Measuring Dislocation Line Charge in GaN using Scanning Precession Electron Diffraction

Published

July 26, 2025

Author(s)

Edwin Supple, Kristine Bertness, Matthew Brubaker, Alexana Roshko

Abstract

Dislocations in GaN have been known for some time to accumulate line charge along their length, affecting the electrical and optical properties of the material: increasing leakage current, inducing premature breakdown, and reducing carrier mobility [1]. Analysis following the Read model, first developed for dislocations in Ge, suggests dislocation line charge densities up to 2 electrons per unit cell length in GaN on edge dislocations only, depending on free carrier concentration and dislocation density [2]. Line charge measurement in the scanning/transmission electron microscope (STEM) using phase contrast techniques such as electron holography and STEM differential phase contrast (DPC) is of interest to validate the model on the individual-dislocation scale. Electron holography experiments have yielded results with reasonable agreement although they consistently find screw and mixed dislocations also have associated charge [3-5]. STEM DPC measures deflection of the direct beam due to electric fields using center of mass (CoM) shift. Dynamical diffraction effects hinder this interpretation by introducing non-uniformity to the direct beam, adding noise to what is already a small signal. Lattice bending, strain, and nonuniform electric fields can all cause these dynamical effects and the dislocations under study necessarily feature all three. STEM precession electron diffraction (PED) to reduce dynamical effects and improve signal:noise, allowing for dislocation charge quantification. STEM PED tilts the converged probe through a cone of tilt angles at each scan pixel, thereby sampling substantially more diffraction conditions and so reducing spurious contrast from dynamical diffraction while maintaining a small convergence angle necessary for CoM shift measurement. We have used STEM PED to measure and quantify electric field surrounding dislocation in GaN. Figure 1a shows the measured center of mass (CoM) shift in plan view from a dataset collected with 2.3 mrad convergence angle and 0.3-degree precession angle. CoM shift angle and intensity is shown according to the inset color wheel. Each dislocation deflects the beam outwards from the dislocation core, indicating charge present along the dislocation. Figure 1b shows the potential map generated from the same dataset. Dislocations are identified as negative potential wells, in agreement with the idea that electrons are collecting there; and a residual background potential slope remains due to low frequency noise in the data. A radial average potential profile is generated for each dislocation and is plotted in Figure 1c, with background subtraction to accommodate the residual slope. Figure 1d shows the mean potential profile from the collection of dislocations is shown in black; the calculated potential well and the convolution of the calculated potential with the beam profile are shown as a dashed line and red line respectively. The convolved calculation at 0.2 electrons per unit cell length is in good agreement with the averaged data.

Proceedings Title

Microscopy and Microanalysis

Conference Dates

July 27-31, 2025

Conference Location

Salt Lake City, UT, US

Pub Type

Conferences

Keywords

TEM, STEM, DPC, dislocations, electric field

Citation

Supple, E. , Bertness, K. , Brubaker, M. and Roshko, A. (2025), Measuring Dislocation Line Charge in GaN using Scanning Precession Electron Diffraction, Microscopy and Microanalysis, Salt Lake City, UT, US (Accessed October 10, 2025)