Search NASA⌕ Search

Engineering topics

Yanguas-Gil, Angel

Publications and source records attributed to Yanguas-Gil, Angel.

At least 19 records

Design principles for lifelong learning AI accelerators

Lifelong learning-an agent's ability to learn throughout its lifetime-is a hallmark of biological learning systems and a central challenge for artificial intelligence (AI). The development of lifelong learning algorithms could lead to a range of novel AI applications, but this will also require the development of appropriate hardware accelerators, particularly if the models are to be deployed on edge platforms, which have strict size, weight and power constraints. Here we explore the design of lifelong learning AI accelerators that are intended for deployment in untethered environments. We identify key desirable capabilities for lifelong learning accelerators and highlight metrics to evaluate such accelerators. We then discuss current edge AI accelerators and explore the future design of lifelong learning accelerators, considering the role that different emerging technologies could play. The Perspective explores the future design of lifelong learning artificial intelligence (AI) accelerators that are intended for deployment in untethered environments, identifying key desirable capabilities for such edge AI accelerators and guidance on metrics to evaluate them.

42 ENGINEERING↗

Methods for low-temperature p-CVD and thermal ALD of magnesium diboride

ALD and p-CVD methods to generate MgB 2 and MgB 2 -containing films in the growth temperature range of 250-300° C. The thermal ALD and p-CVD methods shown herein ensure that the high-temperature-induced roughening, which causes high surface resistances in MgB2 coatings grown by the mentioned conventional techniques, is avoided. The MgB 2 and MgB 2 -containing films exhibit superconductive properties at above 20° K.

Mandia, David Joseph↗

Machine learning and atomic layer deposition: Predicting saturation times from reactor growth profiles using artificial neural networks

In this work, we explore the application of deep neural networks to the optimization of atomic layer deposition (ALD) processes. In particular, we focus on a one-shot optimization problem, where we try to predict the optimal dose time that leads to saturation everywhere in the reactor based on thickness values measured at different points of an ALD reactor after a single trial growth. In order to tackle this problem, we introduce a dataset designed to train neural networks to predict saturation times based on these inputs for a cross-flow ALD reactor. Here, we then explore the predictive ability of artificial neural networks of different depths and sizes using a separate testing dataset to evaluate their accuracies. The results obtained show that networks trained using stochastic gradient descent methods can accurately predict saturation times without requiring any additional information on the surface kinetics. This provides a viable approach to minimize the number of experiments required to optimize new ALD processes in a known reactor, and it highlights the way machine learning can be leveraged for thin film growth and manufacturing. While the datasets and training procedure depend on the reactor geometry, the trained neural networks provide a general surrogate model connecting thickness values and trial dose times with optimal saturation times that can be reused for different ALD processes within the same reactor.

36 MATERIALS SCIENCE↗

Refractory solar selective coatings

Selective receiver coatings provide high performance for concentrated solar power applications. The solar selective coating provides high solar absorptivity (90% or greater) with low IR emissivity (0.1 or less) while maintaining stability at temperatures greater than 700° C. The coating comprises a composite of a mesoporous photonic matrix with a conformal optical coating. One example composite coating includes a mesoporous photonic coating comprising a plurality of particles having sizes between 100 nm and 2000 nm, and a conformal optical coating formed by Atomic Layer Deposition (ALD) that infiltrates the mesoporous structure of the photonic coating and comprises metal nanoparticles and an amorphous dielectric matrix.

14 SOLAR ENERGY↗

Scalable synthesis of supported catalysts using fluidized bed atomic layer deposition

Overcoating layers deposited on the surface of heterogeneous catalysts using atomic layer deposition (ALD) have been shown to increase catalyst activity, lifetime, and selectivity. In this study, we performed Al 2 O 3 ALD and Pd ALD in a commercial fluidized bed reactor on high surface area mesoporous powder supports to create overcoated catalysts with high precursor utilization. We investigated the reaction mechanism for both Al 2 O 3 ALD and Pd ALD using in situ mass spectrometry and developed a mathematical model to understand the precursor saturation behaviors. We characterized the catalyst samples using a variety of techniques to measure the surface area, porosity, composition, and surface chemistry of the overcoated catalysts. Finally, we used propane dehydrogenation as a probe reaction to evaluate the performance of the catalysts prepared by fluidized bed ALD.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Biological underpinnings for lifelong learning machines

