Synchrotron white beam X-ray topography (SWBXT) [1, 2][M. Dudley, in: Mater Res. Soc. Symp. Proc., vol. 307, 1993, p. 213; Encyclopedia of Applied Physics, vol. 21, 1997, p. 533] combined with chemical etching and Nomarski optical microscopy have been employed to investigate the phenomenon of twinning in sulfur-doped, MLEC-grown 〈1 0 0〉 InP single crystals. Results confirm the prediction of Hurle [3, 4][Sir Charles Frank, OBE, FRS; on 80th Birthday tribute, 1991, p. 188; J. Crystal Growth 147 (1995) 239] that twins nucleate in regions where {1 1 1} edge facets are anchored to the three-phase boundary (TPB), and that twinning produces a {1 1 1} external facet on the shoulder region of crystal. The critical factor controlling the point at which the twin nucleates is observed to be the appearance of a {1 1 5} external shoulder facet which is then converted to the {1 1 1} external shoulder facet upon twinning. Based on the observation of both anchored edge facets and twinning at shoulder angles of approximately 74°, ν max, the maximum in the range of angles, ν (where ν is the angle between the edge facet and the extension of the crystal surface, or shoulder), over which edge facets can be expected to be trapped at the TPB, as presented by Hurle, has been increased, from ∼86.5° to ∼112°, by modifying approximations made concerning the relevant physical parameters for the particular case of S-doped InP. A correction is made to an error in Hurle: the most dangerous grow-out angle for the occurrence of twinning in 〈1 0 0〉 oriented crystals, is 74.2° and not 35.5°. Also Hurle predicted that, in crystals with diamond cubic or zincblende structures, the range of ν angles for edge facet pinning should encompass 70.5(3)° for twinning to be possible. This is here extended to the statement that it should encompass either 70.5(3)° and/or 109.47°. Incidences of twinning were also observed to occur by a similar mechanism but one which involved the conversion of a {1 1 4} external shoulder facet to a {1 1 0} one.
Read full abstract