Carrier distribution and relaxation-induced defects of InAs/GaAs quantum dots

(1)

Carrier distribution and relaxation-induced defects of InAs/GaAs quantum dots

J. S. Wang, J. F. Chen, J. L. Huang, P. Y. Wang, and X. J. Guo

Citation: Applied Physics Letters 77, 3027 (2000); doi: 10.1063/1.1323735 View online: http://dx.doi.org/10.1063/1.1323735

View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/77/19?ver=pdfcov

Published by the AIP Publishing

Articles you may be interested in

Intraband carrier dynamics in InAs/GaAs quantum dots stimulated by bound-to-continuum excitation

J. Appl. Phys. 113, 223511 (2013); 10.1063/1.4810859

Size dependent carrier thermal escape and transfer in bimodally distributed self assembled InAs/GaAs quantum dots

J. Appl. Phys. 111, 123522 (2012); 10.1063/1.4729315

Electron emission properties of relaxation-induced traps in InAs/GaAs quantum dots and the effect of electronic band structure

J. Appl. Phys. 102, 043705 (2007); 10.1063/1.2770817

Characterization of electron trap states due to InAs quantum dots in GaAs

Appl. Phys. Lett. 76, 2916 (2000); 10.1063/1.126516

Electrical characterization of partially relaxed In x Ga 1x As /GaAs multiple quantum well structures

Appl. Phys. Lett. 70, 3284 (1997); 10.1063/1.118428

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 140.113.38.11 On: Thu, 01 May 2014 07:39:42

(2)

Carrier distribution and relaxation-induced defects of InAs

Õ

GaAs

quantum dots

J. S. Wang, J. F. Chen, J. L. Huang, and P. Y. Wang

Department of Electrophysics, National Chiao Tung University, Hsinchu, Taiwan, Republic of China X. J. Guo

Department of Chemistry, National Taiwan University, Taipei 106, Taiwan, Republic of China 共Received 2 June 2000; accepted for publication 6 September 2000兲

The carrier distribution and defects have been investigated in InAs/GaAs quantum dots by cross-sectional transmission electron microscopy 共XTEM兲, capacitance–voltage, and deep level transient spectroscopy. Carrier confinement is found for 1.1- and 2.3-monolayer-共ML兲-thick InAs samples. For 2.3 ML sample, XTEM images show the presence of defect-free self-assembled quantum dots. With further increase of the InAs thickness to 3.4 ML, significant carrier depletion caused by the relaxation is observed. In contrast to 1.1 and 2.3 ML samples in which no traps are detected, two broad traps and three discrete traps at 0.54, 0.40, and 0.34 eV are observed in 3.4 ML sample. The traps at 0.54 and 0.34 eV are found to be similar to the traps observed in relaxed In0.2Ga0.8As/GaAs single quantum well structures. By comparing with the XTEM images, the trap at 0.54 eV is identified to be the relaxation-induced dislocation trap in the GaAs layer. © 2000

American Institute of Physics. 关S0003-6951共00兲02245-2兴

The In共Ga兲As/GaAs self-assembled quantum dots 共QDs兲 have attracted significant interest both from the point of view of fundamental physics and for technological applications.1–4 The QDs were obtained by the so-called Stranski–Krastanov growth mode.5 The initial growth stage of the In共Ga兲As/ GaAs heterostructure has been studied by many workers.6–8 For instance, it has been reported that when InAs coverage increases to about 1.75 monolayer共ML兲, InAs strained film will be partially relieved and the growth mode changes from two-dimensional共2D兲 to three-dimensional 共3D兲.7The 2D to 3D transition during the initial stage of growth will not in-troduce any misfit dislocation.7,9,10But as the InAs coverage increases to about 3 ML, island coalescence and defect in-corporation will occur.7,11 Although the QDs structural and optical properties have been studied by atomic force microscopy,7 transmission electron microscopy 共TEM兲,10,11 and photoluminescence 共PL兲,12 there has been no report on the detailed properties of the relaxation-induced defects. In-vestigation into this topic is necessary in order to understand the onset of relaxation and defect formation. In addition, this work is part of an ongoing investigation of the strain relax-ation in In0.2Ga0.8As/GaAs quantum wells.

13,14

The samples, with ultrathin InAs layer sandwiched by 0.3-␮m-thick n-type GaAs layers, were grown on n⫹-GaAs

共100兲 substrates by Varian Gen II molecular beam epitaxy.

