Articles you may be interested inHigh-resolution transmission electron microscopy of as-deposited boron nitride on the edge of ultrathin Si flake An edge effect observed in high-resolution scanning Auger-electron microscopy is modeled by Monte Carlo methods based on the direct simulation of individual inelastic scattering processes. The Monte Carlo method permits the simulation of the spatial distribution of excited electrons generated by inelastic scattering. The results suggest that high-energy excited ("secondary") electrons are a significant source of Auger electrons (L-MM transition) in aluminum and play an important role in causing the edge effect.

Experiments have been conducted to determine the feasibility of the theoretical experiment set forth by Cazaux. The experiment consists of electron bombarding a thin foil of aluminum with a thin specimen mounted on the opposite side. A focused scanning electron beam is used to bombard the aluminum substrate to produce a localized source of AlKα x rays. This focused Al x-ray (hν=1.4 keV) source creates a spatially localized source of photoelectrons in the specimen. The photoelectrons are separated in energy with a cylindrical mirror spectrometer to produce an ESCA spectrum from an area of less than 20 μm in diameter. Two-dimensional ESCA images have been obtained by this method utilizing intensity variations in the amplitude of a selected photoelectron peak to intensity modulate a cathode ray tube monitor as the electron beam is scanned.

This paper reports on silicon nitride formation by low energy implantation of nitrogen or ammonia into silicon. Extensive material investigation on nitrogen-implanted films using various analytical methods, including ellipsometry, chemical etching, transmission electron microscopy (TEM), x-ray photoelectron microscopy (XPS), infrared transmission spectroscopy (IR), and Rutherford backscattering spectroscopy (RBS) is discussed. A new technique to derive the film thickness, density, refractive index, and dielectric constant is developed. This method employs a combination of ellipsometry, capacitance measurement, and nitrogen areal density obtained by RBS. A major finding indicates that the implanted nitride films have a low density compared to CVD nitride. Other material properties are also summarized.Thermal nitridation and nitridation of oxide have been shown to have certain advantages over thermal oxidation (1-3). The thin dielectric obtained is very uniform and exhibits high breakdown strength. In addition, the nitrided layer has a higher dielectric constant than oxide and is impervious to impurity diffusion. The film is also less susceptible to process-induced gate failures (3). With the scaling of the gate dielectric in VLSI, thin nitride is a potential candidate to be investigated.We have shown that thin nitride can be formed by using low energy implantation of N2 in an ion milling machine (4). Since the ion milling machine can deliver a high current density over a large area, the through-put of this process is comparable with single-wafer processing. The implantation of N2 is accompanied by surface sputtering. The sputter-implantation process produces a very uniform layer. Rutherford backseattering spectroscopy (RBS) indicates that the total nitrogen incorporated into the Si surface is about 3.2 -3.5 • 1018 cm -2 at 1.7 keV. The as-implanted layer has a N/Si ratio of 1. A damaged layer about 3 -4 nm thick at the interface extending into the Si substrate is also present. After thermal treatment, the ratio approaches the stoichiometric nitride value of 1.3. The damage to the Si substrate is reduced but not totally removed after annealing at 900~ for 1/2h. A damaged layer about 0.8 nm thick remains. Transmission infrared spectroscopy shows the existence of N-Si bonding in both the as-implanted and the annealed samples. However, the transmission minimum shifts from about 850 to 800 cm -I after heat-treatment. ExperimentsThe substrates used in the studies were (100)-oriented silicon wafers. The native oxide is removed before loading into an ion milling machine (Veeco Microetch System). The system is first pumped down to below 2 • 10 -6 torr and then the implantation is carried out with nitrogen or ammonia at a pressure of 8 • 10 -5 torr. The total dose of the implant is always greater than 0.3 C/cm ~. The samples are then annealed in dry nitrogen or oxidized in wet oxygen. Results Ellipsometry measurements.--Ellipsometry is oftenused to obtain the thickness and the refractive index of a dielectric layer independ...