Chapter 4

Overdischarging Lithium-Ion Batteries

Daniel Hauck and Michael Kurrat

Abstract A main objective of the project LithoRec II is to study the overdischarge of

Li-Ion batteries for the purpose of recycling. For safety and functional reasons, the
batteries need to be overdischarged before undergoing the process steps of disas-
sembly and crushing. In this study, various devices for overdischarging are set up and
investigated together with corresponding battery behavior. Device properties such as
cost, safety, discharge time and discharge flexibility are assessed. Furthermore, the
study focused on parameters like battery heating during discharge and pole reversal,
relaxation amounts and heights as well as short circuit currents. The investigations
show, that it is most advisable to discharge whole battery systems by energy recu-
peration into grid with electronic adjustable loads, because of efficiency and safety
reasons. Overdischarging is not problematic if crucial battery parameters are
observed. It is recommended that overdischarging of automotive traction batteries
should only be done by high-voltage specialists.

4.1 Introduction

4.1.1 Overdischarging Batteries in the Literature

To date, relatively little has been reported on overdischarging lithium-ion batteries

(LIBs). This is possibly due to the fact that this field has been of little interest to
researchers and its importance has not even been questioned. LithoRec recycling as
well as electrical hazards are the key areas of focus forming the motivational
background to this project. A further aspect of studying overdischarging is cell
destruction or accelerated aging which can arise as a result of faulty programming in

D. Hauck (&)
M. Kurrat

Institute of High Voltage Technologies and Electrical Power Systems, TU Braunschweig,
Brunswick, Germany
e-mail: [email protected]

© Springer International Publishing AG 2018 53

A. Kwade and J. Diekmann (eds.), Recycling of Lithium-Ion Batteries,
Sustainable Production, Life Cycle Engineering and Management,
https://doi.org/10.1007/978-3-319-70572-9_4
54 D. Hauck and M. Kurrat

the battery management unit (BMU). Again, there is not much literature on LIBs,
whereas much is known about other battery technologies such as lead-acid batteries
in which undesired overdischarging may occur during stationary storage (Blank
2012). Similar investigations have however been conducted for NiCd and NiMH
batteries (Jossen et al. 2016). The chemical hazard of overdischarging compared to
overcharging, especially for LIBs, is much less which may be one of the reasons for
it not being the focus of much interest in the literature.

4.1.2 Motivation for Studying Overdischarging in LithoRec

For safety and operational capability reasons, overdischarging LIBs is a necessary

aspect of the LithoRec recycling chain, in particular for the later processes of
dismantling and shredding batteries.
Battery systems typically charged to around 400 V (DC) pose a potentially
life-threatening hazard to operators when they are dismantled. Generally, working
under high voltage (more than 60 V DC) is permitted only for high-voltage
(HV) specialists according to DGVU 200-005 (previously BGI 8686). Discharging
a battery system is preferable prior to dismantling it, as a discharged battery system
does not require HV specialists. Furthermore, the discharging of the system does
not entail much effort compared to discharging singular modules or cells. Therefore,
discharging at battery system level should be the preferred option.
Using overdischarged batteries is also very important for their subsequent
shredding. The mechanical pulp of charged batteries may cause local short circuits
which may then lead to local heating or sparking. The electrolyte-gas atmosphere in
a shredder is extremely flammable. As a consequence, only discharged batteries
should be shredded.

4.1.3 Individual Process Steps of Overdischarging

in LithoRec Recycling

Normally, batteries that have been overdischarged are no longer of any use, so that
their subsequent recycling is a reasonable last step in a battery’s life cycle. There is
an increasing effort to emphasize this aspect in industry, research and politics. The
concept of “prolonging or reusing automotive traction batteries after their first
application as second-life batteries” is becoming more widespread. From economic
and more importantly from environmental standpoints, there are significant argu-
ments for promoting this aspiration. To make a decision about which option is most
reasonable—recycling or a second life—the state of an old battery should be
assessed. Figure 4.1 shows a proposed scheme to find out how to meet this
challenge.
4 Overdischarging Lithium-Ion Batteries 55

Fig. 4.1 Flowchart proposing how to deal with LIBs before and after overdischarging

For the purpose of assessing the state of a battery, it is reasonable to classify the
battery into a system, modules and cells. Usually, the BMU which gives data on the
state of a cell, is designed according to master/slave architecture. In battery systems,
it is possible to recapture the lifetime data about the state of the BMU master if
there is access to the information and the respective reading equipment. In this way,
it is possible to undertake sorting and storage according to the charge level, energy
or cell chemistry. However, with modules and cells, such reading of data is not
possible, so that the characterization has to be done manually.1 Ultimately, all
batteries, even so-called second life batteries, need to go through the process of
recycling, meaning that overdischarging and short circuits are inevitable.

1
The manual characterization of the state will not be described here because overdischarging is the
focus of this paper.
56 D. Hauck and M. Kurrat

4.1.4 Focus of the Research

In addressing the subject of “overdischarging LIBs”, several questions are raised,

for example:
• What does overdischarging mean and what happens at pole reversal?
• With which technique or equipment can a battery be overcharged or pole
reversed?
• How long does overdischarging need and how high is the voltage reversal?
• How safe is overdischarging and what risks are to be encountered?
• What is the cost of discharging and how much profit can be gained through
energy recovery?
• At what point of discharging is it safe to connect the battery to a short circuit?
The following sections attempt to answer these and further questions.
Section 4.2 explains the basics of overdischarging LIBs under electrical and elec-
trochemical aspects. Section 4.3 presents the applied battery systems, the applied or
developed instruments for discharging, as well as the measuring setups for
overdischarge investigations. Section 4.4 presents and discusses the results. The
key findings are summarized in Sect. 4.5 and an outlook on future challenges is also
given.

