- Email: [email protected]
Al(Ga)As quantum dots structures
Applied Surface Science 260 (2012) 47–50
Contents lists available at SciVerse ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Submicron Raman and photoluminescence topography of InAs/Al(Ga)As quantum dots structures O.F. Kolomys a,∗ , V.V. Strelchuk a , T.S. Shamirzaev b , A.S. Romanyuk a , P. Tronc c a b c
V. Lashkaryov Institute of Semiconductor Physics National Academy of Sciences of Ukraine, 45 Nauky pr., 03028 Kyiv, Ukraine A.V. Rzhanov Institute of Semiconductor Physics, Siberian Branch of the Russian Academy of Sciences, RU-630090 Novosibirsk, Russia Centre National de la Recherche Scientifique, Ecole Superieure de Physique et de Chimie Industrielles de la Villede Paris, 10 rue Vauquelin, 75005 Paris, France
a r t i c l e
i n f o
Article history: Available online 23 March 2012 Keywords: Micro-Raman scattering Photoluminescence InAs quantum dots Intermixing PACS: 78.30.Fs 68.55.Ln 73.20.Mf
a b s t r a c t Two-period structures with and without vertical coupling between indirect and direct bandgap InAs quantum dots (QDs) both with type I band alignment, grown by molecular-beam epitaxy, were investigated by confocal Raman and photoluminescence (PL) microspectroscopy. The observed blue shift of PL band of the indirect (direct) bandgap QD by 20 (80) meV with decrease of thickness of Ga(Al)As intermediate layer between two InAs QD layers from 30 to 8 nm is considered as caused by increase of elastic strains (decrease of QDs sizes) in QD layers and by coupling between QDs electronic states. Scanning confocal resonant Raman microspectroscopy was applied for non-destructive evaluation of composition at various depths along the thickness of vertical coupling of the upper InAs/AlGaAs and lower InAs/AlAs QDs layers of the sandwich structures. Based on the analysis of determined from the in-depth Raman spectra optical phonons frequencies, the depth distribution of composition in InAlAs and GaAlAs alloy layers formed as a result of strain-driven enhanced interdiffusion was determined. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Heterostructures with the self-organized InAs quantum dots (QDs) grown in the Stranski–Krastanov mode have been extensively investigated because of their potential significance for device application and fundamental physical studies. Due to the unique electronic and optical properties, the Förster resonant exciton energy transfer between two different InAs QDs layers [1] and from the QDs to organic molecules [2] acted as donor and acceptors may be used for fabrication of toxic molecule sensors. At the present time the system of In(Ga)As QDs in GaAs matrix is extensively studied. The system of InAs QDs embedded in AlAs matrix has received much less attention. In comparison with InAs/GaAs QD, the InAs/AlAs QD system is characterized by the increased quantum confinement and the associated longer photoluminescence (PL) decay time and by the PL line shifted into visible spectral range [3]. It is well known that the electron ground state of InAs/AlAs QDs depending on the dot size and the aluminum content may belong to or to Xxy valley of the conductivity zone InAs [4,5]. Another feature arises in In(Ga)As/Ga(Al)As heterostructures when the two constituent compounds have different lattice parameters, namely, strong intermixing of InAs and barrier material due
∗ Corresponding author. Tel.: +380 445256240; fax: +380 445256240. E-mail address: [email protected] (O.F. Kolomys). 0169-4332/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2012.02.149
to strain-driven InAs segregation [4]. InAs–AlAs intermixing in InAs QD may be considerably reduced only by using the growth at low temperatures and by short-time growth interruptions [4]. The local inhomogeneity of InAs/AlAs distribution across InAs QDs was shown by scanning tunnel electron microscopy [6]. Embedding of Al and Ga from GaAlAs matrix into InAs QD and the lower rate of Al/In intermixing than for Ga/In was shown by Raman spectroscopy [7]. These structure inhomogeneities strongly effect electron and vibration energy structure of InAs QDs, and are of relevance for the fabrication of QD based optoelectronic devices. Despite the number of works devoted to study of composition and strains in InAs/AlAs QDs by transmission electron microscopy (TEM) [4], X-ray diffraction [8] and Raman spectroscopy (RS) [7], so far the impact of submicron scale structural, strain and compositional inhomogeneities on their optical and electronic properties remains unknown. Recently confocal micro-Raman spectroscopy was shown to allow investigations of residual stress depth distribution in strained sapphire crystal with lateral resolution of ≈300 nm and axial depth resolution of ≈600 nm [9], and the technique is non-destructive and non-contact. In the present paper, the study the spatial inhomogeneity of composition in two-period structures with self-assembled InAs QDs was carried out using scanning confocal Raman and photoluminescence microspectroscopy. To date, such optical studies of InAs/AlAs-QD with submicron spatial resolution are not available.
48
O.F. Kolomys et al. / Applied Surface Science 260 (2012) 47–50
Fig. 1. Two-period structures with coupled InAs QDs. Thickness of the separating layer between the InAs QD layers d(QD1–QD2) = 8 nm (a) and 30 nm (b).
