Key Characteristics for Thermal Runaway of Li-ion Batteries

Key Characteristics for Thermal Runaway of Li-ion Batteries

Available online at www.sciencedirect.com Available online at www.sciencedirect.com ScienceDirect ScienceDirect Energy Procedia 00 (2018) 000–000 Av...

1MB Sizes 0 Downloads 2 Views

Available online at www.sciencedirect.com Available online at www.sciencedirect.com

ScienceDirect ScienceDirect

Energy Procedia 00 (2018) 000–000 Available online www.sciencedirect.com Available online atatwww.sciencedirect.com Energy Procedia 00 (2018) 000–000

ScienceDirect ScienceDirect

www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia

Energy Procedia 158 Energy Procedia 00(2019) (2017)4684–4689 000–000 www.elsevier.com/locate/procedia

10th th

International Conference on Applied Energy (ICAE2018), 22-25 August 2018, Hong Kong, 10 International Conference on Applied Energy China(ICAE2018), 22-25 August 2018, Hong Kong, China

Key Characteristics for Thermal Runaway of Li-ion Batteries The 15th International Symposium Runaway on District Heating and Cooling Key Characteristics for Thermal of Li-ion Batteries

Xuning Fengaa, Siqi Zhengaa, Dongsheng Renbb, Xiangming Hea,a,*, Li Wangaa, Xiang Liub,c , b,c XuningAssessing Feng , Siqi Zheng ,Maogang Dongsheng Ren , Xiangming He *, Li Wang , Xiang Liu , dof b, demand-outdoor the feasibility using the heat Li , Minggao Ouyang ** Maogang Lid, Minggao Ouyangb,**

temperature function a long-term district heat100084, demand forecast Institute of Nuclear andfor New Energy Technology, Tsinghua University, Beijing China a

a Institute of Nuclearofand New Energy Technology, Tsinghua University, Beijing 100084, China State Key Laboratory Automotive Safety and Energy, Tsinghua University, Beijing 100084, China cb State KeySciences Laboratory Automotive Safety aand Energy,National Tsinghua University,Argonne, Beijing 100084, Chemical andof Engineering Division, Argonne IL 60439, USA a,b,c a bLaboratory, c China c c d Chemical Sciences and Engineering Argonne National Laboratory, China Office, ThermalDivision, Hazard Technology, Shanghai, 200029, Argonne, PR ChinaIL 60439, USA d China Office, Thermal Hazard Technology, Shanghai, 200029, PR China a IN+ Center for Innovation, Technology and Policy Research - Instituto Superior Técnico, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal b Veolia Recherche & Innovation, 291 Avenue Dreyfous Daniel, 78520 Limay, France Abstract c Département Systèmes Énergétiques et Environnement - IMT Atlantique, 4 rue Alfred Kastler, 44300 Nantes, France b

