Content uploaded by Roeland Bisschop
Author content
All content in this area was uploaded by Roeland Bisschop on Oct 18, 2019
Content may be subject to copyright.
SAFETY & TRANSPORT
FIRE RESEARCH
Fire Safety of Lithium-Ion Batteries in Road
Vehicles
Roeland Bisschop, Ola Willstrand, Francine Amon,
Max Rosengren
RISE Report 2019:50
© RISE Research Institutes of Sweden
Fire Safety of Lithium-Ion Batteries in Road
Vehicles
Roeland Bisschop, Ola Willstrand, Francine Amon,
Max Rosengren
2
© RISE Research Institutes of Sweden
Abstract
The demand for lithium-ion battery powered road vehicles continues to increase around
the world. As more of these become operational across the globe, their involvement in
traffic accidents and fire incidents is likely to rise. This can damage the lithium-ion battery
and subsequently pose a threat to occupants and responders as well as those involved in
post-crash operations. There are many different types of lithium-ion batteries, with
different packaging and chemistries but also variations in how they are integrated into
modern vehicles. To use lithium-ion batteries safely means to keep the cells within a
defined voltage and temperature window. These limits can be exceeded as a result of crash
or fault conditions. This report provides background information regarding lithium-ion
batteries and battery pack integration in vehicles. Fire hazards are identified and means
for preventing and controlling them are presented. The possibility of fixed fire suppression
and detection systems in electric vehicles is discussed.
Key words: Lithium-Ion Batteries, Electric Vehicles, Fire Risks, Post-Crash Handling, Risk
Management, Fire Safety
RISE Research Institutes of Sweden AB
RISE Report 2019:50
ISBN: 978-91-88907-78-3
Borås 2019
Cover image: A collage of four different images. Burning heavy truck on a highway,
burning passenger car in an urban area, passenger cars in dense traffic, bus travelling
through an urban area.
3
© RISE Research Institutes of Sweden
Acknowledgements
The project (No. 45629-1) is financed by the Swedish FFI-program (Strategic Vehicle
Research and Innovation) which is a partnership between the Swedish government and the
automotive industry. Partners within this project comprise of RISE Research Institutes of
Sweden, Scania, Volvo Buses, SFVF (Swedish Association of Vehicle Workshops), Fogmaker
International and Dafo Vehicle Fire Protection. All support in the project is gratefully
acknowledged.
4
© RISE Research Institutes of Sweden
Content
Abstract ..................................................................................................... 2
Acknowledgements .................................................................................... 3
Content ...................................................................................................... 4
1 Introduction ........................................................................................ 6
2 Electric Road Vehicles .......................................................................... 7
2.1 Statistics................................................................................................................... 7
2.2 Vehicle Configurations .......................................................................................... 10
2.3 Plug-In Charging ................................................................................................... 14
3 Lithium-Ion Batteries ......................................................................... 15
3.1 Packaging ............................................................................................................... 15
3.2 The Electrochemical Cell ....................................................................................... 17
3.2.1 Electrolyte ...................................................................................................... 18
3.2.2 Separator ........................................................................................................ 19
3.3 Lithium-Ion Batteries in Road Vehicles ............................................................... 20
3.3.1 Lithium-Ion Battery Packs, Modules and Cells ............................................ 20
3.3.2 Passenger Cars with Lithium-Ion Batteries ................................................... 21
3.3.3 Heavy Vehicles with Lithium-Ion Batteries .................................................. 26
4 Fire Risks Associated with Lithium-Ion Batteries .............................. 30
4.1 Thermal Runaway ................................................................................................ 30
4.2 Battery Failure Causes .......................................................................................... 32
4.2.1 Internal Cell Short Circuit ............................................................................. 33
4.2.2 Mechanical Deformation and Impact ........................................................... 34
4.2.3 Charge ........................................................................................................... 36
4.2.4 Discharge ........................................................................................................
37
4.2.5 External Short Circuit ................................................................................... 38
4.2.6 Exposure to High Temperatures ................................................................... 39
4.3 Hazards and Risk Factors ..................................................................................... 40
4.3.1 Chemistry ....................................................................................................... 41
4.3.2 State of Charge and Cell Capacity ................................................................. 42
4.3.3 Thermal Propagation .................................................................................... 44
4.4 Challenges for Responders ................................................................................... 45
4.4.1 Identifying Electric Vehicles ......................................................................... 45
4.4.2 Toxicity of Vented Gases and Fire Water Run-Off ........................................ 46
4.4.3 Fibre Composite Materials .............................................................................47
5 Collisions and Fires ............................................................................ 48
5.1 Documented Incidents ......................................................................................... 48
5
© RISE Research Institutes of Sweden
5.1.1 Trends and Statistics ..................................................................................... 50
5.2 Handling of Damaged Electric Vehicles ................................................................ 51
5.2.1 Fire Hazards .................................................................................................. 52
5.2.2 Electrical Hazards ......................................................................................... 56
6 Safety Solutions ................................................................................. 59
6.1 A Holistic View ..................................................................................................... 59
6.1.1 Battery Cell Level ........................................................................................... 60
6.1.2 Battery Management System (BMS) ............................................................. 60
6.1.3 Battery Module Level .................................................................................... 62
6.1.4 Battery Pack Level ......................................................................................... 62
6.1.5 Vehicle Level ................................................................................................. 63
6.2 Fixed Fire Detection and Suppression Systems ................................................... 64
6.2.1 Detection ....................................................................................................... 64
6.2.2 Suppression ................................................................................................... 65
6.3 Hazard Identification Workshop .......................................................................... 68
6.3.1 Method .......................................................................................................... 68
6.3.2 Results ........................................................................................................... 69
7 Conclusions ........................................................................................ 71
8 References .......................................................................................... 72
Appendix A, Documentation from Workshop ........................................... 91
Prevention ..................................................................................................................... 91
Recovery ....................................................................................................................... 98
Appendix B, Participants of Workshop ................................................... 104
6
© RISE Research Institutes of Sweden
1 Introduction
The demand for electric vehicles (EVs) continues to increase around the world. This is
largely due to regulations related to air quality and environmental issues in combination
with consumer demand and cheaper rechargeable energy storage systems. Furthermore,
significant developments have made these storage systems, especially those belonging to
the lithium-ion family, suited for automotive applications [1].
As more lithium-ion battery (LIB) powered road vehicles become operational across the
globe, their involvement in traffic accidents is likely to rise. As for conventionally fuelled
vehicles, the on-board energy storage system is a risk factor for those involved in, or
responding to, accidents. While the risks associated with conventional vehicles are well-
defined and generally accepted by society; time and education are needed to achieve this
comfort level for LIB powered road vehicles. When it comes to EVs there is a risk that the
LIB may ignite after significant amounts of time after being damaged or reignite after having
been extinguished. This matter not only concerns firefighters, but also those involved in
handling damaged EVs through towing, workshop, scrapyard or recycling activities.
This RISE report, part of current project (No. 45629-1), addresses these and other concerns
through a review of available literature. Fundamental information on EVs and LIBs is
presented, and matters related to fire risks and safety solutions are investigated. This
provides a scientific basis to those who seek to develop their own guidelines and routines
for handling risks associated with LIBs in road vehicles.
Current project will continue to investigate and develop relevant risk management routines
and evaluate fire suppression and emergency cooling systems. For the latter, full-scale
experiments will be performed to evaluate if they can enhance safety when integrated into
LIBs.
7
© RISE Research Institutes of Sweden
2 Electric Road Vehicles
Over the last few years there has been a continuous and strong increase in the number of
electric vehicles on our roads. This is largely due to regulations related to air quality and
environmental issues in combination with consumer demand and cheaper rechargeable
energy storage systems. Furthermore, significant developments have made these storage
systems, especially those belonging to the lithium-ion family, suited for automotive
applications [1].
However, the shift to new and different means of transport and infrastructure is
accompanied by new risks. It is thus important to have a basic understanding about these
vehicles as their involvement in traffic accidents is likely to increase. This chapter addresses
this by providing background information needed to understand electric vehicles. Specific
topics include statistics related to the growing number of electric vehicles as well as their
operating principles and fuelling mechanisms. Together they provide basic insight into the
scope of their market penetration and the unique features that set them apart from other
vehicles.
2.1 Statistics
Data from the International Energy Agency up to 2017, presented in Figure 1, shows that
most of the passenger cars in the world can be found in the Peoples Republic of China
(China), the European Union (EU) and the United States of America (US) [2]. In 2017,
approximately 40 % of all electric passenger cars in the world could be found driving around
in China. Coming in second is the EU with roughly 870 000 electric passenger cars. This is
relatively close to the US, where 760 000 electric passenger cars were recorded for the same
year.
Figure 1 The uptake of electric passenger cars is dominated by China, the US, and the EU [2].
0
500
1000
1500
2000
2500
3000
3500
Number of electric passenger cars
x 1000
Total
China
US
EU
8
© RISE Research Institutes of Sweden
Figure 2 shows how the number of electric passenger cars in the Nordic countries compare
to the rest of the EU according to the European Alternative Fuels Observatory [3]. Together,
the Nordic countries represent the largest market for electric vehicles in the EU, with most
purchases made in Norway and Sweden [4]. The country that stands out the most is Norway.
In 2018, approximately half of all passenger cars sold in Norway were electric [3]. This is
much higher than other Nordic countries, where electric passenger cars sold in Sweden,
Denmark, Finland and Iceland comprised about 8%, 2%, 5% and 20% of all new cars sold
in 2018, respectively [3].
Figure 2 The growth in electric passenger cars in Europe and the Nordic countries [3]
Other vehicle types, such as buses are experiencing similar trends as those observed for
passenger cars, see Figure 3. Currently, this shift is particularly evident for public
transportation solutions in large cities. Influencing factors in this are the cost and weight of
lithium-ion battery packs. Specifically, smaller batteries can be used in local and city traffic
as due to the short routes and frequent stops. In contrast, long haul buses, such as coaches,
require very large and heavy batteries or require continuous charging. It is thus no surprise
that the current growth has been most significant in metropolitan areas.
Figure 3 Number of electric buses operating in the European Union [5].
0
200
400
600
800
1000
1200
Number of electric passenger
cars x 1000
EU
Norway
Sweden
Denmark
Finland
Iceland
0
400
800
1200
1600
2000
Number of electric buses
EU
Netherlands
Germany
Nordic 5
Austria
France
9
© RISE Research Institutes of Sweden
Similar trends are seen when it comes to transportation of goods by electric heavy trucks.
Rechargeable energy storage systems, such as lithium-ion batteries, are still less energy-
dense than fossil-fuel1. This means that a significant charging infrastructure and/or
development in battery technology is needed to sustain continues operation over long
distances. They are currently more suited to short distance delivery in metropolitan areas.
A good example are heavy trucks used to deliver goods inside metropolitan areas or services
to and from ports and rail yards. Typically, these heavy trucks drive short distances with
frequent stops for loading, unloading and charging. This makes them suitable candidates
for electrification.
Other aspects are the increasingly stringent emission and noise requirements on vehicles.
