Parasitic bipolar gain in fully depleted n-channel SOI MOSFET's

Fully depleted SOI MOSFET's include an inherent parasitic lateral bipolar structure with a floating base. We present here the first complete physically based explanation of the bipolar gain mechanism, and its dependence on bias and technological parameters. A simple, one-dimensional physical model, with no fitting parameters, is constructed, and is shown to agree well with simulations and measurements performed on a new type of SOI MOSFET structure. It is shown that parameters which affect the gain, such as SOI layer thickness, body doping concentration and gate and drain voltages, do so primarily by affecting the concentration of holes in the body region. Thus, current gain falls dramatically with increasing drain voltage due to the associated impact ionization driven increase in the hole concentration. Gummel plots of this parasitic bipolar indicate an apparent ideality factor of 0.5 for the hole current, due to the body hole concentration's dependence on drain voltage     

Fabrication of low-loss IR-transmitting Ge30As10Se30Te30 glass fibers

Improved purification and processing techniques have been utilized to fabricate Ge₃₀As₁₀Se₃₀Te₃₀ glass fibers with a minimum loss of O.11 dB/m at 6.6 μm. This is the lowest loss reported for any telluride glass fiber in the infrared region. Furthermore, the fibers exhibit less than 1 dB/m loss between 5.25 and 9.5 μm     

Cure monitoring of thermoset coatings

The combination of favorable features of tensile mode dynamic mechanical analysis and torsional braid analysis leads to a useful method for cure monitoring of thermoset coatings. This sensitive technique allows the observation of cure reactions over a wide time and temperature range, thus including gelation and long-term curing in one measurement. Isothermal baking can be simulated realistically, since evaporation of volatile compounds like solvents or blocking agents is not hindered and the sample film is in direct contact with the heating atmosphere. Examples are given showing the cure behavior of automotive clear coats and electro coats depending on parameters like crosslinker type and catalytic agents.     

Food Process Engineering: The Last 25 Years and Challenges Ahead

In the first part of this contribution, an overview is given of some of the main developments in food process engineering in the last 25 years of the 20th century. This overview is, of course, colored by the personal experience of the authors, but a sincere effort was made to maintain a general perspective. Topics that will be briefly discussed are: progress in understanding how to control food microstructure formation during processing, separation processes, conversion processes and stabilization processes, progress in flavor technology and understanding of flavor retention during processing and release. In the 2nd part, in our view, the most exiting future developments are briefly discussed. The major items here are: processing requirements for functional foods, integrated process design approaches, application of novel‘fields’in food processes,‘precision’processing, supply chain approaches to food manufacturing, and more.     

Digital filter bank design quadratic-constrained formulation

Formulate the filter bank design problem as an quadratic-constrained least-squares minimization problem. The solution of the minimization problem converges very quickly since the cost function as well as the constraints are quadratic functions with respect to the unknown parameters. The formulations of the perfect-reconstruction cosine-modulated filter bank, of the near-perfect-reconstruction pseudo-QMF bank, and of the two-channel biorthogonal linear-phase filter bank are derived using the proposed approach. Compared with other design methods, the proposed technique yields PR filter banks with much higher stopband attenuation. The proposed technique can also be extended to design multidimensional filter banks     

Surface tailoring of an ethylene propylene diene elastomeric terpolymer via plasma‐polymerized coating of tetramethyldisiloxane

In the research presented here, we explore the use of a low‐energy plasma to deposit thin silicone polymer films using tetramethyldisiloxane (TMDSO) (H(CH₃)₂? Si? O? Si? (CH₃)₂H) on the surface of an ethylene propylene diene elastomeric terpolymer (EPDM) in order to enhance the surface hydrophobicity, lower the surface energy and improve the degradation/wear characteristics. The processing conditions were varied over a wide range of treatment times and discharge powers to control the physical characteristics, thickness, morphology and chemical structure of the plasma polymer films. Scanning electron microscopy (SEM) shows that pore‐free homogeneous plasma polymer thin films of granular microstructure composed of small grains are formed and that the morphology of the granular structure depends on the plasma processing conditions, such as plasma power and time of deposition. The thicknesses of the coatings were determined using SEM, which confirmed that the thicknesses of the deposited plasma‐polymer films could be precisely controlled by the plasma parameters. The kinetics of plasma‐polymer film deposition were also evaluated. Contact angle measurements of different solvent droplets on the coatings were used to calculate the surface energies of the coatings. These coatings appeared to be hydrophobic and had low surface energies. X‐ray photoelectron spectroscopy (XPS) and photoacoustic Fourier‐transform infrared (PA‐FT‐IR) spectroscopy were used to investigate the detailed chemical structures of the deposited films. The optimum plasma processing conditions to achieve the desired thin plasma polymer coatings are discussed in the light of the chemistry that takes place at the interfaces. Copyright © 2004 Society of Chemical Industry     

Shortest path algorithms: A computational study with the C programming language

The main purpose of this study is to evaluate the computational efficiency of algorithms for calculating shortest paths when they are correctly coded by using the C programming language. The eight algorithms that we selected for this experiment are the most efficient, either measured in terms of worst-case bounds or marked as such from previous computational studies; they include the redistributive heap algorithm. We suggest computer implementations that use the full power of C. In particular, the network representation and the various data structures used to keep the scan eligible list may be managed by using only additions and no multiplications, while it is not possible with FORTRAN. These capabilities, unique to C, yield several interesting conclusions: one may expect to speed up a shortest path algorithm by a factor of 20%; in some cases, this factor may reach 30%. Interestingly, the level of programming difficulty required to achieve these benefits is not greater than that required by implementations using arrays.