INFLUENCE OF GROWTH INTERRUPTION ON THE FORMATION OF SOLID-STATE INTERFACES

Transcription

1 122 INFLUENCE OF GROWTH INTERRUPTION ON THE FORMATION OF SOLID-STATE INTERFACES I. Busch 1, M. Krumrey 2 and J. Stümpel 1 1 Physikalisch-Technische Bundesanstalt, Bundesallee 100, Braunschweig, Germany 2 Physikalisch-Technische Bundesanstalt, Abbestraße 2-12, Berlin, Germany Abstract In this study the effects of growth interruptions on the formation of the interfaces in GaAs/AlAsmultilayers are investigated. For that purpose a series of different samples has been manufactured with molecular beam epitaxy (MBE). The introduction of growth interruptions of 50 s after the deposition of the layer leads to a change in the morphological properties of the interfaces, in particular their correlation length. These modifications due to the growth interrupt are analysed with diffuse X-ray scattering (DXRS). As a result of the measurements an extension of the lateral correlation length can be proved. By contrast the vertical correlation of the interfaces is not affected. Introduction With the shrinking of the dimensions of solid-state systems its physical properties become more and more influenced by the structure of the interface. Diffuse X-ray scattering (DXRS) is sensitive to the morphological aspects of the interfaces in thin film systems. The diffusely scattered X- rays are detected in the vicinity of the critical angle of reflection. The variation of the scattered radiation is mainly due to the roughness of the interfaces. Because of the small incident (α i ) and exit angles (α f ) the resulting scattering vector q is very small, and the Laue condition for constructive interference is only fulfilled for the interfaces of the super lattice. The coplanar scattering geometry (cf. fig. 1) used in this experiment causes a zero q y component of the scattering vector q. The q x - and -components are adjusted by the two angles α i and α f with respect to the sample surface. The intensity profile of the scattered X-rays at the reciprocal coordinates (q x, ) is the image of the interfaces into the reciprocal space (cf. Fig. 1, right). The areas shaded in grey are not accessible for measurements due to the fact that one of the angles is below zero. This image is the so-called space map of the sample. The structure of this space map is basically generated by the morphology of the interface.

2 This document was presented at the Denver X-ray Conference (DXC) on Applications of X-ray Analysis. Sponsored by the International Centre for Diffraction Data (ICDD). This document is provided by ICDD in cooperation with the authors and presenters of the DXC for the express purpose of educating the scientific community. All copyrights for the document are retained by ICDD. Usage is restricted for the purposes of education and scientific research. DXC Website ICDD Website -

3 123 Offset-Scan θ /2θ -Scan scattering plane Detector Scan k i q Rock Scan a i a f k f -1 / nm q x sample q x Figure 1 The coplanar scattering geometry and the construction of the reciprocal scattering vector q. The incident beam, the reflected beam, and the normal to the surface are within the scattering plane (left side). The path of the different scan modi are pictured on the right side. The θ/2θ scan shown in Fig. 1 (right) is used in X-ray reflectometry (XRR), which is a special case of general DXRS. In XRR the specular reflected X-rays are measured while the condition α i = α f is fulfilled. Only the component perpendicular to the surface ( ) contributes to the scattering vector q. With this scan, the thickness d of the layers of the specimen can be determined with an uncertainty of less than 1 nm. The simultaneously measured diffuse scattered intensity already provides access to the rms-roughness σ of the interface. For an advanced analysis of the lateral structure of the interface topography, the measurement of the off-specular reflections is required, because in this case q has an additional component parallel to the interface (q x ). The measurements of the diffuse scattered X-rays over the complete reciprocal space are carried out with a series of rock scans to produce the space map of the system (cf. Fig. 1, right side and Fig. 3). The simulation of the scattering process with structure models for the multilayer system is done on the basis of the Distorted Wave Born Approximation (DWBA) and fractal models for the rough interfaces [2,3]. This simulation provides the critical parameters for the morphology of the interface, like rms-roughness σ, lateral ξ, and vertical correlation length ξ, etc. Measurements A set of three samples with different growth conditions has been analysed in this study. The GaAs/AlAs multilayer systems are produced by molecular beam epitaxy (MBE). Each system

