Publications

Where there is a significant actuarial basis for decision making (e.g., the occurrence of fires in single-family dwellings), there is little incentive for formal risk management. Formal risk assessments are most useful in those cases where the value of the structure is high, many people may be affected, the societal perception of risk is high, consequences of a mishap would be severe, and the actuarial uncertainty is large. For these cases, there is little opportunity to obtain the necessary experiential data to make informed decisions, and the consequences in terms of money, lives, and societal confidence are severe enough to warrant a formal risk assessment. Other important factors include the symbolic value of the structure and vulnerability to single point failures. It is unlikely that formal risk management and assessment practices will or should replace the proven institutions of building codes and engineering practices. Nevertheless, formal risk assessment can provide valuable insights into the hazards threatening high-value and high-risk (perceived or actual) buildings and structures, which can in turn be translated into improved public health, safety, and security. The key is to choose and apply the right assessment tool to match the structure in question. Design-for-reliability concepts can be applied to buildings, bridges, transportation sys- tems, dams, and other structures. The use of these concepts could have the dual benefits of lowering life-cycle costs by reducing the necessity for maintenance and repair and of enhancing the saiiety and security of the structure's users.

How to ensure the appropriate performance of our built environment in the face of normal conditions, natural hazards, and malevolent threats is an issue of emerging national and international importance. As the world population increases, new construction must be increasingly cost effective and at the same time increasingly secure, safe, and durable. As the existing infrastructure ages, materials and techniques for retrofitting must be developed in parallel with improvements in design, engineering, and building codes for new construction. Both new and renovated structures are more often being subjected to the scrutiny of risk analysis. An international conference, "Assuring the Performance of Buildings and Infrastructures," was held in May 1997 to address some of these issues. The conference was co-sponsored by the Architectural Engineering Division of the American Society of Civil Engineers (ASCE), the American Institute of Architects, and Sandia National Laboratories and convened in Albuquerque, NM. Many of the papers presented at the conference are found within this issue of Techno20~. This paper presents some of the major conference themes and summarizes discussions not found in the other papers.

In rough agreement with experimental values derived from Cu island shapes vs. temperature, ab-initio calculations yield formation energies of 0.27 and 0.26 eV/ step-edge-atom for (100)- and (111)-micro facet steps on Cu(lll), and 0.09 and 0.12 eV per kink in those steps. Comparison to ab-initio results for Al and Pt shows that as a rule, the average formation energy of straight steps on a close-packed metal surface equals -7% of the metal's cohesive energy.

The carrier lifetime in vertical cavity surface emitting lasers is determined for various oxide aperture sizes using turn-on delay measurements. The lifetime is relatively constant for large devices, but decreases for aperture sizes below 9 x 9 µm.

Chemical sensing with a magnetically excited flexural plate wave (mag- FPW) resonator has been demonstrated for the first time. One surface of the resonator was coated with ethyl cellulose to impart sensitivity to volatile solvents such as chloroform, tetrachloroethylene, trichloroethylene, and toluene. The absorbed mass of the analyte causes a shift in the membrane resonance frequency of the two-port mag-FPW resonator. An oscillator circuit is used to track the resonance frequency, providing a convenient means of monitoring analyte concentration levels. Analyte concentrations of 10 ppm were easily detected.

Coulomb driven, magneto-optically induced electron and hole bound states from a series of heavily doped GaAs/Al_0.3Ga_0.7As single heterojunctions (SHJ) are revealed in high magnetic fields. At low magnetic fields ({nu} >2), the photohuninescence spectra display Shubnikov de-Haas type oscillations associated with the empty second subband transition. In the regime of the Landau filling factor {nu} <1 and 1< {nu} <2, we found strong bound states due to Mott type Vocalizations. Since a SHJ has an open valence band structure, these bound states area unique property of the dynamic movement of the valence holes in strong magnetic fields.

