On Bayes and unbiased estimators of loss

We consider estimation of loss for generalized Bayes or pseudo-Bayes estimators of a multivariate normal mean vector, θ. In 3 and higher dimensions, the MLEX is UMVUE and minimax but is inadmissible. It is dominated by the James-Stein estimator and by many others. Johnstone (1988, On inadmissibility of some unbiased estimates of loss,Statistical Decision Theory and Related Topics, IV (eds. S. S. Gupta and J. O. Berger), Vol. 1, 361–379, Springer, New York) considered the estimation of loss for the usual estimatorX and the James-Stein estimator. He found improvements over the Stein unbiased estimator of risk. In this paper, for a generalized Bayes point estimator of θ, we compare generalized Bayes estimators to unbiased estimators of loss. We find, somewhat surprisingly, that the unbiased estimator often dominates the corresponding generalized Bayes estimator of loss for priors which give minimax estimators in the original point estimation problem. In particular, we give a class of priors for which the generalized Bayes estimator of θ is admissible and minimax but for which the unbiased estimator of loss dominates the generalized Bayes estimator of loss. We also give a general inadmissibility result for a generalized Bayes estimator of loss. Research supported by NSF Grant DMS-97-04524.     

Exciton Molecules in Quantum Wells: Influence of the Well Width Fluctuations

The influence of the well width fluctuations on the dependence of the binding energy of excitonic complexes in quantum wells is studied by using the path-integral Monte-Carlo technique. The results are compared with available experimental data and a good agreement is found.Postdoctoral researcher of FWO-Vlaanderen     

Electric Instability in <Emphasis Type="Italic">n</Emphasis>-CdTe:In Layers with <Emphasis Type="Italic">S</Emphasis>-Shaped Voltage-Current Characteristics

A switching of the S-type in the 20–200 μm thick polycrystalline n-CdTe:In layers with resistance of 10³–10⁶ Ω·cm is studied. The electric instability in the layers is found to be due to the electron-thermal breakdown mechanism. The dependence of the switching threshold parameters on the intensity of exposure can be used for fabrication of infrared-radiation controlled electric switches on the basis of n-CdTe:In layers. __________ Translated from Izvestiya Vysshikh Uchebnykh Zavedenii, Fizika, No. 6, pp. 28–30, June, 2005.     

The “corset effect” of spin-lattice relaxation in polymer melts confined in nanoporous media

Linear polyethylene oxides with molecular weightsM _w of 1665 and 10170 confined in pores with variable diameters in a solid methacrylate matrix were studied by proton field-cycling nuclear magnetic resonance relaxometry. The pore diameter was varied in the range of 9–57 nm. In all cases, the spin-lattice relaxation time shows a frequency dependence close toT ₁∞ v^3/4 in the range ofv=3·10^?1-2·10¹ MHz as predicted by the tube-reptation model. This protonT ₁ dispersion essentially reproduces that found in a previous deuteron study (R. Kimmich, R.-O. Seitter, U. Beginn, M. Möller, N. Fatkullin: Chem. Phys. Lett. 307, 147, 1999). As a feature particularly characteristic for reptation, this finding suggests that reptation is the dominating chain dynamics mechanism under pore confinement in the corresponding time range. The absolute values of the spin-lattice relaxation times indicate that the diameter of the effective tubes in which reptation occurs is much smaller than the pore diameters on the time scale of spin-lattice relaxation experimens. An estimation leads to a valued ^*～0.5 nm. The impenetrability of the solid pore walls, the uncrossability of polymer chains (“excluded volume”) and the low value of the compressibility in polymer melts create the “corset effect” which reduces the lateral motions of polymer chains to a microscopic scale of only a few tenths of a nanometer.     

Anionic iron(II) alkoxides as initiators for the controlled ring‐opening polymerization of lactide

Hetero‐bimetallic Fe(II) alkoxide/aryloxides were evaluated as initiators for the ring‐opening polymerization of rac‐lactide. [(THF)NaFe(O^tBu)₃]₂ ( 1 ) and [(THF)₄Na₂Fe(2,6‐diisopropylphenolate)₄] ( 2 ) (THF = tetrahydrofuran) both polymerized lactide efficiently at room temperature, with complex 1 affording better control over the molecular weight parameters of the resultant polymer. At conversions below 70%, a linear increase in molecular weight with conversion was observed, indicative of a well‐controlled polymerization process. Complex 2 is the first example of a dianionic Fe(II) alkoxide and has been structurally characterized to reveal a distorted square planar FeO₄ array in which both Na counterions bridge two aryloxide ligands and are further complexed by two THF ligands. © 2003 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 41: 3798–3803, 2003     

Effect of nickel coating on electrochemical performance of graphite anodes for lithium ion batteries

Among the different materials often studied and proposed as negative electrodes for lithium-ion batteries, graphite anodes are the most used in commercial batteries. For this study, synthetic graphite was tested. During the first discharge 0.2 Li ions were consumed for the formation of the SEI film and the capacity reaches about 387 mAh/g. But at the end of the first charge only 72% of the initial charge was recovered (the reversible capacity is about 279 mAh/g). In order to improve this performance we have deposited metallic nickel on graphite with the intention to obtain a homogeneous thin layer able to modify the nature of the SEI film, to allow the diffusion of lithium ions through the protective layer, and also to increase the performance of graphite electrodes. The results show a decrease of the irreversible capacity loss (16% instead of 28% for pure graphite electrodes) as well as better cycleability for a nickel-deposited graphite electrode with only 11% weight ratio of nickel. On the other hand, an increase of the nickel content decreases this performance.     

effect of surface modification on the properties of hydrogen-sensitive GaAs-based Schottky diodes

The surface of hydrogen-sensitive GaAs Schottky diodes is modified by nonpolishing etching and by producing quantum wells and quantum dots in the space-charge region of the semiconductor. The sensitivity to hydrogen is found to increase by a factor of 8–37 after the etching and by two or three orders of magnitude after the introduction of quantum wells and dots. It is shown that the increased sensitivity is associated with the lowering of the barrier at the Pd/GaAs interface, the retardation of hydrogen diffusion into GaAs due to the presence of strained quantum-size layers, and an increase in the recombination current. The presence of the recombination component is supported by luminescence from the quantum wells and quantum dots, as well as from the GaAs substrate. The etch composition is shown to be a decisive factor in raising the sensitivity.