Professional Documents
Culture Documents
Environmental
Science
View Article Online
PERSPECTIVE View Journal | View Issue
ion batteries
Cite this: Energy Environ. Sci.,
2016, 9, 2219
K. R. Cromptonabc and B. J. Landi*bd
There are inherent safety risks associated with inactive lithium ion batteries leading to greater restrictions and
regulations on shipping and inactive storage. Maintaining all cells of a lithium ion battery at near zero voltage
with an applied fixed load is one promising approach which can lessen (and potentially eliminate) the risk of a
lithium ion battery entering thermal runaway when in an inactive state. However, in a conventional lithium ion
cell, a near zero cell voltage can be damaging if the anode electrochemical potential increases to greater than
the potential where dissolution of the copper current collector occurs (i.e. B3.1 V vs. Li/Li+ at room
temperature). Past approaches to yield lithium ion cells that are tolerant to a near zero volt state of charge will
be reviewed and their anticipated tradeoffs discussed. Additionally, the present work introduces a novel method
to prevent the anode potential from increasing to greater than B3.1 V vs. Li/Li+ during near zero volt storage
by modifying the amount of reversible lithium in a conventional cell using a pre-lithiated anode. Specifically,
LiCoO2/MCMB pouch cells fabricated with a pre-lithiated anode maintained 99% of their original capacity after
Received 20th March 2016, three, 3 day and three, 7 day storage periods at near zero volts under fixed load. A LiCoO2/MCMB pouch cell
Accepted 26th May 2016 fabricated with a pre-lithiated anode also maintained its original discharge performance after three, 3 day
DOI: 10.1039/c6ee00836d storage periods under fixed load at 45 1C. Thus, modifying the amount of reversible lithium in a cell with anode
pre-lithiation is a promising concept to achieve lithium ion cells with high tolerance to near zero volt storage
www.rsc.org/ees without modification to typical cell construction parameters or use of unconventional materials.
Broader context
Recently in February, the U. N. International Civil Aviation Organization put forth a ban on transporting lithium ion batteries as cargo on passenger planes to
take effect in 2016. Such a ban is the result of several aircraft fires, some of which have resulted in the loss of human life, that have been blamed on the lithium
ion batteries that were being shipped as cargo on the planes. The ban is expected to stay in place until at least 2018, when new packaging rules for shipping
lithium ion batteries are designated to be put forth. While an important step in ensuring the safety of aircraft passengers and personnel, this ban and increased
future packaging rules will likely reduce the distribution network of lithium ion batteries and increase costs to manufacturers. In this perspective, the approach
of completely discharging a lithium ion battery to near zero volts (an effectively inert state) is presented and discussed as an alternative pathway to render lithium
ion batteries safe for transport. Past and current approaches to fabricate lithium ion batteries that are tolerant to near zero volt storage, which require modifications
to cell construction parameters and/or use of unconventional materials, are reviewed and discussed. A novel strategy to achieve near zero volt tolerant batteries is
also presented and demonstrated. The novel strategy has the advantages of not using unconventional materials and not requiring modifications to typical cell
construction parameters. These advantages can directly impact rapid adoption of near zero volt tolerant lithium ion cells that can make lithium ion batteries much
safer to transport. In addition, the benefits of this technology could extend to situations like electric vehicle accidents and home energy storage emergencies.
1 Lithium ion battery transportation becoming more paramount. For many portable applications,
electrochemical energy storage using lithium ion batteries is
risks and needs currently the premier method due to the enhanced rechargeable
As society transitions to renewable energy sources and expanded use chemistry. Compared to other designs (i.e. NiCd, Ni–H, etc.) the
of electrical energy, energy storage in a reliable and safe manner is lithium ion chemistry has higher energy density, cycle life and
highly tunable performance characteristics.1,2 There are many
efforts underway to enhance lithium ion battery performance to
a
Microsystems Engineering Department, Rochester Institute of Technology, USA
b
align with future application needs,3 however, as energy density
NanoPower Research Labs, Rochester Institute of Technology, USA.
E-mail: bjlsps@rit.edu
increases, the associated safety risks also increase.4
c
SMART Scholarship Program, USA When in a user-active state, manufacturer defects or abuse
d
Department of Chemical Engineering, Rochester Institute of Technology, USA of lithium ion cells within a battery can lead to a thermal
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2219
View Article Online
runaway event.5–21 Thermal runaway results from several internal cells only to 0–25% of their rated charge capacity (0% represents
exothermic reactions that are initiated by overheating of a cell all rated capacity removed from the cell, but some charge energy
that can result from internal short, rapid charge/discharge, does remain) has been shown to prevent thermal runaway under
overcharge, external heating, or other abuse condition. The physical abuse conditions (i.e. nail penetration)12 and lead to
exothermic reactions include SEI decomposition, electrolyte higher onset temperatures of thermal runaway/exothermic
reaction with the electrodes, decomposition of active materials reactions upon overheating.13,14 Thus, discharging cells to a
and electrolyte decomposition.5–18 Thermal runaway can result low state of charge before storage and shipping could greatly
in a dangerous fire or explosion that can propagate to other reduce the associated safety risks in an inactive state.
nearby cells.19,20 In the case of a large battery consisting of many However, it is well known that accurately measuring the
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
lithium ion cells, or many batteries stored together, propagation state of charge of a lithium ion battery requires sophisticated
can lead to a very dangerous event with severe damage including equipment and substantial amounts of time. Measuring the
explosion, fire, and venting of toxic gases.21 Several research open circuit voltage with a voltmeter is insufficient, because for
efforts to mitigate safety risks while lithium ion batteries are in most conventional lithium ion cells, the open circuit voltage
a user-active state have been reported.21–44 Internal to lithium will change little for B10–90% state of charge. Additionally, the
ion cells, shut-down separators to prevent ion flow upon over- open circuit voltage can vary for a given charge state depending
heating,22,23 cathode coatings to suppress exothermic release,24–28 on the cell history, materials used in the cell, and ambient
non-flammable electrolytes to avoid electrolyte combustion22,28–34,35 conditions. Difficulty in accurately and rapidly determining the
and redox shuttle additives to prevent overcharge22,37–44 have been state of charge presents a substantial challenge to effectively
investigated. External to cells during normal operation, battery implementing state-of-charge limitations on lithium ion batteries
management systems (BMS) to avoid abuse of cells, blocking diodes that are to be transported.
to prevent inadvertent charge or discharge, current limiting fuses to In addition to state of charge determination, discharging
prevent rapid charge or discharge, and bypass diodes to prevent lithium ion cells to low states of charge increases the risk of
overcharge/overdischarge of a ‘‘weak’’ cell in a battery pack have all them entering an overdischarged state during transport due to
been investigated or are currently used.21,22 In the case of entering self-discharge45–57 or cell mismanagement. Such an overdischarged
dangerous operating conditions, positive temperature coefficient state risks damaging internal components of the cells (most
devices (PTCs) to block or reduce current upon overheating and significantly dissolution of the copper current collector of the
current interrupt devices (CIDs) to block current in the case of anode58–64) which can cause significant degradation in the cell’s
over-pressure have been used or are currently in use.21,22 performance or render it unusable.
