ARTICLE IN PRESS

1386-9477/$ - se

doi:10.1016/j.ph

�CorrespondE-mail addr

(J.I. Lee).1Also to be c

Physica E 33 (2006) 280–283

www.elsevier.com/locate/physe

Photoluminescence study on the growth of self-assembled InAsquantum dots: Formation characteristics of bimodal-sized

quantum dots

S.I. Junga, H.Y. Yeoa, I. Yuna,1, J.Y. Leemb, I.K. Hanc, J.S. Kimd, J.I. Leee,�

aDepartment of Electrical and Electronic Engineering, Yonsei University, Seoul 120-749, South KoreabSchool of Nano Engineering, Institute for Nanotechnology Applications, InJe University, Kimhae 621-749, South Korea

cNano Devices Research Center, Korea Institute of Science and Technology (KIST), Seoul 136-791, South KoreadBasic Research Laboratory, Electronics and Telecommunications Research Institute (ETRI), Daejeon 305-350, South Korea

eNano-Surface Group, Korea Research Institute of Standards and Science (KRISS), Daejeon 305-340, South Korea

Received 3 February 2006; received in revised form 20 March 2006; accepted 21 March 2006

Available online 16 May 2006

Abstract

We report a photoluminescence (PL) study on the growth process of self-assembled InAs quantum dots (QDs) under various growth

conditions. Distinctive double-peak feature was observed in the PL spectra of the QD samples grown at the relatively high substrate

temperature. From the excitation power-dependent PL and the temperature-dependent PL measurements, the double-peak feature is

associated with the ground-state transitions from InAs QDs with two different size branches. In addition, the variation in the bimodal

size distribution of the QD ensembles with different InAs coverage is demonstrated.

r 2006 Elsevier B.V. All rights reserved.

PACS: 81.07.Ta; 78.55.Cr; 78.67.Hc

Keywords: Quantum dots; InAs; Bimodal size distribution; Photoluminescence

1. Introduction

In the past several years, a quasi-zero-dimensionalstructure, especially self-assembled semiconductor quan-tum dot (QD) grown by the Stranski–Krastanow (S–K)mode has been widely studied from a fundamental physicspoint of view and for potential device applications [1–3].Even though lot of research work on the device applica-tions of QDs has been reported, the self-assembled processfor the formation of semiconductor QDs has still raisedinteresting questions on the growth dynamics of the QDs,such as shape and size distribution. For example, the QDswith bimodal size distribution were observed at certaingrowth conditions. In the previous reports, InAs/GaAs

e front matter r 2006 Elsevier B.V. All rights reserved.

yse.2006.03.150

ing author. Fax: +82 42 868-5047.

esses: iyun@yonsei.ac.kr (I. Yun), jilee@kriss.re.kr

orresponded to. Fax: +82 2 362 6444.

QDs with bimodal size distribution grown by molecularbeam epitaxy (MBE) have been demonstrated, and theiroptical and structural properties have been investigated interms of various theoretical and experimental approaches[4–7]. However, it is still necessary to understand the effectsof the growth parameters and mechanisms on the growthof bimodal-sized QDs. As a result, the control on theformation of self-assembled QDs can be a major factor fordevice applications such as an infrared photodetector [8]and a semiconductor optical amplifier (SOA) [9].In the present work, we discuss some distinctive features

associated with the bimodal size distribution of InAs/GaAsQDs grown under different growth conditions. Especially,we explain that different growth conditions result indifferent QD size distribution during the formation ofQDs by optical spectroscopy method. In general, if thegrowth temperature and/or In(Ga)As coverage are chan-ged, the size of QDs can be changed, resulting in themodification in the optical properties, especially peak

ARTICLE IN PRESS

Fig. 1. PL spectra from (a) the A sample (growth temperature of 440 1C/

InAs coverage of 2.5 ML), (b) the B sample (480 1C/2.5ML) and (c) the C

sample (480 1C/3.0ML). The dotted lines are Gaussian fitting lineshapes.

S.I. Jung et al. / Physica E 33 (2006) 280–283 281

position [10,11]. However, there was no significant changein peak position of the QD samples investigated in thiswork, and abnormal behavior of the excitation power-dependent and temperature-dependent PL was observedwhich are mostly due to two different QD branches. Wesuggest an innovative model for the growth process of QDswith bimodal size distribution.

