Peter M. Attia
Research Subjects
Peter M. Attia has a total of 150 co-authors. Here are the top co-authors by number of shared publications:
- William C. Chueh
Department of Materials Science and Engineering, Stanford University, Stanford, USA
10 shared publications - Martin Z. Bazant
Massachusetts Institute of Technology
7 shared publications - Stephen J. Harris
Lawrence Berkeley National Laboratory
5 shared publications Amariah Condon
Glimpse
4 shared publications- Amariah Condon
Glimpse Engineering Inc., Somerville, Massachusetts 02143, United States of America
4 shared publications Norman Jin
Department of Materials Science and Engineering, Stanford University, Stanford, California 94305, United States
4 shared publications
Latest Publications
- Open Access
- 26 August 2026
- Published by Elsevier in Cell Reports Physical Science
- Vol. 7 (9)
Bridging the gap between lab discovery and commercial deployment
- 26 August 2026
- Published by Elsevier in Cell Reports Physical Science
- Vol. 7 (9)
Open AccessBridging the gap between lab discovery and commercial deployment
- 7 July 2026
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2026-01 (6) , 675
Minute Differences in Battery Geometry Explain Electrochemical Variability
Device variability in high-volume manufacturing is intrinsic in many technologies. Such is the case in lithium-ion battery cells, where cell-to-cell electrochemical variations impact cell-to-pack integration and battery management systems. Here, we combine computer vision and 120-second X-ray computed tomography (CT) across 295 commercial 21700 cylindrical cells to explain and predict electrochemical performance variability, both at the beginning and end of life. Surprisingly, the cell electrochemical performances (varying by 0.51% in capacity and energy, by 4.32% in power, and by 10.6% in cycle life) correlate strongly to a single descriptor: eccentricity of the cylindrical steel cans (absolute values variations <0.99%). Through interpretable machine learning, these fast CT predict variations in energy and power throughout aging without any electrochemical cycling data. This work illustrates the utility of computer vision and explainable machine learning for revealing causal links between the internal structures of complex devices and their performance and aging, in the case of batteries, without needing electrochemical characteristics as input.
- 7 July 2026
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2026-01 (6) , 675
Minute Differences in Battery Geometry Explain Electrochemical Variability
Device variability in high-volume manufacturing is intrinsic in many technologies. Such is the case in lithium-ion battery cells, where cell-to-cell electrochemical variations impact cell-to-pack integration and battery management systems. Here, we combine computer vision and 120-second X-ray computed tomography (CT) across 295 commercial 21700 cylindrical cells to explain and predict electrochemical performance variability, both at the beginning and end of life. Surprisingly, the cell electrochemical performances (varying by 0.51% in capacity and energy, by 4.32% in power, and by 10.6% in cycle life) correlate strongly to a single descriptor: eccentricity of the cylindrical steel cans (absolute values variations <0.99%). Through interpretable machine learning, these fast CT predict variations in energy and power throughout aging without any electrochemical cycling data. This work illustrates the utility of computer vision and explainable machine learning for revealing causal links between the internal structures of complex devices and their performance and aging, in the case of batteries, without needing electrochemical characteristics as input.
- 7 July 2026
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2026-01 (5) , 632
(Invited) Glimpse: Enabling Battery Quality at Scale
Founded in 2023, Glimpse’s mission is to enable battery quality at scale via high-throughput computed tomography (CT) scanning. In less than three years, Glimpse has acquired over sixty customers—ranging from small battery startups to some of the largest battery producers and OEMs in the world. In this talk, I’ll share Glimpse’s origin story, business and technical objectives, progress to date, and future directions. I’ll also discuss the lessons we’ve learned along the way.
- 7 July 2026
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2026-01 (5) , 632
(Invited) Glimpse: Enabling Battery Quality at Scale
Founded in 2023, Glimpse’s mission is to enable battery quality at scale via high-throughput computed tomography (CT) scanning. In less than three years, Glimpse has acquired over sixty customers—ranging from small battery startups to some of the largest battery producers and OEMs in the world. In this talk, I’ll share Glimpse’s origin story, business and technical objectives, progress to date, and future directions. I’ll also discuss the lessons we’ve learned along the way.
