Publications

Different hardness-assurance tests are often required for advanced bipolar devices than for CMOS devices. In this work, the dose-rate dependence of bipolar current-gain degradation is mapped over a wide range of dose rates for the first time, and it is very different from analogous MOSFET curves. Annealing experiments following irradiation show negligible change in base current at room temperature, but significant recovery at temperatures of 100°C and above. In contrast to what is observed in MOSFET’s, irradiation and annealing tests cannot be used to predict the low-dose-rate response of bipolar devices. A comparison of x-ray-induced and 60Co gamma-ray-induced gain degradation is reported for the first time for bipolar transistors. The role of the emitter bias during irradiation is also examined. Implications for hardening and hardness assurance are discussed. © 1993 IEEE

Vadose-zone moisture transport near an impermeable barrier has been under study at a field site near Albuquerque, NM since 1990. Moisture content and temperature have been monitored in the subsurface on a regular basis; both undergo a seasonal variation about average values. Even though the slab introduces two-dimensional effects on the scale of the slab, moisture and heat transport is predominantly vertical. Numerical simulations, based on the models developed by Philip and de Vries (1957) and de Vries (1958), indicate that the heat flow is conduction-dominated while the moisture movement is dominated by diffusive vapor distillation. Model predictions of the magnitude and extent of changes in moisture content underneath the slab are in reasonable agreement with observation.

A method is proposed for supps'essing the resonances that occur as an item of rotating machinery is spun-up from rest to its operating speed. This proposed method invokes “stiffiness scheduling” so that the resonant frequency of the system is shifted during spin-up so as to be distant from the excitation frequency. A strategy for modulating the stiffness through the use of shape memory alloy is also presented.

A Sandia National Laboratories/AT&T Bell Laboratories Team is developing a soft x-ray projection lithography tool that uses a compact laser plasma as a source of 14 nm x-rays. Optimization of the 14 nm x-rays source brightness is a key issue in this research. This paper describes our understanding of the source as it has been obtained through the use of computer simulations utilizing the LASNEX radiation-hydrodynamics code.

Lightning protection systems (LPSs) for explosives handling and storage facilities have long been designed similarity to those need for more conventional facilities, but their overall effectiveness in controlling interior electromagnetic (EM) environments has still not been rigorously assessed. Frequent lightning-caused failures of a security system installed in earth-covered explosives storage structures prompted the U.S. Army and Sandia National Laboratories to conduct a program to determine quantitatively the EM environments inside an explosives storage structure that is struck by lightning. These environments were measured directly during rocket-triggered lightning (RTL) tests in the summer of 1991 and were computed using linear finite-difference, time-domain (FDTD) EM solvers. The experimental and computational results were first compared in order to validate the code and were then used to construct bounds for interior environments corresponding to severe incident lightning flashes. The code results were also used to develop simple circuit models for the EM field behavior.

Ion-beam-induced charge-collection imaging (IBICC) has been used to study the SEU mechanisms of the Sandia TA670 16K-bit SRAM. Quantitative charge-collection spectra from known regions of the memory cell have been derived with this technique. For 2.4-MeV He ions at normal incidence, charge collection depth for a reverse-biased p+ drain strike is estimated to be 4.8±0.4 μm. Heavy-ion strikes to the reverse-biased p-well result in nearly complete collection of deposited charge to a depth of 5.5±0.5 μm. A charge amplification effect in the n-on drain is identified and is due to either bipolar amplification or a shunt effect in the parasitic vertical npn bipolar transistor associated with the n+/n substrate, p-well, and n+ drain. This effect is present only when the n+ drain is at 0V bias. When coupled with previous SEU-imaging, these results strongly suggest that the dominant SEU mechanism in this SRAM is a heavy-ion strike to the n-on transistor drain. © 1993 IEEE

The problem of constructing trees given a matrix of interleaf distances is motivated by applications in computational evolutionary biology and linguistics. The general problem is to find an edge-weighted tree which most closely approximates (under some norm) the distance matrix. Although the construction problem is easy when the tree exactly fits the distance matrix, optimization problems under all popular criteria are either known or conjectured to be NP-complete. In this paper we consider the related problem where we are given a partial order on the pairwise distances, and wish to construct (if possible) an edge-weighted tree realizing the partial order. In particular we are interested in partial orders which arise from experiments on triples of species. We will show that the consistency problem is NP-hard in general, but that for certain special cases the construction problem can be solved in polynomial time.

