Abstract
High-resolution lattice images of two of the platelets used in the displacement measurements in the previous paper (Part I) have been used in image matching experiments in an effort to determine the structure of the platelets. This process was complicated by the fact that the structure of these platelets has been altered to some degree by radiation damage, as indicated by their non-zero component of displacement parallel to the plane of the platelets. However, since the value of the normal displacement of both platelets was close to the mean value determined in Part I,0.39 a o, it was assumed that the damage was not severe and that the observed component of displacement parallel to the platelet could be simulated by the appropriate rigid-body shear of the atomic coordinates pertaining to the ideal, undamaged models. In addition to three models for the structure of platelets previously proposed in the literature, several new models were developed and the images of all these were computed and compared with experiment. It was found that the best match was provided by a model which contains the equivalent of a single layer of nitrogen atoms equally distributed between two adjacent layers of the structure. This model has a normal displacement which agrees well with the value determined in Part I and it has been shown that the particular arrangement of nitrogen and carbon atoms envisaged in this model minimizes the bond energy of the structure.