Professional Documents
Culture Documents
Learningfrom Overpotentialsin Lithium Ion Batteries
Learningfrom Overpotentialsin Lithium Ion Batteries
The practically available specific energy of Li ion batteries (LIB) is highly depending on the used specific charge/discharge current,
since the respective overpotentials of each electrode affect the two vital specific energy parameters, specific capacity and voltage.
Focusing on the positive composite electrode as the specific energy bottleneck, the overall nature of the overpotential is discussed
for the LiNi1/3 Co1/3 Mn1/3 O2 (NCM) active material. It is shown that the characteristic overpotentials during charge (delithiation)
and discharge (lithiation) is state of charge (SOC) and depth of discharge (DOD) dependent, respectively. It was demonstrated
that the discharge characteristics are intertwined with the previous charge conditions, particularly with the charging time and the
specific charge capacity. Increasing both in parallel can even lead to a deterioration of the subsequent specific discharge capacity.
Furthermore, Li+ transport pathways within the NCM composite electrode are discussed and their influence on the observed
overpotential evaluated. Changes of the overpotential are found to be mainly associated with changes within the NCM crystal
structure, which is experimentally supported by the correlation of the SOC dependent overpotential with the XRD determined c-axis
lattice parameter. Consequently, the Li+ transport within the active material is mostly responsible for limiting the practically available
specific energy.
© 2016 The Electrochemical Society. [DOI: 10.1149/2.0461614jes] All rights reserved.
Manuscript submitted August 24, 2016; revised manuscript received September 27, 2016. Published November 2, 2016.
For economic exploitation of renewable energy sources, the limiting processes as a basis to further increase the specific energy
development of effective and efficient energy storage devices is and specific power of the LIB cell.
indispensable.1 In this regard the lithium ion battery (LIB) is con- For the sake of reduction of complexity, we focused our investiga-
sidered as key technology thanks to its outstanding characteristics in tions on the positive electrode (= cathode), which is considered to be
terms of high energy density and in most cases high energy efficiency.2 the bottleneck for the increase in cell capacity and specific energy.9,10
In general, the output specific energy (Edis ) of any battery cell is Among the various positive electrode materials, the layered oxide
determined by the discharge voltage (Vdis ) as well as the discharge LiNi1/3 Co1/3 Mn1/3 O2 (NCM) was intensively investigated in literature
capacity (Qdis ): and is of high present and future commercial interest, in particular
Q dis tdis when the Ni content is increased.11–19
E dis = Vdis (q) dq = Idis Vdis (t) dt [1] As starting point, the data in this study was evaluated based on our
0 0 previous insights.20 By thoroughly investigation of the specific capac-
where q represents the specific capacity and tdis is the time in which ity loss in the initial cycle (responsible for the decreased Coulombic
the constant current Idis is drawn from the cell (we neglect in this efficiency) for NCM/Li half cell in the potential range of 4.6–3.0 V
discussion that typically the output power and not the output current vs. Li/Li+ and a used specific current of 30 mA g−1 , it could be shown
is the practically relevant parameter).3 During discharge of the cell, an that the initial specific capacity loss is solely the result of incomplete
internal cell resistance (Ri ) reduces Vdis compared to the open-circuit relithiation during discharge.20 Contrary to the common belief where
voltage (Voc ) by an overvoltage (η). the specific capacity loss was attributed to electrolyte oxidation and
surface layer formation,21–23 it could be concluded that for the given
η = Ri Idis [2] conditions the “parasitic charge capacity” (which is related to charge
reactions not resulting into delithiation, e.g. electrolyte oxidation) of
NCM is negligible.20
Vdis = Voc − η (q, Idis ) [3] As a consequence a significant contribution of oxidative electrolyte
According to Eqs. 1 and 3, the output specific energy is strongly decomposition reactions to the total specific charge capacity could be
dependent on the output current, indicating the important role of the neglected and the specific charge capacity could be solely assigned to
occurring kinetics within a working LIB cell.4 These kinetic limi- Li+ removal from the NCM structure, i.e., as true “Li+ delithiation
tations decrease the thermodynamically achievable specific energy capacity”. The specific discharge capacity is analogously called “lithi-
values by lowering Vdis and in turn Qdis . For a given LIB cell, it is con- ation capacity” and the observed specific capacity loss (= the differ-
sequently possible to enhance the practically available specific energy ence between the specific charge and the specific discharge capacity)
by enhancing the responsible kinetics. could be denominated to “non re-lithiation capacity loss”, which does
Several assessments for rate investigations are known in literature.5 regard that those Li+ site vacancies in the NCM host structure that
