Response to ‘‘Comment on ‘Optical and acoustic phonon modes in self-organized Ge quantum dot superlattices’ ’’

In a comment on our recent letter, Yu first pointed out that there was strong alloying between the Ge dots and Si barrier layers because of the appearance of Si–Ge modes as also observed in SiGe alloys. It is correct that the Ge dot samples reported in our letter have some degrees of alloying due to interdiffusion. This was due to the fact that the samples were grown at a high temperature of 600 °C. The reason for using this temperature came from our intention to control the size uniformity. It was found that an optimum temperature occurs at around 600 °C for high-uniform monomodal Ge dots on planar Si substrate. At lower temperatures, the uniformity becomes worse. Moreover, two kinds of dots coexisted on the Si substrate ~pyramid and dome!. Figure 1~a! shows a cross-sectional TEM image of a tenperiod Ge quantum dot sample grown at 550 °C ~sample 1!. Vertically correlated dots are evident. The size variation of the first Ge dot layer, however, is determined to be 20% from AFM measurements @Fig. 1~b!#. The nonuniformity arises from limited diffusion at the lower growth temperature and becomes worse in the vertically correlated multilayers because vertical correlation rearranges the strain distribution, leading to the fact that the upperlayer dots are larger than those in the lower layers @Fig. 1~a!#. On the other hand, at a higher temperature around 600 °C, the dot uniformity is much improved and the size variation decreases to, for example 7%–8% in our samples in the letter. Such uniformity may be desirable, but the interdiffusion between the Ge dots and Si spacers reduces quantum confinement effects and is undesirable for optical applications. Thus, there is always a tradeoff between the uniformity and interdiffusion. Now let us take a look at the Raman spectrum of the sample 1 @Fig. 1~c!#. This figure includes a spectrum for a ten-period Ge dot superlattice grown at 500 °C ~sample 2!. One can easily see Si–Ge vibration modes at around 400 cm from both samples. The relative strength ISi–Ge /IGe–Ge decreases as the temperature decreases. Another phenomenon is that there is a weak feature ~indicated by an arrow in the figure! between Si–Ge mode and Si–Si mode for sample 1. This is due to localized Si–Si motion in the neighborhood of one or more Ge atoms. Such kinds of localized Si–Si optical modes ~Si–Siloc! are often observed in SiGe alloys