Philip Allen - US grants

The conduction of heat and electricity in condensed matter will be studied theoretically in four diverse situations. Two of these, diffusive electrical transport in highly disordered alloys, and heat transport in insulating glasses, will be assisted by large scale computer evaluations of "exact" Kubo formulas designed both to test simpler approximations and to elucidate mechanisms of transport. The alloy problem uses the conventional independent electron model, and the Kubo-Greenwood formula. The glass problem will use the analogous harmonic approximation, and the analog conductivity formula which has not been previously worked out. Among the aims of these projects are "saturation" of resistivity, and the occurrence in metals of negative temperature coefficients of resistivity. Also a rigorous test will be made of the KKR-CPA theory of alloy resistivity. The two other regimes involve quasi-particle transport, in conventional band Fermi liquid metals and in the new oxide superconductors. For the former, band theory will be used to extract coupling constants from the anistropic resistivity of hexagonal metals, and also to analyze the Hall coefficients. For the latter, an effort will be made to deduce the nature of the current carrying quasi-particles and of the scattering mechanisms, by comparison with experimental measurements.

The purpose of the experiments contained in the present research proposal is to examine empirically the impact of normal aging and abnormal aging (the impact of Alzheimer's disease) on perceptual matching, spatial recognition memory, and visual search tasks. The participants will be healthy young and older adults, and older adults with Alzheimer's disease (AD). The goals of the current project are to: (a) test the viability of an internal noise model of cognitive aging, (b) distinguish differences in cognitive functioning related to normal aging from those associated with AD. Experiments 1-3 will examine the performances of healthy young and older adults on perceptual matching tasks (Experiment 1 will also test AD patients). The relative speed and accuracy with which subjects respond to "same" and "different" trials will be used to test the validity of the internal noise model. Experiments 4-6 will examine age differences on a spatial, recognition memory task for order information (Experiment 4 will also test AD patients). The internal noise model of cognitive aging predicts that transposition distance should impact upon reaction time (RT) and/or accuracy. Specifically, Correct rejection RTs and/or false alarm errors should increase as transposition distance decreases, and this effect should be larger for older adults. Experiments 7 and 9 will use more "ecologically valid" task to examine age differences in spatial recognition memory. A "clock face" task will be used where the minute and/or hour hand(s) of a clock face will be shifted relative to where they originally appeared on part of the trials (Experiment 7 will also test AD patients). Experiment 8 will test healthy older adults and AD patients on three paper-and-pencil tasks in order to examine the impact of linguistic deficits, perseveration, and context on AD patients' typically poor performance on production tasks for spatial memory. Using a visual search task, Experiments 10-12 will vary the number of instances of a target letter (i.e., redundancy) on a given trial. Also, half the trials will involve perceptual noise. The internal noise model predicts that older adults should exhibit relatively larger redundancy effects for noise trials than for no-noise trials on this visual search task where redundancy is varied from one to three instances. If the current tasks can separate healthy older adults from AD patients (i.e., Experiments 1, 4, 7-8, and 10), then these tasks could be used as assessment tools for indexing AD.

Theory will be developed for the electronic and vibrational properties of metals, insulators and superconductors, especially crystalline oxides such as copper-and bismuth-oxide superconductors and vanadium-oxide based metal-insulator systems. The underlying motive is to explore the still obscure mechanism for high temperature superconductivity. A secondary motive for the research is to improve understanding of the diverse and poorly understood electrical transport properties of these materials. As a model system barium-bismuth-oxide will be studied in detail. The charge carriers in this insulator are bipolarons. A continuous evolution through an exotic superconducting state to a normal metallic conductor occurs as bismuth is replaced by lead. Polaron and bipolaron models will be developed and tested to model resistivity, Hall coefficient, and superconductivity. Other model systems to be studied are niobium-oxide and vanadium-dioxide. The latter has a metal-insulator transition at a temperature convenient for device applications such as memory elements. The effect of lattice distortion and phonons on the transition will be studied. Thermal conduction will also be studied in a model of amorphous silicon.