2. Experimental Two-period structures with InAs QDs grown by molecular beam epitaxy on semi-insulating (0 0 1)-oriented GaAs substrates using Riber-32P system were investigated. The sandwich structures consisted of two vertically coupled InAs QD layers between layers of AlAs (indirect bandgap QDs (QD1)) and Al0.36 Ga0.64 As (direct bandgap QDs (QD2)) were grown on 170 nm thick GaAs buffer layer. The thickness of Al(Ga)As separating layer between the QD layers was equal to dQD1 ,QD2 = 8 and 30 nm (Fig. 1). Both QDs layers were deposited at 0.04 ML/s rate to a nominal thickness of 2.5 ML. The QDs were formed at a temperature of 510 ◦ C. A 35 nm Al0,36 Ga0,64 As layer and 50 nm GaAs cap layer were grown on top of the structure. More detailed description of the growth process was given in [10]. TEM investigations showed the InAs QDs to have lens-like shape with base diameter ∼15–20 nm and height ∼4–5 nm. Dots density was about 1010 cm−2 . Micro-Raman and PL spectra were measured at T = 10/300 K using Horiba Jobin Yvon T64000 spectrometer equipped with confocal microscope and automated piezo-driven XYZ stage. Discrete lines of Ar–Kr ion laser (exc = 488.0, 647.0 nm) and He–Cd laser (exc = 325.0 nm) with power on sample surface of 1–2 mW were used for excitation. Laser beam was focused on the sample into spot of ∼0.2–0.5 m in diameter. Spatial mapping of the optical spectra was realized by the displacement of the automated stage with spatial step of 0.1 m. 3. Results and discussion Fig. 2a and b shows low-temperature micro-PL spectra of the investigated structures with coupled InAs QDs measured at exc = 325.0 and 647.0 nm. Ultraviolet excitation is efficiently absorbed by the top direct bandgap InAs/Al0,36 Ga0,64 As QDs structure, and according to [4], these dots are responsible for low-energy exciton emission band. For the sample with dQD1,QD2 = 8 nm, inhomogeneously broadened emission line with the peak energy of 1.55–1.75 eV and full-width on half-maximum (FWHM) of 150–200 meV (QD2 PL band on Fig. 2a) corresponds to the ensemble of direct bandgap InAs QDs. The shape of this band is well approximated with two Gaussian functions with maxima at ≈1.67 eV ( = 75 meV) and ≈1.72 eV ( = 68 meV), which we attribute to bimodal QDs size distribution, since it does not disappear at low excitation density. At h ¯ ωexc = 3.81 eV, the emission intensity of the
Fig. 2. Low-temperature micro-PL of two-period structures with coupled InAs QDs. Thickness of the separating layer between InAs QD layers d(QD1–QD2) = 8 nm (a) and 30 nm (b). exc = 325.0 and 647.0 nm (¯hωexc = 3.81 and 1.92 eV, correspondingly). T = 10 K.
bottom layer of indirect bandgap InAs QDs at ≈1.82 eV (QD1 PL band on Fig. 2a) is noticeably less than the intensity of the top layer of direct bandgap InAs QDs, i.e. IQD1 < IQD2 . At h ¯ ωexc = 1.92 eV, the excitation is absorbed in the top and bottom thin direct and indirect bangap InAs QDs layers, and also in 2D AlAs wetting layer (Exy (AlAs) ≈ 2.2 eV [4]). As a result, intensity increase of high-energy PL band at ∼1.82 eV (FWHM = 45 eV) corresponding to exciton recombination in indirect bandgap InAs QDs is observed in the PL spectrum. Shape analysis of this band revealed its low-energy asymmetry to be related to radiation with emission of LO(InAs)-like and AlAs-like phonons of In(Al)As [10]. Drastically different situation has place for the sample with thickness of separating Al(Ga)As layer of dQD1 ,QD2 = 30 nm (Fig. 2b). Although, the bottom layer of InAs QDs was grown at the same conditions for both studied indirect bandgap two-period structures, increase of separating layer thickness dQD1 ,QD2 can change their energy structure. It is known [11], that increase of thickness of top barrier layer leads to relaxation of elastic strains in the QDs layer, and correspondingly, to decrease of energy of QD quantum states, and to shift of the PL line towards low-energy side. Indeed, X-ray studies have shown that for the sample with higher thickness of separating Al(Ga)As layer elastic strains of layers an order of magnitude less than for the sample with a thin layer. It follows that, the difference in the energy position of QD1 band in PL spectra of indirect bandgap InAs QDs (E ≈ 20 meV, Fig. 2) is caused by the differences in the elastic strains in the layer of two investigated structures.