In the case of a user-inactive state, it would be ideal for the However, if lithium ion cells could be modified in such a
individual cells of a lithium ion battery to store no charge way that complete discharge to a near zero volt state of charge
energy and be at a near zero volt state of charge since the did not damage internal components, both the motivation of
battery does not need stored energy to perform a task. No rendering lithium ion batteries safe to transport and effectively
stored charge energy would minimize the amount of energy implementing state-of-charge limitations on cells to be trans-
that could be released in a catastrophic event, which reduces ported could be realized. In a representative scenario, after a
safety risks associated with the inactive battery. Even discharging cell is discharged to a low state of charge, an appropriately
2220 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
tolerant to a near zero volt state of charge. For one, a cell would be
by extension more tolerant to over-discharge during operation that
can occur due to mismanagement of a battery system. Also, as a
result of the higher overdischarge tolerance, individual cells or
batteries could be stored for long periods without regular checking
of the open circuit voltage or requiring trickle charging. Such a
feature could be useful for batteries that are difficult to access such
as those in implanted medical devices or those used in satellites.
Batteries transported in higher risk situations, such as
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2221
View Article Online
(e.g. carbon black, graphite, etc.) prior to being deposited onto The cathode slurry was coated onto a 20 mm aluminum foil
metal foil current collectors. The electrodes are electrically using a doctor blade. The composite was dried overnight in a
isolated from each other in the battery by a polymer separator, vacuum oven at 90 1C and then calendared to a composite
most often consisting of a microporous polypropylene/polyethylene density of B2.2 g cm3.
laminate that allows for lithium ion diffusion and migration.2 A The active materials used in the cathode and anode have a
non-aqueous, ionically conductive and electrically insulating defined specific capacity, which is the amount of lithium they
electrolyte, most commonly consisting of a lithium salt dissolved can reversibly intercalate/de-intercalate per gram of the material
in organic solvent(s), is wetted into the separator/electrodes. over a particular potential range. The LiCoO2 cathode material
The lithium ion cell used in the present study is illustrated used in this study has a specific capacity of B140 mA h g1
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
in Fig. 2a and b. The anode was prepared by mixing MCMB’s when cycled between the potentials 3.0–4.2 V vs. Li/Li+. The
into a slurry with SuperCt Carbon Black, SFG-6 and PVDF graphite anode material used in this study, in the form of
(Solvay) in a 93 : 0.3 : 2.7 : 4 w/w ratio using N-methyl-1-pyrrolidone MesoCarbon Microbeads and SFG-6, has a practical specific
(NMP) as the solvent. The PVDF was dissolved first in the NMP and capacity of 300–330 mA h g1 when cycled between 0.005–1.5 V
then the conductive additive and active material was added in vs. Li/Li+.
succession with intermediate mixing steps. All mixing was done In the present study, the cathode is loaded at B19.3 mg cm2
with a Thinky AR-100 planetary mixer. The slurry was coated onto a of LiCoO2 giving an areal capacity of B2.7 mA h cm2. The
20 mm thick copper foil using a doctor blade. The composite was anode is loaded at B10 mg cm2 of MCMB/SFG-6, giving an
then dried overnight in a vacuum oven at 90 1C and then areal capacity of B3.0 mA h cm2 (conservatively using
calendared to a composite density of B1.2 g cm3. The cathode 300 mA h g1 as the capacity of the active material). The area
was prepared by mixing LiCoO2 (MTI Corp) into a slurry with of the electrodes used is 4.5 cm2, giving a total cathode capacity
SuperCt Carbon Black, SFG-6 and PVDF (Solvay) in a 90 : 5 : 5 w/w of B12.2 mA h and an anode capacity of B13.5 mA h. The 10%
ratio. The mixing procedure was the same as described for excess of anode capacity compared to the cathode is common
the anode using N-methyl-1-pyrrolidone (NMP) as the solvent. practice to reduce the risk of lithium plating on the anode
during cycling.66
A lithium metal reference electrode is used in the present
study to measure the individual electrochemical potential of
the cathode and anode during cycling. The design used in the
present work, as shown in Fig. 2a and b places the reference
electrode on either side of the cathode/anode stack in the
pouch cell, which is a setup that has been used successfully
by others.62,63 As shown in Fig. 2b, the reference electrode used
in this study consists of a lithium metal foil pressed onto
copper mesh that is ultrasonically welded to a copper tab. A
lithium–Cu mesh is placed on both sides of the separator to
ensure good contact with the electrolyte soaked separator when
the cell is compressed between restraint plates during testing,
as depicted in Fig. 2b.
All pouch cells were assembled in an argon glove box
maintained at o1 ppm O2 and o1 ppm water. An electrolyte
of 1.2 M LiPF6 in 3 : 7 volume ratio of ethylene carbon (EC) and
ethyl methyl carbonate (EMC) was injected prior to sealing the
laminate. An Arbin BT2000 cycler with an auxiliary voltage
module was used for cycling the cells. The cell voltage and
the electrochemical potential of each electrode with respect to
the Li metal reference were measured independently through-
out cycling. As depicted in Fig. 2b, for testing the cell was
compressed between restraint plates. A rubber slab was placed
on one side of the cell in the restraint plates to ensure even
pressure and comply with expansion and contraction of the cell
upon cycling. All testing was done at room temperature unless
otherwise specified.
2222 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
step which referenced a cathode potential maximum of 4.2 V vs. experimental 3-electrode cells in the present study, the cathode
Li/Li+. The potential of the cathode vs. Li/Li+ was selected as is charged to the same degree independent of anode variation.
the control for the upper charge limit to ensure that for all The stability of LiCoO2 is known to be sensitive to the degree
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
Fig. 3 (a) Schematic of cell function during charge of conventional LiCoO2/MCMB pouch cell. The pink arrows represent lithium ion flow and the red
arrow represents electron flow. (b) 1st cycle charge cell voltage and electrode potentials of conventional LiCoO2/MCMB pouch cell. (c) Schematic of cell
function during discharge of conventional LiCoO2/MCMB pouch cell. The pink arrows represent lithium ion flow and the red arrow represents electron
flow. (d) 1st cycle discharge cell voltage and electrode potentials of conventional LiCoO2/MCMB pouch cell. (e) 5th cycle charge cell voltage and
electrode potentials of conventional LiCoO2/MCMB pouch cell. (f) 5th cycle discharge cell voltage and electrode potentials of conventional LiCoO2/
MCMB pouch cell.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2223
View Article Online
which it is charged to, particularly when overcharged.67 Therefore, amount of lithium ions will be de-intercalated from the cathode
utilizing the cathode potential vs. Li/Li+ as the upper charge limit as are intercalated into the anode. Then, upon discharge, the
helps eliminate the degree of cathode charge as a variable affecting same amount of lithium that flowed from the cathode to the
the cell’s cycling performance. Prototype cells and cells constructed anode on the charge will de-intercalate from the anode and
for anode pre-lithiation that do not employ a reference electrode intercalate into the cathode. This amount of lithium that can
utilized a cell voltage of 4.1 V as the upper charge limit, which is a flow from the anode to the cathode during discharge will herein
standard charge cutoff voltage for a LiCoO2/MCMB lithium ion cell. be defined as the amount of reversible lithium in the cell.
