See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279069810

Thermal runaway of commercial 18650 Li-ion batteries with LFP and NCA
cathodes - Impact of state of charge and overcharge

Article  in  RSC Advances · June 2015

DOI: 10.1039/C5RA05897J

CITATIONS READS

233 9,482

9 authors, including:

Sebastian Scheikl Gisela Fauler

HyCentA former VARTA Micro Innovation GmbH, Graz
3 PUBLICATIONS   233 CITATIONS    11 PUBLICATIONS   748 CITATIONS   

SEE PROFILE SEE PROFILE

Viktor Hacker
Graz University of Technology
368 PUBLICATIONS   3,658 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Project "e!Polycat": Pt-free anode catalyst systems and composite polysaccharide-membranes for the alkaline DEFC View project

GO DEFC View project

All content following this page was uploaded by Viktor Hacker on 11 January 2016.

The user has requested enhancement of the downloaded file.

 RSC Advances
View Article Online
PAPER View Journal | View Issue

Thermal runaway of commercial 18650 Li-ion

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

batteries with LFP and NCA cathodes – impact of

Cite this: RSC Adv., 2015, 5, 57171
state of charge and overcharge
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

Andrey W. Golubkov,*a Sebastian Scheikl,a René Planteu,a Gernot Voitic,b

Helmar Wiltsche,c Christoph Stangl,d Gisela Fauler,d Alexander Thalera
and Viktor Hackerb

Thermal runaway characteristics of two types of commercially available 18650 cells, based on LixFePO4 and
Lix (Ni0.80Co0.15Al0.05)O2 were investigated in detail. The cells were preconditioned to state of charge (SOC)
values in the range of 0% to 143%; this ensured that the working SOC window as well as overcharge
conditions were covered in the experiments. Subsequently a series of temperature-ramp tests was
performed with the preconditioned cells. Charged cells went into a thermal runaway, when heated
above a critical temperature. The following thermal runaway parameters are provided for each
Received 2nd April 2015
Accepted 22nd June 2015
experiment with the two cell types: temperature of a first detected exothermic reaction, maximum cell
temperature, amount of produced ventgas and the composition of the ventgas. The dependence of
DOI: 10.1039/c5ra05897j
those parameters with respect to the SOC is presented and a model of the major reactions during the
www.rsc.org/advances thermal runaway is made.

external measures (built-in safety devices, forced cooling,

1 Introduction connement).3–5
Li-ion batteries1,2 excel in energy density and cycle life. In the past, accelerated rate calorimetry (ARC) tests with
Unfortunately those benets come with a price: when Li-ion limited maximum temperature6–12 and without limitation13 as
batteries are mistreated with high over-temperature or well as re experiments and mechanical abuse14–17 with
strong overcharge, they can transit into a so-called thermal complete Li-ion cells were done. Recently over-temperature
runaway. During the thermal runaway, the battery tempera- and over-charge tests with large format cells (which may be
ture increases due to exothermic reactions. In turn, the used for automotive applications) were published.18–20 It is
increased temperature accelerates those degradation reac- known that the severity of the thermal runaway event in over-
tions and the system destabilizes. At the end of the thermal temperature experiments increases with increasing SOC.4,21–27
runaway, battery temperatures higher than 1000
C can be It is also known, that a thermal runaway can be triggered by
reached and high amounts of burnable and harmful gases can strong overcharge beyond safe voltage limits of the cell.28–34
be released. Even if the overcharge condition does not trigger a thermal
Because Li-ion batteries are widely used, the possible runaway, safety may be compromised by Li-plating on the
hazards of Li-ion batteries are a key issue for automotive, anode.35
aerospace and consumer electronics industries. The safety In our previous publication36 the safety characteristics of
characteristics of Li-ion battery systems depend (a) on the used three different commercial Li-ion batteries charged to 100%
cell type (geometry, materials), (b) on the initial conditions SOC were investigated. It was demonstrated, that cells with
before misuse (state of charge, ageing effects), (c) on the type cathodes based on iron-phosphate as well as on metal-oxide
of misuse (over-temperature, over-charge) and (d) on material exhibit a thermal runaway in thermal-ramp experi-
ments. The severity of the thermal runaway showed a strong
dependence on the material composition of the cells.
In this publication two cell types are introduced and the
a
Kompetenzzentrum – Das Virtuelle Fahrzeug Forschungsgesellscha mbH, mass inventory of the cells is calculated based on tear down
Inffeldgasse 21a, A-8010 Graz, Austria. E-mail: [email protected] results. The thermal runaway testing method is explained and
b
Institute of Chemical Engineering and Environmental Technology, Graz University of
the outcomes of experiments with discharged, partially
Technology, Inffeldgasse 25/C/II, 8010 Graz, Austria
c
charged, fully charged and over-charged cells are presented.
Institute of Analytical Chemistry and Food Chemistry, Graz University of Technology,
Stremayrgasse 9/III, 8010 Graz, Austria Possible chemical reactions are listed and quantitative calcu-
d
Varta Micro Innovation GmbH, Stremayrgasse 9, 8010 Graz, Austria lations of ventgas generation are made for two cases.

This journal is © The Royal Society of Chemistry 2015 RSC Adv., 2015, 5, 57171–57186 | 57171
 View Article Online

RSC Advances Paper

2 Samples 2.2.1 Binder and conducting agent. The mass ratio of

binder material and conducting agents in the electrode coatings
The two types of commercially available Li-ion batteries, with was not measured. We assume that sodium carboxymethylcel-
the geometrical format 18650, were purchased from two well lulose (CMC) with a degree of carboxymethyl substitution (DS)
known manufacturers. The rst cell, rated to a nominal capacity of 0.7 is used as the anode binder37 and polyvinylidene uoride
of Cnom ¼ 1.1 A h is based on a LixFePO4 (LFP) cathode. The LFP (PVDF) is used as the cathode binder.38 CMC is a cost effective
material is considered as relatively safe. Unfortunately binder material in the anode, but can not be used in the
commercial LFP-based cells have lower capacity and nominal cathode. We suppose that 5% of anode coating and 2.5% (NCA)
voltage compared to metal-oxide based cells. According to the or 5% (LFP) of cathode coating is binder material.
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

datasheet the LFP cell is designed for a maximum discharge Additionally a conducting agent is needed to improve the
current of 30 A and has a cycle life of >1000 full discharge cycles. electrical conductivity between the cathode particles and
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

The second cell has a much higher nominal capacity Cnom ¼ cathode substrate-foil. We suppose that 2.5% (NCA) or 5% (LFP)
3.35 A h and is based on a Lix(Ni0.80Co0.15Al0.05)O2 (NCA) of cathode coating consists of carbon black. We justify the
cathode. To our knowledge, this mass produced cell has the increased amount of binder and conducting agent of the LFP
highest energy density which is commercially available as of cell with its higher power capability.
2013. It is specied to a maximum discharge current of 6.7 A h 2.2.2 LFP particle coating. The active cathode material of
and its cycle life is >300 cycles. the LFP cell consists of LixFePO4. The LixFePO4 particles need to
In the following, the two cell types will be denoted as LFP and be nano structured and carbon coated to achieve good diffusion
NCA for easy reading. of Li-ions and good inter-particle electrical conductivity.39 It is
hard to tell which amount of carbon coating was actually used
2.1 Cell composition, methods in the tested commercial battery. Optimum values of carbon
coating found in the literature vary from 1.5% to 15%.40 We
For the interpretation of the misuse experiment results it is
assume that 10% of the LFP cathode consists of carbon coating.
benecial to know the mass split of the cell components.
Please note, that this might be the upper estimate. One of the
Unfortunately information regarding detailed cell composition
reviewers suggested, that the carbon coating of a commercial
is kept condential by the manufacturers. We had to make a
battery is probably in the range of 1% to 2%.
tear down and an analysis of the cell components for both cell
2.2.3 Electrolyte and SEI. The amount of salt in the elec-
species by ourselves. The following parameters were measured
trolyte could not be measured as well, it is supposed that both
directly using the same methods and equipment as in ref. 36:
cells use the traditional salt LiPF6 with a concentration of
 Mass of the anode and cathode coating, the electrolyte, the
1.1 mol L1. The density of the electrolytes is estimated with
current collector foils, the separator and the housing material.
1.21 kg L1.
 The solvent mass-ratios of the electrolyte. Detected solvents
Vinylene carbonate (VC) is a common solid electrolyte
were dimethyl carbonate (DMC), ethyl methyl carbonate (EMC),
interface (SEI) improving additive.41 We assume that 2% of VC
ethylene carbonate (EC), propylene carbonate (PC) and methyl
was added to the electrolyte.42 During initial charging VC and
phenyl carbonate (MPC).
EC undergo reduction reactions and form the SEI at the surface
 The mole-ratios of the different transition metals and
of the graphite particles of the anode. A fully developed SEI
phosphor in the cathode coating.
prevents further reduction of the electrolyte solvents.43 The SEI
Additionally, separator foils were examined with differential
composition and formation reactions can be complicated41,44,45
scanning calorimetry coupled with thermal gravimetric analysis
and lie beyond the scope of this work. Instead, for further
(DSC-TGA, NETZSCH STA 449 C). Separator samples were rinsed
calculations, we treat the SEI as being made of only four
with diethyl carbonate and dried in a desiccator for 12 hours.
components:
During the test the DSC-TGA was ushed with and the heat
(1) The polymerization product of VC41,46
ramp was set to 10 K min1.