The dynamics of flexible bodies spinning at rates near or above their first natural frequencies is a notoriously difficult area of analysis. Recently, a method of analysis, tentatively referred to as a method of quadratic modes, has been developed to address this sort of problem. This method restricts consideration to configurations in which all kinematic constraints are automatically satisfied through second order in deformation. Besides providing robustness, this analysis method reduces the problem from one that would otherwise require the reformulation of stiffness matrices at each time step to one of solving only a small number of nonlinear equations at each time step. A test of this method has been performed, examining the vibrations of a rotating, inflated membrane.

Parallel computing offers new capabilities for using molecular dynamics (MD) to simulate larger numbers of atoms and longer time scales. In this paper we discuss two methods we have used to implement the embedded atom method (EAM) formalism for molecular dynamics on multiple-instruction/multiple-data (MIMD) parallel computers. The first method (atom-decomposition) is simple and suitable for small numbers of atoms. The second method (force-decomposition) is new and is particularly appropriate for the EAM because all the computations are between pairs of atoms. Both methods have the advantage of not requiring any geometric information about the physical domain being simulated. We present timing results for the two parallel methods on a benchmark EAM problem and briefly indicate how the methods can be used in other kinds of materials MD simulations.

The Vapnik-Chervonenkis (V-C) dimension is an important combinatorial tool in the analysis of learning problems in the PAC framework. For polynomial learnability, we seek upper bounds on the V-C dimension that are polynomial in the syntactic complexity of concepts. Such upper bounds are automatic for discrete concept classes, but hitherto little has been known about what general conditions guarantee polynomial bounds on V-C dimension for classes in which concepts and examples are represented by tuples of real numbers. In this paper, we show that for two general kinds of concept class the V-C dimension is polynomially bounded as a function of the syntactic complexity of concepts. One is classes where the criterion for membership of an instance in a concept can be expressed as a formula (in the first-order theory of the reals) with fixed quantification depth and exponentially-bounded length, whose atomic predicates are polynomial inequalities of exponentially-bounded degree. The other is classes where containment of an instance in a concept is testable in polynomial time, assuming we may compute standard arithmetic operations on reals exactly in constant time. Our results show that in the continuous case, as in the discrete, the real barrier to efficient learning in the Occam sense is complexity-theoretic and not information-theoretic. We present examples to show how these results apply to concept classes defined by geometrical figures and neural nets, and derive polynomial bounds on the V-C dimension for these classes.

AC impedance and x-ray diffraction measurements versus temperature and pressure on several ternary fullerene intercalation compounds are reported. The results support our previously established empirical correlation between superconducting onset temperature and 300K fcc lattice constant. Compounds which do not follow this correlation either phase separate or are otherwise unstable at low T and/or high P. © 1993.

During hydrocarbon reservoir stimulations, such as hydraulic fracturing, the cracking and slippage of the formation results in the emission of seismic energy. The objective of this study was to determine the properties of these induced micro-seisms. A hydraulic fracture experiment was performed in the Piceance Basin of Western Colorado to induce and record micro-seismic events. The formation was subjected to four processes; breakdown/ballout, step-rate test, KCL mini-fracture, and linear-gel mini-fracture. Micro-seisms were acquired with an advanced three-component wall-locked seismic accelerometer package, placed in an observation well 211 ft offset from the fracture well. During the two hours of formation treatment, more than 1200 micro-seisms with signal-to-noise ratios in excess of 20 dB were observed. The observed micro- seisms had a nominally flat frequency spectrum from 100 Hz to 1500 Hz and lack the spurious tool-resonance effects evident in previous attempts to measure micro-seisms. Both p-wave and s-wave arrivals are clearly evident in the data set, and hodogram analysis yielded coherent estimates of the event locations. This paper describes the characteristics of the observed micro- seismic events (event occurrence, signal-to-noise ratios, and bandwidth) and illustrates that the new acquisition approach results in enhanced detectability and event location resolution.