Biological organisms learn from interactions with their environment throughout their lifetime. For artificial systems to successfully act and adapt in the real world, it is desirable to similarly be able to learn on a continual basis. This challenge is known as lifelong learning, and remains to a large extent unsolved. In this Perspective article, we identify a set of key capabilities that artificial systems will need to achieve lifelong learning. We describe a number of biological mechanisms, both neuronal and non-neuronal, that help explain how organisms solve these challenges, and present examples of biologically inspired models and biologically plausible mechanisms that have been applied to artificial systems in the quest towards development of lifelong learning machines. We discuss opportunities to further our understanding and advance the state of the art in lifelong learning, aiming to bridge the gap between natural and artificial intelligence. It is an outstanding challenge to develop intelligent machines that can learn continually from interactions with their environment, throughout their lifetime. Kudithipudi et al. review neuronal and non-neuronal processes in organisms that address this challenge and discuss pathways to developing biologically inspired approaches for lifelong learning machines.

Kudithipudi, Dhireesha↗

Molecular layer etching

A method of etching an organic or hybrid inorganic/organic material. The method etches molecular layer deposition coatings. An etching cycle comprises a first half reaction exposing the coating to a precursor. A second half reaction exposes a second precursor, removing or etching a portion of the coating.

Young, Matthias John↗

Reactor scale simulations of ALD and ALE: Ideal and non-ideal self-limited processes in a cylindrical and a 300 mm wafer cross-flow reactor

Here, we have developed a simulation tool to model self-limited processes such as atomic layer deposition (ALD) and atomic layer etching inside reactors of arbitrary geometry as well the output of in situ quartz crystal microbalance and mass spectrometry. We have applied this model to two standard types of cross-flow reactors: a cylindrical tube reactor and a model 300 mm wafer reactor, and explored both ideal and nonideal self-limited kinetics. The model results are in agreement with experimental results and analytic expressions obtained using a simple plug-flow model for the cylindrical tube reactor. We also extended the simulations to consider two nonideal self-limited processes: soft-saturating processes characterized by a slow reaction pathway and processes where surface by-products can compete with the precursor for the same pool of adsorption sites. Our results show that it is possible to have a self-limited process with saturated yet inhomogeneous growth profiles due to the competition of reactor by-products. This is in agreement with experimental observations for titanium dioxide ALD from titanium tetraisopropoxide and titanium tetrachloride precursors.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Intelligent Agents for the Optimization of Atomic Layer Deposition

Atomic layer deposition (ALD) is a highly controllable thin film synthesis approach with applications in computing, energy, and separations. The flexibility of ALD means that it can access a massive chemical catalogue; however, this chemical and process diversity results in significant challenges in determining processing parameters that result in stable and uniform film growth with minimal precursor consumption. In situ measurements of the ALD growth per cycle (GPC) can accelerate process development but it still requires expert intuition and time-consuming trial and error to identify acceptable processing parameters. This procedure is made more difficult by the presence of experimental noise in the GPC values and the complexity of ALD surface chemistries. Here, a need exists for efficient optimization approaches capable of autonomously determining processing conditions resulting in optimal ALD film growth. In this work, we present the development of three optimization strategies and compare their performance in optimizing four simulated ALD processes. Furthermore, the effect of noise in the GPC measurements on optimization convergence is studied.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Tuning electronic properties in LaNiO 3 thin films by B-site Cu-substitution