2.2 Cell composition, results and discussion

It is not in the scope of this work to compile an exhaustive
material inventory of the two commercial cell types. Neverthe-
less, to obtain some insight into chemical reactions taking place (1)
during cell misuse, it is helpful to make at least rough estima-
tions for cell components that were not accessible to direct
measurements (Table 1). Estimations for the amount and
composition of active material, particle coating, binder, carbon
black and the SEI in the electrode coatings as well as for the
amount of salt, additives and soluble SEI in the electrolyte were
discussed with our project partners. Effects of cell formation (2) The organic Li-carbonate from EC reduction47–49
were considered. The compositions of the separators were
estimated from DSC measurements. 2C3H4O3 (EC) + 2Li+ + 2e / (CH2OCO2Li)2 + C2H4 (2)

57172 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015
 Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Table 1 Mass split of the discharged NCA and LFP cell. Please note, that the mass ratios for the binder, carbon black, the SEI and the salt were not measured; instead rough estimates are given.
Paper

The chemical formulas for the SEI are taken from literature. The composition of the separator is deduced from DSC-TGA scans. For known or estimated materials the molar mass M and the
amount of material n in the cell is given. For polymers M and n of the monomers is given. In the case of CMC the average number of atoms per monomer is used. The Li fractions in the molecular
res
formulas of the electrodes were calculated for discharged cells with consideration of the irreversible capacity loss nirrLi and the residual capacity nLi . In addition, the maximum theoretical
capacity of intercalated Li in the electrodes was calculated using Faraday's laws of electrolysis

NCA cell LFP cell

Mass M n Li-capacity Mass Mass M n Li-capacity

Component Mass ratio (g) Formula (g mol1) (mmol) (A h) ratio (g) Formula (g mol1) (mmol) (A h)

Cathode coating:
Active material 95.0% 17.04 Li0.925 95.6 178.2 4.78 80.0% 7.73 Li0.882FePO4 156.9 49.2 1.32
(Ni0.80Co0.15Al0.05)O2
Particle coating — — — — — 10.0% 0.97 C 12.0 80.5
Carbon black 2.5% 0.45 C 12.0 37.4 5.0% 0.48 C 12.0 40.2
Binder (PVDF) 2.5% 0.45 (C2H2F2)n 64.0 7.0 5.0% 0.48 (C2H2F2)n 64.0 7.5

This journal is © The Royal Society of Chemistry 2015

Sum 100% 17.93 100% 9.66

Anode coating:
Graphite 93.5% 10.92 Li0.008C6 72.1 151.5 4.06 93.5% 4.84 Li0.006C6 72.1 67.2 1.80
Binder (CMC) 5.0% 0.58 (C7.4H10.7O6.4)n 202.1 2.9 5.0% 0.26 (C7.4H10.7O6.4)n 202.1 1.3
SEI, lithium uoride 0.4% 0.04 LiF 25.9 1.7 0.4% 0.02 LiF 25.9 0.8
SEI, lithium carbonate 1.1% 0.13 Li2CO3 73.9 1.7 1.1% 0.06 Li2CO3 73.9 0.8
Sum 100.0% 11.67 100.0% 5.18

Separator:
PP — — — — — 66% 0.76 (C3H6)n 42.1 18.0
PE — — — — — 34% 0.39 (C2H4)n 28.1 13.9
UHMWPE 100% 0.70 (C2H4)n 28.1 25.0 — — — — —
Sum 100% 0.70 100% 1.15

Electrolyte:
SEI, polymer organic 2.0% 0.09 (C6H4O6)n 86.1 1.0 2.0% 0.13 (C6H4O6)n 172.1 0.7
SEI, organic 12.7% 0.56 (CH2OCO2Li)2 162.0 3.5 3.9% 0.25 (CH2OCO2Li)2 162.0 1.5
EC 17.1% 0.75 C3H4O3 88.1 8.5 24.8% 1.59 C3H4O3 88.1 18.1
DMC 49.7% 2.19 C3H6O3 90.1 24.3 33.0% 2.12 C3H6O3 90.1 23.5
EMC 5.3% 0.23 C4H8O3 104.1 2.2 16.5% 1.06 C4H8O3 104.1 10.2
PC — — — — — 8.3% 0.53 C4H6O3 102.1 5.2
MPC 2.7% 0.12 C8H8O3 196.1 0.6 — — — — —
Salt 10.6% 0.46 LiPF6 151.9 3.1 11.5% 0.74 LiPF6 151.9 4.9
Sum 100.0% 4.41 100.0% 6.41

Inactive components:
Housing (metal can) 5.71 10.45
Cathode Al foil 1.20 2.14
Anode Cu foil 2.72 3.86
Other components 1.05 —
Sum 10.69 16.46
Total sum 45.40 38.87
View Article Online

RSC Adv., 2015, 5, 57171–57186 | 57173

RSC Advances
 View Article Online

RSC Advances Paper

(3) The inorganic Li-carbonate Li2CO3 from EC reduction50–52 The effect of the missing lithium nirr Li (proportional to Cirr) in
the cathode is taken into account in further stoichiometric
C3H4O3 (EC) + 2Li+ + 2e / Li2CO3 + C2H4 (3) calculations.
2.2.5 Residual capacity. Commercial Li-ion cells must not
(4) And LiF which can be produced from decomposition of be discharged beyond their rated minimal operation voltage
the salt and the Li-carbonate53 (Vmin(NCA) ¼ 2.5 V and Vmin(LFP) ¼ 2.0 V) during normal
LiPF6 /LiF þ PF5 cycling. If cells are discharged to voltages lower than Vmin
PF5 þ ROH/HF þ RF þ POF3 (4) dissolution of the copper foil may occur,55 because the anode
Li2 CO3 þ 2HF/2LiF þ CO2 þ H2 O potential may reach the oxidation potential56 of Cu. Anodes of
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

cells that are discharged to Vmin are not fully delithiated, instead
We assume that all VC (2% of electrolyte) goes into poly- a small amount of Li stays in the anodes and acts as a safety
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

merization (1) and that the additional SEI components margin to keep the anode potentials below the copper disso-
(CH2OCO2Li)2 : Li2CO3 : LiF are in the ratio 1/2 : 1/4 : 1/4.44 The lution potential. We assume that the residual capacity Cres
components of the SEI are listed (Table 1) as a part of either (which is proportional to the amount of residual Li nres Li ) equals

anode or electrolyte depending on their solubility in the elec- to 1% of the nominal cell capacity:
trolyte solvent.54 To calculate the actual amounts of lithium
Cres ¼ 0.01Cnom (7)
containing SEI we need to take the irreversible capacity loss into
account. nres
Li ¼ 1/FC
res
(8)
2.2.4 Irreversible capacity loss. We think that the most
economical anode material for both manufacturers is surface The amount of residual lithium is considered in further
treated natural graphite. During cell assembly the graphite is in calculation of the lithiation states of both electrodes.
delithiated state and the cathode is in fully lithiated state. At the
rst charging (cell formation) an amount of lithium nirr Li that is
equivalent to 8% of the maximum anode-Li-capacity is trap- 2.3 Available capacities in the electrodes
ped.2 The associated charge Cirr is called irreversible capacity With identied amount of active cathode material nacat and with
loss: known Cirr and Cres the theoretically usable capacity of the
nirr cathode aer cell formation can be calculated
Li ¼ 0:08nC6
a
(5)

Cucat ¼ Fnacat  Cirr  Cres (9)

C irr
¼ Fnirr
Li (6)
and compared to the nominal capacity as given in the data
here F is the Faraday constant and naC6 is the amount of sheet.
graphite units C6 in the anode (in mol). We assumed that all In the case of LFP cell Cucat ¼ 1.16 A h. In theory, LFP material
trapped lithium is integrated and immobilized in the SEI can be fully delithiated, and Cucat should be equal to Cnom. In our
according to the chemical reactions (2)–(4). The calculated work, the calculated Cucat exceeded Cnom. According to the data
Li (NCA) ¼ 12.1 mmol and
values for the NCA and LFP cell are nirr sheet the LFP cell is rated to Cnom ¼ 1.1 A h and the measured
nirr
Li (LFP) ¼ 5.4 mmol respectively. As a consequence, aer capacities in the allowed voltage range were even smaller
formation, the cathode can never again be fully lithiated. Even (Fig. 1). The discrepancy may be caused by incomplete utiliza-
when the cell is fully discharged, nirr Li is missing, and the tion of the LFP material of a real cell or by ageing effects of the
amount of Li per stoichiometric formula in the cathode is <1. cathode.