All cells were discharged to 3.0 V cell voltage, which is a standard In real lithium ion cell function, during the first charge of a
cutoff value for a LiCoO2/MCMB cell. cell side reactions with the electrolyte occur at the surface of the
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
During the charge of a cell, an externally applied voltage anode active material. The breakdown products of these reactions
difference causes lithium ions to migrate from the cathode to form a passivation layer on the anode active particles, called the
the anode. The migration of lithium ions results in lithium ions Solid Electrolyte Interface (SEI).68–73 When the SEI is fully formed,
de-intercalating from the cathode and intercalating into the it prevents any further side reactions between the electrolyte and
anode. In conjunction electrons flow from the cathode to the the anode active materials. During the formation of the SEI layer,
anode through an external circuit. The cell function during lithium ions irreversibly form into reaction products instead of
charge is depicted in Fig. 3a and the measured cell voltage and reversibly intercalating into the anode material. For graphite
electrode potentials during charge are shown in Fig. 3b. As anodes used in the present study, this loss amounts to B4–8%
shown in Fig. 3b, as the cell charges on the first cycle, the of the lithium ions inserted on the first cycle (as measured by
electrochemical potential of the cathode vs. Li/Li+ increases as half-cell data). As a result, the amount of lithium that can be
lithium ions de-intercalate from it. The electrochemical de-intercalated from the anode upon discharge is less than what
potential of the anode, on the other hand, decreases as lithium was initially intercalated, resulting in a decrease in the amount
ions are intercalated into it. As a result, the cell voltage, which of reversible lithium in the cell.
is the difference between the cathode and anode electrochemical The LiCoO2 cathode used in the present study, on the other
potentials, increases until charge is stopped at the cutoff voltage hand, has a first cycle capacity loss of B2–3% (as measured by half-
(when the cathode reaches 4.2 V vs. Li/Li+ in this case). cell data), meaning it can intercalate B97–98% of the lithium ions
When the cell is discharged by applying an external load that were de-intercalated from it on the first charge. Since the loss of
between the electrodes, lithium ions diffuse from the anode to reversible lithium during anode SEI formation is greater than the
the cathode due to the electrochemical potential difference loss of cathode capacity, it can be expected that the amount of
between the electrodes. In conjunction, electrons flow through reversible lithium in the cell will be less than that required to fully
the external load from the anode to the cathode. The cell function intercalate the cathode with lithium upon discharge.
during discharge is depicted in Fig. 3c and the measured cell As shown in Fig. 3d, the electrochemical potential of the
voltage and electrode potentials during charge are shown in anode increases rapidly to 0.785 V vs. Li/Li+ at the end-of-
Fig. 3d. As shown in Fig. 3d, as the cell discharges, the electro- discharge while the electrochemical potential of the cathode
chemical potential of the cathode vs. Li/Li+ decreases as lithium remains high in its normal range at 3.789 V vs. Li/Li+. This
ions intercalate into it. The electrochemical potential of the anode, increase in anode potential indicates that the anode is mostly
on the other hand, increases as lithium ions de-intercalate from it. depleted of reversible lithium while the cathode is not fully
As a result, the cell voltage decreases until discharge is stopped at intercalated with lithium at the end of discharge. This is
the cutoff voltage (3 V cell voltage in this case). consistent with the expectation that the cell will not have
The reference electrode tests used in the present study is enough reversible lithium to fully intercalate the cathode at
what allows for independent measurement of each electrode the end of discharge in this conventional cell.
potential as well as the cell voltage during cell cycling. In order The 5th charge and discharge profiles for the LiCoO2/MCMB
to ensure that the reference electrode measurements are consistent cell are shown in Fig. 3e and f. As observed, the electrode
with the observed cell voltage, the difference between the measured behavior during normal cycling is nearly the same as the first
electrode potentials was compared to the measured cell voltage cycle with only some minor differences. The anode potential is
throughout cycling. An example is shown in Fig. 3b and d and as slightly lower at the beginning of charge due to the SEI being
shown, the calculated cell voltage is nearly identical to the measured already formed. As a result, the cell voltage starts slightly
cell voltage throughout charge and discharge with a magnitude higher. Additionally, the cell cycles with nearly identical charge
of difference of o8 mV. The high degree of consistency observed and discharge capacities, indicating that the amount of rever-
between measured and calculated cell voltage confirms the pouch sible lithium is not being reduced by further SEI formation and
cell assembly with reference electrodes is valid. the cathode is not losing any additional capacity. This is in
contrast to the first cycle, which had a discharge capacity of
3.3 Amount of reversible lithium in a cell B0.5 mA h less than the charge capacity due to loss of
When a conventional LiCoO2/graphite lithium ion cell is reversible lithium during SEI formation. If anode materials
assembled, the cathode is inherently fully intercalated with with higher first cycle loss than graphite such as silicon74–81
lithium while the anode contains no lithium. In an idealized and germanium82–85 are used in alternative cells, the initial loss
view of lithium ion cell function, during charge, the same of reversible lithium is more dramatic.
2224 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
3.4 Electrode potential behavior during cell overdischarge function of time in Fig. 4b. As shown in Fig. 4a and b, during
towards near zero volts by an applied fixed load the C/10 discharge (first B12 mA h or B10 hours) the cell
For the cell described in Section 3.2 and 3.3, after the 5th C/10 current is constant and the electrode potentials exhibit their
discharge to 3.0 V cell voltage, a fixed load of 2.5 kOhm was normal discharge behavior. At B12 mA h, (B10 hours), the cell
applied to the cell by the Arbin cycler to simulate a resistor reaches 3.0 V and the fixed load (2.5 kOhm) step is initiated and
applied between the leads of the cell, similar to what could be the cell undergoes a transient period as it discharges to near
achieved in a practical storage or transportation situation. The zero volts. As shown in Fig. 4b, the cell voltage decreases to
value of the applied load was designed so that at 3.0 V cell o0.4 V in the first hour which under the fixed load corresponds
voltage, the initial current would be equivalent to a C/10 (1.2 mA) to a cell current of o0.16 mA. During the first hour the cathode
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
discharge rate. The fixed load step was applied for a total of three potential decreases initially, but then recovers to B3.6 V vs.
days to represent an extended period of near zero volt storage. Li/Li+ while the anode potential increases rapidly up to B3.4 V vs.
The 3-electrode data for the 5th cycle during discharge is Li/Li+. Both electrodes plateau at these potentials for B2 hours.
shown as a function of discharge capacity in Fig. 4a and a The cathode then decreases and asymptotes to within 10 mV of
Fig. 4 (a) 5th cycle discharge and fixed load step (grey shading) cell voltage and electrode potentials plotted vs. cell capacity of a conventional LiCoO2/
MCMB pouch cell. (b) 5th cycle discharge and fixed load step (grey shading) cell voltage and electrode potentials plotted vs. time of a conventional
LiCoO2/MCMB pouch cell. Red shading represents assigned anode potential range in which copper dissolution occurs. (c) Schematic of LiCoO2/MCMB
cell function in the near zero volt condition. Orange arrows represent copper ion flow. Red arrow represents electron flow through external circuit. (d)
Linear sweep voltammogram of copper foil. Red shaded region represents assigned potential range in which copper dissolution occurs.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2225
View Article Online
anode potential at B3.2 V vs. Li/Li+. Once the cell reaches this zero volt storage, cell design must be modified or different materials
point, the cell voltage is o10 mV and the cell current is o4 mA used to avoid the copper dissolution degradation mechanism.
(C/3000) which can be considered to be a quasi-equilibrium state.