4.2 The Basics of Overdischarging Lithium-Ion Batteries

4.2.1 Definition of Overdischarging

Generally, overdischarging a battery means draining the current after the end-point
voltage, which leads to lasting damage (Bauer 2017). The following figure illus-
trates this situation.
In the example given in Fig. 4.2, overdischarging starts below 2.50 V.
Depending on the discharge current and the temperature, approx. 20% of additional
capacity can be discharged at up to 0 V. In the case that discharging will be
continued to the negative voltage range (below 0 V), a so-called pole reversal will
occur. In this case, the anode and cathode change polarity and thus the anode will
become the cathode and vice versa.

4.2.2 Electrochemical Basis

LIBs are electrochemical storages in which the lithium-ions—the charge carriers

migrate between the anode and the cathode. During discharging, the lithium-ions
are released from the anode and react with the electrons which flow via the external
4 Overdischarging Lithium-Ion Batteries 57

Fig. 4.2 Definition of “overdischarging” and “pole reversal”

load circuit, thus forming a chemical compound at the cathode. While overdis-
charging, all lithium-ions finally migrate from the anode to the cathode. After this
point, forced overdischarging makes the battery try to continue delivering charged
ions. Thus, copper ions dissolve from the anode current collector and take over the
ion transport. This leads, on the one hand, to the formation of dendrites which
pierce the separator leading to short circuits. On the other hand, metallic copper is
deposited on the graphite coating in the separator. Additionally, traces of corrosion
can be formed on the graphite-coated copper current collector (Gellerich 2015).
The latter has to be dealt with during the subsequent processing steps (separa-
tion, leaching by treatment with bases, reprocessing). In addition, temperatures up
to 120 °C have been measured at the cell wall while overdischarging, which is
dangerous due to the possibility of starting an exothermic reaction of the cell
(Gellerich 2015).
In other overdischarging trials, the composition of the gases was measured
during overcharging and overdischarging processes. It was found that the cell
heated up by 29 °C; it expanded in volume and various gases were formed when
batteries were being overdischarged. The largest proportion of gas at 47% was
hydrogen, which is highly inflammable, and can form explosive oxyhydrogen with
air. Other detected gases were: carbon monoxide, carbon dioxide, methane, ethane
and ethylene (Mähliß 2012).
58 D. Hauck and M. Kurrat

4.2.3 Electrical Basis of Overdischarging

In order to discharge a battery, it is necessary to connect an electric load to the main

poles. The electrons from the anode migrate via the external circuit and via the
electric consumer load to the cathode where they react with the lithium-ions. The
electric load can show different types of electrical behavior depending on the dis-
charge device/technique. Essentially, different modes of discharge behavior exist, as
qualitatively shown in Fig. 4.3.
Furthermore, different ways of controlling can also be undertaken successively
depending on the type of discharge equipment. The control types offer different
advantage/disadvantages and play a major part in the selection of the overdis-
charging process. The following section explains other important criteria for the
selection of the overdischarging process.

4.2.4 Criteria for Description and Selection

of Overdischarging Processes

There are a number of different technical ways (see Sect. 4.3.2) of discharging
electrical vehicle (EV) batteries. These options all have advantages and disadvan-
tages. Some of the important factors for selecting and defining the discharging
process are:
• the cost of the discharging device
– the acquisition cost
– the cost of (e.g. current, cooling liquid) or income from (reimbursement
through power recovery) the running process
– maintenance and repair costs
• other costs
– power lines for short circuit of cells, modules or systems
– storage areas for cells, modules or systems
– HV-qualified personnel

Fig. 4.3 Different basic modes of electrical behaviour for overdischarging

4 Overdischarging Lithium-Ion Batteries 59

• speed of discharging
• safety of discharging
• flexibility in the formation of the discharge characteristic curve
• complexity and ease of handling of discharging device
• automation of the discharging device
• taking into account the material recycling share depending on the discharging
strategy ! discharging at high speed might lead—similarly to “fast discharg-
ing”—to a chemical deposition of Li or Cu which cannot be recovered in a good
yield.
• mobility of the discharging device
• sufficient power supply depending on the load (only in the case of power
regeneration).
For an optimal discharging strategy, the criteria listed above should be assessed
in a holistic approach.

4.3 Investigated Batteries and Devices

for Overdischarging

The devices and batteries for discharging, which were used and developed during
the project, will be explained in the following section. Batteries as well as devices
for discharging were studied during the research. The batteries can be divided into
battery cells, modules and systems. The devices for discharging can be differenti-
ated by energy transformation type. The following Fig. 4.4 illustrates this.
While the discharging of cells and modules was tested with devices that were
based on energy transformation into heat, the discharging of systems was addi-
tionally tested by energy transformation through power recovery. The following
section describes in detail both the batteries and devices investigated as well as the
experiments conducted.