Fig. 1 OCV characteristics of the (left) LFP and (right) NCA cells. Measured values in the allowed voltage range and additionally in the overcharge
region are given. SOC points, at which temperature-ramp experiments were done, are marked. Rough estimates, where Li-plating and complete
cathode delithiation may occur, are indicated for the NCA cell. The discontinuities at 100% are caused by relaxations during the time-gaps
between cycling and the overcharge experiments.

57174 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015
 View Article Online

Paper RSC Advances

It is noteworthy that the available capacity of the LFP anode 3.3 SOC set-point
Cand
u
¼ FnaC6 (10) The cell was brought to the desired SOC by charging or dis-
charging, starting from 100% SOC. The coulomb counting
exceeds the Cucat
by 50%. In other words, the anode of the LFP
method was used for SOC calculation and the charge/discharge
cell is overbalanced. This makes sense for a high power cell, as
was stopped when the required SOC was reached. For experi-
it allows high charging currents with reduced risk of
ments with SOC < 100% the cell was discharged outside of the
Li-plating.
reactor. For SOC > 100% the cell was overcharged inside the
In contrast to LFP, the NCA cathodes should not be fully
reactor, for safety reasons. In order to prevent cell heating, the
delithiated during normal operation. Correspondingly, the
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

overcharge current was set to very low values. The SOC set-
theoretically available capacity of the NCA cell of 4.42 A h was
points of all experiments are marked in Fig. 1.
higher than the nominal capacity 3.35 A h. The calculated
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

capacity of the active material in the anode was 4.06 A h. That

means the NCA anode was slightly overbalanced by 21%. 3.4 Thermal-ramp experiment
2.3.1 Separator. The composition of the separator mate- The sealed reactor was evacuated and ushed with inert gas.
rials was deduced from DSC-TGA measurements. The separator The heaters were turned on. The sample inside the reactor was
of the LFP cells showed endothermic (melting) peaks at 132
C heated slowly with a rate of 2
C min1 (NCA) or 4
C min1
and 159
C which are typical for a 3-layered laminate with a (LFP). Cell temperatures, gas temperatures and the pressure in
polyethylene (PE) core between two polypropylene (PP) skin the sealed reactor were recorded. At some point the cell trans-
layers (PP/PE/PP). We assume that the LFP separator consists of ited into thermal runaway and ventgas was released in the
2/3PP and 1/3PE. reactor. The amount of gas inside the reactor nideal
sum was calcu-
The separator of the NCA cell showed only one indistinct lated using the ideal gas law
endothermic peak at 130
C. We assume that the NCA sepa- pV
rator consists of ultra-high molecular weight polyethylene nideal
sum ¼  n0 : (11)
Rqgas
(UHMWPE) membrane.57,58

Here p denotes the pressure in the reactor, V ¼ 0.0027 m3 is

3 Experimental the reactor volume, R is the gas constant, qgas is the gas
temperature in the reactor (in K) and n0 is the initial amount of
In this work a total of 23 thermal ramp experiments with the two
gas in the reactor at the start of the experiment.
cell types were done at different SOC. Each experiment con-
The eqn (11) is only valid, when qgas is equal to the mean gas
sisted of the following steps; the cell underwent a open circuit
temperature in the reactor. During the thermal runaway a
voltage (OCV) check, was charged to the selected SOC and
violent cell venting may take place and hot gases are released
inserted into the sample holder. The sample holder was
into the pressure vessel. In the rst seconds aer venting, when
attached inside a sealed reactor and the thermal ramp experi-
the gas temperature inside the reactor is not homogeneous,
ment was started (the test-rig and thermal ramp method is
nideal ideal
sum may be over or underestimated. Thus, given nsum values
described in ref. 36). Aer the thermal ramp experiment gas
were calculated when the gas temperature was in equilibrium.
samples were taken and analysed.

3.5 Ventgas analysis

3.1 Initial OCV check
Gas samples were taken aer the thermal runaway reaction. If
We applied the same OCV measurement procedure as in ref. 36. no thermal runaway occurred, then the gas samples were taken
Each sample was fully discharged to 0% SOC (2.5 V) and then aer the cell temperature exceeded 250
C. The gas was ana-
fully charged to 100% SOC (LFP: 3.5 V, NCA: 4.2 V). The health lysed with a gas chromatograph system (GC, Agilent Technolo-
status of the cells was checked by comparing the measured gies 3000 Micro GC, two columns, Mol Sieve and PLOTU). A
capacities with the nominal capacity from the manufacturer. thermal conductivity detector (TCD) was used to detect
Typical OCV proles are given in Fig. 1. BaSyTec CTS cell test permanent gases. The GC was calibrated for H2, O2, N2, CO,
system and Heiden Power DC-source-load were used for battery CO2, CH4, C2H2, C2H4 and C2H6. The GC used Ar and He as
cycling. carrier gases.

3.2 Sample preparation 3.6 Role of the inert gas

Aer the OCV check the insulation foil was stripped from the Before each experiment, the reactor was lled with inert gas to
cell and the sample was weighed. Three K-type thermocouples prevent reactions of the vent-gas with the reactor atmosphere.
were spot-welded to the cell housing. Then the sample was We used either N2 or Ar as inert gas. Both gases have advantages
wrapped in a thermal insulation layer and inserted into the and disadvantages.
heating sleeve of the sample holder. Finally, the sample holder  Advantages of using Ar as inert gas: in this case N2 is not
was installed inside the reactor, the electrical connections were present in the reactor. There are no reactions which can
made and the reactor was sealed. produce N2 during thermal runaway. The only possible source

This journal is © The Royal Society of Chemistry 2015 RSC Adv., 2015, 5, 57171–57186 | 57175
 View Article Online

RSC Advances Paper

of N2 in a ventgas sample is leakage from ambient air. There- Only in the last three experiments, aer enough experience was
fore, the presence of N2 (accompanied by O2) in the GC results gained, we were condent to use N2 as the inert gas.
indicates gas leakage. The amount of Ar in the samples could
not be quantied, because it was used as a carrier gas in the GC
setup. 4 Results
 Advantages of using N2 as inert gas: in this case N2 fulls We did 23 thermal-ramp experiments with NCA and LFP cells
two functions. It serves as inert gas and also as an internal set to different SOC. The results are summarized in Tables 2 and
standard. Since the amount of N2 in the reactor is known (VN2 ¼ 3. Typical experiment runs are shown in Fig. 2 and 4. The
0.0027 m3), absolute amounts of other detected gas compo-
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

dependence of the thermal runaway parameters on the SOC is

nents can be derived from their relative GC results rGC
i visualised in Fig. 3 and 5.
nN2 GC
nGC ¼ ri (12)
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

i
rGC
N2
4.1 NCA cells
We tested the thermal stabilities of discharged as well as
The absolute amount of vent-gas nGCsum can be calculated from partially charged, fully charged and over charged NCA cells.
the GC results. Discharged NCA cells (Experiment 1–5) showed no
X
nGC pronounced thermal runaway characteristics. Only small unre-
sum ¼ nGC
i (13)
isN2 markable exothermic peaks were observed between 150
C and
300
C. The amount of gas depended on the timespan which the
cells spend at increased temperature: aer the initial burst plate
The amount of ventgas calculated with the ideal gas eqn (11)
opening of the cell housing the vent-gas was released from the
can be compared with the total amount of gas from GC results
cell into the reactor with an uniform rate. There was no sudden
sum ¼ nsum than it is likely, that all formed gases were
(13). If nideal GC
gas liberation and no violent chemical reaction. CO2 was the
detected by the GC.
major identied component of the vent-gas. Interestingly, the
However, there is also one strong disadvantage of using N2. If
mass loss of the discharged cells of 4.4 g equalled to the mass of
leaks from ambient air occur, leaked N2 falsies the internal
electrolyte in the cells (Table 1).
standard. Therefore, for the most experiments we used Ar as
In Experiment 1 we used N2 as internal standard. The GC
inert gas in the reactor and detected N2 indicated gas leaks.
detected nGCsum ¼ 23.2 mmol of produced gas (Table 3). In

Table 2 Results of thermal ramp experiments with NCA and LFP cells. Here SOC is the state of charge, qo is the onset temperature, qm is the
maximum cell temperature during the experiment, Dm is the mass loss of the cell, nideal
sum is the measured amount of produced vent-gas (11) and
the chemical components are those species that were detected by the GC system. Missing values could not be measured or detected. The ratios
of the detected gases are given in mol%

SOC qR qm Dm nideal
sum H2 CO2 CO CH4 C2H4 C2H6
No. Cell (%) (
C) (
C) (g) (mmol) (%) (%) (%) (%) (%) (%)

