- Email: [email protected]
Confined electronic structure in GaAs quantum dots
ARTICLE IN PRESS
Journal of Luminescence 108 (2004) 379–383
Confined electronic structure in GaAs quantum dots M. Yamagiwaa,*, N. Sumitaa, F. Minamia, N. Koguchib b
a Department of Physics, Tokyo Institute of Technology, Meguro-ku, Tokyo 152-8551, Japan Nanomaterials Laboratory, National Institute for Materials Science, 1-2-1 Sengen, Tsukuba 305-0047, Japan
Abstract The energy levels of electrons and holes confined in GaAs/AlGaAs quantum dots (QDs) are studied using photoluminescence (PL) measurements under high pressure, as well as micro-PL ðmPLÞ measurements of single QDs. From the high-pressure measurements, the band lineups of the GaAs QD are determined. The GaAs QD valence-band heavy-hole ground-state-first excited-state separation is 24 meV: From the mPL measurements, the GaAs QD conduction band electron ground state-first excited state separation is 96 meV: r 2004 Elsevier B.V. All rights reserved. PACS: 78.67.Hc; 73.22.f Keywords: Confined electronic structures; GaAs; Quantum dots; Modified droplet epitaxy; Pressure dependence; Micro-PL
1. Introduction Semiconductor QDs are becoming essential materials for fabricating novel photonic and electronic devices. The appearance of self-aggregated dots, grown by Stranski–Krastanov (SK) epitaxy on a two-dimensional wetting layer has triggered an in-depth research of the physical phenomena related to the QDs. However, selfassembling of QDs with the SK method has its drawback in the necessary presence of strain for triggering the island formation in the epitaxial layer. Strain deeply modifies the electronic structure of the QD. Recently, self-assembling of strainand defect-free GaAs/AlGaAs QDs with no wetting layer have been obtained by the modified *Corresponding author. Tel.: +81-3-5734-2446; fax: +81-35734-2751. E-mail address: [email protected] (M. Yamagiwa).
droplet epitaxy (MDE) method [1]. This material is a promising candidate for investigating fundamental aspects of the physics of zero-dimensional confined structures without strain and a wetting layer. Here, we study the confined electronic structure of GaAs/AlGaAs QDs through high pressure PL measurements and single QD mPL measurements.
2. Experimental techniques This experiment consists of two parts: one is the high-pressure QD PL measurements, and the other is the single QD mPL measurements. The samples used in both experiments are grown by MDE using a Riber-32P molecular-beam epitaxy (MBE) system. The samples are grown on a GaAs ð0 0 1Þ wafer with a 300 nm thick GaAs buffer layer and a 500 nm thick AlGaAs barrier. After QD growth, a
0022-2313/$ - see front matter r 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jlumin.2004.01.080
ARTICLE IN PRESS 380
M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
100 nm thick AlGaAs barrier and 10 nm thick GaAs cap layer are grown. In the high-pressure PL experiment, a sample with an average QD base size of 15 nm and a density of 1:2 1010 cm2 is used. The PL of this sample at 4 K has a QD PL peak centered at 1:721 eV: In the mPL experiment, a lower density sample is desired in order to minimize the number of QDs in the excited/ observed region to one QD. Efforts have been made in this experiment in order to achieve a sample with a density of the order of 108 cm2 : 2.1. Sample preparation The low-density sample is grown using the following droplet formation conditions. At a substrate temperature of 330 C; the minimum Ga supply needed for dot growth (one monolayer) is supplied with a Ga flux of 0.05 monolayers/s. This Ga supply corresponds to the coverage of the As adlayer on a GaAs ð0 0 1Þ cð4 4Þ surface [2]. A sample with an average QD base size of 40 nm and a density of 7 108 cm2 is grown. The PL of this sample at 4 K has a QD PL peak centered at 1:766 eV and a full-width half-maximum (FWHM) of 109 meV: The intensity of the QD PL peak is low compared to that of the AlGaAs PL peak, and reflects its low density.
