Improving the performance at elevated temperature of high voltage graphite/LiNi0.5Mn1.5O4 cells with added lithium catechol dimethyl borate

Article abstract of DOI:10.1149/2.0331702jes

Performance of LiNi_0.5Mn_1.5O₄/graphite cells cycled to 4.8 V at 55^?C with the 1.2 M LiPF₆ in EC/EMC (3/7, STD electrolyte) with and without added lithium catechol dimethyl borate (LiCDMB) has been investigated. The incorporation of 0.5 wt% LiCDMB to the STD electrolyte results in an improved capacity retention and coulombic efficiency upon cycling at 55^?C. Ex-situ analysis of the electrode surfaces via a combination of SEM, TEM, and XPS reveals that oxidation of LiCDMB at high potential results in the deposition of a passivation layer on the electrode surface, preventing transition metal ion dissolution from the cathode and subsequent deposition on the anode. NMR investigations of the bulk electrolyte stored at 85^?C reveals that added LiCDMB prevents the thermal decomposition of LiPF₆.

Full text of DOI:10.1149/2.0331702jes

A128
Journal of The Electrochemical Society, 164 (2) A128-A136 (2017)
Improving the Performance at Elevated Temperature of High
Voltage Graphite/LiNi0.5Mn1.5O4 Cells with Added Lithium
Catechol Dimethyl Borate
Yingnan Dong, Julien Demeaux, Yuzi Zhang, Mengqing Xu,∗ Liu Zhou, Alex D. MacIntosh,
and Brett L. Lucht∗∗,z
Department of Chemistry, University of Rhode Island, Kingston, Rhode Island 02881, USA
Performance of LiNi0.5Mn1.5O4/graphite cells cycled to 4.8 V at 55◦C with the 1.2 M LiPF6 in EC/EMC (3/7, STD electrolyte)
with and without added lithium catechol dimethyl borate (LiCDMB) has been investigated. The incorporation of 0.5 wt% LiCDMB
to the STD electrolyte results in an improved capacity retention and coulombic efficiency upon cycling at 55◦C. Ex-situ analysis
of the electrode surfaces via a combination of SEM, TEM, and XPS reveals that oxidation of LiCDMB at high potential results
in the deposition of a passivation layer on the electrode surface, preventing transition metal ion dissolution from the cathode and
subsequent deposition on the anode. NMR investigations of the bulk electrolyte stored at 85◦C reveals that added LiCDMB prevents
the thermal decomposition of LiPF6.
© The Author(s) 2016. Published by ECS. This is an open access article distributed under the terms of the Creative Commons
Attribution Non-Commercial No Derivatives 4.0 License (CC BY-NC-ND, http://creativecommons.org/licenses/by-nc-nd/4.0/),
which permits non-commercial reuse, distribution, and reproduction in any medium, provided the original work is not changed in any
way and is properly cited. For permission for commercial reuse, please email: oa@electrochem.org. [DOI: 10.1149/2.0331702jes]
All rights reserved.
Manuscript submitted October 21, 2016; revised manuscript received November 28, 2016. Published December 13, 2016. This was
Paper 539 presented at the Honolulu, Hawaii, Meeting of the Society, October 2–7, 2016.
Lithium-ion batteries are widely used for portable electronics and
ion battery electrolyte improves the electrochemical performance
of LiNi0.5Mn1.5O4/Graphite cells cycle to high potential (4.8 V vs.
LiC6/C6) (Figure 1). Ex-situ surface analysis of the cycled electrodes
was conducted to better understand the source of performance en-
hancement.
are currently being incorporated into electric vehicles due to their
high energy density.1,2 However, there is significant interest in further
increasing the energy density of lithium-ion batteries.3 One method
to achieve higher energy density is increasing the operating potential
of the cathode material. Most commercial lithium-ion batteries con-
tain a lithiated transition metal oxide cathode that typically operates
at ∼4.0 V vs. Li/Li+.3,4 Several novel cathode materials with operat-
ing potentials over 4.0 V are currently under investigation, including
LiNiPO4,5 LiCoPO4,5,6 and LiNi0.5Mn1.5O4. While the high operat-
ing potential of the LiNi0.5Mn1.5O4 spinel cathode (4.8 V vs. Li/Li+)
offers high energy density, commercialization has been hampered by
severe capacity fade and poor efficiency.7 The capacity fade is par-
ticularly pronounced at moderately elevated temperatures (>45◦C)
and in full cells employing a graphite anode.7 The failure mecha-
nisms of LiNi0.5Mn1.5O4 cells at high voltage and elevated temper-
ature have been recently investigated.8–14 Electrolyte decomposition,
electrode/electrolyte interface degradation, and transition metal dis-
solution are the leading factors reported for performance fade. One
effective method for improving the performance of high voltage cath-
odes involves the incorporation of SEI (solid electrolyte interface)
and CEI (cathode electrolyte interface) forming electrolyte additives
that are sacrificially oxidized on the surface of electrodes to generate a
passivation film which inhibits transition metal dissolution and further
electrolyte oxidation.