1 NCA 0 — 302 — 65 1.7 94.6 1.6 1.6 0.3 —

2 NCA 0 160 316 4.4 52 1.8 94.7 1.9 1.2 0.4 —
3 NCA 0 160 315 4.5 55 1.2 96 1.5 1.1 0.2 —
4 NCA 0 161 214 4.4 39 0.9 96.2 1.1 1.4 0.3 —
5 NCA 0 150 243 4.4 59 0.8 96.6 1 1.3 0.3 —
6 NCA 25 150 739 5.9 67 15.5 62.7 5.5 8.7 7.5 —
7 NCA 50 140 970 8.5 157 17.5 33.8 39.9 5.2 3.2 0.4
8 NCA 75 140 955 — 217 24.2 20.8 43.7 7.5 3.3 0.5
9 NCA 100 144 904 — 273 22.6 19.7 48.9 6.6 2.4 —
10 NCA 100 138 896 20.5 314 26.1 17.5 44 8.9 2.7 0.9
11 NCA 100 136 933 20.9 244 28.5 22.7 41.5 5.9 1.3 0.3
12 NCA 112 144 — 19.2 252 25.1 18.8 48.1 5.9 2.1 —
13 NCA 120 80 929 — 281 23.5 20.8 48.7 5.4 1.6 —
14 NCA 127 80 983 — 317 28.8 16.2 46.6 6.4 1.3 0.3
15 NCA 132 80 943 17 262 25.8 18.9 49.2 4.7 1.4 —
16 NCA 143 65 1075 20.1 303 26.2 22 43.4 6.9 1.5 —
17 LFP 0 — 251 6.1 55 2.7 93.5 1.8 0.7 0.7 0.7
18 LFP 25 195 231 6.1 31 7.1 85.3 3.1 1.2 3.1 0.2
19 LFP 50 130 283 6.1 32 20.8 66.2 4.8 1.6 6.6 —
20 LFP 75 149 362 6.3 41 21.8 62.6 6.4 1.9 6.3 1
21 LFP 100 140 440 7.1 32 29.4 48.3 9.1 5.4 7.2 0.5
22 LFP 115 155 395 6.2 61 34 52.2 6.4 2.6 4.7 0.1
23 LFP 130 80 448 — 58 30.1 55.8 7.7 6.4 — —

57176 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015
 View Article Online

Paper RSC Advances

Table 3 Thermal runaway parameters of experiments with NCA cells. Experiments with N2 as internal standard were selected and the amounts
of measured gases are given in absolute units (12). The amount of vent-gas nideal GC
sum and nsum was calculated with ideal gas eqn (11) and with results
of the GC (13) respectively

SOC nideal
sum nGC
sum H2 CO2 CO CH4 C2H4
No. (%) (mmol) (mmol) (mmol) (mmol) (mmol) (mmol) (mmol)

1 0 65.4 23.2 0.4 21.9 0.4 0.4 0.1

13 120 281.3 279.1 65.5 57.9 136 15.1 4.6
14 127 317 317.1 91.6 51.6 148.6 20.2 4.1
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

contrast, the amount of ventgas inside the reactor (11) was between 65
C and 80
C. It is an important nding, that over-
sum ¼ 65.4 mmol. We conclude that the GC could
much higher nideal charged NCA cells can proceed straight into thermal runaway
not identify the missing 42.2 mmol of gas, because its setup was when heated above 65
C.
optimized for a limited set of permanent gases. The thermal runaway reactions were accompanied by abrupt
The cells with SOC $ 25% displayed an unmistakable vent-gas releases. Cells with higher SOC produced more vent-gas.
thermal runaway behaviour. When (partially) charged NCA cells Up to 317 mmol of gas were recorded. The gas composition
were heated beyond a critical temperature, self accelerating depended on the SOC as well: the fractions of CO2 decreased and
exothermic reactions started and the cell temperatures the fractions of CO and H2 increased with rising SOC. A clear trend
suddenly increased up to maximum values in the range of for other detected gases (CH4, C2H4 and C2H6) was not observed.
739
C and 1075
C. We used N2 as inert gas in the Experiments 13 and 14 (over-
The onsets of the exothermic reactions were obtained from charged NCA) in the same way as in Experiment 1. The calcu-
the rate plot: the temperature, where a rst clear deviation lated amounts of gas nideal GC
sum and nsum were in good agreement,
towards increased temperature rate was detected, was dened indicating that all produced gases were detected by the GC. In
as the onset temperature qo. For NCA cells with SOC # 100% qo other words, it is likely that the quantitative GC results (Table 3)
was in the range between 136
C and 160
C. Overcharged NCA represent the major vent-gas components for over-charged cells
cells (SOC > 100%) showed much lower onset temperatures and that only smaller amounts of gas may be missing.

Fig. 2 Typical thermal ramp experiments with NCA cells. (a) Cell temperature profiles. (c) Amount of produced vent-gas. The overshoot peaks
are artefacts of the calculation (11) caused by inhomogeneous gas temperature. (b) Cell temperature rate dq/dt vs. cell temperature q, full
temperature range. (d) Close up view of the cell temperature rate.

This journal is © The Royal Society of Chemistry 2015 RSC Adv., 2015, 5, 57171–57186 | 57177
 View Article Online

RSC Advances Paper

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

Fig. 3 Characteristic thermal runaway parameters from all NCA experiments. (a) Onset cell temperature qo. (b) Maximum cell temperature qm.
(c) Amount of produced gas nideal GC
sum . (d) Main detected gas components ri .

4.2 LFP cells The discharged LFP cell (Experiment 17) showed a behaviour
similar to discharged NCA cells. Exothermic reactions could not
In addition to the experiments with NCA cells, we did 7
be detected. Aer the initial burst plate opening of the cell
thermal-ramp experiments with LFP cells at different SOC
(Table 2). housing, the amount of gas increased evenly over time as the cell

Fig. 4Typical experiments with LFP cells: (a) cell temperature profiles. (c) Amount of produced vent-gas. (b) Cell temperature rate dq/dt vs. cell
temperature q, full temperature range. (d) Close up view of the cell temperature rate.

57178 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015
 View Article Online

Paper RSC Advances

This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

Fig. 5 Summary of all experiments with LFP cells at different SOC: (a) onset cell temperature qo. (b) Maximum cell temperature qm. (c) Amount of
produced gas nideal
sum . (d) Main detected gas components ri .
GC

was heated. For the discharged cell, the GC registered essentially of produced gas and its components be explained with a set of
only CO2. We suspect that the GC could not detect all gas chemical equations?
components that were produced by the discharged cell: similar Material naj that is available for the reaction system is listed
to Experiment 1 with a discharged NCA cell, signicant amounts in Table 1. In addition lithium naO2 and oxygen naLi may be
of gas may be missing in the GC results, simply because the used released in heated cells. Part of the material is consumed (by
GC equipment was not capable of detecting them. becoming a reactant nrj of the reaction system).
First mild exothermic reactions were seen for a cell that was
charged to 25% SOC. The reactions were not strong enough to 0 # nrj # naj (14)
evolve into a distinct thermal runaway. Vent gas was produced
continuously with time, likewise to the experiments with dis- The reaction products may consist of gases, uids and
charged cells. solids. A measurable subset of the resulting gaseous products
LFP cells charged to SOC $ 50% showed pronounced nGC
i and the sum of ventgas nideal
sum is given in Table 3. The

thermal runaway reactions. Increasing SOC caused increasing challenge is to nd the right set A of equations and to nd the
maximum temperatures during thermal runaway. The utilisation number b for each equation (how oen is each
maximum temperatures qm ranged from 283
C to 448
C. equation applied) so that the calculated amounts of products
The onset temperature qo was 140
C for cells between 50% npi match the measured values:
SOC and 100% SOC. The cell overcharged to 130% SOC showed
minimize (nGC
i  npi ) > 0 (15)
a exothermic reaction already at 80
C. In contrast to over-
charged NCA cells, the initial exothermic reaction of the over- and
charged LFP cell could not sustain a full thermal runaway. The !
overcharged LFP cell proceeded into thermal runaway only aer X
minimize nideal  npi .0 (16)
it was heated by the heating sleeve beyond 140
C. sum
i¼gaseous
The amount of gas nideal
sum ranged between 31 mmol and
61 mmol and showed no clear dependence on the SOC. With
increasing SOC the relative composition of the detected gases In other words, the difference of calculated and measured
changed to lower CO2 and higher H2 fractions. The fractions of amounts of products is dened as the cost function and the
CO (max. 9.1%) were lower than for NCA cells. system is restricted by the amounts of reactants and products.
The mass loss of the LFP cells ranged from 6.1 g to 7.1 g and The algorithm should minimise the cost function and respect
is comparable to the amount of electrolyte (6.5 g) in this cells. the restrictions.
The mathematical problem was solved using the LIPSOL
linear programming toolbox in Matlab. The set of chemical
5 Discussion equations and two explicit calculations (discharged NCA cell,
Experiment 1 and over-charged NCA cell, Experiment 13) are
It is tempting to pinpoint the main contributors of heat and gas
disclosed in the next subsections:
release during the thermal runaway reactions. Can the amount

This journal is © The Royal Society of Chemistry 2015 RSC Adv., 2015, 5, 57171–57186 | 57179
 View Article Online