To enter some urban cores, vehicles are required to have low emissions whereas the reduced
noise emissions from an electric truck would make it possible to operate quietly at night
which is very attractive to e.g. refuse collectors and last mile distributors. Currently there
are only a few electric heavy trucks operational in today’s market, however, see Figure 3.
This is likely to change, as more electric heavy trucks are set to be released this year as seen
in Table 1.
Figure 4 Number of electric heavy trucks operating in the European Union [6]
1 To give an example, a commercial lithium-ion battery cell LCO type with a nominal voltage of 3.7V and
energy density of 200mAh/g has a specific capacity of 0.74 kWh/kg [1]. That of gasoline and diesel lies
around 13 kWh/kg [250].
0
45
90
135
180
Number of electric heavy trucks
EU
Netherlands
Italy
Nordic 5
Austria
Germany
10
© RISE Research Institutes of Sweden
Table 1 Electric heavy trucks that are yet to be released.
Electric Heavy Trucks Use/Role Availability
Scania L 320 6x2 PHEV [7] Urban, distribution, waste,
construction Market release, 2019
DAF LF Electric [8] Urban, light duty Field test, 2018/2019
DAF CF Electric [8] Urban, medium duty Field test, 2018/2019
DAF CF Hybrid [8] Urban, mid-range Field test, 2018/2019
Volvo FL Electric [9] Urban Market release, 2019
Volvo FE Electric [10] Urban, heavy loads Market release, 2019
Mack LR Electric [11] Urban, refuse collection Field test, 2019
Volvo Vera [12]
Shipping ports and logistics centres,
autonomous, repetitive short trips,
heavy loads.
Unknown
2.2 Vehicle Configurations
There are several significant advantages with electrification. They have proven to reduce
emissions and operate more efficiently than vehicles driven by fossil-fuels. The major issue
with conventional powertrains lies in the power source, the internal combustion engine.
Even the most advanced types for automotive applications operate below 50 % efficiency
[13] [14]. Electric Machines (EM), however, operate at around 95 % efficiency [13].
Other technologies such as regenerative braking provide further efficiency benefits to
electrification. A significant number of vehicles have been hybridised for this exact purpose.
When the vehicle brakes, energy is generated and stored in a small on-board battery. This
energy is subsequently used to power the vehicle. Such vehicles are commonly referred to
as mild hybrids.
There are many different options for driving fully or partially on electric power. Depending
on the definition of an electric vehicle they may be hybrid, plug-in hybrid, range-extended,
battery electric or fuel cell electric. An overview of these, and their common abbreviations
may be seen below in Table 2. Note that these classifications mainly reflect on the way a
vehicle’s powertrain is configured. In this study, vehicles which have a hybrid or fully
electric powertrain are referred to as electric vehicles (EVs).
Table 3 shows a conventionally fuelled, and driven, vehicle. This type of vehicle has an on-
board fuel tank. Fuel is pumped to the ICE and combusts in its cylinders. Subsequently, the
combustion energy propels a crank, which drives a large flywheel connected to a
transmission, which converts the power into the speed and force needed to drive the vehicle.
In doing so, the chemical energy of the fuel has been converted to mechanical work.
The process of combusting fuel to generate mechanical work has a low efficiency. The
efficiency of current ICEs for passenger cars lies in the range of 30-36% [14]. Very efficient
diesel-fuelled ICEs can achieve 39-47% [13] [14].
The amount of fuel stored in passenger cars and heavy vehicles is normally within the range
of 50-100 L and 400-1000 L, respectively [15]. In passenger cars the fuel tank is normally
placed near the rear axle. This provides protection against impact, which is important as
most conventional fuels are extremely flammable.
11
© RISE Research Institutes of Sweden
Table 2 Classification of electric road vehicles.
Vehicle 1st Motor 2nd Motor Acronym
Electric
Range2
[km] [16]
Power Source
Conventional
vehicle
Internal
combustion
engine (ICE)
None ICEV 0 Fossil-fuel
Hybrid electric
vehicle ICE
Electric
machine
(EM)
HEV 0 to 10
Fossil-fuel
combined with
Lead-acid,
NiMH or Li-ion
battery
Plug-in hybrid
electric vehicle
ICE or
electric
machine
(EM)
EM or ICE PHEV 20 to 85
Fossil-fuel
combined with
Li-ion battery
Range extended
electric vehicle EM ICE REEV or
PHEV 75-145
Fossil fuel
combined with
Li-ion battery
Battery electric
vehicle EM None BEV 80 to 400 Li-ion battery
Fuel cell electric
vehicle EM None FCEV 160 to 500
Fuel cell, can be
combined with
Li-ion battery or
supercapacitor
The BEV does not consume any fossil fuel or emit exhaust gas. The BEV powertrain
primarily consists of a traction battery, electric machine and a transmission or final drive
system. This can be seen in Table 4. At the heart of the BEV lies a lithium-ion traction
battery. These have to be significant in size in order to achieve sufficient driving ranges. It
takes up a lot more space than fuel tanks due as lithium-ion batteries generate less energy
per weight unit than gasoline or diesel. Specifically, 0.1-0.2 kWh/kg versus more than 10
kWh/kg for conventional fuels. This also means that the TB make up a large portion of the
vehicles total weight. For example, the battery pack in the Tesla Model S 85 makes up 30%
of its total weight [17].
2 Indicative electric driving range for passenger cars.
12
© RISE Research Institutes of Sweden
Table 3 ICE configuration and system components
Figure 5 ICE configuration
System Application
ICE
The fuel combusts in the cylinders of the ICE,
propelling a crank, which transfers combustion
energy to mechanical work. Efficiency <50% [13]
[14].
Gearing
Transfers mechanical work. Gearing refers to the
transmission, differential, split drive and any
other devices which transfer mechanical power.
Mechanical
power
Typically, rotating shafts and axles due to an
applied torque.
Fuel tank
Generally, for passenger cars, fuel tanks can store
between 50 to 100 L of fuel whereas heavy
vehicles such as trucks and buses store 400 to
1000 L of fuel [15].
Fuel line Typically, in the form of reinforced rubber hoses.
Fuel port Port that connects to fuelling equipment in order
to re-fill the fuel tank.
Table 4 BEV configuration and system components
Figure 6 BEV configuration
System
Application
Traction
battery
Stores electrical energy which can be released and
made available to power the vehicle. Lithium-ion
batteries (LIBs) are the preferred option for the
traction batteries in PHEVs and BEVs.
Electric
machine
Used as a traction motor and
sometimes a generator
[18]. Efficiency ~95% [13].
High
voltage
cables
High voltage cables may be found between the
battery and power electronics as well as between the
power electronics and the electric machines. Their
total weight may be up to 10 kg in hybrid passenger
vehicles [18].
Battery
charger
Electrical power grids provide AC current and
batteries can only store DC current. The electricity
thus needs to be converted. This is achieved by
either an on-board AC/DC converter or by a
converter integrated into the charging station itself
[18].
In the automotive industry, hybrids are vehicles that have an electric powertrain as well as
an ICE system. Doing so allows for significant fuel savings. It allows for regenerative
braking, smaller engines and more efficient operating conditions, as well as the ability for
engine shut-off when idling or at low load conditions [17].
13
© RISE Research Institutes of Sweden
There are different types of hybrids on the market. They can be classified as series, parallel,
or series/parallel hybrids. Each of these has its advantages and disadvantages. Series
hybrids operate on the electric machine. They have a large TB and small IC [17]. As seen in
Figure 7 there is no mechanical connection between the ICE and the wheels. This allows the
ICE to continuously operate at its most efficient engine speeds.
Parallel hybrids have the option to be powered by the EM or ICE independently, or to
employ them simultaneously, see Figure 8. In this case there is a direct connection between
the ICE, the transmission, and the final drive. These conditions give efficient driving at
highway speeds. Usually parallel hybrids have a large ICE and a small TB [17].
Split hybrids, also referred to as series/parallel, combine the best of these configurations.
As can be seen from Figure 9, they allow for vehicles to be powered directly by the ICE, with
the EM assisting, or for the ICE to power a generator that powers the EM. This greater
flexibility does generally come at a higher cost and with a more complex vehicle design.
Plug-in hybrids (PHEVs) follow these principles to the same extent HEVs do. The main
difference is that PHEVs have larger batteries and therefore greater electric driving ranges.
Energy generating systems such as regenerative braking is not enough to charge these large
batteries hence external charging is needed.
Table 5 Hybrid configurations and system components
Figure 7 Series hybrid
Figure 8 Parallel hybrid
Figure 9 Split hybrid
Propulsion
power
converters
Converts power from AC/DC or DC/AC. The DC/AC converter is located
between the battery and electric motor. Hybrid vehicles are also equipped
with an AC/DC converter between the generator and the traction battery
[18]
.
DC/DC
converter
Converter which connects to a 12V battery (passenger cars) that powers
auxiliary equipment. In electric vehicles this battery is charged by the
traction battery
[18].
14
© RISE Research Institutes of Sweden
2.3 Plug-In Charging
As for conventional vehicles, the driving range of EVs is limited by its fuel. In this case
however, rather than filling the fuel tank with liquid, the battery must be supplied with
electricity. There are three different ways of doing this, i.e. by swapping the battery with a
fully charged one, charging wirelessly, or plug-in charging.
Plug-in charging is used to charge the vast majority of EVs in Europe [16]. Charging occurs
by physically connecting a power cable from the EV to the grid. There is an international
standard for conductive charging systems for EVs, namely IEC 61851. This standard defines
four charging modes.
The first charging mode, mode 1, considers the EV to be connected to the grid by using
common household sockets and cables. The current that is supplied is very limited [16]. In
addition to this, there are no protective systems [19]. It is therefore not very relevant for
EVs, and more commonly used to charge light vehicles, e.g. bicycles and scooters [20].
Mode 2 charging also considers charging through household sockets. This type of charging
is considered slow or semi-fast [16]. A special cord is used with built-in charging equipment
and this cord is equipped with a so-called pilot function and a circuit breaker. The pilot
function detects the presence of the vehicle, communicates the maximum allowable
charging current, and controls charging.
The third mode of charging connects the EV to a charging station via a special plug-socket.
This type of charging is specifically designed for EVs and allows charging at higher power
levels. In this case there is communication between the vehicle and the dedicated charging
station, not with the cable. This type of charging has a high degree of safety as these
protection systems are installed in the charging station itself. Among other things, the
power supply must fulfil the requirements set by the on-board charger and those of the
charging station. If not, there is no power transfer between the charging station and the EV.
The final charging mode, and the fastest one, is referred to as Mode 4. Here the EV is
connected to the power grid through a charger inside an offboard charging station [21]. In
this case, the control and protection functions as well as the charging cable are permanently
installed in the charging station. The charging station itself converts AC power to DC power
inside the charging station. For the other charging modes, this conversion is achieved with
AC/DC converters that are inside the EV. As such, Mode 1 – 3 are sometimes referred to as
on-board charging whereas Mode 4 is called off-board charging. Note that not all EVs allow
for DC charging.
15
© RISE Research Institutes of Sweden
3 Lithium-Ion Batteries
The energy of lithium-ion batteries (LIBs) is housed within individual battery cells. Each
cell has one positive and one negative terminal. These terminals are connected to thin metal
foil that has been coated with electrochemically active material. The active material for the
negative and positive side of the battery is referred to as anode and cathode material,
respectively. When the battery is discharged, electricity flows into the anode and out of the
cathode, see Figure 10.