However, the elucidation of the aforementioned kinetic processes is a can be not re-lithiated during the discharge (lithiation) step.20 The
challenge. In a multi component LIB cell there are numerous kinetic total specific capacity loss can have a kinetic (thus reversible) and
processes taking place simultaneously including kinetically determin- an irreversible origin. It is possible to distinguish between these both
ing Li+ transport processes and pathways that can theoretically affect types of specific capacity loss by introducing a constant potential (CP)
the overpotentials. These Li+ transport processes/pathways have al- step during lithiation at the discharge cutoff potential, which enables
ready been intensively investigated from a theoretical point of view.6–8 the recovery of the reversible (recoverable) specific capacity loss. The
However, to the best of our knowledge there is no investigation of these residual (non recoverable) specific capacity loss indicates the irre-
processes according to their significance during practical LIB cell ap- versible part.20 According to these insights, the cathode relevant Li+
plication. In this regard, it is important to determine the kinetically extraction ratio parameter can thus be easily determined by the ra-
tio of the experimentally determined specific charge capacity and the
respective calculated specific theoretical capacity.20
∗ Electrochemical Society Fellow. In this work, the influence of the specific current on the electro-
z
E-mail: jkasnatscheew@uni-muenster.de; martin.winter@uni-muenster.de; chemical behavior of a NCM composite electrode is investigated in
m.winter@fz-juelich.de detail. Literature known kinetic factors are discussed and classified
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
A2944 Journal of The Electrochemical Society, 163 (14) A2943-A2950 (2016)
Figure 1. Potential profiles of the first charge using a fixed upper cutoff po- Figure 2. Potential profiles of the first discharge of NCM/Li half cells at varied
tential of 4.6 V vs. Li/Li+ of NCM/Li half cells at varied specific currents. specific currents with a fixed discharge cutoff potential of 3.0 V vs. Li/Li+ . In
all experiments the NCM/Li cells have been charged with a specific current of
30 mA g−1 and up to a charge cutoff potential of 4.6 V vs. Li/Li+ .
according to their significance, thus offering an overview and practical
assessment for battery application. This diagnostic work should act as
basis to overcome one of the main challenges of the LIB cell, namely cific currents, any issues associated with uneven delithiation/internal
enhancing its gravimetric and volumetric energy at elevated rates. gradients can be excluded. The presence of such internal gradients
is observable at significantly higher rates/specific currents,25,26 so the
Experimental relevance can thus be excluded for the given conditions.
The potential profiles reflecting the overpotential during discharge
The LiNi1/3 Co1/3 Mn1/3 O2 (NCM) electrodes were provided by are depicted in Figure 2. In the experimental setup, the charging con-
BMW Group. NCM working electrodes had an active mass load- ditions of the NCM/Li half cells were fixed to the specific current of
ing of 15 mg cm−2 . Battery grade electrolyte 1 M LiPF6 in EC:EMC 30 mA g−1 and the charge cutoff potential to 4.6 V vs. Li/Li+ , which
(1:1, by wt) (LP50 Selectilyte) from BASF was used in all inves- yields a specific charge capacity of ≈220 mAh g−1 (equal SOC). In
tigations. Galvanostatic charge/discharge cycling experiments were the specific discharge capacity range from 0 to 60 mAh g−1 the over-
performed in a three electrode cell set-up (Swagelok). Lithium metal potential is small and almost constant, contrary to the overpotential
(Rockwood Lithium) was chosen as counter- and reference electrode. observed during charge (= delithiation, Figure 1). Only at the end of
A Maccor Series 4000 was used as battery cell test system. Galvano- the lithiation process at a high depth of discharge (DOD), a strongly
static charge/discharge cycling conditions have been varied in this sloping potential profile down to the discharge cutoff potential and
work and are mentioned in the caption of each figure. The discharge a steadily increasing potential differences (= overpotentials) are ob-
cutoff potential was set to 3.0 V vs. Li/Li+ . In several experiments, a served. The increase in specific capacity loss with increasing specific
constant potential (CP) step at the discharge cutoff potential was set current can be correlated to a “lithiation hindrance” i.e., a kinetically
for 24 h for the initial cycle. For further galvanostatic charge/discharge limited Li+ uptake of NCM, which according to our previous re-
cycling the CP step length was set to 6 h for the next two formation port induces the observed change in overpotential.20 The increase in
cycles and 2 h for the following cycles. specific discharge current leads to higher overpotentials and thus to
Lattice parameters in the XRD measurement were obtained by lower specific discharge capacities, resulting in decreased Coulombic
Rietveld refinement (Topas) using powder diffraction data from a D8 efficiencies.