I. Andrić

*, A. Pina , P. Ferrão , J. Fournier ., B. Lacarrière , O. Le Corre

Abstract The lithium ion batteries are having increasing energy densities, meeting the requirement from industry, especially for the electric The lithium ion batteries increasing energyisdensities, meeting the requirement from industry, especially for the electric vehicles. However, a cell are withhaving a higher energy density more prone to thermal runaway. We analyze the key characteristics during vehicles. However,toahelp cell with more prone to thermal runaway.temperatures We analyze are the key characteristics during thermal runaway bettera higher define energy batterydensity thermalis runaway. Three characteristic regarded as the common Abstract thermal to help betterfor define battery ThreeThe characteristic the common features runaway of thermal runaway all kinds of thermal lithium runaway. ion batteries. underlyingtemperatures mechanismsare forregarded the threeas characteristic features thermal runaway allbykinds of analysis. lithiuminion underlying mechanisms three characteristic temperatures havenetworks been investigated thermal The conclusion the analysis set benchmarks forthe evaluating the thermal District of heating arefor commonly addressed the batteries. literature The asofone of the most effective for solutions for decreasing the temperatures haveemissions been investigated by thermal analysis. The conclusion of the analysis set benchmarks for evaluating the thermal runaway behaviors of commercial lithium-ion batteries, and the proposed methodologies benefits further and development greenhouse gas from the building sector. These systems require high investments which areresearch returned through the heat runaway behaviors ofchanged commercial lithium-ion batteries, and the proposed methodologies benefits further research development of battery safety design for electric vehicles. sales. Due to the climate conditions and building renovation policies, heat demand in the futureand could decrease, ofprolonging battery safety design for electric vehicles. the investment return period. Copyright 2018ofElsevier Ltd.isAll rights reserved. The main©scope this paper to assess the feasibility of using the heat demand – outdoor temperature function for heat demand © 2019 The Authors. Published Elsevier Ltd. Copyright © 2018 Elsevier Ltd. by All rights reserved. Selection and peer-review under responsibility the scientific committee of theas10ath case International Conference Applied Energy forecast. The district of Alvalade, located inofLisbon (Portugal), was used study. The district on is consisted of 665 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) th International Conference on Applied Energy Selection and peer-review under responsibility of the scientific committee of the 10 (ICAE2018). buildings that vary in both construction period and typology. Three weather scenarios (low, medium, high) and three district Peer-review under responsibility of the scientific committee of ICAE2018 – The 10th International Conference on Applied Energy. (ICAE2018). renovation scenarios were developed (shallow, intermediate, deep). To estimate the error, obtained heat demand values were Keywords: ion battery; thermal runaway; thermal analysis; internal short circuit comparedLithium with results fromsafety; a dynamic heat demand model, previously developed and validated by the authors. Keywords: Lithium ion battery; safety; thermal runaway; thermal analysis; internal short circuit The results showed that when only weather change is considered, the margin of error could be acceptable for some applications (the error in annual demand was lower than 20% for all weather scenarios considered). However, after introducing renovation 1.scenarios, Introduction the error value increased up to 59.5% (depending on the weather and renovation scenarios combination considered). 1.The Introduction value of slope coefficient increased on average within the range of 3.8% up to 8% per decade, that corresponds to the decrease in thefield number of heating hours of 22-139h duringlead the heating season (depending the combination of weather Why some accidents involving Li-ion batteries to thermal runaway (TR)on problem while others do notand is Why some field accidents involving Li-ion batteries lead to thermal runaway (TR) problem while others do not is renovation scenarios considered). On the other hand, function intercept increased for 7.8-12.7% per decade (depending on the still unclear [1]. There is little agreement among world experts when proposing a TR test procedure for the Electric coupled scenarios). TheTechnical could used to modify when the parameters forof the scenarios considered, and still unclear [1]. There isvalues little suggested agreement among world experts proposing a TR test procedure for the Electric Vehicle Safety-Global Regulation, tobeobtain consensus onfunction accurately definition TR [2]. The innumerable improve the accuracy of heat demand estimations. Vehicle Safety-Global Technical Regulation, to obtain consensus on accurately definition of TR [2]. The innumerable cell chemistries should be “blamed” for causing the situation. The lack of standardization among cell chemistries

cell chemistries should be “blamed” for causing the situation. The lack of standardization among cell chemistries © 2017 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the Scientific Committee of The 15th International Symposium on District Heating and Cooling. * Corresponding author. Tel.: +86-10-62794226; fax: +86-10-62794226. * E-mail Corresponding Tel.: +86-10-62794226; fax: +86-10-62794226. address:author. [email protected]; [email protected] Keywords: Heat demand; Forecast; Climate [email protected] change E-mail address: [email protected];

1876-6102 Copyright © 2018 Elsevier Ltd. All rights reserved. 1876-6102 Copyright © 2018 Elsevier Ltd. All of rights reserved. committee of the 10th International Conference on Applied Energy (ICAE2018). Selection and peer-review under responsibility the scientific Selection and peer-review under responsibility of the scientific committee of the 10th International Conference on Applied Energy (ICAE2018). 1876-6102 © 2017 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the Scientific Committee of The 15th International Symposium on District Heating and Cooling. 1876-6102 © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of ICAE2018 – The 10th International Conference on Applied Energy. 10.1016/j.egypro.2019.01.736

2

Xuning Feng et al. / Energy Procedia 158 (2019) 4684–4689 Author name / Energy Procedia 00 (2018) 000–000