O.F. Kolomys et al. / Applied Surface Science 260 (2012) 47–50
Somewhat different situation has place for direct bandgap InAs QDs of the top layer of the two-period structure. Since the unstrained thick layer has lower elastic energy, the transition from two-dimensional growth mode to three-dimensional islands occurs at large thicknesses of deposited InAs layer. Therefore, direct bandgap InAs QDs of the top layer of two-period structure grown on thick unstrained separating layer will have smaller sizes in comparison with QDs grown on thin strained separating layer. As a result, the PL band for direct-gap InAs QD2 grown on unstrained thick separating layer (Fig. 2b) will have a high-energy position, compared to sample with thin separating layer (Fig. 2a). And finally, the electronic interaction between vertically coupled InAs QDs, which holds for the sample with a thin Al(Ga)As separating layer [11], causes decrease in the energy of QD quantum states [12] and, accordingly, low-frequency shift of QD emission band. Now let us consider Raman scattering results. We applied indepth scanning method in order to obtain the distribution of composition in two-period InAs QDs structures using non-destructive technique of confocal Raman microspectroscopy. Lateral resolution of the applied technique, which is determined by the Rayleigh criterion, at excitation exc = 488.0 nm makes d = 0.61exc /NA ≈ 330 nm, and axial resolution L = 0.89exc /(NA)2 ≈ 0.53 m [13], where NA = 0.90–numeric aperture of the used microscope objective. It is important to note, that in our case the value of axial resolution is comparable with the thickness of the studied structure. However, the waist of the laser beam under in depth scanning is systematically translated through the nanometer-thick layers of two-period structure toward the substrate material. And resonant Raman scattering can occurs in each of the layers when the excitation energy corresponds to the energy of electronic transitions of impurity centers and bound excitons in epilayers, for which the gigantic oscillator strength was predicted by Rashba [14]. Such resonance phenomenon, for example, allowed investigation of resonance Raman scattering of CdSe/ZnSe nanostructures with the single inset of CdSe with nominal thickness from 0.6 to 3.15 monolayers [15]. That is why, the composition information from the internal volume of the sample could be obtained because of the optical penetration depth for used excitation greater than the overall thickness of the probed structure, and the strong enhancement of the Raman scattering from the focal spot under the resonant conditions is expected. Depth scanning confocal micro-Raman spectra for two-period structure with coupled InAs QDs with d(QD1–QD2) = 8 nm, taken under step by step scanning of focused laser spot in depth of the structure from the surface toward the substrate in normal to (1 0 0) plane direction with step of 100 nm, are presented in Fig. 3. Changing of focused laser spot position in depth of the sample, i.e. changing the excitation depth, leads to intensity redistribution and variation of frequencies of phonon lines from different layers of two-period InAs/Al(Ga)As structures which is observed in the micro-Raman spectra. Since the phonon dispersion curves of InAs, AlAs and AlGaAs optical phonons do not overlap in frequency, these phonons are spatially localized in the corresponding layers of InAs/AlAs/AlGaAs heterostructure. There is also a small possibility for backscattering Raman scattering to register the interface modes with dielectric nature, which appear only: (i) under resonance conditions of excitation, (ii) in the presence of abrupt heterointerfaces between the layers (edges and apices of pyramidal quantum dots), (iii) at existence of differences in dielectric permittivity of the structure layers [16]. TEM investigations of InAs/AlAs QDs grown at Tgrowth = 510 ◦ C showed [3] strong InAs–AlAs intermixing and blurring of heterojunction interfaces between the layers. Therefore, we believe that the appearance of interface modes in the Raman spectra of the investigated structures is unlikely.
49
Fig. 3. Depth scanning confocal micro-Raman spectra for two-period structure with coupled InAs QDs with d(QD1–QD2) = 8 nm, taken under step by step scanning (100 nm) in depth of the structure from the input of exciting spot into GaAs cap layer (curve 1) until the complete exit from the bottom GaAs buffer layer (curve 9). exc = 488.0 nm (Eexc = 2.54 eV). T = 300 K.
In order to estimate the accurate phonon frequencies, all phonon lines were first simulated using Lorentz functions. Then, LO phonon frequencies of the active layers were calculated using equations for Alx In1−x As and Alx Ga1−x As alloys [17]. For Alx In1−x As: InAs-like
:
LO = 238.6 + 20.38x − 46.02x2 , ωInAs
AlAs-like
:
LO = 332.7 + 64.3x + 7.1x2 . ωAlAs
For Alx Ga1−x As: GaAs-like
:
LO ωGaAs = 292.1 − 39.96x,
AlAs-like
:
LO = 361.7 + 55.62x − 15.45x2 . ωAlAs
Frequency positions of the phonon lines were analyzed depending on the probing depth and attributed to solid solutions of certain compositions (see Table 1). Thus, the phonon bands at ≈402.5 cm−1 and 360.8 cm−1 correspond to inelastic scattering on the LO and TO phonons of barrier AlAs layers. The observed low-frequency shift of the LO (AlAs) band compared to the bulk AlAs (403.5 cm−1 ) may be caused by tensile elastic strains. Low-frequency wing of the LO (AlAs) band at ≈396.9 cm−1 corresponds to In0.1 Al0.9 As layer which is formed at the heterointerface of QD1/AlAs containing layer. And corresponding LOInAlAs (InAs-like) phonon is observed at ≈229 cm−1 . The weaker feature appearing at 241 cm−1 is attributed to the LO phonons of the InAs QDs, similar Raman scattering by QD phonons has been observed previously [18]. The weak band at 378.4 cm−1 and 280.6 cm−1 corresponds to
50
O.F. Kolomys et al. / Applied Surface Science 260 (2012) 47–50
Table 1 Data obtained from analysis of scanning confocal micro-Raman spectroscopy for two-period structure with coupled InAs QDs. Technological parameters
GaAs Ga0.7 Al0.3 As InAs QDs Ga0.7 Al0.3 As AlAs InAs QDs – AlAs – GaAs buffer
Obtained from micro-Raman analysis d(QD1–QD2) = 8 nm
d(QD1–QD2) = 30 nm
GaAs Ga0.66 Al0.34 As InAs QDs Iny Gax Al1−x As Ga0.3 Al0.7 As InAs QDs In0.1 Al0.9 As AlAs Al0.6 Ga0.4 As GaAs buffer
GaAs Ga0,7 Al0,3 As InAs QDs Iny Gax Al1−x As Ga0.36 Al0.64 As InAs QDs
component intermixing at the heterointerfaces with formation of solid solutions. Thus, micro-Raman data for the first time with submicron spatial resolution experimentally demonstrate composition changes in depth of two-period structure with coupled InAs QDs. This work was financially supported by the Ukrainian State Program “Nanotechnologies and Nanomaterials” (project No. 3.5.2.6/6), State Agency of Science, Innovations and Informatization of Ukraine (project number M/407-2011), RFFR (project 10-0200240), and NATO CLG (Grant No. 983 878).
