The role of “ZrF4-modification” on the structure and electrochemical performance of Li4Ti5O12 anode material

Article abstract of DOI:10.1016/j.jallcom.2018.02.197

In order to determine the reaction mechanism of the fluoride modification process and its influence on the electrochemical properties of Li₄Ti₅O₁₂ anode material, the “ZrF₄-modified” Li₄Ti₅O₁₂ was prepared via a co-precipitation method. Structural characterization results shows that both Zr⁴⁺ and F^? ions were not incorporated into the lattice structure of Li₄Ti₅O₁₂ after the modification process. Instead, F^? reacts with Li₄Ti₅O₁₂ chemically to generate new impurity phase such as anatase TiO₂ and LiF, while Zr⁴⁺ forms amorphous ZrO₂ nano-particles over the Li₄Ti₅O₁₂ particles. There was no dense coating layer formed on the surface of the Li₄Ti₅O₁₂ particles. These results indicate the reaction mechanism of the ZrF₄-modification process was different from that previously reported (such as “AlF₃, MgF₂, ZnF₂, SrF₂ and CaF₂-modification”). The rate capability and cycling stability of Li₄Ti₅O₁₂ can be enhanced due to the “ZrF₄-modification” process. Specifically, the 2 wt% “ZrF₄-modified” Li₄Ti₅O₁₂ shows the best electrochemical performance.

Full text of DOI:10.1016/j.jallcom.2018.02.197

Journal of Alloys and Compounds 745 (2018) 659e668
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: http://www.elsevier.com/locate/jalcom
The role of “ZrF4-modification” on the structure and electrochemical
performance of Li4Ti5O12 anode material
,
,
,
b
,
,
,
, b, **
Wen Li a b, Xue Bai a b, Lihui Zhang a *, Aijia Wei a b, Xiaohui Li a b, Zhenfa Liu a
a Institute of Energy Resources, Hebei Academy of Sciences, Shijiazhuang, Hebei Province 050081, PR China
b Hebei Engineer Research Center for Water Saving in Industry, Shijiazhuang, Hebei Province 050081, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
In order to determine the reaction mechanism of the fluoride modification process and its influence on
the electrochemical properties of Li4Ti5O12 anode material, the “ZrF4-modified” Li4Ti5O12 was prepared
via a co-precipitation method. Structural characterization results shows that both Zr4þ and Fꢀ ions were
not incorporated into the lattice structure of Li4Ti5O12 after the modification process. Instead, Fꢀ reacts
with Li4Ti5O12 chemically to generate new impurity phase such as anatase TiO2 and LiF, while Zr4þ forms
amorphous ZrO2 nano-particles over the Li4Ti5O12 particles. There was no dense coating layer formed on
the surface of the Li4Ti5O12 particles. These results indicate the reaction mechanism of the ZrF4-modi-
fication process was different from that previously reported (such as “AlF3, MgF2, ZnF2, SrF2 and CaF2-
modification”). The rate capability and cycling stability of Li4Ti5O12 can be enhanced due to the “ZrF4-
modification” process. Specifically, the 2 wt% “ZrF4-modified” Li4Ti5O12 shows the best electrochemical
performance.
Received 20 October 2017
Received in revised form
31 January 2018
Accepted 15 February 2018
Available online 16 February 2018
Keywords:
Li4Ti5O12
ZrF4
Modification
Lithium ion batteries
© 2018 Elsevier B.V. All rights reserved.
1. Introduction
anode material [17], showing that instead of the formation of an
AlF3 coating layer, Al3þ and Fꢀ ions were incorporated into the
Surface modification, especially with fluorides has been devel-
oped to improve the electrochemical performance of electrode
materials for lithium ion batteries (LIBs) [1e16]. Recently, AlF3
[1e9], CaF2 [10,11], SrF2 [12,13], LiF [14], LaF3 [15] and ZrFx [16] have
been widely used to modify cathode materials for improving their
electrochemical performance. The modification process involves
generation of a fluoride precipitate with a metal nitrate and
ammonium fluoride, and then chemical co-precipitation followed
by low-temperature calcination, which leads to the formation of a
fluoride coating layer over an electrode surface. The enhanced
performance was attributed to the fact that the fluoride coating
material which exhibits thermal stability and an inert behavior
with hydrofluoric acid, could prevent the electrode material from
reacting with hydrofluoric acid and effectively enhance its cycling
stability [1e16].
Li4Ti5O12 lattice structure with the production of a small amount of
anatase TiO2 after the modification process. The modified Li4Ti5O12
anodes delivered slightly higher discharge capacity and exhibited
excellent long-term cycling stability compared to the pure
Li4Ti5O12. The improved electrochemical performance could be
attributed to the facts that [17]: (1) co-doping of Al3þ and Fꢀ could
improve both ionic and electronic conductivities of the Li4Ti5O12
;
(2) the in situ generated TiO2 could enhance the ionic diffusion in
the electrode by restraining the growth of Li4Ti5O12 particles and by
providing a short ionic diffusion path through the boundary among
TiO2 and Li4Ti5O12; and (3) the Fꢀ doping could prevent the
Li4Ti5O12 particles from the attack by acidic species present in the
electrolyte, such as hydrofluoric acid. Our previous work showed
that AlF3-modification can usefully restrain the gas-release prob-
lem of LIBs using Li4Ti5O12 as the anode material [18]. The results
indicated that electrolyte reduction decomposition was suppressed
by the coating layer formed on the surface of Li4Ti5O12 after the AlF3
modification process, thereby preventing the gas generation during
cycling tests. To further understand the reaction mechanism of the
fluoride modification process, MgF2 [19] and ZnF2 [20] were both
used to separately modify Li4Ti5O12, respectively. The results show
that Fꢀ reacted with Li4Ti5O12 to produce a new impurity phase
Additionally, AlF3 has been first used to modify the Li4Ti5O12
* Corresponding author. Institute of Energy Resources, Hebei Academy of Sci-
ences, Shijiazhuang, Hebei Province 050081, PR China.