4.3.1 Battery Cells, Modules, and Systems Tested

The tested batteries were provided by Volkswagen AG and AUDI AG who were
the project partners. The batteries were from recent automotive generations that
were produced at the time of the research. An overview of the cells, modules, and
systems used is given in Tables 4.1 and 4.2.
The tables above illustrate the different variations in cell and module connec-
tions, depending on whether the application is for electric vehicles (EVs), hybrid
electric vehicles (HEVs) or plug-in hybrid electric vehicles (PHEVs).
Whereas the system of the VW e-Golf consists of a triple parallel connection
(higher capacity or rather driving range), the HEV and PHEV are only connected in
60 D. Hauck and M. Kurrat

Fig. 4.4 Devices used and their energy transformation concepts to overdischarge cells, modules,
and systems

Table 4.1 Battery cells, modules, systems investigated at the time of the research
BEV PHEV HEV
VW e-Golf VW Passat GTE AUDI Q5 Hybrid
System

Modules

Cells

series (higher voltage or rather power to accelerate/recuperate). This information is

helpful for calculating the maximum tolerable discharge current. It can also be
concluded that the systems exhibit nearly the same voltage, though the hybrid
variants possess much less energy.
Table 4.2 Most important battery data to be known for overdischarging
Property Unit BEV PHEV HEV
VW e-Gotf VW Passat GTE AUDI Q5 Hybrid
Cell Module System Cell Module System Cell Module System
Chemistry – NMC NCA NMC
Voltage Min. V 3 12 264 3 36 288 2.5 45 144
4 Overdischarging Lithium-Ion Batteries

Max. 4.1 16.4 360 4.1 49.2 394 4.3 73.8 309.6
Current Max. discharge A −116.6 −350 −350 −255 −255 −255 −1801 −1801 −1801
C-rate – −4.7 −4.7 −4 7 −9 1 −9.1 −9.1 −581 −581 −581
Nominal energy Wh 89 1.065 23 400 99 1.194 9 548 11.4 205 820
Capacity Ah 25 75 75 28 28 28 3.1 3.1 3.1
Weight Kg 0.72 10.5 320 0.72 10.5 155 <0.53 <9.5 38
Amount – 264 16 & 12 1 96 8 1 72 4 1
Connection – – 3p4s & 3p2s 3p88 s – 12s1p 96s1p – 18s1p 72s1p
1
Represents max. pulse discharge current for 10s
61
62 D. Hauck and M. Kurrat

4.3.2 Applied Devices and Measuring Setup

for Overdischarging

To overdischarge the batteries, special overdischarging setups had to be developed

and constructed and then compared to commercially available loads.
Discharging with constant resistance
Obviously, the simplest way of discharging a battery is with constant resistance.
Those variants, A, B and C, indicated in Figs. 4.5, 4.6 and 4.7 involve the CR
control mode.
(A) Single cell resistor
The cells are connected directly to a resistor. The current flow can be calculated
according to Ohm’s law. The load resistance applied is 1.0 Ω and also 0.5 Ω, so that
the maximum current of 10 A is not exceeded. To conduct the heat away, the load
resistor is placed in a cooling body which is mounted on a cooling plate.

Fig. 4.5 Single cell resistor (variant A)

4 Overdischarging Lithium-Ion Batteries 63

Fig. 4.6 Selectable resistor device (variant B)

(B) Selectable resistors

An extension of the singular resistance connection is the discharger with selectable
resistors. The device has three different resistance settings which, depending on the
battery voltage and the voltage limit set by programmable logic control (PLC), can
be switched on and off as required. Because the voltage drops during the dis-
charging process, the device has the advantage that the discharging current remains
high due to switching to low resistance values in contrast to variant A.
(C) Discharging in conductive liquids
Instead of a battery with a solid conductor for discharging, it is possible to use
conductive liquids. An obvious variant is to use salt water as the conductive liquid.
Figure 4.7 shows a schematic and a real experimental setup for cell discharge.
64 D. Hauck and M. Kurrat

Fig. 4.7 Experimental setup to discharge battery cells in conductive liquids (variant C)

To measure the discharge current within the liquid, a positive cell pole is
insulated with silicone and passed over an ammeter outside the vessel serving as a
substitute pole, and then reintroduced close to the original pole.
This method falsifies the test results to some extent, but is however necessary to
obtain experimental data. For the setup, a solution of 300 g of NaCl in three liters of
water was used. Calculations based on the concentration, specific conductivity and
specific resistance give a resistance of 3.914 Ω for the shortest distance between the
minus and the plus pole substitute. The theoretical discharge current of a fully
charged cell is therefore approx. 1 A.
Discharging with constant current
(D) Discharging using a semiconductor
To discharge battery cells and modules with constant current, a circuit with
metal-oxide-semiconductor field-effect transistor (MOSFET) was develop as shown
in Fig. 4.8.
The device is conceived for batteries up to 80 V. Using the potentiometer, a
maximum discharge current of 50 A at a maximum allowed power of 300 W can be
set. A total of five MOSFETs were mounted on one common cooling body which,
in turn, was cooled using an air blower.
4 Overdischarging Lithium-Ion Batteries 65

Fig. 4.8 Device with MOSFETs to discharge cells and modules with constant current (variant D)

Discharging with constant power

In order to discharge a battery in CP mode, commercially available electronically
adjustable loads were used as seen in Fig. 4.9.
(E) Discharging using energy transformation into heat
To discharge a battery at higher power, a TORKEL 860 from Megger GmbH
(Fig. 4.9 left) was used. It is limited to a maximum input voltage of 80 V with a
maximum current of 110 A or 15 kW. Three control modes of CC, CP and CR are
supported. Discharging occurs through resistors which are cooled down with eight
air blowers. The device can be controlled using a touch panel or the supported user
software.
(F) Discharging using energy recuperation into grid
To discharge a battery with energy recovery, the electronically adjustable load
EA-ELR-9500-30 supplied by EPS Stromversorgung GmbH (Fig. 4.9 right) was

Fig. 4.9 Electronically adjustable loads to discharge modules and systems in CR, CC and CP
modes [left: heat transformation (variant E); right: energy recuperation into grid (variant F)]
66 D. Hauck and M. Kurrat

used. The maximum input voltage was 500 V and the maximum discharge current
was limited to 30 A or 3.5 kW. The four control modes of CC, CV, CP and CR are
supported. For power recovery, the DC energy introduced is converted into a power
synchronized sinusoidal current. The power can be introduced into local and public
power supplies. In this way, there is little heat development compared to other
discharging devices. The device can be regulated using a touch panel or propriety
software. The efficiency at maximum input voltage or power is 94.5%.