The magnetophotoluminescence (MPL) behavior of a GaAs/Al_0.3Ga_0.7As single heterojunction has been investigated to 60T. We observed negatively charged singlet and triplet exciton states that are formed at high magnetic fields beyond the {nu}=l quantum Hall state. The variation of the charged exciton binding energies are in good agreement with theoretical predictions. The MPL transition intensities for these states showed intensity variations (maxima and minima) at the {nu}=l/3 and 1/5 fractional quantum Hall (FQH) state as a consequence of a large reduction of electron-hole screening at these filling factors.

The transport properties of a quasi-three-dimensional, 200 layer quantum well structure are investigated at integer filling in the quantum Hall state. We find that the transverse magnetoresistance R_xx, the Hall resistance R_xy, and the vertical resistance R_zz all follow a similar behavior with both temperature and in-plane magnetic field. A general feature of the influence of increasing in-plane field B_in is that the Hall conductance quantization first improves, but above a characteristic value B^C_in, the quantization is systematically removed. We consider the interplay of the chid edge state transport and the bulk (quantum Hall) transport properties. This mechanism may arise from the competition of the cyclotron energy with the superlattice band structure energies. A comparison of the resuIts with existing theories of the chiral edge state transport with in-plane field is also discussed.

We present here the results of polarized magneto-photoluminescence measurements on a high mobility single-heterojunction. The presence of a doublet structure over a large magnetic field range (2>v>l/6) is interpreted as possible evidence for the existence of a magneto-roton minima of the charged density waves. This is understood as an indication of strong electronic correlation even in the case of the IQHE limit.

Magnetic semiconductors offer a unique possibility for strongly tuning the intrinsic alloy disorder potential with applied magnetic field. We report the direct observation of a series of step-like reductions in the magnetic alloy disorder potential in single ZnSe/Zn(Cd,Mn)Se quantum wells between O and 60 Tesla. This disorder, measured through the linewidth of low temperature photoluminescence spectra drops abruptly at -19, 36, and 53 Tesla, in concert with observed magnetization steps. Conventional models of alloy disorder (developed for nonmagnetic semiconductors) reproduce the general shape of the data, but markedly underestimate the size of the linewidth reduction.

Single-level parallel optimization approaches, those in which either the simulation code executes in parallel or the optimiza- tion algorithm invokes multiple simultaneous single-processor analyses, have been investigated previously and been shown to be effective in reducing the time required to compute optimal solutions. However, these approaches have clear performance limita- tions that prevent effective scaling with the thousands of processors available in massively parallel supercomputers. In more recent work, a capability has been developed for multilevel parallelism in which multiple instances of multiprocessor simulations are coordinated simultaneously. This implementation employs a master-slave approach using the Message Passing Interface (MPI) within the DAKOTA software toolkit. Mathematical analysis on achieving peak efficiency in multilevel parallelism has shown that the most effective processor partitioning scheme is the one that limits the size of multiprocessor simulations in favor of concurrent execution of multiple simulations. That is, if both coarse-grained and fine-grained parallelism can be exploited, then preference should be given to the coarse-grained parallelism. This analysis was verified in multilevel paralIel computatiorud experiments on networks of workstations (NOWS) and on the Intel TeraFLOPS massively parallel supercomputer. In current work, methods for exploiting additional coarse-grained parallelism in optimization are being investigated so that fine-grained efficiency losses can be further minimized. These activities are focusing on both algorithmic coarse-grained parallel- ism (multiple independent function evaluations) through the development of speculative gradient methods and concurrent iterator strategies and on function evaluation coarse-grained parallelism (multiple separable simulations within a function evaluation) through the development of general partitioning and nested synchronization facilities. The net result is a total of four separate lev- els of parallelism which can minimize efficiency losses and achieve near linear scaling on massively parallel computers.

