reprint
solidi
status
physica
pss
Phys. Status Solidi B, 1–4 (2013) / DOI 10.1002/pssb.201248573
b
www.pss-b.com
basic solid state physics
Application of external tensile and compressive strain on a single layer InAs/GaAs quantum dot via epitaxial lift-off
K. M. Omambac*, J. G. Porquez**, J. Afalla, D. Vasquez, M. H. M. Balgos, R. Jaculbia, A. S. Somintac, and A. A. Salvador Condensed Matter Physics Laboratory, National Institute of Physics, University of the Philippines, Diliman Quezon City 1101, Philippines Received 29 November 2012, revised 5 March 2013, accepted 11 March 2013 Published online 17 April 2013 Keywords epitaxial lift-off, III–V semiconductors, photoluminescence, quantum dots, strain author: e-mail
[email protected], Phone: þ63 927 982 6632, Fax: þ632 920 9749
[email protected]
* Corresponding ** e-mail
Tensile and compressive strains via epitaxial lift-off (ELO) techniques were applied on single-layer InAs/GaAs quantum dots (QDs). At low temperatures, due to the difference in thermal expansion coefficients of the ELO film and host substrate, the ELO QDs film bonded to Si and MgO substrates experienced tensile and compressive strain, respectively. At 13 K, we observed that the photoluminescence (PL) spectra of the ELO film bonded to MgO blueshifts by 10 meV while the ELO film bonded to Si redshifts by 8.5 meV with respect to the ground state of the as-grown sample. The estimated tensile and compressive strains at this temperature were determined by monitoring the valence-band splitting of the GaAs PL peak. The film bonded to Si has a light hole (lh) to heavy hole (hh) energy separation of 4.6 meV, resulting to values of strain, " ¼ 6.049 104 and stress, X ¼ 0.746 103 kbar or 74.6 MPa. On the other hand, the film bonded on MgO has an
lh–hh energy separation of 3.7 meV, giving " ¼ 4.8 104 and X ¼ 0.24 103 kbar or 24 MPa. Furthermore, we also observed a reversal of the PL intensity peak between the ground and excited-state transition of the film bonded on silicon only. A plateau-like feature between the two peaks also emerged, indicating the presence of another optical transition, which is enhanced due to application of tensile strain. We associated this peak to the 1LO-phonon replica of the PL transition resulting from the excited state. Based on these observations, this reversal is most likely attributed to the reduction of the carrier-relaxation mechanism from excited states to the ground-state transition upon the application of tensile strain. Finally, the result of this study showed the efficacy of the ELO technique as an alternative way of introducing variable tensile and compressive strain in the InAs/GaAs QD’s heterostructure.
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction InAs quantum dots (QDs) are under intensive investigation for their use in optoelectronic devices and as ideal templates to study near zero-dimensional structures [1–3]. Studies have been undertaken on molecular beam epitaxy (MBE) grown InAs QDs regarding strain effects on size and optical emission of the QDs [4, 5]. A common method used to investigate the strain effects on the dot confinement energy is to grow the InAs QDs samples on different layers from GaAs to InGaAs, or GaNAs or InGaP [6–8]. However, the lattice mismatch between the dots and the template also led to dots of varying shapes and dimensions. Thus, the explicit effect of strain on QDs of the same dimensions was difficult to extract from these
techniques. Recently, some groups reported the use of an InGaAs cap to apply strain to the InAs dots of the same shapes and sizes [9, 10]. However, only compressive strains can be applied using these previous methods. In this work, we were able to apply either tensile or compressive strain on the same set of MBE-grown InAs QDs, such that a more direct relationship of strain on dot can be determined. A thin film of InAs/GaAs QDs were bonded on either Si or MgO substrates via epitaxial lift-off (ELO) [11]. At low temperatures, the InAs QDs experience increased strain due to the difference in thermal expansion coefficients of the ELO film and host substrate [11, 12]. Similarly, another group has successfully applied tensile and ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2