2. Experiment

For a systematic study on the growth conditions duringthe growth of self-assembled QDs, a single-layered InAs/GaAs QDs was grown by using a Riber 32P MBE system.After the deposition of a GaAs buffer with a thickness of300 nm at a substrate temperature of 560 1C, the growthtemperature were lowered (440 and 480 1C), and then theQDs with a nominal InAs thickness of 2.5 and 3.0monolayer (ML) were, respectively, grown on the GaAsbuffer at a growth rate of 0.07ML s�1. The formation of anInAs QD was verified by the observation of the 2D–3Dtransition monitored by in situ reflection high-energyelectron diffraction pattern (RHEED). An undoped25 nm GaAs cap layer was grown after short-periodgrowth interruption of 30 s under an As-rich condition.The InAs QDs with a nominal thickness of 2.5ML weregrown at a substrate temperature of 440 and 480 1C for theA sample and the B sample, respectively. For the C sample,the InAs QDs with a nominal thickness of 3ML weregrown at a substrate temperature of 480 1C.

In PL measurements, an argon ion laser with awavelength of 514.5 nm was used as an excitation sourceto generate electron–hole pairs, and its intensity was4Wcm�2. The luminescence light from the samples wasfocused with collection lenses, dispersed by a 1m singlegrating monochromator and detected by a liquid nitrogen-cooled Ge detector.

3. Results and discussion

The PL spectra from the InAs QD samples grown atdifferent growth temperature and different nominal InAscoverage are shown in Fig. 1. The PL was measured at 16Kand an excitation power of 1mW. In Fig. 1, the PL spectrashow different shapes at relatively low excitation intensity.That is, the PL spectrum of the A sample shows only onepeak with a weak high-energy shoulder, at 1.075 eV with ashoulder near 1.160 eV, which is a typical PL feature ofself-assembled InAs QDs [12,13]. However, the PL spectraof the B and C samples show two well-separated peaks asshown in Figs. 1(b) and (c), respectively, which is mainlydue to the inhomogeneously broadened QDs with twodifferent sizes. In order to confirm the source of the double-peak features in PL spectra for the B and C samples, theexcitation power-dependent PL measurements were carriedout.

Fig. 2 shows the PL spectra of the QD samples under theexcitation power range from 1 to 100mW at 16K. There

are two peaks in PL spectra for the three QD samples atrelatively high excitation power. For the A sample, thepeak at 1.160 eV is newly observed at 5mW and itsintensity is enhanced with an increase in the excitationpower, indicating that the additional peak corresponds tothe excited states of the InAs QDs. However, the PLspectra for the B and C samples show the double-peakfeature even at an excitation intensity of 1mW. With anincrease in the excitation power, the overall intensity isenhanced and the shape is not significantly changed. Therelative intensity between a low-energy side peak and ahigh-energy side peak is almost same for the B and Csamples. This result indicates that the two peaks are largelyrelated to the bimodal size distribution of QDs.An additional behavior supporting the bimodal size

distribution of QD ensembles can be found in thetemperature-dependent PL. Fig. 3 shows the temperaturedependence of the emission peak positions for the low- andhigh-energy side peaks from the PL spectra for the A and Bsamples measured at an excitation power of 100mW. Asclearly shown in Fig. 3(a), all the curves on the low- andhigh-energy side peaks of the A sample almost follow theVarshni equation (dotted lines). However, an anomalousenergy shift has been found on the B sample (Fig. 3(b)).

ARTICLE IN PRESS

Fig. 2. Excitation power-dependent PL spectra from (a) A, (b) B, and (c)

C samples.

Fig. 3. Temperature dependence of PL for (a) A sample and (b) B sample

measured at an excitation power of 100mW. The dotted lines show the

temperature dependence of bulk InAs band gap according to the Varshni

law.

S.I. Jung et al. / Physica E 33 (2006) 280–283282

The curve on the low-energy side peak designated as theground state of large QDs follows the Varshni equationover the whole temperature range, but the high-energy sidepeak does not obey the functional form above 120K.Assuming that the low-energy side peak of the B sample isdue to the ground state of large QDs, we can say that thehigh-energy peak are composed of two contributions, onefrom the excited states of large QDs and the other from theground states of small QDs, as mentioned in Fig. 2(b). Thismay be due to the carrier redistribution by the thermallyactivated carriers between the adjacent QDs with differentsizes through the wetting layer and/or the barrier as acarrier transfer channel. We think that the digression fromthe functional curve may be additional evidence assistingthe above discussions on the bimodal QD size distribution.