Resistors are essential parts of futuristic all-oxide electronic architectures, yet easily overlooked due to their apparent simplicity. In this study, design of thin films with specific resistance spanning six orders of magnitude by partial substitution of Cu 2+ for Ni 3+ in the metallic conductor LaNiO 3 is shown. Substitution is attainable across the whole composition range using atomic layer deposition on LaAlO 3 (100) pc substrates, however with some inclusion of Ruddlesden–Popper RP1 phase (La 2 CuO 4 ) at high levels of Cu-incorporation. The thermal stability of the resistance is based on a metal–insulator transition that evolves from non-existent for LaNiO 3 to above room-temperature for high Cu 2+ substitution. This provides further insight in the metal–insulator transition found for correlated rare-earth nickelates, especially the type of transition seen for ultrathin films of LaNiO 3 that is usually attributed to oxygen vacancy formation. Materials with variable resistivity and metal–insulator-transition temperatures are key in the design of futuristic electronics such as metal–insulator-transition-driven neuromorphic devices.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Atomic layer deposition of HfO 2 films using carbon-free tetrakis(tetrahydroborato)hafnium and water

Thin hafnium oxide films were prepared by atomic layer deposition using a carbon-free precursor, tetrakis(tetrahydroborato)hafnium [Hf(BH 4 )( 4 )], and H 2 O. Film growth was studied using an in situ quartz crystal microbalance and Fourier transform infrared spectroscopy measurements. Self-limiting growth was observed between 100 and 175 degrees C, but the thermal decomposition of the Hf precursor occurred at higher temperatures. The film properties were investigated using x-ray photoelectron spectroscopy, x-ray reflectivity, x-ray diffraction, ellipsometry, time-of-flight secondary ion mass spectrometry, and x-ray absorption spectroscopy. The as-deposited films were found to consist of an amorphous mixture of HfO 2 and B 2 O 3 , and had a lower density and lower refractive index compared to pure HfO 2 thin films. Finally, annealing the films to >750 degrees C yielded crystalline monoclinic HfO 2 with a density of 9g/cm 3 and a refractive index of 2.10.

36 MATERIALS SCIENCE↗

Probing the Atomic-Scale Structure of Amorphous Aluminum Oxide Grown by Atomic Layer Deposition

Atomic layer deposition (ALD) is a well-established technique for depositing nanoscale coatings with pristine control of film thickness and composition. The trimethylaluminum (TMA) and water (H 2 O) ALD chemistry is inarguably the most widely used and yet to date, we have little information about the atomic-scale structure of the amorphous aluminum oxide (AlO x ) formed by this chemistry. This lack of understanding hinders our ability to establish process-structure-property relationships and ultimately limits technological advancements employing AlOx made via ALD. In this work, we employ synchrotron high-energy X-ray diffraction (HE-XRD) coupled with pair distribution function (PDF) analysis to characterize the atomic structure of amorphous AlO x ALD coatings. We combine ex situ and in operando HE-XRD measurements on ALD AlO x and fit these experimental data using stochastic structural modeling to reveal variations in the Al-O bond length, Al and O coordination environment, and extent of Al vacancies as a function of growth conditions. In particular, the local atomic structure of ALD AlO x is found to change with the substrate and number of ALD cycles. The observed trends are consistent with the formation of bulk Al 2 O 3 surrounded by an O-rich surface layer. We deconvolute these data to reveal atomic-scale structural information for both the bulk and surface phases. Overall, this work demonstrates the usefulness of HE-XRD and PDF analysis in improving our understanding of the structure of amorphous ALD thin films and provides a pathway to evaluate how process changes impact the structure and properties of ALD films.

36 MATERIALS SCIENCE↗

Impact of thermodynamic fluctuations and pattern size on the nucleation behavior during area selective deposition

In this work, the authors explore the impact that thermodynamic fluctuations have on the spontaneous appearance of defects during area selective deposition on patterned surfaces that are fully covered by adsorbates under local thermodynamic equilibrium, such as self-assembled monolayers. By using a simple lattice gas model for the adsorbed monolayer, the authors were able to track the spontaneous formation of defects in the monolayer as a function of the pattern width. The results indicate that, for pattern widths of the order of tens of nanometers, roughening effects at the pattern edge can be the leading source for the spontaneous appearance of nucleation defects. This leads to an enhancement of the density of defects that can be up to three orders of magnitude higher than those expected in uniform (not patterned) surfaces. The model also predicts a density of defects that is inversely proportional to the pattern width. Finally, if the dynamic nucleation of defects during area selective deposition is driven by thermodynamic fluctuations, the model predicts that the nucleation rate should be proportional to the total precursor fluency and independent of purge times. Moreover, a tight confinement of the monolayer through a high quality smooth interface in the patterned substrate and strong cohesive interactions between adsorbates should each contribute to a reduction of the overall defect density.