RSC Advances Paper

5.1 Oxygen release from cathode material partly lithiated LFP cathode in a Li-ion battery consists of a mix
of lithiated (LiFePO4) and delithiated (FePO4) particles66 then the
At elevated temperatures cathode materials of Li-ion batteries
oxygen release of a partially charged cathode is given by:
can release oxygen. It is believed, that the exothermic reaction of
oxygen with electrolyte is the main contributor to the extent of ð1  xÞLiFePO4 þ xFePO4 /ð1  xÞLiFePO4
the thermal runway effect.4 The amount of released oxygen 1 1
depends on the lithiation state of the cathode. þ x Fe2 P2 O7 þ O2 (20)
2 4
In ref. 59–62 it is shown, that delithiated Lix(Ni0.80Co0.15-
Al0.05)O2 cathode material undergoes complex phase transitions The absolute amount of O2 from the LFP cell can be calcu-
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

accompanied by formation of O2 in the temperature range from lated with equations similar to (18) and (19).
175
C to 600
C. In a simplied scheme, heated NCA transits Both cathodes materials NCA and LFP can contribute O2
from layered to rock salt structure with O2 release, depending (Fig. 6) which in turn can take part in further exothermic
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

on the lithiation state: degradation reactions. The amount of O2 is higher for deli-
Lix ðNi0:80 Co0:15 Al0:05 ÞO2 /Lix ðNi0:80 Co0:15 Al0:05 ÞO1þx thiated cathodes (battery is charged). Note, that because of the
irreversible capacity loss during formation of actual cells, the
1
þ ð1  xÞO2 (17) cathode can not be fully lithiated by discharge of the cell: even
2
at 0% SOC (battery is discharged) the lithiation factor x < 1 and a
small amount of O2 may be released.
The lithiation state x can be calculated using
nirr
Li þ nLi þ 1=F SOC C
res nom 5.2 Exposure of lithium by the anode
x ¼ 1 (18)
nNCA
a
On the anode side graphite particles can defoliate and expose
intercalated Li at temperatures above 230
C.67,68 The amount of
The amount of liberated O2 is released Li depends on SOC of the battery:

1x a naLi ¼ 1/F SOC Cnom + nres (21)

naO2 ¼ nNCA : (19) Li
2
The NCA cell can release naLi (SOC ¼ 100%) ¼ 126.2 mmol in
This calculation gives naO2 ðSOC ¼ 100%Þ ¼ 69:2 mmol for a the fully charged state and naLi (SOC ¼ 0%) ¼ 1.24 mmol in
fully charged NCA cell and naO2 ðSOC ¼ 0%Þ ¼ 6:7 mmol for a discharged state.
fully discharged NCA cell.
LFP cathodes have better thermal stability than layered metal 5.3 Typical chemical reactions
oxide cathodes such as Lix(Ni0.80Co0.15Al0.05)O2, because of the In this section we compile a list of probable degradation reac-
strong covalent bonds of the phosphate groups of LixFePO4.63 tions which may take place during thermal runaway. The most
Nevertheless a phase transition with O2 release of heated FePO4 signicant chemical reactions may be reactions with O2 and Li:
can be detected by XRD experiments.64,65 If assumed that the partially delithiated cathodes release O2 and partially lithiated
anodes release Li at elevated temperatures (17), (20) and (21).
Both released materials are highly reactive and promote a
number of reactions that are summarized in a previous publi-
cation.36 Additionally, following reactions are considered:
Combustion of the carbon black (conducting additive) or
anode graphite
1
Cþ O2 /CO (22)
2
the water–gas shi reaction

CO + H2O / CO2 + H2 (23)

oxidation of exposed Li with water

2Li + H2O / H2 + Li2O (24)

endothermic decomposition of liquid lithium carbonate (at

high temperatures, qmelt ¼ 720
C) with carbon black69

Fig. 6 Oxygen release at different lithiation states: (a) NCA powder Li2CO3 + C / Li2O2 + 2CO (25)
upon heating up to 600
C (from ref. 59 and 61) and (b) LFP powder in
electrolyte upon heating to 350
C (from ref. 64). hydrolysis of the hexauorophosphate salt70,71

57180 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015
 View Article Online

Paper RSC Advances

LiPF6 + H2O / 2HF + LiF + POF3 (26) or react with HF analogous to80

POF3 + 2H2O / POF(OH)2 + 2HF (27) C2H5OCO2Li + HF / C2H5OH + LiF + CO2 (33)

with the proposed scheme

5.4 Alternative CO2 producing reactions

CO2 was the main gaseous product that was identied in the
ventgas of discharged cells. Little O2 is available in cells at 0%
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

SOC and it is questionable if combustion alone can account for

all CO2. Therefore effort was made to nd further alternative (34)
reactions with CO2 evolution without oxygen involvement.
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

Following reactions were found in the literature:

Ring-opening and polymerisation of EC and PC72–74 e.g.:

Inorganic SEI can react with HF as well.48,53

(28)
Li2CO3 + 2HF / 2LiF + CO2 + H2O (35)

In the presence of impurities such as trace water LiPF6 may

thermal decomposition of the carbonate esters50,75,76 e.g.: react to POF3 that in turn reacts with the electrolyte in a
decarboxylation reaction with CO2 release:50,77,81–83
LiPF6 /LiF þ PF5
(29) PF5 þ ROH/HF þ RF þ POF3 (36)
POF3 þ solvent/CO2 þ phosphate

or What is the most signicant CO2 production mechanisms in

oxygen depleted environment? In the case of the NCA cell
(Table 1) decomposition of all electrolyte solvent (28)–(30) may
translate to 35.0 mmol CO2. The amount of SEI is lower than the
(30) amount of electrolyte solvents and therefore only 8.7 mmol of
CO2 can be produced with eqn (34) and (35). The reactions
involving HF (33)–(36) may be further suppressed by the limited
amount of trace ROH and LiPF6.

Gnanaraj et al. notes75 that the electrolyte salt decomposes 5.5 Gas release of a discharged NCA cell
In the Experiment 1 a discharged NCA cell was subject to a
LiPF6 4 LiF + PF5 (31)
thermal-ramp test and absolute amounts of produced gas
and the resulting strong Lewis base PF5 lowers the onset-
temperatures of solvent decomposition reactions by acid–base Table 4 Model reaction system for the Experiment 1 (discharged NCA
catalysis. The onset temperature for DEC and DMC decompo- cell)
sition lies at 170
C (ref. 75) and 190
C (ref. 77) respectively. Utilization bl
This values are well below the maximal temperatures reached in Reaction l (mmol)
our thermal-ramp experiments. Electrolyte decomposition with
CO2 release was also observed in other research.5,78 DMCliq / DMCgas 11.6
The maximum amount of CO2 generated from purely ECliq / ECgas 8.5
MPCliq / MPCgas 0.6
thermal decomposition of the electrolyte solvents (28)–(30) is (CH2OCO2Li)2 / Li2CO3 + C2H4 + CO2 + 0.5O2 0.1
only limited by the amount of available electrolyte. CMC + 3.175O2 / 7.4CO + 5.35H2O 1
Further CO2 may be produced from the SEI degradation: The PVDF + Li / (CH]CF) + 0.5H2 + LiF 0.9
organic SEI produced by EC reduction (2) can decompose in 0.5O2 + H2 / H2O 7
thermally driven reactions,79 CO + H2O / CO2 + H2 7.2
DMC / CO2 + CH3OCH3 12.5
1 EMC / CO2 + CH3OC2H5 2.2
ðCH2 OCO2 LiÞ2 /Li2 CO3 þ C2 H4 þ CO2 þ O2 (32) DMC + 2Li + H2 / Li2CO3 + 2CH4 0.2
2

This journal is © The Royal Society of Chemistry 2015 RSC Adv., 2015, 5, 57171–57186 | 57181
 View Article Online

RSC Advances Paper

Table 5 Initially available material in the cell as well as material that is Table 4 was calculated with the linear optimisation algo-
consumed as a reactant according to the proposed reaction system in rithm. It gives one possible set of reactions to reproduce the
Table 4 for the Experiment 1 (discharged NCA cell)
measured value of CO2 and the overall amount of produced gas
Compound j Availibility naj (mmol) Consumption nrj (mmol) in the reactor. Because of the elevated temperature all liquid
solvents present inside the Li-ion cell either decompose or
PE 25 0 evaporate (Table 5). The calculation gives rise to new gaseous
O2 6.7 6.7 components and the amount of those components can be
C 946.4 0
CMC 2.9 1
compared to the actual measurements (Table 6).
In this mathematical solution, the missing 42.2 mmol of gas
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

PVDF 7 0.9
DMCliq 24.3 24.3 consist mainly of solvent decomposition products (CH3OCH3,
EMCliq 2.2 2.2 CH3OC2H5) and remaining solvents as well as water in gaseous
ECliq 8.5 8.5
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

state. Such gases can not be found by the GC system due to

MPCliq 0.6 0.6
LiPF6 3.1 0
following reasons: (a) the sampler of the GC runs at room
(CH2OCO2Li)2 3.5 0.1 temperature and therefore the solvents condense and are not
C6H4O6 1 0
Li 1.3 1.3
Li2CO3 1.7 0 Table 7 Model reaction system for the Experiment 13 (over-charged
LiF 1.7 0 NCA cell)