an excitation light source and the light is coupled into a single-mode fiber. The light from the fiber is collimated by a fiber coupler and focused onto the sample by a 50 objective lens to a spot size of 0:6 mm: The position of the excitation beam spot on the sample is controlled by adjusting the position of a concave lens placed between the fiber coupler and the objective lens. The PL is picked up by the same objective and coupled into a different single-mode fiber. The PL spectra from the fiber is observed using a 50 cm spectrometer with a 1800 g=mm grating and a liquid nitrogen-cooled CCD detector. The excitation power is changed by a neutral-density filter.
3. Results and discussion 3.1. High-pressure experiment The shift in the PL peak energies with pressure for bulk GaAs, AlGaAs, and GaAs QDs are shown in Fig. 1. All peaks shift to higher energies, with the bulk GaAs and GaAs QD peaks shifting back to lower energies, beyond a certain pressure, where the G-X conduction band crossover is considered to occur. The AlGaAs emission becomes too weak to detect after 7:8 kbar; reflecting
2.2. High-pressure PL experiment In the high-pressure QD PL measurements, a Merril–Bassett diamond anvil cell is used. A 1:4 ethanol–methanol compound solution serves as the hydrostatic pressure transmitting medium. The cell is directly immersed in liquid helium to obtain data at 4:2 K: A frequency-doubled YAG cw laser ð532 nm; 20 mWÞ is used as an excitation light source and the PL spectra are observed using a 30 cm spectrometer with a 600 g=mm grating and a liquid nitrogen-cooled CCD detector. 2.3. mPL experiment In the single QD mPL measurements, the sample is placed on a continuous-flow Janis ST-500 Microscopy Cryostat and cooled to 7:4 K: A frequencydoubled YAG cw laser ð532 nm; 20 mWÞ is used as
Fig. 1. Pressure dependence of the QD PL spectrum. The dotted lines are linear fits to the data. The solid lines are linear fits to the data with the exciton binding energy considered. The split QD peak after crossover is expressed by the open and closed squares.
ARTICLE IN PRESS M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
a change to an indirect recombination emission. This pressure is regarded as the band crossover pressure for AlGaAs and the pressure coefficient is 12:2 meV=kbar: This value is comparable to the reported pressure coefficient for Al0:3 Ga0:7 As; which is 9:5 meV=kbar [3]. The energies of the bulk GaAs and GaAs QD peaks deviate from a linear pressure dependence near the band crossover. This is due to several effects, including the change in exciton binding energy at high pressure and the direct/indirect recombination emission spectra overlap of the inhomogeneously broadened QD PL peak. The pressure dependence of the exciton binding energy for bulk GaAs is given as dlnðR=R0 Þ=dP ¼ 0:0083ð=kbarÞ; where the bulk GaAs exciton binding energy at ambient pressure is R0 ¼ 4:2 meV [4,5]. The exciton binding energy for GaAs and GaAs QDs before and after the band crossover can be derived using a hydrogenic model as given by R¼
e4 m : 2_2 e2
ð1Þ
Here e is the electronic charge, _ is Plank’s constant, e is the electric permittivity and m is the reduced mass given by m¼
1 1 þ : me mhh
ð2Þ
Here, only the reduced mass of the electron me and heavy hole mhh is considered, because of the non-degeneracy of the valence band in the case of quantum dots. Application of this model to QDs has been reported by Li [6]. e and m are used as parameters to calculate the exciton binding energy. Assuming the bulk GaAs crossover occurs at 30 kbar; and the GaAs QDs crossover occurs at 20 kbar; the exciton binding energy is found to be 5:4 meV for bulk GaAs before crossover, 18:3 meV for bulk GaAs after crossover, 5:3 meV for GaAs QDs before crossover and 19:7 meV for GaAs QDs after crossover. Taking these binding energies into account, the energy gap derived from the emission spectra energy is shown as the solid lines in Fig. 1. From this, the actual crossover pressure for bulk GaAs is found to be 30:0 kbar; and for GaAs QDs, 18:3 kbar: The pressure
381
