TEMStrain    

Ambiguity

Numerical analysis shows that one pattern is not sufficient for determination of a complete strain tensor because the optimization problem is ill–conditioned: adding

to a strain tensor given in the microscope coordinate system with the ’3’ along optical axis of the microscope has an undetectable effect on the simulated pattern; di are arbitrary numbers with magnitude of the order of 1/1000 and a is of the order of 2. In order to alleviate the problem, multiple patterns with different orientation but originating from the same area of the foil can be used. In the case of two diffraction patterns, a unique strain tensor can be obtained if the characteristics a are sufficiently different. If more than two zone axes are used, the solution is, in principle, unique. In practice, the reliability of the results may be still low due to errors in the locations of HOLZ lines.