The thickness of the InAs layer was varied from 1.1, 2.3 to 3.4 ML. The InAs layer was undoped and the total 0.6-␮ m-thick GaAs layers were Si doped with a concentration of 6

⫻1016_cm_⫺3 _{to allow the depletion edge to sweep across} InAs layer under reverse bias. The whole structure was grown at 550 °C without interruption, with an As4 beam equivalent pressure of 1.2⫻10⫺5Torr. The growth rates for GaAs and InAs were 0.98 and 0.23 ML/s, respectively. The growth was monitored by reflection high-energy electron dif-fraction共RHEED兲 and the QDs nucleation was observed for the 2.3 and 3.4 ML samples via the onset of a spotty RHEED

pattern. Schottky diodes were fabricated by evaporating Al on samples with a dot diameter of 1500 ␮m.

Figure 1 shows the room-temperature capacitance– voltage (C – V) data of the studied samples and their corre-sponding apparent carrier concentrations. The measuring fre-quencies were about 105Hz. Carrier-confinement peaks with intensities about 8⫻1016and 1.6⫻1017cm⫺3were observed, respectively, for 1.1 and 2.3 ML samples. When temperature was lowered, the peak intensities increased and widths nar-rowed, suggesting quantum confinement. Self-consistent simulation of C – V profile by Moon et al.15 shows that the broad C – V peak at high temperatures is attributed to the Debye averaging process between 2D and 3D electrons. As temperature is decreased, the 3D electron contribution be-comes negligible and the apparent carrier distribution is mainly determined by a change of the position expectation value of 2D electrons under the sweeping C – V measure-ment. The small width of the C – V peak is the result of a very small change of the position expectation value of 2D electrons. On the other hand, significant carrier depletion far beyond the InAs layer was observed for 3.4 ML sample. This

FIG. 1. Room temperature C – V data and the corresponding apparent carrier concentration for InAs/GaAs samples with InAs layer thickness of 1.1, 2.3, and 3.4 ML. The measuring frequencies were 105_Hz.

APPLIED PHYSICS LETTERS VOLUME 77, NUMBER 19 6 NOVEMBER 2000

3027

(3)

carrier depletion is attributed to strain relaxation when the InAs thickness increases beyond its critical thickness. This critical thickness is consistent with the previously reported value of 3 ML for island coalescence and defect incorporation.7,11Similar carrier depletion has been observed in In0.2Ga0.8As/GaAs quantum well when the In0.2Ga0.8As thickness is beyond its critical thickness.14

Figures 2共a兲 and 2共b兲 show the cross-sectional TEM

共XTEM兲 images of 2.3 and 3.4 ML samples, respectively.

Quantum dots were clearly observed in the InAs layer for 2.3 ML sample. No defects 共dislocations or stacking faults兲 ei-ther in the GaAs layers or in the InAs layer were observed. This indicates that during the initial 3D growth, self-assembled dots were formed coherently.7,9,10 On the other hand, for 3.4 ML sample, the XTEM image in Fig. 2共b兲 shows the presence of threading dislocations and stacking faults in the GaAs cap layer. It is also noted that misfit dis-locations and cross-hatched stacking faults appear near the QDs boundaries. This result indicates that islands coales-cence boundaries are favorable sites for dislocation nucleation.16 In contrast, the GaAs bottom layer is dislocation-free.

To further study their defects, all samples were mea-sured by deep-level transient spectroscopy 共DLTS兲 using a gain-phase analyzer HP4194A under dark condition. The filling pulse was set at 2 s and the measuring frequency was 105_{Hz. While no traps beyond the detection limit were} ob-served for 1.1 and 2.3 ML samples, several prominent traps were observed in 3.4 ML sample as shown in Fig. 3. In order to resolve traps from different depths, different bias voltage and filling pulse height were used: a small step voltage of

