Tests by TEM contrast simulations of the elastic field of a buried (001) low angle twist boundary in silicon

Abstract

A buried (001) low angle twist boundary (misorientation angle 0.48°) in silicon is investigated by the technique of two-beam transmission electron microscopy using bright-field images obtained from diffracting vectors g{400}, g{311} and g{220}, and g{nnn} with n = 3 and 4. To analyse its elastic field, the concept of interfacial "relaxation centres" is assumed, for which the relative displacement field of the two crystals is zero. These centres are located in the middle of each square formed by a dislocation unit cell. This assumption is tested positively for the first time for a double periodic network of screw dislocations. For the image contrast calculations, the elastic field of the network is that of two perpendicular families of screw misfit dislocations in the sense of Bonnet (1981). Since the two-beam bright-field approximation is weak to explain the experimental images taken with g{nnn} and n = 3 and 4, extra N beam image calculations involving N < 7 systematic reflections along g{nnn} have been carried out with a devoted programme. All the computed images prove to be in good agreement with the experimental images, still validating the adopted elastic field.

A dissociated dislocation in an ultrathin silicon plate

“…In expression (20a), the sign before the second monomial in B 2 should be negative. In the system of linear equations (26), line 12, the second monomial in B 3 should be the same monomial but in B 1 . In expression (32c) and function F 6 , the first term of the denominator should be written with a positive exponent (þ 2g|x 2 |).…”

“…Most measurements of S were obtained from weak-beam TEM observations on screw dislocations running in a direction 110 h i roughly parallel to the average plane of the thin plate. They often lie either on a {111} gliding plane [10][11][12][13] or a {001} low-angle twist boundary [18,19,25,26]. Plan view images taken under HRTEM conditions were also used [17] in the hope of visualizing possible kinks along the partials, the plate thickness being extremely small, about 4 nm.…”