** Corresponding author. Institute of Energy Resources, Hebei Academy of Sci-
ences, Shijiazhuang, Hebei Province 050081, PR China.
E-mail addresses: zlhkxy@126.com (L. Zhang), liuzhenfa@hebut.edu.cn (Z. Liu).
https://doi.org/10.1016/j.jallcom.2018.02.197
0925-8388/© 2018 Elsevier B.V. All rights reserved.

660
W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
composed of anatase TiO2 and LiF, while the Mg2þ ions and Zn2þ
produced the MgO and ZnO coating layers on the surface of
Li4Ti5O12 particles, respectively [30,31]. The above mentioned MgO
and ZnO coating layers effectively curbed the electrolyte reduction
decomposition, thereby improving the electrochemical properties.
Li et al. [21] chose SrF2 to modify Li4Ti5O12, and indicated that Sr2þ
assembled in an argon-filled glove box using the as-prepared
Li4Ti5O12 as working electrode, 1 mol Lꢀ1 solution of LiPF6 in
ethylene carbonate, dimethyl carbonate, and diethyl carbonate
(volume ratio 1:1:1) (Capchem Technology (Shenzhen) Co., Ltd.) as
the electrolyte, a microporous polypropylene membrane (Celgard
2400, Celgard Inc., USA) as a separator and lithium foil as a counter
electrode.
and Fꢀ could not incorporate into the bulk phase of the Li4Ti5O12
,
but instead formed a SrF2 coating layer on the surface of the
Li4Ti5O12, which covered the catalytic active sites of the Li4Ti5O12
and suppressed the electrolyte reduction decomposition on the
surface of the Li4Ti5O12, thus improving the electrochemical per-
Galvanostatic charge and discharge experiments were carried
out on an automatic galvanostatic (dis)charge unit (Land CT 2001A,
Wuhan, China) between 0 and 3 V at different rates at 25 ꢁC. The
coin-type half cells were firstly (dis)charged at different C-rates
(0.5, 1, 3, 5, 10 C and 0.5 C) for five cycles, and then discharged and
charged at a 5-C ate for another 195 times. Cyclic voltammetry (CV)
and electrochemical impedance spectroscopy (EIS) tests were
conducted on an electrochemical workstation (Interface 3000,
Gamry, USA). For the CV measurements, the potential range was set
from 0 to 3 V while the scan rate was set at 0.2 mV sꢀ1. The EIS
measurements were performed in the frequency range of
10 mHze100 kHz before discharge-charge tests for the first time, at
first charging up to 1.6 V for the second time and at 21st charging
up to 1.6 V for the third time with a current density of 1 C rate in the
voltage range of 0e3 V, respectively.
formance. Recently, we used CaF2 to modify commercial Li4Ti5O12
,
and found that Fꢀ ions reacted with Ca2þ ions to form microscale
CaF2 crystals stacked on the surface of Li4Ti5O12 particles instead of
a CaF2 coating layer which not only reduced the electrode polari-
zation, but also partially suppressed the reductive decomposition,
both of which were beneficial to the promotion of the electro-
chemical performance of Li4Ti5O12 [22].
Summarily, the reaction mechanism of the fluoride modification
process for Li4Ti5O12 highly depends on the composition of the
coating materials. So far, ZrF4 has not been used to modify Li4Ti5O12
anode material and the reaction mechanism of the ZrF4-modifica-
tion process is not clearly known. In this paper, the “ZrF4-modified”
Li4Ti5O12 was obtained via the above-mentioned co-precipitation
method for the first time to our knowledge, and the reaction
mechanism of the ZrF4-modification process is further investigated
systematically.
