OceanSpace Research Performance Progress Report-FINAL Advancing optical imaging and classification to enhance biodiversity monitoring.
Explore the source record for details and available documents.
09 BIOMASS FUELS↗
Engineering topics
Arnold, William
Publications and source records attributed to Arnold, William.
Toward a Holistic Performance Evaluation of Large Language Models Across Diverse AI Accelerators
Artificial intelligence (AI) methods have become critical in scientific applications to help accelerate scientific discovery. Large language models (LLMs) are being considered a promising approach to address some challenging problems because of their superior generalization capabilities across domains. The effectiveness of the models and the accuracy of the applications are contingent upon their efficient execution on the underlying hardware infrastructure. Specialized Al accelerator hardware systems have recently become available for accelerating Al applications. However, the comparative performance of these AI accelerators on large language models has not been previously studied. In this paper, we systematically study LLMs on multiple AI accelerators and GPUs and evaluate their performance characteristics for these models. We evaluate these systems with (i) a micro-benchmark using a core transformer block, (ii) a GPT-2 model, and (iii) an 1,I,M-driven science use case, GenSLM. We present our findings and analyses of the models' performance to better understand the intrinsic capabilities of AI accelerators. Furthermore, our analysis takes into account key factors such as sequence lengths, scaling behavior, and sensitivity to gradient accumulation steps.
Emani, Murali↗
Highly Conductive Iodine and Fluorine Dual-Doped Argyrodite Solid Electrolyte for Lithium Metal Batteries
Sulfide-type argyrodite solid electrolytes (SEs) with halide doping have attracted serious interest. Although other halides such as Cl and Br have been found to enhance Li-ion transport in argyrodites, the direct synthesis of efficiently conductive Li 6 PS 5 I without postprocessing has rarely been investigated. In this work, we report the one-step synthesis of highly conductive Li 6 PS 5 I with an impressive ionic conductivity of 2.5 × 10 –4 S cm –1 at room temperature through a solvent-based method. Moreover, by introducing F – to partially replace I – , hybrid-doped argyrodites Li 6 PS 5 F x I 1–x (x = 0.25, 0.5, 0.75) have been synthesized. Li 6 PS 5 F 0.25 I 0.75 achieves the highest conductivity of 3.5 × 10 –4 S cm –1 due to the energetic preference for anion-disordering among F – /I – and S 2– , which facilitate faster Li transport as supported by density functional theory (DFT) calculations. With higher F content in argyrodites, Li 6 PS 5 F 0.75 I 0.25 displays the best electrochemical stability toward Li metal, as evidenced by long-term stable cycling in Li symmetric cells up to 1100 h. Solid-state Li metal batteries with an active cathode of Li 4 Ti 5 O 12 (LTO) display an initial specific capacity of 140 mAh g –1 and remain at 105 mAh g –1 after 200 cycles, suggesting great battery cycling performance. Furthermore, this research has developed new compositions in the argyrodite SE family which could lead to advancements in the development of solid-state Li metal batteries.
25 ENERGY STORAGE↗
Solvent-free and low temperature synthesis of chalcogenide Na superionic conductors for solid-state batteries
Sodium chalcogenide ionic conductors are attractive candidates as solid electrolytes (SEs) in solid-state Na metal batteries. They show the advantages of high ionic conductivity of 10 –4 –10 –2 S cm –1 at room temperature and great chemical stability in air. However, simple, efficient, and scalable approaches for the synthesis of chalcogenide solid electrolytes (SEs) are required. In this work, we report a solvent-free mixing to form dry intermediate products, which are subjected to different treatments (electron-beam assisted method or low temperature heating (≤150 °C)) to produce pure phase of Na 3 SbS 4-y Se y (0 ≤ y < 2) chalcogenides. Heavy Se-doping in Na 3 SbS 4 results in the tetragonal-to-cubic phase transition as well as a significant change of Sb-S bonding in Raman spectra. Among all chalcogenide SEs, Na 3 SbS 3 Se showed the highest ionic conductivity of 3.75 × 10 –4 S cm –1 at room temperature, 47% higher than that of pristine Na 3 SbS 4 . Moreover, the Se-dopant also enhanced the electrochemical stability towards Na metal in solid-state batteries. The solid-state Na||FeS 2 battery with Na 3 SbS 3 Se SE displayed long-term cycling ability up to 1,000 cycles within the voltage window of 1.0–2.7 V and retained a specific capacity of 105 mAh g –1 after 600 cycles. As a result, this technique promotes the practical applications of chalcogenide SEs in solid-state batteries.
