Flowing damage in ion-implanted amorphous silicon

Using molecular-dynamics simulations, we have studied the creation and evolution of damage in crystalline and amorphous silicon following the implantation of energetic keV ions. A method is proposed to identify anomalous atoms based on a weighted combination of local, atomic-scale properties, which applies to both Si phases. For crystalline Si, the passage of the ions causes compact amorphous regions to form, while no evidence for melting is observed. The relaxation of the amorphouslike regions proceeds initially by the rapid recrystallization of smaller clusters and isolated atoms, followed by a long period of steplike changes in the number of defects due to spontaneous annealing of damage pockets at the crystalline-amorphous interface. In amorphous Si, the initial stage of damage annealing (which lasts a few picoseconds) resembles closely that observed in crystalline Si; on larger time scales, however, the damage is found to “percolate,” or flow, through the system, inducing damage away from the collision cascade, thus causing an overall “derelaxation” of the material.

Figure 1

Distribution of defects in c-Si (left and center columns) and a-Si (right column) at times indicated. Only defective atoms are shown, as identified using the topological method (left column for c-Si) or the probability-function method (center column for c-Si, right column for a-Si). Atoms colored in purple (darker) are those that will return to a normal state within the following few time steps.

Figure 2

Total number of defects vs time (logarithmic scale) in the crystalline target. The arrows point to the different configurations presented in Fig. 1.

Figure 3

Distribution of atoms whose potential energy exceeds $-$3.25 eV in the a-Si target at $t=0.2$ ps. The black line indicates the trajectory of the implanted ion.

Figure 4

Bond-angle distributions (dots) and probability functions (lines) for both c-Si (blue) and a-Si (red).

Figure 5

Distribution of the continuous coordination $C_i$ for c-Si and a-Si. The left peak in the c-Si curve is due to surface atoms (which are smoothed out in the a-Si curve).

Figure 6

Amount of damage as a function of time in the c-Si target using the probability approach with threshold values of 0.40 and 0.168 (red area) and the topological approach (solid black line). Blue and purple areas correspond to the number of anomalous atoms which are the same as, or first neighbor to, topologically identified defects, respectively.

Figure 7

Number of anomalous atoms as a function of time in the a-Si target using various individual and collective thresholds.