Fig. 4 schematically shows the plausible diagramdemonstrating the evolution of the bimodal sized QDsunder different growth conditions. Fig. 4(a) shows agrowth process for conventionally grown QDs with a sizedistribution. At a growth temperature of 440 1C, Indesorption just begins to affect the transition duringgrowth of the InAs QDs [14]. In contrast, increasing thegrowth temperature from 440 to 480 1C with a fixed InAscoverage (2.5ML) leads to the formation of the bimodal

size distribution from one Gaussian distribution in QD sizeshown in Fig. 4(b). That is, the formation of relativelysmall QDs can be explained mainly due to the more Indesorption from the large QDs at high growth temperature[14–17]. As a result, the numbers of large and small QDsbecome changed, and the high-energy peak in the PLspectrum in Fig. 1(b) was newly observed due to theground states of small QDs. As the growth time for the QDlayer is increased, the evolution of the two QD branchesalso shows the different behavior, schematically shown inFig. 4(c). As the InAs coverage increases from 2.5 to3.0ML at a fixed growth temperature (480 1C), the numberof large QDs may be larger than that of small QDs, whichwas indirectly confirmed by PL in Fig 1(c). That is, the PLintensity of the low-energy peak in Fig. 1(c) becomesstronger and the high-energy side peak becomes relativelyweaker without significant change in the energy positions.We suppose that the result can be largely due to the over-growth effect on longer growth time [7].

4. Conclusion

In conclusion, we investigated the formation character-istics of self-assembled InAs QDs with bimodal size

ARTICLE IN PRESS

Fig. 4. A schematic diagram for the formation of bimodal-sized QDs. (a)

A growth process for the conventionally grown QDs with a size

distribution, (b) the formation of bimodal size distribution from one

Gaussian distribution of QDs, and (c) the variation of the QDs with

bimodal size distribution.

S.I. Jung et al. / Physica E 33 (2006) 280–283 283

distribution. Increasing the growth temperature and theInAs coverage leads to different growth behavior of QDs,changing the optical properties. We suggest an innovativemodel on the formation of InAs QDs with different growthconditions. That is, the formation of bimodal-sized QDscan be largely related to the In desorption at a high growthtemperature and over-growth effect for longer growth time.

References

[1] R. Leon, P.M. Petroff, D. Leonard, S. Farad, Science 267 (1995) 1966.

[2] S.I. Jung, J.J. Yoon, H.J. Park, Y.M. Park, M.H. Jeon, J.Y. Leem,

C.M. Lee, E.T. Cho, J.I. Lee, J.S. Kim, J.S. Son, J.S. Kim, D.Y. Lee,

I.K. Han, Physica E 26 (2005) 100.

[3] J.S. Kim, J.H. Lee, S.U. Hong, H.-S. Kwack, C.W. Lee, D.K. Oh,

ETRI. J. 26 (2004) 475.

[4] M. Grundmann, O. Stier, D. Bimberg, Phys. Rev. B 52 (1995)

11969.

[5] J. Ibanez, R. Leon, D.T. Vu, S. Chaparro, S.R. Johnson, C. Navarro,

Y.H. Zhang, Appl. Phys. Lett. 79 (2001) 2013.

[6] J.S. Kim, P.W. Yu, J.Y. Leem, J.I. Lee, S.K. Noh, J.S. Kim, G.H.

Kim, S.K. Kang, S.I. Ban, S.G. Kim, Y.D. Jang, U.H. Lee, J.S. Yim,

D. Lee, J. Cryst. Growth 234 (2002) 105.

[7] S.J. Lee, S.K. Noh, J.W. Choe, E.K. Kim, J. Cryst. Growth 267

(2004) 405.

[8] W.J. Choi, J.D. Song, S.H. Hwang, J.I. Lee, J.H. Kim, J.I. Song,

E.K. Kim, A. Chovet, Physica E 26 (2005) 366.

[9] M. Sugawara, T. Akiyama, N. Hatori, Y. Nakata, H. Ebe,

H. Ishikawa, Meas. Sci. Technol. 13 (2002) 1683.

[10] H.L. Wang, D. Ning, S.L. Feng, J. Cryst. Growth 209 (2000)

630.

[11] G. Saint-Girons, G. Patriarche, A. Mereuta, I. Sagnes, J. Appl. Phys.

91 (2002) 3859.

[12] D. Leonard, M. Krishnamurthy, C.M. Reaves, S.P. Denbaars,

P.M. Petroff, Appl. Phys. Lett. 63 (1993) 3203.

[13] L. Goldstein, F. Glas, J.Y. Marzin, M.N. Charasse, G. Le Roux,

Appl. Phys. Lett. 47 (1985) 1099.

[14] H. Lee, R.R. Lowe-Webb, W. Yang, P.C. Sercel, Appl. Phys. Lett. 71

(1997) 2325.

[15] W.J. Schaffer, M.D. Lind, S.P. Kowalczyk, R.W. Grant, J. Vac. Sci.

Technol. B 1 (1983) 688.

[16] C.T. Foxon, B.A. Joyce, J. Cryst. Growth 44 (1978) 75.

[17] D. Leonard, M. Krishnamurthy, S. Fafard, J.L. Merz, P.M. Petroff,

J. Vac. Sci. Technol. B 12 (1994) 1063.