25 ENERGY STORAGE↗
Wet chemical synthesis and properties of argyrodite sulfide solid electrolytes for solid state lithium batteries
The commercialization of the lithium-ion battery (LIB) in 1991 was responsible for the explosion in portable electronic technologies that has been seen over the past 30 years. With the advent of electric vehicles and other high-powered technologies, there is tremendous demand for LIBs with higher energy density and high safety. To achieve this, new electrode materials must be explored. The obvious choice of anode material would be pure metal lithium, which has a theoretical specific capacity of 3860 mAh g-1 . Unfortunately, metal lithium anodes have not been widely commercialized due to their tendency to react violently with the flammable liquid electrolytes used in today’s batteries. Battery safety can best be achieved by adopting solid electrolytes in place of liquid electrolytes. Solid electrolytes are nonvolatile and nonflammable, safely allowing for the combination of high-capacity cathode materials with a Li metal anode. Argyrodite sulfide solid electrolytes such as halogen-doped Li6PS5X (X = Cl, Br, I) are noted for their high ionic conductivity. But before sulfides can be commercially adopted, they possess several disadvantages which must be addressed, including time- and energy-consuming synthesis processes, poor electrochemical stability, and intrinsically poor air stability. This dissertation seeks to address each of these challenges through materials design an synthesis strategies. In this work, we pioneer a solvent-based approach for the synthesis of argyrodite solid electrolytes Li7PS6 and Li6PS5Xinstead of a stringent solid-state synthesis. Nontoxic ethanol is employed as the solvent, enabling a rapid synthetic approach to produce argyrodite solid electrolytes with high phase purity and compositional flexibility. Compared with Li7PS6, halogen doping (i.e. X = F, Cl, Br, I) not only increases the ionic conductivity, but also enhances the electrochemical stability at the interface towards Li metal. Specifically, F-doped argyrodites produce a robust SEI layer containing LiF, contributing to enhanced interfacial stability. Finally, to address the air instability challenge, argyrodite-incorporated composite solid electrolytes (CSEs) are designed and prepared to produce stable and flexible membranes that are demonstrated in solid-state Li metal batteries. These advances push argyrodite sulfide solid electrolyte research further and pave the way for the proliferation of next generation lithium metal batteries.
25 ENERGY STORAGE↗
Synthesis of Fluorine-Doped Lithium Argyrodite Solid Electrolytes for Solid-State Lithium Metal Batteries
Solid-state lithium metal batteries (SSLMBs) that utilize novel solid electrolytes (SEs) have garnered much attention because of their potential to yield safe and high-energy-density batteries. Sulfide-based argyrodite-class SEs are an attractive option because of their impressive ionic conductivity. Recent studies have shown that LiF at the interface between Li and SE enhances electrochemical stability. However, the synthesis of F-doped argyrodites has remained challenging because of the high temperatures used in the state-of-the-art solid-state synthesis methods. In this work, for the f irst time, we report F-doped Li 5+y PS 5 F y argyrodites with a tunable doping content and dual dopants (F-/Cl- and F-/Br-) that were synthesized through a solvent-based approach. Among all compositions, Li 6 PS 5 F 0.5 C l0.5 exhibits the highest Li-ion conductivity of 3.5 x 10 -4 S cm -1 at room temperature (RT). Furthermore, Li symmetric cells using Li 6 PS 5 F 0.5 Cl 0.5 show the best cycling performance among the tested cells. X-ray photoelectron spectroscopy and ab initio molecular dynamics simulations revealed that the enhanced interfacial stability of Li 6 PS 5 F 0.5 C l0.5 SE against Li metal can be attributed to the formation of a stable solid electrolyte interphase (SEI)-containing conductive species (Li 3 P), alongside LiCl and LiF. These findings open new opportunities to develop high-performance SSLMBs using a novel class of F-doped argyrodite electrolytes.
25 ENERGY STORAGE↗
Highly efficient interface stabilization for ambient-temperature quasi-solid-state sodium metal batteries
Solid-state sodium (Na) batteries (SSSBs) using sulfide-based solid electrolytes (SSEs) hold tremendous promise due to their high theoretical specific capacity, enhanced safety and abundant resources. However, detrimental interfacial issues between SSEs and Na metal present a major challenge to the advancement of sulfide-based SSSBs. To address interfacial issues, we demonstrate an efficient approach by incorporating an ionic liquid electrolyte ((PYR/Na)TFSI) as interlayer to stabilize the Na metal/SSE interface. The presence of the (PYR/Na)TFSI interlayer enables the formation of a stable solid electrolyte interphase (SEI) to prevent the harmful reactions and inhibit Na dendrites. Combination of ab initio molecular dynamics simulations and X-ray photoelectron spectroscopy revealed that this stable SEI is largely composed of reduced products of TFSI -, such as NaF and CF 3 . As a result, the symmetric cells exhibited stable Na plating/striping cycling for 300 h at 0.1 mA cm -2 . In addition, FeS2||Na quasi-solid-state batteries delivered an impressive specific capacity of over 300 mAh g -1 under the current density of 20 mA g -1 at room temperature. Under a higher current density (100 mA g -1 ), such batteries performed with long-term cycling stability and maintained a specific capacity of around 103 mAh g -1 after 330 cycles. This work demonstrates the novel perspective of using an ionic liquid interlayer to address interfacial issues, contributing to the advancement of high-performance SSSBs for the next-generation energy storage systems.