⫺0.5 V was superimposed upon a direct current bias from 0, ⫺0.5, ⫺1, ⫺1.5, ⫺2, ⫺2.5, ⫺3, ⫺3.5 to ⫺4 V. From Fig. 3,

three discrete traps E1, E2, and E5 and two broad traps E3 and E4 can be seen. They are共1兲 E1 trap, observed at ⬃300 K in Figs. 3共a兲–3共b兲; 共2兲 E2 trap, observed at ⬃275 K in Figs. 3共c兲–3共d兲; 共3兲 E3 trap, observed around 135 K in Fig. 3共e兲; 共4兲 E4 trap, observed in Figs. 3共f兲–3共g兲; and 共5兲 E5

trap, observed at ⬃235 K in Fig. 3共h兲. The corresponding Arrhenius plots of the discrete traps E1, E2, and E5 are shown in Fig. 4. The activation energies 共cross sections兲 were determined to be 0.54 eV (6.20⫻10⫺17cm2), 0.40 eV (2.19⫻10⫺18cm2), and 0.34 eV (2.48⫻10⫺18cm2) for E1, E2, and E5 traps, respectively. The E3 trap appears as a broad continuum of states in the temperature range of 80– 200 K, implying the presence of a defect band within the band gap. We speculate that this band is due to structural defects rather than point defects. Similar signals like E3 trap were observed in Figs. 3共c兲–3共d兲 from 80 to 200 K. As for the broad E4 trap, it may consist of E5 trap judging from their closeness in temperature. Due to their broadness, no reliable activation energies could be obtained for E3 and E4

FIG. 2. The XTEM images for InAs/GaAs samples with InAs thickness of

共a兲 2.3 and 共b兲 3.4 ML.

FIG. 3. The DLTS spectra for InAs/GaAs sample with InAs thickness of 3.4 ML. The measuring frequency is 105Hz and t2/t1⫽1.53 s/0.153 s.

FIG. 4. The Arrhenius plot of traps E1, E2, and E5. Other works are in-cluded for comparison, including the trap observed in In0.2Ga0.8As/GaAs

single quantum well with 1000-Å-thick InGaAs and traps from Refs. 19 and 20. The inset shows the DLTS peak amplitudes of E1 and E2 traps vs filling-pulse duration time.

3028 Appl. Phys. Lett., Vol. 77, No. 19, 6 November 2000 Wanget al.

(4)

traps. The origin of E3 and E4 traps is not clear.

We believe that all the traps detected in 3.4 ML sample are due to relaxation-induced defects rather than electron emission from the QDs as previously observed in InAs/GaAs17and InP/In0.5Ga0.5P

18

materials. The photolumi-nescence emissions 共not shown here兲 at 10 K related to the QDs were at 1.10 and 1.13 eV for 2.3 and 3.4 ML samples, respectively. Due to such small difference of 30 meV, any carrier emission from the QD states in 3.4 ML sample is likely to be observed in 2.3 ML sample. Since no DLTS signals were observed in 2.3 ML sample which contains QDs, we exclude the possibility that the trapping signals ob-served in 3.4 ML sample are the result of electron emission from the QDs. We speculate the reason why no electron emission from the QDs was observed was probably due to that the maximum measurable emission rate in our DLTS system is limited to⬃10 s⫺1.

In the E1 trap, the DLTS signals mainly come from the region close to the top of the GaAs cap layer. Comparing with the XTEM images, we assign this E1 trap to be the threading dislocation trap. We previously observed a similar trap at 0.53 eV (␴⫽1.1⫻10⫺16cm2) in relaxed In0.2Ga0.8As/GaAs single quantum well, as shown by the solid circles in Fig. 4. This result suggests that E1 trap is independent of In composition and is the relaxation-induced dislocation trap in the GaAs bulk layer. A similar trap at 0.58 eV has also been observed by Buchwald et al.19 in relaxed In0.083Ga0.917As/GaAs single heterojunction sample, further supporting this argument. Wosinski20also observed a similar trap 共ED1兲 at 0.68 eV in plastic deformed bulk GaAs. Al-though there is a slight difference in the activation energy, this trap is close to E1 trap in the Arrhenius plot in Fig. 4. Moreover, the DLTS peak amplitude of E1 trap was found to have a logarithmic dependence on the duration time of filling pulse in the range of 10⫺6– 10⫺2s, as shown in the inset of Fig. 4. This logarithmic dependence is characteristic of dislocation-related traps and is attributed to the Coulombic repulsion of the carriers captured at the traps along the lin-early arrayed dislocation lines.20

From the turning bias of ⫺1.7 V in the C – V curve as shown in Fig. 1, E2 trap seems to come mainly from the GaAs cap region below the Schottky depletion edge and the QDs. A logarithmic dependence of its DLTS peak amplitude on the duration time of filling pulse from 10⫺4 to 9

⫻10⫺2_{s was observed, as shown in the inset of Fig. 4}

sug-gesting that E2 trap is likely another kind of dislocation trap.

