Microelectronic Engineering 19 (1992) 891-894 891 Elsevier

Interface Propert ies of Strained I n G a A s / I n P Q u a n t u m Wells

Grown by L P - M O V P E

R. Schwedler, B. Gallmann, K. Wolter, A. Kohl, K. Leo, H. Kurz, S. Juillaguet~ J. Carnassel~ J. P. Laurenti~ F. H. Banmann c

Institute of Semiconductor Electronics, RWTH Aachen, FRG

~Groupe d'Etude des Semiconducteurs, Universit~ des Sciences et Techniques du Languedoc, Montpellier, France

bCentre Lorrain d'Optique et Electronique des Solides, Universit~ de Metz, France

CAT&T Bell Laboratories, Holmdel, NY, USA

Abs t rac t

We have analysed ultrathin (5-10 monolayers) Inl_xGa~As/InP (0.17 < x < 1) quan- tum wells grown by low-pressure metal organic vapour phase epitaxy using optical spec- troscopy and transmission electron microscopy. We find, for all compositions, evidence for a complex interface structure at both the lower and the upper interface. Both origi- nate from interdiffusion of arsenic and phosphorus at growth time. The influence of the interface structure on the optical transition energies is discussed in detail.

1. I N T R O D U C T I O N

Low pressure metal organic vapour phase epitaxy (LP-MOVPE) enables the growth of high-quality quantum well (QW) structures with precisely controlled well and barrier widths. A growth interruption sequence (GIS) is employed to smooth interfaces. The smoothing is confirmed by spectrally discrete transitions due to fluctuations of the well width. However, despite the necessity of the GIS to obtain good interfaces, the interface quality is far from being ideally abrupt: The GIS is known to introduce unintentional layers of InAsP and InGaAsP at the heterointerfaces, whose composition and thickness depend on the particular GIS [1-3]. Additionally, even in the case of an optimized GIS, the width of the optical transition lines corresponds to an interface roughness of the order of one monolayer (ML).

The useful lateral resolution of spectroscopic methods is ultimately limited by the ex- citon volume which corresponds to several thousand lattice cells. The transition energies observed represent averaged material properties on this scale. Sample inhomogeneities on

0167-9317/92/$05.00 © 1992 - Elsevier Science Publishers B.V. All rights reserved.

892 R. Schwedler et al. / Strained InGaAs/InP quantum wells

1.15

m 1,1 > ©

LJ 1.05 >, O~ L G) c 1

I

LLI 0.95

0.9

' I ' I ' I ' I ' I ' i . ~ 1 ' / J / ' <

.5 iVlk InAso.55Po.35 3%L/Lo~M~// //<

Variation of In L_xGaxAs O x G a ~ ~ / 5 N 1 L <

O o,

3ML ~ / v V a r i a t i o n of this^ ~_P

0.1 0.2 0.3 OA ,0.5 0.6 0.7 0.8 0.9

Ga - content XGa

Figure h PL peak energies of Inl_~Ga~As/InP MQWs (nominal well thickness Lz = 5 ML) for various Ga-contents of the QW regions. The solid lines have been calculated assuming 4 MLs of Inl_~Ga~As and 4 additional MLs of InAs0.65P0.3s. Well width fluctuations of =k 1 ML have been taken into account for both the Inl_~Ga~As and the InAs0.65P0.35.

length scales smaller than the excitation spot size may also inhibit unambiguous inter- pretation of the optical spectra. Thus, supplementary microscopic techniques are needed to describe in detail the formation of the interfaces. Here, high resolution transmission electron microscopy (HRTEM) is a powerful tool due to its ability to resolve single atomic layers in a plane perpendicular to the QWs [4].

2 . E X P E R I M E N T A L

The samples analysed are strained azld unstrained Inl_~Ga=As/InP multiple quantum wells (MQW) grown on semi-insulating InP substrates by L P - M O V P E at 640°C. The Gallium content has been caxefully calibrated by sputtered neutrals mass spectroscopy [1,5]. In this work, we will discuss two sets of samples: i) a series of MQWs (10 periods) with Gallium contents 0.17 < xa= _< 1.00 and a nominal well thickness of 5 monolayers (MLs) is used for optical characterization and it) two samples with a nominal well width of 10 MLs (an= = 0.47 and xG= = 0.73, respectively) axe used for HRTEM microscopy.

HRTEM pictures were taken in chemical lattice imaging (CLI) mode. Details of this method are described in ref. 4. Colour encoded pictures have been generated by matching the observed lattice patterns to patterns of undisturbed InP and Inl_~Ga=As obtained from the same sample. Because of this encoding process, the CLI pictures give information on intermediate compositions of interface layers composed of InP and Inl_~Ga~As. They do not contain, e.g. quaxltitative information on possible layers of InAsl_~P~.

