Short-Time Oxidation Behavior of Low-Carbon, Low-Silicon Steel in Air at 850–1180 °C: II. Linear to Parabolic Transition Determined Using Existing Gas-Phase Transport and Solid-Phase Diffusion Theories

Existing gas-phase transport and solid-phase diffusion theories are used to calculate the oxidation kinetics of low carbon and low silicon steel in flowing air at 850–1180 °C. The linear-to-parabolic transition scale thickness derived is proportional to the parabolic rate constant and inversely proportional to the linear rate constant. Calculated parabolic rate constants are consistent with experimental results, whereas calculated linear rate constants are 40% smaller and calculated linear-to-parabolic transition scale thicknesses are 20–40% greater than the experimental results. Introduction of a correction factor cannot simultaneously resolve the discrepancies. Various sources causing the discrepancies are discussed. One important issue identified is that the conditions for deriving the equation for calculating the linear rate constant are different from those in laboratory furnaces. The linear-to-parabolic transition scale thickness derived is also a function of gas composition, gas velocity and sample length. An intrinsic gas-phase mass-transfer factor, independent of gas velocity and sample size, is recommended to express the gas-phase mass-transfer coefficient for calculating the linear rate constant.     

On the temperature sensitivity of 1.5-/spl mu/m GaInNAsSb lasers

We analyze the temperature sensitivity of 1.5-/spl mu/m GaInNAsSb lasers grown on GaAs. Building on the method of Tansu and coworkers, we find evidence that the characteristic temperatures for the threshold current T/sub 0/ and external efficiency T/sub 1/ are balanced by a combination of monomolecular recombination and temperature destabilizing mechanism(s) near room temperature. At elevated temperatures, the destabilizing process(es) dominate, due to increased threshold current density J/sub th/. While it is difficult to definitively identify carrier leakage, Auger recombination, or a combination of the two as the responsible mechanism(s), results indicate that carrier leakage certainly plays a role. Evidence of intervalence band absorption was also found; T/sub 1/ was reduced, but J/sub th/ and T/sub 0/ were not significantly degraded. Conclusions are corroborated by supporting measurements of the Z-parameter with bias, spontaneous emission spectrum, and band-offsets. Spontaneous emission measurements show evidence of weak Fermi-level pinning within the active region at threshold, indicating a form of carrier leakage. This is consistent with the characteristic temperature analysis and a leakage mechanism is proposed. This process is partially responsible for the greater temperature sensitivity of device parameters and the poor internal efficiency. Methods for reducing the effects of each parasitic mechanism are also described.     

CPE: a parallel library for financial engineering applications

The Clustertech parallel environment is an object-oriented C++ library that uses abstractions to simplify parallel programming for financial engineering applications. The message passing interface ensures CPE's portability and performance over a wide range of parallel cluster and symmetric multiprocessing machines.     

3D electroplated microstructures fabricated by a novel height control method

 A 3D electroplating process by means of pixel-wise, step height control of selective electroplating has been successfully demonstrated. A total of eight rectangular pixels of 80 × 20 μm² in area and 1.1–9.3 μm in height have been fabricated with a height difference of 1 μm in adjacent pixels. The process requires only two masks in a single electroplating process. Nickel electroplating is selected as the demonstrating material and is performed at 50 °C with current density of 400 A/m² for 30 min. A theoretical model in electroplating process is discussed to assist the design of 3D microstructures and verified with the experimental data. As such, this process has potential applications in making pixel-wise, 3D microstructures with precise height control. Received: 10 August 2001/Accepted: 24 September 2001     

Face representation using independent component analysis

This paper addresses the problem of face recognition using independent component analysis (ICA). More specifically, we are going to address two issues on face representation using ICA. First, as the independent components (ICs) are independent but not orthogonal, images outside a training set cannot be projected into these basis functions directly. In this paper, we propose a least-squares solution method using Householder Transformation to find a new representation. Second, we demonstrate that not all ICs are useful for recognition. Along this direction, we design and develop an IC selection algorithm to find a subset of ICs for recognition. Three public available databases, namely, MIT AI Laboratory, Yale University and Olivette Research Laboratory, are selected to evaluate the performance and the results are encouraging.     

Determinant spreading associated with demyelination in a nonhuman primate model of multiple sclerosis

Definition of the immune process that causes demyelination in multiple sclerosis is essential to determine the feasibility of Ag-directed immunotherapy. Using the nonhuman primate, Callithrix jacchus jacchus (common marmoset), we show that immunization with myelin basic protein and proteolipid protein determinants results in clinical disease with significant demyelination. Demyelination was associated with spreading to myelin oligodendrocyte glycoprotein (MOG) determinants that generated anti-MOG serum Abs and Ig deposition in central nervous system white matter lesions. These data associate intermolecular "determinant spreading" with clinical autoimmune disease in primates and raise important issues for the pathogenesis and treatment of multiple sclerosis.     

Germ-line mutations in the neurofibromatosis 2 gene: correlations with disease severity and retinal abnormalities

Neurofibromatosis 2 (NF2) features bilateral vestibular schwannomas, other benign neural tumors, and cataracts. Patients in some families develop many tumors at an early age and have rapid clinical progression, whereas in other families, patients may not have symptoms until much later and vestibular schwannomas may be the only tumors. The NF2 gene has been cloned from chromosome 22q; most identified germ-line mutations result in a truncated protein and severe NF2. To look for additional mutations and clinical correlations, we used SSCP analysis to screen DNA from 32 unrelated patients. We identified 20 different mutations in 21 patients (66%): 10 nonsense mutations, 2 frameshifts, 7 splice-site mutations, and 1 large in-frame deletion. Clinical information on 47 patients from the 21 families included ages at onset and at diagnosis, numbers of meningiomas, spinal and skin tumors, and presence of cataracts and retinal abnormalities. We compared clinical findings in patients with nonsense or frameshift mutations to those with splice-site mutations. When each patient was considered as an independent random event, the two groups differed (P < or = .05) for nearly every variable. Patients with nonsense or frameshift mutations were younger at onset and at diagnosis and had a higher frequency and mean number of tumors, supporting the correlation between nonsense and frameshift mutations and severe NF2. When each family was considered as an independent random event, statistically significant differences between the two groups were observed only for mean ages at onset and at diagnosis. A larger data set is needed to resolve these discrepancies. We observed retinal hamartomas and/or epiretinal membranes in nine patients from five families with four different nonsense mutations. This finding, which may represent a new genotype-phenotype correlation, merits further study.     

Chronic lymphatic leukemia. Peroral cladribine as primary treatment

Ten patients with newly diagnosed B-chronic lymphocytic leukaemia were treated with cladribin orally for five days every four weeks with a median of four series. This is the first reported clinical study where a purine analogue is administered orally. The tumour reducing effect was fast. Eight out of 10 patients responded with a partial or clinical complete remission. Two of these were in molecular biological complete remission. With an observation time of 22 months we have seen no serious side effects so far. A randomized study (including a long term follow up) between chlorambucil, fludarabin and cladribin is needed to clarify the future role of cladribin in B-CLL treatment.