We have designed and tested a prototype dish/Stirling hybrid-receiver combustion system. The system consists of a pre-mixed natural-gas burner heating a pin-finned sodium heat pipe. The design emphasizes simplicity, low cost, and ruggedness. Our test was on a 1/6^th -scale device, with a nominal firing rate of 18kWt, a power throughput of 13kWt, and a sodium vapor temperature of 750°C. The air/fuel mixture was electrically preheated to 640°C to simulate recuperation. The test rig was instrumented for temperatures, pressures, flow rates, overall leak rate, and exhaust emissions. The data verify our burner and heat-transfer models. Performance and post-test examinations validate our choice of materials and fabrication methods. Based on the 1/6^th -scale results, we are designing a till-scale hybrid receiver. This is a fully-integrated system, including burner, pin-fin primary heat exchanger, recuperator (in place of the electrical pre-heater used in the prototype system), solar absorber, and sodium heat pipe. The major challenges of the design are to avoid pre-ignition, achieve robust heat-pipe performance, and attain long life of the burner matrix, recuperator, and flue-gas seals. We have used computational fluid dynamics extensively in designing to avoid pre-ignition and for designing the heat-pipe wick, and we have used individual component tests and results of the 1/6^th -scale test to optimize for long life. In this paper, we present our design philosophy and basic details of our design. We describe the sub-scale test rig and compare test results with predictions. Finally, we outline the evolution of our full-scale design, and present its current status.

Successful techniques have been developed for simulating some experimental shipboard fires. The experimental fues were staged in Holds 4 and 5 of the Mayo Lykes, a test ship operated by the United States Coast Guard Fire and Safiety Test Detachment at Little Sand Island in Mobile Bay, Alabama. The tests simulated an engine-room or galley fire in the compartment adjacent to simulated hazardous cargo. The purpose of these tests was to determine the effect the fires in Hold 4 had on the cargo in Holds 4 and 5. The simulation is done with CFX, a commercial computational fluid dynamics code. Analyses show that simulations can accurately estimate a maritime fire environment for radioactive materials packaging. Radiative heat transfer dominates the hold-fue environment near the hot bulkhead. Flame temperatures between 800 and 1000°C give heat fluxes and temperatures typical of the measured fire environment for the simulated radioactive materials package. The simulation predicted the occurrence of flow patterns near the calorimeter (simulated radioactive materials package) similar to those observed during the experiment. The simulation was also accurate in predicting a heated fluid layer near the ceiling that increases in thickness as time passes.

We have constructed a humidity-controlled chamber in which deflections of polysilicon cantilever beams are observed by interferometry, resulting in in-situ adhesion measurements within a fracture mechanics framework. From adhesion energy measurements for uncoated hydrophilic beams, we demonstrate an exponential dependence of adhesion on relative humidity (RH). We can explain this trend with a single-asperity model for capillary condensation. For coated hydrophobic beams, adhesion is independent of RH up to a threshold value which depends on the coating used. However, we have found that exposure to very high RH ({ge}90%) ambients can cause a dramatic increase in adhesion, surprisingly with a stronger effect for perfluorodecyltrichlorosilane (FDTS, C{sub 10}H{sub 4}F{sub 17}SiCl{sub 3}) than octadecyltrichlorosilane (ODTS, C{sub 18}H{sub 37}SiCl{sub 3}). Newly developed computational mechanics to measure adhesion in the presence of an applied load allow us to explore how the adhesion increase develops. We believe that water adsorption at silanol sites at the FDTS/substrate interface, possibly exacerbated by coupling agent migration, leads to water islanding and the subsequent adhesion increase at very high RH levels.

We report the first experimental observation of non-classical morphological equilibration of a corrugated crystalline surface. Periodic rippled structures with wavelengths of 290-550 nm were made on Si(OO1) by sputter rippling and then annealed at 650 - 750 °C. In contrast to the classical exponential decay with time, the ripple amplitude, A_{lambda}(t), followed an inverse linear decay, A_{lambda}(t)= A_{lambda}(0)/(1 +k_{lambda}t), agreeing with a prediction of Ozdemir and Zangwill. We measure the activation energy for surface relaxation to be 1.6±0.2 eV, consistent with an interpretation that dimers mediate transport.