Advance (Bruker) with Bragg-Brentano-Geometry and Ni-filtered Cu In its entirety, the lithiation (discharge) kinetics seem to be slower
Kα radiation. The corresponding Li concentration of the electrodes than the delithiation (charge) kinetics as the value of the specific
was determined by ICP-OES. discharge capacity (<200 mAh g−1 ) is lower than the value of the
specific charge capacity (= 220 mAh g−1 ). However, as seen in Figure
2, over a wide potential range the lithiation kinetics are good up to a
Results and Discussion
certain DOD threshold value. One can conclude, that the overpotential
For clarity reasons, the specific current dependent overpotential increase at the end of the lithiation process is in majority responsible
behavior of the delithiation (charge) and lithiation (discharge) pro- for the incomplete lithiation, thus leading to the specific capacity
cess is discussed individually. Here, we charge the NCM up to 4.6 V loss (non relithiation capacity loss). This phenomenon is only present
vs. Li/Li+ which is a substantially higher cutoff potential than the for cycles after a specific discharge current increase compared to
typically used 4.2 V vs. Li/Li+ .20,24 As depicted in Figure 1, the indi- the previous cycle, as it is the case for the investigated initial cycle,
vidual potential profiles (including the overpotentials) are dependent where formally the specific current is higher compared to the pristine
not only on the specific current, but also on the specific charge capac- state (zero specific discharge current). Keeping the specific discharge
ity (delithiation capacity; see introductory remarks) and thus on the current constant during galvanostatic charge/discharge cycling, the
state of charge (SOC). The highest overpotential is observed at the specific discharge capacity is practically equal to the specific charge
beginning of the delithiation process, which then steadily decreases capacity leading to a CE of almost 100% (independent of the applied
until a minimum is reached at a specific charge capacity of ≈140 mAh constant specific current). For the case that the specific discharge
g−1 . After this point, the difference in overpotential slightly increases current is even lower than the previous cycle specific discharge current,
again. Hence, the internal cell resistance during the initial charging the CE can even be higher than 100%.20
process apparently reaches its minimum at a specific charge capac- Obviously, a lower overpotential leads to a higher lithiation amount
ity of ≈140 mAh g−1 , which corresponds to a Li+ extraction ratio into the available vacant Li+ sites, which increases the specific dis-
of ≈50% when considering that the theoretical specific capacity of charge capacity (Figure 2). By analogy, a low overpotential during
NCM is 278 mAh g−1 .20 Since this trend is also visible at lower spe- charge also leads to a higher delithiation amount, which increases
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
Journal of The Electrochemical Society, 163 (14) A2943-A2950 (2016) A2945
Figure 3. Specific charge and discharge capacities in the first cycle as a func-
tion of the specific charge current for NCM/Li half cells in the potential range Figure 4. Detailed insight in the specific capacity losses in the first cycle of
of 3.0 V – 4.6 V vs. Li/Li+ at a fixed specific discharge current of 30 mA g−1. NCM/Li half cells as a function of the applied specific charge current. The
specific discharge current was fixed to 30 mA g−1 . Cells were cycled in the
potential range of 3.0 V – 4.6 V vs. Li/Li+ .
the specific charge capacity (Figure 1). The increase of vacant Li+
sites by a higher delithiation degree should in turn also lead to a set charge cutoff potential of 4.6 V vs. Li/Li+ . Electrolyte oxidation
subsequent higher lithiation degree and thus to a higher specific dis- would counteract the electrode potential increase during delithiation
charge capacity.21 Referring to Figure 1, the obtained specific charge leading to a higher specific charge capacity increase.20 Overall, only
capacities are strongly dependent on the applied specific currents. a marginal and thus insignificant part of electrolyte oxidation can be
The subsequent specific discharge capacities under fixed discharging expected in the specific charge capacity increase of 11.2 mAh g−1 .