4.3.3 Experimental Setup and Research Questions

The investigations are related to the functionality and applicability of the dis-
charging device. Each discharging device has its technical specifications with
respect to voltage, current and power. The choice of which device is to be used for
battery discharge depends on these specifications. As in the previous section, the
devices and setups for overdischarging offer various features in terms of control
modes. As battery behavior when overdischarging is the focus of this research, no
detailed description of the experiments to investigate the behavior of the devices
will be given at this point. Instead, the main parameters relevant to overdischarging
and the experiments conducted to determine these parameters will be described.
The experiments will be divided according to battery size (cells, modules and
systems) and then described accordingly.
Discharging devices used and experimental setup
To investigate the discharging behavior of battery cells, a MOSFET device (variant
D) was used. It offers the advantage that the discharge occurs at constant current. To
discharge the module, the electronically adjustable load EA-ELR-9500-30 (variant
F) was used. To discharge the system, besides variant F, the TORKEL 860 battery
load tester (variant E) was also used as, with this device, a significantly higher
power can be achieved. The basic setup for all variants is shown in the following
circuit diagram in Fig. 4.10.

Fig. 4.10 Circuit diagram of basic setup for all discharge variants
4 Overdischarging Lithium-Ion Batteries 67

To measure the current, the battery was connected to the load resistor (discharge
device) via a shunt resistor. An additional high voltage switch in series enabled the
safe switching of the battery to the discharge circuit. Parallel to the battery and to
the shunt, a short circuit was set up to enable the battery to be short circuited with
an additional high current switch.
Research issues
With the aid of an experimental setup, various questions are to be investigated
such as:
• How does the ambient/cell temperature influence discharge behavior?
• How much energy can be extracted in the overdischarging range and how long
does this take?
• How high are the short circuit currents depending on the electrical connection?
• How much does the voltage relaxation vary depending on the short circuit
duration?
• How much heat is developed while overdischarging?
• What effect does pole reversal have on the voltage relaxation?

4.4 Results and Discussion

Investigations related to the functional behavior and applicability of the discharging

devices (Sect. 4.4.1) as well as the behavior and characteristics of batteries when
they are overdischarged (Sect. 4.4.2) are reported in the next sections.

4.4.1 Investigations Related to the Functional Behavior

and Applicability of the Discharger

Discharging with constant resistance

The discharging device variants A–C all function following the control mode at
constant resistance. In Figs. 4.11 and 4.12 (representing variants A and B), the
current/voltage characteristic curves for the discharge of two 3.1 Ah cell are
illustrated.
The initial voltage of both the cells amounts to 3.67 V and a 50% state-of-charge
(SOC). The discharging process in variant A is conducted at a constant resistance
value of 0.5 Ω. The discharge current falls proportionally to the discharge voltage.
A voltage drop to 1 V is only observed after approx. 30 min.
With variant B, varying resistances for the discharge are used. After about
10 min, the voltage drops dramatically and the discharging resistance is switched to
a lower value, so that the discharge current remains relatively high. Due to the low
68 D. Hauck and M. Kurrat

Fig. 4.11 Overdischarge current/voltage characteristic curve for a 3.1 Ah cell with variant A

Fig. 4.12 Overdischarge current/voltage characteristic curve for a 3.1 Ah cell with variant B

capacity in the low voltage range, a rapid drop to 0 V occurs, resulting in the
initiation of a short circuit of the cell. A direct comparison of both variants A and B
makes it clear that switching different resistances with a PLC leads to a significantly
faster discharge and is therefore more efficient.
Apart from these variants with a solid load resistor, there is variant C which uses
a liquid load resistor comprising a conductive salt/water solution. The discharge
curve for the current and voltage are shown in the following figure (Fig 4.13).
The data points were taken every second and with an accuracy of 10 mV/
10 mA. At the start of the measurements, the current drifted to a value of approx.
0.9 A which fell to approx. 0.6 A in the following 3 min. During the experiment
(until the 22nd minute), the current strength began to fluctuate and dropped
4 Overdischarging Lithium-Ion Batteries 69

Fig. 4.13 Overdischarge current/voltage characteristic curve for a 3.1 Ah cell with variant C

continuously until it reached 0.1 A. Finally, using a metal bar, the cell was lifted out
of the liquid within the vessel a few times and immersed again (22nd to 25th
minute). This was done to remove potential resistance layers on the electrode and
thereby enable an increase in the current flow. Since the current did increase,
although only for a short period, this shows that both the layer and the gas buildup
on the electrode influence the discharge process negatively. The experiment was
therefore interrupted (25th to 28th minute) and the cell was removed from the
liquid. A strong dissolution of the original silver-colored coating of the substitute
pole (banana plug—the material is not known) was visible and a layer of black
coating was built up as shown in Fig. 4.14 (middle left).
In order to achieve better conductivity from the cathode side, the remnants of the
plug were wrapped in nickel foil. Subsequently, in a second trial, the battery cell
was reimmersed in the solution (37th minute). The current then reached a value of
0.8 A for 5 min (until the 42nd minute), before it started dropping continuously in
the following 6 min (until the 48th minute), reaching a value of approx. 0.6 A. In
the following 6 min (until the 54th minute), the current began to fluctuate and
reached its minimum at 0.1 A. The cell was then removed from the solution and the
data collection was stopped at the 68th minute. After the removal of the cell, it was
seen that the nickel foil was completely dissolved and the black covering layer on
the anode was even more distinct. Liquid samples were taken after each immersion
trial (Fig. 4.14 middle right) and sent to the battery research center MEET for