Depending on the cell geometry, the current collectors is pressed or rolled together with
polymer separators and submerged in electrolyte. This is an electrically conductive media
that allows for lithium-ions to be transported from one side to the other. The transfer of
lithium-ions from one side to the other, through a separating material, results in chemical
reactions that generate an electrical current. The direction of current depends on whether
the battery is discharged or charged. In the case of charge, it flows from the anode to
cathode, see Figure 10. The opposite happens when the battery is discharged.
Figure 10 Working principle of a lithium-ion battery when discharged.
3.1 Packaging
Packaging material is used to protect the electrochemical components of the lithium-ion
battery cell. This packaging can be of different materials and shapes. They are usually
distinguished by the shape of the package. As such, LIB cells are thus sometimes referred to
as being cylindrical, prismatic or pouch cells. These are shown in Figure 11.
16
© RISE Research Institutes of Sweden
Cylindrical
Pouch
Prismatic
Figure 11 Exterior housing types that are common for lithium-ion battery cells.
Cylindrical cells have a very high mechanical stability as their shape distributes forces, due
to internal pressure increase, evenly over the circumference. Their shape makes it however
harder to package them together in an efficient manner as a significant amount of space is
lost when arranging them in a rectangular shape. This, however, make it easier for air to
flow freely around a group of cylindrical cells resulting in easier thermal management [22].
Prismatic cells are commonly used for automotive traction batteries. Their prismatic shape
makes them easier to integrate in a battery pack than cylindrical cells, see Figure 12. This
can make it more challenging to regulate their temperatures. The contents of prismatic cells
follow the principle for cylindrical cells. Instead of rolled up, several layers of current
collector packages are put on top of each other. As a result, prismatic cells tend to be tightly
packed, which results in relatively high mechanical stresses on the prismatic package [22].
Pouch cells store their content inside a flexible foil pouch rather than inside a rigid
container. In this case, the current collector assembly is stacked inside the pouch package,
rather than rolled. This gives most of the space inside the package to be used for
electrochemical material and thus allows for a high energy density per pouch cell. Their flat
shape also allows for very high packaging efficiency of 90-95 % when it comes to integrating
them in battery packs [23]. As a result, temperature management also becomes more
important for this cell type, as it is more difficult to dissipate heat. Their soft construction
can also be a drawback as it makes them vulnerable to external mechanical damage.
Furthermore, pouch cells require a support structure as they are not mechanically rigid.
Figure 12 The packaging shape of the battery cells influences pack density and heat dissipation.
17
© RISE Research Institutes of Sweden
3.2 The Electrochemical Cell
A LIB package consists of different electrochemical materials. These include anode,
cathode, separator and electrolyte. Each of them plays a role in the batteries’ properties
concerning specific energy, life, safety and cost.
The type of cathode material is often used to classify lithium-ion batteries in groups. This is
mostly because their chemistry is one of the main variables in their construction. There are
many different options available, see Table 6. Note that cathode types presented here only
are a selection of the most common and commercialised types. For a more complete
overview refer to Nitta et al. [1].
Lithium Cobalt Oxide (LCO), as seen in Table 6, is common in a very large number of
consumer devices such as smartphones. It offers relatively high capacity and voltage
compared to other cathode materials and it is relatively easy to manufacture [24]. There are
however significant safety concerns, especially at high temperature and overcharge
conditions.
Introducing new technologies, such as EVs, to a market dominated by conventionally
fuelled vehicles comes with a major barrier, i.e. fear of the unknown. Compromising on
safety and accepting the risk of EVs catching fire due to LIB failure, even when abused, is
thus not an option. The automotive industry therefore generally avoids chemistries that are
known to have low thermal stability. Instead, they opt for safer cathode materials such as
Lithium Iron Phosphate (LFP), Lithium Nickel Manganese Cobalt Oxide (NMC), Lithium
Manganese Oxide (LMO) or blends of different cathode materials.
Performance is the major influencing factor when manufacturers choose certain cathode
materials while not making any compromises on safety. To achieve high-performance and
fast acceleration, a battery needs to be able to supply a lot of power. When longer driving
ranges are needed the focus shifts to achieving a high energy density. Normally both high
power and high energy density are desirable and today NMC, or Lithium Nickel Cobalt
Aluminium Oxide (NCA), paired with graphite anodes is the most common [25] [26].
The characteristics of cathode materials can be modified further by blending different
cathode materials. Such materials are referred to as hybrid or blended cathode materials.
This technology was developed by commercial automotive battery suppliers and consist of
a mixture of two or more different active materials [27]. Blending allows for different
cathode materials to complement each other. It combines the best properties of the
individual active materials and helps to reduce the shortcomings of the parent materials.
Julien et al. showed that drawbacks of LFP, i.e. relatively low energy density, could be
overcome by blending it with NMC [28]. Simultaneously, the material had better thermal
stability than NMC by itself, due to the positive influence of LFP.
18
© RISE Research Institutes of Sweden
Table 6 Overview of the properties of common cathode materials.
Specific Energy [29] Voltage at 50% SOC [29] Life [17] Safety [17] Cost [17]
LFP 160 Ah/kg 3.4 V High High Medium
LMO 100-120 Ah/kg 4 V Low Medium Low
LCO 155 Ah/kg 3.9 V Medium Low Medium
NCA 180 Ah/kg 3.7 V Medium Low High
NMC 160 Ah/kg 3.8 V High Medium High
The number of options when it comes to anode materials are more limited. There are two
types of anode materials commercially available, namely those comprised of different
carbon configurations and Lithium Titanate Oxide (LTO) [1]. The former offers a good
balance between cost, availability, energy and power density as well as cycle life whereas the
latter provides better performance when it comes to thermal stability, charge/discharge rate
and cycle life but suffers when it comes to cost, cell voltage and cell capacity [1].
3.2.1 Electrolyte
The individual components inside LIBs are soaked in an electrically conductive solution
referred to as electrolyte. This media allows for ions to be transported between the positive
and negative electrodes. It plays a very large role in the safety and general performance of
LIBs. There are many different types of compositions possible and available, yet not all of
them are compatible with other battery components or able to hold charge.
The vast majority of electrolytes for LIBs are nonaqueous solutions [30] [31], i.e. water is
not the solvent. Electrolytes used for LIBs normally consist of Lithium
Hexafluorophosphate (LiPF6) salts and organic carbonate solvents such as Ethylene
Carbonate (EC). The composition of the solutions has mostly remained the same over the
last decade. Xu [31] argues that this is due to three key factors; the electrolyte components
being very sensitive, additives having become more effective, and the fact that the battery
industry has been reluctant to change the existing supply chain.
Electrolyte components for LIBs are sensitive. Their operating temperature is limited, and
typically lies between -20 ºC and +50 ºC [32]. When exposed to environments that are not
within this range of safe operation, they could be permanently damaged. This starts with
decomposition reactions of the interphase layer between the anode and electrolyte, referred
to as solid electrolyte interphase (SEI). Herstedt [33] found that the onset of these reactions
is strongly dependent to the salt that is used. Electrolytes systems with lithium
tetrafluoroborate (LiBF4) or lithium hexafluorophosphate (LiPF6) salts, these reactions start
at around 60-80 ºC and 80-100 ºC, respectively. For lithium triflate (LiTf) and lithium
bisimide (LiTFSI)3 systems the decomposition reactions start at 110-120 ºC and 125-135 ºC,
respectively. This is potentially followed by other exothermic reactions inside the LIB.
3 Lithium bis(trifluoromethanesulfonyl)imide
19
© RISE Research Institutes of Sweden
Another major issue with the current electrolytes considered for LIBs is its flammability. As
seen in Table 7, not all electrolyte constituents are equally flammable. The most flammable
solvent is Ethyl Acetate (EA). Among other things, this is due to the fact that it has a very
low flashpoint. When exposed to temperatures below zero, EA releases enough vapour to
sustain burning if ignited by a spark or flame. Note however that in comparison to gasoline,
a convential fuel that has been around for more than a century, this solvent is relatively safe.
Additives and electrolyte components have been shown to be able to lower the flammability
and slow down the thermal degradation of electrolyte [32]. Their main drawback is however
that they can reduce performance [34]. Alternative electrolytes are being developed.
Specifically, nonaqueous fluoro-compounds, ionic liquids and polymeric electrolytes [31]
[25] [30]. None of these, except for certain polymeric electrolytes, have been
commercialised on a large scale yet. The polymeric electrolytes currently available offer
improved mechanical strength and less potential for leakage of toxic fluids [35] yet remain
limited to the same safety window as conventional electrolyte [36].
Table 7 Flammability data for the electrolyte solvent in LIB cells and data for conventional
automotive fuels for comparison.
Organic Electrolyte Solvents
Boiling
Temperature
[ºC]
Autoignition
Temperature
[ºC]
Flash
Point
[ºC]
Flammable
Limits,
Lower /
Upper [%]
Ethyl Acetate (EA) [37] [38] 77 427 -3 2.2 / 9
Dimethyl Carbonate (DMC) [37]
[38] 91 458 16 4.22 / 12.87
Ethyl Methyl Carbonate (EMC) [37]
[38] 110 440 24 -/-
Diethyl Carbonate (DEC) [37] [38] 126 445 25 1.4 / 14.3
Ethylene Carbonate (EC) [37] [38] 248 465 143 3.6 / 16.1
Propylene Carbonate (PC) [37] [38] 242 455 132 1.8 / 14.3
Gasoline [39] 30 to 210 > 350 < -40 1.4 / 7.6
Diesel [40] >180 240 >61.5 0.7 / 5
3.2.2 Separator
The separator prevents the positive and negative electrode from contacting each other while
enabling as many conducting ions as possible to flow through it. From a safety point of view,
the former is very important. If the separator would be breached or contracts significantly,
there is a risk that an internal short-circuit materialises. Weber et al. [41] therefore stress
that separators must possess high strength characteristics, negligible thermal expansion
and high melting point.
LIBs with organic electrolytes typically use microporous separators [42]. These can be
fabricated from material such as polyethylene (PE) and polypropylene (PP) [43]. These
20
© RISE Research Institutes of Sweden
types of separators have a melting point around 125-130 ºC and 155-160 ºC, respectively
[37] [44]. If the separator melts, the barrier between the positive and negative electrode is
breached and an internal short circuit occurs, which may trigger a large heat output followed
by uncontrollable chemical reactions and generation of massive amount of gas which could
result in a cell case explosion if not vented [32] [44]. Separators may also be ceramic or
composite based. This material offers improvement in terms of mechanical strength,
thermal resistance, performance and cell life [41]. According to Nesler et al. [45] this
technology needs more time to develop before it can be commonly used for EVs.
3.3 Lithium-Ion Batteries in Road Vehicles
Lithium-ion batteries are the preferred energy storage solution for modern EVs. Their
unmatched properties such as high cycle life, high energy density, and high efficiency makes
them very suitable for automotive applications [1]. They can be small and be used for start-
stop systems in EVs, or they can be much larger and used to power the powertrain.
