In recent years, ion implantation has become one of the key techniques in semiconductor fabrication. The annealing of the damage produced during implantation is, however, not fully understood. Ion implantation at high temperatures allows the time-resolved study of implantation-enhanced diffusion. During the process, point defects are generated by the ion implantation and consumed by recombination in the bulk as well as by diffusion to the surface and recombination there. With increasing temperatures, the recombination of point defects, which are acting as diffusion vehicles, results in reduced effective diffusion. Profiles processed above 900 show marked uphill diffusion at the surface caused by large gradients of the point defect concentrations. This uphill diffusion affirms the generally accepted pair diffusion theories. Since the point defects are in steady state even after process times which are short compared to the total process time, we are able to give a qualitative analysis of the dose dependence of the diffusion. By extensive numerical simulations, we could estimate the product of bulk recombination rate and equilibrium concentrations of self-interstitials and vacancies as well as the interface recombination velocity for the self-interstitials. The results obtained are in qualitative agreement with previous work of others. The results demonstrate, in fact, clearly the advantages of the method presented. But due to experimental problems concerning the temperature measurement, which have not been fully resolved up to now, the results have to be considered as crude estimates.

Modeling and numerical simulation of crystal growth of Si film and heat transport in 3D structure were made for optimization of physical and geometrical parameters used during laser recrystallization. Based on simulations a new concept called micro-absorber was introduced for obtaining defect-free Si films.

This paper is concerned with the stress simulation of a LOCOS structure during not only oxidation but also the subsequent cooling down based on viscoelastic stress modeling. A viscoelastic model is successfully applied to the oxide, nitride and silicon substrate for a LOCOS structure. Thermal stress is also taken into account during the cooling down process. The viscoelastic deformation problem of all the three materials for the LOCOS structure are solved by a two-dimensional finite element method. It is the first time to show that the stress values after cooling down to room temperature are much higher than those right after oxidation. It is also shown that varying the cooling down rates results in the different stress values after cooling down.

In this paper, a three part algorithm is employed to obtain stable convergence during stress dependent oxidation simulation using the finite element method is presented. By introducing (1) a reduced integration formulation, (2) an averaging procedure for the mid-side node velocities at the Si/SiO₂ interface, and (3) a three-node element to discretize the oxidant diffusion equation, major improvements in achieving stable convergence are realized during stress dependent oxidation simulation. This technique is generally applicable for an oxidation simulator using the finite element method.

We present a new apparent bandgap narrowing model for semiconductor device simulation. The new model is derived from revised data of previous measurements on the apparent bandgap narrowing by using a corrected intrinsic carrier concentration. The revised values reveal sufficient agreement with our theoretical calculation. The new model is implemented in a triangular mesh device simulator TRIMEDES. Simulated BJT current-voltage and current-temperature characteristics using the proposed model reveal excellent agreement with measurements.

Low-field mobilities and velocity versus electric field relations are among the key input parameters for drift-diffusion simulations of field-effect and bipolar transistors. For example, most device simulations that treat scattering from ionized impurities contain mobilities or velocity versus field relations based on the Born approximation (BA). The BA is insensitive to the sign of the charged impurity and is especially poor for ionized impurity scattering because of the relatively strong scattering of long-wavelength carriers, which have low energies, and therefore violate the validity condition for the BA. Such carriers occur at high symmetry points in the Brillouin zone and are critical for device behavior. There has been a tendency in the past to assume that majority and minority mobilities are equal. This assumption can lead to incorrect interpretations of device data and thereby misleading design strategies based on such simulations. We have calculated the majority electron and minority hole mobilities in GaAs at 300 K for donor densities between 510¹⁶ and 110¹⁹ cm^-3 and the majority hole and minority electron mobilities for acceptor densities between 510¹⁶ and 110²⁰ cm^-3. We have included all the important scattering mechanisms for GaAs: acoustic phonon, polar optic phonon, nonpolar optic phonon (holes only), piezoelectric, ionized impurity, carrier-carrier, and plasmon scattering. The ionized impurity and carrier-carrier scattering processes have been calculated with a quantum mechanical phase-shift analysis to obtain more accurate matrix elements for these two scattering mechanisms. We compare the total scattering rate for majority electrons due to ionized impurities based on exact phase shifts and on the BA used by Brooks-Herring. We also present additional data that show the differences between the exact phase-shift analyses and the BA for majority electron scattering rates as functions of carrier energy and scattering angle. These results show that the calculated low-field mobilities are in good agreement with experiment, but they predict that at high dopant densities minority mobilities should increase with increasing dopant density for a short range of densities. This effect occurs because of the reduction of plasmon scattering and the removal of carriers from carrier-carrier scattering because of the Pauli exclusion principle. Some recent experiments support this finding. These results are important for device modeling because of the need to have reliable values for the minority mobilities and velocity-field relations.

