Determining the Enthalpy of Formation of A Si Interstitial Using Quantitative Tem and Sims
- PDF / 777,779 Bytes
- 8 Pages / 417.6 x 639 pts Page_size
- 68 Downloads / 170 Views
at. Res. Soc. Symp. Proc. Vol. 532 ©1998 Materials Research Society
this experiment at various temperatures, the enthalpy of formation of the Si self-interstitial can also be determined. Since the dislocations can be used to compute a number of atoms trapped, these studies can give quantitative information about the recombination and diffusion of defects in the crystal. Preliminary experiments were carried out to determine the feasibility of this experimental procedure and also to optimize the various implant and anneal conditions in order to achieve maximum sensitivity and minimum statistical error. We will describe our experimental procedure then outline our results and discuss their implications. Si+ implants
Low temperature, short time anneal V Y\ /%V
A N/ \V X%% A
EO 00000
High temperature, long time anneal
(C1/C1*)
C1 SCI
Fig. 1: Schematic diagram of the experimental procedure used to determine CI* EXPERIMENTAL Use of Stereology in Optimizing of Quantitative Transmission Electron Microscopy Recently there has been much work on the quantification of trapped interstitials in extended defects arising from the ion implantation and annealing of silicon. Several studies have shown that these defects can be a source of interstitials for transient enhanced diffusion and can also be a sink for interstitials. The two major types of defects quantified are {3111 defects and dislocation loops. Several methods have been used to quantify total number of trapped interstitials in the dislocation loops. The most commonly used methods are tracing, tracing combined with image processing, and the use of Gaussian distributions to estimate size.4 112
Stereology is based on geometric probability, which has been in use for several decades. According to stereology, the probability that any given point is within a dislocation loop is the same as the area fraction covered by the loops. Therefore, a grid is used, and the fraction of grid nodes within the loops is the same as the fractional area covered by the loops. Since the loops are all extrinsic the trapped interstitial content in the loops can now be determined. Several grid/image magnification combinations were studied and the procedure outlined below minimized the time/error combination. Weak beam, dark field transmission electron microscopy is used to produce a negative at a magnification of 50,000 X. An 8x10 inch print is made at a magnification of approximately 4X, resulting in a total magnification of approximately 200,000X. A 0.5cm spaced grid (1333 nodes) is placed over the photograph. Then every node that lies inside a loop is marked. The number of marked points is divided by the total number of points (1333 in this case). This fraction is multiplied by the planar atomic density of the I111 } plane (-1.6E1 5/cm 2) to yield the number of trapped interstitials. Following many tests the variation for a single investigator counting the same image was found to be less than ± 3El 2/cm 2. In addition if two people count the same image the variation again is less than ± 4
Data Loading...