3. Results and discussion
XRD patterns of all samples are shown in Fig. 1(a)-(c) and the
standard XRD patterns of Si, rutile TiO2, anatase TiO2, LiF and
Li4Ti5O12 are shown in Fig. 1d. Internal standard and calibration
curves are used to analyze the lattice parameters of all materials,
and the diffraction peaks of standard Si are marked with asterisks
($). Among all investigated samples, the major diffraction peaks
were in accordance with the standard patterns of Li4Ti5O12 with
PDF number of 72e0426, suggests that the modification process
does not change the basic Li4Ti5O12 structure. The lattice parame-
ters of all samples (shown in Table 1) were obtained by the Rietveld
refinement and the obtained lattice constants for all those samples
are 8.3568(5), 8.3568(6), 8.3568(8), 8.3569(3), 8.3568(7), and
8.3569(4) for pure Li4Ti5O12, 1ZFLTO, 2ZFLTO, 3ZFLTO, 2FLTO and
2ZLTO, respectively. Previous research showed that the substitution
of Ti4þ with Zr4þ ions for Li4Ti5O12 increases the lattice parameter
due to the larger ionic radius of Zr4þ (0.072 nm) compared to Ti4þ
(0.0605 nm) [23e25]. Moreover, the introduction of Zr4þ ions in
Li4Ti5O12 can form the anatase TiO2. However, in our case, no lattice
parameter changes and anatase TiO2 were detected for 2ZLTO
sample, suggesting that Zr4þ ions do not incorporate the Li4Ti5O12
lattice after the modification process. After the introduction of
Fꢀinto the Li4Ti5O12 particles, Fꢀ will incorporate the [O]32e site,
causing the transition of a certain amount of Ti ions from Ti4þ to
Ti3þ due to the charge compensation [26]. It is well known that the
incorporation of Fꢀ into the Li4Ti5O12 lattice causes the decrease of
the lattice parameter, based on Vegard's law. Although the smaller
radius of Fꢀ ions (0.133 nm) compared to that of O2ꢀ (0.140 nm) will
cause the shrinkage of the lattice, while the radius of Ti3þ
(0.067 nm) being larger than that of Ti4þ (0.0605 nm) will expand
the lattice, the former plays a leading role in determining the lattice
parameter and thus leads a decrease of the lattice constants due to
the incorporation of F into the structure [27]. The lattice parameter
2. Experimental
Li4Ti5O12 materials used in this work were obtained from
Sichuan Xingneng New Materials Co. For those coated materials,
Li4Ti5O12 powder was first dispersed in the deionized water. Next,
ammonium fluoride (NH4F, AR) and magnaliumnitrate non-
ahydrate [Zr(NO3)4$5H2O, AR] were added into the above obtained
Li4Ti5O12 suspension based on the fixed stoichiometric molar ratio
of Zr4þ/Fꢀ ¼ ¼, and various coating amounts were chosen. The
Li4Ti5O12 powder in the suspension was then collected and dried at
80 ꢁC for 10 h. Finally, the ZrF4-modified Li4Ti5O12 samples used in
this work were obtained by calcining the above drying powder at
400 ꢁC for 5 h under argon atmosphere. All ZrF4-modified Li4Ti5O12
samples were designated mZFLTO, where m denotes the weight
ratio of ZrF4/Li4Ti5O12 (i.e. m ¼ 1, 2, and 3). The Li4Ti5O12 was also
modified with Zr4þ and Fꢀ, denoted 2MLTO and 2FLTO, respectively.
Powder XRD patterns were collected over a 2q
range of 10ꢁe90ꢁ
to identify the phase composition and crystal lattice parameters of
all the Li4Ti5O12 samples using CuK X-rays (1.5406 Å at 40 kV and
a
40 mA) on an Ultima IV (Rigaku) X-ray diffractometer. The partic-
ulate morphology and the coating layer of the Li4Ti5O12 samples
using
a scanning electron microscope (SEM, Sirion 200 FEI
Netherlands) and a transmission electron microscope (TEM, JEOL
JEM-2100plus). Chemical composition of the coating layer for the
modified- Li4Ti5O12 samples were determined by X-ray photo-
electron spectroscopy (XPS, PHI5600 Physical Electronics).
Electrochemical characterization of the Li4Ti5O12 samples were
performed by assembling 2032 coin-cells, which were consist of a
Li4Ti5O12 electrode disc and a piece of lithium metal separated by a
polymer separator. The Li4Ti5O12 electrode used in this work was
made by mixing 85 wt% as-prepared samples (pure Li4Ti5O12 and
the modified Li4Ti5O12 samples) with 10 wt% conductive carbon
(Super-P) and 5 wt% polyvinylidene fluoride binder using N-methyl
pyrrolidone as a solvent to form a slurry. Then, the obtained slurry
was coated on a Cu foil by painting. Afterward, the working elec-
trode was dried in a vacuum oven at 105 ꢁC for 12 h to remove any
residual solvent and the adsorbed moisture. The coin-cell was
of 2FLTO has almost the same value as that for the pure Li4Ti5O12
,
indicating that no Fꢀ ions incorporate into the Li4Ti5O12 structure
after the modification process. Additionally, the lattice parameters
of mZFLTO samples are quite similar to that of the pure Li4Ti5O12
,
showing that both Zr4þ and Fꢀ are not incorporated into the
Li4Ti5O12 lattice after the modification process. Moreover, 2FLTO
shows extra peaks indexed to anatase TiO2 and LiF due to the fact
that Fꢀ has chemically reacted with Li4Ti5O12 [19]. Interestingly,

W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
661
Fig. 1. XRD patterns of al samples in the range of (a) 10e90ꢁ, (b) 10e40ꢁ, (c) 34e50ꢁ, (d) standard XRD patterns of Si, rutile TiO2, anatase TiO2, LiF and Li4Ti5O12
.
Table 1
Lattice parameters of all samples.
Samples
Lattice parameters
/Å
Interplanar spacing
d/Å
a
V/Å3
Li4Ti5O12
1ZFLTO
2ZFLTO
3ZFLTO
2FLTO
8.35685
8.35686
8.35688
8.35693
8.35687
8.35694
583.61685
583.61894
583.62313
583.63361
583.62104
583.63570
4.82483
4.82484
4.82485
4.82487
4.82484
4.82488
2ZLTO
these mZFLTO samples also exhibit diffraction peaks due to anatase
and LiF, and the intensity of which increases with both the
increasing of m value and the introduction of Fꢀ amount. Sum-
marily, Zr4þ is chemically stable with Li4Ti5O12 while Fꢀ is chemi-
cally unstable with Li4Ti5O12 during the “ZrF4-modification”
process. It is noteworthy that no diffraction peaks indexed to ZrF4
were observed for all of the mZFLTO samples, providing that no ZrF4
was formed during the modification process.