This work describes a new analytical MOSFET model for analog circuit simulation based on the charge-sheet model. The current equation consists of diffusion and drift components, therefore I_ds is a smooth function of the applied voltages. Since the original charge-sheet model is valid only for long-channel transistors, it has been further developed to describe quarter-micron MOSFETs by introducing the lateral electric field E_y into the theory. The new model includes these field contributions self-consistently, and describes the drain current of MOSFETs from long to quarter-micron channel lengths with a single model parameter set without discontinuities in derivatives of the drain current I_ds. The mobility reduction due to E_y is described by an empirical equation with physical parameter values taken from literature. Only two fitting parameters, the impurity scattering and the surface roughness scattering in the mobility equation, are added to the physical parameters. The subdiffusion lengths are also taken as fitting parameters. Though the new model reduces the number of fitting parameters totally to four, it reproduces measured I_ds excellently for MOSFETs with all channel lengths. The model has been included in the parameter extraction program JANUS, which extracts model parameters automatically. The algorithm for parameter extraction is summarized.

In this paper, a "hydrodynamic" version of the three-dimensional code HFIELDS-3D is used to achieve a detailed knowledge on the distribution of the substrate current inside a recessed-oxide MOSFET. The physical model features a temperature-dependent formulation of the impact-ionization rate, allowing non-local effects to be accounted for. The discretization strategy relies on the Box Integration scheme and uses suitable generalizations of the Scharfetter-Gummel technique for the energy-balance equation. The simulation results show that the narrow-channel effect has a different impact on drain and substrate currents. Further three-dimensional effects, such as the extra heating of the carriers at the channel edge, are demonstrated.

A Monte Carlo calculation is performed to examine the transport coefficients of the electron gas under an inhomogeneous electric field. The expressions constructed from the M. C. results are then incorporated into the hydrodynamic formulation to calculate the internal characteristics of a silicon BJT device. The calculated results agree well with the Monte Carlo prediction.

A spherical-harmonics expansion method is used to find approximate numerical solutions of the Boltzmann Transport Equation in the homogeneous case. Acoustic and optical phonon scattering, ionized impurity scattering as well as an energy band structure fitting the silicon density of states up to 2.6 eV above the conduction-band edge are used in the model. Comparisons with Monte Carlo data show excellent agreement, and prove that detailed information on the high-energy tail of the distribution function can be obtained at very low cost using this methodology.

In two-dimensional simulation of thin-base RHET, we combined three different simulation methods--the Schrödinger equation, the Monte Carlo simulation, and two-dimensional device simulation within a drift and diffusion model. We found that, in the thin-base RHET, the potential distribution differs from that expected from the thick-base RHET. In the thin-base RHET, the potential of the intrinsic base region does not equal that of the base electrode because the intrinsic base region is depleted and the negative emitter voltage (V_EB0) raises the potential of both the intrinsic base and the nondoped region under the intrinsic base. There are also modified by the collector voltage. We also show emitter current-voltage characteristics, transfer ratio, and transit time calculated using this method and compare them with results for the one-dimensional case.

We have developed a new model to analyze the thermal failure mechanism due to electrical-over-stress (EOS) for two-dimensional planar pn-junction structures where the failure power is proportional to about 1/5 power of the failure time. We adopted a pseudo two-dimensional numerical simulation method where a pn-junction diode is divided into small elements and represented by a circuit network composed of many minute resistors and diodes. The failure mechanism studied by Wunsch and Bell, that is one of many studies for one-dimensional pn-diodes, is not valid for the case of two-dimensional pn-junction, such as a planar type junction. On the contrary, the failure mechanism was found to be much correlative with the junction structure, especially the impurity concentration in the substrate in the two-dimensional case. When the impurity concentration in the substrate is high enough (e.g. N_sub10¹⁷[cm^-3]), the breakdown occurs at the whole junction. The heat transfer is one-dimensional and the failure power is proportional to about 1/2 power of the failure time, which is well known results reported by many researchers: e.g. Wunsch &Bell. On the other hand, when the impurity concentration in the substrate is low enough (e.g. N_sub10¹⁶[cm^-3]), the breakdown occurs locally at the junction edge. The heat transfer is two-dimensional and the failure power is in proportion to about 1/5 power of the failure time.