Large battery packs are usually found in PHEVs and BEVs whereas HEVs require less
energy capacity and thus fewer batteries. In this section the focus is vehicles that are
designed to fit large battery packs. It is important to consider this as the examples given
may not necessarily apply to, or be relevant for, HEVs.
3.3.1 Lithium-Ion Battery Packs, Modules and Cells
When speaking of LIB in the automotive industry there are several distinct levels of
components that need to be understood. These are shown in Figure 13. The most basic level
is the lithium-ion cell. Consumer devices such as smartphones usually consist of a single
battery cell. Their voltage is thus restricted to what one cell can provide, i.e. roughly 4 V.
A much greater amount of stored energy can be obtained by connecting battery cells, and
modules, together in series or parallel. LIB cells for automotive applications are normally
connected together, in series and/or parallel, to form a module. The number of cells per
module varies, but generally adds up to less than 60 V per module. Voltages greater than 30
VAC or 60 VDC are considered harmful for humans and defined as high voltages within the
vehicle industry [21]. Restricting the voltage of battery modules is thus beneficial from a
handling and shipping perspective. Finally, the battery modules are connected to form
battery packs to meet the needed energy and power. Note that in some systems, several
battery packs are coupled together to create the whole battery system. In doing so,
applications such as passenger cars, heavy vehicles and electric ships can reach capacities
around 10-100, 10-400 and 500-4000 kWh, respectively.
21
© RISE Research Institutes of Sweden
Figure 13 General construction of a battery pack.
3.3.2 Passenger Cars with Lithium-Ion Batteries
Many battery cells need to be integrated into an EV in order to achieve the needed power
and energy. The overall goal in EV design is to achieve the largest possible battery pack while
maintaining an appropriate safety level.
A common approach is to install the battery pack inside stiffened and reinforced
compartments or areas less prone to be affected in crash conditions [46], see Figure 14 and
Figure 15. The latter can be referred to as the “safe zone” of a passenger car [47]. This zone
normally considers the area in the center of the chassis, between the wheelbase. By
integrating the LIB pack in this area, automotive manufacturers aim to eliminate the
possibility that the battery is affected by crash or impact conditions.
Figure 14 “Safe-zone” based on [48]
Figure 15 Battery layout for a Nissan Leaf [49]
For passenger cars there are three main configurations in which the “safe-zone” is utilized.
Most common are the “Floor” and “T” configurations [50] where the battery is distributed
in a square or rectangular area, as the one shown in Figure 16 or arranged in the shape of
the letter “T” as seen in Figure 17. The third option can be referred to as the “Rear” solution
illustrated by Figure 18. Here the battery pack is in the rear of the vehicle and in some cases
stacked upwards.
Figure 16 The “Floor” solution
Figure 17 The “T” solution
Figure 18 The “Rear” solution
22
© RISE Research Institutes of Sweden
The “floor” type uses all of the available space in the “safe zone”. The entire battery pack is
located underneath the passenger compartment. This provides more interior space for
passengers and luggage but also allows for high energy storage. One of the drawbacks of this
arrangement is that there is less ground clearance and that there is a larger target for ground
debris [50]. See Table 8 for an overview of several EVs that consider the “Floor”
configuration.
Table 8 Selection of EVs that employ the “Floor” solution to integrate their battery packs.
Nissan Leaf, EV Type: BEV
Figure 19 and Figure 20 show the “floor” battery pack in the Nissan Leaf. The pack varies in its
shape as more of the battery cells are placed underneath the front and rear seats. This model
employs pouch cells in its battery pack. These flat cells are oriented horizontally, like a stack of
paper, in the thinner sections of the pack. Underneath the seats they come up higher, as they are
oriented vertically, like paper in a filing cabinet.
Figure 19 Nissan Leaf, copied from [51].
Figure 20 Battery pack, copied from [52]
Tesla Model S, EV Type: BEV
The configuration found on Tesla’s is particularly flat in comparison to other vehicle models.
Tesla refers to their solution as a “skateboard” battery pack. This thin pack ensures maximum
available interior space.
Figure 21 Tesla Model S, copied from [53]
Renault Zoe, EV Type: BEV
Figure 22 and Figure 23 show the battery pack of the Renault Zoe. This pack is located
underneath the floor of the passenger compartment. The total capacity of this pack is 41 kWh at a
weight of 300 kg [54], roughly 20% of the total weight of the vehicle.
Figure 22 Renault Zoe ZE40, copied from [54].
Figure 23 Battery pack, copied from [55]
23
© RISE Research Institutes of Sweden
BMW i3, EV Type BEV/PHEV
The “Floor” solution for the BMW i3 may be seen in Figure 24. There are different versions
available of this model, with one of them being a PHEV (or REEV). Normally, vehicles have their
fuel tanks in between the rear wheels. As shown in Figure 25, this is not the case here. The fuel
tank, indicated by the red arrow, is placed in front of the battery pack.
Figure 24 BMW i3, copied from [56]
Figure 25 REEV, copied from [57]
The “T” solution arranges the battery modules in a T-shape within the safe zone, as
illustrated by Figure 17. This configuration allows for greater clearance between the ground
and the battery pack. This is achieved by reducing the passenger area. It is rather narrow
and usually protected by the front axle of the vehicle [58]. This ensures protection of the
battery pack against frontal collision and side impact [50]. Several EVs with the T-shape, or
similar configuration can be seen in Table 9.
The “Rear” solution makes use of the available space between the rear wheels of the
vehicle. Typically, this type of configuration is found in small vehicles or hybrids, as they
require less storage capacity. To increase the available energy, some EVs make use of the
space behind or above the rear wheels. A selection of EVs that follow this configuration is
seen in Table 10.
24
© RISE Research Institutes of Sweden
Table 9 Selection of EVs that employ the “T” solution to integrate their battery packs.
Volkswagen e-Golf, EV Type: BEV
Volkswagen combines a T-shape together with the space underneath the seats and floor for the
battery pack in the Volkswagen e-Golf. This pack has an energy capacity of 24.2 kWh [59] and
may be see in Figure 26 and Figure 27. This battery pack makes up a large portion of the vehicles
total weight, namely 20 %.
Figure 26 Volkswagen e-Golf, copied from [60].
Figure 27 Battery pack, copied from [59].
Chevrolet Volt / Opel Ampere, EV Type: PHEV
The Chevrolet Volt (Opel Ampere in the EU [61]) may be seen in Figure 28 and Figure 29. The
battery pack itself consist of vertically arranged pouch cells (e.g. paper in a filing cabinet).
Figure 28 Chevrolet Volt, copied from [62].
Figure 29 The battery pack, copied from [63].
Volvo XC60, EV Type: PHEV Mitsubishi Outlander, EV Type: PHEV
The battery pack in the Volvo XC60 PHEV is a
variant of the “T” solution. In this case one part
of the “T” is made up of the battery pack, and
the other of the fuel tank, see Figure 30.
The configuration used in the Mitsubishi
Outlander, seen in Figure 31, follows that of
the Volvo XC60. Its design is less
linear/rectangular, but it follows the same
principle. That is that the “T” is made up of the
battery pack and fuel tank combined.
Figure 30 Volvo XC60 PHEV, copied from [64]
Figure 31 Mitsubishi Outlander, copied from
[65].
25
© RISE Research Institutes of Sweden
Table 10 Selection of EVs that employ the “Rear” solution to integrate their battery packs
Chevrolet Spark, EV Type: BEV
As seen in Figure 32, the battery pack is located around the rear axle. The modules are positioned
in a way that results into two modules being located underneath the rear seating area and two of
them protruding from below the rear of the car booth, see Figure 33.
Figure 32 Chevrolet Spark, copied from [66]
Figure 33 Battery pack, copied from [67]
Mitsubishi Colt EV, EV Type: BEV
The battery pack for this vehicle is indicated by the arrow in Figure 34. It is positioned slightly in
front of the rear axle.
Figure 34 Mitsubishi Colt EV, copied from [68]
Volkswagen Passat, EV Type: PHEV Kia Niro, EV Type PHEV
In the cases seen in Figure 35 and Figure 36, a short yet wide battery pack is used. The pack itself
is mounted in between the wheels, close to the rear axle. The fuel tank of these vehicles is
installed closely behind this pack.
Figure 35 Volkswagen Passat, copied from [69].
Figure 36 Kia Niro PHEV, copied from [70]
26
© RISE Research Institutes of Sweden
As seen in Table 11, different passenger car manufacturers consider different types of
chemistries and battery cell types. In general battery chemistries are considered that
provide a balance between energy and power density as well as safety. It is interesting to
note that many of the considered vehicles employ blended cathodes.
Table 11 Summary of different LIB pack configurations for BEV and PHEV passenger cars.
Passenger
cars: BEV
Battery Pack Battery Cell
Energy
Cap.
[kWh]
Configuration Type Chemistry [Anode/Cathode]
Nissan Leaf
(2015) 30 [71] Floor [72] Pouch [71] C/LMO-NCA [71]
Renault Zoe
(2017) 41 [54] Floor [54] Pouch [71] C/NMC [71]
Volkswagen
e-Golf (2016) 36 [71] Floor / T-
shape [59] Prismatic [71] C/LMO-NCA-NMC [71]
BMW i3
(2017) 33 [71] Floor [73] Prismatic [71] C/LMO-NCA-NMC [71]
Tesla Model
S (2012)
60-100
[71]
Skateboard
[53] [74] Cylindrical [71] C/NCA [71]
Mitsubishi
Outlander
(2015)
12 [75] Floor [65] Prismatic [75]+
[76] C/LFP [75]
Volkswagen
Passat GTE
(2015)
9.9 [77] Rear [69] Prismatic [59] -/-
Volvo XC60
(2017) 10.4 [78] Linear [64] Pouch [79] NMC [79]
Volkswagen
Golf GTE
(2015)
8.7 [80] Rear [81] Prismatic [59] C/LMO-NCA-NMC [82]
Kia Niro
(2017) 1.56 [83] Rear [70] Pouch [84] -/-
Chevrolet
Volt (2016) 18.4 [85] T-shape [63] Pouch [84] +
[85] C/LMO-NMC [85]
3.3.3 Heavy Vehicles with Lithium-Ion Batteries
Heavy vehicles such as buses and heavy trucks are also being electrified. Their layout and
design with respect to their ability to protect the battery in traffic accidents is presented in
this section. This general understanding is needed to identify hazards associated with
damaged heavy EVs.
3.3.3.1 Buses
Buses do not necessarily follow the configurations presented for passenger cars. Rather than
integrating the battery pack underneath the vehicle, bus manufacturers such as Volvo Bus,
Solaris, BYD and VDL opt for placing them on top of their vehicles. This is shown in Figure
37 and Table 12. Placing the battery on top of the vehicle requires fewer modifications to be
made to existing buses. It also facilitates movement of passengers and optimises the
27
© RISE Research Institutes of Sweden
occupant space. Other benefits include the fact that the batteries are easier exposed to air,
allowing them to be cooled by the moving vehicle, and are more easily accessible for certain
charging systems.
There are however some drawbacks of this strategy. Placing relatively heavy battery packs
on top of a vehicle makes it more difficult to obtain a low centre of gravity. In addition, roof
mounted solutions require protection from debris and moisture accumulation. This needs
to be considered, as was illustrated by a recall of certain bus models in the US in 2011 [86].
