Band alignment tuning of InAs quantum dots with a thin AlGaAsSb capping layer
Yu-An Liao, Wei-Ting Hsu, Shih-Han Huang, Pei-Chin Chiu, Jen-Inn Chyi, and Wen-Hao Chang
Citation: Applied Physics Letters 102, 173104 (2013); doi: 10.1063/1.4803013
View online: http://dx.doi.org/10.1063/1.4803013
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/102/17?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Effects of GaAs(Sb) cladding layers on InAs/AlAsSb quantum dots Appl. Phys. Lett. 102, 023107 (2013); 10.1063/1.4776221
Observation of band alignment transition in InAs/GaAsSb quantum dots by photoluminescence J. Appl. Phys. 111, 104302 (2012); 10.1063/1.4717766
Temperature dependent and time-resolved photoluminescence studies of InAs self-assembled quantum dots with InGaAs strain reducing layer structure
J. Appl. Phys. 106, 013512 (2009); 10.1063/1.3159648
Thermal peculiarity of AlAs-capped InAs quantum dots in a GaAs matrix J. Appl. Phys. 104, 104303 (2008); 10.1063/1.3020521
Effects of a thin InGaAs layer on InAs quantum dots embedded in InAl(Ga)As Appl. Phys. Lett. 83, 3785 (2003); 10.1063/1.1623947
Band alignment tuning of InAs quantum dots with a thin AlGaAsSb
capping layer
Yu-An Liao,1,2Wei-Ting Hsu,1Shih-Han Huang,1Pei-Chin Chiu,2Jen-Inn Chyi,2 and Wen-Hao Chang1,a)
1
Department of Electrophysics, National Chiao Tung University, Hsinchu 300, Taiwan
2
Department of Electrical Engineering, National Central University, Jhongli 320, Taiwan
(Received 19 February 2013; accepted 10 April 2013; published online 29 April 2013)
We investigate the optical properties of InAs quantum dots (QDs) capped with a thin AlxGa1xAsSb
layer. As evidenced from power-dependent and time-resolved photoluminescence (PL) measurements, the GaAsSb-capped QDs with type-II band alignment can be changed to type-I by adding Al into the GaAsSb capping layer. The evolution of band alignment with the Al content in the AlGaAsSb capping layer has also been confirmed by theoretical calculations based on 8-band k p model. The PL thermal stability and the room temperature PL efficiency are also improved by AlGaAsSb capping. We demonstrate that using the quaternary AlGaAsSb can take the advantages of GaAsSb capping layer on the InAs QDs while retaining a type-I band alignment for applications in long-wavelength light emitters.VC 2013 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4803013]
Self-assembled InAs quantum dots (QDs) with a GaAsSb capping layer (CL) have attracted much attention recently in the study of improving the performance of long-wavelength QD-based devices.1–3The major role of the GaAsSb CL is to reduce the strain inside the QDs, similar to the conventional InGaAs strain reducing layers.4The presence of Sb atoms in the CL is also helpful for suppressing the QD decomposition during overgrowth5,6and thereby preserving the island height as compared to GaAs-capped QDs. It has been shown that both effects can lead to a redshifted emission wavelength and an improved photoluminescence (PL) efficiency for Sb content less than 12%–14%. As the Sb content exceeds 14%, the InAs-GaAsSb heterointerface becomes a type-II band alignment,7,8 leading to spatially separated electrons (in the QDs) and holes (in the GaAsSb CL), and resulting in a much longer recombi-nation lifetime.9,10Although the type-II InAs/GaAsSb QDs are promising for memory11 and photovoltaic devices,12,13 the degraded recombination efficiency is however detrimental for light emitting devices. Several works have been devoted to the tailoring of the optical properties of GaAsSb-capped InAs QDs, such as varying the Sb composition in the GaAsSb CL,6,7 post-growth thermal treatments,14,15varying the GaAsSb CL thickness,16 graded Sb content in CL,17 or using quaternary GaAsNSb.18However, since the effects of strain reduction and decomposition suppression are proportional to the Sb content in the CL,6it seems unlikely to take the advantages of GaAsSb CL while retaining a type-I QD band alignment. In this con-text, replacing the GaAsSb by an AlGaAsSb CL appears to be a promising alternative. Figure1(a)shows a contour map of the unstrained valence band offset (VBO) between AlxGa1xAs1ySby and InAs [i.e., EVðAlxGa1xAs1ySbyÞ
