1. Task Force 3: Electrolyte leakage Last update- 14/10/2014 1

2.  Organization of Task force  Issues raised during China EVS meeting  Present requirement and purpose  Approach of Task Force  Definitions  Aqueous Electrolytes  Issues  Proposal  Non-Aqueous Electrolytes  Issues  Proposal  Conclusion and future work Content: 2

3. Scope: Issues related to electrolyte leakage (except venting) Objective : Reply all the queries on this issue, provide justification for the requirements in GTR draft and if required propose test procedure. Organization of Task force: Scope & Objective 3

4. China (CATARC) European Union (JRC) USA (NHTSA) South Korea ( KATRI) France (UTAC) OICA (Alliance, Daimler, JLR, PSA, GM, Toyota, Nissan, JAMA, VW, Scania, SK Continental, Volvo, Renault) Organization of Task force: Members 4 Contributions Research data /analysis: JRC Comments: JRC, NHTSA, KATRI, CATARC, OICA

5. End of January 2014: Initial comments on the issue by exchange of emails 28 th February 2014: 1st meeting with co-sponsors on progress of task force 11 th March 2014: 1st audio meeting of TF to discuss the initial proposals 21 st March 2014: 1st face to face meeting in Paris 11 th April 2014: 2nd audio meeting of TF to finalize the task force conclusion on aqueous electrolyte REESS 25 th April 2014: 2 nd meeting with co-sponsors on progress of task force 8th of May 2014: Comments on the ‘status report’ by exchange of emails 12 th May 2014: 2 nd face to face meeting in Washington (validation of conclusion) 13 th May 2014: Presentation of task force progress report in the 5 th GTR-EVS meeting 15 th October 2014: Face to face meeting in Brussels 19 th November 2014: Presentation of task force progress report in the 6 th GTR-EVS meeting 5 Organization of Task force: Work plan

6. Electrolyte Leakage: Issues raised during last EVS meeting 6  Three category of questions 1.Leakage detection: How to distinguish leakage? What is an appropriate coating? 2.Leakage (spillage) amount measurement : How is leakage measured? How to quantify electrolyte leakage amount (7 % volume or 5 litters)? How to differentiate ‘electrolyte leakage’ from ‘coolant’? How to measure the electrolyte vapor (in case required to)? Electrolyte leakage currently defined as liquid leakage. This poses possible difficulty for batteries using volatile electrolytes (e.g. Li-ion). How is liquid electrolyte leakage measured and distinguished from electrolyte lost due to vapors or evaporation of spilled electrolyte? 3.Venting gas: How to differentiate smoke from combustion with electrolyte vapors at venting, for example in the thermal cycling test?

7. Acceptance criteria in present GTR draft : 7 Test itemObservation time LeakageSpillageRuptureFireExplosionIsolation Resistance Retention Vibration1h√-√√√√- Thermal choc1h√-√√√√- Fire resistanceUp to 3 h----√-- External short circuit √-√√√√- Over discharge√-√√√√- Over charge√-√√√√- Mechanical shock 1h√√√√√√√ Mechanical integrity 1h√√√√√√√ Post-crash Vehicle 30 min√√√√√√√ The ‘flammability aspect’ and the ‘electric choc aspect’ of electrolyte is already covered in the GTR draft The fire test does not require leakage /spillage criteria and hence out of scope of this TF Hence the objective of the task force is to look in to the chemical risk (corrosive & toxic nature of electrolytes)

8. Test itemsPurpose of the testPresent Requirements Vibration-The user is supposed to continue to use the vehicle after the event. -In this case, stringent requirements should be applied -No evidence of electrolyte leakage Thermal shock and cycling External short circuit protection -The proposed test procedure is to confirm the operation of protective function. -In this case, stringent requirements should be applied Overcharge protection Over-discharge protection Over-temperature protection Mechanical integrity-Same as vehicle post-crash-No evidence of electrolyte leakage Mechanical shock REESS requirements for whole vehicle post-crash -The user is supposed to stop using the vehicle until certain repair/maintenance is conducted once subject to the event, presuming the battery would not be re- used for any other purpose than vehicle propulsion. -In this case, the requirement relevant to the accident situation, in order to avoid additional risk to the occupants and the surrounding people, should be applied. -Until 30 min after the impact, there shall be no electrolyte leakage from the REESS into the passenger compartment -no more than 7 % by volume of the REESS electrolyte capacity spilled from the REESS to the outside of the passenger compartment. Electrolyte Leakage: Present requirement & Purpose In-use Post-crash