Reaction l Utilization bl (mmol)

DMC + 1.5O2 / 3CO + 3H2O 10.4

components were measured (Table 3). Gas analysis with GC gave DMC / C2H4 + CO2 + H2O 1.1
23.2 mmol of CO2 and small amounts of H2, CO and hydrocar- EC + O2 / 3CO + 2H2O 8.5
bons. The overall amount of produced gas nideal MPCliq + 4.5O2 / 8CO + 4H2O 0.6
sum inside the
C + 0.5O2 / CO 60.9
heated reactor was 65.4 mmol. This means that the GC system
CO + 0.5O2 / CO2 12.3
was unable to detect 42.2 mmol of unknown gas components. (CH2OCO2Li)2 + 2Li / 2Li2CO3 + C2H4 3.5
What is the source of CO2 and what is the nature of the not C6H4O6 + 2O2 / 6CO + 2H2O 1
identied gas components? The cathode material of a discharged CMC + 3.175O2 / 7.4CO + 5.35H2O 2.9
cell is not fully lithiated and may release a small amount of O2 PVDF + O2 / 2CO + 2HF 7
CO + H2O / CO2 + H2 15.5
(19). The released O2 can participate in a combustion reaction,
2Li + H2O / H2 + Li2O 57.5
but the amount of released O2 is not sufficient to produce DMC / CO2 + CH3OCH3 5.3
all measured CO2 ðnaO2 ¼ 6:7 mmol vs: nGC CO2 ¼ 21:9 mmolÞ. We EMC / CO2 + CH3OC2H5 2.2
needed to consider alternative reactions in order to account for DMC + 2Li + H2 / Li2CO3 + 2CH4 7.5
the measured amounts of gases. Li2CO3 / Li2O + CO2 3.1
Li2CO3 + 2HF / 2LiF + CO2 + H2O 13.1
LiPF6 + H2O / 2HF + LiF + POF3 3.1
POF3 + 2H2O / POF(OH)2 + 2HF 3.1

Table 6 Produced material in the Experiment 1 (discharged NCA cell):

calculated values of the model reaction system according to Table 4,
values quantified by the GC system and the overall measured amount
of produced gas inside the reactor Table 8 Initially available material in the cell as well as material that is
consumed as a reactant according to the proposed reaction system in
Calculated From GC nideal
sum Table 7 for the Experiment 13 (over-charged NCA cell)
Compound i npi (mmol) nGC
i (mmol) (mmol)
Compound j Availibility naj (mmol) Consumption nrj (mmol)
O2 0 0.1 —
–[CH]CF]– 0.9 — — PE 25 0
Li2CO3 0.3 — — O2 81.6 81.6
DMCgas 11.6 — — C 946.4 60.9
ECgas 8.5 — — CMC 2.9 2.9
MPCgas 0.6 — — PVDF 7 7
CH3OCH3 12.5 — — DMCliq 24.3 24.3
CH3OC2H5 2.2 — — EMCliq 2.2 2.2
H2O 5.3 — — ECliq 8.5 8.5
CO 0.4 0.4 — MPCliq 0.6 0.6
CO2 21.9 21.9 — LiPF6 3.1 3.1
H2 0.4 0.4 — (CH2OCO2Li)2 3.5 3.5
CH4 0.4 0.4 — C6H4O6 1 1
C2H2 0.1 0.1 — Li 151 137
Sum of gaseous 63.9 23.2 65.4 Li2CO3 1.7 1.7
products LiF 1.7 0

57182 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015
 View Article Online

Paper RSC Advances

Table 9 Produced material in the Experiment 13 (over-charged NCA To nd one of the possible solutions explaining the
cell): calculated values of the model reaction system according to measured gas composition the equations in Table 7 were used.
Table 7, values quantified by the GC system and the overall measured
In this exemplary mathematical solution all electrolyte solvents,
amount of produced gas inside the reactor
organic SEI, lithium carbonate and the released O2 were
Calculated From GC nideal
sum consumed (Table 8). The major products in the calculation were
Compound i npi (mmol) nGC
i (mmol) (mmol) the gases as measured by the GC as well as the solids LiF and
Li2O. The overall amount of measured gas nideal sum and the
LiF 29.3 — —
POF(OH)2 3.1 — —
amounts of the individual gas components nGC i could be
— — reproduced by the calculation (Table 9). The quantity of LiF and
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Li2O 60.6
CH3OCH3 5.3 — — Li2O was not measured and therefore not veried by the
CH3OC2H5 2.2 — — experiments.
CO 136 136 —
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

The major reactions which were responsible for the gas and
CO2 52.6 57.9 —
H2 65.5 65.5 —
heat production during thermal runaway are summarized in a
CH4 15.1 15.1 — simplied picture (Fig. 7). In this scheme the released oxygen
C2H4 4.6 4.6 — triggers a chain of exothermic reactions. Because of O2 insuffi-
Sum of gaseous 281.3 279.1 281.3 ciency incomplete combustion of organic material takes place.
products The resulting H2O reacts with the exposed Li with H2 produc-
tion. Simultaneously H2 and CO2 are produced with the water–
gas shi reaction. In the end the main gases are CO, CO2 and
injected into the GC column and (b) the present GC setup is not H2.
designed to identify and quantify any components other than Although the calculation shows good agreement of
H2, O2, N2, CO, CO2, CH4, C2H2, C2H4 and C2H6. measured and computed amounts of gas it has some aws:
(a) the full amount of CO2 could not be reproduced (b) it is not
considered, that the separator material must decompose and
5.6 Gas release of a charged NCA cell add additional gas volume at temperatures >900
C and (c) in
reality the reactants are not distributed homogeneously when
The situation changes when the cells in thermal ramp experi-
the reactions take place, instead material is violently expelled
ments are charged. High amounts of oxygen and lithium
from the cell into the reactor during thermal runaway. Further
become available and the cells go into distinct thermal
work is needed to take those effects into account.
runaway. In the Experiment 13 (Table 3) an overcharged NCA
cell was tested and the vent-gases were quantied by the GC
system using an internal N2 standard.
The cell in Experiment 13 was overcharged to a capacity of 6 Conclusions and outlook
4.03 A h (120% SOC). The lithiation factor of the cathode was
We measured the thermal runaway characteristics of commer-
xcat
Li ¼ 0.08 and the calculated oxygen release (17) was
cial Li-ion cells in destructive thermal ramp experiments in
naO2 ¼ 81:6 mmol. The lithiation factor of the anode was xand
Li ¼
inert atmosphere. Our samples were 23 NCA and LFP based
1.00 and the amount of intercalated Li on the anode side
Li-ion batteries with the geometrical format 18650 charged to
equalled naLi ¼ 151 mmol. In other words, the anode was fully
different SOC. The main ndings of this work are:
lithiated to the maximum theoretical Li capacity. The cell
(1) The cell material and cell design (e.g. high energy density
produced 281.3 mmol of ventgas during thermal runaway and
vs. high power density) have a high inuence on the maximum
high amounts of CO, H2, CO2 were detected.
cell temperature and on the released gases in thermal runaway
conditions (Table 10). Charged NCA cells showed a drastic
thermal runaway behaviour. NCA cells could reach maximum
temperatures of 1075
C and they released up to 317 mmol of
gas (equal to 7.1 L at standard conditions). Charged LFP cells
exhibited a less pronounced thermal runaway: maximum cell
temperatures as high as 448
C were observed and the LFP cells
released up to 61 mmol of gas.
(2) Discharged cells showed no thermal runaway upon
heating up to 250
C. Both cell types needed to be at least
partially charged in order to go into thermal runaway.
(3) The severity of the thermal runaway increased with
increasing SOC.
(4) The thermal runaway reactions produced high amounts
of CO, H2 and CO2 thus making the gas ammable and
Fig. 7 Proposed main reaction system for a thermal runaway of a potentially toxic. The gas composition depended on the cell
(partially) charged or over-charged Li-ion battery. type and SOC. NCA cells produced more CO and H2 than LFP

This journal is © The Royal Society of Chemistry 2015 RSC Adv., 2015, 5, 57171–57186 | 57183
 View Article Online

RSC Advances Paper

Table 10 Comparison of the two tested cell types. Cell specifications including misuse of battery packs for electric vehicles, airliners
and averaged experimental results and for home storage of solar energy.
Our future work in the next three years will include (a)
LFP NCA
additional testing of 18650 cells in an improved test rig, (b)
Voltage (50% SOC) V 3.32 3.68 experiments with large automotive Li-ion cells in a new large
Nominal capacity Ah 1.1 3.35 test rig, (c) bottom up thermodynamic calculations of the
Cell mass g 38.87 45.40 chemical reaction systems and (d) top down FEM simulation of
Max continuous discharge A 30 6.7
failure propagation and the reaction kinetics in large battery
Cycle life >1000 >300
packs.
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Min. SOC for a pronounced % 50 25

thermal runaway


Onset temperature qo (100% SOC) C 140 139
Max. temperature qm (100% SOC)

