Principles of wet chemical processing in ULSI microfabrication

Fine patterning technology for integrated device manufacturing requires properties such as surface cleanliness, surface smoothness, complete uniformity, and complete etching linearity in wet chemical processing. An improved chemical composition for buffered hydrogen fluoride (BHF:NH₄F+HF+H₂O) is determined based on fundamental research into the chemical reaction mechanism of BHF and SiO ₂. Advanced wet chemical processing based on investigation of chemical reaction mechanisms and properties of liquid chemicals, concentrating on the SiO₂ patterning process by BHF, is described. The principles of wet chemical processing in silicon technology are based on the following: the determination of the dominant reaction (etching) species, the influence of the solubility of the etching products in BHF on etching uniformity and linearity, stability of chemical composition without solid-phase segregation, and an improvement of the wettability of liquid chemicals on the wafer surface by the addition of a surfactant     

Surface active buffered hydrogen fluoride having excellentwettability for ULSI processing

By incorporating selected hydrocarbon surfactants, a surface-active BHF (buffered hydrogen fluoride) has been tailored to achieve the following requirements: (1) the same etch rate as that of conventional BHF; (2) low contact angle; (3) nonsegregation; (4) nonfoaming; (5) low particulate count; (6) few impurities (possibility of purification); (7) low particulate adhesion on the wafer surface; (8) no surface residues; (9) excellent surface smoothness; and (10) high SiO ₂/Si etching selectivity. In order to satisfy these requirements, surfactants must satisfy the following characteristics: (1) good solubility in BHF; (2) hydrophilic property at the wafer surface, (3) nondecomposition in BHF; (4) nonreaction with BHF; and (5) sufficient lowering of contact angle at the critical micelle concentration (CMC). Aliphatic amines satisfy these requirements but have foaming problems. The requirements have been achieved using a binary surfactant system consisting of a combination of aliphatic amine and aliphatic alcohol or aliphatic acid     

Optimizing Spatial filters for Robust EEG Single-Trial Analysis

Due to the volume conduction multichannel electroencephalogram (EEG) recordings give a rather blurred image of brain activity. Therefore spatial filters are extremely useful in single-trial analysis in order to improve the signal-to-noise ratio. There are powerful methods from machine learning and signal processing that permit the optimization of spatio-temporal filters for each subject in a data dependent fashion beyond the fixed filters based on the sensor geometry, e.g., Laplacians. Here we elucidate the theoretical background of the common spatial pattern (CSP) algorithm, a popular method in brain-computer interface (BCD research. Apart from reviewing several variants of the basic algorithm, we reveal tricks of the trade for achieving a powerful CSP performance, briefly elaborate on theoretical aspects of CSP, and demonstrate the application of CSP-type preprocessing in our studies of the Berlin BCI (BBCI) project.     

Bonding ability evaluation of bone cement on the cortical surface of rabbit’s tibia

A composite bone cement designated G2B1 that contains β tricalcium phosphate particles was developed as a bone substitute for percutaneous transpedicular vertebroplasty. In this study, both G2B1 and commercial PMMA bone cement (CMW1) were implanted into proximal tibiae of rabbits, and their bone-bonding strengths were evaluated at 4, 8, 12 and 16 weeks after implantation. Some of the specimens were evaluated histologically using Giemsa surface staining, contact microradiography (CMR) and scanning electron microscopy (SEM). Histological findings showed that G2B1 contacted bone directly without intervening soft tissue in the specimens at each time point, while there was always a soft tissue layer between CMW1 and bone. The bone-bonding strength of G2B1 was significantly higher than that of CMW1 at each time point, and significantly increased from 4 weeks to 8 and 12 weeks, while it decreased significantly from 12 weeks to 16 weeks. Bone remodeling of the cortex under the cement was observed especially for G2B1 and presumably influenced the bone bonding strength of the cement. The results indicate that G2B1 has bioactivity, and bone bonding strength of bioactive bone cements can be estimated fairly with this experimental model in the short term.     

Spontaneous contractures of both abductor digiti minimi muscles in a patient with systemic lupus erythematosus

A case is reported of spontaneous contracture of both little fingers in a patient with systemic lupus erythematosus. Operative release of abductor digiti minimi produced a good result. It seems likely that the cause of the contracture of ADM may have been myositis related to SLE.     

A new discriminative kernel from probabilistic models

Recently, Jaakkola and Haussler (1999) proposed a method for constructing kernel functions from probabilistic models. Their so-called Fisher kernel has been combined with discriminative classifiers such as support vector machines and applied successfully in, for example, DNA and protein analysis. Whereas the Fisher kernel is calculated from the marginal log-likelihood, we propose the TOP kernel derived; from tangent vectors of posterior log-odds. Furthermore, we develop a theoretical framework on feature extractors from probabilistic models and use it for analyzing the TOP kernel. In experiments, our new discriminative TOP kernel compares favorably to the Fisher kernel.     

Treatment of osteomyelitis with antibiotic-soaked porous glass ceramic

We have developed a new drug delivery system using porous apatite-wollastonite glass ceramic (A-W GC) to treat osteomyelitis. A-W GC (porosity, 70% and 20% to 30%), or porous hydroxyapatite (HA) blocks (porosity 35% to 48%) used as controls, were soaked in mixtures of two antibiotics, isepamicin sulphate (ISP) and cefmetazole (CMZ) under high vacuum. We evaluated the release concentrations of the antibiotics from the blocks. The bactericidal concentration of ISP from A-W GC was maintained for more than 42 days, but that from HA decreased to below the detection limit after 28 days. The concentrations of CMZ from both materials were lower than those of ISP. An in vivo study using rabbit femora showed that an osseous concentration of ISP was maintained at eight weeks after implantation. Osteoconduction of the A-W GC block was good. Four patients with infected hip arthroplasties and one with osteomyelitis of the tibia have been treated with the new delivery system with excellent results.     

On Taxonomies for Multi-class Image Categorization

We study the problem of classifying images into a given, pre-determined taxonomy. This task can be elegantly translated into the structured learning framework. However, despite its power, structured learning has known limits in scalability due to its high memory requirements and slow training process. We propose an efficient approximation of the structured learning approach by an ensemble of local support vector machines (SVMs) that can be trained efficiently with standard techniques. A?first theoretical discussion and experiments on toy-data allow to shed light onto why taxonomy-based classification can outperform taxonomy-free approaches and why an appropriately combined ensemble of local SVMs might be of high practical use. Further empirical results on subsets of Caltech256 and VOC2006 data indeed show that our local SVM formulation can effectively exploit the taxonomy structure and thus outperforms standard multi-class classification algorithms while it achieves on par results with taxonomy-based structured algorithms at a significantly decreased computing time.