conditions (30 mA g−1 ) are depicted in Figure 3 as a function of the Furthermore, the LiPF6 /carbonate based electrolytes have been proven
specific charge current. As expected, the lower the specific charge to be almost completely electrochemically stable at the upper cutoff
current, the higher is the specific charge capacity. In literature, a pro- potential of 4.6 V vs. Li/Li+ .20 The assumption of sufficient oxida-
portional correlation between the charge capacity and the subsequent tive stability of the electrolyte is also supported by other reports, that
discharge capacity was reported.21 As depicted in Figure 3, in our claimed an electrochemical stability up to 4.8 V vs. Li/Li+ ,28 or even
investigations this trend holds only true for specific charge currents up to 5.0 V vs. Li/Li+ .29–32 Consequently, it is reasonable to exclude
higher than ≈75 mA g−1 , thus at low specific charge capacities. How- any significant contributions from electrolyte oxidation to the specific
ever, at higher specific charge capacities, i.e., using lower specific charge capacity and to assign the specific charge capacity increase
charge currents, the corresponding specific discharge capacity does to an increase of the delithiation capacity; and the specific capacity
not follow the trend of the specific charge capacity. The obtained spe- loss to an incomplete re-lithiation of the vacant Li+ sites, as previ-
cific discharge capacity reaches a maximum and then even decreases ously reported.20 The specific capacity loss can be either induced by
at very low specific charge currents (1.5 mA g−1 ). It can be concluded kinetic limitations (i.e., being a reversible, recoverable capacity loss),
that lowering the specific charge current, resulting in an increased spe- such as a more resistive cathode electrolyte interphase (CEI) layer,
cific charge capacity, is only beneficial for a corresponding increase in or the formation of electrochemically inactive structural domains in
the subsequent specific discharge capacity down to a certain threshold the NCM host material at high lithium extraction amounts (visible as
value. non recoverable capacity loss indicating the irreversible part). This
As depicted in Figure 3, a maximum specific discharge capacity of inactive domains can also impede the lithiation reaction thus addi-
187.0 mAh g−1 observed for a specific charge current of 15.0 mA g−1 tionally increasing the reversible capacity loss and increase the total
(and a fixed specific discharge current of 30.0 mA g−1 ) corresponds to specific capacity loss.20 It is repeated here, that it is possible to dis-
a specific charge capacity of 226.0 mAh g−1 . At an one order of mag- tinguish between both types of specific capacity loss by introducing
nitude lower specific charge current of 1.5 mA g−1 , a lower specific a constant potential (CP) step during lithiation at the discharge cutoff
discharge capacity of 177.7 mAh g−1 is observed despite to a higher potential, which enables the recovery of the reversible part of the total
specific charge capacity of 237.2 mAh g−1 . Consequently, the addi- specific capacity loss. The residual, non recoverable specific capacity
tional specific charge capacity increase of 11.2 mAh g−1 (237.2–226.0 loss indicates the irreversible part.20
mAh g−1 ) is even lower than the total (reversible + irreversible) spe- The corresponding specific capacity losses are summarized in
cific capacity loss in the first cycle, which increases by 20.5 mAh g−1 Figure 4 as a function of the specific charge current. The aforemen-
(from 39 to 59.5 mAh g−1 , when the charge current is lowered from tioned increase of the total specific capacity loss of 20.4 mAh g−1
15.0 to 1.5 mA g−1 ). The obtained results cannot be simply explained when the specific charge current is decreased from 15.0 to 1.5 mA
by the common attribution of the total specific capacity loss to solely g−1 is mostly of irreversible nature. The additional specific charge
parasitic reactions, such as oxidative electrolyte decomposition, since capacity increase of 11.2 mAh g−1 is accompanied by an irreversible
the increase in specific charge capacity should be at least as high as specific capacity loss increase of 15.3 mAh g−1 . Hence, only 5.1 mAh
the increase in specific capacity loss following this reason.21,22,27 Fol- g−1 (20.4–15.3 mAh g−1 ) of the total specific capacity loss increase is
lowing common understanding, electrolyte oxidation increases solely reversible. The reversible specific capacity loss alone would not lead to
the charge capacity.20 The increase in specific charge capacity of 11.2 a specific discharge capacity decrease of 9.3 mAh g−1 (187.0−177.7
mAh g−1 at low specific currents is rather attributable to an increase of mAh g−1 ) displayed in Figure 3. Hence, a solely kinetic hindrance,
the delithiation capacity (= increase in lithium extraction ratio) than by e.g. the presence of a more resistive CEI layer, can be excluded
to electrolyte oxidation. One might conclude, that at low overpoten- as main reason for the specific discharge capacity decay in Figure
tials, as it is the case for low specific charge currents (Figure 1), a 3. It is therefore concluded that structural changes of the cathode at
higher delithiation capacity should be observed, since with decreas- high delithiation degrees lead to an increased total specific capacity
ing overpotential more delithiation must occur in order to achieve the loss. These findings are in agreement with our previous work, where
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
A2946 Journal of The Electrochemical Society, 163 (14) A2943-A2950 (2016)
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
Journal of The Electrochemical Society, 163 (14) A2943-A2950 (2016) A2947
Figure 6. Positive composite electrode of a LIB including the kinetic processes and Li+ pathways discussed in the text.