4 124 consists of 50 double layers of 2.65 nm AlAs and 1.73 nm of GaAs 1 which are grown on a (100) GaAs substrate with no miscut. Finally, the system is covered by a 10 nm cap layer of GaAs. One of the samples (P90-12) is prepared in a continuous growth mode, i.e. the new layer is grown immediately after the deposition of preceding layer. The other samples have a growth interruption of 50 s after the deposition of each GaAs layer (P92-13) and after the deposition of AlAs (P91-11), respectively. Due to the growth interruption, an enlarged diffusion along the interface occurs which results in a change in the morphology. The DXRS measurements serve to determine the morphological structure of these three samples and the comparison of the different growth conditions on the formation of the interface. The measurements are carried out at the four-crystal monochromator (FCM) beamline of the PTB at BESSY-II, at a photon energy of E=8048 ev (Cu-K α radiation). First, the reflectograms of the samples are measured to determine the correct thickness of the multilayer films. These values are used as necessary input data for the simulation models (cf. Fig. 2). To achieve a large dynamical range of approx. nine orders of magnitude, two detectors are used: a silicon photodiode and a photon-counting silicon drift detector (SDD). The SDD is indispensable for the low intensities of the diffuse scattered radiation reflectance reflectance / a. u. 2,00 2,25 2,50 2,75 3,00 3,25 3,50 3,75 4,00 1 These thicknesses are the nominal values for the MBE process.

5 125 Figure 2 θ/2θ scan of sample produced by MBE in the continuous-growth mode. The system consists of 50 double layer GaAs/AlAs with a thickness of d=4.17 nm. The diagramm on the left side shows a part of the scan to present the high frequency oscillation in detail. The analysis of the reflectogram delivers a thickness of a single GaAs/AlAs double layer of d P90-12 = 4.17 nm for the continuously grown multilayer and d P92-13 = 4.29 nm for the sample with growth interruption after the GaAs layer. The good reproduction of the layer thickness of the samples during the MBE process is proved by the high frequency oscillations due to the side maxima in an XRR scan. The high resolution scans are carried out with a step width of θ = Space Map The value of the rms-roughness, extracted from the XRR measurements, are almost identical for the three samples (σ 0.5 nm), but the lateral structure of the interfaces varies significantly. This is documented by the difference in the space maps of the systems, particularly in the changes of the shape of the intensity maxima in the diffuse part (cf. Fig. 3). 4,5 4,0 5,0 3,5 4,5 3,0 2,5 4,0 2,0 3,5 1,5-0,20-0,15-0,10-0,05 0,00 0,05 0,10 0,15 0,20 q x -0,3-0,2-0,1 0,0 0,1 0,2 0,3 q x Figure 3 Space maps of the samples with a growth interruption of 50 s after the deposition of AlAs (left) and with a growth interruption of 50 s after GaAs layer (right). The intensity of the scattered radiation varies over five (left) and four (right) orders of magnitude, respectively. The colorcode runs from high (red) over medium (yellow) to low intensity (blue). In the case of sample P91-11 with a growth interruption after AlAs (the space map on the left side of Fig. 3), the lateral correlation length ξ has a value of some ten nanometers (the large ex-

6 126 tension of the banana in q x direction), the vertical correlation length ξ is in the region of some hundred nanometers (the concentration of the maximum in direction), i.e. the replication of surface features through the next layer is almost perfect. Sample P92-13, with a growth interruption of 50 s, shows a high concentration of the coherent scattered radiation in a small area of the reciprocal space (cf. fig. 3, right). The lateral correlation length of this sample is greatly extended to some hundred nanometers. The replication of the interface topography stays at the same high level as before. The value of ξ is again some hundred nanometers. Conclusions The study demonstrates the destruction-free analysis of the influence of different growth conditions on the formation of solid-state interfaces. As an example, interruption during the growth process is chosen to demonstrate the DXRS technology. The enlarged two-dimensional diffusion during the growth interruptions in the MBE process enables the emergence of long-range structures (islands, step edges, etc.) at the interfaces. The analysis of all measurements of the three samples and the modelling of the multilayer systems in computer simulations on the basis of DWBA will provide more precise and quantitative data. These data will be the starting point for investigations of the transport processes in comparable systems. References [1] G. Bernatz; S. Nau; R. Rettig; H. Jänsch; W. Stolz, J. Appl. Phys. 86 (1999) 6752 [2] S.K. Sinha et al., Phys. Rev. B 38 (1988) 2297 [3] I. Busch and J. Stümpel, J. Appl. Surf. Sci (2003) 201