Defects and dislocation densities in InP wafers


Indium Phosphide wafers are frequently used as substrates for the growth of infrared laser and LED heterostructures. Wafers are commercially available and may easily be obtained in large quantity for production. Defects in substrates are generally specified as a dislocation density, but the inhomogeneity of their distribution and the range of sizes is not known from one wafer to the next. When growing a new device on such a wafer, it is imperative to understand how the substrate defects affect the heterojunctions above them that comprise the active layers of the device. At a minimum, the quality of the wafers should be measurable in order to determine whether they are of sufficient quality to be used in the production of reliable devices with near-zero latent defects.

Klar’s spectroscopic mapping system is ideally suited for capturing the variation of the bandgap across a wafer and identifying the size, distribution, and nature of any defects.

To demonstrate the use of the Klar instrument for such studies, we acquired a square 10mm commercial (100) InP wafer from MSE supplies and scanned a portion of it to map out the bandgap variation. Using ImageJ, we were able to determine that the defect density in our sample was within a factor of two of that specified by the supplier (~ 50,000/sq.cm.). This tool also captures the distribution of sizes of the defects.

Energy map of a 2mm x 2mm section of a commercial InP wafer using a 635nm laser excitation source. The dotted box on the map shows a region containing one of the larger defects (see below).

Energy map of a 2mm x 2mm section of a commercial InP wafer using a 635nm laser excitation source. The dotted box on the map shows a region containing one of the larger defects (see below).

Spectra from inside (top) and outside (bottom) the region containing a dislocation causing an increase in the bandgap by about 0.5%. The peak intensity of the emission from the dislocation is reduced by over 30%, a consequence of the change in the transition dynamics at the defect’s location in the crystal.

Spectra from inside (top) and outside (bottom) the region containing a dislocation causing an increase in the bandgap by about 0.5%. The peak intensity of the emission from the dislocation is reduced by over 30%, a consequence of the change in the transition dynamics at the defect’s location in the crystal.