EVðInAsÞ, where EV is the valence band maximum] as func-tions of the Al (x) and the Sb (y) contents according to the ma-terial parameters in Ref.19. If the quaternary AlGaAsSb alloy is used for capping the InAs/GaAs QDs, the band alignment can be separated into the type-I and type-II regions by the
boundary of zero VBO (solid line). Although the boundary line would be changed by the inhomogeneous stain distribution and the quantum confinement of the QDs, it is evident that the band alignment can be restored to type I by adding Al into the CL when the Sb content exceeds 0.14. Furthermore, InAs/ AlGaAsSb QDs can offer stronger electron confinement when the VBO is zero for certain Al and Sb contents in the CL, which is also preferable for the development of QD-based in-termediate-band solar cells.
In this letter, we demonstrate the tuning of band alignment and optical properties of InAs/GaAs QDs using a thin quater-nary AlGaAsSb CL. As evidenced from power dependent PL and time-resolved PL (TRPL) measurements, the GaAsSb-capped QDs with type-II band alignment can be changed to type-I by adding Al into the GaAsSb CL. The evolution of band alignment with the Al content in the CL is also compared with theoretical calculations based on 8-band k p model.
The samples were grown on GaAs substrates by molecu-lar beam epitaxy. After the growth of a 200 nm thick GaAs buffer layer on the substrate, a layer of self-assembled InAs QDs (2.7 monolayers) was deposited at 500C and subse-quently capped with a 5 nm thick AlxGa1xAs1ySby CL.
Four samples with nominal Al contents of x¼ 0, 0.1, 0.2, and 0.3 have been grown. The nominal Sb content isy¼ 0.2 for all samples. It is worth to mention that the growth rate for the AlGaAsSb layer in all samples was kept the same (500 nm/h) in order to minimize variations in the Sb incorpo-ration rate and to mitigate Sb segregations. A sample with GaAs capped InAs QDs was also grown as a reference sam-ple of type-I QDs. Finally, all samsam-ples were capped by a 50 nm GaAs layer. Atomic force microscopy revealed that uncapped surface QDs are lens shaped, with an average height of 8 nm, an average diameter of 20 nm, and an areal density of about 3 1010cm2. PL was excited by an Arþ laser (488 nm), analyzed by a 0.5 m monochromator and detected by an InGaAs photomultiplier tube. TRPL measure-ments were performed using a 50 ps pulsed laser diode (635 nm/2.5 MHz). The decay traces were recorded using the a)
Electronic mail: [email protected]
0003-6951/2013/102(17)/173104/4/$30.00 102, 173104-1 VC2013 AIP Publishing LLC
time-correlated single photon counting technique with a tem-poral resolution of150 ps.
Figure 1(c) shows the PL spectra for the samples measured at T¼ 12 K under a low excitation power of Pex¼ 10 lW. The QD PL peak blueshifts systematically
with increasing Al contentx. For the GaAsSb-capped sample (x¼ 0), the InAs-GaAsSb interface is expected to exhibit a type-II band alignment.6–8 Since the lattice constant of AlxGa1xAsSb is similar to that of GaAsSb, strain
redistribu-tion caused by adding Al into the CL can be excluded. The blueshift in PL peak can thus be attributed to the reduction in the VBO at the InAs-AlxGa1xAsSb interface.
