锂电老化机理和诊断分析全英文.pdf

www.inl.gov Diagnostic Testing and Analysis Toward Understanding Aging Mechanisms and Related Path Dependence Kevin L. Gering, PhD Energy Storage complemented by Equivalent Circuit Models (ECMs). INL Essential Modeling Tools Key Elements are Patent Pending Life Data +Mech. Models =Early Diagnosis and Predictive Analyses Cell 80 Cell 14 Time, weeks 020406080100120140160 Conductance Fade, % 0 10 20 30 40 50 60 70 Data with 0.005 Ohm offset Data with No offset IR Ohmic drop, EIS data Legend: A: IR Ohmic term B: SEI, chem/electronic C: SEI, phys/structure D: Rct,o (intrinsic) Total Net SEI Contribution (B+C) A B D C Modeling Cell Conductance Results from two-model synergy (MSM + -BV Kinetics) Key insights into cell operation and rate-based mathematics allows accurate modeling and high-fidelity diagnostic analysis of conductance behavior in electrochemical cells. Cell conductance has a principal influence on attainable power, decreasing over the life of a cell. Based on data for EIS semicircle RHS edge, Gen2 cells cycle-life tested at 25 C. INL Kinetics Modeling Based on an improved form of Butler-Volmer expression that is well-suited for Li-ion systems. Model gives extremely accurate predictions over (T, I, tpulse , cell aging) when coupled with an advanced set of rate expressions. Current, A 0.0010.010.1110 Total Resistance, Ohm 0.0 0.5 1.0 1.5 2.0 2.5 3.0 after 1 second 2 seconds 3 seconds 4 seconds 5 seconds 6 seconds 7 seconds 8 seconds 9 seconds 10 seconds 0.5 second 0.5 s 10 s Solid Curves: Model Predictions Cell 80 (20% Power Fade) Rtotalat 30 oC along pulse timeline Approximate trend line for io Current, A 0.0010.010.1110 Total Resistance, Ohm 0.0 0.5 1.0 1.5 2.0 2.5 3.0 after 1 second 2 seconds 3 seconds 4 seconds 5 seconds 6 seconds 7 seconds 8 seconds 9 seconds 10 seconds 0.5 second 0.5 s 10 s Solid Curves: Model Predictions Cell 80 (20% Power Fade) Rtotalat 30 oC along pulse timeline Approximate trend line for io Cell 80 (20% Power Fade) Rtotalat 30 oC along pulse timeline Approximate trend line for io Current, A 0.0010.010.1110 Total Resistance, Ohm 0.01 0.1 1 10 oC -40 -30 -20 -10 0 20 30 10 Symbols: Test Data Solid Curves: Model Rtotalat 10 s over temperature Current, A 0.0010.010.1110 Total Resistance, Ohm 0.01 0.1 1 10 oC -40 -30 -20 -10 0 20 30 10 Symbols: Test Data Solid Curves: Model Rtotalat 10 s over temperature 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 2 4 6 8 0.01 0.1 1 10 Total Resistance, Ohm Time During Pulse, s Current, A Cell 80 at 30 oC 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 2 4 6 8 0.01 0.1 1 10 Total Resistance, Ohm Time During Pulse, s Current, A Cell 80 at 30 oC .and many other performance quantities. Time, Weeks 050100150200250 C1/1 Capacity Fade Term, % 0 10 20 30 40 50 Data Overall Model Results Related to Li+ Related to Active Sites Offsets due to C/1 rate limitations Test data covers 140 weeks Time, Weeks 050100150200250 C1/1 Capacity Fade Term, % 0 10 20 30 40 50 Data Overall Model Results Irreversible total Irreversible, Li+ Irreversible, Active Sites Reversible total Reversible, Li+ Reversible, Active Sites The rate of lithium consumption is high during initial time (continued SEI formation; side reactions), but tapers off considerably by 30 weeks. Reversible contributions to fade dominate at early time, are tied primarily to active sites, and undergo a maximum at around 150 weeks. In comparison, irreversible losses grow steadily over the time period. Capacity fade is dominated by mechanisms that impact active sites, initially by reversible mechanisms through about 180 weeks, then by irreversible mechanisms thereafter. Under these test conditions the theoretical limit of capacity loss is effectively met by about 200 weeks. Modeling Capacity Loss Over Life (Gen2 Li-ion Cells under cycL