Estimation of diffusion coefficient by photoemission electron microscopy in ion-implanted nanostructures

ROM 2011-3
Author: R. Batabyala, S. Patraa, A. Roya, S. Roya, L. Bischoffb, B.N. Deva,*
Institute: (a) Department of Materials Science, Indian Association for the Cultivation of Science, Kolkata, India / (b) Research Centre Dresden-Rossendorf, 01314 Dresden, Germany
Publication: Applied Surface Science 256 (2009) 536–540

We have fabricated parallel stripes of nanostructures in an n-type Si substrate by implanting 30 keV Ga+ ions from a focused ion beam (FIB) source. Two sets of implantation were carried out. In one case, during implantation the substrate was held at room temperature and in the other case at 400° C. Photoemission electron microscopy (PEEM, a FOCUS GmbH product) was carried out on these samples. The implanted parallel stripes, each with a nominal dimension of 4000 nm × 100 nm, appear as bright regions in the PEEM image. Line scans of the intensities from the PEEM image were recorded along and across these stripes. The intensity profile at the edges of a line scan is broader for the implantation carried out at 400° C compared to room temperature. From the analysis of this intensity profile, the lateral diffusion coefficient of Ga in silicon was estimated assuming that the PEEM intensity is proportional to Ga concentration. The diffusion coefficient at 400° C has been estimated to be 1.3×10-15 m2/s. Across the stripes an asymmetric diffusion profile has been observed, which has been related to the sequence of implantation of these stripes and the associated defect distribution due to lateral straggling of the implanted ions.

Several PEEM images are shown in Fig. 1. Fig. 1(a) with a field of view (FOV) of 150 µm shows a stripe-patterned structure with its surrounding frame structure. Fig. 1(b) and (c) show PEEM images (FOV: 30 µm) of Ga-implanted stripe patterns, fabricated at RT and at 400° C substrate temperature, respectively. The Ga-implanted regions are brighter than the surrounding regions of the substrate which means that the photoelectron ejection is much more intense from the Ga-implanted regions. Ga implantation into an n-type substrate would initially compensate and then transform the implanted region into p-type. p-type regions appear brighter in PEEM image. Here a sharp stripe pattern is observed for implantation at RT (Fig. 1(b)) while a fuzzy pattern is observed for implantation at 400° C (Fig. 1(c)).

Fig. 2 shows the PEEM intensity profiles of nanostripe patterns along the length of the Ga-doped stripes. Because the intensity of PEEM images depends on the work function or photothreshold of the material, as the photothreshold of a semiconductor is modified by dopant concentration, the intensity is also related to dopant concentration. Hence the lateral intensity profile of the nanostructure is used to determine the lateral diffusion coefficient of the dopant. The intensity profile data have been fitted with the standard diffusion model in order to determine the lateral diffusion coefficient.

We have observed an asymmetric behaviour in the intensity profiles across the implanted stripes at RT; the profiles are even more complex for implantation at 400 8C. These profiles are shown in Fig. 3(a). An interpretation of these intensity profiles is given via the illustration in Fig. 3(b). This asymmetry can be explained in terms of the initial FIB beam profile itself as well as the defects (vacancies and interstitials) created during Ga+ ion implantation. 30 keV Ga+ ions produce ~540 vacancies/ion. The lateral distribution of defects due to the lateral straggling of the ion beam is responsible for this asymmetric intensity (concentration) profile which occurs both at RT and 400 8C. Fig. 3(b) (top) illustrates the first Ga-implanted stripe (red strip) and the defect and Ga distribution due to lateral straggling (gray region around the red stripe). This distribution is symmetric on both sides of the stripe. Presence of defects causes enhanced diffusion. However, diffusion of Ga into these defected regions would produce a symmetric concentration profile and in turn symmetric intensity profile in PEEM. When the second stripe is implanted [Fig. 3(b) (middle)], again a symmetric defect concentration profile is created on both sides of the stripe. However, additional defects are already present on one side of the stripe, i.e. between the second and the first stripe. Therefore, during implantation of the second stripe there will be more enhanced diffusion of Ga towards the first stripe. Thus the Ga concentration profile would be more broadened towards the first stripe. This process would continue for all the later implanted stripes [Fig. 3(b) (bottom)]. The expected Ga concentration profiles due to this process are shown in the right side of Fig. 3(b).

We have presented a method for the estimation of diffusion coefficient in nanoscale systems using photoemission electron microscopy (PEEM). p-type stripes have been fabricated by implanting 30 keV Ga+ ions into an n-type Si crystal using a focused ion beam (FIB) system. The implanted stripes are of nominal dimension of 4000 nm × 100 nm. These stripes appear bright in the PEEM image. The brightness (PEEM intensity) appears to fall sharply at the edges of the stripes for the implantation carried out at room temperature. The stripes obtained by implantation at a substrate temperature of 400° C have appeared fuzzy in PEEM image. The intensity profile at the edges of the stripes appears broadened. By analyzing the intensity profiles from the PEEM images from the samples implanted at RT and at 400° C, the lateral diffusion coefficient of Ga in Si has been estimated. The knowledge of lateral diffusion coefficient is necessary to estimate how closely two doped nanostructures can be fabricated so that they do not interfere with each other.