ionic conductivity of the electrolyte.22 Moreover, it has been reported conductivity displays a similar delithiation overpotential but a slightly
that even when only 1% (by weight) conductive carbon is used, the earlier begin of the potential slope resulting in a higher specific ca-
electronic conductivity is still above the ionic conductivity value of pacity loss and lower specific discharge capacity. This indicates that
the electrolyte.43 Additionally, the significance of the electronic bulk a too slow ion transport can indeed affect the electrochemical perfor-
conductivity can be disproved by comparing different layered oxide mance at higher specific currents, but only at higher DODs during the
active materials. When adding Cobalt to LiNiO2 yielding for example lithiation process. This characteristic range seems to be sensitive with
LiNi0.8 Co0.2 O2 , the electronic conductivity decreases by an order of respect to additional impedances induced by e.g. insufficient ionic
magnitude.45 Nevertheless, despite decreased electronic conductivity, transport or by other impeding processes within the cell, that depend
the overall electrochemical performance in terms of rate capability on the ion concentration in the electrolyte, as may be the ionic charge
even increases.12 This further reveals that performance changes at transfer from the electrolyte into the electrode.4,6 Overall, this leads to
increased current rates are not primarily related to electronic conduc- the conclusion that ionic transport within the commonly used 1.0 M
tivity limitations but rather to factors, which will be discussed in the LiPF6 /organic carbonate solvent based electrolytes cannot be made
following. responsible for the delithiation overpotential, but indeed can slightly
The two orders of magnitude lower ionic conductivity value of affect the lithiation kinetics at higher DOD in the case of small LiPF6
a common LiPF6 /organic carbonate based electrolyte may indicate a molarities.
more significant impact on the overpotential. However, when decreas- Both, the porosity and the inactive materials (binder, conductive
ing the ionic conductivity value by around a third via decreasing the carbon) within the positive composite electrode can kinetically affect
LiPF6 molarity from 1.0 M to 0.2 M,44 no apparent changes in the the Li+ transport. The influence of the porosity on the Li+ transport
potential profile are observable, when using a specific current of 30 has already been intensively reported by Zheng et al.22 Decreasing
mA g−1 (Figure 7). However, when increasing the specific current to the porosity via calendaring of the electrode has been found to be
150 mA g−1 , slight changes are becoming evident. Compared to the beneficial regarding mechanical strength and practical energy density.
1.0 M electrolyte solution, the 0.2 M electrolyte with the lower ionic However, an extreme porosity decrease can also deteriorate the elec-
trochemical performance. With respect to the initial charge/discharge
cycle, an apparent overpotential increase was observed merely for the
electrode with no porosity (0%). The discharge rate test confirmed a
significant specific discharge capacity dependence just for porosities
lower than 20%. It is worth noting that in order to observe any effects
the authors used a high active material mass loading of 30 mg cm−2 ,
which is higher than the mass loading of commercial electrodes.22
Indeed, electrodes with lower mass loadings reveal better rate perfor-
mance at elevated discharge currents.46,47 It is therefore assumed that
electrode porosity cannot significantly influence the Li+ ion transport
as long typical porosities and mass loadings are used.
The influence of the inactive components within a composite cath-
ode on the impedance was clarified by Liu et al.43 The authors revealed
that two factors can increase the impedance within a composite pos-
itive electrode: on the one hand, the ratio of conductive carbon to
binder and on the other hand the ratio of the inactive to active ma-
terial. The most beneficial mass compositional ratio for conductive
carbon (acetylene black) to binder (PVDF) was found to be higher
than 0.6:1. The advantage of this optimized compositional ratio is
related to its low impedance even at small overall inactive ingredient
Figure 7. 1st and 4th cycle of NCM/Li half cells using the specific currents of mass loading of below 5%. With commonly used positive electrodes,
30 (1st cycle) and 150 mA g−1 (4th cycle) in the potential range from 3.0 V – applying compositional ratios above 0.6:1 the measured impedances
4.6 V vs. Li/Li+ for a 0.2 and 1.0 M LiPF6 /organic carbonate solvent-based display the lowest values, especially at a sufficient inactive compo-
electrolyte. nent mass loading. The impedance increase associated with altering
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
A2948 Journal of The Electrochemical Society, 163 (14) A2943-A2950 (2016)
Figure 8. Left: Schematic showing the de-solvation process during lithiation (discharge) of layered NCM. The solvent characteristic Li+ solvation strength can
retard the lithiation process resulting in an overpotential (activation energy EA ) increase of this reaction (charge transfer). Right: 1st cycle potential profile of a
NCM/Li half cell containing 1.0 M LiPF6 in different electrolyte solvents. The varied solvation strengths do not significantly influence the lithiation potential,
which indicates sufficient de-solvation kinetics.