There have been several reports where electrolyte additives have
improved the performance of cathodes cycled to high potential.15–18,23
Lithium bis(oxalato borate) (LiBOB) has been reported to be one
of the better additives for LiNi0.5Mn1.5O4 cathodes.16,17,20–25 The re-
lated additive, lithium difluorooxalato borate (LiDFOB)27–31 has also
been reported to improve the properties of Li1.2Ni0.15Mn0.55Co0.1O2
cathodes cycled to high potential19 In addition to the lithium oxalato
borates,26,32 we have recently reported on the beneficial effect of the
incorporation of lithium tetralkylborates as Additives for Designed
Surface Modification (ADSM) to function as functional group deliv-
ery agents to modify the cathode surface.8,9
Experimental
Synthesis and characterization.—2.75 g of catechol (99%,
Aldrich) was dissolved in 100 mL of diethyl ether (anhydrous, ≥99%,
¯
Aldrich). 30 mL of n-butyl lithium (1.6 M in hexanes, ACROS) was
added to the solution drop by drop. The reaction mixture was stirred
for 24 hours the product precipitated and collected via filtration in
a N2-filled glove-box. The salt obtained was subsequently washed
with a small amount of diethyl ether (anhydrous, ≥99%, Aldrich) and
stored under vacuum overnight to remove solvent. The product was
characterized as lithium catechol by 1H NMR. 1.25 g of lithium cate-
chol was suspended in 50 mL of ether, and 2.5 mL of trimethyl borate
was added to the suspension. A slurry mixture was obtained. The so-
lution was stirred for 24 hours and the mixture composed of lithium
tetramethyl borate and lithium catechol dimethyl borate was filtered
in the N2-filled glove box. The mixture was separated by adding ex-
cess dimethyl carbonate to dissolve lithium catechol dimethyl borate
(lithium tetramethyl borate has poor solubility in DMC). The filtrate
was collected and the DMC was removed under vacuum, lithium
catechol dimethyl borate (LiCDMB, Scheme 1) was obtained as a
In this manuscript, a new asymmetric lithium borate, lithium cat-
echol dimethyl borate (LiCDMB) is synthetized via a simple two-
step reaction. Incorporation of LiCDMB into a standard lithium
−
∗
Figure 1. Chemical structure of lithium salt anions: (a) B(O(CO2)2O)2
Electrochemical Society Student Member.
Electrochemical Society Member.
(BOB−), (b) BF2(O(CO2)2O)− (DFOB−), and (c) a new non-fluorinated anion
B(OMe)2(O(C6H4)O)− (CDMB−).
∗∗
zE-mail: blucht@chm.uri.edu
Downloaded on 2016-12-13 to IP 80.82.78.170 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, 164 (2) A128-A136 (2017)
A129
Scheme 1. Synthetic route for lithium catechol dimethyl borate.
1
white crystal. The product was characterized by H and 11B nuclear
magnetic resonance (NMR) spectroscopy.
croscopy (SEM, JEOL5900). The cycled electrodes were exposed to
ultrasound in DMC solvent for 3 h to allow homogenous dispersion
of the active materials in the solution, and then the dispersed solution
was cast on a copper TEM grid (500 mesh) and dried overnight in a
vacuum oven. The TEM grids were quickly transferred into the TEM
chamber. Imaging was conducted using a JEOL JEM-2100F TEM
(Pebody, MA) at 160 eV. The diameter of the beam was 5 nm, and
low-dose imaging was employed to minimize electron-beam-induced
changes to organic components in the surface layer.
Electrochemical test and characterization.—Battery grade car-
bonate solvents, lithium hexafluorophosphate (LiPF6) and 1.2 M
LiPF6 in EC/EMC (3/7 v/v) (STD electrolyte) were obtained from
a commercial source. The additive was added as weight percent of the
total mass of electrolyte.
Lithium catechol dimethyl borate (LiCDMB) was synthesized and
added as 0.5% (wt) to the STD electrolyte. The composite cathode and
anode electrodes were obtained from a commercial source. The com-
posite LiNi0.5Mn1.5O4 electrode is composed of active material (92%),
conductive carbon (4%) and PVDF binder (4%). The composite an-
ode electrode is composed of graphite (ConocoPhilips, 95.7%) along
with conductive carbon (0.5%) and CMC & SBR binder (3.8%). The
cathode loading is 15.9 mg/cm2 and loading of anode is 7.1 mg/cm2.
2032-type coin cells were built with cathode (d = 14.7 mm) and
graphite anode (d = 15.0 mm), a Setela E20MM (d = 19 mm) separa-
tor, and 40 μL of electrolyte in each cell in an Argon-filled glove box
with a water content less than 0.1 ppm. Carbon black electrode (Super
C65, d = 15 mm) half-cells were built with 100 μL of electrolyte, a
glass fiber (Whatman, d = 15.6 mm) separator and a Setela E20MM
(d = 19 mm) separator. Cells were cycled on an Arbin Instruments
batter cycler and the temperature was controlled with Fisher Scientific
Isotemp Incubators.
Thermal stability.—Samples for NMR spectroscopy were pre-
pared in a glove box filled with high purity Ar followed by flame
sealing in glass NMR tubes under reduced pressure. Sealed samples
were heated in a silicon oil bath at 85◦C. Samples were weighed be-
fore and after storage to confirm seal. NMR analyses were conducted
on a Bruker 300 MHz NMR spectrometer. 19F NMR spectra were ref-
erenced to LiPF6 at −65.0 ppm and 31P NMR spectra were referenced
to LiPF6 at −145.0 ppm, as described previously.33–35
Results and Discussion
Characterization of lithium catechol dimethyl borate
(LiCDMB).—The as-synthesized product is purified via crys-
tallizations, and characterized by NMR spectroscopy in D2O (1H,
11B). The corresponding 1H and 11B NMR spectra are depicted
in Figure 2. The singlet peak at 3.3 ppm is characteristic of the
methoxyl group (-OCH3) of the product, a small peak characteristic
of residual wash solvent (DMC) can be observed at 3.8 ppm, and
the peak ascribed to the residual H in D2O is observed at 4.8 ppm.