An essential requirement for both Vertical Seismic Profiling (VSP) and Cross-Hole Seismic Profiling (CHSP) is the rapid acquisition of high resolution borehole seismic data. Additionally, full wave-field recording using three-component receivers enables the use of both transmitted and reflected elastic wave events in the resulting seismic images of the subsurface. To this end, an advanced three-component multi-station borehole seismic receiver system has been designed and developed by Sandia National Labs (SNL) and OYO Geospace. The system acquires data from multiple three-component wall-locking accelerometer packages and telemeters digital data to the surface in real-time. Due to the multiplicity of measurement stations and the real-time data link, acquisition time for the borehole seismic survey is significantly reduced. The system was tested at the Chevron La Habra Test Site using Chevron's clamped axial borehole vibrator as the seismic source. Several source and receiver fans were acquired using a four-station version of the advanced receiver system. For comparison purposes, an equivalent data set was acquired using a standard analog wall-locking geophone receiver. The test data indicate several enhancements provided by the multi-station receiver relative to the standard receiver; drastically improved signal-to-noise ratio, increased signal bandwidth, the detection of multiple reflectors, and a true 4: 1 reduction in survey time.

The earth`s ionosphere consists of an ionized plasma which will interact with any electromagnetic wave propagating through it. The interaction is particularly strong at vhf and uhf frequencies but decreases for higher microwave frequencies. These interaction effects and their relationship to the operation of a wide-bandwidth, synthetic-aperture, space-based radar are examined. Emphasis is placed on the dispersion effects and the polarimetric effects. Results show that high-resolution (wide-bandwidth) and high-quality coherent polarimetrics will be very difficult to achieve below 1 GHz.

SAFSIM (System Analysis Flow SIMulator) is a FORTRAN computer program that provides engineering simulations of user-specified flow networks at the system level. It includes fluid mechanics, heat transfer, and reactor dynamics capabilities. SAFSIM provides sufficient versatility to allow the simulation of almost any flow system, from a backyard sprinkler system to a clustered nuclear reactor propulsion system. In addition to versatility, speed and robustness are primary goals of SAFSIM development. The current capabilities of SAFSIM are summarized and some sample applications are presented. It is applied here to a nuclear thermal propulsion system and nuclear rocket engine test facility.

Many designs use EPLDs (Erasable Programmable Logic Devices) to implement control logic and state machines. If the design is slow, timing through the EPLD is not crucial so designers often treat the device as a black box. In high speed designs, timing through the EPLD is critical. In these cases a thorough understanding of the device architecture is necessary. Lessons learned in the implementation of a high-speed design using the Altera EPM5130 are discussed.

A new Assembly Test Chip, ATC04, designed to measure mechanical stresses at the die surface has been built and tested. This CMOS chip 0.25 in. on a side, has an array of 25 piezoresistive stress sensing cells, four resistive heaters and two ring oscillators. The ATCO4 chip facilitates making stress measurements with relatively simple test equipment and data analysis. The design, use, and accuracy of the chip are discussed and initial results are presented from three types of stress measurement experiments: four-point bending calibration, single point bending of a substrate with an ATC04 attached by epoxy, and stress produced by a liquid epoxy encapsulant.

The feasibility of utilizing a groundbased laser without an orbital mirror for space debris removal is examined. Technical issues include atmospheric transmission losses, adaptive-optics corrections of wavefront distortions, laser field-of-view limitations, and laser-induced impulse generation. The physical constraints require a laser with megawatt output, long run-time capability, and wavelength with good atmospheric transmission characteristics. It is found that a 5-MW reactor-pumped laser can deorbit debris having masses of the order of one kilogram from orbital altitudes to be used by Space Station Freedom. Debris under one kilogram can be deorbited after one pass over the laser site, while larger debris can be deorbited or transferred to alternate orbits after multiple passes over the site.

Flux-flow based devices such as the superconducting flux flow transistor and magnetically controlled long junctions have been made from thin films of TICaBaCuO and YBaCuO. The devices are based on the magnetic control of flux flow in their respective structures: a long junction or an array of weak links. The equivalent circuits of the two devices are similar: a low impedance input control line, an output impedance of 3-20 Q and an active current-controlled element. The long junctions have tended to be slower, have lower gain and be somewhat less noisy than their counterparts. The performance of circuits such as narrowband and distributed amplifiers (SO GHz bandwidths, noise figures < 3 dB), phase shifters (continuous with < 2 dB loss 4-40 GHz), logic gates (2-3 ps gate delays) and memories made using these devices will be compared and analyzed. © 1993 IEEE

