] [PubMed]
29. B. L. Liang, P. S. Wong, N. Nuntawong, A. R. Albrecht, J. Tatebayashi, T. J. Rotter, G. Balakrishnan, and D. L. Huffaker, “Optical properties of patterned InAs quantum dot ensembles grown on GaAs nano-pyramids,” Appl. Phys. Lett. 91(24), 243106 (2007). [CrossRef]
30. P. S. Wong, B. L. Liang, V. G. Dorogan, A. R. Albrecht, J. Tatebayashi, X. He, N. Nuntawong, Y. I. Mazur, G. J. Salamo, S. R. J. Brueck, and D. L. Huffaker, “Improved photoluminescence efficiency of patterned quantum dots incorporating a dots-in-the-well structure,” Nanotechnology 19(43), 435710 (2008). [CrossRef] [PubMed]
31. Yu. I. Mazur, B. L. Liang, Zh. M. Wang, G. G. Tarasov, D. Guzun, and G. J. Salamo, “Development of continuum states in photoluminescence of self-assembled InGaAs/GaAs quantum dots,” J. Appl. Phys. 101(1), 014301 (2007). [CrossRef]
32. S. Kanjanachuchai, M. Xu, A. Jaffré, A. Jittrong, T. Chokamnuai, S. Panyakeow, and M. Boutchich, “Excitation transfer in stacked quantum dot chains,” Semicond. Sci. Technol. 30(5), 055005 (2015). [CrossRef]
33. Y. Wang, X. Sheng, Q. Guo, X. Li, S. Wang, G. Fu, Y. I. Mazur, Y. Maidaniuk, M. E. Ware, G. J. Salamo, B. Liang, and D. L. Huffaker, “Photoluminescence study of the interface fluctuation effect for InGaAs/InAlAs single quantum well with different thickness,” Nanoscale Res. Lett. 12(1), 229 (2017). [CrossRef] [PubMed]
34. M. Gurioli, S. Sanguinetti, E. Grilli, S. Taddei, A. Vinattieri, M. Colocci, P. Frigeri, and S. Franchi, “Disorder–induced localized states in InAs/GaAs multilayer quantum dots,” Appl. Phys. Lett. 83(11), 2262–2264 (2003). [CrossRef]
35. A. Tackeuchi, Y. Nakata, S. Muto, Y. Sugiyama, T. Usuki, Y. Nishikawa, N. Yokoyama, and O. Wada, “Time-resolved study of carrier transfer among InAs/GaAs multi-coupled quantum dots,” Jpn. J. Appl. Phys. 34(11A), L1439 (1995). [CrossRef]
36. Yu. I. Mazur, V. G. Dorogan, M. E. Ware, E. Marega, P. M. Lytvyn, Z. Y. Zhuchenko, G. G. Tarasov, and G. J. Salamo, “Effective of dimensionality and morphology on polarized photoluminescence in dot-chain structures,” J. Appl. Phys. 112, 084314 (2012). [CrossRef]