the compositional ratios does not affect the performance at moderate without any electrochemical treatment.34,62 Nevertheless, the reported
currents, which consequently disproves its significant influence on the associated impedances were insignificantly low compared to the total
observed overpotentials.48 cell impedance excluding such surface layer significance. Hence, the
With regard to charging of the graphite anode, the charge transfer reported surface layer changes are more likely attributable to intrin-
process, which includes the Li+ de-solvation from a carbonate based sic thermal or chemical (e.g., by electrode surface impurities such
solvent, was considered as a rate determining step.38,39 Even for the as moisture or hydroxide) electrolyte decomposition than to oxidative
positive electrode, some reports assign the Li+ de-solvation process a electrochemical decomposition.7,23,30,63,64 The fact that LiF is reported
rate determining step due to the observed large activation energy (EA ) to be the surface layer (or CEI layer) dominating species supports this
values.49–55 With respect to a full charge/discharge cycle of the NCM assumption, since LiF cannot be an oxidation product of LiPF6 .65–67
cathode, the Li+ de-solvation process can only influence the lithiation Furthermore, it was reported that there is no potential dependence for
process (discharge) as schematically depicted in Figure 8 (left). Since the resistance associated with the surface layer.7 On the one hand this
no de-solvation is required during delithiation, an influence on the neglects any electrochemical origin of the surface layer and on the
delithiation overpotential during charge can be excluded. As described other hand disconfirms the observed significant influence on the SOC
above, a significant lithiation hindrance is only observed beyond a dependent overpotential in Figure 1. Considering these thoughts, it is
certain DOD at the end of discharge (Figure 2), which has been unlikely to assign changes in electrochemical performance to changes
attributed to structurally reasoned lithiation hindrances.20 of the cathode surface layer (or CEI) caused by electrolyte oxida-
Rather fundamental studies focused on the Li+ de-solvation phe- tion, especially for the initial charge/discharge cycle, when using a
nomenon from a theoretical point of view.56,57 In one report the coor- common LiPF6 carbonate based electrolyte.
dination strength of LIB electrolyte solvents to Li+ was determined.58 The Li+ transport through solids may be assumed as the slowest
It was claimed that Li+ is equally coordinated in ethylene carbonate Li transport pathway. In NCM cathodes, the Li+ resides in the 3b
+
(EC) and γ-butyrolactone (GBL) solvents and experiences a stronger layer (Li+ layer) on the Li+ sites, octahedrally coordinated by oxygen
coordination in propylene carbonate (PC). A possible influence of the anions. Meanwhile, the transition metal ions reside in an 3a layer
aforementioned solvents, including fluoroethylene carbonate (FEC) (TM layer) on the TM sites, also octahedrally coordinated by oxygen
and vinylene carbonate (VC) as solid electrolyte interphase (SEI)59 anions.13 The Li+ movement during lithiation and delithiation occurs
forming electrolyte additives, on the overpotential was investigated within this two dimensional Li+ layer via octahedral and tetrahedral
in this work. The charge/discharge potential curve of the 1st cycle vacancy hopping.18,68 The strength of the Coulombic interactions and
for selected 1.0 M LiPF6 electrolytes is depicted in Figure 8 (right). respective coordination of Li+ by surrounded oxygen anions dictates
The lithiation overpotential is similar for each solvent, thus indicating the ease of the Li+ movement. The bigger the interlayer distance,
a rather insignificant impact of the Li+ de-solvation process on the the easier the movement.13 This Li+ transport property is affected
overpotential. by the oxidation state i.e. with the SOC. The SOC determines the
According to literature, a further performance influencing process crystallographic parameter and thus the interlayer distances for the
is associated with Li+ transport through the CEI.15,60,61 CEI formation Li+ layer, as well. Applying the HSAB concept, the oxidation of the
is attributed to electrochemically induced changes of a native surface corresponding TM ions during charging of NCM leads to harder TM
passivation layer7,30 via precipitation of oxidatively decomposed elec- ionic species, which in turn better attracts the oxygen ions in the a
trolyte components. However, such CEI formation is unlikely to oc- axis and simultaneously stronger repels the oxygen ions from each
cur for the here applied electrochemical conditions, because a major other in the c axis.12,69 The resulting steady increase of the c axis
electrochemical oxidation of the commonly used LiPF6 in carbon- lattice parameter during charge increases the responsible interlayer
ates based electrolytes was disproved by our previous work.20 Other distance for Li+ movement. Such SOC dependent c axis development,
reports claimed an electrochemical stability of LiPF6 in carbonate as measured by XRD, is depicted in Figure 9. It is shown that the c axis
based electrolytes even above 5.0 V vs. Li/Li+ .29–32 The fact that for lattice parameter, as indication for the interlayer distance and thus Li+
the high voltage spinel LiNi1/2 Mn3/2 O4 (LNMO) a rather low irre- mobility within NCM, is increasing up to a maximum as a delithiation
versible specific capacity loss was observed at the cutoff potential of degree of 0.5 is reached. After this threshold value, the c axis lattice
4.95 V vs. Li/Li+ supports the assumption of a rather electrochem- parameter decreases again. This might be attributed to the presence
ically stable LiPF6 carbonate solvent based electrolyte.20 Even if a of Ni2+ ions within the Li+ layer due to the NCM characteristic
CEI would be formed on LNMO according to the definition above, its Li/Ni mixing phenomenon.70 Beyond a certain electrode potential the
influence on the overpotential contribution in the initial cycle cannot divalent Ni2+ starts to irreversibly oxidize and become trivalent.71,72
be large due to the observed good electrochemical data and low po- The harder trivalent Ni species probably counteracts the c axis lattice
tential hysteresis in the initial cycle.20 Indeed, surface changes have parameter increase. These c axis lattice parameter developments are
been observed, but already after simple exposure to the electrolyte in agreement with previous reports for NCM based electrodes.12,73
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
Journal of The Electrochemical Society, 163 (14) A2943-A2950 (2016) A2949
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
A2950 Journal of The Electrochemical Society, 163 (14) A2943-A2950 (2016)
19. D. Andre, S. J. Kim, P. Lamp, S. F. Lux, F. Maglia, O. Paschos, and B. Stiaszny, J 47. K. Amine, J. Liu, I. Belharouak, S. H. Kang, I. Bloom, D. Vissers, and G. Henriksen,
Mater Chem A, 3, 6709 (2015). J Power Sources, 146, 111 (2005).