In order to examine the band alignments, we have per-formed power dependent PL measurements. Figure2(a)shows the PL peak energy of QD ground state as a function of the cube root of excitation powerP1=3
ex . The PL peak of GaAsSb-capped QDs shows a large energy blueshift with increasingPex and exhibits a nearly linear dependence onP1=3
ex , signifying its
type-II character.2,8,10Such an energy blueshift is not observed in the type-I GaAs-capped QDs. For the AlGaAsSb-capped QDs, the PL peak blueshifts at low excitation powers but becomes nearly unchanged under higher excitation conditions. As the Al contentx in the CL is increased, the energy blueshift becomes less significant. This behavior can be explained by the gradual evolution from type-II to type-I recombination with the increasing Al contentx in the CL. Increasing Al con-tent in the CL tends to reduce the VBO at the QD-CL interface, which becomes unable to confine holes in the CL, resulting in type-I like behaviors. It is remarkable for the AlGaAsSb-capped QDs withx¼ 0.3, where the PL peak energy is nearly independent of excitation power, indicating that QD-CL inter-face has changed to a type-I band alignment.
The QD band alignments can be further investigated by TRPL measurements. Figure 2(b) shows the TRPL decay
FIG. 1. (a) The contour map of the unstrained VBO between AlxGa1xAs1ySbyand InAs as functions of the Al (x) and the Sb (y)
con-tents. (b) A schematic of band alignment for AlGaAsSb-capped InAs/GaAs QDs. (c) The PL spectra measured atT¼ 12 K for the GaAs-capped and
AlGaAsSb-capped InAs QDs with different Al contents (x). FIG. 2. (a) The ground state PL peak energies as a function of P1=3ex. (b)
Time-resolved PL spectra for the investigated samples. (c) The ground state PL peak energies as a function of Al contents (x). (d) The estimated wave function overlaps according to the measured decay lifetimes.
FIG. 3. (a)-(d) The calculated wave function distributions for the hole ground state on the (1–10) plane for different Al contents (x¼ 0; 0:1; 0:2; 0:3) in the CL. (e)–(h) The calculated band structures along the growth direction through the center of the QD (solid line, A) and through the CL near the QD base (dotted line, B).
traces of the investigated samples measured at T ¼ 12 K. The measured decay lifetime for GaAs-capped QDs in the reference sample is 0.8 ns, comparable to the typical reported value of 1 ns. In contrast, the GaAsSb-capped sample exhibits a much longer lifetime of 15.7 ns due to its type-II band alignment. Increasing Al content in the CL leads to a significant shortening in decay lifetimes, indicating that more hole wave functions penetrate into the QDs. The deduced decay lifetimes are 14.5, 6.5, and 2.2 ns for samples withx¼ 0.1, 0.2, and 0.3, respectively. Because the radiative recombination lifetime is inversely proportional to the square of the overlap integral of the electron and hole wave functions and proportional to the emission energy, the meas-ured lifetimes can thus be a measure of the electron-hole overlap in the QDs. If we assume that the overlap in the GaAs-capped type-I QDs is 100%, we estimated that the electron-hole overlaps in the AlxGa1xAsSb-capped QDs are
27%, 27%, 40%, and 70% for samples withx¼ 0, 0.1, 0.2, and 0.3, respectively.