Fig. 2 displays the SEM images of pure Li4Ti5O12 and modified
Li4Ti5O12 samples. The pure Li4Ti5O12 particles show a well-
crystallized structure and a smooth surface. In contrast, the
2ZFLTO, 2FLTO, and 2ZLTO particles present a rough surface and
there are some tiny particles on the surface of the Li4Ti5O12 parti-
cles, which might be the generated impurity phase. The tiny par-
ticles on the 2FLTO particles should be generated anatase TiO2 and
LiF during the modification process [19]. XPS measurements were
performed to clarify the generated impurity phases on the 2ZFLTO
particles and the corresponding results are shown in Fig. 3. Not only
were Ti, O, and Li elements detected by XPS, but Zr and F elements
were also found. The Ti 2p spectrum of the 2ZFLTO sample in Fig. 3a
consists of a Ti 2p3/2 peak at 458.32 eV and a Ti 2p1/2 peak at
463.98 eV, which can be assigned to Ti4þ. Moreover, the peak at
529.87 eV in the O 1s spectrum shown in Fig. 3c also proves the
presence of O2ꢀ. The combination of Ti and O spectra prove the
formation of TiO2. The peaks observed at 685.15 eV in the F 1s
spectrum shown in Fig. 3d and 50.32 eV in the Li 1s spectrum
shown in Fig. 3f are both related to LiF. These XPS results further
demonstrate that TiO2 and LiF exist on the 2MFLTO particles, which
is in good agreement with the XRD results. The pair of Zr 3d3/2
peaks located at 182.15 and 184.49 eV appear in the Zr 3d spectrum
shown in Fig. 3e are in good agreement with ZrO2. The peak at
531.58 eV in the O 1s XPS spectrum in Fig. 3c also certifies the
presence of O2ꢀ. Therefore, the Zr4þ ions form ZrO2 instead of ZrF4
after the modification process. Previous XRD results show that no
diffraction peaks belonging to ZrO2 were detected; that is, the ZrO2
impurity clarified here is most likely to exist on the Li4Ti5O12 par-
ticles as an amorphous state. In order to clarify the existence of
amorphous ZrO2 on the surface of Li4Ti5O12 particles, high-

662
W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
Fig. 2. SEM images of (a) pure Li4Ti5O12, (b) 2ZFLTO, (c) 2FLTO and (d) 2ZLTO sample.
resolution TEM (HRTEM) was carried out in 2ZFLTO and 2ZLTO
samples. As shown in Fig. 4, the 2ZFLTO and 2ZLTO particles are
covered with nano-particles showing no crystal plane but an
amorphous state, which is consistent with previous XRD results.
Moreover, instead of being homogenously distributed, these nano-
particles are isolatedly distributed on the surface of Li4Ti5O12 par-
ticles. No dense ZrO2 coating layer was detected on the surface of
the 2ZFLTO and 2ZLTO particles.
formed on the surface of Li4Ti5O12 particles, and that the ZrO2 nano-
particles on the Li4Ti5O12 could not cover all of the active sites on
the surface of the Li4Ti5O12 particles. It is noteworthy that the peak
intensity for the electrolyte reduction decomposition of 2ZFLTO
becomes obviously weaker than that of pure Li4Ti5O12, indicating
that the electrolyte reduction decomposition is restrained partially
due to the ZrO2 nano-particles formed on the Li4Ti5O12. Moreover,
the mZFLTO samples also show an extra pair of redox peaks in the
range of 1.6e2.1 V and the current response of this redox reaction
increases with increasing modification content, which is ascribed
to the Liþ insertion into and de-insertion out of anatase TiO2
[17e19]. However, the 2ZLTO samples exhibit no extra pair of redox
peaks in the range of 1.6e2.1 V due to the fact that no anatase TiO2
formed during the modification process. The CV test results further
testify to the fact that Fꢀ will react with Li4Ti5O12 first to form
anatase TiO2 for the mZFLTO samples. Fig. 5b exhibits the galva-
nostatic charge-discharge curves of pure LTO, mZFLTO and 2ZLTO
samples at a 0.5 C-rate in the range of 0e3 V. The galvanostatic
charge-discharge curves of all samples demonstrate a long range of
flat voltage at approximately 1.5 V during the discharge process and
at approximately 1.6 V during charge process, which are assigned to
the Liþ insertion and de-insertion of two-phase reaction and are
also correspond to the one pair of strong redox peaks in the CV
curves shown in Fig. 5. Moreover, the galvanostatic charge-
discharge curves of all samples exhibit a short range of flat
voltage at approximately 0.7 V during the discharge process due to
the electrolyte reduction decomposition, suggesting that the elec-
trolyte reduction decomposition could not be suppressed effec-
tively after the modification process. It is worth noting that 2ZFLTO
presents a range of flat voltage at approximately 0.7 V that is
shorter than that of pure LTO, further proving that the electrolyte
reduction is partially suppressed. Thus, both the results from the CV
curves and galvanostatic charge-discharge curves illustrate that the
electrolyte reduction decomposition could not be suppressed
effectively, due to the fact that there is no dense coating layer
In order to study the effects of the ZrF4-modification process, the
electrochemical properties of the pure Li4Ti5O12 and modified
Li4Ti5O12 samples were tested systematically. Fig. 5a shows the CV
curves of the pure Li4Ti5O12, mZFLTO and 2ZLTO samples in the
range of 0e3 V at 0.2 mV sꢀ1. As shown in Fig. 5a, the cathodic and
anodic peaks at respective approximately 1.7 and 1.5 V, respec-