coefficient is found to be 10:7 meV=kbar for bulk GaAs and 9:4 meV=kbar for GaAs QDs. The bulk GaAs pressure coefficient is in good agreement with the reported value of 10:6 meV [7]. After the band crossover, the pressure coefficient of the AlGaAs X conduction band is found to be 1:6 meV=kbar; which is also in good agreement with the reported value of 1:7 meV=kbar [3]. Fig. 2 shows the pressure dependence of the QD PL spectrum FWHM. A drop in FWHM can be seen near the crossover pressure, accompanied by a marked drop in emission intensity. The 10% difference in the pressure coefficient for bulk GaAs and GaAs QDs, as well as the drop in FWHM and emission intensity can be explained by the size distribution of the QDs; the smaller QDs at the higher energy side of the peak have a smaller pressure coefficient than the larger QDs, which have a pressure coefficient closer to that of bulk GaAs. A similar structure dependence of pressure coefficient has been reported for superlattices [8]. Near the crossover pressure, as more GaAs QD G conduction bands and the AlGaAs X conduction band cross over with pressure, the number of QDs contributing to the indirect recombination emission increases. After the band crossover, the GaAs QD PL spectrum FWHM changes from that of an inhomogeneously broadened peak to that of the sharp excitonic spectrum of an AlGaAs X conduction band electron to QD G valence band
Fig. 2. Pressure dependence of the QD PL spectrum FWHM. The line is a guide to the eye.
ARTICLE IN PRESS 382
M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
recombination. This type II recombination emission has a long radiative lifetime, and consequently has a low luminescent efficiency. This is supported by the results of time-resolved PL measurements taken at 3 and 20 kbar: At 3 kbar; an emission with a lifetime of 358 ps can be observed, while at 20 kbar; the emission decay time exceeds the upper limits of the streak camera. This is due to the weak emission intensity as well as the long relaxation time of an indirect recombination emission. At pressures higher than the GaAs QDs crossover, the GaAs QD peak splits into two peaks with the same pressure constant. The energy separation between these two peaks is 24 meV: This is considered to be the recombination emission of AlGaAs electrons with a GaAs QD heavy-hole ground state and the first excited state. From the pressure coefficients of AlGaAs, bulk GaAs and GaAs QDs, the energy shift necessary for the band crossover for bulk GaAs and GaAs QDs can be calculated. This gives the GaAs QDs conduction-band ground state confinement energy DEen¼1 and valence-band ground state confinement n¼1 energy for heavy holes DEhh at ambient pressure. The GaAs QDs valence-band excited state conn¼2 finement energy for heavy holes DEhh at ambient pressure is given by the energy separation between the two GaAs QDs peaks after crossover. Using the AlGaAs G-X conduction-band energy difference, the GaAs QDs conduction- and valenceband offsets DEC and DEV at ambient pressure can also be determined. The reconstructed band lineups of the GaAs QDs are shown in Fig. 3.
Fig. 3. Reconstructed band lineups of the GaAs QDs.
Fig. 4. Excitation intensity dependence of the mPL spectrum of a single QD.
3.2. mPL experiment The intensity dependence of the mPL spectrum of a single QD is shown in Fig. 4. The energy range is between the bulk GaAs and AlGaAs peak. The AlGaAs peak is at 1:938 eV and is not shown because of its high intensity compared to the QD peak. A sharp peak at low excitation power can be seen at 1:738 eV: As the excitation power increases, various spectral lines appear. At an excitation power of 12 mW; spectral lines from multi-excitonic energy levels can be observed. A peak can be seen at 1:834 eV: By observing the PL
spectrum at high excitation power while changing the position of the excitation/observation spot on the sample with the concave lens, it is found that the intensities of the two peaks at 1.738 and 1:834 eV changes in unison with each other, with the higher energy peak intensity being constantly lower than that of the lower energy peak. As this higher energy peak appears only at high excitation power, it is assumed that the peak at 1:738 eV is from the ground state electron–hole recombination emission from a single QD, while the peak at 1:834 eV is from the first excited state emission.