9. Prepare a list of potential risks associated with existing electrolytes => discussion completed TF members agreed to distinguish the REESS in to two categories based on the types of electrolytes Aqueous electrolyte Non-aqueous electrolyte TF member agreed to distinguish in-use and post-crash requirements The discussions will be in two steps: first complete the discussion on REESS based on aqueous electrolytes (by end March) => discussion completed and then discuss the particularities of REESS with non-aqueous electrolytes. => still under discussion Approach : 9

10. Aqueous electrolyte: An aqueous electrolyte is an electrolyte based on water solvent for the compounds (e.g. acids, bases) providing conducting ions after its dissociation. => agreed ? Non-aqueous electrolyte: A non-aqueous electrolyte is an electrolyte based on organic solvent for the compounds (e.g. salts) providing conducting ions after its dissociation. => agreed ? Leakage: refer JRC comments in slide 18, should we distinguish liquid leakage form possible vapor/gas leakage ? Definitions : 10 These definitions are agreed ? To be discussed

11. IssuePotential RiskProposed Solution 1Leakage/spillage in liquid form Irritant, Corrosive Large amounts, order of liters, expected No evidence of electrolyte leakage/spillage Well understood and documented, e.g. FMVSS 305, UNECE R100.02, UNECE R94/95/12, IEEE 1578 Visual inspection as described in the current GTR text may be used for electrolyte leakage/spillage detection 2Vapor from leakage/spillage No significant amount expected No action needed (low volatility of water) 3Volatile gas Expected in normal operation Flammable gas (e.g.H 2 ) Well understood and documented (e.g. UNECER100.02, EN62485-3, SAEJ1718) Venting (proposal from Japan) 4Leakage/spillage in liquid form Irritant, Corrosive Large amounts, order of liters, expected Until 30 min after the impact, there shall be no electrolyte leakage from the REESS into the passenger compartment No more than 7 % by volume of the REESS electrolyte capacity spilled from the REESS to the outside of the passenger compartment. Visual inspection may be used for electrolyte leakage/spillage detection 5Vapor from leakage/spillage No significant amount expected No action needed (low volatility of water) Aqueous Electrolytes :Issues In-use Post-crash TF3 experts agrees that proposed solution is adequate for issues related to Aqueous Electrolytes

12. Add the following clarification on the leakage detection: [Based on FMVSS 305] Good engineering judgment should be used to fulfil the requirement of an appropriate coating. One of the possible solutions might be an absorbent paper which surrounds the REESS casing. If in this case an electrolyte leakage occurs, the absorbent paper would get stains and wetted. The electrolyte leakage would be simple to recognize by visible inspection. Add the following clarification on spillage/leakage measurement: [Based on FMVSS 305] The spilled amount of electrolyte can be measured by usual techniques of determination of liquid volumes after collecting the spillage. For containers containing both Stoddard (colored coolant) and electrolyte, the fluids shall be allowed to separate by specific gravity then measured. Aqueous Electrolytes : Proposal Final text will be proposed in the next GTR-EVS meeting

13. IssuePotential RiskProposed Solution 1Leakage/spillage in liquid form Flammable, Toxic, Corrosive Small amounts, order of millilitres, expected No evidence of electrolyte leakage/spillage 2Vapor from leakage/spillage Toxic / flammable gas No vapors are expected as no electrolyte leakage/spillage is allowed [JRC= > Detection of vapors is needed to ensure the requirement of no leakage/spillage of electrolyte is fulfilled => need discussion ] 3Volatile gas Not expected in normal operation Venting (proposal from Japan) 4Leakage/spillage in liquid form Flammable, Toxic, Corrosive Small amounts, order of millilitres, expected For REESS based on non-aqueous electrolytes, there should not be any spillage/leakage ‘outside vehicle’ or ‘inside passenger’ compartment Visual inspection may be used for electrolyte leakage/spillage detection 5Vapor from leakage/spillage Toxic / flammable gas Under discussion Non-Aqueous Electrolytes : Issues In-use Post-crash NEW 1 Justification: 1. JRC analysis shows that 7% criteria for li-ion battery may lead to dangourouse situation. Spilling ca. 1 L of dimethyl carbonate results in a PAC-3 concentration level in a volume of vehicle +3 m-thick layer around it. To be discussed

14. Non-Aqueous Electrolytes : Vapor from leakage/spillage Back ground: JRC: For non-aqueous electrolytes the amount of vapors generated are not negligible given the high volatility of their components. They may come not only from a liquid spillage but from a damaged cell (vapors will diffuse out of casing). Risk analysis : Theoretical analysis : refer JRC comments to OICA proposal (slide 18,19,20) In-field data: At this stage no ‘in field data’ available on hazards due to toxic gas released during post crash event (with out spillage/leakage) Experimental data : is there any experimental data available simulating the post crash situation of vehicle ??? State of art – test procedure to measure the toxic gas after crash (if required ): is there any test procedure available for toxic gas measurement post-crash of complete vehicle? => ???? Is the test procedure repeatable & reproducible ? What are gas we need to measure ? What are the limits ? Other issue: How to distinguish with venting gas? To be discussed