solidi
status
physica
pss
b
K. M. Omambac et al.: Application of external strain on a single layer InAs/GaAs quantum dot
compressive strain by transferring GaAs membranes with embedded InGaAs QDs using a piezoelectric actuator via PMMA resist [13]. However, this method required a high voltage (1100 V) to introduce strain and the resulting shifting is only 4 meV. 2 Methodology A sample was grown through MBE consisting of 0.2 mm sacrificial AlAs, 1.5 mm GaAs buffer, InAs/GaAs QDs, and a 20-nm silicon-doped GaAs cap. The dots were grown in an epiready h100i oriented semiinsulating GaAs substrate. The dots had a growth rate of 0.136 ML s1 at 520 8C with an In:As4 ratio of 1:261. The InAs QD layer was initiated by opening the In and As shutter for 45 s followed by a 10-s As-interrupt and finally, the QDs were formed for another 10 s. Another sample, intended for atomic force microscopy (AFM) analysis, was grown using the same growth parameters but without a GaAs cap. Two 5 mm 5 mm pieces were cut from the sample. Roomtemperature (RT) and variable-temperature CW-photoluminescence (PL) spectroscopy were performed on both samples. The samples then underwent ELO [11, 12] processes and bonded to either MgO or Si substrate. During bonding of the epitaxial film (GaAs with embedded QDs) and the host substrates (Si and MgO), a water droplet was place in between, creating weak bonding utilizing the van der Waals force or surface tension forces. After the ELO processes the same PL measurements were performed on the InAs/GaAs QDs films now bonded on Si and MgO substrates. The samples were excited by a 488-nm Arþ laser with a power density set to 5 W cm2. The PL signal was then fed through a 0.5-m spectrometer mounted with an LN2-cooled Ge detector; data were gathered using standard lock-in techniques. For low-temperature measurements, the samples were mounted on the cold finger of a closed-cycle helium cryostat. A Lakeshore 340 temperature controller was used to stabilize the temperature inside the cryostat. To estimate the strain induced by the difference in thermal expansion coefficients, PL of the GaAs film were carried out at 13 K for both ELO films bonded on Si and MgO samples. An AFM tapping mode measurement was carried out on the uncapped sample using an NT-MDT Solver Pro-M with a lateral step size resolution of 12 nm and a vertical resolution of 1 nm to estimate the size and density of the dots. 3 Results and discussion This section discusses the AFM characterization (inset Fig. 1) of the uncapped sample self-assembled QDs. The average QD height and diameter is 9 and 20 nm, respectively, with average dots density of 2.6 10 cm2. The base of our grown dots is relatively large; however, it is known that the QD ground-state energy is more dependent on height rather than the lateral dimension [14, 15]. Figure 1 shows the RT-PL spectra of the as-grown and ELO-bonded samples. Films bonded to Si and MgO show similar features to that of the as-grown sample, indicating that the ELO processes did not introduce additional strain to the film at this temperature. The PL spectra of the as-grown sample show two distinct ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1 The RT (300 K) PLspectra of the as-grown sample and the samples bonded on MgO and Si. The spectra show two distinct peaks. The inset is the atomic force micrograph of the uncapped InAs/GaAs QDs. AFM measurement shows a single dot distribution with a 9-nm average height.