Some buses do integrate the battery pack underneath the passenger space. An example of
this is the Proterra Catalyst. Their battery pack is located below the floor of the bus as also
seen in Table 12. In doing so this bus model can integrate enough batteries to obtain energy
capacities of up to 660 kWh [87].
Chinese electric buses are also commonly equipped with a large number of batteries to
achieve high energy capacities. An example of this is the BYD K9. This bus has been present
in Europe since 2013. Its configuration is intended to supply enough energy storage capacity
for full-day operation. They do not consider a “floor” configuration, instead they achieve a
high capacity by integrating several different battery packs throughout the vehicle as seen
in Figure 37 and Table 12.
The Volvo, VDL and Solaris buses reserve less space for their battery packs. As a result, their
energy capacity is less than the BYD K9 and Proterra Catalyst. To sustain their operation,
they rely on opportunity charging at e.g. bus-stops. One benefit of having fewer batteries is
that the vehicle carries less weight. This can allow for lighter construction and greater
efficiency.
The Optare Versa has its battery pack in the rear of the vehicle as also seen in Figure 37 and
Table 12. This is a relatively simple installation when compared to the roof mounted option,
as that method requires special fixtures and equipment.
Table 12 Battery packs in electric buses
Figure 37 Position of the battery packs on selected buses
BYD K9 A+C+E [88]
Volvo 7900 C [89]
VDL Citea B [90]
Solaris Urbino B [91]
Optare Versa D+E [92]
Proterra
Catalyst F [93] [94]
28
© RISE Research Institutes of Sweden
The types of batteries that are considered by the buses discussed in this section are
presented in Table 13. Note that LFP chemistries appear to be relatively common for buses.
LIBs of this chemistry have a lower energy capacity per kg than other chemistries such as
NMC, which is common for electric passenger cars. There is however more space available
on buses, hence this plays less of a role. The use of LFP cells allows them to reap the benefits
of a more stable battery chemistry while still being able to achieve high energy and power
densities.
Table 13 Selected electric bus models currently operating in Europe and their characteristics.
Buses:
BEV or
PHEV
Battery Pack Battery Cell
Energy Capacity
[kWh] Configuration Type Anode/Cathode
Volvo
7900
76 [95]
150 - 250 [96] Roof (rear) [89] - -/LFP
BYD K9 216-345 [97] Roof (rear) + rear
and front [88] Prismatic [98] -/LFP [99]
Solaris
Urbino 80-240 [91] Roof (front) [91] Pouch [100] LTO/- [101]
VDL Citea
60-250 [90] Roof (front) [90]
Prismatic
[90] + [102]
or Pouch
[90] + [103]
LTO/-
or -/LFP [90]
Optare
Versa 92-138 [104] Rear [92] Cylindrical
[105] + [106]
-/Lithium Iron
Magnesium Phosphate
[92]
Proterra
Catalyst
94 -440 (35 ft.) [107]
94 -660 (40 ft.) [87] Floor [94] - -
3.3.3.2 Heavy Trucks
There are not a lot of heavy trucks with lithium-ion batteries on the market yet. Therefore,
only limited data is available on how lithium-ion battery packs are integrated, see Table 14.
Contrary to buses, the placement of battery packs in heavy trucks appears to be more
restricted. To give an example, consider the Scania L 320 6x2 PHEV [5] heavy truck. Here
the battery pack is located behind the front wheel axle on the side of the driver. A similar
configuration may be found in the electric heavy trucks that were announced by DAF this
year [6]. Their press release images [21] show that the two battery packs used in the full
electric models are located behind the front axle. One of them is located on the driver side
and the other on the passenger side, see Figure 38. The hybrid DAF LE Hybrid has a single
battery pack. In this case the fuel tank and battery pack are mounted on opposite sides of
the driveshaft.
Lithium-ion batteries may potentially be integrated in truck trailers in the future. Some
companies are working on developing truck trailers with solar panels. Their idea is to store
excess energy produced by these panels in lithium-ion batteries [108]. This energy can
then be used e.g. to power refrigerated trailers.
29
© RISE Research Institutes of Sweden
Table 14 Selected heavy truck models and their battery pack characteristics.
Heavy Trucks: BEV or
PHEV
Battery Pack
Energy Capacity [kWh]
Configuration
Scania L 320 6x2 [7] 18.4 (limited to 7.4) Behind front wheel axle, left side of
the vehicle.
DAF LF Electric [8] Up to 222 -
DAF CF Electric [8] 170 Behind front wheel axle, both sides of
the vehicle
DAF CF Hybrid [8] 85 Behind front wheel axle, left side of
the vehicle.
Volvo FL Electric [9] 100 - 300 -
Volvo FE Electric [10] 200 - 300 -
Mack LR Electric [11] Unknown -
Volvo Vera [12] 300 [109] -
.
Figure 38 Potential placement of battery packs in heavy trucks.
30
© RISE Research Institutes of Sweden
4 Fire Risks Associated with Lithium-
Ion Batteries
As more LIB powered vehicles become operational across the globe, their involvement in
traffic incidents is likely to rise as their presence on the road increases. There is a chance,
as in conventionally fuelled vehicles, that the energy stored on-board can become a danger
to the safety of those involved in an incident. The risks associated with conventional vehicles
are well-defined and generally acceptable by society; however, time and education are
needed to achieve this comfort level for LIB powered EVs.
Videos and news reports of fire and smoke shooting out of phones and laptops as well as
hoverboards while being ridden or while being charged have given LIBs notoriety. These
cases clearly illustrate what can happen to LIBs when there are limited systems in place that
warrant their safe operation. Recently a study was performed in the Netherlands by the
Food and Consumer Product Safety Authority on the fire safety of hoverboards [110]. Here
significant safety lapses were identified among 30 different types of hoverboards. Some of
these products lacked temperature regulation, had limited fire-resistance housing or
allowed its LIB to be charged indefinitely. Simply charging such LIBs can lead to fire.
4.1 Thermal Runaway
The primary safety concern with LIBs originates from the individual battery cells that make
up the battery pack. The battery cell may release gas when abused, which can ignite or cause
an explosion. Abuse conditions are met when the safe operating window is not kept, as is
illustrated in Figure 39. Once the battery’s voltage or temperature limits are exceeded,
certain chemical reactions may be triggered inside the battery [44]. This may lead to an
internal short circuit or increase of the internal temperature by other mechanisms. The
battery cell can subsequently fail by venting flammable gas, burn, explode or become a
projectile.
Figure 39 Illustration of the limited window of operation for a LIB cell.
The hazardous events arise when certain mechanisms are triggered. This behaviour is due
to the components that make up the LIB, as there is a combination of flammable fuel,
potential oxidisers and heat generation during usage. When exothermic chemical reactions
31
© RISE Research Institutes of Sweden
are generating more heat than is being dissipated, the LIB enters a so-called thermal
runaway [44]. Thermal runaway is triggered by a chain of chemical reactions inside the
battery resulting in accelerated increase of internal temperature, see Table 15. Specifically,
decomposition of SEI (Solid Electrolyte Interface) layer4 and reactions between electrolyte
and anode is followed by melting of the separator and breakdown of the cathode material.
The outcome can be that of complete combustion of the LIB accompanied by the release of
gas, flying projectiles and powerful jet flames [37].
Doughty and Monitor [111] classify these events leading to thermal runaway in several
stages. First the onset of heating is triggered, which corresponds to the decomposition of
the SEI layer at the anode. The rate of self-heating is still controllable at this point and is
practically defined as 0.2ºC/min by Doughty and Monitor. However, if this heat is not
dissipated further reactions will be triggered that accelerate self-heating. This is referred to
as the acceleration stage. The final stage is that of thermal runaway. Doughty and Monitor
characterise this as the point where a self-heating rate of 10ºC/min or greater is obtained.
Note that the point at which this event is triggered is strongly dependent on the battery
design, structure and material.
Table 15 Self-heating and decomposition reactions of LIBs.
Process
Onset
Temperature
[ºC]
Notes
ONSET STAGE
Decomposition of
SEI layer at anode
80-120 [112]
80-100 [33]
> 70 [113]
• Determines the minimum temperature
where chain-like thermal
decompositions are irreversibly triggered
[112] [111].
• Self-heating rate of 0.2ºC/min [110].
• Highly dependent on the electrolyte salt
used [33]. The data presented considers
electrolyte with LiPF6 as these are most
common.
ACCELERATION STAGE
Reaction of the
lithiated anode with
organic solvents in
the electrolyte after
decomposition of
SEI layer
> 1105 [113]
• Temperature rise may be up to 100 ºC
[114]
• Flammable hydrocarbon gases (ethane,
methane and others) are released [115].
Separator starts to
melt [37] [44]
> 125 (PE)
> 155 (PP)
• This causes an internal short circuit and
further increases the self-heating rate.
Reaction between
intercalated lithium
and binder6
> 160 [113]
• Only occurs if there is anode material left
to react with [114].
• Temperature depends on the considered
binder material. [113]
4 The interface between electrolyte and current collectors. This is where electron exchange occurs.
5 If using carbon-based anode.
6 Binder materials bind the active material particles and current collector together [249].
32
© RISE Research Institutes of Sweden
Process
Onset
Temperature
[ºC]
Notes
RUNWAY STAGE
Decomposition of
the cathode
material.
LFP > 140 [26],
218 [116],
212, 287 [117],
• Usually the main source of heat
generation and cause of thermal
runaway [112].
• The heat of reaction varies greatly. Xiang
et al. recorded a range of 35 to 458 J/g
for different cathode materials between
50-225ºC [116].
• Releases oxygen [115]. Higher charge
level increases the amount of oxygen
released.
LCO > 168 [116]
LMO > 110 [116],
> 190 [113]
NMC > 212 [117]
NCA > 183 [117],
139 [118]
Decomposition of
electrolyte solvents
> 180 [113]
> 202 [116]
• Exothermal reactions. The heat of
reaction comprises 258 J/g between 50-
225ºC [116].
COMBUSTION
Combustion of
solvent [37] [38]
Autoignition >
427
Flashpoint > -3
• The released oxygen facilitates the
required conditions for the combustion
of flammable organic electrolytes [119].
• Flashpoint ignition requires an ignition
source, e.g. a spark or flame from the
LIB.
Combustion of
solids Varies
• Contribution of plastic oxidation in fire
calorimetry tests was estimated equal to
that of the electrolyte in terms of heat
release [120].
• Highly charged LIBs are a big safety
concern due to combustible lithiated
anode materials [119].
• Some ignition data of solids may be
found in [121].
4.2 Battery Failure Causes
The catastrophic loss of a cell can result in even more severe consequences such as damage
to other system elements, and/or human injury or death. Failure of a cell may be the result
of poor cell design or manufacturing flaws, external abuse (thermal, mechanical, electrical),
poor battery assembly design or manufacture, poor battery electronics design or
manufacture, or poor support equipment (i.e. battery charging/discharging equipment)
design or manufacture. The primary battery risks are generally a result of external or
internal short circuits, high or low temperatures, overcharge or over-discharge. These
mechanisms can result in exothermic reactions within the battery. When temperatures
become sufficiently high, or there is an ignition source present that ignites the flammable
gases released by the battery, the fire triangle seen in Figure 40. is completed.