The quasi-equilibrium state remains for the final 66 hours of the 3.5 Electrode asymptotic potential
fixed load step. The electrode potentials slightly decrease during As shown in Fig. 4b, after a period of time the electrode
this time by about 0.1 V to B3.1 V vs. Li/Li+ and the cell voltage potentials asymptote to within a small potential difference of
decreases to B3 mV. Over the course of the 3 day fixed load step, each other and no longer change significantly (cell reaches
the cell discharged an additional B0.5 mA h. quasi-equilibrium state). It is therefore convenient to define a
The initial decrease and subsequent recovery of the electro- single potential that is representative of the minimally varying
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
chemical potential of the cathode during the transient period potential of both electrodes when the cell is in the near zero
may not be expected, but similar results have been observed in volt, quasi-equilibrium state. This quantity will be defined herein
past work.86 The decrease and recovery of the cathode potential as the electrode asymptotic potential (EAP) and is described by
may be due to several non-ideal effects such as changing ionic eqn (1) where VC is the cathode potential vs. Li/Li+ and VA is the
activities, the effects of reaction rate constants or non-faradaic anode potential vs. Li/Li+.
processes as the cell current decreases. Regardless, the cathode
VC VA
potential remaining in its normal range for several hours and EAP ¼ þ VA when ðVC VA Þ o c (1)
2
intercalating 0.5 mA h of additional lithium at B3.6 V vs. Li/Li+
during the fixed load step supports that the cathode can still The threshold potential difference c is chosen to represent
intercalate more lithium in its normal potential range after the when the cell reaches a quasi-equilibrium state in which the
normal discharge. electrode potentials no longer change significantly. c may have
The rapid rise in potential of the anode when the fixed load to be defined differently for different cell configurations, fixed
is applied is consistent with the anode being mostly depleted of load values and active materials used. The value of c is chosen
intercalated lithium ions and highly polarizable at the end of to be 10 mV in the present study. Once the electrode potential
discharge. The major concern with the rise in anode potential is difference reaches c, the EAP drifts by only B0.1 V during the
that onset of copper dissolution into the electrolyte can occur remaining 66 hours of the fixed load period, indicating that
as depicted in Fig. 4c. 10 mV is a sufficient choice for c in the present case.
Fig. 4d quantifies the copper dissolution process in a linear
sweep voltammogram, which uses a polished copper working
electrode, polished platinum counter electrode and lithium reference
4 Past and current concepts and
electrode in the same configuration as shown in Fig. 2a and b. A approaches to near zero volt tolerance
Princeton Applied Research Versastat 3 was used for the voltammetry of lithium ion cells
and swept the potential of the copper foil in the positive direction
at 10 mV s1 from the open circuit voltage (OCV) of B3.1 V vs. There have been several approaches to make lithium ion cells
Li/Li+. As can be seen, an oxidative current from copper foil tolerant to the general case of overdischarge. Many approaches
initiates near the OCV of B3.1 V vs. Li/Li+ and reaches current to general overdischarge protection can be applied to a near
densities of B0.01 mA h cm2 at 3.3 V vs. Li/Li+, B0.1 mA h zero volt storage scenario and will be included in the following
cm2 at 3.4 V vs. Li/Li+ and B1.0 mA h cm2 at 3.5 V vs. Li/Li+. sections. However, approaches to protect a cell from overdischarge
The oxidative current is primarily attributed to copper ions to a negative cell voltage, such as redox shuttle additives,39,44
stripping off the surface of the copper foil resulting in soluble cannot be applied to a fixed load, near zero volt scenario and thus
cations, consistent with previous studies.62,87–89 are not considered.
It is important to note here that determined values of the
onset potential for copper dissolution can vary since the choice of 4.1 Zero crossing potential
an oxidative current density threshold is arbitrary and non-faradaic Many approaches to near zero volt storage or overdischarge
process can contribute to the oxidative current. Additionally, the protection have made use of predicting a zero crossing potential
exact value of the onset of copper dissolution can be affected by (ZCP)90–93 at which the electrode potentials meet when a cell is
factors such as ambient conditions and electrolyte composition. In discharged to zero volts. Although rigorous descriptions of how
the present work, 3.1 V vs. Li/Li+ is assigned for interpretation the ZCP is predicted are absent, examination of prior work
purposes as less than this value, no oxidative current is expected suggests that the ZCP has been predicted based on attempting
based on the linear sweep voltammogram shown in Fig. 4d, and to match and extrapolate half-cell (vs. Li metal) constant current
therefore, no copper dissolution is expected to be occurring. discharge profiles of each electrode. For fixed load storage of
The adverse effects from copper ions being present in the cells at near zero volts, the predicted ZCP is regarded as the
electrolyte will lead to side reactions with the electrolyte, potential of the electrodes throughout the duration of the fixed
competitive reduction processes with lithium ions, and copper load storage period, analogous to the EAP.
dendrite formation which can lead to internal shorting. Each of The ZCP method is unlikely to accurately represent the true
these effects can cause significant damage to the cell and reduce its transient behavior of the electrode potentials and EAP of a
capacity and performance.58–64 Thus, for cells to be tolerant to near cell when a fixed load is applied. Loss due to SEI formation,
2226 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
incomplete intercalation of the anode (due to excess anode can tolerate fixed load, near zero volt105 and overdischarge106
used to prevent plating), electrode composite variation and any conditions. However, lower bulk electrical conductivity, coulombic
other non-ideal effects that arise can be difficult to account for. efficiency issues from substantial SEI formation on nanoscale
Additionally, in a practical near zero volt storage situation surface, and higher cost are disadvantages of the nanocarbon-
where a fixed load is applied, the cell current decreases to very based current collectors compared to copper.104
low levels (e.g. BC/1000–C/10 000 rate) when the cell voltage In addition to replacing copper current collectors with a
approaches zero volts. As a result, half-cell discharge profiles more stable material, a similar concept to protect cells during
resulting from more typical, constant discharge rates (e.g. C/10, near zero volt/overdischarge conditions is to passivate copper
C/20) are not good predictors of the transient behavior of the current collectors and prevent dissolution at high potentials vs.
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
electrode potentials or the EAP of the cell when the cell voltage Li/Li+. One approach is the use of succinonitrile as an electrolyte
(and cell current) approaches zero. Lastly, under fixed load additive to passivate the copper current collector and prevent its
storage at near zero volts, the concept of a ZCP is not physically corrosion.62 However, succinonitrile additive has also been
accurate as the electrode potentials do not ever cross each other shown to significantly increase the impedance of cells during
and the cell never actually reaches zero volts. Rather, the electrode cycling107 which can limit rate capability. Formation of nitrile
potentials asymptote towards an intermediate potential (defined compounds on the surface of copper foil before electrode
as the EAP in Section 3.5) and the cell gets to a near zero volt state. fabrication108 has also shown promise, but it is unclear what
Thus, use of a ZCP prediction is insufficient for accurately effect a nitrile coating might have on the charge transfer resistance
characterizing electrode behavior during near zero volt storage between the copper current collector and anode composite. Future
under fixed load. Instead, approaches to near zero volt storage approaches to inhibit corrosion of the copper current collector,
should make use of reference electrode measurements to accurately with minimal effect on battery performance, may present a viable
and definitively characterize the behavior of the electrode potentials. pathway to near zero volt storage tolerant cells.