Shear bands and faults are ubiquitous features of brittle rock deformation at a variety of length scales. Despite the prevalence of these features, understandhg of their inception remains rudimentary. Laboratory experiments suggest a casual association of localization of deformation (faulting) with peak stress, but more detailed examination reveals that localization can precede or follow the peak. Rudnicki and Rice (1975, hereafter abbreviated as RR) have suggested a the- ory of the inception of localization as a bifurcation or nonuniqueness of the so- lution for homogeneous deformation. They predict a strong dependence of local- ization on deformation state. In particular, they predict that localization can occur prepeak for deformation states near deviatoric pure shear and does not occur until well after peak for axisymmetric compression. This prediction is roughly in ac- cord with the true triaxial experiments of Mogi (1967, 1971). More recently, Ord et al. (1991) and Wwersik et al. (1991) have reported observations of localization prior to peak stress in plane strain experiments. The predictions of RR depend strongly on the constitutive properties of the rock and detailed comparison has been impeded by inadequate knowledge of those properties. Even the idealized constitutive model used by RR requires knowledge of the evolution of the constitutive properties with inelastic deformation that is not readily obtainable from the typical axisymmetric compression test. Although it is conceptually advantageous to consider inelastic deformation at fixed mean stress, the mean stress changes throughout the axisymmetric compression test. In this paper, we present a synthesis of a number of axisymmetric compres- sion tests to extract a detailed implementation of the constitutive framework used by RR. The resulting constitutive relation is then used to -predict the response for plane strain. Conditions for localization of deformation derived by RR are evalu- ated for both plane strain and axisymmetric compression.

Structured conversation diagrams, or conversation specifications, allow agents to have predictable interactions and achieve predefined information-based goals, but they lack the flexibility needed to function robustly in an unpredictable environment. We propose a mechanism that combines a typical conversation structure with a separately established policy to generate an actual conversation. The word "policy" connotes a high-level direction external to a specific planned interaction with the environment. Policies, which describe acceptable procedures and influence decisions, can be applied to broad sets of activity. Based on their observation of issues related to a policy, agents may dynamically adjust their communication patterns. The policy object describes limitations, constraints, and requirements that may affect the conversation in certain circumstances. Using this new mechanism of interaction simplifies the description of individual conversations and allows domain-specific issues to be brought to bear more easily during agent communication. By following the behavior of the conversation specification when possible and deferring to the policy to derive behavior in exceptional circumstances, an agent is able to function predictably under normal situations and still act rationally in abnormal situations. Different conversation policies applied to a given conversation specification can change the nature of the interaction without changing the specification.

The Instrumentation and Telemetry Departments at Sandia National Laboratories have been exploring the instrumentation of sealed canisters where the flight application will not tolerate either the presence of a chemical power source or penetration by power supply wires. This paper will describe the application of a low power micro-controller based instrumentation system that uses magnetic coupling for both power and data to support a flight application.

Junction field effect transistors (JFET) are fabricated on a GaN epitaxial structure grown by metal organic chemical vapor deposition (MOCVD). The DC and microwave characteristics of the device are presented. A junction breakdown voltage of 56 V is obtained corresponding to the theoretical limit of the breakdown field in GaN for the doping levels used. A maximum extrinsic transconductance (g_m) of 48 mS/mm and a maximum source-drain current of 270 mA/mm are achieved on a 0.8 µ m gate JFET device at V_GS= 1 V and V_DS=15 V. The intrinsic transconductance, calculated from the measured g_m and the source series resistance, is 81 mS/mm. The f_T and f_max for these devices are 6 GHz and 12 GHz, respectively. These JFETs exhibit a significant current reduction after a high drain bias is applied, which is attributed to a partially depleted channel caused by trapped hot-electrons in the semi-insulating GaN buffer layer. A theoretical model describing the current collapse is described, and an estimate for the length of the trapped electron region is given.