AlAs
References
GaAs buffer
[1] T.S. Shamirzaev, D.S. Abramkin, D.V. Dmitriev, A.K. Gutakovskii, Appl. Phys. Lett. 97 (2010) 263102. [2] T.S. Shamirzaev, D.V. Dmitriev, L.L. Sveshnikova, P. Tronc, JETP 94 (2011) 828. [3] T.S. Shamirzaev, A.M. Gilinsky, A.I. Toropov, A.K. Bakarov, D.A. Tenne, K.S. Zhuravlev, C. von Borczyskowski, D.R.T. Zahn, JETP Lett. 77 (2003) 459. [4] T.S. Shamirzaev, A.V. Nenashev, A.K. Gutakovskii, A.K. Kalagin, K.S. Zhuravlev, M. Larsson, P.O. Holtz, Phys. Rev. B 78 (2008) 085323. [5] T.S. Shamirzaev, A.V. Nenashev, K.S. Zhuravlev, Appl. Phys. Lett. 92 (2008) 213101. [6] P. Offermans, P.M. Koenraad, J.H. Wolter, K. Pierz, M. Roy, P.A. Maksym, Phys. Rev. B 72 (2005) 165332. ˜ [7] J. Ibánez, R. Cuscó, L. Artús, M. Henini, A. Patané, L. Eaves, Appl. Phys. Lett. 88 (2006) 141905. [8] R.D. Twesten, D.M. Follstaedt, S.R. Lee, E.D. Jones, J.L. Reno, J. Mirecki Millunchick, A.G. Norman, S.P. Ahrenkiel, A. Mascarenhas, Phys. Rev. B 60 (1999) 13619. [9] T. Wermelinger, C. Borgia, C. Solenthaler, R. Spolenak, Acta Mater. 55 (2007) 4657. [10] T.S. Shamirzaev, A.M. Gikinsky, A.K. Kalagin, A.V. Nenashev, K.S. Zhuravlev, Phys. Rev. B 76 (2007) 155309. [11] N.N. Ledentsov, V.M. Ustinov, V.A. Shchukin, P.S. Kop’ev, Zh I. Alferov, D. Bimberg, Semiconductors 32 (1998) 343. [12] G.S. Solomon, J.A. Trezza, A.F. Marshall, J.S. Harris Jr., Phys. Rev. Lett. 76 (1996) 952. [13] T.R. Corle, G.S. Kino, Confocal Scanning Optical Microscopy, Academic Press, London, 1996; N.J. Everall, Appl. Spectrosc. 63 (2009) 245A. [14] E.I. Rashba, FTP 8 (1974) 1241 (in Russian). [15] M.Ya. Valakh, S.V. Ivanov, N. Mestres, J. Pascual, T.V. Shubina, S.V. Sorokin, V.V. Streltchuk, G. Pozina, B. Monemar, Semicond. Sci. Technol. 17 (2002) 173. [16] Yu. A. Pusep, G. Zanelatto, S.W. da Silva, J.C. Galzerani, P.P. Gonzalez-Borrero, A.I. Toropov, P. Basmaji, Phys. Rev. B 58 (1998) R1770. [17] W.H. Weber, R. Merlin, Raman Scattering in Materials Science, Berlin, Springer, 2000, 492 pp. [18] S.-F. Ren, D. Lu, G. Qin, Phys. Rev. B 63 (2001) 195315. [19] A. Ramam, S.J. Chua, J. Appl. Phys. 80 (1996) 7157.