33
© RISE Research Institutes of Sweden
Figure 40 The fire triangle for lithium-ion batteries.
4.2.1 Internal Cell Short Circuit
The most hazardous failure cause is that of an internal cell short circuit [122]. This
catastrophic event may occur very suddenly and without previous warning. This can be a
result of manufacturing defects or physical damage due to dendrite growth or mechanical
deformation [122] [37]. When the internal short circuit occurs, the resulting damage is often
severe. The cell discharges its energy through the short circuit. When electric current passes
through conducting material, it produces heat. This mechanism may be referred to as Joule
heat generation. In this local area, the rapid heating can trigger further self-heating and
thermal runaway [123] [122].
That internal short circuit raises the most concern is also said by Ahlberg Tidblad [124]. It
is made clear that this is particularly disturbing when taking into consideration that this
type of failure occurs in batteries that comply with industry standards. This is due to
manufacturing errors, such as burrs, misalignment of the electrode package or punctured
separators. The primary cause relates to the presence of particles in or on the cathode [124].
Zhao et al. [122] studied the behaviour of large format LIB cells, i.e. those used for
automotive applications, and their behaviour during an internal cell short circuit. They
explain the mechanism as creating a current loop within an electrode layer where the short
circuit is found. When the loop is formed, energy is discharged through this electrode layer,
however, this also stresses all other layers, which generate a large amount of current due to
the short. This heat up the complete battery cell.
Santhanagopalan et al. [125] present four probable types of internal cell shorts. That is when
there is contact between negative current collector to positive current collector, negative
current collector to cathode, positive current collector to anode and cathode to anode. These
are classified into the different types given by Figure 41.
The third type, Type 3, is the most hazardous [125]. The anode material has namely low
resistivity compared to the cathode, which allows for high current flow. This means that a
lot of heat will be generated at the anode. Simultaneously, the onset temperature for self-
heating reactions are lowest at the anode, as was discussed in Chapter 4.1. These factors
34
© RISE Research Institutes of Sweden
combined are thus most likely to trigger self-heating mechanisms which can lead to thermal
runaway.
The remaining short circuit types pose less of a threat according to Santhanagopalan et al
[125]. Type 1 does result in a large amount of heat being generated, increasing the external
cell temperature up to 100ºC. However, the current collector materials are good conductors
of heat, meaning that the generated heat can be dissipated fast enough to prevent further
reactions. Type 2 has the lowest amount of localised heating of all types. This is not enough
to trigger any self-heating mechanisms. Finally, Type 4, is the most likely internal short
circuit type to occur in a battery’s life. However, the resulting current flow is low and is thus
not considered a major threat. The result will namely be restricted to a small temperature
rise above ambient temperature. It is important to keep the duration of these internal short
circuit events in mind. For example, even Types 1, 2 or 4 may trigger a thermal runaway if
they are sustained over a long period [125].
Figure 41 There are four different types of internal short circuit paths possible. Not all of them are
equally hazardous [125].
4.2.2 Mechanical Deformation and Impact
Mechanical deformation may also initiate an internal short circuit and potentially result in
fire, see Figure 42. Severe deformation may be a result of certain crash or ground impact
conditions. Severe deformations of the battery pack must be avoided. The high voltage
system may be damaged, causing short circuits and arcing and it may also result in the
leakage of flammable and conductive liquids. According to Trattnig and Leitgeb [46] the
worst-case scenario in a car crash would be the combination of venting gases or leaking
fluids with ignition sources such as electrical arcs or hot surfaces. This could lead to a rapid
scenario that must be delayed for the, potentially trapped, passengers to escape the vehicle
safely.
35
© RISE Research Institutes of Sweden
The severity of the outcome of an internal short circuit, resulting from crash conditions,
depends on a multitude of factors. It involves the interaction between mechanical contact,
heat generation and electrical discharge which may or may not result in thermal runaway
[126]. This was discussed in Section 4.2.1.
Figure 42 Mechanical deformation leading to thermal runaway [126]
Battery packs are usually placed in reinforced and stiff areas of passenger cars, see
Section 3.3. Zhu et al. note however that these packs are still vulnerable to penetration in
side collisions, small overlap crashes as well as penetration due to road debris impacts [127].
They also mention that forces from the rapid deceleration of the vehicle in a crash may be
high enough to result in an external short circuit, causing further damage.
There is not a lot of test data available on EVs that have been crash tested with their battery
pack. This can be motivated by the fact that testing this combination is accompanied by
many hazards for the test facilities. Safe handling and disposal of damaged battery packs is
not straight-forward either, as is discussed in Chapter 5. As such, physical testing is avoided
meaning that much of the data available is obtained from numerical simulations [127].
Xia et al. developed a general numerical model that models the indentation process of LIBs
due to ground impact [50]. Their study showed, among other things, that there is no
possibility that battery cells are damaged due to the impact of flying stones, e.g. gravel.
However, road debris with certain geometrical characteristics can perforate the battery
under certain conditions. They mention that it is almost impossible to fully prevent
penetration of the shield for all ground objects. Once the shield is perforated other layers
will fracture shortly after. This could put individual LIB cells in contact with the ruptured
shield or the road debris.
The EVERSAFE project provided insight into the impact resistance of EVs. This project was
funded by the EU and focused on determining the needed safety requirements for EVs. Part
of their work considered the response of EVs under certain crash conditions through both
physical and virtual testing [58]. Here, they created a model that simulated undercarriage
impact, based on the aforementioned work by Xia et al. [50], who considered a Toyota Yaris
EV with the “T” battery configuration. In their study, EVERSAFE considered the worst
possible conditions for ground impact. That is a “Floor” battery configuration, i.e. the
configuration with the lowest ground clearance, combined with the complete removal of the
vehicle’s front axle. They found that this configuration was indeed vulnerable to ground
36
© RISE Research Institutes of Sweden
impact, as significant loads were recorded inside the battery for certain impact sizes, shapes
and speeds.
The EVERSAFE project also identified and defined critical impact conditions and high-risk
conditions for EVs [58] [128]. Two scenarios were of particular interest with respect to the
battery, namely longitudinal and lateral impact. Of the longitudinal scenarios considered by
EVERSAFE, rear impact was determined to pose the highest risk due to limited legal
requirements which may result in that EVs without a fuel tank do not have to demonstrate
their crash safety for this crash scenario, which leads to that these EVs do not demonstrate
their ability to protect the battery pack in physical rear impact testing. Lateral scenarios
consider impacts to the side of a vehicle. These conditions are most likely to result in
deformation or intrusion of the battery pack and its protective structure. Of the different
side impact tests, side pole impact [129] was deemed most hazardous for EVs.
Another EU project, named OSTLER, performed the Euro NCAP side pole test [129] on a
Toyota Yaris EV as part of their work [130]. At a velocity of 50 km/h they found a significant
intrusion of the battery pack of 154 mm. The EVERSAFE project performed a similar test
on a first-generation Mitsubishi iMiEV at a speed of 35 km/h [128]. They observed no
damage to the battery pack and did not detect battery chemicals or gases.
In addition, Justen and Schöneburg from the Mercedes Car Group presented results from a
crash safety assessment of their hybrid- and electric vehicles [48]. Although they found
major battery intrusions during crash testing there was no thermal or electric reactions
resulting in no fire or explosion. In Chapter 5.1 documented incidents resulting in fire are
presented. There are also examples of real incidents with high force collision impact without
fire [131].
Note that the cases discussed in this section primarily consider passenger cars, as most
available information considers those cases. Studies concerning the crash behaviour of LIBs
in heavy vehicles such as busses and heavy trucks could not be identified.
4.2.3 Charge
LIBs are designed to receive and store a certain amount of energy over a specific amount of
time. When these limits are exceeded, as a result of charging too quickly or overcharging,
the cell performance may degrade, or the cell may even fail.
The charge level of batteries is normally defined in terms of state of charge (SOC). Their
operational limits may be defined from 0-100%, which means that a battery at 100% SOC
is considered fully charged to its rated capacity. However, full capacity of the battery
normally goes beyond its rated capacity, both at upper and lower limits.
Overcharging may be realized when the cell voltage is incorrectly detected by the charging
control system, when the charger breaks down or when the wrong charger is used [44].
When overcharging, the anode material can become overly lithiated. As a result, lithium
intercalation ceases and lithium metal deposits on the anode. These deposits may grow into
metallic fingers commonly referred to as dendrites. As they grow, they can reach the point
where they penetrate the separator and cause an internal short circuit [132]. The opposite
happens at the cathode. Here overcharging may result in it becoming de-lithiated to the
point where the cathode decomposes thermally and generates heat.
37
© RISE Research Institutes of Sweden
Brand et al. considered the onset of self-heating due to overcharge abuse of four battery cells
[117]. They found that the cells which considered LFP cathode and C anode material were
less resistant to overcharge. When they were fully charged and slightly overcharged, 100 %
SOC and 105 % SOC, respectively, self-heating mechanisms were triggered. Other cell types,
including NMC and NCA with carbon anodes, were also tested. These were more resistant
to overcharge as self-heating occurred at 135 % SOC and 130 % SOC, respectively.
When electric current passes through conducting material, it produces heat so called Joule
heat. This means that high current, which can be associated with faster charging rates,
increases the heat that is generated inside the battery cell. At a high enough current level
there is a risk that the battery cell easily fails [44]. Too high charging voltage can also lead
to the destabilisation of the cathode structure which may lower the temperature at which
the cathode starts to decompose.
The effect of the overcharge conditions, i.e. charging at high charge rates, was demonstrated
by Tobishima and Yamaki [44]. They found that at high charge rates of 2C7 the safety vent
and anode cap housing would open simultaneously, with the cell exploding. Overcharge
tests were also performed by Larsson et al. in [133] and [134]. In the former study, one out
of four LFP cells that were overcharged with 2C resulted in fire. Wang et al. [115]
summarised the outcome of several overcharge abuse tests. They mention that in general,
abuse can occur when charging at 0.5C and above.
Low temperature charging, e.g. below 0°C, should be avoided to prevent fast initiation and
growth of lithium dendrites capable of forming internal short circuits. Recall that during
the charging process, lithium-ions move from cathode to the anode. They are then stored in
the layered structure of the anode. Charging at low temperatures affects this kinetic process
within the LIB cell. As a result, the lithium-ions may form metallic lithium instead of
intercalating into the anode. These quickly initiate dendrites [135]. In turn this can cause
internal short circuits.
4.2.4 Discharge
When the LIB is discharged, lithium-ions flow from the negative current collector and anode
to the positive current collector and cathode. If the level of discharge becomes too great
however, the negative current collector, which consists of copper, can dissolve. As a result,
small conductive copper particles are released in the electrolyte which increase the risk for
an internal short circuit [132]. It can also lead to the evolution of hydrogen and oxygen, cell
venting and plating on the cathode.
Overdischarge abuse occurs when discharging battery cells below their minimum voltage.