4.2 Anode current collectors that are stable at high potentials 4.3 Secondary electrode materials with intermediate
vs. Li/Li+ charge/discharge potentials
One strategy to avoid copper dissolution in a near zero volt state Use of secondary active materials in the cathode that have an
has been to employ alternative anode current collectors which intercalation potential below the copper dissolution potential
do not undergo dissolution at higher potentials vs. Li/Li+. can protect a cell in a near zero volt scenario. Such a secondary
Voltammetry studies, similar to the one shown in Fig. 4c, have material can force the cathode potential to be less than the
been done previously87,88,94,95 on potential current collector copper dissolution potential during overdischarge. In a near
replacement materials. The voltammetry studies can determine zero volt storage situation, this effect can keep the anode potential
the electrochemical stability of a current collector replacement less than the copper dissolution potential, thus preventing copper
in the potential range of interest. Potential anode current dissolution. Li2MoO486,93,109 as an additive to a LiCoO2 cathode
collector replacements must demonstrate increased stability has been demonstrated to work in this regard and protect cells
at high potentials vs. Li/Li+ (i.e. 43.5 V) as well as low potentials during overdischarge. Several other secondary cathode materials,
vs. Li/Li+ (i.e. 5–1000 mV). Some metals, like aluminum,85,94–100 like LiS2 and VxOy, with the appropriate intercalation potentials
that are stable at high potentials vs. Li/Li+ will alloy with exist and have been patented.86,110
lithium at low potentials in conventional LiPF6 electrolytes. Use of secondary active materials in the anode with inter-
This would lead to pulverization of the anode current collector mediate charge/discharge potentials has also been patented90
during normal cycling. Titanium, titanium alloys, nickel, nickel as a method to protect cells during overdischarge. Specifically,
alloys and stainless steel have all been patented90,91 as potential the secondary active material would have an intermediate
current collector replacements because they meet the high and de-intercalation potential that is greater than the potential
low potential stability requirements. where the cathode can begin to degrade and less than the
Titanium foils are showing promise in commercially developed copper dissolution potential. The secondary active material
near zero volt storage capable cells101 as an anode current collector. would contain additional active lithium that, upon overdischarge,
However, tradeoffs exist with titanium foils in that they are typically would begin de-intercalating lithium as soon as the graphite was
thicker than standard copper foils and can cost substantially more depleted of lithium. In concept, this will keep the anode potential
due to increased extraction and processing costs.102 These tradeoffs less than the copper dissolution potential during overdischarge. It
can reduce both the volumetric and gravimetric energy density will also prevent the cathode potential from decreasing to the point
of the cells while increasing cell cost. Additionally, bulk titanium is where it can begin to degrade. The stability of the many suggested
more than an order of magnitude more resistive than copper, secondary materials in the low anode potential range during normal
which may limit the rate capability of cells, especially in large operation, namely metal oxides, is a major operational concern in
format or wound cells. pursuing this approach.111
Carbon nanotube103,104 and graphene free-standing electrodes Overall, no widely available validation or experimental data
may also be potential current collector replacements due to their exists on many of the patented secondary active materials.
high chemical stability. Anodes made purely of other carbon Thus, the tolerance to multi-day near zero volt storage and
allotropes have already been demonstrated to generate cells that impact on performance during normal operation is unclear.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2227
View Article Online
Additionally, adding an amount of secondary active materials increase implementability of near zero volt tolerant cells which
with intermediate charge/discharge potentials will likely decrease can impact the rapid and widespread adoption of lithium ion
cell energy density by lowering the average discharge voltage cells that are safer to transport.
and/or electrode specific capacity. Based on the discussion in Sections 3 and 4, it can be
expected that anode pre-lithiation prior to cell assembly will
4.4 Secondary cathode active materials with high first cycle increase the amount of reversible lithium in a cell, which
loss should change the behavior of the electrode potentials during
Another strategy towards fabricating cells with tolerance to near zero volt storage. Hypothetically, a cell having more
overdischarge is to generate a lithium ion cell with a higher reversible lithium than necessary to fully lithiate the cathode
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
ratio of reversible lithium to cathode capacity by using high first upon discharge will prevent the anode potential from increasing
cycle loss cathode additives. Li2NiO2 can deliver B300 mA h g1 to the copper dissolution potential upon fixed load discharge to
of lithium on the first charge and then its decomposition near zero volts. If the amount of pre-lithiation is minimal,
products can only intercalate B70–100 mA h g1 when the damage to the cathode, which can result from a low cathode
discharge is stopped at 3.0 V vs. Li/Li+.112–114 Use of Li2NiO2 as a potential during near zero volt storage,90 can also be minimized
secondary cathode material has been shown to protect cells or mitigated. Thus, anode pre-lithiation is a potentially affective
from overdischarge.63 The higher ratio of reversible lithium to approach to realize lithium ion cells that are highly tolerant to
cathode capacity leads to the anode not being depleted and the near zero volt storage.
cathode approaching full intercalation at the end of discharge. Anode pre-lithiation does not introduce unconventional
Additionally, the breakdown products of Li2NiO2 have a lower materials or necessitate change in the battery construction
discharge potential and act as a secondary active material in the parameters, so such an approach can avoid the anticipated
cathode like those described in Section 4.3. As a result, the tradeoffs of the current and past approaches to near zero volt
anode potential does not reach the copper dissolution potential storage tolerance described in Section 4. Additionally, anode
during overdischarge. pre-lithiation in concept could be applied to a range of different
Although a promising approach, Li2NiO2 as a secondary active active material combinations, since there are no operational
material in the cathode has been shown to lead to substantial gas concerns related to secondary active material stability or reactions
evolution during initial charging and at high temperatures.114 with the primary active material.
Gas evolution can cause severe and potentially dangerous swelling Pre-lithiation of a graphite anode has also been shown to
of cells during the conditioning process or during normal use. It have the additional benefit of improving performance of con-
has also been shown that Li2NiO2 additive will increase the fade ventional cells117–119 which can help to reduce the cost/benefit
rate of a LiCoO2/graphite cell.63 Additionally, the breakdown ratio. Also, it may eliminate the need for formation cycling,120
products of Li2NiO2 will likely decrease the energy density of the which could serve to reduce manufacturing costs. Thus, addition
cell due to their lower capacity and discharge voltage compared of anode pre-lithiation to the battery manufacturing process may
to many common cathode materials like LiCoO2. not substantially increase the cost of lithium ion cells and could
Thus, operational concerns exist regarding the use of Li2NiO2 improve general performance in addition to enabling near zero
as a secondary active material in the cathode to protect the cell volt storage tolerance.
during overdischarge. These concerns include potential for Additionally, several methods of pre-lithiation with potential
substantial gas evolution or side reactions, effect on cycling for industrial scalability exist. Use of Stabilized Lithium Metal
stability, and impact on energy density. Other secondary active Powder (SLMP)121 has been shown to be an effective way to
materials like Li6CoO4115 or additives like Li2O116 that have a pre-lithiate anodes.122 Electrochemical pre-lithiation such as bath
high first cycle charge/discharge capacity ratio may protect cells pre-lithiation has also been developed and is a promising option
during overdischarge in the same manner as Li2NiO2. However, that is scalable.123–125 Therefore, industrially scalable pre-lithiation
in all cases where secondary active materials are used to increase methods exist and could be utilized for large scale production.
the amount of reversible lithium, there are significant concerns
5.1 Pre-lithiation process and resulting effect on behavior of
regarding the material impact on cell energy density and
electrode potentials during normal charge and discharge
operational stability which warrant development of an improved
approach to near zero volt storage tolerance. In the present work, a lab-scale pre-lithiation method using a
sacrificial cathode is used for demonstration of the approach. A
LiCoO2/MCMB cell is built with a pre-lithiated anode. All cells
5 Novel approach to near zero volt described herein are fabricated with the same construction
storage capability using pre-lithiation parameters described in Section 3.1. Fig. 5 illustrates the
of anodes process that was used (in a dry Argon glove box maintained
ato1.0 ppm oxygen and o1.0 ppm water) to increase the
In light of the numerous anticipated performance and cost amount of reversible lithium present in a cell for near zero volt
tradeoffs of past and current approaches to near zero volt storage evaluation.
storage tolerance, it is highly desirable to identify a new method Initially, a LiCoO2/MCMB pouch cell was constructed (with-
that avoids these anticipated tradeoffs. Such a method would out a lithium reference electrode) and was cycled at a C/10 rate
2228 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
twice from 3.0–4.1 V cell voltage then partially charged by vs. 3.789 V conventional). The lower cathode potential at the
0.24 mA h as depicted in Fig. 5a. At this state of cycling, the end of discharge indicates that at this point the cathode is
SEI layer on the anode will be formed and the graphite will be approaching full intercalation as depicted in Fig. 6c. The anode
partially (B2% of its capacity) lithiated, whereas the cathode is potential is also B350 mV lower in the RLE cell than in the
partially depleted of lithium (see Fig. 5a). conventional cell at the end of discharge (0.355 V RLE vs.
