The electron-density distribution in silicon

The electron-density distribution in crystalline silicon was refined with a three-parameter density model which was originally designed by Brill. The same data sets were used as by other authors [Price, Maslen & Mair (PMM) (1978). Acta Cryst. A34, 183-193; Hansen & Coppens (1978). Acta Cryst. A34, 909-921] in refining their multipole models. The data sets are (1) the 15 Mo Kα room-temperature data of Aldred & Hart (AH) [Proc. R. Soc. London Ser. A (1973), 322, 223- 238, 239-254], (2) the 15 AH data plus the 222 and 442 data of Roberto & Batterman [Phys. Rev. B (1970), 2, 3220-3226] and Trucano & Batterman [Phys. Rev. B (1972), 6, 3659-3666], respectively, (3) the 17 data of (2) plus four data measured by Hattori, Kuriyama, Katagawa & Kato (HKKK) [J. Phys. Soc. Jpn (1965), 20, 988-996]. R for these data sets, obtained with the present density model, is 0.12, 0.13 and 0.30%, respectively. This compares well with the values obtained by the other authors with their multipole models. In the deformation density, PMM found a Si-Si bond peak of height 0.13 e Å^-3 with data set (1), whereas Yang & Coppens [Solid State Commun. (1974), 15, 1555- 1559] found a peak height of 0.29 e Å^-3 with data set (3). With data set (2) a peak height of 0.20 e Å^-3 is found in agreement with the result of PMM. However, the increase of the peak height to 0.29 e Å^-3 with data set (3) is an inaccuracy which arises from the fact that the four HKKK reflections were not measured accurately enough.