Response to “Comment on ‘Raman scattering from a self-organized Ge dot superlattice’ ” [Appl. Phys. Lett. 75, 3572 (1999)]

~301 cm in our case! does not mean that the peak mu come from Si acoustic phonons. In order to prove our ass ment, Raman scattering measurements were performe the dot sample and an identical Si substrate using the s experimental Raman system with an identical data collec time. In addition, different polarization configurations a cording to selection rules were used to better distinguish signals from the dot sample and the Si substrate. Figu shows the observed results. The spectrum from the sam ~top solid curve! was recorded in the 001(100,010)001 ̄ backscattering geometry. This configuration was chosen to m mize the acoustic phonon peak at around 303 cm 21 from Si substrate. The spectrum from the Si substrate ~bottom solid curve! was recorded in the 001(110,110)001 ̄ backscattering configuration in order to enhance the Si acoustic pho peak. A peak at 301 cm 21 from the sample in the top solid curve is about six times stronger than the Si acoustic pho peak at 303 cm from the substrate in the bottom sol curve. The strain on multilayered Si induced by the form tion of Ge dots changes the symmetry of the localized ~around the dots !. Because of this effect, the 303 cm 21 Si acoustic phonon peak may show up even though the sam is under the 001(100,010)001 ̄ configuration. Thus the ob served 301 cm Raman line from the dot sample may in clude the contribution from the Si acoustic phonons. Limit work on this issue 4 seems to indicate that the intensity of th Si acoustic phonon peak does not change significantly w and without the existence of strain. The only dominant sig in our case is from the Ge–Ge mode. In addition, the app ance of the Si–Ge mode at 403 cm 21 ~top solid curve in Fig. 1! suggests imperfect Si–Ge interfaces due to high gro temperature and/or the formation of Ge dots. Otherwise, peak should be forbidden in the 001(100,010)001 ̄ backscattering configuration. This also supports the existenc the Ge–Ge mode. In fact, the only concern here may be which parts, SiGe wetting layers or Ge dots, mostly contr ute to the Ge–Ge mode. Existing work 5 indicated that the