C 440 911
Nomenclature
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

Produced gas nideal

sum (100% SOC) mmol 32 277
Detected CO2 (100% SOC) % 48.3 20.0
Dm Mass loss of the cell, caused by temperature ramp
Detected CO (100% SOC) % 9.1 44.8
Detected H2 (100% SOC) % 29.4 25.7 experiment (g)
Impact of overcharge Weak Strong q Cell temperature (
C)
qm Maximum cell temperature during the temperature ramp
experiment (
C)
qo Onset temperature of the thermal runaway (
C)
qgas Gas temperature inside the reactor (K)
cells. Discharged cells generated mainly CO2. Increased SOC
Cirr Charge associated with nirr
Li (A h)
led to increased amounts of CO and H2.
Cnom Typical cell capacity as specied in the datasheet (A h)
(5) The SOC and the cell type had less effect on the onset
Cres Charge associated with nres
Li (A h)
temperature, as long as no Li plating occurred. Overcharge may
Cuand Theoretically usable capacity of the anode (A h)
cause metallic Li deposition on the anode which compromises Cucat Theoretically usable capacity of the cathode (A h)
the thermal stability. The onset temperature of overcharged F Faraday constant (F ¼ 96 485 A s mol1)
cells decreased dramatically from 140
C to values as low as nai Amount of substance i in a pristine cell, at the start of a
65
C! thermal ramp experiment (mol)
(6) In three experiments, the absolute amounts of gases from npi Theoretically calculated amount of product i, which is
NCA cells were quantied. It is shown, that it is theoretically produced by chemical reactions during the thermal
possible to explain the absolute amounts of the measured gases runaway (mol)
with a set of chemical degradation reactions and with known naj Amount of material j in the cell, that is available for
amounts of initial material in the cell. chemical reactions during the thermal runaway (mol)
(7) We think that the main reactions in charged cells are nacat Amount of either LFP or NCA units in the cathode
combustion of carbonous material and Li oxidation. Both are nGC
i Absolute amount of gas component i in the reactor,
strong exothermic reactions which contribute to the energy calculated from GC results (mol)
release during the thermal runaway of a Li-ion battery. The nGC
sum Amount of gas produced by a cell during a temperature
amounts of O2 and Li available to degradation reactions ramp experiment, calculated from GC results (mol)
depend on the SOC as well as on the amount and type of active nideal
sum Amount of gas produced by a cell during a temperature
cathode and anode material. Higher SOC increases the O2 ramp experiment, calculated with the ideal gas law (mol)
release of the cathode and the amount of intercalated Li in the nirr
Li Amount of irreversibly trapped Li in the anode caused by
anode. In over-charged cells these amounts increase further initial cell formation (mol)
nres
Li Amount of residual Li in the anode of a cell which is
and deposition of highly reactive metallic Li may occur on the
discharged to Vmin (mol)
anode.
nrj Theoretically calculated amount of reagent j in the cell,
(8) It is proposed that both, the cathode and anode side
which is consumed by chemical reactions during the
participate in the reaction system. Therefore experiments with
thermal runaway (mol)
only one electrode may not cover the full picture.
n0 Initial amount of gas in the reactor at the start of the
Many open questions concerning the safety of Li-ion experiment (mol)
batteries remain. The industry needs scaling rules to evaluate nN2 Actual amount of N2 in the reactor (mol)
the safety of large battery systems with hundreds of cells based P Gas pressure in the reactor (Pa)
on results of misuse experiments with individual cells. Many R Gas constant (R ¼ 8.314 J mol1 K1)
test results exist for small 18650 cells but we think that more rGC
i Result of GC measurement: fraction of gas component i
effort must be made to understand the thermal runaway in the GC sample (mol%)
behaviour of large cells with capacities as high as 60 A h. It is yet T Time (s)
to prove, if specic amount of gas and heat are the same for V Volume of the reactor (m3)
small and large cells. The risks of re and toxicity (including Vmin Minimum cell voltage as specied in the datasheet (V)
HF) of vent gas must be quantied for real life applications x Lithiation factor of the anode or cathode

57184 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015
 View Article Online

Paper RSC Advances

20 P. Huang, Q. Wang, K. Li, P. Ping and J. Sun, Sci. Rep., 2015,

Acknowledgements
5, 7788.
VIRTUAL VEHICLE Research Center is funded within the 21 H. Wang, A. Tang and K. Wang, Chin. J. Chem., 2011, 29, 27–
COMET Competence Centers for Excellent Technologies pro- 32.
gramme by the Austrian Federal Ministry for Transport, Inno- 22 C.-Y. Wen, C.-Y. Jhu, Y.-W. Wang, C.-C. Chiang and
vation and Technology (BMVIT), the Federal Ministry of C.-M. Shu, J. Therm. Anal. Calorim., 2012, 109, 1297–1302.
Science, Research and Economy (BMWFW), the Austrian 23 T.-Y. Lu, C.-C. Chiang, S.-H. Wu, K.-C. Chen, S.-J. Lin,
Research Promotion Agency (FFG), the province of Styria and C.-Y. Wen and C.-M. Shu, J. Therm. Anal. Calorim., 2013,
the Styrian Business Promotion Agency (SFG). The COMET 114, 1083–1088.
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

programme is administrated by FFG. We would furthermore 24 C.-Y. Jhu, Y.-W. Wang, C.-M. Shu, J.-C. Chang and H.-C. Wu,
like to express our thanks to our supporting scientic project J. Hazard. Mater., 2011, 192, 99–107.
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

partners, namely “Varta Micro Innovation GmbH” and to the 25 C.-Y. Jhu, Y.-W. Wang, C.-Y. Wen, C.-C. Chiang and
Graz University of Technology. The work was made possible by C.-M. Shu, J. Therm. Anal. Calorim., 2011, 106, 159–163.
the principal industrial project partners “BASF SE” and by 26 H. Ishikawa, O. Mendoza, Y. Sone and M. Umeda, J. Power
“MAGNA STEYR Battery Systems GmbH & Co OG”. Sources, 2012, 198, 236–242.
27 E. P. Roth and D. H. Doughty, J. Power Sources, 2004, 128,
308–318.
References 28 S.-i. Tobishima and J.-i. Yamaki, A consideration of lithium
cell safety, 1999.
1 K. Xu, Chem. Rev., 2004, 104, 4303–4417. 29 D. Belov and M.-H. Yang, Solid State Ionics, 2008, 179, 1816–
2 Z. J. Zhang and P. Ramadass, in Encycl. Sustain. Sci. Technol., 1821.
Springer New York, New York, NY, 2012, pp. 319–356. 30 Y. Zeng, K. Wu, D. Wang, Z. Wang and L. Chen, J. Power
3 J. Wen, Y. Yu and C. Chen, Mater. Express, 2012, 2, 197–212. Sources, 2006, 160, 1302–1307.
4 D. Doughty and E. P. Roth, Electrochem. Soc. Interface, 2012, 31 T. Ohsaki, T. Kishi, T. Kuboki, N. Takami, N. Shimura,
21, 37–44. Y. Sato, M. Sekino and A. Satoh, J. Power Sources, 2005,
5 E. P. Roth and C. J. Orendorff, Electrochem. Soc. Interface, 146, 97–100.
2012, 21, 45–49. 32 J. Jeevarajan, in Lithium-Ion Batter. Adv. Appl., Elsevier, 2014,
6 D. Abraham, E. P. Roth, R. Kostecki, K. McCarthy, pp. 387–407.
S. MacLaren and D. Doughty, J. Power Sources, 2006, 161, 33 C. Orendorff, Electrochem. Soc. Interface, 2012, 21, 61–65.
648–657. 34 R. a. Leising, M. J. Palazzo, E. S. Takeuchi and K. J. Takeuchi,
7 Z. Chen, Y. Qin, Y. Ren, W. Lu, C. Orendorff, E. P. Roth and J. Electrochem. Soc., 2001, 148, A838.
K. Amine, Energy Environ. Sci., 2011, 4, 4023. 35 Z. Li, J. Huang, B. Yann Liaw, V. Metzler and J. Zhang, J.
8 G. Nagasubramanian and C. J. Orendorff, J. Power Sources, Power Sources, 2014, 254, 168–182.
2011, 196, 8604–8609. 36 A. W. Golubkov, D. Fuchs, J. Wagner, H. Wiltsche, C. Stangl,
9 H. Maleki, G. Deng, A. Anani and J. Howard, J. Electrochem. G. Fauler, G. Voitic, A. Thaler and V. Hacker, RSC Adv., 2014,
Soc., 1999, 146, 3224. 4, 3633.
10 P. Ping, Q. Wang, P. Huang, J. Sun and C. Chen, Appl. Energy, 37 J.-H. Lee, U. Paik, V. a. Hackley and Y.-M. Choi, J.
2014, 129, 261–273. Electrochem. Soc., 2005, 152, A1763.
11 P. Röder, B. Stiaszny, J. C. Ziegler, N. Baba, P. Lagaly and 38 S.-L. Chou, Y. Pan, J.-Z. Wang, H.-K. Liu and S.-X. Dou, Phys.
H.-D. Wiemhöfer, J. Power Sources, 2014, 268, 315–325. Chem. Chem. Phys., 2014, 16, 20347–20359.
12 M. Fleischhammer, T. Waldmann, G. Bisle, B.-I. Hogg and 39 Z. Chen and J. R. Dahn, Reducing Carbon in LiFePO4/C
M. Wohlfahrt-Mehrens, J. Power Sources, 2015, 274, 432–439. Composite Electrodes to Maximize Specic Energy, Volumetric
13 D. H. Doughty, E. P. Roth, C. C. Cras, G. Nagasubramanian, Energy, and Tap Density, 2002.
G. Henriksen and K. Amine, J. Power Sources, 2005, 146, 116– 40 J. Wang and X. Sun, Energy Environ. Sci., 2012, 5, 5163.
120. 41 H. Ota, Y. Sakata, A. Inoue and S. Yamaguchi, J. Electrochem.
14 J. Lamb and C. J. Orendorff, J. Power Sources, 2014, 247, 189– Soc., 2004, 151, A1659.
196. 42 J. C. Burns, R. Petibon, K. J. Nelson, N. N. Sinha, A. Kassam,
15 G. Trattnig and W. Leitgeb, in Automot. Batter. Technol., B. M. Way and J. R. Dahn, J. Electrochem. Soc., 2013, 160,
Springer, 2014, pp. 19–37. A1668–A1674.
16 X. Liu, S. I. Stoliarov, M. Denlinger, A. Masias and K. Snyder, 43 S. Zhang, J. Power Sources, 2006, 162, 1379–1394.
J. Power Sources, 2015, 280, 516–525. 44 E. Peled, D. Golodnitsky, A. Ulus and V. Yut, Electrochim.
17 P. Ping, Q. Wang, P. Huang, K. Li, J. Sun, D. Kong and Acta, 2004, 391–395.
C. Chen, J. Power Sources, 2015, 285, 80–89. 45 K. Ushirogata, K. Sodeyama, Y. Okuno and Y. Tateyama, J.
18 X. Feng, M. Fang, X. He, M. Ouyang, L. Lu, H. Wang and Am. Chem. Soc., 2013, 135, 11967–11974.
M. Zhang, J. Power Sources, 2014, 255, 294–301. 46 L. El Ouatani, R. Dedryvere, C. Siret, P. Biensan, S. Reynaud,
19 F. Larsson and B.-E. Mellander, J. Electrochem. Soc., 2014, P. Iratcabal and D. Gonbeau, J. Electrochem. Soc., 2009, 156,
161, A1611–A1617. A103.

