Finite element analysis of advanced multilayer capacitors

We establish a systematic methodology to design and analyse electromagnetic components such as advanced multilayer ceramic capacitors (MLCCs) using the finite element (FE) method. We employ a coupled formulation to compute the interaction between the electric and magnetic fields. Unlike a linear distribution of current assumed in the circuit model, an accurate electrostatic solution to model the entire advanced MLCCs (4 × 4 × 27 = 432 cells) is presented. The FE solution is used to compute the lumped parameters for a range of frequencies. These lumped parameters are then used to compute the parasitic elements of the MLCCs. We introduce two algorithms to efficiently analyse the behaviour of a capacitor with changing frequency. The lower frequency (much below the self‐resonant frequency of the capacitor) algorithm separates the effect of the electric and magnetic fields and reduces the computational effort required to solve the FE problem, whereas, the high‐frequency algorithm couples the effect between the electric and the magnetic fields. We use these algorithms in conjunction with a new multiple scale technique to effectively determine the small values of R, L and C in MLCCs. The formulation, the implementation, and the numerical results demonstrate the efficacy of the present FE formulation and establish a systematic methodology to design and analyse advanced electromagnetic components. Copyright © 2003 John Wiley & Sons, Ltd.     

Quantum mechanical effects on noise properties of nanoelectronic devices: application to Monte Carlo simulation

A discussion about the quantum mechanical effects on noise properties of ballistic (phase-coherent) nanoscale devices is presented. It is shown that quantum noise can be understood in terms of quantum trajectories. This interpretation provides a simple and intuitive explanation of the origin of quantum noise that can be very salutary for nanoelectronic engineers. In particular, an injection model is presented that, coupled with a standard Monte Carlo algorithm, provides an accurate modeling of quantum noise. As a test, the standard results of noise in tunneling junction devices are reproduced within this approach.     

Mass Transfer in a Flat Gas/Liquid Interface using non‐Newtonian Media

Gas/liquid mass transfer has been investigated using a stirred vessel gas/liquid contactor using non‐Newtonian media and carbon dioxide as absorbent and gas phase, respectively. The volumetric mass transfer coefficients at different operational variables have been determined. Non‐Newtonian media (liquid phase) were prepared as aqueous solutions of sodium carboxymethyl cellulose salt. The influence of the rheological properties, polymer concentration, stirring rate, and gas flow rate on mass transfer was studied for these liquid phases. Kinematic viscosity and density experimental data were used to calculate the average molecular weight corresponding to the polymer employed. The Ostwald model has been used to fit the rheological behavior of aqueous solutions of the polymer employed as absorbent phase. Reasonably good agreement was found between the predictions of the proposed models and the experimental data of mass transfer coefficients.     

Asymptotic performance of single parity-check product codes

This article investigates the asymptotic performance of single parity-check (SPC) product codes (PCs) from a decoding point of view. Specifically, the probability of bit error is bounded before and after the decoding of each dimension, similar to the analysis of "iterated codes" by Elias (1954). It is shown that the asymptotic probability of bit error can be driven to zero as the number of dimensions, and hence the block length, increases at signal-to-noise ratios (SNRs) within 2 dB of capacity over the additive white Gaussian noise (AWGN) channel.     

Full-diversity, high-rate space-time block codes from division algebras

We present some general techniques for constructing full-rank, minimal-delay, rate at least one space-time block codes (STBCs) over a variety of signal sets for arbitrary number of transmit antennas using commutative division algebras (field extensions) as well as using noncommutative division algebras of the rational field /spl Qopf/ embedded in matrix rings. The first half of the paper deals with constructions using field extensions of /spl Qopf/. Working with cyclotomic field extensions, we construct several families of STBCs over a wide range of signal sets that are of full rank, minimal delay, and rate at least one appropriate for any number of transmit antennas. We study the coding gain and capacity of these codes. Using transcendental extensions we construct arbitrary rate codes that are full rank for arbitrary number of antennas. We also present a method of constructing STBCs using noncyclotomic field extensions. In the later half of the paper, we discuss two ways of embedding noncommutative division algebras into matrices: left regular representation, and representation over maximal cyclic subfields. The 4/spl times/4 real orthogonal design is obtained by the left regular representation of quaternions. Alamouti's (1998) code is just a special case of the construction using representation over maximal cyclic subfields and we observe certain algebraic uniqueness characteristics of it. Also, we discuss a general principle for constructing cyclic division algebras using the nth root of a transcendental element and study the capacity of the STBCs obtained from this construction. Another family of cyclic division algebras discovered by Brauer (1933) is discussed and several examples of STBCs derived from each of these constructions are presented.     

Optimal mean-square prediction of the mobile-radio fading envelope

Long-range prediction of the mobile-radio fading envelope is an enabling technology for many fading compensation approaches. Because the fading envelope is well modeled as a bandlimited process, it has special predictability properties. In this paper, we find a linear predictor that is optimal in the mean-square sense when the predictor impulse response is energy constrained. This solution may be used to determine the minimum mean squared error of a prediction based on past values that are corrupted with estimation errors.