- Open Access
- 10 April 2026
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 173 (7) , 070515
Visualization and Detection of Internal Damage Caused by External Mechanical Loads in Cylindrical Li-Ion Batteries
Detecting and preventing lithium-ion battery failures remains a critical challenge, as failures degrade performance, damage materials, and pose safety risks. While catastrophic outcomes under extreme impacts or compression—such as short circuits, thermal runaway, and fires—have been extensively studied, batteries in real-world applications are more often exposed to moderate stresses. These conditions produce small internal changes with weak initial electrochemical signatures that can accumulate into severe damage over time. This work investigates how single-impact and compressive forces of varying intensities and orientations affect the structure and electrochemical behavior of 18650 NMC/graphite cylindrical cells. X-ray computed tomography (CT) provides direct visualization of damage to both active and inactive components, while open-circuit potential, impedance spectroscopy, and charge-discharge cycling collectively capture electrochemical consequences. The results reveal immediate yet modest capacity loss even after a single mechanical event, with damaged cells later showing cycling performance comparable to intact cells if only minor damage is introduced. However, repeated loading produces cumulative degradation, further aggravated by subsequent electrochemical cycling. CT imaging highlights structural changes invisible to conventional diagnostics, while electrochemical tests expose early functional impacts. Together, these complementary methods advance understanding of how moderate mechanical damage initiates, evolves, and compromises long-term reliability in lithium-ion batteries.
- 10 April 2026
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 173 (7) , 070515
Open AccessVisualization and Detection of Internal Damage Caused by External Mechanical Loads in Cylindrical Li-Ion Batteries
Detecting and preventing lithium-ion battery failures remains a critical challenge, as failures degrade performance, damage materials, and pose safety risks. While catastrophic outcomes under extreme impacts or compression—such as short circuits, thermal runaway, and fires—have been extensively studied, batteries in real-world applications are more often exposed to moderate stresses. These conditions produce small internal changes with weak initial electrochemical signatures that can accumulate into severe damage over time. This work investigates how single-impact and compressive forces of varying intensities and orientations affect the structure and electrochemical behavior of 18650 NMC/graphite cylindrical cells. X-ray computed tomography (CT) provides direct visualization of damage to both active and inactive components, while open-circuit potential, impedance spectroscopy, and charge-discharge cycling collectively capture electrochemical consequences. The results reveal immediate yet modest capacity loss even after a single mechanical event, with damaged cells later showing cycling performance comparable to intact cells if only minor damage is introduced. However, repeated loading produces cumulative degradation, further aggravated by subsequent electrochemical cycling. CT imaging highlights structural changes invisible to conventional diagnostics, while electrochemical tests expose early functional impacts. Together, these complementary methods advance understanding of how moderate mechanical damage initiates, evolves, and compromises long-term reliability in lithium-ion batteries.
- Open Access
- 11 March 2026
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 173 (5) , 050519
Electrolyte Motion Induced Salt Inhomogeneity in Si-Containing Cylindrical Cells – A Case Study
Electrolyte motion induced salt inhomogeneity (EMSI) is a failure mechanism in cylindrical and prismatic lithium-ion batteries that can lead to rapid capacity loss at high rates and eventually Li plating and cell failure. If a cell displaying EMSI is allowed to rest for several weeks before the onset of Li plating, capacity recovery is possible. This work studies the effect of varied anode silicon oxide (SiO x ) content (0 to 15% Si by weight) on EMSI in cylindrical 18650 cells by measuring capacity loss, changes in electrolyte resistance, and bulk movement of electrolyte via both moment of inertia measurements and X-ray CT imaging. With increased SiO x content, cells pump more electrolyte, display increasing electrolyte resistance with cycle number and show accelerated capacity loss. This trend did not hold true for the cell with the most SiO x , which showed electrolyte pumping without EMSI. This was possible due to a low porosity anode, which generates electrolyte pumping from both cathode and anode and does not lead to inhomogeneous salt distributions. This cell was also underfilled, likely reducing the amount of electrolyte motion. A simple model is presented for predicting the rest times required for a cell to equilibrate the electrolyte salt gradients, which agrees well with the experimental data. Three days is a typical “relaxation half-life” for these cells, with full capacity recovery after 30 1 C/1 C cycles taking up to two weeks.