ARTICLE IN PRESS M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
383
The energy separation between these peaks is 96 meV; in accordance with calculated values [9]. A very weak peak at 1:755 eV can also be observed at the highest excitation power. This peak is at 17 meV higher energy than the main QD peak, and is close in value to the QD valence band’s heavy-hole energy level separation found by the high pressure experiment. This may be due to the recombination of the QD conduction-band ground-state electron and the valence-band excited-state hole. This recombination is caused by the coupling between optically allowed and forbidden transitions induced by Coulomb interaction between electrons and holes [10].
coefficient of the G conduction band before crossover for Al0:35 Ga0:65 As and Al0:38 Ga0:62 As is approximately 0:1 meV=kbar [12]. These differences are within the resolution of the spectrometer, and thus the same band lineups can be expected for both samples.
4. Conclusion
References
From the high-pressure experiment, the band lineups of the GaAs QDs are determined. The GaAs QD valence-band heavy-hole ground state and the excited-state energy difference is 24 meV: From the mPL experiment, the GaAs QD conduction-band electron ground state and the first excited state energy difference is found to be 96 meV: Due to the slight difference in Al concentration x in the Alx Ga1x As barrier layer for the samples in the high pressure experiment and the mPL experiment (x ¼ 0:38 for the high pressure experiment sample and x ¼ 0:35 for the mPL experiment sample), a direct comparison of the results of these experiments may not give a rigorous energy-level structure [11] for GaAs QDs. However, the difference in the pressure coefficient of the X conduction band after crossover for Al0:35 Ga0:65 As and Al0:38 Ga0:62 As is approximately 0:05 meV=kbar [8]. The difference in the pressure
[1] K. Watanabe, N. Koguchi, Y. Gotoh, Jpn. J. Appl. Phys. 39 (2000) L79. [2] A. Ohtake, J. Nakamura, S. Tsukamoto, N. Koguchi, A. Natori, Phys. Rev. Lett. 89 (2002) 206102. [3] W.P. Roach, M. Chandrasekhar, H.R. Chandrasekhar, F.A. Chambers, Phys. Rev. B 44 (1991) 13404. [4] A.R. Goni, ˜ A. Cantarero, K. Syassen, M. Cardona, Phys. Rev. B 41 (1990) 10111. [5] A.R. Goni, ˜ A. Cantarero, U. Schwarz, K. Syassen, A. Chevy, Phys. Rev. B 45 (1992) 4221. [6] G.H. Li, A.R. Goni, ˜ K. Syassen, O. Brandt, K. Ploog, Phys. Rev. B 50 (1994) 18420. [7] J. Phillips, P. Bhattacharya, U. Venkateswaran, Appl. Phys. Lett. 74 (1999) 1549. [8] J.D. Lambkin, A.R. Adams, D.J. Dunstan, P. Dawson, C.T. Foxon, Phys. Rev. B 39 (1989) 5546. [9] S. Sanguinetti, K. Watanabe, T. Kuroda, F. Minami, Y. Gotoh, N. Koguchi, J. Cryst. Growth 242 (2002) 321. [10] U. Hohenester, R. Di Felice, E. Molinari, F. Rossi, Appl. Phys. Lett. 75 (1999) 3449. [11] D.E. Aspnes, Phys. Rev. B 14 (1976) 5331. [12] N. Lifshitz, A. Jayaraman, R.A. Logan, R.G. Maines, Phys. Rev. B 20 (1979) 2398.
Acknowledgements This work was supported partially by the Grantin-Aid for Scientific Research from the Ministry of Science, Education and Culture of Japan.
