|
Using a Variable Temperature Scanning Tunnel Microscope (STM) we observed a phase transition in progress on an atomic level and determined that point defects play a crucial role in this process. Today it is widely accepted that defects and impurities affect not only transport and optical properties of a system but also its critical behavior.
As a representative of a two-dimensional phase transition we have chosen the recently discovered Ö3xÖ3 to 3x3 phase transition in the ?-phase of Sn/Ge(111). This is a transition from a normal metal room temperature to a surface charge density wave (CDW) low temperature phase. We found that point defects (Ge substitutional atoms and vacancies) produce decaying waves with 3x3 periodicity even at room temperature. This perturbation in charge distribution affects only nearest and second nearest neighbors of the defects. When the temperature is lowered the extent of these 3x3 waves grows continuously and the resultant STM image is a superposition of waves from each defect. With other words, the decay length of the waves increases with decreasing temperature and becomes infinite around 70 K. This singularity defines the phase transition temperature of the Sn/Ge(111) system, Tc=70K. Below 70 K the surface consists of 3x3 (charge density wave) domains separated by sharp, only few unit cells wide, domain boundaries. This phase can no longer be described as a superposition of decaying waves and is therefore different from the high temperature phase. Tc is significantly lower than the 210 K deduced previously from LEED intensity measurements which cannot distinguish between defect induced perturbations with 3x3 symmetry described above and a true 3x3 reconstruction of the Sn film. Defects are very important in the charge density wave phase too. They predefine the position and orientation of the domains. One of the most interesting things, which we observed, is the correlation of the defects below the phase transition temperature. In the low temperature phase, Ge substitutional atoms do not occupy charge density maxima inside one domain. At room temperature and even at temperatures as low as 165 K, however, the defects are randomly distributed. We verified this observation by a rigorous statistical analysis and concluded that Ge atoms (i. e. the defects in the Sn film) moved to correlated positions at a temperature around the phase transition temperature. We attribute this somewhat surprising observation to a collective interaction between the Ge defects mediated by density waves in the Sn layer.
|
|
|
