Comparison of BB PL intensity (b) and (d) to Cu (a) and Fe (c) densities. The transition metal density was measured by . The increased recombination activity of the transition metal precipitates is visible as a reduced BB PL intensity (b) and (d).
The recombination active areas of the copper precipitates are proportional to their actual areas (a). The overlapping confidence bands around the linear fits for iron and copper show no significant dependence of the recombination active area on the element. (b) shows the excerpt of the iron precipitates from (a). While the scattered data cannot show a linear dependency for iron, at least a positive trend is visible from the linear fit. The recombination active area of a precipitate is defined on the upper left side of (a). The errors for the area estimation are defined by adding or removing one pixel layer around the precipitate.
Comparison between (a), measured by analyzing Fano resonances (Ref. 19) and stress (b). Both maps are calculated from one measurement which guarantees a perfect spatial alignment. The bright straight lines mark the areas of high compressive stress, where the recombination activity tends to be reduced (particularly at the left line, where the compressive stress is maximal). At high tensile stress the recombination activity tends to be increased. The correlation between recombination activity and stress might be reduced by dislocations around the precipitates and the grain boundaries at the triple point. The area, where the precipitate intersects the surface is marked by the dotted line. Within this area, the Raman measurements are not reliable.
Recombination activity, represented by the maximum at the precipitates, plotted against the maximum tensile stress at the precipitate. The graph shows a positive correlation, which is not significantly dependent on the impurity element (Fe or Cu). Also the iron data points show a positive correlation between stress and . The error bars indicate the uncertainty of SMAPS of about 8 MPa and an uncertainty in the contrast of about 10%. The fact, that the data points for iron are systematically on the right side of the simulation curve, could be due to the systematic overestimation of stress by SMAPS. The simulation (geometry on the right side, details in Sec. IV) shows a good agreement with the experimental data, considering the simplicity of the simulation model.
Article metrics loading...
Full text loading...