Journal of Luminescence 108 (2004) 379–383
Confined electronic structure in GaAs quantum dots M. Yamagiwaa,*, N. Sumitaa, F. Minamia, N. Koguchib b
a Department of Physics, Tokyo Institute of Technology, Meguro-ku, Tokyo 152-8551, Japan Nanomaterials Laboratory, National Institute for Materials Science, 1-2-1 Sengen, Tsukuba 305-0047, Japan
Abstract The energy levels of electrons and holes confined in GaAs/AlGaAs quantum dots (QDs) are studied using photoluminescence (PL) measurements under high pressure, as well as micro-PL ðmPLÞ measurements of single QDs. From the high-pressure measurements, the band lineups of the GaAs QD are determined. The GaAs QD valence-band heavy-hole ground-state-first excited-state separation is 24 meV: From the mPL measurements, the GaAs QD conduction band electron ground state-first excited state separation is 96 meV: r 2004 Elsevier B.V. All rights reserved. PACS: 78.67.Hc; 73.22.f Keywords: Confined electronic structures; GaAs; Quantum dots; Modified droplet epitaxy; Pressure dependence; Micro-PL
1. Introduction Semiconductor QDs are becoming essential materials for fabricating novel photonic and electronic devices. The appearance of self-aggregated dots, grown by Stranski–Krastanov (SK) epitaxy on a two-dimensional wetting layer has triggered an in-depth research of the physical phenomena related to the QDs. However, selfassembling of QDs with the SK method has its drawback in the necessary presence of strain for triggering the island formation in the epitaxial layer. Strain deeply modifies the electronic structure of the QD. Recently, self-assembling of strainand defect-free GaAs/AlGaAs QDs with no wetting layer have been obtained by the modified *Corresponding author. Tel.: +81-3-5734-2446; fax: +81-35734-2751. E-mail address: [email protected] (M. Yamagiwa).
droplet epitaxy (MDE) method [1]. This material is a promising candidate for investigating fundamental aspects of the physics of zero-dimensional confined structures without strain and a wetting layer. Here, we study the confined electronic structure of GaAs/AlGaAs QDs through high pressure PL measurements and single QD mPL measurements.
2. Experimental techniques This experiment consists of two parts: one is the high-pressure QD PL measurements, and the other is the single QD mPL measurements. The samples used in both experiments are grown by MDE using a Riber-32P molecular-beam epitaxy (MBE) system. The samples are grown on a GaAs ð0 0 1Þ wafer with a 300 nm thick GaAs buffer layer and a 500 nm thick AlGaAs barrier. After QD growth, a
0022-2313/$ - see front matter r 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jlumin.2004.01.080
ARTICLE IN PRESS 380
M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
100 nm thick AlGaAs barrier and 10 nm thick GaAs cap layer are grown. In the high-pressure PL experiment, a sample with an average QD base size of 15 nm and a density of 1:2 1010 cm2 is used. The PL of this sample at 4 K has a QD PL peak centered at 1:721 eV: In the mPL experiment, a lower density sample is desired in order to minimize the number of QDs in the excited/ observed region to one QD. Efforts have been made in this experiment in order to achieve a sample with a density of the order of 108 cm2 : 2.1. Sample preparation The low-density sample is grown using the following droplet formation conditions. At a substrate temperature of 330 C; the minimum Ga supply needed for dot growth (one monolayer) is supplied with a Ga flux of 0.05 monolayers/s. This Ga supply corresponds to the coverage of the As adlayer on a GaAs ð0 0 1Þ cð4 4Þ surface [2]. A sample with an average QD base size of 40 nm and a density of 7 108 cm2 is grown. The PL of this sample at 4 K has a QD PL peak centered at 1:766 eV and a full-width half-maximum (FWHM) of 109 meV: The intensity of the QD PL peak is low compared to that of the AlGaAs PL peak, and reflects its low density.
an excitation light source and the light is coupled into a single-mode fiber. The light from the fiber is collimated by a fiber coupler and focused onto the sample by a 50 objective lens to a spot size of 0:6 mm: The position of the excitation beam spot on the sample is controlled by adjusting the position of a concave lens placed between the fiber coupler and the objective lens. The PL is picked up by the same objective and coupled into a different single-mode fiber. The PL spectra from the fiber is observed using a 50 cm spectrometer with a 1800 g=mm grating and a liquid nitrogen-cooled CCD detector. The excitation power is changed by a neutral-density filter.