As depicted in Fig. 5b and c, the cell was then disassembled 0.785 V conventional). The lower anode potential at the end
and the pre-lithiated anode was reassembled into a cell with a of discharge indicates that at the end of discharge the anode is
fresh LiCoO2 cathode which is intrinsically fully lithiated (and a not fully depleted of reversible lithium as depicted in Fig. 6c.
reference electrode in this step – not shown in Fig. 5c). Since Thus, the process of using a pre-lithiated anode manifests
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
the cathode is fully lithiated, the SEI is already formed on the itself in changes to the behavior of the electrode potentials
anode and the anode has some lithium intercalated into it, it during normal charge and discharge. This is expected to impact
can be expected that the resulting cell will have more reversible the resulting transient behavior of the electrode potentials and
lithium than is required to fully intercalate the cathode upon the cell’s EAP during near zero volt storage under fixed load.
discharge. This will be referred to herein as a reversible lithium
excess (RLE) cell condition.
The RLE cell with a lithium reference electrode was conditioned 5.2 Electrode potentials during near zero volt storage of an
by cycling once at a C/20 rate and then four times at a C/10 rate. RLE cell
The cell was charged and discharged in the same manner as After the C/10 discharge to 3.0 V cell voltage on the 5th
described in Section 3.2 between 3.0 V full cell voltage and a conditioning cycle of the RLE cell, a fixed load of 2.5 kOhms
cathode potential of 4.2 V vs. Li/Li+. The first cycle charge and was applied to the cell to discharge it to near zero volts (i.e.
discharge curves of the RLE cell, including the electrode potentials actual voltage reaches 1–3 mV). The fixed load was applied for
as measured by a lithium reference electrode, are shown in Fig. 6a three days to represent an extended near zero volt storage period.
and b, respectively. It is observed that the increase of the amount The 3-electrode data during the 5th cycle discharge is shown
of reversible lithium has changed the behavior of the electrode as a function of discharge capacity in Fig. 7a and a function of
potentials during normal charge and discharge. At the start of the time in Fig. 7b. As shown in Fig. 7b, during the C/10 discharge
charge, the anode potential is B700 mV vs. Li/Li+ lower than the (first 10 hours) the cell current is constant and the electrode
conventional cell (i.e. 0.374 V RLE vs. 1.105 V conventional) and as potentials exhibit their normal discharge behavior. At B10 hours,
a result the cell voltage is B700 mV vs. Li/Li+ higher during the the cell reaches 3.0 V and the fixed load (2.5 kOhm) step is
initial stage of charge (i.e. 3.460 V RLE vs. 2.773 V conventional). initiated. Under fixed load, the RLE cell, like the conventional cell
The lower anode potential at the beginning of charge is consistent described in Section 3.4, initially undergoes a transient period as it
with it being partially lithiated upon cell assembly. It is important discharges to near zero volts. As shown in Fig. 7b, the cell voltage
to note that at the end of charge the anode potential is greater than decreases to o0.1 V in the first two hours which under the fixed
0 V vs. Li/Li+ (i.e. B50 mV), and therefore it still has sufficient load corresponds to a cell current of o0.04 mA. During the first
capacity to intercalate all lithium during the charge. This indicates two hours the cathode potential initially decreases to B1.3 V vs.
that, despite pre-lithiation of the anode prior to final cell assembly, Li/Li+ and then increases to B1.4 V vs. Li/Li+. The anode
no lithium plating is likely to be occurring. potential initially increases to B0.6 V vs. Li/Li+, then decreases
At the end of discharge, the cathode potential is B350 mV to B0.5 V vs. Li/Li+ as the cell current decreases. The anode potential
lower in the RLE cell than in the conventional cell (3.360 V RLE then increases to 41.0 V vs. Li/Li+ after 2 hours at fixed load.
Fig. 5 (a) Schematic of 0.24 mA h charge of a conditioned, discharged cell. (b) Schematic depicting disassembly of cell and discarding cathode.
(c) Schematic of reassembly for partially lithiated anode with fresh cathode in a new cell.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2229
View Article Online
2230 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
discharge capacity after the third, 3 day near zero volt storage
period. Additionally, only a minor fade in the average discharge
voltage of 25 mV was observed. Such a drastic improvement in
performance retention over the conventional cell is attributed
to the fact that as measured by the reference electrode, the
potential of the anode never exceeds 2.0 V vs. Li/Li+ during each
near zero volt storage period. This is significantly less than the
assigned copper dissolution potential of 3.1 V vs. Li/Li+ and as
such no copper dissolution is expected to be occurring.
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2231
View Article Online
Fig. 8 (a) Cycling schedule flow chart depicting the duration, number of cycles and near zero volt storage periods of conventional and RLE
LiCoO2/MCMB pouch cells. (b) Discharge profiles of a conventional LiCoO2/MCMB pouch cell prior to zero volt storage and after one, two and
three, 3 day near zero volt storage periods. (c) Discharge profile of the conventional LiCoO2/MCMB pouch cell electrode potentials as measured by a
lithium metal reference prior to near zero volt storage and after three, 3 day near zero volt storage periods. (d) Discharge profiles of an RLE
LiCoO2/MCMB pouch cell prior to zero volt storage and after one, two and three, 3 day near zero volt storage periods. (e) Discharge profile of the RLE
LiCoO2/MCMB pouch cell electrode potentials as measured by a lithium metal reference prior to near zero volt storage and after three, 3 day near zero
volt storage periods.
2232 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
5.4 Effect of near zero volt storage on rate capability and long
term cycling
Conventional and RLE prototype LiCoO2/MCMB pouch cells
were constructed (without reference electrodes) with the same
construction parameters as described in Section 3.1 and 5.1,
respectively. The cycling schedule for the cells is shown in the
flow chart in Fig. 10a. The cells were cycled between 3.0–4.1 V
and conditioned with 1 cycle at a C/20 rate followed by 4 cycles
at a C/10 rate. After the 5th cycle discharge the RLE cell was
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2233
View Article Online
2234 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
As shown in Fig. 11c, during the first 2 hours of the fixed After the first two hours of fixed load storage, the electrode
load step, the cell voltage quickly decreases to o20 mV. The potentials increase to B1.7 V vs. Li/Li+, where after B11 hours
cathode potential decreases to B1.2 V vs. Li/Li+ and then of the fixed load step, they asymptote to within 10 mV of each
recovers up to B1.5 V vs. Li/Li+. The anode potential steadily other at B1.75 V vs. Li/Li+, realizing the EAP for this cell and
increases to B1.5 V vs. Li/Li+. reaching a quasi-equilibrium state. The electrode potentials
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
Fig. 11 (a) Cycling schedule flow chart summarizing room temperature conditioning, initial cycling at 45 1C, and near zero volt storage testing at
45 1C of RLE LiCoO2/MCMB pouch cell. (b) 7th cycle discharge and fixed load step (grey shading) cell voltage and electrode potentials plotted vs. cell
capacity of RLE LiCoO2/MCMB pouch cell at 45 1C. (c) 7th cycle discharge and fixed load step (grey shading) cell voltage and electrode potentials plotted
vs. time of RLE LiCoO2/MCMB pouch cell. (d) Discharge profiles of an RLE LiCoO2/MCMB pouch cell prior to zero volt storage, and after one, two and
three, 3 day near zero volt storage periods at 45 1C. (e) Discharge profile of the RLE LiCoO2/MCMB pouch cell electrode potentials as measured by a
lithium metal reference prior to near zero volt storage, and after three, 3 day near zero volt storage periods, all run at 45 1C.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2235
View Article Online
stay at B1.75 V vs. Li/Li+ for the remainder of the fixed load reported approaches have anticipated tradeoffs with regards to
storage period. performance and cost of lithium ion cells.
