Search

Abstract: The relative sensitivity factor (RSF) plays a critical role in the quantification of nitrogen concentration in oxynitride gate oxide measured using secondary ion mass spectrometry (SIMS). The RSF depends primarily on the matrixes where the nitrogen resides under conditions in which the parametric settings of the SIMS equipment remain unchanged. In this article, different types of oxynitride dielectric films have been prepared under several different fabrication processes including decoupled plasma nitridation, thermal nitridation by nitric oxide (NO) and nitrous oxide (N2O), which resulted in the formation of significantly different chemical bonding and micro-structures for the oxynitride. The nitrogen RSF values of these oxynitride films were determined using the ratio of the species SiN− and 30Si− from the Time-of-Flight SIMS. It was found that the overall RSF values from the point-by-point algorithm ranged from 1.15×1021 to 2.41×1021 atoms/cm3. These account for about 2 times the difference in the RSF values. [Copyright &y& Elsevier]

Abstract: Roughness development is one of the most often addressed issues in the secondary ion mass spectrometry (SIMS) ultra-shallow depth profiling. The effect of oxygen flooding pressure on the roughness development has been investigated under the bombardment of 500eV O2 + beam with simultaneous sample rotation. Oxygen flooding had two competing effects on the surface roughening, i.e., enhancement of initiating roughening and suppression of roughening development, which were suggested to be described by the onset depth z on and transient width w tr of surface roughening. Both z on and w tr decreased as oxygen flooding pressure increased. As the result, surface roughening was most pronounced at the intermediate pressure from 4.4E−5Pa to 5.8E−5Pa. The surface roughening is negligible while without flooding or with flooding at the saturated pressure. No flooding is preferable for depth profiling ultra-shallow B implantation because of the better B profile shape and short analysis time. [Copyright &y& Elsevier]

Abstract: As the wafer size continues to increase and the feature size of the integrated circuits (IC) continues to shrink, process control of IC manufacturing becomes ever more important to reduce the cost of failures caused by the drift of processes or equipments. Characterization tools with high precision and reproducibility are required to capture any abnormality of the process. Although Secondary ion mass spectrometry (SIMS) has been widely used in dopant profile control, it was reported that magnetic sector SIMS, compared to quadrupole SIMS, has lower short-term repeatability and long-term reproducibility due to the high extraction field applied between sample and extraction lens. In this paper, we demonstrate that CAMECA Wf can deliver high long-term reproducibility because of its high-level automation and improved design of immersion lens. The relative standard deviation (R.S.D.) of the relative sensitivity factors (RSF) of three typical elements, i.e., boron (B), phosphorous (P) and nitrogen (N), over 3 years are 3.7%, 5.5% and 4.1%, respectively. The high reproducibility results have a practical implication that deviation can be estimated without testing the standards. [Copyright &y& Elsevier]

Abstract: Ultra-thin silicon oxynitride (SiO x N y ) is the leading candidate to replace pure silicon oxide (SiO2) before high k dielectrics come into place because oxynitrides demonstrate several properties superior to those of the conventional gate oxides. The performance of the transistor was reported to depend on the N dose and its distribution in the gate oxide. Therefore, accurate characterization of SiO x N y is prerequisite to control the quality of the ultra-thin nitrided gate oxide. However, secondary ion mass spectrometry (SIMS) faces big challenges in analyzing ultra-thin gate oxide because of surface effect and matrix effect. In this work, MCs2 + (M stands for matrix element) was detected to reduce the matrix effect in depth profiling the ultra-thin oxynitride. However, N profile was very close to the top surface if the oxynitirde was fabricated by decoupled plasma nitridation (DPN). With the conventional approach, the N dose was overestimated and the N profile was distorted near the top surface. To obtain a reliable N profile, the oxynitride was capped with a thin layer of oxide. The N profile of interest was hence in the region of sputtering equilibrium, i.e. the surface effect was minimized. As the results, reliable N dose and profile have been obtained. [Copyright &y& Elsevier]