scattering on LOAlGaAs (AlAs-like) and LOAlGaAs (GaAs-like) phonons of the upper AlGaAs layers. Intensive peak at 268.2 cm−1 , which increases with moving of the focal point into the structure and disappears near the substrate, apparently, is responsible for the scattering on LOInAlGaAs (GaAs-like) phonons of quaternary In(Ga,Al)As compound, which is formed at the interface of the layer containing QD2/AlGaAs with low Al content [19]. The LO(GaAs) phonon band at ≈290 cm−1 corresponding to covering GaAs and thick buffer layers (Fig. 2, curves 1 and 9, correspondingly) is also present in the spectrum. The sequence of the layers in depth of the investigated structure is shown in Table 1. Carried out X-ray diffraction studies confirmed the presence of such sequence of solid solutions layers in the studied two-period structures. This result in the first approximation indicates the compensation of the low-frequency offset of LO-phonon caused by the confinement effect by the high-frequency shift of the phonon line due to the compressive strains in the layers of the investigated twoperiod structure. A detailed analysis of the composition and strains in the layers of two-period structures carried out on the basis of Xray diffraction and micro-Raman scattering data will be published in a separate paper. From the data given in Table 1 it can be seen, that for the investigated two-period structures with coupled InAs QDs, the processes of interdiffusion at the heterointerfaces reinforced by inhomogeneous distribution of elastic strains leads to
Contents lists available at SciVerse ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Submicron Raman and photoluminescence topography of InAs/Al(Ga)As quantum dots structures O.F. Kolomys a,∗ , V.V. Strelchuk a , T.S. Shamirzaev b , A.S. Romanyuk a , P. Tronc c a b c
V. Lashkaryov Institute of Semiconductor Physics National Academy of Sciences of Ukraine, 45 Nauky pr., 03028 Kyiv, Ukraine A.V. Rzhanov Institute of Semiconductor Physics, Siberian Branch of the Russian Academy of Sciences, RU-630090 Novosibirsk, Russia Centre National de la Recherche Scientifique, Ecole Superieure de Physique et de Chimie Industrielles de la Villede Paris, 10 rue Vauquelin, 75005 Paris, France
a r t i c l e
i n f o
Article history: Available online 23 March 2012 Keywords: Micro-Raman scattering Photoluminescence InAs quantum dots Intermixing PACS: 78.30.Fs 68.55.Ln 73.20.Mf
a b s t r a c t Two-period structures with and without vertical coupling between indirect and direct bandgap InAs quantum dots (QDs) both with type I band alignment, grown by molecular-beam epitaxy, were investigated by confocal Raman and photoluminescence (PL) microspectroscopy. The observed blue shift of PL band of the indirect (direct) bandgap QD by 20 (80) meV with decrease of thickness of Ga(Al)As intermediate layer between two InAs QD layers from 30 to 8 nm is considered as caused by increase of elastic strains (decrease of QDs sizes) in QD layers and by coupling between QDs electronic states. Scanning confocal resonant Raman microspectroscopy was applied for non-destructive evaluation of composition at various depths along the thickness of vertical coupling of the upper InAs/AlGaAs and lower InAs/AlAs QDs layers of the sandwich structures. Based on the analysis of determined from the in-depth Raman spectra optical phonons frequencies, the depth distribution of composition in InAlAs and GaAlAs alloy layers formed as a result of strain-driven enhanced interdiffusion was determined. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Heterostructures with the self-organized InAs quantum dots (QDs) grown in the Stranski–Krastanov mode have been extensively investigated because of their potential significance for device application and fundamental physical studies. Due to the unique electronic and optical properties, the Förster resonant exciton energy transfer between two different InAs QDs layers [1] and from the QDs to organic molecules [2] acted as donor and acceptors may be used for fabrication of toxic molecule sensors. At the present time the system of In(Ga)As QDs in GaAs matrix is extensively studied. The system of InAs QDs embedded in AlAs matrix has received much less attention. In comparison with InAs/GaAs QD, the InAs/AlAs QD system is characterized by the increased quantum confinement and the associated longer photoluminescence (PL) decay time and by the PL line shifted into visible spectral range [3]. It is well known that the electron ground state of InAs/AlAs QDs depending on the dot size and the aluminum content may belong to or to Xxy valley of the conductivity zone InAs [4,5]. Another feature arises in In(Ga)As/Ga(Al)As heterostructures when the two constituent compounds have different lattice parameters, namely, strong intermixing of InAs and barrier material due
∗ Corresponding author. Tel.: +380 445256240; fax: +380 445256240. E-mail address: [email protected] (O.F. Kolomys). 0169-4332/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2012.02.149
to strain-driven InAs segregation [4]. InAs–AlAs intermixing in InAs QD may be considerably reduced only by using the growth at low temperatures and by short-time growth interruptions [4]. The local inhomogeneity of InAs/AlAs distribution across InAs QDs was shown by scanning tunnel electron microscopy [6]. Embedding of Al and Ga from GaAlAs matrix into InAs QD and the lower rate of Al/In intermixing than for Ga/In was shown by Raman spectroscopy [7]. These structure inhomogeneities strongly effect electron and vibration energy structure of InAs QDs, and are of relevance for the fabrication of QD based optoelectronic devices. Despite the number of works devoted to study of composition and strains in InAs/AlAs QDs by transmission electron microscopy (TEM) [4], X-ray diffraction [8] and Raman spectroscopy (RS) [7], so far the impact of submicron scale structural, strain and compositional inhomogeneities on their optical and electronic properties remains unknown. Recently confocal micro-Raman spectroscopy was shown to allow investigations of residual stress depth distribution in strained sapphire crystal with lateral resolution of ≈300 nm and axial depth resolution of ≈600 nm [9], and the technique is non-destructive and non-contact. In the present paper, the study the spatial inhomogeneity of composition in two-period structures with self-assembled InAs QDs was carried out using scanning confocal Raman and photoluminescence microspectroscopy. To date, such optical studies of InAs/AlAs-QD with submicron spatial resolution are not available.
48
O.F. Kolomys et al. / Applied Surface Science 260 (2012) 47–50
Fig. 1. Two-period structures with coupled InAs QDs. Thickness of the separating layer between the InAs QD layers d(QD1–QD2) = 8 nm (a) and 30 nm (b).