An automatic optimization system for semiconductor devices has been built-up by fully interfacing an optimizer and a device-analysis code supplemented with sensitivity analysis. The device-analysis code is thought of as a part of a pipeline of simulators. The latters are regarded as subprocesses by the optimizer, which controls their I/O stream. The action of the pipeline is iterated until the optimum set of design parameters is determined. An important feature of the system is that all the derivatives required in the sensitivity analysis are calculated analytically, this providing a substantial improvement in both the numerical accuracy and computational efficiency, and making the scheme attractive from the application standpoint. A few examples of optimization of MOS devices are shown and the performance is reported, indicating that a system of this kind can usefully be exploited in a design environment.

We have developed an efficient general-purpose two-dimensional device simulation system which consists of a solver, and pre- and post-processors. This system can easily handle any complicated device having a non-rectangular shape. It can also be applied to compound semiconductor devices with heterojunctions, including optical devices such as laser diodes. In order to handle any device, a new program for construction of device geometry is developed as a preprocessor. It has an efficient graphic interface to reduce the time required to input data for simulations, which is a very time consuming task for complicated devices. A new efficient data structure representing device geometry is introduced in the program. During postprocessing, any physical quantity can be displayed on the multi-window screen. In addition, a general-purpose solver for basic semiconductor equations is implemented in the system. Using this system, any device can be successfully analyzed in a unified manner and the turn-around time for the simulation is significantly reduced.

This paper describes a unified process and device simulation system named P &D Workbench (Process and Device Workbench). The P &D Workbench is an EWS (Engineering Work Station) based system which is connected with MFCs (Main Frame Computers) via networks and can easily execute 2-dimensional process, device, topography and capacitance simulations. Since the P &D Workbench has a supervisor, data-base and excellent user interface using Japanese menu functions and mouse operations, a handling time can be dramatically reduced. The supervisor controls the simulation sequence and file transfer, and manages jobs and files both on EWSs and MFCs, so that plural simulations of splitting conditions can be automatically executed. Short TAT (Turn Around Time) is achieved by selecting an appropriate platform depended on a problem size and MFCs' CPU loads. The effects of the P &D Workbench are shown in examples applied to the development of a 4M-DRAM.

A yellow light emitting display using neon-hydrogen-argon mixture as filling gas is presented. Strong "monochromatisation" of the emitted light is reported for the first time on the wavelength λ585.3 nm. Experimental results on the dependence of the "monochromatisation effect" is given for various pressure values and filling gas composition. It is underlined the existence of a process of selective population of the upper level 3p[1/2]₀ of the transition corressponding to the wavelength 585.3 nm. The obtained results are discussed in relation with the reported results on yellow light laser in which a discharge in neon-hydrogen mixture is used for laser radiation generation at λ585.3 nm. The proposed explanation of different authors on the upper level population through radiative or dissociative recombination of neon ions is discussed and a new hypothesis is advanced for the strong monochromatisation observed in neon-hydrogen or neon-hydrogen-argon filled displays. According to this hypothesis, in the feeding process of the upper level 3p[1/2]₀ are taking part the neon metastable states too. If such an assumption will come true, cyclic processes in yellow light generation might appear.

To realize a highly efficient small antenna, high-Tc superconductors are adopted to fabricate both a self-resonating helical radiator and a quarter-wave matching circuit. The actual gain and bandwidth measured at 478 MHz using a 1/45-wavelength radiator were respectively 1.5 dBi and 0.35%, indicating that this type of antenna has a high radiation efficiency and a fairly wide bandwidth. It is also confirmed through experiments and theoretical simulations that a decrease in the surface resistance of the radiator more effectively improves the radiation efficiency than a decrease in the surface resistance of the matching circuit.

A new monolithic transmit-receive GaAs FET switch has been developed, named the FET series-shunt connected TR switch and capable of switching high rf transmitting power. Both insertion loss and isolation limitations of this type TR switch have been analyzed using the switching cutoff frequency of the control FET, and the formula for calculating the rated power is provided. A unique feature of this switch is that the power handling of the switch is not limited by the FET gate break-down voltage but is limited by the saturation current, so higher handling power capability is available by using FETs with a larger gate periphery. A design example of the TR switch at a rated power of 8 W in the transmit mode as well as the results of an X band switch are presented.

In this paper, the eigen-function weighted boundary integral epuation method (EW-BIEM) is applied to analyze the dispersion characteristics of various planar transmission lines with finite metallization thickness, such as the micro-strip lines, conductor-backed coplanar waveguides and micro-coplanar striplines for the first time. Due to the choice of the eigen-functions as weighted functions instead of Green's function, the computational time is shortened to a great extent and the singularity problems are also avoided. The difficulty in treating strip thickness can be overcome by considering the 90 edge on the strip as a 90 circular arc whose radius tends to zero. The computational results clearly demonstrate that the effects of finite strip thickness on the propagation properties of these transmission lines can be treated easily and efficiently with this method.