Ferrochelatase (EC 4.99.1.1), the terminal enzyme of the heme biosynthetic pathway, catalyzes Fe²⁺ chelation into protoporphyrin IX. Resonance Raman and W-visible absorbance spectroscopes of wild type and engineered variants of murine ferrochelatase were used to examine the proposed structural mechanism for iron insertion into protoporphyrin by ferrochelatase. The recombinant variants (i.e., H207N and E287Q) are enzymes in which the conserved amino acids histidine-207 and glutamate-287 of murine ferrochelatase were substituted with asparagine and glutamine, respectively. Both of these residues are at the active site of the enzyme as deduced from the Bacillus subtilis ferrochelatase three-dimensional structure. Addition of free base or metalated porphyrins to wild type ferrochelatase and H207N variant yields a quasi 1:1 complex, possibly a monomeric protein-bound species. In contrast, the addition of porphyrin (either free base or metalated) to E287Q is sub-stoichiometric, as this variant retains bound porphyrin in the active site during isolation and purification. The specificity of porphyrin binding is confirmed by the narrowing of the structure-sensitive resonance Raman lines and the vinyl vibrational mode. Resonance Raman spectra of free base and metalated porphyrins bound to the wild type ferrochelatase indicate a nonplanar distortion of the porphyrin macrocycle, although the magnitude of the distortion cannot be determined without first defining the specific type of deformation. Significantly, the extent of the nonplanar distortion varies in the case of H207N- and E287Q-bound porphyrins. In fact, resonance Raman spectral decomposition indicates a homogeneous ruffled distortion for the nickel protoporphyrin bound to the wild type ferrochelatase, whereas both a planar and ruffled conformations are present for the H207N-bound porphyrin. Perhaps more revealing is the unusual resonance , 3 Raman spectrum of the endogenous E287Q-bound porphyrin, which has the structure-sensitive lines greatly upshifted relative to those of the free base protoporphyrin in solution. This could be interpreted as an equilibrium between protein conformers, one of which favors a highly distorted porphyrin macrocycle. Taken together these findings suggest that the mode of porphyrin distortion in murine ferrochelatase is different from that reported for yeast ferrochelatase, which requires metal binding for porphyrin distortion.

The first synthesis of an octahalotetraalkylporphyrin [2,3,7,8,12,13,17,18 -octabromo-5,10,15,20- tetrakis(trifluoromethyl)porphinato nickel(II)] is reported; this perhalogenated porphyrin has several novel properties including a very nonplanar ruffled structure with an unusually short Ni- N distance, an extremely red-shifted optical spectrum, and hindered rotation of the trifluoromethyl groups ({Delta}G_278K =47 kJ mol^-1).

Under existing methods of probabilistic risk assessment (PRA), the analysis of fire-induced circuit faults has typically been conducted on a simplistic basis. In particular, those hot-short methodologies that have been applied remain controversial in regards to the scope of the assessments, the underlying methods, and the assumptions employed. To address weaknesses in fire PRA methodologies, the USNRC has initiated a fire risk analysis research program that includes a task for improving the tools for performing circuit analysis. The objective of this task is to obtain a better understanding of the mechanisms linking fire-induced cable damage to potentially risk-significant failure modes of power, control, and instrumentation cables. This paper discusses the current status of the circuit analysis task.

Sandia National Laboratories has undertaken an ambitious, multiyear effort to greatly improve our parachute system modeling and analysis capabilities. The impetus for this effort is twofold. First, extending the stockpile lifetime raises serious questions regarding the ability of the parachutes to meet their requirements in the future due to material aging. These aging questions cannot currently be answered using available tools and techniques which are based upon the experience of expert staff and full-scale flight tests and are, therefore, not predictive. Second, the atrophy of our parachute technology base and the loss of our experienced staff has eroded our ability to respond to any future problems with stockpiled parachutes or to rapidly design a new parachute system on an experience base alone. To assure a future in-house capability for technical oversight of stockpile nuclear weapon parachutes, Sandia must move from our present empirically based approach to a computationally based, predictive methodology. This paper discusses the current status of the code development and experimental validation activities. Significant milestones that have been achieved and those that are coming up in the next year are discussed.