20. J. Kasnatscheew, M. Evertz, B. Streipert, R. Wagner, R. Klopsch, B. Vortmann, 48. G. Liu, H. Zheng, A. S. Simens, A. M. Minor, X. Song, and V. S. Battaglia, J Elec-
H. Hahn, S. Nowak, M. Amereller, A. C. Gentschev, P. Lamp, and M. Winter, Phys trochem Soc, 154, A1129 (2007).
Chem Chem Phys, 18, 3956 (2016). 49. T. Abe, H. Fukuda, Y. Iriyama, and Z. Ogumi, J Electrochem Soc, 151, A1120 (2004).
21. P. Gao, G. Yang, H. D. Liu, L. Wang, and H. S. Zhou, Solid State Ionics, 207, 50 50. Y. Yamada, Y. Iriyama, T. Abe, and Z. Ogumi, Langmuir, 25, 12766 (2009).
(2012). 51. T. Abe, F. Sagane, M. Ohtsuka, Y. Iriyama, and Z. Ogumi, J Electrochem Soc, 152,
22. H. H. Zheng, L. Tan, G. Liu, X. Y. Song, and V. S. Battaglia, J Power Sources, 208, A2151 (2005).
52 (2012). 52. Y. Yamada, F. Sagane, Y. Iriyama, T. Abe, and Z. Ogumi, J Phys Chem C, 113, 14528
23. Y. Park, S. H. Shin, H. Hwang, S. M. Lee, S. P. Kim, H. C. Choi, and Y. M. Jung, J (2009).
Mol Struct, 1069, 157 (2014). 53. T. Doi, Y. Iriyama, T. Abe, and Z. Ogumi, Anal Chem, 77, 1696 (2005).
24. D. R. Gallus, R. Schmitz, R. Wagner, B. Hoffmann, S. Nowak, I. Cekic-Laskovic, 54. I. Yamada, T. Abe, Y. Iriyama, and Z. Ogumi, Electrochem Commun, 5, 502 (2003).
R. W. Schmitz, and M. Winter, Electrochim Acta, 134, 393 (2014). 55. I. Yamada, Y. Iriyama, T. Abe, and Z. Ogumi, J Power Sources, 172, 933 (2007).
25. J. Nanda, J. Remillard, A. O’Neill, D. Bernardi, T. Ro, K. E. Nietering, J.-Y. Go, and 56. S. Yanase and T. Oi, J. Nucl. Sci. Technol., 39, 1060 (2002).
T. J. Miller, Adv Funct Mater, 21, 3282 (2011). 57. T. Fukushima, Y. Matsuda, H. Hashimoto, and R. Arakawa, Electrochem Solid St, 4,
26. J. Liu, M. Kunz, K. Chen, N. Tamura, and T. J. Richardson, The Journal of Physical A127 (2001).
Chemistry Letters, 1, 2120 (2010). 58. Y. Matsuda, T. Fukushima, H. Hashimoto, and R. Arakawa, J Electrochem Soc, 149,
27. H. Zheng, Q. Sun, G. Liu, X. Song, and V. S. Battaglia, J Power Sources, 207, 134 A1045 (2002).
(2012). 59. M. Winter, Z Phys Chem, 223, 1395 (2009).
28. S. C. Yin, Y. H. Rho, I. Swainson, and L. F. Nazar, Chem Mater, 18, 1901 (2006). 60. X. Liu, P. He, H. Li, M. Ishida, and H. Zhou, J Alloy Compd, 552, 76 (2013).
29. J. M. Tarascon and D. Guyomard, Solid State Ionics, 69, 293 (1994). 61. R. Holze, L. J. Fu, H. Liu, C. Li, Y. P. Wu, E. Rahm, and H. Q. Wu, Solid State Sci,
30. K. Xu, Chem Rev, 104, 4303 (2004). 8, 113 (2006).
31. D. Guyomard and J. M. Tarascon, Solid State Ionics, 69, 222 (1994). 62. Y. Matsuo, R. Kostecki, and F. McLarnon, J Electrochem Soc, 148, A687 (2001).