Testing at 25 C) Excellent fit with Wfinite, which captures interfacial transport ROhmicRct anode Cdl anode Rct cathode Cdl cathode finiteinfinite WW Equivalent Circuit Modeling 00.10.20.30.4 -0.3 -0.2 -0.1 0 0.1 Z Z Cell #80, -10 0C, experiment Fit Result 10-310-210-1100101102103104105 10-2 10-1 100 Frequency (Hz) |Z| Cell #80, -10 oC, experiment Fit Result 10-310-210-1100101102103104105 -50 0 50 100 Frequency (Hz) theta Nyquist PlotBode Plot Collaborations Hawaii Natural Energy Institute. Involved in diagnostic analysis of cell performance data to determine path dependence effects related to aging conditions tied to PHEV test protocol. HNEI work is coordinated by Prof. Bor Yann Liaw. University of California at Pomona. This work applies an advanced form of density functional theory (DFT) and rigorous treatment of electrolyte properties to determine how the SEI structure and chemistry affects local electrochemical behavior and efficiency and local thermodynamic behavior of electrolyte species. Findings of this work will allow greater diagnostic analysis of interfacial limitations in Li-ion cells. UC-Pomona work is coordinated by an acknowledged expert in statistical thermodynamics, Prof. Lloyd Lee. Argonne National Lab. Provides oversight and coordination on key issues regarding the ABRT program. Battery testing and modeling tasks are complementary between INL and ANL. DT is progressing with Sanyo Y cells, looking at various issues of cell performance and path dependence of aging. FY 2010 will be a pivotal year in continuing/completing PHEV-relevant cell testing and diagnostic analysis. Early results are useful for assessing BOL trends and initial estimates of parameters for aging models. More extensive data over time is needed to surmise probable degradation mechanisms regarding capacity, impedance, etc. INL and HNEI have developed key computational tools used to model, diagnose, and predict performance and aging of electrochemical cells. These tools are targeting mechanisms of cell degradation, related path dependence, and chief causes and conditions of performance loss. Thermal cycling should be considered as a standard aging condition for batteries intended for vehicle applications (HEV, PHEV, EV), and could be useful as an accelerated aging condition. The immediate benefits of this work will be (1) to provide more realistic and accurate life predictions by accounting for the influence of thermal cycling effects and related path dependence on aging mechanisms, and (2) provide a basis for improving battery development and management. Summary / Conclusions Future Work We will continue to monitor aging trends for our path dependence studies over the next several months. Mechanistic analyses and modeling of mature data sets will be performed at the completion of this work to determine the extent of path dependence of cell aging, wherein existing INL and HNEI modeling tools will be applied. Demonstrate INL diagnostic/predictive modeling capabilities through software that integrates key modules regarding performance over life. We will quantify the impact of thermal cycling on Sanyo Y cell aging. Future path dependence studies could involve other duty-cycles (e.g., FUDS, DST), other temperature parameters defined for a particular city or region, and other Li-ion cell chemistries. Pending cell availability, we will perform DT on ABRT Gen4 cells to elucidate path dependence of aging for that cell chemistry. Acknowledgements DOE Vehicle Technologies Program David Howell, DOE-EERE, VTP Jeff Belt, INL David Jamison, INL Christopher Michelbacher, INL Tim Murphy, INL Sergiy Sazhin, INL Mikael Cugnet, HNEI Matthieu Dubarry, HNEI Bor Yann Liaw, HNEI Lloyd L. Lee, UC-Pomona This work was performed for the United States Department of Energy under contract DE-AC07-05ID14517. HNEI