4685

further complicates the situation. Particle morphology, doping variations and coating strategy etc., all have significant influence on the thermal stability of cell materials [3]. Moreover, the cell format also varies, including pouch, prismatic and cylindrical. Before reaching consensus on the definition of TR, it is pertinent to consider a few common features from extensive experimental observations. Another parameter to consider is the internal-short-circuit (ISC) resistance, which usually accompanies the occurrence of TR [4]. When modelling the TR behavior using joule heating equations, the total amount and the release rate of the heat generated by ISC always does not match the experimental data. A commonly employed compromise is to tune the coefficients in the model to eliminate the error in the reaction dynamics [5]. We have collected massive data since we made progress in the technique on testing the thermal runaway features of large format lithium ion batteries [6]. We have obtained new insights from hundreds of adiabatic thermal runaway results. We present this paper as an overview of a series of scientific papers planned on this topic. 2. Experimental 2.1. Calorimetry We are equipped with three kinds of accelerating rate calorimetries (ARC) manufactured by Thermal Hazard Technology (THT), including the es-ARC, EV-ARC, and ARC-EV+. Thermal runaway tests were conducted under the heat-wait-seek-exotherm mode. The ARC builds an adiabatic boundary condition around the cell during the exotherm mode, ensuring accurate measurement of the heat generation during TR. The ARC test is usually used to evaluate the thermal runaway features of a full cell. The voltage and the impedance (1kHz) are also measured during ARC test. We also have a differential scanning calorimetry (DSC), manufactured by Netzsch®, including the STA 409PC/4/H Luxx, DSC 214 Polyma. Kinetics at the cell component level are evaluated by setting proper temperature scanning rates. The DSC test is usually used to evaluate the thermal properties of the cell and its chemistries. The DSC result helps reveal the mechanisms behind the ARC curves. The samples are components scratched from commercial cells, and the mass ratio for the powder of “cathode:electrolyte:anode” is “3.6mg:2μL:6.3mg”, same as that in the porous electrode. The DSC scans the samples from 50oC to 600oC with a scanning rate of 20oC·min-1. 2.2. The battery thermal runaway database

Fig. 1 Some tested samples of commercial lithium-ion batteries

Varies kinds of samples (Fig. 1), with capacities ranging from 1Ah to 50Ah, come from the world’s leading lithiumion battery manufacturers. We also collect qualified data from literature [8]. Various kinds of cell chemistries have been tested: for cathodes we consider LiCoO2(LCO), LiFePO4(LFP), LiMn2O4(LMO), and LiNixCoyMnzO2(NCM) etc.; for anodes we consider graphite, mesocarbon-microbead(MCMB), hard carbon, and Li4Ti5O12(LTO) etc.; for separators, we study polyethylene(PE), polypropylene(PP), PE|PP|PE multilayer, and polyimide(PI) based membranes, with or without a ceramic coating. The number and diversity of the cell studied persuades our conclusions that the common features identified from the database are reliable.

Xuning Feng et al. / Energy Procedia 158 (2019) 4684–4689 Author name / Energy Procedia 00 (2018) 000–000

4686

3

3. Results and discussion 3.1. The characteristic temperatures of battery thermal runaway We identify three characteristic temperatures {T1, T2, T3} for each case as shown in Fig. 2 from a database comprised of hundreds of ARC test results. The definitions of {T1, T2, T3} are discussed in this section. We can “control” the battery safety response of the cell as long as we understand the factors that contribute to the origin of {T1, T2, T3}. T1 is the onset temperature for the detectable self-heat generation; the mechanism for this heat source is usually attributed to SEI decomposition. T2 is the trigger temperature for TR. Understanding T2 is quite critical for safe battery design, because a higher T2 usually means better overall thermal stability, and the battery can thus be more likely to pass standard abuse test. However, there is still no quantifiable definition of T2. The magnitude of dT·dt-1 might quoted as a metric, but this value varies for different cell chemistries. More importantly, the critical reaction that is responsible for releasing a large amount of heat instantaneously, remains unclear. T3 is the maximum temperature that the battery can reach during a TR. The difference ∆T=T3-T2 is directly linked to the total heat released during a TR (∆HTR) as shown in Eqn. (1), where M is the cell mass and Cp is the heat capacity. The TR propagation response largely depends on T2 and T3, the accurate measurement of which requires specific techniques especially for large format samples. ∆HTR=M·Cp·∆T= M·Cp·(T3-T2) (1)

Fig. 2 Three characteristic temperatures {T1, T2, T3} in an ARC test, and their unsolved mechanisms.

3.2. Statistics on the ARC test results

Fig. 3 Statistics for the ARC test results for 33 battery samples with different cell chemistries. (a) T3 vs T2. (b) max{dT·dt-1} vs energy density.