As for E5 trap, from the applied bias, it seems to residue near the QDs. From the XTEM images, a lot of misfit dislocations were observed near the QD boundaries. In addition, we have observed a similar relaxation-induced trap at 0.33 eV(␴

⫽1.4⫻10⫺18_cm2_{) along the InGaAs/GaAs interface in a} re-laxed In0.2Ga0.8As/GaAs sample

14

with a very thick InGaAs layer 共1000 Å兲. This highly relaxed InGaAs/GaAs sample contains a lot of misfit dislocations. Consequently, we tenta-tively assign E5 trap to be the misfit dislocation trap. More experimental data are needed in order to make a more con-clusive argument.

The authors would like to thank the National Science Council of the Republic of China for financially supporting this research under Contract No. NSC-89-2112-M-009-024.

1_{H. Drexler, D. Leonard, W. Hansen, J. P. Kotthaus, and P. M. Petroff,} Phys. Rev. Lett. 73, 2252共1994兲.

2_{M. Grundmann, O. Stier, and D. Bimberg, Phys. Rev. B 52, 11969}_共1995兲. 3

Q. Xie, A. Madhukar, P. Chen, and N. P. Kobayashi, Phys. Rev. Lett. 75, 2542共1995兲.

4_{F. Heinrichsdorff, M. H. Mao, N. Kirstaedter, A. Krost, and D. Bimberg,} Appl. Phys. Lett. 71, 22共1997兲.

5

I. N. Stranski and L. Von Krastanow, Akad. Wiss. Lit. Mainz Abh. Math. Naturwiss. Kl. llb 146, 797共1939兲.

6_{S. Guha, A. Madhukar, and K. C. Rajkumar, Appl. Phys. Lett. 57, 2110}

共1990兲.

7_{J. M. Moison, F. Houzay, F. Barthe, L. Leprince, E. Andre, and O. Vatel,} Appl. Phys. Lett. 64, 196共1994兲.

8_{D. J. Bottomley, Appl. Phys. Lett. 72, 783}_共1998兲.

9_{C. W. Snyder, B. G. Orr, D. Kessler, and L. M. Sander, Phys. Rev. Lett.} 66, 3032共1991兲.

10

D. Leonard, M. Krishnamurthy, C. M. Reaves, S. P. Denbaars, and P. M. Petroff, Appl. Phys. Lett. 63, 3203共1993兲.

11_{M. Tabuchi, S. Noda, and A. Sasaki, in Science and Technology of}

Me-soscopic Structures, edited by S. Namba, C. Hamaguchi, and T. Ando

共Springer, Tokyo 1992兲, p. 379.

12

V. Turck, F. Heinrichsdorff, M. Veit, R. Heitz, M. Grundmann, A. Krost, and D. Bimberg, Appl. Surf. Sci. 123, 352共1998兲.

13_{J. F. Chen, P. Y. Wang, C. Y. Tsai, J. S. Wang, and N. C. Chen, Appl.} Phys. Lett. 75, 2461共1999兲.

14

J. F. Chen, P. Y. Wang, J. S. Wang, and N. C. Chen, J. Appl. Phys. 87, 1369共2000兲.

15_{C. R. Moon, B. D. Choe, S. D. Kwon, and H. Lim, Appl. Phys. Lett. 72,} 1196共1998兲.

16_{S. V. Ghaisas and A. Madhukar, J. Vac. Sci. Technol. B 7, 264}_共1989兲. 17

C. M. A. Kapteyn, M. Lion, R. Heitz, D. Bimberg, P. N. Brunkov, B. V. Volovik, S. G. Konnikov, A. R. Kovsh, and V. M. Ustinov, Appl. Phys. Lett. 76, 1573共2000兲.

18_{S. Anand, N. Carlsson, M. E. Pistol, L. Samuelson, and W. Seifert, Appl.} Phys. Lett. 67, 3016共1995兲.

19

W. R. Buchwald, J. H. Zhao, M. Harmatz, and E. H. Poindexter, Solid-State Electron. 36, 1077共1993兲.

20_{T. Wosinski, J. Appl. Phys. 65, 1566}_共1989兲.

3029

Appl. Phys. Lett., Vol. 77, No. 19, 6 November 2000 Wanget al.