In the unlikely event where four battery cells are in series, and one of them is completely
discharged (0 V), this could lead to the empty cell being discharged even further [117]. In
this case the polarity of the cell reverses. Brand et al. considered this scenario in their study
of over-discharge abuse on C/LFP, C/NMC and C/NCA cells. They discharged the batteries
from 100 % SOC at a 1C rate but did not measure significant temperature increases (max.
47.5ºC) or observe damage to the cell casing.
7 This refers to the charge and discharge rate of the battery. A 2C charge rate means that the current for
charging is twice as high as the batteries capacity to store electrical charge. 1C is the current needed to fully
charge the battery in one hour.
38
© RISE Research Institutes of Sweden
Overdischarge abuse tests on C/NMC pouch cells with a capacity of 25 Ah were performed
by Guo et al. [136]. They identified the different stages of failure during overdischarge
conditions. At -10% SOC (of full capacity, which means reversed polarity of the cell) the SEI
layer on the anode began to decompose, followed by the dissolution of the copper current
collector at -12% SOC. Charge levels below -12% resulted in internal short-circuits, where
their intensity increased with decreasing charge levels. Guo et al. also mention that this risk
is greater when battery cells that are connected in series [136].
Overdischarge can occur when discharging a battery where the charge levels of its individual
cells is not in balance. Normally safety systems are in place to prevent this. However, it is
still possible that this occurs if these safety systems fail and the battery is misused [132]. In
case it has been stored for long periods of time so that self-discharge has an effect, charging
may cause problems if individual cells reach too low SOC. However, self-discharge cannot
by itself cause overdischarge in the sense of reversed polarity.
4.2.5 External Short Circuit
An external short circuit is another form of electric abuse that may destabilise the battery.
This event may occur in case the battery is exposed to, for example, severe mechanical
deformation and impact, immersion in water, corrosion and electric shock during
maintenance.
The response of stainless-steel prismatic C/LCO cells when exposed to an external short was
investigated by Leisner et al. [137]. They observed a very high current peak and an internal
cell temperature of 132°C for a C/LCO cell. Note that these cells were not equipped with
current limiting or temperature trip safety devices.
External short circuit tests were performed by Davidsson et al. on three different cell types
[138]. This was achieved with a contactor that was limited to 10 000 A. Short circuit of a cell
with hard-plastic packaging material corresponded to an initial current of 3200 A being
registered. The pressure inside the cell then increased significantly and the cell burst into
pieces. A pouch cell, with a metal foil enclosure, expanded significantly after an initial
current of 1800 A followed by cell rupture. The last battery, with metal casing, was not
affected by the short circuit. No activity was observed after the initial current of 200 A was
measured. It is unclear whether the considered cells had built-in fuses or safety vents.
Wang et al. summarised the results of external short circuit tests [115]. The test method
considered connecting a resistor across the terminals to allow current flow to heat up the
considered battery cell. They mention that although there is internal heating, there is also
significant heat dissipation of the external circuit. They did not mention whether this was
enough to prevent self-heating mechanisms from being triggered.
Larsson et al. performed external short circuit abuse testing on LIB cells [133], [134]. In the
former test the cell expanded 20 to 30 seconds after the short circuit had been initiated.
Then the measured current dropped while the cells ventilated for 2 minutes. External cell
temperatures of up to 100ºC were recorded followed by discharge to 43% SOC. The terminal
tabs burnt off during this test for one of the considered battery cells and thus broke the
external short circuit.
The external short circuit resistance of independent and series connected 10Ah pouch cells
was studied by Kriston et all. [139]. Short circuit was initiated by connecting the battery
39
© RISE Research Institutes of Sweden
terminals using different external resistances. They classified the behaviour that followed
into three stages. First high currents are recorded. This is followed by a current drop,
increase in cell temperature, vaporisation of electrolyte, pressure build-up and venting of
the cells. Finally, as the active material discharges, the current drops. Note that thermal
runaway or the release of significant smoke was not observed for the studied cells. The
reader is referred to Kriston et al for videos and detailed images of the tests [139].
4.2.6 Exposure to High Temperatures
One of the limiting factors of LIB cell safety is its thermal stability. When exposed to high
temperatures internal degradation mechanisms and exothermic reactions may lead to
problems. When the external temperature of the battery is higher than the internal
temperature, it is heated instead of cooled. Once the battery warms up to certain
temperature levels, decomposition mechanisms are triggered causing the battery to
generate further heat. As shown previously in Table 15, the true problem then arises when
the Runaway Stage is reached.
Resistance to high external temperatures may be assessed by external heating in oven or by
an external fire. Larsson et al. considered external heating by oven in [133], [134] and [140].
Here LIB cells were mounted in an oven that was heated to 300°C in a set amount of time.
In [140] this method was employed to assess hard prismatic LCO-graphite cells. This study
found that all cells underwent thermal runaway at temperatures above 190°C and were
releasing smoke and gas. Note that this temperature refers to the last point before the
temperature increases tremendously. For roughly half of the studied cases, accumulated
gases in the oven ignited and exploded. This occurred approximately 15 seconds after
thermal runaway was initiated. Another study by Larsson et al. [133] found that thermal
runaway of a cylindrical Samsung 18650 cell was observed at approximately 220°C. This
resulted in an immediate fire and an extreme rate of temperature increase. Furthermore,
shortly before thermal runaway, the cell discharged burning electrolyte. The same study
also considered LFP pouch cells. Here they observed no or very weak signs of thermal
runaway.
Instead of placing the battery cells in an oven, they can be exposed to external fire. In a
similar fashion to what was discussed before, this may trigger a thermal runaway event.
Larsson et al. studied this in [134], [141], [142], [143] by exposing different LIB cells to
propane burner.
The complete battery pack may also be exposed to an external fire. This could be the result
of fuel leak for example, which accumulates underneath the LIB pack and ignites. Over time
this heat may penetrate a battery pack, initiate cell failure, and spread further within the
pack. To mitigate this risk EVs must pass fire resistance testing, i.e. UNECE Reg. No. 100
[144]. The amount of time in which the battery pack is exposed to external flames is 2
minutes. This test is similar to the test conducted on gasoline tanks. In the test the size of
the fire is determined by the geometry of the battery or tank respectively. When there is no
evidence of explosion during these 2 minutes or the following observation period, where the
test object is to reach ambient temperatures or has its temperature decrease for at least 3
hours, this test can be considered passed. Note that the test may be performed on either the
full-scale level (EV), or component level (LIB pack). In the case of the former, recorded tests
have shown that a very high fire resistance can be achieved. The LIB pack has been found
to not contribute to the fire for 25-40 minutes when integrated in an EV. This resistance
40
© RISE Research Institutes of Sweden
drops when the battery is considered separately. Then the time may reduce down to 2-11
minutes [145] [146] [147] [148] [149] [150].
Exposure to high temperatures may also be the result of manufacturing faults such as loose
battery cell connectors. Beauregard investigated a PHEV destroyed by fire in 2008 [151].
They found that the likely cause of this event were loose connectors. In combination with a
vibrating vehicle, this led to the build-up of heat. In turn the battery cells short circuited
which eventually resulted in the vehicle burning down.
Finally, it is important to consider that there can be negative implications to raising the
ambient temperature of the LIB. Although this may not directly trigger negative reactions
it does reduce the safety margin. When close to the edge of this margin, internal short circuit
reactions that would not otherwise trigger further-self heating reactions may push a battery
cell over the edge [125].
4.3 Hazards and Risk Factors
When a battery does fail this may have several different outcomes, e.g. venting, fire or even
explosion. These different hazards have been classified by the European Council for
Automotive Research and Development (EUCAR), see Table 16. Here an explosion is the
most severe event. When heating LIBs their internal pressure builds up and eventually the
cell cracks and/or ventilates or explodes. It is cell explosion that is referred to in Table 16.
In addition, if the released gas can accumulate to create an explosive environment which is
ignited it leads to an explosion. This type of explosion is usually not addressed by battery
testing, except in some more recently developed tests.
In 2015, Hendricks et al. developed a comprehensive method of analysing the failure modes,
mechanisms, and effects (FMMEA) of LIBs [132]. The FMMEA produces a risk
prioritization number that combines the likelihood of occurrence, the severity, and the
detectability of the failure for a specific battery system. Their article included a resulting
table which summaries an FMMEA of LIBs focused on internal failure modes within a
battery cell.
The factors that affect the severity of these hazards are varying and complex. Among other
things, they can be linked to the battery chemistry, its charge level and the failure cause.
This section focusses on battery chemistry and charge level. In addition, the risk for failures
to propagate from one cell to the next is discussed.
41
© RISE Research Institutes of Sweden
Table 16 European Council for Automotive Research and Development (EUCAR) hazard levels and
descriptions [152]
Hazard Level Description Classification Criteria and Effect
0 No effect No effect. No loss of functionality.
1
Passive
protection
activated
No defect; no leakage; no venting, fire, or flame; no
rupture; no explosion; no exothermic reaction or thermal
runaway. Cell irreversibly damaged. Repair of protection
device needed.
2 Defect/damage
No leakage; no venting, fire, or flame; no rupture; no
explosion; no exothermic reaction or thermal runaway.
Cell irreversibly damaged. Repair needed.
3 Leakage,
∆ <50%
No venting, fire, or flame8; no rupture; no explosion.
Weight loss <50% of electrolyte weight (electrolyte =
solvent + salt)
4 Venting,
∆ ≥50%
No fire or flame8; no rupture; no explosion. Weight loss ≥
50% of electrolyte weight (electrolyte = solvent + salt).
5 Fire or flame No rupture; no explosion (i.e., no flying parts).
6 Rupture No explosion, but flying parts of the active mass.
7 Explosion Explosion (i.e., disintegration of the cell).
4.3.1 Chemistry
The thermal runaway and the heat and fire development in batteries varies with battery
chemistry. A study performed by Maleki et al. [153] concluded that exothermic reactions
between electrolyte and cathode material at elevated temperatures are the main
contributors to thermal runaway. Doughty and Pesaran [111] state that the order of thermal
stability for cathode materials follows LFP>LMO>NCM>NCA>LCO, in decreasing order. It
is important to note that thermal stability refers to the amount of heat that is generated per
unit time when exothermic reactions have been triggered. It does not reflect on the
temperature at which they are triggered.
Abuse testing by Larsson et al. [37] has shown that thermal runaway is initiated after the
temperature of the battery cell reaches 150-200ºC. They also showed that LIBs with an LFP
cathode has a less severe thermal runaway event than a LIB with LCO cathode [133].
Xiang et al. [116] investigated the thermal stability of LiPF6-based electrolyte9, both
independently and while being in contact with various cathode materials. They found that
the electrolyte yields strong exothermic reactions below 225ºC. Following this, they looked
at the LiPF6-based electrolyte in combination with several cathode materials. This showed
8 “The presence of flame requires the presence of an ignition source in combination with fuel and oxidizer in
concentrations that will support combustion. A fire or flame will not be observed if any of these elements are
absent. For this reason, we recommend that a spark source be use during tests that are likely to result in
venting of cell(s). We believe that “credible abuse environments” would likely include a spark source. Thus,
if a spark source were added to the test configuration and the gas or liquid expelled from the cell was
flammable, the test article would quickly progress from level 3 or level 4 to level 5” [152].