Fig. 4.14 Experimental setup (left), cell poles (middle left), liquid sample (middle right), brown
liquid (right)
70 D. Hauck and M. Kurrat

Table 4.3 Concentration Concentration (ppm)

analysis of the different
Unshaken Shaken
samples
Ni Cu Ni Cu
Trial 1 370 4.498 1.252 15.838
Trial 2 436 217 18.183 7.072
Trial 3 212 99 4.233 1.577

analysis. To find out whether the battery would continue to discharge, a third trial
was conducted. The silicon insulation on the cathode was removed and the cell, tied
to a string, was immersed as shown in Fig. 4.14 (right). The open vessel was
allowed to stand and monitored for about 30 min. In Fig. 4.14 (right), it can clearly
be seen that the liquid turned brown and some material was found floating on the
surface. Apart from this, gas formation was observed. The loss of contact of the
electrode to the liquid may be the cause of the fluctuation of the current observed.
The cell was again removed from the vessel and a cell voltage of 3.78 V was
measured. Compared to the off-load voltage of 4.01 V after the second trial, a
significant drop in voltage of 0.23 mV was evident. This shows that in fact a further
and in this case a faster discharge has taken place. After another discharge (fourth
trial) and the subsequent removal of the cell, it was seen that the anode pole had
completely disconnected from the cell and the experiment was ended.
Table 4.3 shows an extract of the results of the MEET X-ray fluorescence
analysis of the samples from the first three trials of samples taken without and after
agitation.
Generally, a significant difference is observed between samples taken before and
after agitation. This can only be explained as being caused by precipitation or
floating materials (from Ni and Cu salts and hydroxides). In the agitated samples, a
high Ni content was found in trial 2, which can be explained as having been caused
by using Ni foil on the banana plug. The material used on the banana plug was not
investigated. With Cu, a continuous drop in concentration was evident. At the start
(trial 1), a large amount of Cu was probably dissolved as Cu salts in solution
(Eq. 4.1), which then led to the formation of Cu(I) oxide and Cu(II) oxide layers on
the electrode (Eq. 4.2a, b). This explains the dark coating on the electrode.

Cu þ 2H2 O þ 2NaCl ! CuCl2 þ 2NaOH þ H2 " ð4:1Þ

Cu þ H2 O ! CuO þ H2 " ð4:2aÞ

2Cu þ H2 O ! Cu2 O þ H2 " ð4:2bÞ

The visible dissolution of the banana plug (Ni electrode) led to an increase of
nickel ions in the solution. The cause of the dissolution is the formation of
hydrochloric acid (see Eqs. 4.3a, 4.3b to 4.4):
4 Overdischarging Lithium-Ion Batteries 71

2NaCl ! Cl 2 þ 2Na þ þ 2e ð4:3aÞ

Cl2 þ H2 O ! HCl þ HOCl ð4:3bÞ

Ni þ 2HCl ! NiCl 2 þ H2 " ð4:4Þ

In the third and fourth trials, the insulation of the cathode was removed and
therefore the original cell potential made of aluminum was exposed. The reaction of
the cathode side can be represented in Eqs. 4.5 and 4.6.

Al ! Al3 þ þ 3e ð4:5Þ

3CuCl2 þ 2Al ! 3Cu þ 2AlCl3 ð4:6Þ

In analogy to Eq. 4.4, the dissolution of the aluminum cathode can be described
by Eq. 4.6. The experiment has shown that electrolysis at cell level strongly
influences the discharge characteristic curve and may even lead to the dissolution of
electrodes. Therefore, an opening of the cell and hence the formation of hydro-
fluoric acid (HF) cannot be excluded. The necessary extent of measures to neu-
tralize HF must be estimated. In addition, in order to guarantee safety, the amount
of hydrogen (H2) has to be analyzed to exclude the potential risk of oxyhydrogen
formation. At the module or system level, these effects are less likely because, on
the one hand, higher voltages are used and, on the other hand, there are larger
electrode surfaces. However, to allow a more stable discharging process at cell
level, it is necessary to test the use of other liquids and additives. One way of
minimizing the electrolysis process could be the application of a more resistant
coating on the electrode surfaces. Also, the use of media other than liquids, e.g.
conductive sands or polymers, should be considered. These substances would make
the processes easier to handle and would rule out danger from HF and
oxyhydrogen.
Discharging with constant current
In order to discharge a battery with constant current, the specially developed
MOSFET device (variant D) was used. A typical characteristic curve of the current/
voltage for the 3.1 Ah cell with this discharge mode is shown in Fig. 4.15.
The initial voltage of the cell is 3.67 V or 50% SOC. Due to the preset, there is a
higher current at the start when the MOSFET device is switched on. The current is
then set to a constant 6 A. After about 12 min, the lower voltage limit is reached
and after about 17 min, the cell is short circuited because it is close to 0 V.
If a higher discharge current were set, discharging would take place at a much
higher speed. Another advantage over variant A can be seen in the overdischarging
area, where the discharge current remains constant and leads to a fast decline in
voltage.
72 D. Hauck and M. Kurrat

