Detailed Analysis and Computationally Efficient Modeling of Ultra-Shallow Dopant Profiles Obtained by Low Energy B, Bf 2
- PDF / 383,237 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 38 Downloads / 183 Views
ABSTRACT With increasing levels of integration, future generations of integrated circuit technology will require extremely shallow dopant profiles. Ion implantation has long been used in semiconductor material processing and will be a vitally important technique for obtaining ultra-shallow dopant profiles. However, implant channeling for low energy ion implantation must be understood and minimized. We report the results of a detailed experimental analysis of 275 ultra-shallow boron, BF2, and arsenic as-implanted profiles, and the development of an accurate and computationally efficient model for ultra-shallow implants. The ultra-shallow dopant profiles have been modeled by using the Dual-Pearson approach, which employs a weighted sum of two Pearson functions to simulate the profiles. The computationally efficient model covers the following range of implant parameters: implant species B, BF2, As; implant energies from 1 keV to 15 keV; any dose; tilt angles from 00 to 100; all rotation angles (0'-360'). This experimental analysis is important for the development of scaled devices with ultra-shallow junctions, and the computationally efficient model will enable process simulators to predict ultra-shallow as-implanted profiles accurately. INTRODUCTION BF2+ ion implantation has developed into one of the popular techniques used for the formation of ultra-shallow p+-n junctions. Its primary advantage over the conventional B+ ion implantation is the realization of extremely shallow boron profiles due to the much lower energy of boron, as a result of the energy partition between boron and fluorine [1, 21. Implantation of heavy arsenic ions is widely used for fabrication of shallow n+-p junctions. Thus a better understanding of the dependence of dopant profiles obtained by low energy B, BF2, and As ion implantation is essential. Also, an accurate and computationally-efficient model for the simulation and optimization of low energy ion implantation is necessary. EXPERIMENTAL PROCEDURE One hundred and seventy five 125 mm (diameter), (100) single-crystal silicon wafers were used for the experiments. Since a native oxide layer of even a few nanometer thickness can affect the ultra-shallow implant profiles, HF (hydrofluoric acid) etching was performed for all of the wafers just prior implantation, thus ensuring75essentially an oxide-free silicon surface [3]. The waer wrto the iplnedwih11 wafers were implanted with B, BF2, or As- at implant energies of 5 or 10 keV, and doses 12 -215 ranging between 2x1012 cm and 2xl015 cm-2 for a wide range of implant angles using an electrostatically-scanned Eaton NV-6200AV ion implanter. The crystal cut error was found to be less than 10 from Therma-Wave maps. The half angle beam divergence of this implanter has been determined to be less than 1'. The tilt angle for ion implantation is defined as the angle between the 769 Mat. Res. Soc. Symp. Proc. Vol. 396 01996 Materials Research Society
incident ion beam vector and a vector perpendicular to the wafer with the axis of rotation being the flat
Data Loading...