tively, are attributed to the redox of Ti4þ/Ti3þ, and the reduction
and oxidation peaks below 0.6 V are caused by a multi-step resto-
ration of Ti4þ [17e22]. It is worth noting that there is an obvious
irreversible peak appearing at approximately 0.7 V during the
reduction process for all samples, which is due to the electrolyte
reduction decomposition on the Li4Ti5O12 particles. It has been
proved that the irreversible peak disappears completely at
approximately 0.7 V for the MgF2- or ZnF2-modified Li4Ti5O12
samples, due to the amorphous MgO or ZnO coating layer formed
during the modification process which could suppress the elec-
trolyte reduction decomposition on the Li4Ti5O12 particles detected
in our previous study. It is also demonstrates that the irreversible
peak at approximately 0.7 V of the CaF2-modified LTO sample partly
disappears, which is attributed to the fact that the stacked CaF2
crystals formed during the modification process could cover part of
the active sites on the surface of the Li4Ti5O12 particles thereby
restraining the electrolyte reduction decomposition. However, the
irreversible peaks at approximately 0.7 V for the mZFLTO and 2ZLTO
samples almost does not disappear, suggesting that the electrolyte
reduction decomposition could not be suppressed effectively. This
could be ascribed to the fact that there is no dense coating layer

W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
663
Fig. 3. (a) XPS spectrum, (b) Ti 2p, (c) O 1s, (d) Li 1s, (e) Zr 3d and (f) F 1s spectrum of 2ZrFLTO sample.

664
W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
Fig. 4. HR-TEM images of (aeb) 2ZFLTO and (ced) 2ZLTO samples.
Fig. 5. (a) Cyclic voltammetry curves at a scan rate of 0.2 mV sꢀ1 and (b) galvanostatic
charge-discharge curves at 0.5 C-rate of the pure Li4Ti5O12 and mZFLTO samples in the
range of 0e3 V.
Fig. 6. (a) Charge capacity retention at different C-rates and (b) the second-cycle
charge capacity as a function of the C-rates in the range of 0e3 V.

W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
665
formed over the Li4Ti5O12 particles.
Fig. 6a shows the charge capacity of the pure Li4Ti5O12 and the
mZFLTO samples at various C rates in the voltage window between
0 and 3 V, and the second-cycle charge capacity at each rate is
compared in Fig. 6b (1 C ¼ 250 m Ah gꢀ1). As it shown in Fig. 6, the
2ZFLTO and 3ZFLTO samples exhibit better rate capability than pure
Li4Ti5O12, suggesting that improved rate capability can be obtained
by ZrF4-modification. It is worth mentioning that the 2ZFLTO
sample delivers the best rate capability among all samples. Table 2
shows the charge capacity at different C rates of pure LTO and
mZFLTO samples. Pure Li4Ti5O12 exhibits charge capacities of 235.5,
219.9, 204.8, 182.9 and 170.5 mAh gꢀ1 at C rates of 0.2, 0.5, 1, 3, and
5, respectively. The corresponding charge capacities of 2ZFLTO are
242.1, 225.7, 215.9, 193.6 and 182.0 mAh gꢀ1, respectively, showing
an improvement of 2.8%, 2.6%, 5.4%, 4.9% and 6.7% at various rates
respectively. One possible reason for the improved charge capacity
of 2ZFLTO is that the anatase TiO2 formed during the modification
process can act as an anode material, providing extra capacity
during the lithium intercalation process. Fig. 7 shows the galva-
nostatic discharge-charge curves of pure Li4Ti5O12 and 2ZFLTO at
different C-rates in the voltage range of 0e3 V. It can been seen
from Fig. 7 that there is a smaller potential difference between
discharge and charge platforms of 2ZFLTO than that of pure
Li4Ti5O12, especially at high C rates, showing that ZrF4-modification
is conducive to reduce the electrode polarization.
EIS was applied to investigate the electrochemical kinetics be-
haviors at the interface of electrolyte/electrode interface for a better
understanding of rate performance. Before the discharge-charge
tests, the EIS of both pure Li4Ti5O12 and 2ZFLTO was performed
on fresh cells. The EIS of both samples was then measured for the
second time, first charging up to 1.6 V and for the third time at 21st
charging up to 1.6 V, with a current density rate of 1 C in the range
of 0e3 V. Next, the coin cells of both samples were first charged up
to 1.6 V with a current density rate of 1 C followed by EIS mea-
surements. Finally, after 20 cycles under the current density of 1 C,
the same batteries were charged up to 1.6 V again, and EIS mea-
surements were started. All the EIS experimental results were
simulated by Z-view software using equivalent circuits inset
Fig. 8aec. The equivalent circuit (shown in Fig. 8aec) consists of
several components: the solution resistance (Rs), theþresistance of
Fig. 7. Galvanostatic discharge-charge curves of (a) pure Li4Ti5O12 and (b)2ZFLTO at
different C-rates in the range of 0e3 V.
comparison, Nyquist plots of cells first charged up to 1.6 V show
two semicircles of HFS and LFS, and for those fresh cells only one
semicircle of LFS. In the low-frequency range, a slope line is
observed for those Nyquist plots of all samples, corresponding to
the resistance of Liþ diffusion in active materials (LFL). The Liþ
diffusion coefficient (D) can be calculated from the low-frequency
plots according to the following equation [29,30].