biomedical equipment purchase;

9417755 Allen The ability of local density approximation band theory to predict properties of matter will be explored. Algorithms will be tested to make global searches for ground-state crystal structures. Magnetic excitations and electrical resistivity of metals will be calculated. Vibrational energy relaxation in carbon-sixty will be modeled. %%% Theory will be used to predict properties of crystalline matter starting from fundamental concepts. Specifically, methods for predicting the crystal structure, magnetism and electrical conductivity of materials will be developed and applied to various real materials of current interest. These methods, if successful will have wide applicability.

DESCRIPTION (adapted from investigator's abstract): The purpose of the proposed research is to test a theory of age differences in processing resources for semantic and episodic memory. Salthouse has hypothesized that processing speed is a processing resource. Allen has proposed that neural noise as measured by entropy provides an even more basic measure of age differences in processing resources. Specifically, increased neural noise reduces the available processing resources for cognitive processes. The goal of the present project is to determine why differential age differences occur for episodic and semantic memory tasks across peripheral and central processing stages by examining whether practice and/or task type affects these age differences. The present proposal contains eight experiments that are designed to determine how entropy (a measure of neural noise), processing speed, and processing accuracy are related to age differences in processing resources across different processing stages (peripheral vs. central) and tasks (episodic vs. semantic memory). Experiments 1-2 will examine the time-accuracy functions of healthy young and older adults on lexical decision and memory search tasks that vary exposure duration. Accuracy data will be used to derive entropy values. Experiments 3-4 will use Pashler's psychological refractory period paradigm to examine age differences in processing resources as the stimulus onset asynchrony is varied between two tasks (Experiment 3 will use dot location followed by a lexical decision task: Experiment 4 will use dot location followed by lexical decision or memory search tasks). Experiment 4 will also examine the effect of practice on age differences in processing resources by using four sessions per participant. Experiments 5-8 will consist of four within subject experiments in which the same stimuli (words or numbers) are used for both semantic and episodic memory tasks. Experiments 5-8 will rule out stimulus confounds that were present in earlier studies on this topic.

Allen 9725037 This is a renewal award to conduct theoretical research on metallic, or nearly metallic, oxides and on glasses. The normal modes of thermal vibration in glassy solids have been characterized in computer studies of up to 4096 atoms. These models yield realistic vibrations for amorphous silicon. Four new projects are planned: (1) Clarify the concept of diffusivity of normal modes; (2) Find connections between low energy "resonant"modes and low energy structural rearrangements which couple strongly to the resonant modes; (3) Develop a theory of the viscosity of the thermal vibrations and apply this to the study of internal friction in glasses; (4) Develop a new theory of hopping conduction of heat when localized states are concentrated. Although most important oxides are insulating, some are very good metals and the most interesting ones are on the borderline. The transition from metallic to insulating has usually two equally strong and very different driving forces, Coulomb correlation and lattice distortion. The former causes great theoretical difficulty even in the simplest fully interacting model, whereas the latter, solvable in principle, has multiple manifestations depending on the chemical species and crystal structure. A study of BaBiO3, where Coulomb effects are minimal, is currently being completed. To extract the common physics of many oxides, a simplified model due to Rice and Sneddon is used, where the only lattice degrees of freedom is the displacement of oxygen atoms parallel to the metal-oxygen bond. This not only captures the breathing distortion that drives the Bi(3+)/Bi(5+) disproportionation (and causes insulating behavior), but also captures the principle Jahn- Teller displacement which couples to the metal-insulator behavior of Eg metals and T2g metals and many isoelectronic materials. %%% This theoretical research will continue studies of metal oxide materials and their unusual properties. These materials are intr iguing due to their widely varied properties. Some conduct electricity, like normal metals, while others are insulating. Some are on the borderline between metallic properties and insulating properties. The research supported by this grant will investigate this difficult, but technologically important class of materials. ***