peaks. The two peaks are separated by 66 meV with one located at 1.016 eV and another at 1.082 eV. To determine the origin of these PL peaks, excitation intensity dependence was measured at 13 K. The excitation intensity dependence PL spectra of the single-layer InAs QDs is shown in Fig. 2. As the excitation intensity increases from 2 to 15 W cm2, an increase in the PL signal of the higher energy is observed. The increase in the PL signal of the higher energies is due to the state-filling processes in QDs [2–4]. According to state filling, as the excitation intensity increases, lower-energy states become saturated and higherenergy states will begin to populate. Therefore, we assign the peaks as transitions from the ground state (1.016 eV) and the first excited state (1.082 eV). Upon cooling down to low temperatures, the film bonded to MgO experienced compressive strain, while the film bonded to Si experienced tensile strain due to difference in the host substrates thermal expansion coefficients. A substrate that has a larger thermal expansion coefficient
Figure 2 13 K PL spectra of the as-grown sample with different excitation intensities. As the excitation intensity increases, lowerenergy states become saturated and higher-energy states will begin to populate. The spectra show that the second and third peaks are excited states. www.pss-b.com
Original Paper Phys. Status Solidi B (2013)
Figure 3 13 K PL spectra of both as-grown sample and the ELO QDs bonded on MgO and Si substrates with their corresponding energy shifting due to strain. Gaussian fits on the QDs bonded on Si reveals the presence of another transition below the first excited-state transition. The inset shows the GaAs PL peak of the as-grown and the ELO films bonded on MgO and Si with embedded InAs QDs. Estimates of strain were made by looking at the magnitude of shifting and the valence-band splitting.
than the epitaxial layer (MgO) resulted in compressive strain being induced on the layer, while the opposite occurred for a substrate that has a smaller thermal expansion coefficient (Si). Thermal expansion coefficients of Si, GaAs, and MgO are known to be 3 106, 5.8 106, and 8 106 K1, respectively. The 13 K PL spectra of the samples are shown in Fig. 3. The PL spectrum of as-grown QD sample shows the expected blueshifting of the two main peaks, now located at 1.080, 1.138 eV. A third peak also emerges at 1.202 eV, which is attributed to the second excited state. In comparison, the PL spectra of the ELO film bonded to MgO are now well resolved and shifted. The PL peak assigned to the ground state blueshifts by 10 meV, while the peaks attributed to the first and second excited states are shifted even more by 40 and 58 meV, respectively, with respect to the as-grown sample. This is expected in highly confined systems where excited energy states are more strongly affected by changes in well sizes, band offsets, and bandgaps introduced by strain. The 13 K PL of the ELO film bonded to Si shows an 8.5 meV redshifting [20, 21] of the PL signal assigned to the ground-state transition. In addition to the redshifting, a plateau-like feature between the two peaks also emerges, indicating the presence of another optical transition, which is enhanced due to application of tensile strain. Also, the relative PL intensities assigned to the ground and first excited state has reversed. Gaussian fitting applied to the PL spectra of the ELO film on Si reveals that this transition is approximately 31 meV below the transition of the first excited state. Since the dots are far apart, it is likely that this emerging PL feature is related to intradot transitions. The LO phonon, however, of InAs QDs is reported to be 29 meV, but www.pss-b.com
3