The behavior of the electrode potentials at elevated temperature In the present study, anode pre-lithiation was demonstrated to
is very similar to that at room temperature. The cell took B5 hours be a promising alternative approach to render lithium ion cells
longer to reach the defined threshold of 10 mV at elevated tempera- tolerant to near zero volt storage. LiCoO2/MCMB pouch cells
ture, but otherwise the transient behavior and EAP are similar. fabricated with a pre-lithiated anode to generate an RLE cell
The discharge performance retention of the RLE cell at 45 1C condition demonstrated high tolerance to three 3 day and three
is plotted in Fig. 11d. As shown, the discharge performance 7 day storage periods at near zero volts. Both cells maintained 99%
retention is very similar to that at room temperature. After of their discharge capacity and only minor average discharge voltage
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
three, 3 day storage periods at 45 1C under fixed load at near decreases of r25 mV were observed. Additionally, storage of an RLE
zero volts, the cell discharged with 103% of the cell capacity cell at near zero volts for a 3 day period did not noticeably diminish
prior to near zero volt storage under fixed load with no decrease its rate capability up to a 5C discharge rate or diminish its long term
in average discharge voltage. The strong performance retention cycling stability up to 500, 30% DOD LEO cycles. LiCoO2/MCMB
is attributed to the cell’s EAP remaining at B1.6–1.7 V vs. Li/Li+ pouch cells fabricated with a pre-lithiated anode also demonstrated
for each fixed load storage period, so no copper dissolution is high tolerance to near zero volt storage at 45 1C, with no loss of
expected to be occurring during any of the near zero volt discharge capacity or decrease in average discharge voltage. The
storage periods and the SEI is likely not oxidatively stripped. approach of using pre-lithiated anodes to enable near zero volt
Fig. 11e shows the measured electrode potentials during storage tolerance in lithium ion cells is promising to be highly
discharge before and after the one, two and three, 3 day storage implementable since an anode pre-lithiation step was the only
periods under fixed load at near zero volts at 45 1C. As shown requirement and scalable pre-lithiation methods exist. No
the cathode performance fades similar to the cell tested at change to conventional cell construction parameters is required
room temperature, with a lower average discharge voltage in and no notable change in cell performance was observed.
the second half of discharge. The fade in cathode discharge The present study utilized reference electrode measurements
performance indicates that the low electrochemical potential which monitor the true behavior of the electrode potentials during
experienced by the cathode during near zero volt storage may near zero volt storage under fixed load. A transient period is observed
be causing minor damage to the cathode material. in which the electrode potentials migrate towards each other
The anode discharge performance, on the other hand, followed by a quasi-equilibrium state when the electrode potentials
changes in contrast with results observed at room temperature. asymptote to a semi-constant intermediate potential (the EAP). The
The average discharge potential of the anode decreases after reference measurements informed on the effect of anode pre-
three, 3 day storage periods under fixed load at near zero volts. lithiation to keep the anode potential below the copper dissolution
The change may be attributed to the effects of the VC electrolyte potential during extended near zero volt storage. Future work in the
additive or the elevated temperature on the SEI layer of the area of near zero volt storage tolerance should make use of reference
anode during near zero volt storage. The decreased anode electrode measurements rather than relying on predictions of a ZCP
potential during cell discharge results in an overall increase method based on half-cell cycling data of the electrodes.
in the cell voltage. Reference electrode measurements indicate Demonstration of scalable pre-lithiation methods to enable near
that the electrodes in RLE type LiCoO2/MCMB cells undergo no zero volt storage tolerance and precise optimization of the amount
severe detrimental effects from the elevated temperature, near of pre-lithiation of the anode based on reference measurements can
zero volt storage. Thus, anode pre-lithiation is a viable approach to further improve and expand upon the results shown here.
near zero volt storage tolerance at elevated temperatures which can Modifying the cathode to stabilize it against the near zero volt
impact batteries used in implanted medical devices and cells storage conditions of an RLE cell may also further improve near
experiencing high temperatures during transit or storage. zero volt storage tolerance. The anode pre-lithiation approach
can also be applied to lithium ion cells using other common
cathode materials such as LiFePO4 and LiNi0.8Co0.15Al0.05O2 (NCA).
6 Conclusions and outlook Overall, implementable approaches to achieve near zero volt
storage capability of lithium ion cells without sacrificing cell
Near zero volt storage tolerance has the potential to substantially quality in the coming years can help to enhance safety during
increase the safety of lithium ion batteries during storage and storage and shipment and have several other potential safety
shipment as well as have several other benefits across medical, benefits. Increased safety will help to increase use and public
aerospace, EV, home storage and other applications. Previous reception while decreasing costs associated with lithium ion
commercial approaches to accomplish this goal have included batteries as society moves toward a more electrified infrastructure.
replacement of the copper current collector with a more electro-
chemically stable material like Ti. Other studied strategies have
included stabilizing copper with succinonitrile electrolyte additive, Acknowledgements
forming a nitrile compound on the surface of copper before cell
assembly, and utilizing secondary active materials with intermediate The authors appreciate insightful discussions and experimental
discharge potentials or high first cycle loss. However, all past assistance from Dr Michael W. Forney of Aerospace Corporation
2236 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
as well as Jason W. Staub, Dr Matthew J. Ganter, Anthony P. 20 P. Ping, Q. Wang, P. Huang, K. Li, J. Sun, D. Kong and
Leggiero and Dr Christopher M. Schauerman of the Rochester C. Chen, J. Power Sources, 2015, 285, 80–89.
Institute of Technology. The authors acknowledge financial 21 D. Doughty and E. P. Roth, Electrochem. Soc. Interface, 2012,
support from the U.S. Government and the Air Force Research 37–44.
Laboratory under agreement number FA9453-14-1-0239. The 22 P. G. Balakrishnan, R. Ramesh and T. Prem Kumar,
U.S. Government is authorized to reproduce and distribute J. Power Sources, 2006, 155, 401–414.
reprints for Governmental purposes notwithstanding any copy- 23 H. Lee, M. Yanilmaz, O. Toprakci, K. Fu and X. Zhang,
right notation thereon. The views and conclusions contained Energy Environ. Sci., 2014, 7, 3857–3886.
herein are those of the authors and should not be interpreted as 24 C.-C. Lin, H.-C. Wu, J.-P. Pan, C.-Y. Su, T.-H. Wang, H.-S.
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
necessarily representing the official policies or endorsements, Sheu and N.-L. Wu, Electrochim. Acta, 2013, 101, 11–17.
either expressed or implied, of Air Force Research Laboratory 25 J. Cho, T. G. Kim, C. Kim, J. G. Lee, Y. W. Kim and B. Park,
(AFRL) or the U.S. Government. K. R. C. acknowledges financial J. Power Sources, 2005, 146, 58–64.
support from the SMART Scholarship Program. 26 J. Cho, J. Power Sources, 2004, 126, 186–189.