9 Found in the vast majority of commercial LIBs [37].
42
© RISE Research Institutes of Sweden
that LCO can release oxygen at elevated temperatures and further induce the combustion
reaction of LiPF6-based electrolyte.
Xiang et al. [116] also investigated cells with LFP cathodes. They found that this cathode
material can inhibit the decomposition of electrolyte and yield a less severe thermal
runaway event. Specifically, its reaction heat measured 35 J/g between 20-225ºC. In
comparison, the electrolyte by itself or together with either LCO or LMO cathodes resulted
in 258 J/g, 358 J/g and 308 J/g, respectively.
Xiang et al. [116] also found that the onset temperature for decomposition reactions of
cathode materials was highest for LFP, i.e. 218ºC. Other tested cathodes such as LCO and
LMO yielded onset temperature around 168 ºC and 110ºC, respectively. At 202 ºC,
polymeric products in the LiPF6-based electrolyte started to decompose. Note that LMO,
which is considered safer than LCO [111], was found to have a lower onset temperature. This
is because safety is often connected to thermal stability and not by onset temperature. Xing
el. al. [115] argues that the reaction heat released below 225 ºC is the key indicator for
thermal stability, which was found to be lower for LMO than LCO.
Brand et al. [117] recorded the onset temperatures for different LIB cells using accelerated
arc calorimetry. They found that self-heating with temperature rates higher than 5ºC/min.
occurred at temperatures of 212 ºC and 287 ºC for the LFP cells. The onset temperatures for
the NMC and NCA cells was found to correspond to 212 and 183 ºC, respectively.
From a fire and heat generation perspective, LFP is the preferred option. It may however
not be as favourable when considering the release of toxic gases or the risk for explosion.
Larsson argues that this may be the negative side-effect of the suppression effect that LFP
has [37]. The mixture of gases emitted from LIBs namely tends to be more toxic when it is
not burning. In addition, the gases can accumulate and experience a delayed ignition
resulting in a gas-explosion if it occurs in a confined place such as a room, building, parking
garage etc [154].
4.3.2 State of Charge and Cell Capacity
The capacity and state of charge (SOC) affects, among other things, the behaviour of a LIB
leading up to and during thermal runaway. Battery cells with high capacity, such as those
used for automotive applications, generate more heat when in use. This is due to the higher
current flow within the cell. This makes them more vulnerable as self-heating reactions will
be triggered faster, increasing the likelihood of thermal runaway [125]. At greater charge
levels, the extent of lithiation on the anode is much greater. This material is highly reactive
and has been shown to increase the likelihood of thermal runaway [125] [118].
Larsson et al. performed abuse tests on batteries with varying levels of charge in [134] and
[141] by exposing them to external fire. Here they showed that a higher charge level
corresponds to a more rapid total energy release and a higher peak energy release rate. A
lower charge level yielded a lower energy release rate spread out over a longer time.
However, the charge level of the batteries did not have a significant effect on the total
amount of energy released.
The release of toxic gasses is also affected by the SOC level of a LIB [141]. Larsson et al.
showed that lower SOC yielded higher amounts of hydrogen fluoride (HF) to be released.
Similar results were found by Ribière et al. in [120]. They conclude that the measured
43
© RISE Research Institutes of Sweden
quantity of HF indicates a SOC dependence and that the maximum concentration was
achieved at zero percent SOC. This may indicate that a larger portion of HF is consumed by
more severe fires with higher temperatures, such as those associated with high SOC levels.
Ouyang et al. investigated the fire hazard associated with lithium-ion batteries under
overcharge conditions [155]. They performed abuse experiments on two different cell types,
NMC and LFP. These cells were charged to different levels ranging between 4.2 V to 5 V and
abused. The abuse considered slowly heating the cells with an electric heater. Among other
things, Ouyang et al. analysed several safety parameters such as those related to the onset
of thermal runaway (TR) and radiated heat [155]. For the readers convenience, these results
by Ouyang et al. have been copied and presented in Table 17 and Table 18.
The results shown in Table 17 presents the effect the SOC level has on the response of the
abused cell [155]. This response is presented in terms of the time/temperature that is
needed for cell rupture, ignition and thermal runaway. Their results show that cells at higher
SOC levels go through the different stages faster, with a particularly violent thermal
runaway and ejection for high SOC. They also mentioned that when thermal runaway and
ejection occurred, it was particularly violent for a high SOC. Similar behaviour was recorded
by Ribière et al. [120]. High SOC yielded rapid energy release whereas lower SOC levels
showed less severe thermal runaway and slower burning of the battery.
Golubkov et al. investigated the impact of SOC and overcharge on commercial LIB cells with
LFP and NCA cathodes [118]. They found that a minimum charge level was needed for
thermal runaway to be initiated. Heating fully discharged cells to 250 ºC did not result in
thermal runaway. At least 50 % and 25 % SOC were needed for the considered LFP and NCA
cells, respectively, for this mechanism to be triggered. At 100 % SOC, significant self-heating
occurred when both cells heated to ~140ºC. When overcharged to 143 % SOC, this drops
down to as low as 65ºC. There was however a significant difference in the subsequently
recorded maximum temperatures. That is, maximum cell temperatures of 440ºC and 911ºC
for LFP and NCA, respectively.
Table 17 Specifications of the battery surface temperature during abuse testing by Ouyang et al. [155].
Cell
Cut-Off
Voltage
[V]
Time to
Cracks
[s]
Temp. at
Cracks
[ºC]
Time to
Ignition
[s]
Temp. at
Ignition
[ºC]
Time to
Thermal
Runaway
[s]
Temp. at
Thermal
Runaway
[ºC]
Max.
Temp.
[ºC]
NMC
4.2
197
127
239
158
317
232
553
4.5
196
129
230
162
280
226
606
4.8
191
133
222
160
273
228
630
5.0
190
132
219
163
262
230
673
LFP
4.2
201
115
300
182
358
229
571
4.5
202
115
266
175
310
218
585
4.8
185
121
259
178
290
224
630
5.0
181
127
251
181
280
227
647
Ouyang et al. also measured the radiative heat flux of the tested batteries [155]. This result
is presented in Table 18. The information concerns the amount of energy or heat that is
being radiated by the considered battery cells. The higher the radiative heat flux, the faster
surrounding objects will warm up. This also means that the time after which other battery
cells may fail reduces. The time needed to ignite an object relates to the amount of energy
released onto the object per unit time, i.e. the heat flux. The larger this value is, the shorter
44
© RISE Research Institutes of Sweden
the amount of time needed to ignite another surface. Ouyang et al. [154] show that the NMC
releases more energy than the LFP battery. More importantly, they show that the radiated
energy increases significantly for higher SOC.
Table 18 Detailed data on heat flux in abusive testing performed by Ouyang et al. [155].
Cell
Cut-Off Voltage [V]
Peak Heat Flux [kW/m
2
]
Total Radiative Heat [kJ/m
2
]
NMC
4.2
1.81
25.9
4.5
3.08
26.7
4.8
6.51
41.4
5.0
7.63
41.9
LFP
4.2
1.98
26.7
4.5
4.77
34.7
4.8
6.72
36.3
5.0
1.99
17.9
4.3.3 Thermal Propagation
Thermal propagation refers to the case where a single battery cell failure spreads to
neighbouring cells. The greater the number of cells involved, the larger the amount of gas
and energy that may be released. The risk for significant fire propagation increases
accordingly. It is very important to understand and prevent this failure which may originate
from a single cell and result in thermal runaway of a large pack of cells [156] . Note that EVs
may hold a very large number of cells in a battery pack and due to limited space and
optimized energy density in the packs, non or small spacing between cells and modules are
generally a fact. This is beneficial for thermal propagation.
Lamb et al. [157] investigated failure propagation in LIB modules. Cylindrical and pouch
C/LCO cells were considered and arranged as a triangle or stack, respectively, to create a
battery module. The cells were either connected in series or in parallel. Thermal runaway
was then initiated in one of the cells in each module by mechanical nail penetration. They
found that the significant air gaps around cylindrical cells limit the heat transfer between
them during a thermal runaway. Cells connected in parallel resulted in a stronger
propagation due to heat transfer along the terminals combined with short circuit. Heat
transfer between cells played a more significant role for the pouch module. Thermal
runaway propagated throughout the modules, regardless of whether the connection was in
series or in parallel.
With regards to thermal propagation it is important to consider the charged state of the
battery. This was discussed in more detail previously, but the general trend is that the energy
release rate for charged cells is much higher than discharged cells. According to Hewson
and Domino [158], this is the reason why regulations require that batteries to be transported
or handled should be below some critical charge state.
Before venting, the amount of heat a battery cell can generate is partly limited by the amount
of oxygen inside the cell. When the battery cell does vent, fresh oxygen supply is made
available. According to Santhanagopalan et al. [123] this could enable for up to 2 or 3 times
more heat to be released in comparison to when the cell does not vent. They therefore
propose to restrict the oxygen availability inside battery packs so that less heat is generated
by a failing cell, subsequently reducing the risk for propagation. However, it is important to
consider the flammability limits of the respective gases that are released so that explosions
are avoided.
45
© RISE Research Institutes of Sweden
4.4 Challenges for Responders
In 2013 Long et al. conducted full scale fire tests on two battery types using a mock-up
vehicle shell used for firefighter training purposes [147]. One of the goals of these tests was
to determine whether there are special requirements for firefighting operations involving
electric vehicles compared to conventional ICE vehicles. The batteries were placed in
relatively easily accessible locations in the vehicle: in the rear cargo storage compartment,
either in plain view or under a mock “floorboard”. The firefighters observed that the biggest
challenge was to supply water to the source of the fire. They could cool the outside of the
battery pack, but they could not reach the burning cells unless there was a way to inject the
water inside the pack. The fires reignited multiple times in 5 of the 6 tests.
With regard to firefighting operations, researchers have also found that normally there is
no danger to firefighters for electric shock due to using water as an extinguishing agent [147]
[159]. Two series of fire tests have included suppression of LIB fires with water mists [37]
[141]. In both cases the total HF emissions were similar whether water mist was used or not,
but HF production increased significantly while water mist was being applied to the fire.
Exposure to HF could thus be a hazard for the fire service if water mist is used as the
suppression agent or possible other water-based agents, however, very little research has
been conducted on this. Additional information about the toxicity of the gases emitted by
LIB is found in section 4.4.2 below.
Egelhaaf [146] found that a very large amount of smoke was emitted after the batteries were
extinguished and recommended that a larger than normal area should be blocked off
compared to an ICE vehicle fire.
For electrical vehicles there is not only the threat of a fire immediately after a crash, but also
the risk of a delayed event. This could occur during post-crash handling, including towing
and workshop activities. In addition, there is a risk of reignition significant amounts of time
after first extinguishment. These risks connected to handling of damaged EVs are
elaborated in greater details in section 5.2.
4.4.1 Identifying Electric Vehicles
One of the biggest challenges for responders is to identify the type of vehicle they are dealing
with [160]. Grant [159] states that it can be hard to distinguish EVs from ICE vehicles due
to their similar exterior characteristics. It is very important to understand what type of
vehicle is being dealt with in order to make an appropriate assessment of its associated