Statistics from the ARC test data reflect patterns that may help quantitatively describe a TR. T3 tends to be higher for cells with higher energy density (Fig. 3a). The maximum dT·dt-1 that is reported around T2 has a high correlation with the energy density of the battery (Fig. 3b). Therefore setting different thresholds of dT·dt-1 according to the energy density of the battery might be appropriate for the online monitoring of TR. The second order derivatives of

4

Xuning Feng et al. / Energy Procedia 158 (2019) 4684–4689 Author name / Energy Procedia 00 (2018) 000–000

4687

the temperature rise will be further evaluated, in order to confirm that the level rise in dT·dt-1 is a reliable metric for the quantitative definition of battery TR. 3.3. Interpreting the thermal runaway mechanisms by reaction pathway To trace the origins of T2 and T3, we try to comb the complex “Reaction Pathway” during a TR, as shown in Fig. 4. Guided by the notion of First Principle, every component within the cell, as shown in Fig. 4(a), has a unique failure behavior at extreme temperatures that we call a “Single reaction” (Fig. 4b). Some of the products, like O2 and PF5-, are highly reactive to trigger further secondary reactions. Calorimetric tests can be designed easily to track the progress of these single and secondary reactions. However, the data becomes more complex when we inspect the “Mixed reactions” for the system of porous electrodes. The active material in the cathode/anode particles interact with components in the electrolyte, once the SEI is lost. There is little interaction between the cathode and the anode side, if the separator is intact. Fig. 4(c) contains the heat generated from the cathode/anode interacting with the electrolyte measured by DSC. Separate kinetic equations for the cathode/anode reactions have been used to model the TR behavior [7]. However, this methodology fails to establish a model to fit the TR data for the full cell shown in Fig. 2, the predicted T3 will be much lower than the measured data, due to the mismatch in the mechanism of heat generation. Further consideration that the separator collapses at extreme temperature bifurcates the analysis, according to the occurrence of an ISC, as shown in Fig. 4(d). Case A is that electrolyte can still transform Li-ions, and an electrical circuit with low resistance can be formed between the cathode and electrode. Case B is that no electrolyte exists due to evaporation, and redox reactions occur between the oxide cathode and reductive anode. Interestingly, the heat flow measured by DSC for Case B is quite close to the heat release behavior between T2 and T3.

Fig. 4 . Untangling the “Reaction Pathways” according to our current understanding. (a) A 3D perspective of a lithium ion battery, with NCM+LMO | 1M LiPF6 1:1:1 EC:DMC:DEC | Graphite cell chemistry. (b) Single reactions for the cell chemistries, and secondary reactions caused by intermediate products generated from single reactions. (c) Typical mixed reactions that may cause the sharp temperature rise from T2 (triggering) to T3 (maximum), cases categorized by the integrity of separator.

DSC data of cell materials (cathode, anode, electrolyte, separator, etc.) and their combinations are used to fit the ARC data of the full cell in Fig. 2, as shown in Fig. 5. The good fit of the DSC data and ARC data indicates that the reaction pathways we guess in Fig. 5 might be correct. Based on all the experiments conducted, we believe that this should be the best guess ever. According to the collapse of separator (at T2=192.4℃) and the redox reaction between the cathode and anode (249.2℃), the TR process is divided into three stages. Stage I: At low temperatures (
4688

Xuning Feng et al. / Energy Procedia 158 (2019) 4684–4689 Author name / Energy Procedia 00 (2018) 000–000

5

Nevertheless, the heat flow of “cathode+electrolyte” is low at this stage and should be omitted. The “anode+electrolyte” system generates approximately 17% of the total heat released during TR in Stage I. Stage II: Once the separator collapse at T2, some of the electrolyte still exist inside the cell (the boiling point for EC is 248℃), while others are evaporated. Internal short circuit (ISC) occurs in Stage II releasing electric energy stored inside the battery intensively. But the ISC cannot release all the electric energy inside the battery, and will cease once the temperature rises to 249.2℃, when most of the electrolyte gasifies, thereby cutting off the circuit for internal short. Only 9% of the total heat released during TR is caused by ISC. Stage III: The redox reaction between the cathode and anode starts and release large amount of heat, 74% in this case. Obviously, the T2 is determined by the collapse temperature of the separator, whereas the T3 is determined by the redox reaction between the cathode and anode.

Fig. 5 The thermal runaway mechanisms underlies the characteristic temperatures.

