T. Smy

A lateral magnetotransistor structure with a linear response to the magnetic field

An experimental study and analytical model of a novel magnetotransistor are presented. This device displays some very promising features. A linear response to the magnetic field is experimentally demonstrated and very high sensitivities are measured, on the order of 3000%/T. Previous comparable magnetotransistors have reported sensitivities on the order of 150%/T. A theoretical explanation of the very high sensitivity is proposed, involving carrier deflection as the dominant operating principle. >

Suppressed sidewall injection magnetotransistor with focused emitter injection and carrier double deflection

We present a new linear magnetic field sensor (MFS) made in standard CMOS technology with a sensitivity of 1.26 percent/mT (1 percent/T ≡ 0.01 T-1). The device is a dual-collector lateral magneto-transistor (LMT) with suppressed injection of the emitter sidewalls, confinement of the injection to the center bottom of the emitter-base junction, and double deflection of carriers. Desirable collector current levels can be set without major loss of sensitivity by choosing from a wide range of operating points.

Simulation of elevated temperature aluminum metallization using SIMBAD

A ballistic deposition model, SIMBAD, has been extended to simulate physical vapor deposition onto substrates at elevated temperatures. The model has been expanded to account for the effect of film curvature on surface diffusion. The effects on via coverage and filling have been simulated for aluminum films, and complete planarization of a 1:1 aspect ratio via is predicted for a temperature of 550 degrees C. Via aspect ratio and sidewall taper can also strongly affect coverage and filling. Biased sputtering has also been incorporated into the model and shows that a primary effect is a substantial reduction in the temperature required to achieve full planarization. However, void formation and substrate damage are problems predicted to occur under some bias sputter conditions. >

Simulation of density variation and step coverage for a variety of via/contact geometries using SIMBAD

The two-dimensional simulation by ballistic deposition of sputtered metal deposited over a 1.5- mu m step and a large variety of vias and contacts is presented. The sizes of the vias and contacts are varied from 1 to 3 mu m, and three different sidewall topographies are simulated. In addition, simulated film growth over a stacked via/contact is presented. The step coverage of each film is determined, and surface profiles are provided at different growth intervals. The use of SIMBAD, a ballistic deposition technique, provides information unattainable through the use of conventional film deposition simulations. In addition to the step coverage available from conventional simulations, density profiles of the simulated films are produced and the columnar microstructure is analyzed. Finally, conclusions are inferred as to the quality of real films deposited over each via geometry. >

Simulation of the microstructure of chemical vapor deposited refractory thin films

A ballistic deposition model (SIMBAD) has been extended to provide qualitative cross‐sectional depictions of the microstructure present in chemical vapor deposited (CVD) films. The model qualitatively depicts the pronounced columnar structure typical of refractory metal, nitride, and silicide films−especially when deposited over integrated circuit topography. The important factors affecting thin film microstructure are seen to be flux shadowing, precursor surface diffusion, and a nonunity sticking coefficient. While the conformal step coverage typical of refractory CVD films is primarily due to a low sticking coefficient, the detailed columnar structure is the result of all three of these mechanisms. The angular distribution of the incident precursor flux is important to the shadowing mechanism, and a sticking coefficient‐dependent angular distribution relevant to CVD is presented. Variations of the model to represent selective deposition (including selectivity loss) are also shown.

A ballistic deposition model for films evaporated over topography

Abstract MgF2 films were evaporated at angles of incidence of 5° and 40° onto oxide lines 1.5 μm wide on a silicon wafer. The film properties observed by scanning electron microscopy studies of cleaved cross-sections were compared with those predicted by simulation by a ballistic deposition model. Good agreement was found for the surface profiles of the films, the orientation of the columnar microstructure and the location of porous low density regions.

Study of the copper reflow process using the GROFILMS simulator

Copper has been proposed as a possible replacement for aluminum as the primary microelectronics metallization material. As the feature sizes of ultralarge scale integration (ULSI) technologies reach sub-0.25 μm in dimension, new metallization processes must be developed to properly fill the extreme topographies. This article uses GROFILMS (a thin film process simulator) to study the Cu reflow process. Comparison of simulation with experimental scanning and transmission electron micrographs show excellent agreement with regards to profile evolution and internal structure. GROFILMS is further used to study material dependent properties such as the underlayer wetting characteristics of Cu. Results show that during both the deposition and the anneal procedures, Cu’s ability to wet the underlayer can determine success or failure of the reflow process. Further, the effects of the film microstructure such as grain boundary grooving and facetting are simulated, observed in real films, and discussed in terms of pr...

Spatial and angular nonuniformities from collimated sputtering

A Monte Carlo collimated sputtering model, SIMSPUD, has been extended to examine the effects of collimation on the uniformity of spatial and angular distributions of flux incident to the substrate. The model predicts that substantial periodic variations can occur in the film thickness due to the proximity of the collimator. Experimental measurements closely confirm these predictions. In addition, the model predicts considerable periodic nonuniformity in the angular distribution over the same lengthscale. This will cause substantial variations in sidewall and bottom coverage inside topography as demonstrated using the film growth simulation, SIMBAD. Finally, collimation is seen to effectively transfer nonuniform sputter target erosion profiles to film thickness profiles, especially at higher aspect ratios. All three of these nonuniform effects have the potential to adversely affect the yield, performance, and reliability of films deposited using collimated sputtering.

Simulation by ballistic deposition of local density variation and step coverage for via metallization

The simulation is presented of ballistic deposition of sputtered metal deposited over a via. The step coverage of the film is determined and surface profiles provided at different film thicknesses. The use of a ballistic deposition technique provides additional information unattainable through the use of conventional film deposition simulations. A density profile of the film is produced and the columnar microstructure analyzed.<<ETX>>

Integrating system and feature scale models to study copper reflow

The reflow of sputtered copper is a microelectronics metallization procedure compatible with damascene processing [Murarka et al., MRS Bull. 18, 46 (1993)]. During the reflow process, copper films are deposited via traditional sputter technologies and subsequently made to fill submicron trenches and vias through surface diffusion enhanced by annealing [Gardner and Fraser, in 1995 Proceedings of the VLSI Multilevel Interconnection Conference, edited by T. E. Wade (VMIC, Tampa, FL, 1995), p. 287]. This article studies the copper reflow processing parameters through simulation using the process simulators SIMSPUD and GROFILMS. Simulation results are presented as depictions of the film on the feature scale and as an analysis of the process parameter space. In particular, factors such as initial film profiles and anneal temperatures are shown to be important considerations in developing a successful Cu reflow process. The effects of the deposition procedure on the reflow process are presented regarding system configuration, copper film thickness, and trench aspect ratios. The deposition system parameters are shown to be important when multiple feature sizes are to be filled concurrently.The reflow of sputtered copper is a microelectronics metallization procedure compatible with damascene processing [Murarka et al., MRS Bull. 18, 46 (1993)]. During the reflow process, copper films are deposited via traditional sputter technologies and subsequently made to fill submicron trenches and vias through surface diffusion enhanced by annealing [Gardner and Fraser, in 1995 Proceedings of the VLSI Multilevel Interconnection Conference, edited by T. E. Wade (VMIC, Tampa, FL, 1995), p. 287]. This article studies the copper reflow processing parameters through simulation using the process simulators SIMSPUD and GROFILMS. Simulation results are presented as depictions of the film on the feature scale and as an analysis of the process parameter space. In particular, factors such as initial film profiles and anneal temperatures are shown to be important considerations in developing a successful Cu reflow process. The effects of the deposition procedure on the reflow process are presented regarding system ...