Fig. 4.15 Overdischarge current/voltage characteristic curve for a 3.1 Ah cell with variant D

Discharging with constant power

In order to discharge a battery in CP mode, the TORKEL 860 device (variant E) and
EA-ELR-9500-30 (variant F) are used. Figure 4.16 presents characteristic curves
showing the current, voltage and power of a VW e-Golf battery system with
TORKEL 860 at constant power.
The battery system is discharged at a constant power of 15 kW. After about
62 min, the power diminishes due to loss of voltage, and the maximum possible
discharging current through the load is reached (55 A). After 66 min, the current
also decreases because the battery system cannot deliver higher current due to the

Fig. 4.16 Overdischarge current/voltage characteristic curve for a VW e-Golf battery system with
variant E
4 Overdischarging Lithium-Ion Batteries 73

lack of missing energy. From 90th minute onward, the system is short circuited and
the measurement stopped.
Assessment of discharging variants
The presented ways of controlling or discharging variants have different advantages
and disadvantages. Using the criteria from Sect. 4.2.4, the variants can be assessed.
In the following Table 4.4, four crucial criteria were used for assessment.
Those discharge devices using constant resistance (variants A, B and C) are low
cost and are considered safe due to the limited maximum current and the lower
power. Due to the proportionality of current and voltage, the discharge time is high
and the discharge curve inflexible. Discharging a cell in liquid is assessed as being
unfeasible, as concluded from the results of the test. However, at the system level,
project partners could achieve positive results.
The discharge via the developed MOSFET device (variant D) does not show any
significant disadvantages. It is slightly more expensive than the passive resistance
discharge devices, but due to it being controlled with constant current, it is much
more efficient with respect to the speed of discharge. Manually controlling the
power is possible by using the potentiometer and so this variant can be assessed as
being semiflexible. Safety is also rated as being mediocre as the discharge current
can be regulated, however there is still the danger of manually over loading with too
high current. By using a temperature controlled current input, it would be possible
to eliminate those deficits.
The discharge devices commercially available (variants E and F) are more
expensive than discharge devices which were developed in-house. However, they
have the advantage of allowing different ways of controlling (CR, CC and CP), and
they allow random discharge curves to be programmed. By monitoring the voltage,

Table 4.4 Assessment of the discharge devices with four crucial criteria
Discharge Discharge device Cost of Discharge Flexibility Safety
mode discharge time discharge during
device profile discharge
CR A Single cell Low High Inflexible Safe
resistor
B Selectable Middle Middle Middle Middle
resistors
C Conductive Low Middle Inflexible Middle
liquid
CC D MOSFET Middle Short Middle Middle
CR, CC, CP E Energy High Short Very Very safe
conversion flexible
into heat
F Energy High Short Very Very safe
recovery in flexible
grid
74 D. Hauck and M. Kurrat

current and temperature, the discharging, with additional switch-off criteria can be
considered as being relatively safe.

4.4.2 Investigation of Battery Behavior in Cases

of Overdischarging, Pole Reversal, Short Circuit
and Voltage Relaxation

Influence of ambient/cell temperature on over discharge and relaxation

behavior
Temperature is known to have a strong impact on the inner resistance of a battery
and thus its discharge behavior. In order to assess the influence of temperature
during overdischarging or the relaxation of voltage after a short circuit, three 50
Ah NCM cells were discharged with 30 A at −20, 0 and 40 °C; they were then short
circuited and relaxed as shown in Fig. 4.17.
It can clearly be seen that at the beginning of the discharge process, the cell voltage
with the lowest temperature (−20 °C) drops the most. This is due to the increased
viscosity of the electrolyte along with its worsened ionic conductivity (Jossen and
Weydanz 2006). The inner cell resistance is about 26.5 mΩ after one second,
compared to 11.2 mΩ (0 °C) and 1.3 mΩ (40 °C). At the moment of the short
circuit (at 0.5 V), this cell was able to discharge only about 10.5 Ah, instead of 11.5
Ah (0 °C) and 12 Ah (40 °C). After stopping the short circuit after 1 h, it can
clearly be seen that cell voltages at the lowest temperature relax the fastest and
highest. After four hours of relaxation, the cell at −20 °C even increased above a

Fig. 4.17 Overdischarge and

relaxation curves for 50
Ah NCM cells at different
temperatures
4 Overdischarging Lithium-Ion Batteries 75

2.5 V cell voltage, which means that it came back to “normal” voltage range. That
might be dangerous for the user because a defect-free battery is presumed, although
it might clearly already be overdischarged. The temperature-dependent relaxation
effect is again traced back to increased inner resistance. In contrast, the cell at the
lowest temperature still has the highest capacity due to the fact that less energy was
taken out in the former discharging step. This conclusion is generally valid and can
be transferred to other cell types or modules and systems.
Influence of short circuit duration, cell capacity and cell connection on
relaxation behavior
Apart from temperature, especially the preceding short circuit duration exerts a
direct effect on relaxation behavior. In the next trial, three 3.1 Ah cells and three 25
Ah cells were each discharged, short circuited and relaxed. The short circuit was
undertaken at 0.1 V for all cells and removed after different times (10 min, 1 h and
24 h). The relaxed voltage was then measured. The relaxation curves of the six cells
are plotted in Fig. 4.18.
The figure above demonstrates the dependence of relaxation behavior on the
length of the short circuit. After a short circuit lasting 10 min, the voltage of the
3.1 Ah cell as well as of the 25 Ah cell increased up to above 2.3 V. If the short
circuit time is longer, lower voltages will accordingly be achieved: 1–1.5 V at 1 h
and 0.1–0.2 V at 24 h. In order to ensure safe relaxation, cells should be short
circuited for at least 24 h before opening a short circuit again. Only in this way it
can be guaranteed that the cells will not become an ignition source during the
subsequent step of shredding.
It can also be seen that the 3.1 Ah cells exhibit a steeper increase in voltage than
at the beginning of the relaxation compared to the 25 Ah cells. Similar to tem-
perature, this can be traced back to the higher inner cell resistance of the 3.1 Ah
cells.