3. Results and discussion 3.1. High-pressure experiment The shift in the PL peak energies with pressure for bulk GaAs, AlGaAs, and GaAs QDs are shown in Fig. 1. All peaks shift to higher energies, with the bulk GaAs and GaAs QD peaks shifting back to lower energies, beyond a certain pressure, where the G-X conduction band crossover is considered to occur. The AlGaAs emission becomes too weak to detect after 7:8 kbar; reflecting
2.2. High-pressure PL experiment In the high-pressure QD PL measurements, a Merril–Bassett diamond anvil cell is used. A 1:4 ethanol–methanol compound solution serves as the hydrostatic pressure transmitting medium. The cell is directly immersed in liquid helium to obtain data at 4:2 K: A frequency-doubled YAG cw laser ð532 nm; 20 mWÞ is used as an excitation light source and the PL spectra are observed using a 30 cm spectrometer with a 600 g=mm grating and a liquid nitrogen-cooled CCD detector. 2.3. mPL experiment In the single QD mPL measurements, the sample is placed on a continuous-flow Janis ST-500 Microscopy Cryostat and cooled to 7:4 K: A frequencydoubled YAG cw laser ð532 nm; 20 mWÞ is used as
Fig. 1. Pressure dependence of the QD PL spectrum. The dotted lines are linear fits to the data. The solid lines are linear fits to the data with the exciton binding energy considered. The split QD peak after crossover is expressed by the open and closed squares.
ARTICLE IN PRESS M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
a change to an indirect recombination emission. This pressure is regarded as the band crossover pressure for AlGaAs and the pressure coefficient is 12:2 meV=kbar: This value is comparable to the reported pressure coefficient for Al0:3 Ga0:7 As; which is 9:5 meV=kbar [3]. The energies of the bulk GaAs and GaAs QD peaks deviate from a linear pressure dependence near the band crossover. This is due to several effects, including the change in exciton binding energy at high pressure and the direct/indirect recombination emission spectra overlap of the inhomogeneously broadened QD PL peak. The pressure dependence of the exciton binding energy for bulk GaAs is given as dlnðR=R0 Þ=dP ¼ 0:0083ð=kbarÞ; where the bulk GaAs exciton binding energy at ambient pressure is R0 ¼ 4:2 meV [4,5]. The exciton binding energy for GaAs and GaAs QDs before and after the band crossover can be derived using a hydrogenic model as given by R¼
e4 m : 2_2 e2
ð1Þ
Here e is the electronic charge, _ is Plank’s constant, e is the electric permittivity and m is the reduced mass given by m¼
1 1 þ : me mhh
ð2Þ
Here, only the reduced mass of the electron me and heavy hole mhh is considered, because of the non-degeneracy of the valence band in the case of quantum dots. Application of this model to QDs has been reported by Li [6]. e and m are used as parameters to calculate the exciton binding energy. Assuming the bulk GaAs crossover occurs at 30 kbar; and the GaAs QDs crossover occurs at 20 kbar; the exciton binding energy is found to be 5:4 meV for bulk GaAs before crossover, 18:3 meV for bulk GaAs after crossover, 5:3 meV for GaAs QDs before crossover and 19:7 meV for GaAs QDs after crossover. Taking these binding energies into account, the energy gap derived from the emission spectra energy is shown as the solid lines in Fig. 1. From this, the actual crossover pressure for bulk GaAs is found to be 30:0 kbar; and for GaAs QDs, 18:3 kbar: The pressure
381