0089492
Allen
This grant supports theoretical research on models which mimic the behavior of real three-dimensional solids. Recent developments in mesoscopic physics have provided powerful new computer algorithms for studying electrical or heat transport in strongly disordered systems. The phenomenon of "resistivity saturation" will be studied, in order to test the hypothesis that "band-mixing" effects are involved and cause the excess conduction beyond the naive prediction of Boltzmann transport theory. A series of models will be studied numerically with increasingly complex band structures. An analog problem in lattice heat transport is the thermal conductivity of glasses. A realistic atomistic model for amorphous silicon will be treated using the tools of mesoscopic physics.

There are many important oxide materials with the perovskite (ABO3) or related crystal structures, including ferroelectrics, high temperature superconductors, ionic conductor electrolytes, etc. Although most perovskites are insulators, a few are metals, and quite a few have metal-to-insulator transitions as a function of doping or of temperature. The systems Ba1-xKxBiO3 and La1-xCaxMnO3 are particularly nice examples. The pure end-member (x=0) compounds are prototype examples of materials which are insulating because of a broken symmetry in the ground state, namely charge ordering in BaBiO3 , orbital ordering in LaMnO3 . Both systems become interesting metals when the doping level x increases to a critical value 0.2-0.4. Good models exist for the electons which participate in this metal-to-insulator transition. It is proposed to study self-trapped states in these materials. In particular, the opposite end-member (x=1) CaMnO3 , when lightly doped with electrons (x=1-epsilon) will be chosen as a model system for study of the competition between spin-polaron effects and lattice-polaron effects. The lowest electronic excitations of both BaBiO3 and LaMnO3 in this model are self-trapped excitons. These will be studied and their novel optical properties (adsorption spectra, resonant Raman spectra, etc.) will be predicted. When more heavily doped, holes should self-organize into planar anti-phase boundary charged defects (the three-dimensional version of "stripes"). These will be studied and the properties predicted. The metal-insulator transition as a function of doping (polaron-glass collapse) will be modeled and carefully characterized.
%%%
This grant supports theoretical research on the electrical and thermal properties of oxide materials which are of great fundamental interest due to their complex behavior as the temperature and/or concentration of additives is changed. These materials exhibit a wealth of effects including ferroelectricity, high temperature superconductivity and ionic conduction which may lead to widespread applications.
***

NIRT: Molecular Assembly for Hybrid Electronics
Abstract

This proposal was received in response to Nanoscale Science and Engineering initiative, NSF 02-148, category NIRT. The extension of integrated circuits into sub-10-nm range promises enormous benefits for computing, networking, and signal processing. However, fabrication of such devices using current paradigms based on CMOS and current VLSI technology are not possible. We believe that this crisis may only be resolved by a radical paradigm shift, which would simultaneously change the approach to fabrication of electron devices and to VLSI circuit architecture. Our approach is to use a biologically inspired approach called "Self-Evolving Neuromorphic Networks". This approach is based on artificial models of the neocortex and is structured to have a high degree of parallelism and intrinsic redundancy. In this approach molecular circuit elements, "self-assembled" by molecular chemistry, can be allowed to grow randomly, forming circuit elements (molecular transistors), which connect lithographically patterned metal grids. However, the random aspect of molecular self-assembly has to be carefully understood and controlled. At present, there is no detailed understanding of this process. It is this crucial gap that we address in this proposal. The devices that we are proposing need molecular wires that can switch into and out of a conductive state. The molecules bridge the metal wires with inherent randomness. Our aims are to predict and control the bridging and switching, through deterministic chemistry of the molecule-metal interaction, as well as through a statistical analysis of the assembly process. To accomplish these aims, we have a diverse team, which will interact strongly across the engineering/chemistry/physics boundaries. As part of the outreach of this project we plan to use the NIRT as a forum in which we will provide new types of educational settings for students (undergraduate and graduate) and high school teachers, and adopt a flexible program of research guided by feedback between theory and experiment, chemistry and physics and engineering.