Theoretical calculations based on eight-band kp model20
have been carried out in order to understand the evolution of hole wave function distribution with the Al contentx in the CL. We consider the InAs QD as a truncated pyramid with {101} facets and having a base length b¼ 14 nm, a height h¼ 3.5 nm, and a 5 nm AlxGa1xAs0.8Sb0.2 CL covering
thereon in a conformal way. The inhomogeneous strain distri-bution and the strain-induced piezoelectric polarization have also been included. The calculated wave function distributions for the hole ground state on the (110) plane for different Al contents in the CL are depicted in Figs. 3(a)–3(d). For the GaAsSb-capped QDs, the hole wave function is localized in the CL and close to the QD base along the [1–10] direction, where the potential is a minimum for the hole. With increasing x, the hole wave function penetrates gradually into the QD and eventually well-localized in the QD forx¼ 0:3. This can be attributed to the change of VBO by introducing Al into the CL, as can be seen from the calculated band structures along the growth direction shown in Figs.3(e)–3(h), where the potential minimum for the hole has been moved from the CL to the QD for x 0:2. The calculated transition energy and wave
function overlap as function of Al content x is displayed in Figs.2(c)and2(d). They fit very well with the measured PL energy shift and the estimated wave function overlaps. For x < 0.1, the energy shift rate follows very well with the VBO change rate (4.1 meV/Al%) due to its type-II character. When x exceeds 0.2, increasing x only leads to higher confined poten-tials for both the electron and hole, which have less effect on the transition energy of type-I QDs.
We have also performed temperature-dependent PL measurements in order to understand the thermal stability of QD emission property after AlGaAsSb capping. Figure 4 shows the Arrhenius plot of the integrated PL intensity for GaAs-capped and Al0.3Ga0.7AsSb-capped InAs QDs. For the
conventional GaAs-capped QDs, the PL intensity started to drop at T > 120 K. In contrast, the PL intensity for the Al0.3Ga0.7AsSb-capped sample can persist up toT¼ 200 K.
The thermal activation energy for PL quenching also increases from 371 to 505 meV, implicating the improved thermal stability of QD PL after AlGaAsSb capping. We would like to mention that the Al0.3Ga0.7AsSb-capped
sam-ple exhibits a large enhancement in the room temperature PL intensity (7) as compared with that of the GaAs-capped reference sample. Such an improvement in the optical prop-erties makes the AlGaAsSb capped InAs QDs very promis-ing for long wavelength applications.
In conclusion, the optical properties of AlGaAsSb-capped InAs QDs have been investigated by PL and TRPL measurements. The original type-II band alignment in GaAsSb-capped InAs QDs can be restored to type-I by add-ing Al into the CL. Furthermore, the AlGaAsSb CL also improves the PL thermal stability and the room temperature PL efficiency. We demonstrate that using a quaternary AlGaAsSb CL can take the advantages of GaAsSb CL (i.e., strain reduction and decomposition suppression) on the InAs QDs while retaining their type-I QD characters.
This work was supported in part by MOE-ATU program and the National Science Council of Taiwan under Grant No. NSC-101-2628-M-009-002-MY3.
1
J. M. Ripalda, D. Granados, Y. Gonzalez, A. M. Sanchez, S. I. Molina, and J. M. Garcıa,Appl. Phys. Lett.87, 202108 (2005).
2H. Y. Liu, M. J. Steer, T. J. Badcock, D. J. Mowbray, M. S. Skolnick, P.
Navaretti, K. M. Groom, M. Hopkinson, and R. A. Hogg,Appl. Phys. Lett.
86, 143108 (2005).
3
K. Nishikawa, Y. Takeda, K. Yamanaka, T. Motohiro, D. Sato, J. Ota, N. Miyashita, and Y. Okada,J. Appl. Phys.111, 044325 (2012).
4K. Nishi, H. Saito, S. Sugou, and J.-S. Lee, Appl. Phys. Lett.
74, 1111 (1999); V. M. Ustinov, N. A. Maleev, A. E. Zhukov, A. Yu. Egorov, A. V. Lunev, B. V. Volovik, I. L. Krestnikov, Yu. G. Musikhin, N. A. Bert, P. S. Kop’ev, Zh. I. Alferov, N. N. Ledentsov, and D. Bimberg,ibid.74, 2815 (1999); N.-T. Yeh, T.-E. Nee, J.-I. Chyi, T. M. Hsu, and C. C. Huang,ibid.