0
the charge transfer reaction (R1 ), the resistance of Li diffusion in
the surface film layer (R2), the resistance of electronic property
change of active material (R3), the Warburg impedance (Zw) related
to the Liþ diffusion in solid state of active materials, the constant
phase element of capacitance corresponding to R1 (CPE1), the
constant phase element of capacitance corresponding to R2 (CPE2)
and the constant phase element of capacitance corresponding to R3
(CPE3). As illustrated, Nyquist plots of both samples at different
state in the high-frequency range exhibit different features. Nyquist
plots of both cells after 20 cycles and charging to 1.6 V possess three
semicircles corresponding to the resistance of Liþ diffusion in the
surface film layer (HFS), resistance of the electronic property
change of the active material (MFS), and resistance of the charge-
R2T2
2A2n4F4C2s2
D ¼
(1)
where A is the surface area of the electrode; n the number of the
electrons per molecule attending the electronic transfer reaction; F
the Faraday constant; C the concentration of Li ions in the Li4Ti5O12
electrode; R the gas constant; T the working temperature, which is
room temperature in our experiment; and
u
s
the slope of the line Z'~
ꢀ0.5, which can be obtained from the line of Z'~ uꢀ0.5 (shown in
transfer reaction (LFS), respectively [28] (Fig. 8c). As
a
Fig. 8def) [29,30]. The impedance parameters are listed in Table 3.
For the fresh cells, the absence of HFS could be due to the fact that
there is almost no SEI film formed on the pure Li4Ti5O12 and 2ZFLTO
electrodes. The LFS value of 2ZFLTO (108.33
U) is higher than that of
Table 2
pure LTO (60.54 ), which could be ascribed to the fact that the
U
Charge capacity at different C-rates of pure Li4Ti5O12 and mZFLTO samples.
impurity particles formed on the surface of 2ZFLTO are electroni-
Sample
Charge capacity at different C-rates (mAh gꢀ1
)
cally unconducive. Moreover, 2ZFLTO shows a slight higher Liþ
diffusion coefficient (6.94 ꢂ 10ꢀ14 cm2 sꢀ1
)
than that of pure
0.2 C
0.5 C
1 C
3 C
5 C
Li4Ti5O12 (3.60 ꢂ 10ꢀ14 cm2 sꢀ1), indicating that the modification
process is likely to promote the Liþ diffusion in the solid state of
active of active materials. After discharge-charge cycling tests, a SEI
film will be formed on the surface of the electrode material, which
Li4Ti5O12
1ZFLTO
2ZFLTO
3ZFLTO
235.5
239.9
242.1
236.9
219.9
223.6
225.7
221.4
204.8
207.6
215.9
210.9
182.9
183.4
193.6
191.8
170.5
169.6
182.0
177.4

666
W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
Fig. 8. (aec) EIS curves and equivalent circuits (insert) and (def) -Z00 plotted against ꢀ0.5 at low frequency region for pure LTO and 2ZFLTO sample (aeb) Before discharge-charge
u
cycling test; (ced) Fist charging up to 1.6 V at 1 C rate; (eef) 21st charging up to 1.6 V at 1 C rate; HFS (resistance of Liþ diffusion in the surface film layer); MFS (resistance of
electronic property change of active material); LFS (resistance of the charge transfer reaction); LFL (resistance of Liþ diffusion in solid state of active materials).
Table 3
The impedance parameters of EIS.
Before cycling test
Li4Ti5O12
First charging up to 1.6V at 1C rate
21st charging up to 1.6 V at 1 C rate
2ZFLTO
Li4Ti5O12
2ZFLTO
Li4Ti5O12
2ZFLTO
HFS (
MFS (
LFS (
U
U
)
)
/
/
/
/
93.92
/
94.23
13.98
1.30E-12
51.54
/
42.19
2.54
22.23
6.01
33.78
3.57
8.73
7.73
9.14
2.00
U
)
60.18
84.04
3.60E-14
54.16
30.24
6.94E-14
s
(
U
cm2 sꢀ0.5
D (cm2 sꢀ1
)
)
3.94E-11
1.99E-11
6.35E-11
is displayed as the semicircle of HFS in the Nyquist plots of both
samples. The modification sample 2ZFLTO shows lower HFS and LFS
values than those of pure Li4Ti5O12 (51.54 versus 93.92
versus 94.23 U
U
, and 42.19
), and show larger Liþ diffusion coefficients than the

W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
667
latter (3.94 ꢂ 10ꢀ11 versus 1.30 ꢂ 10ꢀ12 cm2 sꢀ1) at first charging up
Table 4
The charge capacity retention pure Li4Ti5O12 and mZFLTO samples at 5 C rate.
to 1.6 V. 2ZFLTO delivers a lower HFS (8.73 versus 22.23 U), MFS
(7.73 versus 6.01 U), and LFS (2.00 versus 3.57 U) values than those
Sample
Charge capacity in the range
of 0e3 V (mA h gꢀ1
Capacity retention (%)
of pure LTO. Additionally, the Liþ diffusion coefficient of 2ZFLTO is
higher than that of pure Li4Ti5O12 (6.35 ꢂ 10ꢀ11 versus
1.99 ꢂ 10ꢀ11 cm2 sꢀ1) at 21st charging up to 1.6 V. The EIS results
prove that both the electronic and ionic conductivity were effec-
tively improved after modification. The surface modification can
facilitate the charge-transfer reaction occurring at the interface
between Li4Ti5O12 particles and the electrolyte; specifically, it can
promote the transport of Liþ ions and/or electrons, as follows.