32. D. Guyomard and J. M. Tarascon, J Electrochem Soc, 140, 3071 (1993). 63. Y. Park, S. H. Shin, S. M. Lee, S. P. Kim, H. C. Choi, and Y. M. Jung, J Mol Struct,
33. A. Manthiram, J. Choi, and W. Choi, Solid State Ionics, 177, 2629 (2006). 1069, 183 (2014).
34. F. Lin, I. M. Markus, D. Nordlund, T. C. Weng, M. D. Asta, H. L. L. Xin, and 64. M. Balasubramanian, H. S. Lee, X. Sun, X. Q. Yang, A. R. Moodenbaugh,
M. M. Doeff, Nat. Commun., 5 (2014). J. McBreen, D. A. Fischer, and Z. Fu, Electrochem Solid St, 5, A22 (2002).
35. X. Qi, B. Blizanac, A. DuPasquier, M. Oljaca, J. Li, and M. Winter, Carbon, 64, 334 65. T. Eriksson, A. M. Andersson, A. G. Bishop, C. Gejke, T. Gustafsson, and
(2013). J. O. Thomas, J Electrochem Soc, 149, A69 (2002).
36. S. F. Lux, F. Schappacher, A. Balducci, S. Passerini, and M. Winter, J Electrochem 66. A. M. Andersson, D. P. Abraham, R. Haasch, S. MacLaren, J. Liu, and K. Amine, J
Soc, 157, A320 (2010). Electrochem Soc, 149, A1358 (2002).
37. E. Kramer, T. Schedlbauer, B. Hoffmann, L. Terborg, S. Nowak, H. J. Gores, 67. D. Aurbach, K. Gamolsky, B. Markovsky, G. Salitra, Y. Gofer, U. Heider, R. Oesten,
S. Passerini, and M. Winter, J Electrochem Soc, 160, A356 (2013). and M. Schmidt, J Electrochem Soc, 147, 1322 (2000).
38. K. Xu, A. von Cresce, and U. Lee, Langmuir, 26, 11538 (2010). 68. Y. Wei, J. X. Zheng, S. H. Cui, X. H. Song, Y. T. Su, W. J. Deng, Z. Z. Wu,
39. K. Xu and A. von Cresce, J Mater Chem, 21, 9849 (2011). X. W. Wang, W. D. Wang, M. M. Rao, Y. Lin, C. M. Wang, K. Amine, and F. Pan, J
40. N. Imanishi, M. Fujiyoshi, Y. Takeda, O. Yamamoto, and M. Tabuchi, Solid State Am Chem Soc, 137, 8364 (2015).
Ionics, 118, 121 (1999). 69. J. K. Ngala, N. A. Chernova, M. M. Ma, M. Mamak, P. Y. Zavalij, and
41. S. Levasseur, M. Menetrier, E. Suard, and C. Delmas, Solid State Ionics, 128, 11 M. S. Whittingham, J Mater Chem, 14, 214 (2004).
(2000). 70. J. M. Kim and H. T. Chung, Electrochim Acta, 49, 937 (2004).
42. F. Croce, F. Nobili, A. Deptula, W. Lada, R. Tossici, A. D’Epifanio, B. Scrosati, and 71. C. Pouillerie, L. Croguennec, P. Biensan, P. Willmann, and C. Delmas, J Electrochem
R. Marassi, Electrochem Commun, 1, 605 (1999). Soc, 147, 2061 (2000).
43. G. Liu, H. Zheng, S. Kim, Y. Deng, A. M. Minor, X. Song, and V. S. Battaglia, J 72. T. Kawaguchi, K. Fukuda, K. Tokuda, M. Sakaida, T. Ichitsubo, M. Oishi, J. i. Mizuki,
Electrochem Soc, 155, A887 (2008). and E. Matsubara, Phys Chem Chem Phys, 17, 14064 (2015).
44. A. Nyman, M. Behm, and G. Lindbergh, Electrochim Acta, 53, 6356 (2008). 73. O. Dolotko, A. Senyshyn, M. J. Muhlbauer, K. Nikolowski, and H. Ehrenberg, J
45. I. Saadoune and C. Delmas, J Mater Chem, 6, 193 (1996). Power Sources, 255, 197 (2014).
46. X. J. Liu, G. Y. Zhu, K. Yang, and J. Q. Wang, J Power Sources, 174, 1126 74. J. B. Goodenough and K. Mizuchima; Pat., U. S., Ed. United States, 1981; Vol.
(2007). 4,302,518.
Downloaded on 2017-10-04 to IP 128.176.90.84 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).