This journal is © The Royal Society of Chemistry 2015 RSC Adv., 2015, 5, 57171–57186 | 57185
 View Article Online

RSC Advances Paper

47 G. Gachot, S. Grugeon, G. G. Eshetu, D. Mathiron, P. Ribière, 66 W.-J. Zhang, J. Power Sources, 2011, 196, 2962–2970.
M. Armand and S. Laruelle, Electrochim. Acta, 2012, 83, 402– 67 O. Haik, S. Ganin, G. Gershinsky, E. Zinigrad, B. Markovsky,
409. D. Aurbach and I. Halalay, J. Electrochem. Soc., 2011, 158,
48 M. Onuki, S. Kinoshita, Y. Sakata, M. Yanagidate, Y. Otake, A913.
M. Ue and M. Deguchi, J. Electrochem. Soc., 2008, 155, A794. 68 Y. He, Z. Tang, Q. Song, H. Xie, Q. Xu, Y. Liu and G. Ling,
49 D. Aurbach, A. Zaban, Y. Ein-Eli, I. Weissman, O. Chusid, Thermochim. Acta, 2008, 480, 15–21.
B. Markovsky, M. Levi, E. Levi, A. Schechter and E. Granot, 69 J.-w. Kim and H.-g. Lee, Metall. Mater. Trans. B, 2001, 32, 17–
J. Power Sources, 1997, 68, 91–98. 24.
50 G. Gachot, P. Ribière, D. Mathiron, S. Grugeon, M. Armand, 70 R. Yazami and A. Martinent, in Fluorinated Mater, Energy
This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

J. B. Leriche, S. Pilard and S. Laruelle, Anal. Chem., 2011, 83, Conversion, Elsevier Science, 1st edn, 2005, ch. Fluorinate,
478–485. pp. 173–194.
Open Access Article. Published on 22 June 2015. Downloaded on 30/10/2015 07:21:32.

51 R. Spotnitz and J. Franklin, J. Power Sources, 2003, 113, 81– 71 L. Terborg, S. Weber, F. Blaske, S. Passerini, M. Winter,
100. U. Karst and S. Nowak, J. Power Sources, 2013, 242, 832–837.
52 H. Yoshida, T. Fukunaga, T. Hazama, M. Terasaki, 72 L. Vogdanis and W. Heitz, Makromol. Chem., Rapid Commun.,
M. Mizutani and M. Yamachi, J. Power Sources, 1997, 68, 1986, 7, 543–547.
311–315. 73 L. Vogdanis, B. Martens, H. Uchtmann, F. Hensel and
53 C. Doh, D. Kim, J. Lee, D. Lee, B. Jin, H. Kim, S. Moon, W. Heitz, Makromol. Chem., Rapid Commun., 1990, 191,
Y. Hwang and A. Veluchamy, Bull. Korean Chem. Soc., 2009, 465–472.
30, 783. 74 S. E. Sloop, J. B. Kerr and K. Kinoshita, J. Power Sources, 2003,
54 K. Tasaki and S. J. Harris, J. Phys. Chem. C, 2010, 114, 8076– 119–121, 330–337.
8083. 75 J. S. Gnanaraj, E. Zinigrad, L. Asraf, H. E. Gottlieb,
55 H. Maleki and J. Howard, J. Power Sources, 2006, 160, 1395– M. Sprecher, M. Schmidt, W. Geissler and D. Aurbach, J.
1402. Electrochem. Soc., 2003, 150, A1533.
56 H. He, Y. Liu, Q. Liu, Z. Li, F. Xu, C. Dun, Y. Ren, M.-x. Wang 76 G. Gachot, S. Grugeon, I. Jimenez-Gordon, G. G. Eshetu,
and J. Xie, J. Electrochem. Soc., 2013, 160, A793–A804. S. Boyanov, A. Lecocq, G. Marlair, S. Pilard and S. Laruelle,
57 X. Huang, J. Solid State Electrochem., 2010, 15, 649–662. Anal. Methods, 2014, 6, 6120.
58 C. T. Love, J. Power Sources, 2011, 196, 2905–2912. 77 T. Kawamura, A. Kimura, M. Egashira, S. Okada and
59 S.-m. Bak, K.-w. Nam, W. Chang, X. Yu, E. Hu, S. Hwang, J.-I. Yamaki, J. Power Sources, 2002, 104, 260–264.
E. A. Stach, K.-B. Kim, K. Y. Chung and X.-Q. Yang, Chem. 78 G. G. Botte and T. J. Bauer, J. Power Sources, 2003, 119–121,
Mater., 2013, 25, 337–351. 815–820.
60 K.-W. Nam, S.-M. Bak, E. Hu, X. Yu, Y. Zhou, X. Wang, L. Wu, 79 M. N. Richard and J. Dahn, J. Electrochem. Soc., 1999, 146,
Y. Zhu, K.-Y. Chung and X.-Q. Yang, Adv. Funct. Mater., 2013, 2068.
23, 1047–1063. 80 W. Kong, H. Li, X. Huang and L. Chen, J. Power Sources, 2005,
61 I. Belharouak, W. Lu, D. Vissers and K. Amine, Electrochem. 142, 285–291.
Commun., 2006, 8, 329–335. 81 A. Hammami, N. Raymond and M. Armand, Nature, 2003,
62 H. J. Bang, H. Joachin, H. Yang, K. Amine and J. Prakash, J. 424, 635–636.
Electrochem. Soc., 2006, 153, A731. 82 C. L. Campion, W. Li and B. L. Lucht, J. Electrochem. Soc.,
63 Z. Li, D. Zhang and F. Yang, J. Mater. Sci., 2009, 44, 2435– 2005, 152, A2327.
2443. 83 C. L. Campion, W. Li, W. B. Euler, B. L. Lucht, B. Ravdel,
64 P. Röder, N. Baba, K. Friedrich and H.-D. Wiemhöfer, J. J. F. DiCarlo, R. Gitzendanner and K. M. Abraham,
Power Sources, 2013, 236, 151–157. Electrochem. Solid-State Lett., 2004, 7, A194.
65 J. Kim, K.-Y. Park, I. Park, J.-K. Yoo, J. Hong and K. Kang, J.
Mater. Chem., 2012, 22, 11964.

57186 | RSC Adv., 2015, 5, 57171–57186 This journal is © The Royal Society of Chemistry 2015

View publication stats