25 ENERGY STORAGE↗
Fast Na diffusion and anharmonic phonon dynamics in superionic Na 3 PS 4
The design of new solid electrolytes (SEs) hinges on identifying and tuning relevant descriptors. Phonons describe the atomic dynamics in crystalline materials and provide a basis to encode possible minimum energy pathways for ion migration, but anharmonic effects can be large in SEs. Identifying and controlling the pertinent phonon modes coupled most strongly with ionic conductivity, and assessing the role of anharmonicity, could therefore pave the way for discovering and designing new SEs via phonon engineering. In this work, we investigate phonons in Na 3 PS 4 and their coupling to fast Na diffusion, using a combination of neutron scattering, ab initio molecular dynamics (AIMD), and extended molecular dynamics based on machine-learned potentials. We identify that anharmonic soft-modes at the Brillouin zone boundary of the anharmonically stabilized cubic phase constitute key phonon modes that control the Na diffusion process in Na 3 PS 4 . We demonstrate how these strongly anharmonic phonon modes enable Na-ions to hop along the minimum energy pathways. Further, the quasi-elastic neutron scattering (QENS) measurements, supplemented with ab initio and machine-learned molecular dynamics simulations, probe the Na diffusivity and diffusion mechanism. These results offer detailed microscopic insights into the dynamic mechanism of fast Na diffusion and provide an avenue to search for further Na solid electrolytes.
42 ENGINEERING↗
Li 0.625 Al 0.125 H 0.25 Cl 0.75 O 0.25 Superionic Conductor with Disordered Rock-Salt Structure
Solid-state Li-ion conductors are of broad interest in electrochemical energy storage, especially in solid-state Li batteries that serve as a promising alternative for the next-generation safe and high-energy-density batteries. Exploring solid-state superionic conductors is significant for the development of solid-state Li batteries with high performance. Herein, we report a disordered rock-salt (A 1 B 1 )-structured solid electrolyte (Li 0.625 Al 0.125 H 0.25 )(Cl 0.75 O 0.25 ) (abbr. LAHCO) that was synthesized using Li 2 OHCl and LiAlCl 4 as precursors. Neutron diffraction reveals that Li, Al, and H atoms occupy the A sites and O and Cl atoms occupy the B sites in the A 1 B 1 structure for pure LAHCO. The LAHCO compound with excess LiAlCl 4 shows the highest Li + ionic conductivity of ~10 –4 S cm –1 at room temperature due to the disordering induced by configurational entropy as well as the entropy of mixing. Moreover, LAHCO–LiAlCl 4 solid electrolyte exhibits a stable polarization voltage under a current density of 5–50 μA cm –2 in Li symmetric cells. Furthermore, this work not only explicates the importance of Li-ion conductors with a rock-salt structure but also contributes toward the development of solid-state Li-ion conductors for broad applications.
25 ENERGY STORAGE↗
Phase-Transition Interlayer Enables High-Performance Solid-State Sodium Batteries with Sulfide Solid Electrolyte
All-solid-state sodium (Na) batteries (ASSSBs) using sulfide-based solid electrolytes (SEs) have attracted intensive attention due to their superior safety, high energy density and low cost. However, interfacial issue is one of the biggest challenges to achieve high-performance sulfide-based ASSSBs due to the serious reactions between active Na metal and sulfide SEs at the interface. To address the interfacial challenges, we propose in this paper a simple and efficient approach by introducing a polymer electrolyte as an interlayer to stabilize the interface. Na 3 SbS 4 as a model sulfide SE is used to demonstrate the interlayer strategy to stabilize the interface by preventing the detrimental reactions and inhibiting Na dendrites. As a result, stable Na plating/stripping was observed in Na symmetric cells under the current density of 0.1 mA cm –2 . Moreover, ASSSBs with Na metal and TiS 2 electrode delivered long-term stability over 300 cycles remaining a specific capacity above 100 mAh g –1 , and FeS 2 ||Na cells exhibited an impressive specific capacity of up to 200 mAh g –1 after the 20th cycles. This work demonstrates an efficient strategy to address interfacial challenges between sulfide SEs and Na metal, which contributes to the development of ASSSBs in next-generation energy storage systems.