coefficient is found to be 10:7 meV=kbar for bulk GaAs and 9:4 meV=kbar for GaAs QDs. The bulk GaAs pressure coefficient is in good agreement with the reported value of 10:6 meV [7]. After the band crossover, the pressure coefficient of the AlGaAs X conduction band is found to be 1:6 meV=kbar; which is also in good agreement with the reported value of 1:7 meV=kbar [3]. Fig. 2 shows the pressure dependence of the QD PL spectrum FWHM. A drop in FWHM can be seen near the crossover pressure, accompanied by a marked drop in emission intensity. The 10% difference in the pressure coefficient for bulk GaAs and GaAs QDs, as well as the drop in FWHM and emission intensity can be explained by the size distribution of the QDs; the smaller QDs at the higher energy side of the peak have a smaller pressure coefficient than the larger QDs, which have a pressure coefficient closer to that of bulk GaAs. A similar structure dependence of pressure coefficient has been reported for superlattices [8]. Near the crossover pressure, as more GaAs QD G conduction bands and the AlGaAs X conduction band cross over with pressure, the number of QDs contributing to the indirect recombination emission increases. After the band crossover, the GaAs QD PL spectrum FWHM changes from that of an inhomogeneously broadened peak to that of the sharp excitonic spectrum of an AlGaAs X conduction band electron to QD G valence band
Fig. 2. Pressure dependence of the QD PL spectrum FWHM. The line is a guide to the eye.
ARTICLE IN PRESS 382
M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
recombination. This type II recombination emission has a long radiative lifetime, and consequently has a low luminescent efficiency. This is supported by the results of time-resolved PL measurements taken at 3 and 20 kbar: At 3 kbar; an emission with a lifetime of 358 ps can be observed, while at 20 kbar; the emission decay time exceeds the upper limits of the streak camera. This is due to the weak emission intensity as well as the long relaxation time of an indirect recombination emission. At pressures higher than the GaAs QDs crossover, the GaAs QD peak splits into two peaks with the same pressure constant. The energy separation between these two peaks is 24 meV: This is considered to be the recombination emission of AlGaAs electrons with a GaAs QD heavy-hole ground state and the first excited state. From the pressure coefficients of AlGaAs, bulk GaAs and GaAs QDs, the energy shift necessary for the band crossover for bulk GaAs and GaAs QDs can be calculated. This gives the GaAs QDs conduction-band ground state confinement energy DEen¼1 and valence-band ground state confinement n¼1 energy for heavy holes DEhh at ambient pressure. The GaAs QDs valence-band excited state conn¼2 finement energy for heavy holes DEhh at ambient pressure is given by the energy separation between the two GaAs QDs peaks after crossover. Using the AlGaAs G-X conduction-band energy difference, the GaAs QDs conduction- and valenceband offsets DEC and DEV at ambient pressure can also be determined. The reconstructed band lineups of the GaAs QDs are shown in Fig. 3.
Fig. 3. Reconstructed band lineups of the GaAs QDs.
Fig. 4. Excitation intensity dependence of the mPL spectrum of a single QD.
3.2. mPL experiment The intensity dependence of the mPL spectrum of a single QD is shown in Fig. 4. The energy range is between the bulk GaAs and AlGaAs peak. The AlGaAs peak is at 1:938 eV and is not shown because of its high intensity compared to the QD peak. A sharp peak at low excitation power can be seen at 1:738 eV: As the excitation power increases, various spectral lines appear. At an excitation power of 12 mW; spectral lines from multi-excitonic energy levels can be observed. A peak can be seen at 1:834 eV: By observing the PL
spectrum at high excitation power while changing the position of the excitation/observation spot on the sample with the concave lens, it is found that the intensities of the two peaks at 1.738 and 1:834 eV changes in unison with each other, with the higher energy peak intensity being constantly lower than that of the lower energy peak. As this higher energy peak appears only at high excitation power, it is assumed that the peak at 1:738 eV is from the ground state electron–hole recombination emission from a single QD, while the peak at 1:834 eV is from the first excited state emission.