15. Conclusion and future work Completed : List of issues related to electrolytes and associated potential risks Discussions on aqueous electrolyte : Leakage detection and measurement techniques are clarified ????? Future work : Discussions on risk associated with non-aqueous electrolytes Propose final text for agreed topics Required time: ????? To be discussed

16. Annex

17. Electrolyte Leakage: Additional issues raised during TF3 meetings TopicIssueComments Observation period:JRC proposed to increase the observation period from 30 minutes to 60 minutes This issue is not specific to this task force and need feedback from other experts (TF4 for example ) Roll-over testKorean proposal: The electrolyte leakage requirement shall be met in case of rollover after vehicle impact tests (same as that of FMVSS 305) This comments is related to vehicle crash test procedure and this is not part of this GTR-EVS ‘No fire’ requirementJRC proposed to use ignition source to check potential flammability of the emitted gas from leaked electrolyte This issue is not specific to this task force and is more related to test procedure Spillage/leakageChina and JRC proposed to use only one definition To be discussed during the F2F meeting

18. REFERENCES [1] JRC comment: None of the standards distinguishes between electrolyte spillage and leakage and these 2 terms are used indistinctively. Introducing separated definitions in the GTR might lead to confusion and ambiguity. Literature information: definitions found in relevant standards: - FMVSS 305=> ELECTROLYTE SPILLAGE: The fall, flow, or run of propulsion battery electrolyte in, on, or from the vehicle, including wetness resulting from capillary action. - SAE J2464=> RELEASE: any spilling, leaking, pumping, pouring, emitting, emptying, discharging, injecting, escaping, leaching, dumping, or disposing into the environment. - IEEE 1578=> ELECTROLYTE RELEASE: Any escape of electrolyte from B, whether in vapour, liquid or gel form. - UN38.3. => LEAKAGE: visible escape of electrolyte or other material from a cell or battery or the loss of material (except battery casing, handling devices or labels) from a cell or battery such that there is loss of mass. - UNECE R100.02 => Both terms are used, however no differentiation is specified. JRC proposal: We propose to define and use only one term accounting for the loss of electrolyte. [2] Literature information: - Electrolytes that may be used in LIB vary considerably; they may meet the definition of one or more hazard classes, such as Class 3 (Flammable), Division 6.1 (Poison) or Class 8 (corrosive) [NREL Current status of environmental health and safety issues of Li ion EV batteries, Laura J. Vimmerstedt, Shan Ring, and Carol J. Hammel. Sep. 1995. NREL/TP-463-7673 (page 36)] and MSDS of various solvents (e.g. EC, DMC, PC, ACN, etc.) - The most commonly used electrolyte salts is LiPF6 will decompose to form HF if mixed with water or exposed to moisture. Leakage of free electrolyte from cells can result in deposition of the electrolyte salt. [Celina Mikolajczak et al., Lithium-ion batteries hazard and use assessment. Springer, 2011, page 13]. - Substances that react when in contact with water (e.g. LiPF6) evolving dangerous quantities of flammable or toxic gas are classified in division 4.3 (dangerous when wet material) under the DOT ((U.S. Department of Transportation) regulations. From NREL Current status of environmental health and safety issues of Li ion EV batteries, Laura J. Vimmerstedt, Shan Ring, and Carol J. Hammel. September 1995. NREL/TP-463-7673 (page 36)]. - Electrolyte leakage poses two potential safety hazards: human contact with the electrolyte and electrolyte residue, and short circuiting of adjacent electronic systems. [Celina Mikolajczak et al., Lithium-ion batteries hazard and use assessment. Springer, 2011, page 45].