76, 1567 (2000); W.-H. Chang, H.-Y. Chen, H.-S. Chang, W.-Y. Chen, T. M. Hsu, T.-P. Hsieh, J.-I. Chyi, and N.-T. Yeh,ibid.86, 131917 (2005); P.-C. Chiu, W.-S. Liu, M.-J. Shiau, J.-I. Chyi, W.-Y. Chen, H.-S. Chang, and T.-M. Hsu,ibid.91, 153106 (2007).
5
J. M. Ulloa, I. W. D. Drouzas, P. M. Koenraad, D. J. Mowbray, M. J. Steer, H. Y. Liu, and M. Hopkinson,Appl. Phys. Lett.90, 213105 (2007).
6J. M. Ulloa, R. Gargallo-Caballero, M. Bozkurt, M. del Moral, A.
Guzman, P. M. Koenraad, and A. Hierro,Phys. Rev. B81, 165305 (2010).
7
H. Y. Liu, M. J. Steer, T. J. Badcock, D. J. Mowbray, M. S. Skolnick, F. Suarez, J. S. Ng, M. Hopkinson, and J. P. R. David,J. Appl. Phys.99, 046104 (2006).
8C. Y. Jin, H. Y. Liu, S. Y. Zhang, Q. Jiang, S. L. Liew, M. Hopkinson, T. J.
Badcock, E. Nabavi, and D. J. Mowbray,Appl. Phys. Lett.91, 021102 (2007). FIG. 4. Arrhenius plot of the integrated PL intensity for GaAs-capped and
AlGaAsSb-capped InAs QDs withx¼ 0:3. The inset shows the PL spectra measured at room temperature.
9Y. D. Jang, T. J. Badcock, D. J. Mowbray, M. S. Skolnick, J. Park, D. Lee,
H. Y. Liu, M. J. Steer, and M. Hopkinson,Appl. Phys. Lett.92, 251905 (2008).
10W.-H. Chang, Y.-A. Liao, W.-T. Hsu, M.-C. Lee, P.-C. Chiu, and J.-I.
Chyi,Appl. Phys. Lett.93, 033107 (2008).
11
A. Marent, M. Geller, A. Schliwa, D. Feise, K. P€otschke, D. Bimberg, N. Akc¸ay, and N. €Oncan,Appl. Phys. Lett.91, 242109 (2007).
12R. B. Laghumavarapu, A. Moscho, A. Khoshakhlagh, M. El-Emawy, L. F.
Lester, and D. L. Huffaker,Appl. Phys. Lett.90, 173125 (2007).
13
W.-S. Liu, H.-M. Wu, F.-H. Tsao, T.-L. Hsu, and J.-I. Chyi,Sol. Energy Mater. Sol. Cells105, 237 (2012).
14Y.-A. Liao, W.-T. Hsu, P.-C. Chiu, J.-I. Chyi, and W.-H. Chang, Appl. Phys. Lett.94, 053101 (2009).
15J. M. Ulloa, J. M. Llorens, B. Alen, D. F. Reyes, D. L. Sales, D. Gonzalez,
and A. Hierro,Appl. Phys. Lett.101, 253112 (2012).
16
W.-T. Hsu, Y.-A. Liao, F.-C. Hsu, P.-C. Chiu, J.-I. Chyi, and W.-H. Chang,Appl. Phys. Lett.99, 073108 (2011).
17A. Hospodkova, M. Zıkova, J. Pangrac, J. Oswald, K. Kuldova, J.
Vyskocˇil, and E. Hulicius,J. Cryst. Growth370, 303 (2013).
18
J. M. Ulloa, D. F. Reyes, M. Montes, K. Yamamoto, D. L. Sales, D. Gonzalez, A. Guzman, and A. Hierro, Appl. Phys. Lett. 100, 013107 (2012).
19
I. Vurgaftman, J. R. Meyer, and L. R. Ram-Mohan,J. Appl. Phys.89, 5815 (2001).
20Numerical calculations were performed using the nextnano33D
nanode-vice simulator.