Fig. 4.18 Relaxation curves

for three 25 Ah and three 3.1
Ah cells for different short
circuit times
76 D. Hauck and M. Kurrat

Apart from studying cell capacity, the goal was also to test the connection of the
cells with respect to relaxation behavior. The systems are usually made up of
several modules which themselves are connected of several cells. The modules can
be handled by humans with respect to their weight. They could therefore be con-
sidered as the smallest exchangeable unit in future generations of automotive
vehicles, because of the large effort required to dismantle a single cell.
In order to test the influence of the connection on the relaxation behavior, several
modules were overdischarged and their relaxation behavior was measured. Three
modules from a VW e-Golf with a 3p4s connection were tested again with different
short circuit times (10 min, 1 h and 24 h). The modules were discharged with the
MOSFET device with a current of 15 A starting with a voltage of 15 V. At a
module voltage of 0.5 V, the modules were short circuited. The relaxation curves
are shown in Fig. 4.19.
As expected and similar to the cell measurements, the relaxation voltage of the
modules grows increasingly as the short circuit time is reduced. What is notable is
the high relaxation to above 8 V (tendency still increasing) after four hours and
after a short circuit of 10 min. As with the cells, only the 24 h short circuit time
leads to a minimal relaxation voltage of the module.
Looking at the short circuit currents just before each short circuit, important
differences can also be found. After 10 min of relaxation, 2.22 A were measured
and at 60 min of short circuit time, there were 1.69 A. After 24 h, only 0.02 A of
short circuit current was measured.
High relaxation voltages might be dangerous to humans. To test the remaining
energy and the remaining current, one module in 2p4s configuration with a
relaxation voltage of 2.8 V was again short circuited. The current and voltage
curves can be seen in Fig. 4.20.
At the time of the short circuit, a current of 103 A was measured. After about
1.5 min of short circuit duration, a current of 3.1 A was still flowing. The relatively
high flows show that modules should also be short circuited for at least 24 h.

Fig. 4.19 Relaxation curves

for three 3p4s modules for
different short circuit times
4 Overdischarging Lithium-Ion Batteries 77

Fig. 4.20 Short circuit current for a 2p4s module at 2.8 V relaxation voltage

The discharge of whole systems makes sense from a technical as well as from an
economic view. However, the system discharge carries the highest risk because
single cells without an active BMU cannot be controlled. The devices TORKEL
860 (variant E) and EA-ELR-9500-30 (variant F) were used to discharge the sys-
tems. The discharge curves with both variants are illustrated for two VW e-Golf
systems in Fig. 4.21.
Figure 4.21 shows that the discharge with a higher power of 15 kW (variant E)
occurs faster than with 3.5 kW (variant F). It is worth mentioning that in the case of
variant F, about 94% of the energy could be fed back into the power grid.
With variant E, a short circuit can already start after 1.5 h, whereas with variant
F, this is only the case after 3.5 h. In order to assess safety, a thermal imaging
camera was also used to measure temperature, as seen in Fig. 4.22.
The photo shows the back part of the battery system because this becomes
warmer than the front. The surrounding temperature is around 9 °C and the
warmest spots reach about 30 °C. The warming-up is clearly visible but is not
dangerous. In order to assess the relaxation behavior in the case of one system, the
short circuit was removed after about 10 min. After only a few seconds, the voltage
jumped to about 30 V, which would be a danger if the system were touched. After
altogether 25 min of relaxation, the voltage reached a value of 130 V which means
an even higher danger if touched. This clearly demonstrates that systems have to be
short circuited over a longer period of time (ideally several days) in order to avoid a
possibly dangerously high relaxation voltage above 60 V.
Influence of pole reversal on relaxation behavior
Pole reversal may cause extreme situations in overdischarging. To obtain a pole
reversal of individual cells, it is not sufficient to use an adjustable resistor for the
load. It is necessary to have an additional current source which displays a constant
discharge current independent of the battery voltage. The following Fig. 4.23
78 D. Hauck and M. Kurrat

Fig. 4.21 Overdischarge current/voltage characteristic curves for VW e-Golf battery systems with
15 kW (variant E) and 3.5 kW (variant F)

Fig. 4.22 Heat distribution

for a VW e-Golf battery
system during
overdischarging at 9 °C
ambient temperature
4 Overdischarging Lithium-Ion Batteries 79