The singlet peak observed at 6.7 ppm is attributed to the aromatic
protons of the product. The integrated ratio of methyl protons to
aromatic protons is 4:6, which matches the structure of LiCDMB.
A single peak characteristic of the product at 7.7 ppm is observed
in the 11B NMR spectrum, 11B chemical shift of boric acid located
between −20-0ppm as a function of pH,51 suggesting LiCDMB
LiNi0.5Mn1.5O4/graphite cells were cycled at 25◦C initially with
the following cycling protocol: C/20 for the first cycle; C/10 for the
second and third cycles; and then C/5 for the remaining cycles at 25◦C.
After cycling at 25◦C for a total of 20 cycles, cells were transferred
to 55◦C and C/5 cycling was continued for an additional 30 cycles.
Cells were charged with a CC-CV mode, constant current charge to
4.8 V followed with a constant voltage charge step at 4.8 V vs. LiC6/C6
until the current decreased to 10% of the applied charging current. The
cells were discharged to 4.25 V vs. LiC6/C6 at same constant current
◦
(CC mode). Coin cells were sealed with epoxy resin prior to 55 C
cycling and there was no evidence for cell leakage after cycling at
55◦C. Cells were built in triplicate. Cell to cell variation was less than
3%. Electrochemical impedance spectroscopy (EIS) was performed
on a Bio-Logic Instrument after formation, 25 ◦C and 55 ◦C cycling at
100% SOC. The perturbation is 10 mV with the frequency range 1000
kHz–20 mHz. Cycled cells were disassembled in an argon glove-box,
and cycled anodes/cathodes were harvested and rinsed with anhydrous
dimethyl carbonate (DMC, Sigma, extra dry 99%) 3 times to remove
residual electrolyte, followed by vacuum drying overnight at room
temperature.
1
doesn’t decompose into boric acid in D2O. The H and 11B NMR
spectra support the isolation of a pure compound.
Ex-situ surface analysis of the discharged electrodes was con-
ducted. XPS measurements were carried out using a ThermoFisher
K-Alpha spectrometer, under focused monochromatized Al Kα radi-
ation (ν = 1486.6 eV). Cells were disassembled in the glove box and
electrode samples were rinsed 3 times with DMC and dried under vac-
uum at room temperature for 10 minutes. Samples were then sealed in
a vial under controlled atmosphere of the glove box and stored for 24
hours. A transfer case (ThermoFisher) was used to avoid any contact
with air/moisture. Peaks were recorded with constant pass energy of
50 eV with an energy resolution of 50 meV and charge neutralization.
Peak positions and areas were optimized by a weighted least squares
fitting method using 70% Gaussian, 30% Lorentzian line shapes using
the Avantage (ThermoFisher) software.
The discharged electrodes were briefly (15 s) exposed to air during
transfer to the SEM and TEM vacuum chamber. Surface morphology
of the cycled electrodes was characterized by scanning electron mi-
Figure 2. (Top) 1H and (bottom) 11B NMR spectra of lithium catechol
dimethyl borate (LiCDMB) in D2O.
Downloaded on 2016-12-13 to IP 80.82.78.170 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A130
Journal of The Electrochemical Society, 164 (2) A128-A136 (2017)
Figure 3. (a) 19F (a) and (b) 31P NMR spectra of electrolytes before storage and after a 8-day storage at 85◦C: (in black) the STD and (in red) the STD + 0.5%
LiCDMB electrolytic solutions.
Thermal stability.—The 19F and 31P NMR spectra of the STD elec-
Cathodic linear sweep voltammetry of Super C65/Li cells is pre-
sented in Figure 5. For the STD electrolyte, the reduction peak of EC
at 0.65 V vs. Li/Li+ can be clearly observed.37 For the electrolyte
containing LiCDMB, the reduction peak for EC is observed at similar
trolyte and STD with 0.5% added LiCDMB before and after storage
for 8-days at 85◦C are presented in Figure 3. The spectra are consistent
with the addition of LiCDMB inhibiting the thermal decomposition
of the LiPF6 electrolyte. In both cases, the fresh electrolytes contain
a single set of peaks in the 19F and 31P NMR spectra characteristic
of LiPF6.35 The 19F spectrum of the STD electrolyte after 8 days of
storage at 85◦C reveals new peaks around −85 ppm characteristic of
fluorophosphates (OPFx(OR)y), in addition a small peak for LiF is
observed at – 155 ppm. After 8 days of storage at 85◦C the 19F NMR
spectrum of the electrolyte with added LiCDMB has a much lower
concentration of peaks characteristic of OPFx(OR)y, consistent with
an inhibition of LiPF6 decomposition (Figure 3a). The 31P spectra
further support the inhibition of electrolyte decomposition with added
LiCDMB. The 31P spectrum of the STD electrolyte stored at 85◦C
for 8 days contains new peaks around −25 ppm characteristic of O =
PFx(OR)y. While the same peaks are present in the samples containing
added LiCDMB, the intensity of the peaks is significantly diminished
consistent with inhibition of the thermal decomposition of LiPF6.36,37
Electrochemical stability.—Electrochemical stability of both the
STD and the LiCDMB electrolyte have been evaluated on carbon
black electrodes with linear sweep voltammetry at high and low
potential.10,36
Anodic linear sweep voltammetry of Super C65/Li cells is pre-
sented in Figure 4. Additive oxidation is clearly observed above
3.5 V vs. Li/Li+ as increased current, compared to the STD elec-
trolyte. Additional oxidation peaks are observed at 4.0 V vs. Li/Li+
and 4.4 V vs. Li/Li+ for the borate-containing electrolyte. Increased
current is observed for the LiCDMB electrolyte up to 5.6 V vs. Li/Li+.