R. Schwedler et al. / Strained InGaAs/InP quantum wells 893

1n.27 Ga.73 As in InP In 27Ga 7:~As

1. well

Figure 2: HRTEM picture of an Ino.27Gao.~3As/InP QW with a nominal well thickness L, of 10 ML.

3. R E S U L T S A N D D I S C U S S I O N

The PL peak positions obtained from the complete set of samples are summarized in Fig. 1. We find a monotonous dependence of the PL peak energies on Ga-content. Noteworthy is the almost constant line splitting over the whole range of compositions.

Details of the electronic states of the Inx_=Ga=As/InP QWs have been obtained by transfer matrix calculations [3]. Calculations based on the assumption of ideal InGaAs/- InP QWs completely fail to explain the experimental data. These calculations predict i) a much stronger increase of the E H I energy with increasing xca, and if) a decreasing line splitting when xa~ approaches unity [6]. A significant improvement is achieved by assuming that, in all structures, an average layer of about 4 MLs of InAs0.6sP0.35 is present (Fig 1). Moreover, if fluctuations of the widths of both the Inl_=Ga=As as well as the InAs0.6sP0.35 layers of one ML are taken into account, the same model explains the line splittings.

However, from the growth process, we expect formation of a more complicated inter- face structure which is caused by intermixing at both interfaces. This has been checked by HRTEM (Fig. 2). Most striking feature is a large asymmetry of the interface structure. The lower interfaces (InP to Ina_=Ga=As) are rather sharp and show little lateral rough- ness, whereas the upper interfaces have a transition region of several MLs and considerable lateral roughness on a scale of tens of/~.

From the HRTEM data, we have derived both the lateral roughness and the averaged vertical composition profile of the interfaces. In the lattice matched sample, we find a lower interface with approx. 3 MLs width and a roughness (standard deviation) of 0.65 MLs averaged over the picture's extent. The upper interface exhibits a much larger width of 9 MLs with a roughness of 1.3 MLs. In comparison, evaluation of the HRTEM picture of the Ga-rich sample gives an almost identical lower interface with a width of 3 MLs and a roughness of 0.55 MLs. The upper interface of this sample is more broadened (11 MLs) compared to the lattice matched case. Surprisingly, the roughness of the upper interface is again 1.35 ML.

Therefore, our results demonstrate, that broadening of the QWs interfaces does not necessarily lead to an enhanced interface roughness. The similar PL linewidth observed from both samples is consistent with this observation.

894 R. Sehwedler et al. / Strained lnGaAs/lnP quantum wells

4. C O N C L U S I O N

We have analysed HRTEM and spectroscopic data on Inl_,Ga~As/InP MQWs. We uti- lized a simple four-layer model assuming the QW to consist of the InP barriers, an Inl_~Ga~As-layer, and an additional layer of InAsl_~P. which accounts for the growth induced interface structure. This model is sufficient to explain the optical transition en- ergies and the interface roughness observed both in HRTEM and in optical spectroscopy.

A C K N O W L E D G E M E N T

This work was supported by the Deutsche Forschungsgemeinschaft under contract no. Ku540/ l l -1 and by the DAAD within the PROCOPE program.

R e f e r e n c e s

1 J. Hergeth, D. Grfitzmacher, F. Reinhardt, P. Balk. J. Crys. Growth 107, 537 (1991).

2 T. Y. Wang, E. H. Reihlen, H. R. Jen, G. B. Stringfellow. J. Appl. Phys. 66, 5376 (1989).

3 S. Juillaguet, J. P. Laurenti, R. Schwedler, K. Wolter, J. Camassel, H. Kurz. In: Non-Stoichiometry in Semiconductors, Edited by K. J. Bachmann, H.-L. Hwang, C. Schwab, p. 155, Elsevier, 1992.

4 A. Ourmazd, D. W. Taylor, J. Cunningham, C. W. Tu. Phys. Rev. Lett. 62, 933 (1989).

5 K. Wolter, R. Schwedler, B. Gallmann, Ch. Jaekel, M. Stollenwerk, J. Camassel, J. P. Laurenti, S. Juillaguet. In: Proceedings ESSDERC 91, Montreux, p. 593, Elsevier, 1991.

6 R. Schwedler, B. Gallmann, K. Wolter, D. Grfitzmacher, M. Stollenwerk, H. Kurz. In: Non-Stoichiometry in Semiconductors, Edited by K. J. Bachmann, H.-L. Hwang, C. Schwab, Elsevier, 1992.