36 MATERIALS SCIENCE↗

Molecular Layer Etching of Metalcone Films Using Lithium Organic Salts and Trimethylaluminum

Advances in semiconductor device manufacturing are limited by our ability to precisely add and remove thin layers of material in multistep fabrication processes. Recent reports on atomic layer etching (ALE) have provided the means for the precise removal of inorganic thin films deposited by atomic layer deposition (ALD), opening new avenues for nanoscale device design. Here, we report on a new technique for the precise removal of metal organic thin films deposited by molecular layer deposition (MLD), which we term molecular layer etching. This etching process employs sequential exposures of lithium organic salt (LOS) and trimethylaluminum (TMA) precursors to produce self-limiting etching behavior. We employ quartz crystal microbalance experiments to demonstrate (i) etching of alucone films preloaded with LOS upon TMA exposures and (ii) layer-by-layer etching of alucone films using alternating exposures of LOS and TMA. We also identify the selectivity of these etching mechanisms. We probe the mechanism for the layer-by-layer etching of alucone using a quartz crystal microbalance and Fourier transform infrared spectroscopy and identify that the etching proceeds via heterolytic cleaving of Al-O bonds in alucone upon LOS exposure followed by methylation to produce volatile species upon TMA exposure. The etching process results in the removal of 0.4 nm/cycle of alucone at 160 °C and up to 3.6 nm/cycle of alucone at 266 °C in ex situ etching experiments on silicon wafers. This halogen-free etching process enables etching of MLD films and provides new fabrication pathways for the control of material geometries at the nanoscale.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Dye Nanoaggregate Structures in MK-2, N3, and N749 Dye···TiO 2 Interfaces That Represent Dye-Sensitized Solar Cell Working Electrodes

The working electrode of a dye-sensitized solar cell (DSSC) consists of dye molecules adsorbed onto nanoparticles of a semiconductor such as TiO 2 . A reliable prediction of the DSSC photovoltaic performance of a given dye requires in-depth knowledge about the precise structure of the dye···TiO 2 interface. In this work, X-ray reflectometry (XRR) and grazing-incidence small-angle X-ray scattering (GISAXS) are employed to determine the dye···TiO 2 interfacial structure and associated dye aggregation behavior of three high-performance DSSC dyes, i.e., a metal-free organic dye, MK-2, and the two archetypal ruthenium-based organometallic dyes, N3 and N749 (Black Dye). Results show that all three dyes form nanoaggregates in dye···TiO 2 interfaces. We determine the dye nanoaggregate separations, sizes, distribution densities, and the extent of short-range order within each dye self-assembly in the longitudinal and lateral directions. Dye···TiO 2 composites fabricated using dye solutions of varying concentrations are analyzed. We find that nanoaggregates of the three dyes are separated by several hundred nanometers (158-203 nm) in dye···TiO 2 interfaces that have been fabricated using concentrated dye solutions (0.5 mM or 1.0 mM). MK-2 and N749 dyes also display smaller interparticle separations. Dye nanoparticle diameters are of the order of 156-198 nm, sizes that are comparable to the largest interparticle separations. Thus, no extraneous dye particles can be fitted into gaps between particles, so the dye self-assembly is saturated. Self-assemblies of all three dyes exhibit both lateral and longitudinal short-range order; N3 displays a particularly short coherence length along the TiO 2 surface, with extensive structured disorder along the longitudinal direction. The operation of DSSC working electrodes would therefore seem to be dependent on a dye self-assembly that may exhibit several levels of structural granularity and dye aggregation effects.