Figure 4. Anodic linear sweep voltammetry at 25◦C of Super C65/Li cells
(sweep rate of 0.1 mV.s−1) using (in black) the EC/EMC (3/7) 1.2 M LiPF6
and (in red) EC/EMC (3/7) 1.2 M LiPF6 + 0.5% LiCDMB electrolytes.
Downloaded on 2016-12-13 to IP 80.82.78.170 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, 164 (2) A128-A136 (2017)
A131
Figure 5. Cathodic linear sweep voltammetry at 25◦C of Super C65/Li cells
(sweep rate of 0.1 mV.s−1) using (in black) the EC/EMC (3/7) 1.2 M LiPF6
and (in red) EC/EMC (3/7) 1.2 M LiPF6 + 0.5% LiCDMB electrolytes.
potential and intensity. This suggests that the presence of the LiCDMB
additive does not affect EC reduction at low potential. There is no ev-
idence for reduction of the additive in the 3.0 V–0.7 V vs. Li/Li+
potential range.
Investigation of the electrochemical stability at both high and low
potentials of the LiCDMB electrolyte suggests reactivity at high po-
tential. Additive oxidation is observed above 3.5 V vs. Li/Li+, before
the first redox couple of the high voltage spinel situated at 4.3 V vs.
Li/Li+ (Mn+III/Mn+IV). This reactivity may result in the generation
of a cathode passivation layer which inhibits Mn2+ dissolution and
further electrolyte oxidation especially for cells cycled at 55◦C.
Cycling performance.—Cycling performance at 25◦C and 55◦C of
LiNi0.5Mn1.5O4/Graphite cells using the STD, LiCDMB and LiBOB
electrolytes is presented in Figure 6. Additive concentration of 0.5%
(wt) of LiCDMB is found to be optimal for improved performance of
high voltage cells. As seen from Figure 6a, the cell with the borate
additive shows better capacity retention than the standard cell after
30 cycles at 55◦C. After 30 cycles at 55◦C, the cell with the STD
electrolyte retains only 52% of its original capacity while addition
of 2.5% LiBOB improves the capacity retention to 61%, as previ-
ously reported.17,23 However, a retention of 76% of the original capac-
ity is observed with added LiCDMB. Cells containing the LiCDMB
electrolyte have the best performance suggesting that LiCDMB is a
promising additive for LiNi0.5Mn1.5O4/Graphite cells.
Coulombic efficiencies of the cells with the STD, LiCDMB and
LiBOB electrolytes are presented in Figure 6b. The first cycle effi-
ciency at C/20 is higher for the cell containing the STD electrolyte
than the cell containing the LiCDMB electrolyte. The difference may
be due to additive oxidation at high potential (Figure 4). However, af-
ter formation cycling the efficiency of cells containing the LiCDMB
electrolyte is better than the cells containing the STD electrolyte, sup-
porting a beneficial effect of the borate additive upon cycling. Upon
cycling at 55◦C, the differences in coulombic efficiency are enhanced
and the cell containing the LiCDMB electrolyte has ∼3% higher ef-
ficiency than the cell containing the STD electrolyte. Despite a small
decrease in the first cycle discharge capacity and efficiency, the long
term cycling performance at 25◦C and 55◦C is significantly improved
with the LiCDMB electrolyte. While the cell with 2.5% of added Li-
BOB had better capacity retention than the cell containing the STD
electrolyte, no improvement to the coulombic efficiency is observed.
Figure 6. (a) Cycling retention and (b) coulombic efficiency of
LiNi0.5Mn1.5O4/Graphite cells (cutoff potentials at 25◦C and 55◦C: 4.80 V–
4.25 V vs. LiC6/C6) using the STD electrolyte (in black), STD with 0.5% (wt)
added LiCDMB (in red) and STD with 2.5% (wt) added LiBOB (in blue).
4.8 V) are measured at different stages upon cycling. The corre-
sponding EIS Nyquist plots are depicted in Figure 7. After formation
cycling, the impedance of the cell with STD electrolyte is found to be
similar to the cell containing the LiCDMB electrolyte (Figure 7a).38
After 20 cycles at 25◦C, the EIS of the cell with the STD electrolyte
and the cell with the LiCDMB electrolyte remain very similar (Figure
7b), consistent with similar specific capacity (Figure 6). However, a
significant change in EIS is observed upon cycling at 55◦C (Figure 7c).
Impedance of the cell with the STD electrolyte is almost twice as large
as the cell with the LiCDMB electrolyte. While additional electrolyte
oxidation is observed for cells containing LiCDMB electrolyte during
formation cycling, the presence of the oxidation products of LiCDMB
result in better capacity retention, efficiency, and lower impedance.
SEM/TEM imaging of electrodes.—SEM imaging of the
LiNi0.5Mn1.5O4 and graphite electrodes after cycling at 55◦C has been
conducted in order to investigate electrode surface morphology. SEM
micrographs of the fresh LiNi0.5Mn1.5O4 cathodes and cathodes cy-
cled with and without added LiCDMB are depicted in Figure 8. The
fresh electrode consists of secondary spherical particles of ca. 8 μm.