3.4. The role of internal short circuit during thermal runaway Further comparative tests are conducted to verify that the heat release by ISC has little influence on the formation of T3, in other words the ISC contribute little to the sharp temperature rise from T2 to T3. Fig. 6(a) shows the details of ARC test results in Fig. 1 with impedance (R) at 1kHz measured. R rises from approximately 120℃, which indicates the vaporization of solvent components (the boiling point of DMC is 91℃, EMC 110℃, DEC 126℃). R drops at T2 indicating the occurrence of ISC. Fig. 6(b) shows an ARC test results for a battery sample with no separator and electrolyte. The sample is prepared by 1) vacuuming most of the solvents within the battery, 2) removing the separator between the electrodes. Therefore there is always ISC between the cathode and anode, the heat generation of which advances the T1 from 78.2℃ to 57.6℃. However, as there is little electrolyte inside, the R of the sample is large, and the heat generation of ISC is limited. As there is no separator left inside, T2 is determined by the redox reaction between the cathode and anode. T2=229.4℃, quite near to the triggering temperature, 249.2℃, of the redox reaction. T3=877.2℃

6

Xuning Feng et al. / Energy Procedia 158 (2019) 4684–4689 Author name / Energy Procedia 00 (2018) 000–000

4689

is still high, convincing us that ISC does not make major contributions in the total heat generation during TR. The higher T3 can be explained because the mass (M) of the battery is reduced by removing separator and electrolyte from the cell. The mass is reduced by approximately 13%, therefore the 18% of temperature rise for T3-T2 is reasonable.

Fig. 6 Test results interpreting the mechanism for the sharp rise at T2 during TR. (a) ARC test results for a commercial 20Ah NCM+LMO | Graphite cell, with separator and electrolyte. (b) ARC test results for the cell, after the separator is removed and electrolyte vacuumed.

Summary Three characteristic temperatures {T1, T2, T3} are regarded as the key common features of battery TR, and will benefit further research investigating the specific TR mechanisms of all kinds of cell chemistries. The formation of T2 is determined by the collapse of the separator, whereas that of T3 is caused by the redox reactions between the cathode and anode. The ISC occupies very few of the total heat generation during TR, though it helps trigger the redox reactions after the separator collapse. We have briefly summarized our recent progress in investigating the TR mechanisms in a lithium ion battery based on our thermal analysis database. Criticism and suggestions are both welcomed for improving the database. It would be beneficial if research groups can share their ARC data to build a “Genome database” for battery safety together. We wish that the proposed methodologies: 1) the graphical illustration of “Reaction Pathway”; and 2) the “Energy Release Diagram” proposed in Ref. [8], can benefit the research on battery TR mechanisms, and help make the lithium ion batteries safer in application. Acknowledgements This work is supported by the National Natural Science Foundation of China (No. 51706117 and U1564205); the China Postdoctoral Science Foundation funded project (No. 2017M610086). The first author thanks the support of “Young Elite Scientist Sponsorship Program” from China Association for Science and Technology. References [1] Doughty DH, Pesaran AA. Vehicle battery safety roadmap guidance. National Renewable Energy Laboratory, 2012, pp24. [2] Report No. 1406-601, in 14th EVS-GTR session, Ottawa, Sept. 25th to 27th, 2017. https://wiki.unece.org/display/trans/EVS+14th++session (Retrieved at May 13, 2018). [3] Schipper F, Erickson EM, Erk C, et al. Review-Recent advances and remaining challenges for lithium ion battery cathodes. J. Electrochem. Soc. 2017;164(1):A6220-A6228. [4] Feng X, Weng C, Ouyang M, et al. Online internal short circuit detection for a large format lithium ion battery. Appl. Energ. 2016;161:168180. [5] Coman PT, Darcy EC, Veje CT, et al. Numercial analysis of heat propagation in a battery pack using a novel technology for triggering thermal runaway. Appl. Energ. 2017;203:189-200. [6] Feng X, Fang M, He X, et al. Thermal runaway features of large format prismatic lithium ion battery using extended volume accelerating rate calorimetry. J. Power Sources. 2014;255:294-301. [7] MacNeil DD, Larcher D, Dahn JR. Comparison of the reactivity of various carbon electrode materials with electrolyte at elevated temperature. J. Electrochem. Soc. 1999;146(10):3596-3602. [8] Feng X, Ouyang M, Liu X, et al. Thermal runaway mechanism of lithium ion battery for electric vehicles: A review. Energy Storage Materials. 2018;10:246-267.