27 J. Cho, Electrochim. Acta, 2003, 48, 2807–2811.
28 J. Y. Shi, C. W. Yi and K. Kim, J. Power Sources, 2010, 195,
6860–6866.
Notes and references
29 G. Nagasubramanian and K. Fenton, Electrochim. Acta,
1 M. Endo, C. Kim, K. Nishimura, T. Fujino and K. Miyashita, 2013, 101, 3–10.
Carbon, 2000, 38, 183–197. 30 K. Xu, M. S. Ding, S. Zhang, J. L. Allen and T. R. Jow,
2 G. M. Ehrlich, Lithium-Ion Batteries, in Handbook of Batteries, J. Electrochem. Soc., 2002, 149, A622.
3rd edn, 2002. 31 K. Xu, S. Zhang, J. L. Allen and T. R. Jow, J. Electrochem.
3 E. Stura and C. Nicolini, Anal. Chim. Acta, 2006, 568, 57–64. Soc., 2002, 149, A1079.
4 E. M. Erickson, C. Ghanty and D. Aurbach, J. Phys. Chem. 32 M. Otsuki and T. Ogino, Lithium-Ion Batter. Sci. Technol.,
Lett., 2014, 5, 3313–3324. 2009, 121, 275–289.
5 G.-H. Kim, A. Pesaran and R. Spotnitz, J. Power Sources, 33 H. Ota, A. Kominato, W. J. Chun, E. Yasukawa and S. Kasuya,
2007, 170, 476–489. J. Power Sources, 2003, 119–121, 393–398.
6 Y. F. Chen and J. W. Evans, J. Electrochem. Soc., 1996, 143, 34 J. Arai, J. Appl. Electrochem., 2002, 32, 1071–1079.
2708–2712. 35 X. Wang, E. Yasukawa and S. Kasuya, J. Electrochem. Soc.,
7 S. Al-Hallaj, H. Maleki, J. S. Hong and J. R. Selman, J. Power 2001, 148, A1066.
Sources, 1999, 83, 1–8. 36 M. S. Ding, K. Xu and T. R. Jow, J. Electrochem. Soc., 2002,
8 R. Spotnitz and J. Franklin, J. Power Sources, 2003, 113, 149, A1489.
81–100. 37 J. R. Dahn, J. Jiang, L. M. Moshurchak, M. D. Fleischauer,
9 T. M. Bandhauer, S. Garimella and T. F. Fuller, J. Electrochem. C. Buhrmester and L. J. Krause, J. Electrochem. Soc., 2005,
Soc., 2011, 158, R1. 152, A1283.
10 H. Maleki, J. Electrochem. Soc., 1999, 146, 3224. 38 M. N. Golovin, J. Electrochem. Soc., 1992, 139, 5.
11 T. G. Zavalis, M. Behm and G. Lindbergh, J. Electrochem. 39 J. Chen, C. Buhrmester and J. R. Dahn, Electrochem. Solid-
Soc., 2012, 159, A848. State Lett., 2005, 8, A59–A62.
12 M. Reichert, J. Haetge, D. Berghus, C. Wendt, V. Meier, 40 T. J. Richardson, J. Electrochem. Soc., 1996, 143, 3992.
U. Rodehorst, S. Passerini, F. Schappacher and M. Winter, 41 S. R. Narayanan, S. Surampudi, A. I. Attia and C. P.
ECS Trans., 2014, 61, 87–103. Bankston, J. Electrochem. Soc., 1991, 138, 2224–2229.
13 A. W. Golubkov, S. Scheikl, R. Planteu, G. Voitic, H. Wiltsche, 42 C. Buhrmester, J. Chen, L. Moshurchak, J. Jiang, R. L.
C. Stangl, G. Fauler, A. Thaler and V. Hacker, RSC Adv., 2015, Wang and J. R. Dahn, J. Electrochem. Soc., 2005, 152,
5, 57171–57186. A2390.
14 O. S. Mendoza-hernandez, H. Ishikawa, Y. Nishikawa, 43 M. Adachi, J. Electrochem. Soc., 1999, 146, 1256.
Y. Maruyama and M. Umeda, J. Power Sources, 2015, 280, 44 C. Buhrmester, L. Moshurchak, R. L. Wang and J. R. Dahn,
499–504. J. Electrochem. Soc., 2006, 153, A288.
15 A. Hammami, N. Raymond and M. Arman, Nature, 2003, 45 J. P. Schmidt, A. Weber and E. Ivers-Tiffée, J. Power Sources,
424, 635–636. 2015, 274, 1231–1238.
16 H. Maleki, G. Deng, A. Anani and J. Howard, J. Electrochem. 46 M. D. Levi, C. Wang and D. Aurbach, J. Electrochem. Soc.,
Soc., 1999, 146, 3224–3229. 2004, 151, A781.
17 H. Yang, H. Bang, K. Amine and J. Prakash, J. Electrochem. 47 E. Markevich, M. D. Levi and D. Aurbach, J. Electrochem.
Soc., 2005, 152, A73. Soc., 2005, 152, A778.
18 Q. Wang, P. Ping, X. Zhao, G. Chu, J. Sun and C. Chen, 48 R. Yazami and Y. F. Reynier, Electrochim. Acta, 2002, 47,
J. Power Sources, 2012, 208, 210–224. 1217–1223.
19 J. Lamb, C. J. Orendorff, L. A. M. Steele and S. W. Spangler, 49 C. Wang, X. W. Zhang, a. J. Appleby, X. Chen and F. E. Little,
J. Power Sources, 2014, 283, 517–523. J. Power Sources, 2002, 112, 98–104.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2237
View Article Online
2238 | Energy Environ. Sci., 2016, 9, 2219--2239 This journal is © The Royal Society of Chemistry 2016
View Article Online
104 B. J. Landi, M. J. Ganter, C. M. Schauerman, C. D. Cress 131 L. Chen, K. Wang, X. Xie and J. Xie, J. Power Sources, 2007,
and R. P. Raffaelle, J. Phys. Chem. C, 2008, 112, 7509–7515. 174, 538–543.
105 J. R. Dahn, U. Vonsacken, M. W. Juzkow and H. Aljanaby, 132 L. Hu, Z. Xue, K. Amine and Z. Zhang, J. Electrochem. Soc.,
J. Electrochem. Soc., 1991, 138, 2207–2211. 2014, 161, A1777–A1781.
106 S. Hossain, Y. Saleh and R. Loutfy, J. Power Sources, 2001, 133 O. Matsuoka, a. Hiwara, T. Omi, M. Toriida, T. Hayashi,
96, 5–13. C. Tanaka, Y. Saito, T. Ishida, H. Tan, S. S. Ono and
107 G.-Y. Kim, R. Petibon and J. R. Dahn, J. Electrochem. Soc., S. Yamamoto, J. Power Sources, 2002, 108, 128–138.
2014, 161, A506–A512. 134 M. Nie, J. Demeaux, B. T. Young, D. R. Heskett, Y. Chen,
108 Y. S. Kim, S. H. Ahn and M. Y. Son, US Pat., US 8192872B2, A. Bose, J. C. Woicik and B. L. Lucht, J. Electrochem. Soc.,
Published on 26 May 2016. Downloaded by INIST - CNRS (Chemistry) on 2/23/2022 9:48:00 AM.
This journal is © The Royal Society of Chemistry 2016 Energy Environ. Sci., 2016, 9, 2219--2239 | 2239