These spherical particles are composed of a primary rod structure sev-
eral hundred nanometers in length. The fresh cathode particle surface
is clean and smooth. After cycling with the STD electrolyte at 55◦C,
Electrochemical impedance spectroscopy.—Electrochemical
impedance spectra of cells at a full state of charge (100% SOC,
Downloaded on 2016-12-13 to IP 80.82.78.170 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A132
Journal of The Electrochemical Society, 164 (2) A128-A136 (2017)
Figure 9. SEM micrographs of the graphite electrode and graphite anodes
harvested from cells after overall cycling at 25◦C and 55◦C using the STD and
STD + 0.5% LiCDMB electrolytes.
Figure 7. EIS spectra at full state of charge (4.8 V vs. LiC6/C6) of
Li1-xNi0.5Mn1.5O4/LixC6 cells using the (in black) STD and (in red) STD
+ 0.5% LiCDMB electrolyte after (a) formation cycling, (b) 20 cycles at
25◦C, and (c) 30 additional cycles at 55◦C.
the structure of the cathode particles is severely damaged. The particle
damage likely results in increased impedance and reduced capacity
retention of cells, as discussed above. For the cells cycled with the
LiCDMB electrolyte, the original secondary and primary structure of
the LiNi0.5Mn1.5O4 particles is maintained after cycling at 55◦C. The
surface of the primary particle is covered by a thick CEI. Maintenance
of the original structure of the cathode and formation of a passivating
CEI contributes to enhanced cycling performance of the cells at 55◦C.
SEM micrographs of the fresh graphite electrode and graphite an-
odes cycled with or without added LiCDMB are depicted in Figure
9. The surfaces of the fresh graphite electrode particles have sharp
edges, and are primarily composed of flake-like structures. After cy-
cling with the STD electrolyte, the graphite surface becomes very
rough and non-uniform with many small particles due to the depo-
sition of electrolyte decomposition products on the surface of the
graphite. Upon cycling with the LiCDMB electrolyte, a more uniform
smooth layer is observed on the graphite surface.
Electrodes extracted from graphite/LiNi0.5Mn1.5O4 cells cycled
at 55◦C have been analyzed by TEM (Figure 10). Both fresh
LiNi0.5Mn1.5O4 and graphite electrodes have sharp edges. After cy-
cling with the STD electrolyte, inhomogeneous coverage of the cath-
ode surface is observed. Alternatively, the surface of the cathode
cycled with the LiCDMB electrolyte has a more uniform surface film
(Figure 4). TEM images of the graphite electrode cycled with the STD
electrolyte reveal significant concentrations of electrolyte decompo-
sition products. The SEI is grainy and uneven, as observed by SEM
above. The graphite anode cycled with the LiCDMB electrolyte has a
thinner but more continuous surface layer. Thus, TEM images are in
agreement with SEM images: addition of LiCDMB results in a thicker
CEI on the high voltage cathode, and a thinner but more uniform SEI
on the graphite anode.
XPS.—Relative atomic concentrations.—In order to correlate
cycling performance of graphite/LiNi0.5Mn1.5O4 cells with surface
chemistry of the electrodes, XPS has been conducted on both
LiNi0.5Mn1.5O4 and graphite electrodes fresh and after cycling.
Figure 8. SEM micrographs of the LiNi0.5Mn1.5O4 electrode and
LiNi0.5Mn1.5O4 cathodes harvested from cells after overall cycling at 25◦C
and 55◦C using the STD and STD + 0.5% LiCDMB electrolytes.
Downloaded on 2016-12-13 to IP 80.82.78.170 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, 164 (2) A128-A136 (2017)
A133
Figure 10. TEM images of LiNi0.5Mn1.5O4 and graphite electrodes harvested from cells after 50 cycles at 25◦C and 45◦C using the STD and STD + 0.5%
LiCDMB electrolytes.
Relative atomic concentrations of elements detected on the
LiNi0.5Mn1.5O4 cathodes fresh and cycled at 55◦C in both the STD
and LiCDMB electrolytes are presented in Figure 11. The cathode
cycled with the STD electrolyte has a reduced concentration of C
1s and inorganic O 1s (metal oxide) due to the deposition of organic
components of the CEI. Increased concentration of organic O 1s (C-O,
C=O) confirms the generation of a new cathode-electrolyte interface.
Relative atomic concentrations of elements detected on the cathode
cycled with the LiCDMB electrolyte indicates that the amount of or-
ganic species deposited on the surface is further increased compared
to the cathode cycled with the STD electrolyte. A significant increase
of organic O 1s along with decrease of the Mn 2p and inorganic O
1s suggest significant coverage of the electrode surface. Therefore,
addition of the LiCDMB additive to the STD electrolyte allows the
deposition of a thicker passivating CEI, likely due to the oxidation of
LiCDMB at high potentials (Figure 4).
Relative atomic concentrations of elements detected on the
graphite anode cycled at 55◦C with both the STD and LCDMB elec-
trolytes are compared in Figure 12 to the fresh graphite electrode.
The graphite electrode cycled with the STD electrolyte has a decrease
in the concentration of C, and increases in the concentrations of O,
F, P, and Mn consistent with the generation of an SEI. A high con-
centration of Mn is detected on the graphite electrode cycled with
the STD electrolyte, consistent with manganese dissolution from the
Figure 11. Relative atomic concentrations of elements present on the
LiNi0.5Mn1.5O4 electrode and LiNi0.5Mn1.5O4 cathodes after overall cycling
at 25◦C and 55◦C in the STD and STD + 0.5% LiCDMB electrolytes.
Figure 12. Relative atomic concentrations of elements present on the graphite
electrode and graphite anodes after overall cycling at 25◦C and 55◦C in the
STD and STD + 0.5% LiCDMB electrolytes.