- 11 March 2026
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 173 (5) , 050519
Open AccessElectrolyte Motion Induced Salt Inhomogeneity in Si-Containing Cylindrical Cells – A Case Study
Electrolyte motion induced salt inhomogeneity (EMSI) is a failure mechanism in cylindrical and prismatic lithium-ion batteries that can lead to rapid capacity loss at high rates and eventually Li plating and cell failure. If a cell displaying EMSI is allowed to rest for several weeks before the onset of Li plating, capacity recovery is possible. This work studies the effect of varied anode silicon oxide (SiO x ) content (0 to 15% Si by weight) on EMSI in cylindrical 18650 cells by measuring capacity loss, changes in electrolyte resistance, and bulk movement of electrolyte via both moment of inertia measurements and X-ray CT imaging. With increased SiO x content, cells pump more electrolyte, display increasing electrolyte resistance with cycle number and show accelerated capacity loss. This trend did not hold true for the cell with the most SiO x , which showed electrolyte pumping without EMSI. This was possible due to a low porosity anode, which generates electrolyte pumping from both cathode and anode and does not lead to inhomogeneous salt distributions. This cell was also underfilled, likely reducing the amount of electrolyte motion. A simple model is presented for predicting the rest times required for a cell to equilibrate the electrolyte salt gradients, which agrees well with the experimental data. Three days is a typical “relaxation half-life” for these cells, with full capacity recovery after 30 1 C/1 C cycles taking up to two weeks.
- Open Access
- 3 March 2026
- Published by Springer Nature in Experimental Mechanics
- Vol. 66 (7) , 1-14
Core Collapse Instability of Spirally Wound Jelly-Roll Lithium-Ion Batteries
Background: Cylindrical lithium-ion batteries are tend to have an inward deformation during charge and discharge, which is called “core collapse”. This structural instability remains poorly quantified, and a deeper understanding of the underlying mechanics is essential for improving the structural integrity and safety of rechargeable cells. Objective: This study aims to investigate the deformation and buckling mechanisms of cylindrical lithium–ion battery jelly rolls under electrochemical cycling, with a focus on the interplay between geometry, interfacial constraint, and internal stress evolution. Methods: X-ray computed tomography (CT) imaging was employed to characterize the three-dimensional deformation of commercial cells, while a detailed finite element (FE) model explicitly incorporating current collectors, coating layers, and separators was developed to simulate chemo-induced stresses. The lithiation and delithiation processes were represented through an equivalent thermal-expansion surrogate to evaluate pre-buckling stress states. Results: CT analysis revealed negligible volumetric and hoop strains, confirming strong geometric constraint from the spiral architecture. Simulations showed that interface compatibility between current collectors and electrodes governs hoop-dominated stress fields, and that separator wrinkling can precede electrode buckling due to low bending stiffness and expansion mismatch. Conclusions: The results elucidate how hoop compression, interface constraint, and structural asymmetry collectively trigger jelly-roll instability. The findings establish a mechanistic foundation for analyzing confined expansion and guiding future multi-physics modeling of battery structural reliability.