14 SOLAR ENERGY↗

The chemical physics of sequential infiltration synthesis—A thermodynamic and kinetic perspective

Sequential infiltration synthesis (SIS) is an emerging materials growth method by which inorganic metal oxides are nucleated and grown within the free volume of polymers in association with chemical functional groups in the polymer. SIS enables the growth of novel polymer-inorganic hybrid materials, porous inorganic materials, and spatially templated nanoscale devices of relevance to a host of technological applications. Although SIS borrows from the precursors and equipment of atomic layer deposition (ALD), the chemistry and physics of SIS differ in important ways. These differences arise from the permeable three-dimensional distribution of functional groups in polymers in SIS, which contrast to the typically impermeable two-dimensional distribution of active sites on solid surfaces in ALD. In SIS, metal-organic vapor-phase precursors dissolve and diffuse into polymers and interact with these functional groups through reversible complex formation and/or irreversible chemical reactions. In this perspective, we describe the thermodynamics and kinetics of SIS and attempt to disentangle the tightly coupled physical and chemical processes that underlie this method. Here, we discuss the various experimental, computational, and theoretical efforts that provide insight into SIS mechanisms and identify approaches that may fill out current gaps in knowledge and expand the utilization of SIS.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Formation of Unsaturated Hydrocarbons and Hydrogen: Surface Chemistry of Methyltrioxorhenium(VII) in ALD of Mixed-Metal Oxide Structures Comprising Re(III) Units

We present the full investigation of the atomic layer deposition (ALD) of a mixed rhenium-aluminum oxide, namely ReAl 2 O 3 CH 3 , a material with tunable resistance, comprising the building unit of conductive rhenium oxides, ReOx. The deposition, involving methyltrioxorhenium(VII) (MeReO 3 , MTO) and trimethylaluminum (TMA), was analyzed by employing complementary in situ diagnostic quartz-crystal microbalance (QCM), Fourier-transform infrared (FT-IR) spectroscopy, and quadrupole mass spectrometry (QMS) to explore and reveal the underlying growth mechanism of this material. A proposed mechanism includes reductive elimination steps, thereby creating a stable Re(III)-containing thin film, making this ALD process unique regarding its growth. In addition, as proven by QMS, the surface reactions include the formation of hydrogen and unsaturated hydrocarbons. From this straightforward process, an extraordinarily high growth rate of 4.5 angstrom cycle -1 at temperatures as low as 150 °C was obtained. This material was found to exhibit highly promising electrical properties in terms of low thermal coefficient of resistance (TCR) in combination with high resistivity. By blending thin films of ReAl 2 O 3 CH 3 with additional layers (1, 2, or 3) of Al 2 O 3 , we were able to fine-tune the electrical resistivity in the range of 3.9 x 10(6) -1.5 x 10 11 Ω.cm. Simultaneously, the TCR was lowered to about -0.014 degrees °C -1 , making this material highly resistive over a broad temperature range and a promising candidate for advanced detector applications, e.g., multichannel plates (MCPs).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Memristor design rules for dynamic learning and edge processing applications

The ability to dynamically learn and adapt to changes in the environment is one of the hallmarks of biological systems. In this work, we identify the subset of the design space of memristive materials that is optimal for dynamic learning applications. Using an architecture inspired on the learning center of the insect brain, we implement a model system consisting of a discrete implementation of spiking neurons where dynamic learning takes place on a set of plastic synapses formed by memristor pairs in a crossbar array. Using two separate benchmarks, one comprising the dynamic learning of the Modified National Institute of Standards and Technology dataset and another one targeting one shot learning, we have identified the key properties that memristive materials should have to be optimal dynamic learners. The results obtained show that a fine degree of control of the memristor internal state is key to achieve high classification accuracy during dynamic learning but that within this optimal region learning is extremely robust both to device variability and to errors in the writing of the internal state, in all cases allowing for 2 sigma variations greater than 40% without significant loss of accuracy, hence overcoming two of the perceived limitations of memristors. By additionally requiring that learning takes place concurrently to information processing, we are able to derive a set constraints to the memristor dynamics.

36 MATERIALS SCIENCE↗