Downloaded on 2016-12-13 to IP 80.82.78.170 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A134
Journal of The Electrochemical Society, 164 (2) A128-A136 (2017)
Figure 13. (On the left) O 1s and (on the right) C 1s core spectra of (a) the fresh LiNi0.5Mn1.5O4 electrode and LiNi0.5Mn1.5O4 cathodes cycled at 55◦C with the
(b) STD and (c) STD + 0.5% LiCDMB electrolytes.
LiNi0.5Mn1.5O4 during cycling at 55◦C and deposition at low potential
on the anode. The graphite anode cycled with the LiCDMB electrolyte
shows increased concentrations of O 1s, and F 1s (LiF + LixPOyFz)
due to the generation of the SEI. The decreased concentration of Mn
2p is significant, consistent with the LiCDMB electrolyte inhibiting
Mn dissolution from the high voltage spinel (Figure 11).
285.6 eV (C-H), 286.5 eV (C-O), ca. 288 eV (C=O), and 289.9 eV
(O-C=O) are observed in the C 1s spectra.
The C 1s and O 1s XPS spectra of the graphite electrodes are
depicted in Figure 14. The C 1s spectrum of the fresh graphite shows
high intensity of the C-C peak at 284.3 eV, along with COx peaks of
the CMC binder.42 The anode cycled with the STD electrolyte con-
tains C 1s peaks characteristic of lithium alkyl carbonates and lithium
carbonate from carbonate solvent reduction as expected from SEI
formation.42,43 Similar peaks are also observed in the C 1s spectrum
of the anode cycled with the LiCDMB electrolyte. An additional C 1s
aromatic shake-up satellite is observed at 291.5 eV.42,43
XPS element spectra.—The O 1s and C 1s core spectra of the
LiNi0.5Mn1.5O4 electrodes are depicted in Figure 13. The O 1s spec-
trum of the fresh cathode is dominated by the metal oxide at 529.0
eV.39,40 The cathode cycled with STD electrolyte contains the same
O-M (M = Mn, Ni) peak at 529.0 eV, along with new peaks at higher
binding energy which correspond to electrolyte decomposition prod-
ucts on the cathode surface characteristic of C=O, C-O, and O-C=O
bonds respectively at 531.2 eV, 532.5 eV, and 533.0 eV.39,40,44–46 The
differences are greater for the cathode cycled with LiCDMB elec-
trolyte (Figure 10). The peak characteristic of the metal oxide at 529.0
eV is very weak while the peaks characteristic of electrolyte decom-
position products have high intensity. In addition, a new small peak is
observed at ∼535 eV which results from the oxidation of LiCDMB
and is likely the result of shake up satellites from the presence of aro-
matic species on the cathode surface.39,41–43,49,50 This suggests greater
coverage of the cathode surface in the presence of added LiCDMB
(Figure 4).
The O 1s spectrum of the fresh graphite electrode contains O 1s
peaks of the CMC binder, along with the Auger peak of sodium at 536
eV.40 The O 1s spectrum of the graphite electrode cycled with the STD
electrolyte contains peaks characteristic of electrolyte decomposition
products in the SEI such as lithium alkyl carbonates and lithium
carbonate: 531.2 eV, 532.5 eV, and 533.2 eV.39,41–43 The O 1s spectrum
of the graphite electrode cycled with the LiCDMB electrolyte contain
similar O 1s peaks. Nevertheless, a fourth peak is observed at 535.2
eV, consistent with shake up satellites from the deposition of aromatic
species and correlates with the C 1s peak at 291.5 eV.42,43 The addition
of LiCDMB results in the deposition aromatic species on the graphite
anode which are not present on the electrode cycled with the STD
electrolyte.
The Li 1s, Mn 3p, and Ni 3p spectra of the graphite anodes cy-
cled in electrolytes with and without added LiCDMB are presented
in Figure 15. The Li 1s, Mn 3p, and Ni 3p spectra of the anode
cycled with the STD electrolyte contain peaks at 49 eV (Mn+IV),
48 eV (Mn+III), 69 eV (Ni+IV),44 and LiF at 56 eV (Figure 15b).
This indicates that transition metal dissolution from the cathode sur-
face is occurring followed by deposition on the anode damaging the
SEI (Figure 13b).47,48 Much weaker intensity XPS peaks associated
with Ni and Mn are observed on the graphite anode cycled with
the LiCDMB electrolyte (Figure 15c). The reduced concentration
of Ni and Mn on the anode likely results from the generation of
a cathode passivation layer composed of the oxidation products of
LiCDMB which inhibits Mn and Ni dissolution. While the concen-
tration of B is surprisingly low, the B concentration is unfortunately
The C 1s core spectra of the cathodes reveal significant differences
in the electrode surfaces for the different electrolytes. The fresh elec-
trode contains C 1s peaks characteristic of C-C (284.3 eV) and C-H
(285.6 eV), along with peaks of the PVdF binder at 286.5 eV (-CH2-)
and 290.7 eV (-CF2-).39–41,44 The C 1s spectrum of the cathode cycled
with the STD electrolyte shows the deposition of organic species that
comprise the CH2 (285.6 eV), C-O (286.5 eV), C=O (ca. 288 eV),
and O-C=O (289.9 eV) groups.39,41,43,44 The low intensity of the PVdF
peaks at 290.7 eV (-CF2-) and at 286.5 eV (-CH2-) suggests that the
active material is mostly covered by the constituents of the CEI. The
XPS spectra of the cathode cycled with the LiCDMB electrolyte has
a thicker surface film since the peaks associated with PVdF are no
longer visible in the C 1s spectrum which confirms coverage of the
cathode material. Functional groups of the CEI at 284.3 eV (C-C),
Downloaded on 2016-12-13 to IP 80.82.78.170 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, 164 (2) A128-A136 (2017)
A135
Figure 14. (On the left) C 1s and (on the right) O 1s core spectra of (a) the fresh graphite electrode and graphite anodes cycled at 55◦C with the (b) STD and (c)
STD + 0.5% LiCDMB electrolytes.
difficult to determine due to overlap of the B1s peak with the P2s
peak.