Automatic motion planning of a spray cleaning robot with collision avoidance is presented in this paper. In manufacturing environments, electronic and mechanical components are traditionally cleaned by spraying or dipping them using chlorofluorocarbon (CFC) solvents. As new scientific data show that such solvents are major causes for stratospheric ozone depletion, an alternate cleaning method is needed. Part cleaning with aqueous solvents is environmentally safe, but can require precision spraying at high pressures for extended time periods. Operator fatigue during manual spraying can decrease the quality of the cleaning process. By spraying with a robotic manipulator, the necessary spray accuracy and consistency to manufacture high-reliability components can be obtained. Our motion planner was developed to automatically generate motions for spraying robots based on the part geometry and cleaning process parameters. For spraying paint and other coatings a geometric description of the parts and robot may be sufficient for motion planning, since coatings are usually done over the visible surfaces. For spray cleaning, the requirement to reach hidden surfaces necessitates the addition of a rule-based method to the geometric motion planning.

The geochemical properties of a porous sand and several tracers (Ni, Br, and Li) have been characterized for use in a caisson experiment designed to validate sorption models used in models of reactive transport. The surfaces of the sand grains have been examined by a combination of techniques including potentiometric titration, acid leaching, optical microscopy, and scanning electron microscopy with energy-dispersive spectroscopy. The surface studies indicate the presence of small amounts of carbonate, kaolinite and iron-oxyhydroxides. Adsorption of nickel, lithium and bromide by the sand was measured using batch techniques. Bromide was not sorbed by the sand. A linear (Kd) or an isotherm sorption model may adequately describe transport of Li; however, a model describing the changes of pH and the concentrations of other solution species as a function of time and position within the caisson and the concomitant effects on Ni sorption may be required for accurate predictions of nickel transport.

For problems where media properties are measured at one scale and applied at another, scaling laws or models must be used in order to define effective properties at the scale of interest. The accuracy of such models will play a critical role in predicting flow and transport through the Yucca Mountain Test Site given the sensitivity of these calculations to the input property fields. Therefore, a research program has been established to gain a fundamental understanding of how properties scale with the aim of developing and testing models that describe scaling behavior in a quantitative manner. Scaling of constitutive rock properties is investigated through physical experimentation involving the collection of suites of gas permeability data measured over a range of discrete scales. Also, various physical characteristics of property heterogeneity and the means by which the heterogeneity is measured and described and systematically investigated to evaluate their influence on scaling behavior. This paper summarizes the approach that is being taken toward this goal and presents the results of a scoping study that was conducted to evaluate the feasibility of the proposed research.

Experimental results exploring gravity-driven wetting front instability in a pre-wetted, rough-walled analog fracture are presented. Initial conditions considered include a uniform moisture field wetted to field capacity of the analog fracture and the structured moisture field created by unstable infiltration into an initially dry fracture. As in previous studies performed under dry initial conditions, instability was found to result both at the cessation of stable infiltration and at flux lower than the fracture capacity under gravitational driving force. Individual fingers were faster, narrower, longer, and more numerous than observed under dry initial conditions. Wetting fronts were found to follow existing wetted structure, providing a mechanism for rapid recharge and transport.

In an attempt to achieve completeness and consistency, the performance-assessment analyses developed by the Yucca Mountain Project are tied to scenarios described in event trees. Development of scenarios requires describing the constituent features, events, and processes in detail. Several features and processes occurring at the waste packages and the rock immediately surrounding the packages (i.e., the near field) have been identified: the effects of radiation on fluids in the near-field rock, the path-dependency of rock-water interactions, and the partitioning of contaminant transport between colloids and solutes. This paper discusses some questions regarding these processes that the near-field performance-assessment modelers will need to have answered to specify those portions of scenarios dealing with the near field.

Experiments investigating the behavior of individual, gravity-driven fingers in an initially dry, rough-walled analog fracture are presented. Fingers were initiated from constant flow to a point source. Finger structure is described in detail; specific phenomena observed include: desaturation behind the finger-tip, variation in finger path, intermittent flow structures, finger-tip bifurcation, and formation of dendritic sub-fingers. Measurements were made of finger-tip velocity, finger width, and finger-tip length. Non-dimensional forms of the measured variables are analyzed relative to the independent parameters, flow rate and gravitational gradient.