)
1
st cycle
100st cycle
Li4Ti5O12
1ZFLTO
2ZFLTO
3ZFLTO
170.5
169.7
182.7
177.4
160.0
161.4
179.0
167.9
93.8
95.1
98.5
94.6
cycling performance of Li4Ti5O12
.
(1) Regarding Li-ion transport, the EIS results indicate that the
formed ZrO2 nano-particles could effectively accelerate the
diffusion of Li ions through SEI film, by reducing the HFS. The
reason is most likely due to the fact that the formed ZrO2
nano-particles on the surface of Li4Ti5O12 particles could
partly prevent the direct contact between Li4Ti5O12 and the
electrolyte, and cover the catalytic sites on the Li4Ti5O12
surface for the decomposition of electrolyte, thereby
restraining the formation of SEI film.
(2) Regarding electron transport, it has been reported that sur-
face modification on electrode materials can not only
maintain the electron pathways between active materials
and carbon additives [31,32], but also improve the adhesion
of electrode materials to the current collector [31,33],
thereby promoting the electron transport among them. In
our case, therefore, it can be considered that ZrF4-modifica-
tion acts in a similar way to improve electron transport and
contribute to the reduced charge-transfer resistance of
4. Conclusions
The reaction mechanism of the ZrF4-modification process was
different from other such processes, e.g., AlF3, MgF2, ZnF2, SrF2, and
CaF2 modification which had been previously reported [17e22].
After the ZrF4-modification process, Fꢀ reacts with Li4Ti5O12
chemically to generate a new impurity phase, such as anatase TiO2
and LiF, while Zr4þ forms amorphous ZrO2 nano-particles over the
Li4Ti5O12 particles. The nano-particles cannot cover all the Li4Ti5O12
particles, with the nano-particles layer over the Li4Ti5O12 particles
being non-continuous. Because no dense ZrO2 coating layer was
formed on the surface of the Li4Ti5O12 particles, the catalytic sites
on the Li4Ti5O12 surface for the decomposition of electrolyte could
not be covered completely. However, the new phase formed over
the Li4Ti5O12 particles could partly suppress the decomposition of
electrolyte, thereby restraining SEI formation, promoting electron
transport across the electrode/electrolyte interface and thus
reducing the charge-transfer resistance. Meanwhile, ZrF4-modifi-
cation is conducive to reducing the electrode polarization. There-
fore, the rate capability of Li4Ti5O12 electrode materials can be
improved by ZrF4 modification. In addition, the formation of both
ZrO2 nano-particles and LiF over the Li4Ti5O12 particles is beneficial
to preventing the electrolyte reduction decomposition, and to
suppressing the gas generation, which is conducive to the cycling
performance of Li4Ti5O12. The 2ZFLTO sample, in particular, it de-
livers the best electrochemical performances among all the modi-
fied samples.
2ZFLTO compared with that of pure Li4Ti5O12
.
Fig. 9 shows the cycling performances of pure Li4Ti5O12 and
mZFLTO samples at a 5-C rate, and the corresponding results are
listed in Table 4. After 100 cycles, the charge capacity retention of
pure Li4Ti5O12 is 93.8%, while that of the 2ZFLTO samples is 98.5%.
The modification sample shows slight higher capacity retention
than the pure sample, suggesting that 2-wt% ZrF4-modification can
improve the cycling performance of Li4Ti5O12. This illustrates that
the Zr4þ and Fꢀ co-modification process benefits enhancement the
cycling stability at a high C rate, due to the fact that the formation of
both ZrO2 nano-particles and LiF over the 2ZFLTO particles are
beneficial to preventing the electrolyte reduction decomposition,
and to suppressing the gas generation, which is conducive to the
Acknowledgments
This work was carried out with financial support from Hebei
Province Applied Basic Research Program-Key Basic Research
Project (17964407D), Science and Technology Program of Hebei
Province (17394411D), Project of Hebei Academy of Science
(171407) and Project of Hebei Academy of Science (171413).
References
[1] Y.K. Sun, J.M. Han, S.T. Myung, S.W. Lee, K. Amine, Significant improvement of
high voltage cycling behavior AlF3-coated LiCoO2 cathode, Electrochem.
Commun. 8 (2006) 821e826.
[2] Y.K. Sun, S.W. Cho, S.T. Myung, K. Amine, J. Prakash, Effect of AlF3 coating
amount on high voltage cycling performance of LiCoO2, Electrochim. Acta 53
(2007) 1013e1019.