An XPS investigation of LiNi0.5Mn1.5O4 and graphite electrodes
cycled with STD electrolyte with or without added LiCDMB indicates
that incorporation of LiCDMB results in the generation of a surface
film on the high voltage cathode. The presence of the novel surface
film inhibits Mn and Ni dissolution and subsequent deposition on
the graphitic anode. The presence of the cathode passivation layer re-
sults in enhanced cycling performance of the LiNi0.5Mn1.5O4/Graphite
cells.
Conclusions
Performance of LiNi0.5Mn1.5O4/Graphite cells cycled up to 4.8 V at
55◦C with 1.2 M LiPF6 in EC/EMC (3/7 v/v, STD electrolyte) with and
without 0.5% (wt) lithium catechol dimethyl borate (LiCDMB) has
been investigated. Upon cycling at 55 ◦C, cells with 0.5% LiCDMB
have improved capacity retention and better cycling efficiency. After
cycling the electrodes were extracted from the cells and ex-situ sur-
face analysis was conducted via a combination of SEM, TEM, and
XPS. Analysis of the cathode reveals the presence of LiCDMB reac-
tion products which generates a thick passivation layer on the high
voltage spinel. The novel cathode electrolyte interface (CEI) prevents
the dissolution of transition metal ions and subsequent migration and
deposition on the graphite anode. Analysis of the anodes reveals that
incorporation of LiCDMB results in the formation of a thin but con-
tinuous SEI with a much lower concentration of Mn than observed on
anodes cycled with the STD electrolyte. Incorporation of LiCDMB
has also been shown to inhibit the thermal decomposition of the STD
electrolyte at 85◦C. The incorporation of the novel additive, LiCDMB,
results in improved performance and changes to the electrode surface
chemistry for LiNi0.5Mn1.5O4/Graphite cells cycled to high potential.
Acknowledgment
The authors thank the Department of Energy Office of Basic En-
ergy Sciences EPSCoR Implementation award (DE–SC0007074) for
support of the XPS.
References
1. J. M. Tarascon and M. Armand, Nature, 414, 359 (2001).
2. M. Armand and J. M. Tarascon, Nature, 451, 652 (2008).
3. A. Manthiram, J. Phys. Chem. Lett., 2, 176 (2011).
4. A. Kraytsberg and Y. Ein-Eli, Adv. Energy Mater., 2, 922 (2012).
5. J. Wolfenstine and J. Allen, J. Power Sources, 136, 150 (2004).
Figure 15. Li 1s, Mn 3p, and Ni 3p core spectra of (a) the LiNi0.5Mn1.5O4
electrode and graphite anodes cycled in LiNi0.5Mn1.5O4/Gr cells with (b) the
STD and (c) STD + 0.5% LiCDMB electrolytes.
Downloaded on 2016-12-13 to IP 80.82.78.170 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A136
Journal of The Electrochemical Society, 164 (2) A128-A136 (2017)
6. J. L. Allen, T. R. Jow, and J. Wolfenstine, J. Power Sources, 196, 8656 (2011).
7. A. Manthiram, K. Chemelewski, and E.-S. Lee, Energy Environ Sci., 7, 1339 (2014).
8. M. Xu, L. Zhou, Y. Dong, U. Tottempudi, J. Demeaux, A. Garsuch, and B. L. Lucht,
ECS Electrochem. Lett, 4, A83 (2015).
9. M. Xu, L. Zhou, Y. Dong, Y. Chen, J. Demeaux, A. D. MacIntosh, A. Garsuch, and
B. L. Lucht, Energy Environ. Sci., 9, 1308 (2016).
31. Z.-M. Xue, B.-B. Sun, W. Zhou, and C.-H. Chen, J. Power Sources, 196, 8710 (2011).
32. W. Xu, A. J. Shusterman, R. Marzke, and C. A. Angell, J. Electrochem. Soc, 151,
A632 (2004).
33. C. L. Campion, W. Li, and B. L. Lucht, J. Electrochem. Soc, 152, A2327 (2005).
34. W. Li, C. Campion, B. L. Lucht, B. Ravdel, J. DiCarlo, and K. M. Abraham, J.
Electrochem. Soc, 152, A1361 (2005).
10. J. Demeaux, D. Lemordant, M. Caillon-Caravanier, H. Galiano, and
B. Claude-Montigny, Electrochim. Acta, 89, 163 (2013).
11. B. Michalak, B. B. Berkes, H. Sommer, T. Bergfeldt, T. Brezesinski, and J. Janek,
Anal Chem., 88, 2877 (2016).
12. M. He, L. Boulet-Roblin, P. Borel, C. Tessier, P. Nova´k, C. Villevieille, and E. J. Berg,
J. Electrochem. Soc, 163, A83 (2015).
35. B. Ravdel, K. M. Abraham, R. Gitzendanner, J. DiCarlo, B. Lucht, and C. Campion,
J. Power Sources, 119, 805 (2003).
36. J. Demeaux, M. Caillon-Caravanier, H. Galiano, D. Lemordant, and
B. Claude-Montigny, J. Electrochem. Soc, 159, A1880 (2012).
37. M. Nie, D. Chalasani, D. P. Abraham, Y. Chen, A. Bose, and B. L. Lucht, J. Phys.
Chem. C, 117, 1257 (2013).