25 ENERGY STORAGE↗
Interface stability of LiCl-rich argyrodite Li 6 PS 5 Cl with propylene carbonate boosts high-performance lithium batteries
In solid-state lithium (Li)-metal batteries (SSLMBs), sulfide-based Li-ion conductors represent one of the most popular solid electrolytes (SEs). However, the development of sulfide-based SSLMBs is significantly hampered by interfacial issues including large solid-solid contact resistance and serious side reactions at the SE/electrode interface. To address these issues, here, we demonstrated a simple and efficient strategy by using LiCl-rich argyrodite Li 6 PS 5 Cl (Li 6 PS 5 Cl-LiCl) SE and trace amount of propylene carbonate (PC) at the SE/electrode interface to facilitate the formation of stable and robust solid electrolyte interphase (SEI). Additionally, the formed SEI not only serves as a buffer layer to passivate the interfacial reactions and suppress Li dendrite growth, but also acts as a bridge for Li-ion conduction to reduce the contact resistance. As a result, the Li||Li symmetric cell exhibited long-term electrochemical cycling stability over 1000 h at a current density of 0.2 mA cm -2 . Furthermore, the assembled Li||Li 4 Ti 5 O 12 (LTO) batteries delivered a specific capacity of 175 mAh g–1 at 0.2 C, and remained an excellent specific capacity of 116 mAh g –1 after 200 cycles at a high current rate of 1 C. These features indicate a feasible strategy to enhance the interfacial property of high-performance SSLMBs.
25 ENERGY STORAGE↗
Stable and Flexible Sulfide Composite Electrolyte for High-Performance Solid-State Lithium Batteries
Not provided.
Materials Science↗
Regulated lithium ionic flux through well-aligned channels for lithium dendrite inhibition in solid-state batteries
The utilization of solid electrolyte (SE) to suppress lithium (Li) dendrites is promising but still far from satisfactory due to the inhomogeneous Li plating/stripping. In this paper, we demonstrated a novel strategy to inhibit Li dendrites via regulating Li-ion flux in SE by using vertically aligned channels. The ion-insulating walls facilitated uniform distribution of Li-ion flux through the channels, leading to a homogeneous Li deposition, thus alleviating Li dendrite formation. As a result, symmetric cells with this SE exhibited excellent long-term stability (1000 h) against Li metal. In addition, Li4Ti5O12 (LTO)/Li cell with the developed SE achieved good battery performance over 100 cycles. The mechanism for dendrite suppression was further investigated by phase-field simulation. This work provides a novel strategy by manipulating uniform Li-ion flux to fabricate SE to inhibit Li dendrites and facilitates the development of high-performance rechargeable Li batteries.
25 ENERGY STORAGE↗
Heat Integration Optimization and Dynamic Modeling Investigation for Advancing the Coal-Direct Chemical Looping Process
The purpose of the project is to address the optimization and startup operation of a modular coal direct chemical looping (CDCL) combustion system integrated with a steam cycle for power generation to reduce the risks involved in further scale-up of the technology. The modular reactor design of the CDCL process provides flexibility in the fabrication of the reactor and in its operating capacity (i.e. turndown ratio) at the cost of a more complex heat exchange network (HEN) design and integration. To address the technology gaps and advance the efficiency and economic feasibility of the CDCL technology, the project will perform a detailed and comprehensive analysis of the integration of a modular CDCL reactor system and a steam cycle system under both static and transient conditions via HEN process performance simulations and system dynamic modeling, respectively. The scope of work consists of 1) Experimental and computational studies of the CDCL combustor reactor 2) Comprehensive static (i.e. steady-state) system HEN design analysis in CDCL 550 MWe commercial unit for power generation and 3) Dynamic modeling of site specific design of 10MWe CDCL large pilot plant. The project team has successfully developed and validated a kinetic model for the oxidation of oxygen carriers in the combustor using the unreacted shrinking core model (UCSM). The model is capable of capturing the oxidation kinetics of fully or partially reduced oxygen carrier particles. A computational fluid dynamics (CFD) model is developed to simulate the hydrodynamics, heat transfer, and chemical reaction occurring in the CDCL combustor. The model is developed in MFIX and ANSYS Fluent. Key aspects of CDCL combustor operation, including heat transfer, oxygen carrier oxidation, and the transport of oxygen carrier particles, are simulated using this CFD model. The HEN for a commercial scale 550 MWe CDCL power plant is simulated and optimized using ASPEN Plus. Practical design considerations are incorporated based on industrial experiences. The performance and cost for the commercial CDCL plant is updated based on these analyses. A dynamic model for the 10 MWe CDCL pilot plant is developed in ProTRAX simulation software. The model is based on the pilot plant design developed in project DE-FE0027654 “10 MWe CDCL Large Pilot Plang – Pre-FEED Study” and the steam cycle data obtained from Dover Light & Power plant. The transient behaviors during pilot plant load variation are simulated using the dynamic model.
01 COAL, LIGNITE, AND PEAT↗