The sensitivity and selectivity of polyvinyl alcohol (PVA) / carbon black composite films have been found to vary depending upon the hydroxylation percentage ("-OH") of the polymer. These chemiresistors made from PVA films whose polymer backbone is 88% hydroxylated (PVA88) have a high sensitivity to water, while chemiresistors made from PVA75 have a higher sensitivity to methanol. The minor differences in polymer composition result in films with different Hildebrand volubility parameters. The relative responses of several different PVA-based chemiresistors to solvents with different volubility parameters are presented. In addition, polyvinyl acetate (PVAC) films with PVA88 are used in an array to distinguish the responses to methanol-water mixtures.

A better understanding of the fraction of contaminants irreversibly sorbed by minerals is necessary to effectively quantify bioavailability. Ferrihydrite, a poorly crystalline iron oxide, is a natural sink for sorbed contaminants. Contaminants may be sorbed/occluded as ferrihydrite precipitates in natural waters or as it ages and transforms to more crystalline iron oxides such as goethite or hematite. Laboratory studies indicate that Cd, Co, Cr, Cu, Ni, Np, Pb, Sr, U, and Zn are irreversibly sorbed to some extent during the aging and transformation of synthetic ferrihydrite. Barium, Ra and Sr are known to sorb on ferrihydrite in the pH range of 6 to 10 and sorb more strongly at pH values above its zero point of charge (pH> 8). We will review recent literature on metal retardation, including our laboratory and modeling investigation of Ba (as an analogue for Ra) and Sr adsorption/resorption, during ferrihydrite transformation to more crystalline iron oxides. Four ferrihydrite suspensions were aged at pH 12 and 50 °C with or without Ba in 0.01 M KN03 for 68 h or in 0.17 M KN03 for 3424 h. Two ferrihydrite suspensions were aged with and without Sr at pH 8 in 0.1 M KN03 at 70°C. Barium or Sr sorption, or resorption, was measured by periodically centrifuging suspension subsamples, filtering, and analyzing the filtrate for Ba or Sr. Solid subsamples were extracted with 0.2 M ammonium oxalate (pH 3 in the dark) and with 6 M HCl to determine the Fe and Ba or Sr attributed to ferrihydrite (or adsorbed on the goethite/hematite stiace) and the total Fe and Ba or Sr content, respectively. Barium or Sr occluded in goethite/hematite was determined by the difference between the total Ba or Sr and the oxalate extractable Ba or Sr. The percent transformation of ferrihydrite to goethite/hematite was estimated from the ratio of oxalate and HC1 extractable Fe. All Ba was retained in the precipitates for at least 20 h. Resorption of Ba reached a maximum of 7 to 8% of the Ba2+ added for samples aged in 0.01 and 0.17 M KN03 after 68 and 90 h of aging, respectively. About 3% of the Ba2+ added was readsorbed from 90 to 3424 h of aging in 0.17 M KN03. The amount of Ba sorbed by ferrihydrite or adsorbed on goethite (oxalate-extractable) decreased from 70 to 40% of the Ba2+ added after 68 h in 0.01 M KNO3 and from 80 to 20% of the Ba2+ added after 400 h in 0.17 M KN03. The Ba occluded in goethite (HCl-extractable) in 0.01 M KN03 increased rapidly to 30% of the Ba2+ added in the first 0.4 h and then to 50% of the Ba2+ added after 68 h. In 0.17 M KN03, Ba occluded in goethite increased from 60% of the Ba2+ added by 68 h and to 75% of the Ba2+ added after 3424 h. After 68 h at 70°C, ferrihydrite transformation was 99% compIete and was slightly inhibited with Sr present during the first few hours. Occlusion of Sr in ferrihydrite or Sr reversibly adsorbed decreased from 96 to 4o/0 after 86 h. Occlusion of Sr in hematite/goethite increased from 4 to 40% after 68 h. Resorption of Sr increased from 0.2 to 50% after 68 h. At least 90% of the Ba and 25% of the Sr added to the ferrihydrite suspensions were retained by the iron oxides during the aging periods in this study. At least 75% of the Ba and 15% of the Sr were irreversibly sorbed during ferrihydrite transfomnation to goethite and/or hematite.