19. [3] JRC comment: For non-aqueous electrolytes the amount of vapours generated are not negligible given the high volatility of their components. They may come not only from a liquid spillage but from a damaged cell (vapours will diffuse out of casing). Literature information: - MSDS of various solvents (e.g. EC, DMC, PC, ACN, etc.) -Organic liquids that may be used in LIB are volatile. In poorly ventilated conditions, vapours from these liquids may reach dangerous concentrations, such that the gas displaces too much air. If an exposed individual receives insufficient oxygen, unconsciousness or death may result. Similarly, decomposing or burning any electrolyte organic liquid may produce asphyxiants (CO, CO2). [Ref. NREL Current status of environmental health and safety issues of Li ion EV batteries, Laura J. Vimmerstedt, Shan Ring, and Carol J. Hammel. September 1995. NREL/TP-463-7673 (page 13)]. -If cells with water based electrolyte are punctured or damaged, leakage of the electrolyte can pose a corrosive hazards; however does not pose a flammability hazard. In comparison, leakage or venting of Li ion cells will release flammable vapours. [Celina Mikolajczak et al., Lithium ion batteries hazard and use assessment, Springer, 2011, page 86]. -Many B systems contain components that may be considered hazardous when the battery is damaged. Under normal operation these systems do not emit toxic substances. In the event of catastrophic failure, such as an external fire, or crash there may be the possibility of hazardous levels of toxic emissions. It is understood that the B and vehicle manufacturers should be aware of any hazards that may exist under circumstances other than catastrophic failure and design for prevention of theses hazards. SAE J2289 -The concentration of the released hazardous substances shall be scaled to the full REESS pack for quantitative comparison and scaled to a volume appropriate to human exposure in the vehicle. SAE J2464. -To prevent explosion, fire or toxicity hazards, the following requirements apply when hazardous gases and other substances can be emitted by the REESS. Refer to the later version of applicable National/International standards or regulations for the maximum allowed accumulated quantity of hazardous gases and other substances. ISO 6469-1 JRC proposal: Vapors generated from a release constitute a real hazard and solutions to this have to be evaluated, (please refer to [5]) [4] JRC comment: Venting in aqueous electrolyte batteries is allowed and does not create a hazardous situation, as the systems continues being functional. However, venting in non-aqueous systems imply that the cells are experiencing some sort of irreversible failure. Although it is recognized that venting prevents a violent event in case of pressure build -up inside a B, a LIB that has vented will continue gassing the electrolyte creating a hazardous situation on the medium/long term. Literature information: - Sealed lead-acid batteries generally incorporate relief valves capable of providing overpressure relief throughout the life of the batteries. The valves are actuated at pressures of about 5 atm and close again at lower pressures. The vents on lithium batteries are different. They are not valves, but rupture disks that provide no more than one-time overpressure relief. Once the valves are actuated, the cells become inoperative, but not necessarily safe. [S.C. Levy, P. Bro, Battery Hazards and Accident Prevention, Plenum Press, 1994, page 95 ].

20. [4] (continued) JRC proposal: We propose to include clear definitions for venting for aqueous and non-aqueous systems and it needs to be clearly specified that venting only to the outside of a vehicle is allowed. Related hazards need to be considered. [5] JRC comment: In order to ensure that the requirement on no electrolyte release (in the in-use tests) or the requirement on limited electrolyte release (in the post-crash tests) is fulfilled a new detection method is needed as proposed visual detection may not be adequate for the electrolyte release detection and quantification. We recognize that most of the standards in this topic (gas emissions, post-crash situation) relate to the traditional lead acid technologies. However the EVS GTR should be able to point out the issue and find solutions to it, as non-aqueous technologies are emerging in the EV and HEV market. Literature information: -SAE J2464 describes under hazardous substance monitoring examples of analytical techniques aimed at identify hazardous species. Examples of acceptable analytical techniques are described in the EPA methods TO-15 and TO-17 for the determination of VOCs in air : (www.epa.gov/ttnamti1/files/ambient/airtox/to-15r.pdf and www.epa.gov/ttnamti1/files/ambient/airtox/to-17r.pdf).www.epa.gov/ttnamti1/files/ambient/airtox/to-15r.pdfwww.epa.gov/ttnamti1/files/ambient/airtox/to-17r.pdf -USABC related standards point out the need to quantify these species. -SAEJ2289 mentions how in case of crash of HEVs there may be the possibility of hazardous levels of toxic emissions. -SAE J2464, USABC and Freedom car standards point out the need to test hazardous substances and to calculate if ERPG-2 levels could be achieved. ERPG-2 levels are the maximum airborne concentration levels below which most individuals could be exposed for up to one hour without experiencing or developing serious or irreversible health effects or symptoms. Emergency Response Planning Guidelines, 1998 ERPG Complete Set,” Stock No. 303-EA-98, American Industrial Hygiene Association.,Fairfax, Virginia. -In the case of crash test organisations and based on research consideration should be given to having a gas monitor to check for flammable or toxic gases near the crashed vehicle. [Michael Paine et.al. SAFETY PRECAUTIONS AND ASSESSMENTS FOR CRASHES INVOLVING ELECTRIC VEHICLES NHTSA, National highway traffic safety administration, 22nd Enhanced safety vehicles conference, Paper Number 11-0107] JRC proposal: A new method based on the quantitative detection of the air-borne electrolyte components and products of their decomposition should be proposed to substitute where appropriate currently adopted visual inspection.