13. D. Lu, M. Xu, L. Zhou, A. Garsuch, and B. L. Lucht, J. Electrochem. Soc, 160, A3138
(2013).
38. D. Aurbach, J. Power Sources, 89, 206 (2000).
39. R. Dedryve`re, D. Foix, S. Franger, S. Patoux, L. Daniel, and D. Gonbeau, J. Phys.
14. T. Yoon, S. Park, J. Y. Mun, J. H. Ryu, W. C. Choi, Y. S. Kang, J. H. Park, and
S. M. Oh, J. Power Sources, 215, 312 (2012).
15. L. Yang, T. Markmaitree, and B. L. Lucht, J. Power Sources, 196, 2251 (2011).
16. M. Xu, B. Knight, and B. L. Lucht, Electrochem. Solid State Lett., 15, A28 (2012).
17. M. Xu, N. Tsiouvaras, A. Garsuch, H. A. Gasteiger, and B. L. Lucht, J. Phys. Chem.
C, 118, 7363 (2014).
Chem. C, 114, 10999 (2010).
40. L. Bodenes, R. Dedryvere, H. Martinez, F. Fischer, C. Tessier, and J. P. Peres, J.
Electrochem. Soc, 159, A1739 (2012).
41. H. Bouayad, Z. Wang, N. Dupre`, R. Dedryve`re, D. Foix, S. Franger, J. F. Martin,
L. Boutafa, S. Patoux, D. Gonbeau, and D. Guyomard, J. Phys. Chem. C, 118, 4634
(2014).
18. I. A. Shkrob, Y. Zhu, T. W. Marin, and D. P. Abraham, J. Phys. Chem. C, 117, 23750
(2013).
42. L. El Ouatani, R. Dedryve`re, C. Siret, P. Biensan, S. Reynaud, P. Iratc¸abal, and
D. Gonbeau, J. Electrochem. Soc, 156, A103 (2009).
19. Y. Zhu, Y. Li, M. Bettge, and D. P. Abraham, Electrochim. Acta, 110,191 (2013).
20. W. Xu and C. A. Angell, Electrochem. Solid-State Lett., 4, E1 (2001).
21. W. Xu and C. A. Angell, Electrochem. Solid-State Lett., 4, L3 (2001).
22. U. Lischka, U. Wietelmann, and M. Wegner, Pat. DE 19829030 C1, Germany (1999).
23. M. Xu, L. Zhou, Y. Dong, Y. Chen, A. Garsuch, and B. L. Lucht, J. Electrochem.
Soc, 160, A2005 (2013).
43. L. El Ouatani, R. Dedryve´re, C. Siret, P. Biensan, and D. Gonbeau, J. Electrochem.
Soc, 156, A468 (2009).
44. J. Pires, A. Castets, L. Timperman, J. Santos-Pen˜a, E. Dumont, S. Levasseur,
C. Tessier, R. Dedryve`re, and M. Anouti, J. Power Sources, 296, 413 (2015).
45. R. Dedryve`re, H. Martinez, S. Leroy, D. Lemordant, F. Bonhomme, P. Biensan, and
D. Gonbeau, J. Power Sources, 174, 462 (2007).
24. S. Tsujioka, H. Takase, M. Takahashi, and H. Sugimoto, Jap. Pat. JP 2001247306 A,
46. R. Dedryve´re, D. Foix, S. Franger, S. Patoux, L. Daniel, and D. Gonbeau, J. Phys.
Japan (2001).
Chem. C, 114, 10999 (2010).
25. S. Tsujioka, H. Takase, M. Takahashi, and H. Sumgioto, Jap. Pat. JP 2001302675 A,
Japan (2001).
26. S. Tsujioka, H. Takase, M. Takahashi, and Y. Isono, Eur. Pat. EP 1308449 A2, Europe
(2003).
27. S. S. Zhang, Electrochem. Commun., 8, 1423 (2006).
28. M. Xu, L. Zhou, L. Hao, L. Xing, W. Li, and B. L. Lucht, J. Power Sources, 196,
6794 (2011).
47. L. Baggetto, N. Dudney, and G. Veith, Electrochim. Acta, 90, 135 (2013).
48. K. Carroll, D. Qian, C. Fell, S. Calvin, G. Veith, M. Chi, L. Baggettod, and Y. Meng,
Phys. Chem. Chem. Phys., 15, 11128 (2013).
49. M. Jung, U. Baston, T. Porwol, H. -J. Freund, and E. Umbach,
https://arxiv.org/ftp/cond-mat/papers/0408/0408663.pdf , (2004).
50. A. Majumdar, S. Chandra Das, T. Shripathi, J. Heinicke, and R. Hippler, Surface
Science, 609, 53 (2013).
29. Z.-M. Xue, J.-F. Zhao, J. Ding, and C.-H. Chen, J. Power Sources, 195, 853 (2010).
30. Z.-M. Xue, W. Zhou, J. Ding, and C.-H. Chen, Electrochim. Acta, 55, 5342 (2010).
51. K. Ishihara, A. Nagasawa, K. Umemoto, H. Ito, K. Saitole, and Znorg. Inorg. Chem.,
33, 3811 (1994).
Downloaded on 2016-12-13 to IP 80.82.78.170 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).