- 3 March 2026
- Published by Springer Nature in Experimental Mechanics
- Vol. 66 (7) , 1-14
Open AccessCore Collapse Instability of Spirally Wound Jelly-Roll Lithium-Ion Batteries
Background: Cylindrical lithium-ion batteries are tend to have an inward deformation during charge and discharge, which is called “core collapse”. This structural instability remains poorly quantified, and a deeper understanding of the underlying mechanics is essential for improving the structural integrity and safety of rechargeable cells. Objective: This study aims to investigate the deformation and buckling mechanisms of cylindrical lithium–ion battery jelly rolls under electrochemical cycling, with a focus on the interplay between geometry, interfacial constraint, and internal stress evolution. Methods: X-ray computed tomography (CT) imaging was employed to characterize the three-dimensional deformation of commercial cells, while a detailed finite element (FE) model explicitly incorporating current collectors, coating layers, and separators was developed to simulate chemo-induced stresses. The lithiation and delithiation processes were represented through an equivalent thermal-expansion surrogate to evaluate pre-buckling stress states. Results: CT analysis revealed negligible volumetric and hoop strains, confirming strong geometric constraint from the spiral architecture. Simulations showed that interface compatibility between current collectors and electrodes governs hoop-dominated stress fields, and that separator wrinkling can precede electrode buckling due to low bending stiffness and expansion mismatch. Conclusions: The results elucidate how hoop compression, interface constraint, and structural asymmetry collectively trigger jelly-roll instability. The findings establish a mechanistic foundation for analyzing confined expansion and guiding future multi-physics modeling of battery structural reliability.
- Open Access
- 12 February 2026
- Published by Royal Society of Chemistry (RSC) in EES Batteries
- Vol. 2 (2) , 683-699
Coupling X-ray computed tomography with digital volume correlation to study core collapse in lithium-ion batteries
X-ray computed tomography combined with digital volume correlation to study battery core collapse and feed numerical models.
- 12 February 2026
- Published by Royal Society of Chemistry (RSC) in EES Batteries
- Vol. 2 (2) , 683-699
Open AccessCoupling X-ray computed tomography with digital volume correlation to study core collapse in lithium-ion batteries
X-ray computed tomography combined with digital volume correlation to study battery core collapse and feed numerical models.
- Open Access
- 1 December 2025
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 172 (12) , 120509
Geometric Changes in Cylindrical Batteries as a Function of State of Charge
High-resolution, high-throughput CT scanning of commercial lithium-ion and sodium-ion cylindrical batteries provides quantitative detail into subtle geometrical changes occurring as a function of state-of-charge. Precise quantification via computer vision algorithms reveals a core area change with reversible and irreversible components in EVE lithium-ion cells, hysteresis in can eccentricity in Samsung lithium-ion cells, and the absence of significant geometrical changes in HAKADI sodium-ion cells. The difference in metrics between charge and discharge is introduced to remove the influence of tabs. The results demonstrate the potential of CT as a quasi-in-situ imaging technique for capturing sensitive geometric changes during cycling.
- 1 December 2025
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 172 (12) , 120509
Open AccessGeometric Changes in Cylindrical Batteries as a Function of State of Charge
High-resolution, high-throughput CT scanning of commercial lithium-ion and sodium-ion cylindrical batteries provides quantitative detail into subtle geometrical changes occurring as a function of state-of-charge. Precise quantification via computer vision algorithms reveals a core area change with reversible and irreversible components in EVE lithium-ion cells, hysteresis in can eccentricity in Samsung lithium-ion cells, and the absence of significant geometrical changes in HAKADI sodium-ion cells. The difference in metrics between charge and discharge is introduced to remove the influence of tabs. The results demonstrate the potential of CT as a quasi-in-situ imaging technique for capturing sensitive geometric changes during cycling.
- Open Access
- 1 November 2025
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 172 (11) , 110536
Design and Performance of the BYD LFP/Graphite 4680 Cylindrical Cell
The increasing demand for large-format lithium-ion batteries in electric vehicles has accelerated the development of cell formats that offer higher energy density and improved pack-level efficiency. Among these, the 4680 cylindrical cell format, pioneered by Tesla, has gained widespread adoption due to its potential to simplify battery pack architecture while enhancing gravimetric energy density. Other cell manufacturers, including SamSung, Panasonic, BYD, and others, have adopted the 46XX format, incorporating various chemistries. This study presents a comprehensive analysis of BYD’s 4680 lFP/graphite cylindrical lithium-ion cells. Detailed physical teardown, X-ray imaging, electrochemical testing, and electron imaging were performed to evaluate both cell- and materials-level attributes. The teardown and X-ray imaging reveal an uncommon multi-tab configuration and homogeneously double-sided electrode coatings, with no silicon additives in the graphite anode. Electrochemical testing indicates energy densities of 374.6 Wh l −1 and 150.5 Wh kg −1 . Furthermore, hybrid pulse power characterization and electrochemical impedance spectroscopy show impressive rate capability, with low cell-level DC area-specific resistance (6–17 Ω·cm 2 across various SOCs) and an area-specific charge transfer resistance of approximately 5.9 Ω·cm 2 . These findings highlight a new direction for the 4680-cell format in both design and performance, establishing a strong foundation for future developments in large-format cylindrical battery technology.