strain and confinement can increase this to 33 meV as observed by Grundmann et al. [18]. In fact, similar studies for LO phonon replica were observed by Heitz et al. [19] for dots with the ground-state transition energy around 1.10 eV. Furthermore, as discussed below, the energy separation between the excited state and the emerging optical transition does not vary significantly at different temperatures. Thus, we associate this peak to a 1LO-phonon replica of the strained InAs QDs located between the excited and groundstate transitions. The detail of the origin of this additional PL peak is under further investigation. Estimates of the strain at this temperature were made by looking at the valence-band splitting [16–21] of the GaAs PL peak. Shown in the inset of Fig. 3 are the GaAs PL spectra of the as-grown film and the film bonded on MgO and Si substrates. At 13 K, the GaAs PL peak of the film bonded on silicon redshifts by 12 meV, while the film bonded on MgO blueshifts by 2.8 meV only with respect to the unstrained GaAs peak. The biaxial tensile and compressive strain can be further divided into two components of stress. The biaxial components are responsible for the lifting of the valence band degeneracy while the hydrostatic components shift the energy transitions. The magnitude of the valence-band splitting energy [16, 21] is given to be DE1C-1HH DE1C-LH ¼ 2jbj½ðC11 þ 2C12 Þ=C11 e ðin strainÞ or DE1C-1HH DE1C-LH ¼ 2jbj½ðS11 þ S12 ÞX ðin stressÞ;
(1)
where b is the deformation potential and Cij and Sij are elastic constants. In Eq. (1) the corresponding valence-band splitting for tensile and compressive strain are 4.7 and 3.5 meV, respectively. The calculated strain and stress of the film bonded on silicon (MgO) are e(strain) ¼ 6.049 104 (4.86 104) and X(stress) of 0.746 103 kbar (0.24 103 kbar) or 74.6 MPa (24 MPa). To understand the PL intensity reversal of the QD ELO film bonded on Si, a temperature-dependent PL for the asgrown and ELO film bonded to Si were performed. Shown in Fig. 4 are the temperature-dependent PL spectra of the as-grown and ELO film bonded on Si. The as-grown film shows that as the temperature decreases from 300 to 13 K, the PL intensities associated to ground- and excited-states dots increase monotonically; however, at 160 K there is a rapid increase in the intensity of the first excited-state transition being comparable to the ground state at 50 K and remains the same to the lowest temperature (13 K). On the other hand, the temperature-dependent PL of the ELO film bonded on Si shows a different response. As temperature is decreased, the applied tensile strain on the ELO film bonded to Si increases. At 170 K, the PL peak intensity of the ground state starts to monotonically decrease as the temperature is further decreased to 13 K. At the same time there is a monotonic increase in the PL peak intensity of the excited-state ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
4
solidi
status
physica
pss
b
K. M. Omambac et al.: Application of external strain on a single layer InAs/GaAs quantum dot
application of tensile strain. Finally, the result of this study shows the efficacy of the ELO technique as an alternative way of introducing variable tensile and compressive strain in the InAs/GaAs QD’s heterostructure. Acknowledgements This work is supported in part by grants from the DOST-PCIERD/GIA, UP Systems, and UP-OVCRD. The authors would like to acknowledge Bess Singidas for AFM analysis.
References
Figure 4 (a) Temperature-dependent PL spectra of the as-grown film. (b) Temperature-dependent PL spectra of InAs QDs bonded on silicon substrate. A reversal in PL intensity was first observed at 50 K. Arrows indicate the decrease and increase of the PL intensity of the film bonded on Si with increasing tensile strain as the temperature is decreased.