Tremendous technological advances in miniaturization have enabled more and more transistors to be packed onto a silicon chip. However, the reduction of feature size on chips is limited not just by the resolution of the fabrication process, but also by the problem of quantum and classical fluctuations. Consequently, below a limiting dimension that we have nearly reached, a new paradigm that goes beyond conventional solid state electronics has to be developed for the next generation of electronics devices. In this project, we propose a new paradigm that is based on an artificial model of the neocortex: In which molecular circuits are assembled in a manner similar to the synaptic connections present in the brain. Our paradigm if realized offers the possibility of the design of the next generation of computational devices, with speeds that, in theory, could be 10 orders of magnitude faster than the fastest existing parallel supercomputer.

This project is collaboration between chemists, materials science researchers, computer scientists and geophysicists to advance the state-of-the-art in the numerical modeling of the properties of minerals, primarily at very high temperatures and pressures. It will create a virtual community laboratory. The intent is to refine existing first-principles computational mineral physics numerical simulation codes, to develop new first-principles codes, to develop novel visual human-computer interface tools, to make these widely available through a web portal and supporting grid infrastructure, and to use these numerical tools to investigate the structural, thermodynamic and thermoelastic properties of minerals in the deep Earth. An additional component looks at some of the possible properties of ice on Titan, a moon of Saturn. The project is motivated by a number of grand challenge problems in understanding Earth and planetary structures including: developing a better understanding of the variation in thermo-mechanical properties of silicate melts characteristic of magma; the behavior of hydrogen in near-anhydrous silicates; the properties of possible iron alloys in the outer core; and the properties of water-ice in environments such as those that may exist on Titan. Other topics to be addressed include the properties of solid solutions involving magnesium, iron and aluminum silicates. Part of the work involves developing improved parallelized methods for solving sparse linear algebra problems and three-dimensional fast Fourier transforms. To facilitate collaboration, the project includes research on the development of a grid-based architecture for the submission, execution and analysis of numerical chemistry calculations. The innovative aspect of this will be the development of a message-relaying framework to augment the communication and collaboration mechanisms currently available. This will involve the augmentation of the NaradaBrokering system (NB) to provide support for plug-in web services through a WSRF-compliant API. NB will also be extended to support the GSI security protocol. Common access to resources will be provided through web portals. Additional development will include the creation of collaborative (multi-user) visualization tools, supported by the NB infrastructure, specifically designed to be useful in browsing a snapshot of a complex simulation while that simulation is in progress. IT work also includes the development of task automation and more sophisticated grid schedulers. To test the collaborative environment, graduate students and post-docs at two non-partner institutions will be trained in the use of the community modeling system. To facilitate use of the virtual laboratory and its tools, tutorial workshops will be held. The project includes international partners from Italy and the UK, and international workshops on the numerical simulation issues are planned.

This engineering education research initiation grant will support an investigation into how new technologies for gesture and body recognition can help students better learn engineering concepts that are strongly tied to spatial and visual understanding. The project will engage researchers from engineering, computer science, and psychology in a well-structured experiment using validated instruments to test if the new technologies improve learning in biomedical engineering.

The broader significance and importance of this project will primarily be to determine if new, affordable, and commercially available technologies can impact learning of difficult engineering concepts. The strong partnership between researchers in different disciplines will build networks in cross-disciplinary research. Additionally an outreach effort to students traditionally under-represented in STEM at local high schools is planned, which can improve recruitment of these students to engineering programs. This project overlaps with NSF's strategic goals of transforming the frontiers by making investments that lead to emerging new fields of engineering, or shifting existing fields. Additionally NSF's goal of innovating for society is enabled by supporting the development of innovative learning systems.