ARTICLE IN PRESS M. Yamagiwa et al. / Journal of Luminescence 108 (2004) 379–383
383
The energy separation between these peaks is 96 meV; in accordance with calculated values [9]. A very weak peak at 1:755 eV can also be observed at the highest excitation power. This peak is at 17 meV higher energy than the main QD peak, and is close in value to the QD valence band’s heavy-hole energy level separation found by the high pressure experiment. This may be due to the recombination of the QD conduction-band ground-state electron and the valence-band excited-state hole. This recombination is caused by the coupling between optically allowed and forbidden transitions induced by Coulomb interaction between electrons and holes [10].
coefficient of the G conduction band before crossover for Al0:35 Ga0:65 As and Al0:38 Ga0:62 As is approximately 0:1 meV=kbar [12]. These differences are within the resolution of the spectrometer, and thus the same band lineups can be expected for both samples.
4. Conclusion
References
From the high-pressure experiment, the band lineups of the GaAs QDs are determined. The GaAs QD valence-band heavy-hole ground state and the excited-state energy difference is 24 meV: From the mPL experiment, the GaAs QD conduction-band electron ground state and the first excited state energy difference is found to be 96 meV: Due to the slight difference in Al concentration x in the Alx Ga1x As barrier layer for the samples in the high pressure experiment and the mPL experiment (x ¼ 0:38 for the high pressure experiment sample and x ¼ 0:35 for the mPL experiment sample), a direct comparison of the results of these experiments may not give a rigorous energy-level structure [11] for GaAs QDs. However, the difference in the pressure coefficient of the X conduction band after crossover for Al0:35 Ga0:65 As and Al0:38 Ga0:62 As is approximately 0:05 meV=kbar [8]. The difference in the pressure
[1] K. Watanabe, N. Koguchi, Y. Gotoh, Jpn. J. Appl. Phys. 39 (2000) L79. [2] A. Ohtake, J. Nakamura, S. Tsukamoto, N. Koguchi, A. Natori, Phys. Rev. Lett. 89 (2002) 206102. [3] W.P. Roach, M. Chandrasekhar, H.R. Chandrasekhar, F.A. Chambers, Phys. Rev. B 44 (1991) 13404. [4] A.R. Goni, ˜ A. Cantarero, K. Syassen, M. Cardona, Phys. Rev. B 41 (1990) 10111. [5] A.R. Goni, ˜ A. Cantarero, U. Schwarz, K. Syassen, A. Chevy, Phys. Rev. B 45 (1992) 4221. [6] G.H. Li, A.R. Goni, ˜ K. Syassen, O. Brandt, K. Ploog, Phys. Rev. B 50 (1994) 18420. [7] J. Phillips, P. Bhattacharya, U. Venkateswaran, Appl. Phys. Lett. 74 (1999) 1549. [8] J.D. Lambkin, A.R. Adams, D.J. Dunstan, P. Dawson, C.T. Foxon, Phys. Rev. B 39 (1989) 5546. [9] S. Sanguinetti, K. Watanabe, T. Kuroda, F. Minami, Y. Gotoh, N. Koguchi, J. Cryst. Growth 242 (2002) 321. [10] U. Hohenester, R. Di Felice, E. Molinari, F. Rossi, Appl. Phys. Lett. 75 (1999) 3449. [11] D.E. Aspnes, Phys. Rev. B 14 (1976) 5331. [12] N. Lifshitz, A. Jayaraman, R.A. Logan, R.G. Maines, Phys. Rev. B 20 (1979) 2398.
Acknowledgements This work was supported partially by the Grantin-Aid for Scientific Research from the Ministry of Science, Education and Culture of Japan.