transition. At 90 K, the intensity of the excited state is now comparable to the ground state and becomes the dominant feature in the PL spectrum at lower temperatures coupled with the emergence of another peak. A plateau-like feature also starts at 90 K. Gaussian fitting shows that this transition did not significantly shift its PL peak position as the temperature is decreased to 13 K from 90 K. The slight shifting, however, is attributed to the increasing tensile strain when the temperature is decreased. With the increasing tensile strain and confinement [19], the intensity of the 1LO phonon replica also increases. Based on these observations, this reversal is most likely attributed to the reduction of the carrier-relaxation mechanism from excited states to the ground-state transition upon the application of tensile strain. As tensile strain increases, carrier-relaxation mechanisms are further reduced resulting in the decrease in the ground-state PL peak intensity at the same time, followed by the increase of PL peak intensity in the excited states due to the increase in carrier density. 4 Conclusions In this paper, we have demonstrated the direct effect of tensile and compressive strain of the single-layer InAs QDs by utilizing the ELO technique. In particular, we observed that the PL spectra corresponding to the ground-state transition of the ELO film bonded to MgO (Si) blueshifts (redshifts) by as much as 10 meV (8.5 meV) with respect to the as-grown sample. Estimates for both tensile and compressive strain were made by utilizing the valence-band splitting of the GaAs PL peak. The results show that ELO technique allows us to apply strain, which can considerably shift over a wide range of energy and a stress as high as 0.746 103 kbar for tensile stress and 0.24 103 kbar for compressive stress. Furthermore, we also observed that the strain reduces the carrier-relaxation mechanism from an excited state to ground-state transition upon the ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[1] D. Bimberg, M. Grundmann, F. Heinrichsdorff, N. N. Ledentsov, V. M. Ustinov, A. E. Zhukov, A. R. Kovsh, M. V. Maximov, Y. M. Shernyakov, B. V. F. M. Volovik, A. F. Tsatsul’nikov, P. S. Kop’ev, and Zh. I. Alferov, Thin Solid Films 367, 235–249 (2007). [2] M. V. Maximov, Y. M. Shernyakov, A. F. Tsatsul’nikov, A. V. Lunev, A. V. Sakharov, V. M. Ustinov, Zh. I. Alferov, A. Yu. Egorov, A. E. Zhukov, A. R. Kovsh, P. S. Kop’ev, L. V. Asryan, and Zh. I. Alferov, J. Appl. Phys. 83, 5561 (1998). [3] P. Bhattacharya, X. H. Su, S. Chakrabarti, G. Ariyawansa, and A. G. U. Perera, Appl. Phys. Lett. 86, 191106 (2005). [4] F. Boxberg and J. Tulkki, Rep. Prog. Phys. 70, 1425–1471 (2007). [5] L. Seravalli, M. Minelli, P. Frigeri, S. Franchi, G. Guizzetti, M. Patrini, T. Ciabattoni, and M. Geddo, J. Appl. Phys. 101, 024313 (2007). [6] A. Takata, R. Oshima, Y. Shoji, and Y. Okada, J. Cryst. Growth 311(7), 1774 (2009). [7] T. Raz, N. Shuall, G. Bahir, D. Ritter, D. Gershoni, and S. N. G. Chu, Appl. Phys. Lett. 85, 3578 (2004). [8] X. Q. Zhang, S. Ganapahty, H. Kumano, K. Uesugi, and I. Suemune, J. Appl. Phys. 92, 6813 (2002). [9] S. Saravanan and T. Harayama, IEICE Electron. Exp. 5, 53 (2008). [10] S. Tonomura and K. Yamaguchi, J. Appl. Phys. 104, 054909 (2008). [11] E. Yablonovitch, J. P. Harbison, and R. Bhat, Appl. Phys. Lett. 51, 2222 (1987). [12] C. M. N. Mateo, J. J. Iban˜ez, J. G. Fernando, J. C. Garcia, K. M. Omambac, R. B. Jaculbia, M. Defensor, and A. A. Salvador, J. Appl. Phys. 104, 103537 (2008). [13] F. Ding, R. Singh, J. D. Plumhof, T. Zander, V. Krapek, Y. H. Chen, M. Benyoucef, V. Zwiller, K. Dorr, G. Bester, A. Rastelli, and O. G. Schmidt, Phys. Rev. Lett. 104, 067495 (2010). [14] K. Nishi, H. Satio, S. Sugou, and J. S. Lee, J. Appl. Phys. Lett. 74, 1111 (1999). [15] H. Satio, K. Nishi, and S. Sugou, J. Appl. Phys. Lett. 73, 2742 (1998). [16] F. H. Pollak and M. Cardona, Phys. Rev. 172, 816 (1968). [17] Y. Okada, S. Ohta, A. Kawabata, H. Shimomura, and M. Kawabe, J. Electron. Mater. 23 (1994). [18] M. Grundmann, O. Stier, and D. Bimberg, Phys. Rev. B 52, 11969 (1995). [19] R. Heitz, M. Viet, N. N. Ledentsov, A. Hoffmann, D. Bimberg, V. M. Ustinov, P. S. Kop’ev, and Zh. I. Alferov, Phys. Rev. B 56, 10435 (1996). [20] Z. Jin, S. Yang, B. Liu, M. Li, X. Wang, Z. Li, G. Du, and S. Liu, Opt. Mater. 14(3), 211 (2000). [21] C. M. N. Mateo, A. T. Garcia, F. R. M. Ramos, K. I. Manibog, and A. A. Salvador, J. Appl. Phys. 101, 073519 (2007). www.pss-b.com