2. Experimental Two-period structures with InAs QDs grown by molecular beam epitaxy on semi-insulating (0 0 1)-oriented GaAs substrates using Riber-32P system were investigated. The sandwich structures consisted of two vertically coupled InAs QD layers between layers of AlAs (indirect bandgap QDs (QD1)) and Al0.36 Ga0.64 As (direct bandgap QDs (QD2)) were grown on 170 nm thick GaAs buffer layer. The thickness of Al(Ga)As separating layer between the QD layers was equal to dQD1 ,QD2 = 8 and 30 nm (Fig. 1). Both QDs layers were deposited at 0.04 ML/s rate to a nominal thickness of 2.5 ML. The QDs were formed at a temperature of 510 ◦ C. A 35 nm Al0,36 Ga0,64 As layer and 50 nm GaAs cap layer were grown on top of the structure. More detailed description of the growth process was given in [10]. TEM investigations showed the InAs QDs to have lens-like shape with base diameter ∼15–20 nm and height ∼4–5 nm. Dots density was about 1010 cm−2 . Micro-Raman and PL spectra were measured at T = 10/300 K using Horiba Jobin Yvon T64000 spectrometer equipped with confocal microscope and automated piezo-driven XYZ stage. Discrete lines of Ar–Kr ion laser (exc = 488.0, 647.0 nm) and He–Cd laser (exc = 325.0 nm) with power on sample surface of 1–2 mW were used for excitation. Laser beam was focused on the sample into spot of ∼0.2–0.5 m in diameter. Spatial mapping of the optical spectra was realized by the displacement of the automated stage with spatial step of 0.1 m. 3. Results and discussion Fig. 2a and b shows low-temperature micro-PL spectra of the investigated structures with coupled InAs QDs measured at exc = 325.0 and 647.0 nm. Ultraviolet excitation is efficiently absorbed by the top direct bandgap InAs/Al0,36 Ga0,64 As QDs structure, and according to [4], these dots are responsible for low-energy exciton emission band. For the sample with dQD1,QD2 = 8 nm, inhomogeneously broadened emission line with the peak energy of 1.55–1.75 eV and full-width on half-maximum (FWHM) of 150–200 meV (QD2 PL band on Fig. 2a) corresponds to the ensemble of direct bandgap InAs QDs. The shape of this band is well approximated with two Gaussian functions with maxima at ≈1.67 eV ( = 75 meV) and ≈1.72 eV ( = 68 meV), which we attribute to bimodal QDs size distribution, since it does not disappear at low excitation density. At h ¯ ωexc = 3.81 eV, the emission intensity of the
Fig. 2. Low-temperature micro-PL of two-period structures with coupled InAs QDs. Thickness of the separating layer between InAs QD layers d(QD1–QD2) = 8 nm (a) and 30 nm (b). exc = 325.0 and 647.0 nm (¯hωexc = 3.81 and 1.92 eV, correspondingly). T = 10 K.
bottom layer of indirect bandgap InAs QDs at ≈1.82 eV (QD1 PL band on Fig. 2a) is noticeably less than the intensity of the top layer of direct bandgap InAs QDs, i.e. IQD1 < IQD2 . At h ¯ ωexc = 1.92 eV, the excitation is absorbed in the top and bottom thin direct and indirect bangap InAs QDs layers, and also in 2D AlAs wetting layer (Exy (AlAs) ≈ 2.2 eV [4]). As a result, intensity increase of high-energy PL band at ∼1.82 eV (FWHM = 45 eV) corresponding to exciton recombination in indirect bandgap InAs QDs is observed in the PL spectrum. Shape analysis of this band revealed its low-energy asymmetry to be related to radiation with emission of LO(InAs)-like and AlAs-like phonons of In(Al)As [10]. Drastically different situation has place for the sample with thickness of separating Al(Ga)As layer of dQD1 ,QD2 = 30 nm (Fig. 2b). Although, the bottom layer of InAs QDs was grown at the same conditions for both studied indirect bandgap two-period structures, increase of separating layer thickness dQD1 ,QD2 can change their energy structure. It is known [11], that increase of thickness of top barrier layer leads to relaxation of elastic strains in the QDs layer, and correspondingly, to decrease of energy of QD quantum states, and to shift of the PL line towards low-energy side. Indeed, X-ray studies have shown that for the sample with higher thickness of separating Al(Ga)As layer elastic strains of layers an order of magnitude less than for the sample with a thin layer. It follows that, the difference in the energy position of QD1 band in PL spectra of indirect bandgap InAs QDs (E ≈ 20 meV, Fig. 2) is caused by the differences in the elastic strains in the layer of two investigated structures.
