Home | Legals | Data Protection | Sitemap | KIT

Application

Comprehensive 3D Characterization of Complex Dislocation Structures in Silicon

X-Ray diffraction laminography (XDL) has been applied to silicon (Si) samples with considerable lateral extension like typical industrial wafers. By combination of XDL with X-Ray white beam topography (XWBT) and (circularly polarized) visible light differential interference contrast (CDIC) microscopy a comprehensive characterization of a complex 3D dislocation structure was obtained, see Fig. 1 [Hänschke 2017]. Here the 3D paths of the individual dislocations could be resolved with a resolution of about 3 micrometer, the assiciated Burgers vector distribution could be determined, and these bulk features could be linked to the corresponding surface features like surface steps or mechnical damage.
 

<img src="cr7_STROBOS_page.jpg" height="436" width="830" alt="Fig. 1: Comprehensive Characterization of a dislocation pattern in silicon linking the (mesoscopic) 3D dislocation structure to the (microscopic/atomistic) Burger's Vector (BV) distribution and surface features.">
Fig. 1: Comprehensive Characterization of a dislocation pattern in silicon linking the (mesoscopic) 3D dislocation structure to the (microscopic/atomistic) Burger's Vector (BV) distribution - here color-coded - and surface features.

 

 

Fig. 2: Dislocation structures emerging Franck-Read
sources in Si a) snap-shots of XDL measurement in between
heating cycles b) visualization of selected individual  3D
dislocation paths and corresponding BV (indicated by color-code)

 

Quasi 4D Imaging of Dislocation Dynamics in Silicon

During the first experiments of the collaborators within the STROBOS-CODE project, the evolution of dislocation patterns in Si originating from Franck-Read sources was observed by performing a gradual annealing alternating with

XDL-scans (see Fig. 2.a). During the respective anneling cycles the samples were heated above the brittle-ductile transition temperature (for Si 600 - 700° C) for a time windows of about 8 to 20 seconds using a mirror furnace. In this way, it was possible to image the step-wise propagation of single dislocation lines in 3D (see Fig. 2.b). Moreover, it was allowed to segment individual dislocation paths to follow the movement and investigate the development of each dislocation in 3D volume. Quantitative 3D information were extracted from reconstruction data to analyse dislocation dynamics due to thermally driven forces (See Fig. 3).

Within the STROBOS-CODE project, the tools for data analysis will be further refined and the relation between the observed dislocation dynamics and the driving mechanical and thermal stresses will be investigated in unprecedented detail.

 

Fig. 3: View on  the evolution of one segemented dislocation and analysis of segment (secrew) lengths from 3D reconstrontuction

 

Applicability of XDL to Higher-Absorbing Semiconductors

In a first feasibility study at the example of gallium arsenide (GaAs) and cadmium telluride (CdTe), the project partners applied XDL successfully to higher-absorbing semiconductor materials.
The strain fields around two indents on a GaAs wafer are clearly visible in the weak beam XDL projection displayed in Figure 4. These indents are imbedded in a network of dislocations, forming cellular structures of strain fields. Figure 5 shows these kind of cellular dislocation networks in 3D. During the algebraic reconstruction process the so called weak beam contrast has been inverted.
Unlike Figure 4 and 5 the contrast mechanism in Figure 6 is the so called integrated intensity contrast. With this kind of diffraction contrast it was possible to reveal grain boundaries within the bulk of a CdTe wafer. The work on the field of higher-absorbing semiconducors will be continued.

 

Network of dislocations in the bulk of indented GaAs
Fig. 4: XDL projection of dislocation networks in the bulk of indented GaAs

 

<img src="GaAs_01_1.jpg" height="1272" width="1989" alt="Image of a 3D reconstruction volume rendering of a cellular dislocation network in a GaAs wafer.">

Fig. 5: Image of a 3D reconstruction volume rendering of a cellular dislocation network in a GaAs wafer.

 

 

Grain boundaries within CdTe
Fig. 6: Integrated intensity image showing grain boundaries within a CdTe wafer.

 

Present and Future

We aim for routine application of the quasi 4D Imaging scheme (gradual annealing with intermediate XDL-scans) to Si, GaAs and various other promissing crystalline materials of scientific and inustrial interest. Imaging techniques are an important tool to gain further insight into the complicated processes regarding defects - especially dislocations - that may occur during crystal growth and (industrial) wafer handling.