- 1 November 2025
- Published by The Electrochemical Society in Journal of the Electrochemical Society
- Vol. 172 (11) , 110536
Open AccessDesign and Performance of the BYD LFP/Graphite 4680 Cylindrical Cell
The increasing demand for large-format lithium-ion batteries in electric vehicles has accelerated the development of cell formats that offer higher energy density and improved pack-level efficiency. Among these, the 4680 cylindrical cell format, pioneered by Tesla, has gained widespread adoption due to its potential to simplify battery pack architecture while enhancing gravimetric energy density. Other cell manufacturers, including SamSung, Panasonic, BYD, and others, have adopted the 46XX format, incorporating various chemistries. This study presents a comprehensive analysis of BYD’s 4680 lFP/graphite cylindrical lithium-ion cells. Detailed physical teardown, X-ray imaging, electrochemical testing, and electron imaging were performed to evaluate both cell- and materials-level attributes. The teardown and X-ray imaging reveal an uncommon multi-tab configuration and homogeneously double-sided electrode coatings, with no silicon additives in the graphite anode. Electrochemical testing indicates energy densities of 374.6 Wh l −1 and 150.5 Wh kg −1 . Furthermore, hybrid pulse power characterization and electrochemical impedance spectroscopy show impressive rate capability, with low cell-level DC area-specific resistance (6–17 Ω·cm 2 across various SOCs) and an area-specific charge transfer resistance of approximately 5.9 Ω·cm 2 . These findings highlight a new direction for the 4680-cell format in both design and performance, establishing a strong foundation for future developments in large-format cylindrical battery technology.
- Open Access
- 21 October 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Minute differences in battery geometry explain electrochemical variability
Device variability in high-volume manufacturing is intrinsic in many technologies. Such is the case in lithium-ion battery cells, where cell-to-cell electrochemical variations impact cell-to-pack integration and battery management system. Here, we combine computer vision and 120-second X-ray computed tomography (CT) across 295 commercial 21700 cylindrical cells to explain and predict electrochemical performance variability, both at the beginning and end of life. Surprisingly, the cell electrochemical performances (varying by 0.51 % in capacity and energy, by 4.32 % in power, and by 10.6 % in cycle life) correlate strongly to a single descriptor: eccentricity of the cylindrical steel cans (absolute values variations <0.99 %). Through interpretable machine learning, these fast CT predict variations in energy and power throughout aging without any electrochemical cycling data. This work illustrates the utility of computer vision and explainable machine learning for revealing causal links between the internal structures of complex devices and their performance and aging, in the case of batteries, without needing electrochemical characteristics as input.
- 21 October 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessMinute differences in battery geometry explain electrochemical variability
Device variability in high-volume manufacturing is intrinsic in many technologies. Such is the case in lithium-ion battery cells, where cell-to-cell electrochemical variations impact cell-to-pack integration and battery management system. Here, we combine computer vision and 120-second X-ray computed tomography (CT) across 295 commercial 21700 cylindrical cells to explain and predict electrochemical performance variability, both at the beginning and end of life. Surprisingly, the cell electrochemical performances (varying by 0.51 % in capacity and energy, by 4.32 % in power, and by 10.6 % in cycle life) correlate strongly to a single descriptor: eccentricity of the cylindrical steel cans (absolute values variations <0.99 %). Through interpretable machine learning, these fast CT predict variations in energy and power throughout aging without any electrochemical cycling data. This work illustrates the utility of computer vision and explainable machine learning for revealing causal links between the internal structures of complex devices and their performance and aging, in the case of batteries, without needing electrochemical characteristics as input.