O.F. Kolomys et al. / Applied Surface Science 260 (2012) 47–50
Somewhat different situation has place for direct bandgap InAs QDs of the top layer of the two-period structure. Since the unstrained thick layer has lower elastic energy, the transition from two-dimensional growth mode to three-dimensional islands occurs at large thicknesses of deposited InAs layer. Therefore, direct bandgap InAs QDs of the top layer of two-period structure grown on thick unstrained separating layer will have smaller sizes in comparison with QDs grown on thin strained separating layer. As a result, the PL band for direct-gap InAs QD2 grown on unstrained thick separating layer (Fig. 2b) will have a high-energy position, compared to sample with thin separating layer (Fig. 2a). And finally, the electronic interaction between vertically coupled InAs QDs, which holds for the sample with a thin Al(Ga)As separating layer [11], causes decrease in the energy of QD quantum states [12] and, accordingly, low-frequency shift of QD emission band. Now let us consider Raman scattering results. We applied indepth scanning method in order to obtain the distribution of composition in two-period InAs QDs structures using non-destructive technique of confocal Raman microspectroscopy. Lateral resolution of the applied technique, which is determined by the Rayleigh criterion, at excitation exc = 488.0 nm makes d = 0.61exc /NA ≈ 330 nm, and axial resolution L = 0.89exc /(NA)2 ≈ 0.53 m [13], where NA = 0.90–numeric aperture of the used microscope objective. It is important to note, that in our case the value of axial resolution is comparable with the thickness of the studied structure. However, the waist of the laser beam under in depth scanning is systematically translated through the nanometer-thick layers of two-period structure toward the substrate material. And resonant Raman scattering can occurs in each of the layers when the excitation energy corresponds to the energy of electronic transitions of impurity centers and bound excitons in epilayers, for which the gigantic oscillator strength was predicted by Rashba [14]. Such resonance phenomenon, for example, allowed investigation of resonance Raman scattering of CdSe/ZnSe nanostructures with the single inset of CdSe with nominal thickness from 0.6 to 3.15 monolayers [15]. That is why, the composition information from the internal volume of the sample could be obtained because of the optical penetration depth for used excitation greater than the overall thickness of the probed structure, and the strong enhancement of the Raman scattering from the focal spot under the resonant conditions is expected. Depth scanning confocal micro-Raman spectra for two-period structure with coupled InAs QDs with d(QD1–QD2) = 8 nm, taken under step by step scanning of focused laser spot in depth of the structure from the surface toward the substrate in normal to (1 0 0) plane direction with step of 100 nm, are presented in Fig. 3. Changing of focused laser spot position in depth of the sample, i.e. changing the excitation depth, leads to intensity redistribution and variation of frequencies of phonon lines from different layers of two-period InAs/Al(Ga)As structures which is observed in the micro-Raman spectra. Since the phonon dispersion curves of InAs, AlAs and AlGaAs optical phonons do not overlap in frequency, these phonons are spatially localized in the corresponding layers of InAs/AlAs/AlGaAs heterostructure. There is also a small possibility for backscattering Raman scattering to register the interface modes with dielectric nature, which appear only: (i) under resonance conditions of excitation, (ii) in the presence of abrupt heterointerfaces between the layers (edges and apices of pyramidal quantum dots), (iii) at existence of differences in dielectric permittivity of the structure layers [16]. TEM investigations of InAs/AlAs QDs grown at Tgrowth = 510 ◦ C showed [3] strong InAs–AlAs intermixing and blurring of heterojunction interfaces between the layers. Therefore, we believe that the appearance of interface modes in the Raman spectra of the investigated structures is unlikely.
49
Fig. 3. Depth scanning confocal micro-Raman spectra for two-period structure with coupled InAs QDs with d(QD1–QD2) = 8 nm, taken under step by step scanning (100 nm) in depth of the structure from the input of exciting spot into GaAs cap layer (curve 1) until the complete exit from the bottom GaAs buffer layer (curve 9). exc = 488.0 nm (Eexc = 2.54 eV). T = 300 K.
In order to estimate the accurate phonon frequencies, all phonon lines were first simulated using Lorentz functions. Then, LO phonon frequencies of the active layers were calculated using equations for Alx In1−x As and Alx Ga1−x As alloys [17]. For Alx In1−x As: InAs-like
:
LO = 238.6 + 20.38x − 46.02x2 , ωInAs
AlAs-like
:
LO = 332.7 + 64.3x + 7.1x2 . ωAlAs
For Alx Ga1−x As: GaAs-like
:
LO ωGaAs = 292.1 − 39.96x,
AlAs-like
:
LO = 361.7 + 55.62x − 15.45x2 . ωAlAs
Frequency positions of the phonon lines were analyzed depending on the probing depth and attributed to solid solutions of certain compositions (see Table 1). Thus, the phonon bands at ≈402.5 cm−1 and 360.8 cm−1 correspond to inelastic scattering on the LO and TO phonons of barrier AlAs layers. The observed low-frequency shift of the LO (AlAs) band compared to the bulk AlAs (403.5 cm−1 ) may be caused by tensile elastic strains. Low-frequency wing of the LO (AlAs) band at ≈396.9 cm−1 corresponds to In0.1 Al0.9 As layer which is formed at the heterointerface of QD1/AlAs containing layer. And corresponding LOInAlAs (InAs-like) phonon is observed at ≈229 cm−1 . The weaker feature appearing at 241 cm−1 is attributed to the LO phonons of the InAs QDs, similar Raman scattering by QD phonons has been observed previously [18]. The weak band at 378.4 cm−1 and 280.6 cm−1 corresponds to
50
O.F. Kolomys et al. / Applied Surface Science 260 (2012) 47–50
Table 1 Data obtained from analysis of scanning confocal micro-Raman spectroscopy for two-period structure with coupled InAs QDs. Technological parameters
GaAs Ga0.7 Al0.3 As InAs QDs Ga0.7 Al0.3 As AlAs InAs QDs – AlAs – GaAs buffer
Obtained from micro-Raman analysis d(QD1–QD2) = 8 nm
d(QD1–QD2) = 30 nm
GaAs Ga0.66 Al0.34 As InAs QDs Iny Gax Al1−x As Ga0.3 Al0.7 As InAs QDs In0.1 Al0.9 As AlAs Al0.6 Ga0.4 As GaAs buffer
GaAs Ga0,7 Al0,3 As InAs QDs Iny Gax Al1−x As Ga0.36 Al0.64 As InAs QDs
component intermixing at the heterointerfaces with formation of solid solutions. Thus, micro-Raman data for the first time with submicron spatial resolution experimentally demonstrate composition changes in depth of two-period structure with coupled InAs QDs. This work was financially supported by the Ukrainian State Program “Nanotechnologies and Nanomaterials” (project No. 3.5.2.6/6), State Agency of Science, Innovations and Informatization of Ukraine (project number M/407-2011), RFFR (project 10-0200240), and NATO CLG (Grant No. 983 878).