Fig. 4.23 Temperature rise and pole reversal for a 25 Ah NCM cell

shows the discharge curves of cell voltage and temperature at a constant 30 A

discharge rate until a pole reversal occurs.
The figure shows that the cell reaches 0 V after 45 min. Subsequently, the cell is
discharged for a further 15 min and then the poles are reversed. The temperature
profiles show that the cell pole warms up significantly (50 °C) at reversal, so that a
longer pole reversal may be hazardous. After 60 min, the measurement was stop-
ped. It was noted that the voltage relaxes for a short moment and then remains at
0 V. In the case of a possible electrical hazard through a high relaxation voltage,
especially at system level, pole reversal may be of assistance.
During discharging modules or systems, a pole reversal of individual cells may
be inevitable. This is one reason why pole reversal needs further investigation.
A typical example is the discharge of cells connected in series, where a cell may
already exhibit lower capacity or a small state-of-health (SOH) value. The fol-
lowing figure demonstrates the discharge of a module (2p4s), where one cell
contains low power (Fig 4.24).
The figure clearly shows that the discharge profile of cell 7/8 drops sharply after
20 min and then reaches the overdischarging range, whereas the other cells remain
in the “normal” range. Since only four cells are connected in series, the voltage drop
observed is also visible in the module voltage profile which shows a profile similar
to the cell voltage profile. The temperature of cell 7/8 also increases significantly,
similar to that seen in the cell measurements, especially in the overdischarging
range. Due to the fact that every two cells are connected in parallel, only half the
current flows, namely 15 A, through each cell. Due to this, at the point of pole
reversal, a heat increase of only 15 K is observed in comparison to 25 K observed
in single cell measurements. In the course of pole reversal, the cell reaches −2 V
before the voltage returns in the direction of 0 V. The increase in cell voltage is
clearly visible in the module voltage profile. After 80 min, the other cells’ voltages
reach 0 V, at which point the discharge process can be terminated.
In this experimental case, pole reversal in combination with progressive dis-
charging is not critical and can, with sufficient expertise, be read from the module
voltage profile. The measurements show that both cell monitoring and temperature
monitoring of individual cells during overdischarging may be useful in preventing
80 D. Hauck and M. Kurrat

Fig. 4.24 Cell pole reversal and heat rise while overdischarging a 2p4s battery module

critical situations. The behavior of the battery may differ and therefore pose other
hazards for higher discharge currents, other circuits or other cell types.

4.5 Summary and Outlook

When recycling was undertaken in the past, overdischarging did not have to be
carried out due to the pyrometallurgical recycling processes. Overdischarging was
not relevant for furnaces, as the residual voltage or energy remaining in a battery at
combustion were unimportant. For this reason, little research has been done about
overdischarging lithium-ion batteries, as they are more relevant to
non-pyrometallurgical recycling. Overdischarging LIBs is a necessary step in the
LithoRec recycling process. Factors such as safety, ease of handling and efficiency
are the key parameters that need to be considered in the realization of overdis-
charging. Inexpensive variants (A–D) which were developed by the project par-
ticipants as well as commercially available devices (E–F) were tested and validated.
The exploratory work in this project essentially led to the development of a
knowledge base about the behavior of cells, modules and systems when overdis-
charging in various situations. The influence of this behavior on cell properties as
well as the residual hazard potential were also studied. Important parameters in
these investigations on overdischarging behavior are the discharge times/capacities,
the short circuits that arise, the amount of the relaxation voltages, the influence of
the surrounding temperatures and the types of switching as well as behavior during
pole reversal.
4 Overdischarging Lithium-Ion Batteries 81

All in all, it can be claimed that overdischarging poses no severe hazard, as long
as the discharge current and thus the battery temperatures allowed are not exceeded.
All discharge variants have their advantages and disadvantages and using them all
can therefore be justified.
The relevant questions for the operators of recycling plants are the costs, effi-
ciency and the degree of automation of the process steps. The recommendation here
is discharging a battery system using an electric load with energy recovery (variant
F) and high power, so that the remaining electrochemically stored energy in the
battery can also be “recycled”. The discharge in liquids at cell level has been proved
to be unviable, but as a system, this process showed promising results. The latter
variant, moreover, offers an interesting solution, especially under the aspect of
deactivating vehicles after accidents when the battery is damaged and contacting is
complicated.
Furthermore, it could be demonstrated that with pole reversal, little or no voltage
relaxation takes place. Especially from the aspect of HV safety due to electrical
hazards as well as the low process speed of recycling, the relaxation behavior of a
battery may be of importance. The influence of various discharge profiles or a pole
reversal on the recovery efficiency of individual raw material could not be inves-
tigated within this project. It is assumed that there might be an “ideal discharge
profile” for the optimum recovery of material.
To make discharging batteries more efficient in the future, the following is
recommended. Vehicle manufacturers could implement a release order entitled
“ready for recycling” in their BMUs, so that a monitoring system integrated into a
BMU and a connection to the system’s internal protection mechanism are possible.
Opening a battery system lid and installing HV circuits inside the opened system
may lead to the serious risk of an electric shock and should be avoided. Ideally,
there should be a standardized plug for HV contacts and BMUs as well as a
standardized software application for discharging, so that this process step can be
conducted safely and automatically. Due to the multiple types of battery systems
that are currently on the market, it is unfortunately not expected that this will be
realized in the near future.

References

Bauer S (2017) AkkuWelt‚ Erste Auflage. Würzburg: Vogel Business Media GmbH
Blank T (2012) Deep discharge behavior of lead-acid batteries and modeling of stationary battery
energy storage systems. 978-1-4673-1000-0/12/$31.00 ©2012 IEEE
Gellerich W (2015) Akkumulatoren - Grundlagen und Praxis, Zweite edn. Shaker Media, Aachen
Jossen A, Phan TB, Svoboda V (2016) Tiefentladung von Batterien - Ursachen, Mechanismen,
Lebensdauereinfluss. https://www.mikrocontroller.net/attachment/34643/Tiefentladung.pdf
Jossen A, Weydanz W (2006) Moderne Akkumulatoren richtig einsetzen. Reichard Verlag,
Untermeitingen
Mähliß J (2012) Aufbau, Funktionsweise und Gefährdungspotenzial von Li-Ionen-Zellen.
Elektronik - ecodesign. 2012