Surface Diffusion of Large Ag Clusters on Ag(100)
- PDF / 1,220,134 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 89 Downloads / 184 Views
both show that the diffusion coefficient should generally decrease with increasing cluster size. This leads to an intuitive expectation that very large 2D clusters should be rather immobile. However, we recently reported that even very large 2D clusters of Ag (N - 102 to 103) undergo measurable diffusion on a Ag(100) surface at room temperature.15 Experimental Characteristics of Cluster Diffusion. Some of the STM data, which led to the rather surprising conclusion that large clusters can diffuse, are shown in Fig. 1. Experimental details are available elsewhere.1S The bright spots in the image are Ag dusters in the middle of a large, smooth terrace; the black streak at the right is a monatomic step edge. The Ag islands adopt an approximately square shape, although irregularities such as rounded corners and crooked edges are common. The Ag dusters in Fig. 1 range in size from 25 to 300 atoms (230 to 2500 A2). Comparison of Fig. la with lb clearly shows that the Ag clusters have moved in the interval between images, in this case 150 min. Over a period of several hours, the root mean square displacement of the dusters is on the order of 102 A. The diffusion of these dusters did not appear to be an artifact of tip-sample perturbation, although such perturbations certainly can occur. For instance, efforts to obtain atomic-scale resolution on small islands (N < 30) invariably perturbed the island. In order to avoid this problem, we limited examination to islands of N 20001-
15001-
10001-
500
0 I
0
I
I
500
1000
I
1500
2000
A
(a) Figure 1.
0
500
1000
1500
2000
(b)
STM images obtained following deposition of 0.015 monolayers (ML) Ag on Ag(100) at room temperature. Deposition rate is 3 x 10-3 ML/s. (a) t = 0 min. (b) t = 150 min.
16
> 100; further, we sacrificed spatial resolution to minimize perturbation. We also examined the effect of different raster conditions on the measured diffusion coefficient, and could find no effect.15 Examples of the key experimental data quantifying cluster motion are given in Fig. 2a. It shows the mean value of the ratio of the square of the displacement, divided by the time interval between observations, , as a function of the mean time between observations, , for 5 different clusters. In Fig. 2a, r is taken to be the total displacement, d, of the cluster's center-of-mass from its point of origin after a time, t. Diffusion coefficients were extracted from data such as those
(b)
(a)
300 250 200
-
N=1 00
150
Data Loading...