AlAs
References
GaAs buffer
[1] T.S. Shamirzaev, D.S. Abramkin, D.V. Dmitriev, A.K. Gutakovskii, Appl. Phys. Lett. 97 (2010) 263102. [2] T.S. Shamirzaev, D.V. Dmitriev, L.L. Sveshnikova, P. Tronc, JETP 94 (2011) 828. [3] T.S. Shamirzaev, A.M. Gilinsky, A.I. Toropov, A.K. Bakarov, D.A. Tenne, K.S. Zhuravlev, C. von Borczyskowski, D.R.T. Zahn, JETP Lett. 77 (2003) 459. [4] T.S. Shamirzaev, A.V. Nenashev, A.K. Gutakovskii, A.K. Kalagin, K.S. Zhuravlev, M. Larsson, P.O. Holtz, Phys. Rev. B 78 (2008) 085323. [5] T.S. Shamirzaev, A.V. Nenashev, K.S. Zhuravlev, Appl. Phys. Lett. 92 (2008) 213101. [6] P. Offermans, P.M. Koenraad, J.H. Wolter, K. Pierz, M. Roy, P.A. Maksym, Phys. Rev. B 72 (2005) 165332. ˜ [7] J. Ibánez, R. Cuscó, L. Artús, M. Henini, A. Patané, L. Eaves, Appl. Phys. Lett. 88 (2006) 141905. [8] R.D. Twesten, D.M. Follstaedt, S.R. Lee, E.D. Jones, J.L. Reno, J. Mirecki Millunchick, A.G. Norman, S.P. Ahrenkiel, A. Mascarenhas, Phys. Rev. B 60 (1999) 13619. [9] T. Wermelinger, C. Borgia, C. Solenthaler, R. Spolenak, Acta Mater. 55 (2007) 4657. [10] T.S. Shamirzaev, A.M. Gikinsky, A.K. Kalagin, A.V. Nenashev, K.S. Zhuravlev, Phys. Rev. B 76 (2007) 155309. [11] N.N. Ledentsov, V.M. Ustinov, V.A. Shchukin, P.S. Kop’ev, Zh I. Alferov, D. Bimberg, Semiconductors 32 (1998) 343. [12] G.S. Solomon, J.A. Trezza, A.F. Marshall, J.S. Harris Jr., Phys. Rev. Lett. 76 (1996) 952. [13] T.R. Corle, G.S. Kino, Confocal Scanning Optical Microscopy, Academic Press, London, 1996; N.J. Everall, Appl. Spectrosc. 63 (2009) 245A. [14] E.I. Rashba, FTP 8 (1974) 1241 (in Russian). [15] M.Ya. Valakh, S.V. Ivanov, N. Mestres, J. Pascual, T.V. Shubina, S.V. Sorokin, V.V. Streltchuk, G. Pozina, B. Monemar, Semicond. Sci. Technol. 17 (2002) 173. [16] Yu. A. Pusep, G. Zanelatto, S.W. da Silva, J.C. Galzerani, P.P. Gonzalez-Borrero, A.I. Toropov, P. Basmaji, Phys. Rev. B 58 (1998) R1770. [17] W.H. Weber, R. Merlin, Raman Scattering in Materials Science, Berlin, Springer, 2000, 492 pp. [18] S.-F. Ren, D. Lu, G. Qin, Phys. Rev. B 63 (2001) 195315. [19] A. Ramam, S.J. Chua, J. Appl. Phys. 80 (1996) 7157.
scattering on LOAlGaAs (AlAs-like) and LOAlGaAs (GaAs-like) phonons of the upper AlGaAs layers. Intensive peak at 268.2 cm−1 , which increases with moving of the focal point into the structure and disappears near the substrate, apparently, is responsible for the scattering on LOInAlGaAs (GaAs-like) phonons of quaternary In(Ga,Al)As compound, which is formed at the interface of the layer containing QD2/AlGaAs with low Al content [19]. The LO(GaAs) phonon band at ≈290 cm−1 corresponding to covering GaAs and thick buffer layers (Fig. 2, curves 1 and 9, correspondingly) is also present in the spectrum. The sequence of the layers in depth of the investigated structure is shown in Table 1. Carried out X-ray diffraction studies confirmed the presence of such sequence of solid solutions layers in the studied two-period structures. This result in the first approximation indicates the compensation of the low-frequency offset of LO-phonon caused by the confinement effect by the high-frequency shift of the phonon line due to the compressive strains in the layers of the investigated twoperiod structure. A detailed analysis of the composition and strains in the layers of two-period structures carried out on the basis of Xray diffraction and micro-Raman scattering data will be published in a separate paper. From the data given in Table 1 it can be seen, that for the investigated two-period structures with coupled InAs QDs, the processes of interdiffusion at the heterointerfaces reinforced by inhomogeneous distribution of elastic strains leads to