[3] K.S. Lee, S.T. Myung, D.W. Kim, Y.K. Sun, AlF3-coated LiCoO2 and Li[Ni1/3Co1/
3Mn1/3]O2 blend composite cathode for lithium ion batteries, J. Power Sources
196 (2011) 6974e6977.
[4] G.M. Song, Y. Wu, G. Liu, Q. Xu, Influence of AlF3 coating on the electro-
chemical properties of LiFePO4/graphite Li-ion batteries, J. Alloys Compd. 487
(2009) 214e217.
[5] B.C. Park, H.B. Kim, S.T. Myung, K. Amine, I. Belharouak, S.M. Lee, Y.K. Sun,
Improvement of structural and electrochemical properties of AlF3-coated Li
[Ni1/3Co1/3Mn1/3]O2 cathode materials on high voltage region, J. Power Sour-
ces 178 (2008) 826e831.
[6] S.H. Lee, C.S. Yoon, K. Amine, Y.K. Sun, Improvement of long-term cycling
performance of Li[Ni0.8Co0.15Al0.05]O2 by AlF3 coating, J. Power Sources 234
Fig. 9. Cycling stability of pure Li4Ti5O12 and mZFLTO samples at 5 C rate in the range
of 0e3 V.

668
W. Li et al. / Journal of Alloys and Compounds 745 (2018) 659e668
(2013) 201e207.
modification on the structure and electrochemical performance of Li4Ti5O12
anode material, Nano 12 (2017) 17500541e17500549.
[21] X. Li, X. Zhao, P.X. Huang, M.S. Wang, Y. Huang, Y. Zhou, Y.H. Lin, M.Z. Qu,
Z.L. Yu, Enhanced electrochemical performance of SrF2-modified Li4Ti5O12
composite anode materials for lithium-ion batteries, J. Alloys Compd. 693
(2017) 61e69.
[22] W. Li, A.J. Wei, L.H. Zhang, X.H. Li, X. Bai, Z.F. Liu, Structure and enhanced
electrochemical performance of the CaF2-modified Li4Ti5O12 anode material,
J. Electroanal. Chem. 791 (2017) 196e203.
[23] J.G. Kim, M.S. Park, S.M. Hwang, Y.U. Heo, T. Liao, Z. Sun, J.H. Park, K.J. Kim,
G. Jeong, Y.J. Kim, J.H. Kim, S. Dou, Zr4þ Doping in Li4Ti5O12 anode for lithium-
ion batteries: open Liþ diffusion paths through structural imperfection,
ChemSusChem 7 (2014) 1451e1457.
[24] X. Li, M.Z. Qu, Z.L. Yu, Structural and electrochemical performances of
Li4Ti5ꢀxZrxO12 as anode material for lithium-ion batteries, J. Alloys Compd.
487 (2009) L12eL17.
[25] R.D. Shannon, Revised effective ionic radii and systematic studies of inter-
atomic distances in halides and chalcogenides, Acta Crystallogr. 32 (1976)
751e767.
[7] Y.K. Sun, M.J. Lee, C.S. Yoon, J. Hassoun, K. Amine, B. Scrosati, The role of AlF3
coatings in improving electrochemical cycling of Li-enriched nickel-manga-
nese oxide electrodes for Li-ion batteries, Adv. Mater. 24 (2012) 1192e1196.
[8] J.M. Zheng, M. Gu, J. Xiao, B.J. Polzin, P.F. Yan, X.L. Chen, C.M. Wang, J.G. Zhang,
Functioning mechanism of AlF3 coating on the Li- and Mn-rich cathode ma-
terials, Chem. Mater. 26 (2014) 6320e6327.
[9] F. Ding, W. Xu, D.W. Choi, W. Wang, X.L. Li, M.H. Engelhard, X.L. Chen,
Z.G. Yang, J.G. Zhang, Enhanced performance of graphite anode materials by
AlF3 coating for lithium-ion batteries, J. Mater. Chem. 22 (2012)
12745e12751.
[10] J.G. Li, X.M. He, R.S. Zhao, Electrochemical performance of SrF2-coated
LiMn2O4 cathode material for Li-ion batteries, Trans. Nonferrous Met. Soc.
China 17 (2007) 1324e1327.
[11] J.G. Li, L. Wang, Q. Zhang, X.M. He, Electrochemical performance of SrF2-
coated LiNi1/3Co1/3Mn1/3O2 cathode materials for Li-ion batteries, J. Power
Sources 190 (2009) 149e153.
[12] K. Xu, Z.F. Jie, R.H. Li, Z.X. Chen, S.T. Wu, J.F. Gu, J.Z. Chen, Synthesis and
electrochemical properties of CaF2-coated for long-cycling Li[Mn1/3Co1/3Ni1/3
]
O2 cathode materials, Electrochim. Acta 60 (2012) 130e133.
[26] Y. Ma, B. Ding, G. Ji, J.Y. Lee, Carbon-encapsulated F-Doped Li4Ti5O12 as a high
rate anode material for Liþ batteries, ACS Nano 7 (2013) 10870e10878.
[27] X. Bai, W. Li, A.J. Wei, X.H. Li, L.H. Zhang, Z.F. Liu, Preparation and electro-
chemical properties of Mg2þ and F- co-doped Li4Ti5O12 anode material for use
in the lithium-ion batteries, Electrochim. Acta 222 (2016) 1045e1055.
[28] J. Chong, J.P. Zhang, H.M. Xie, X.Y. Song, G. Liu, V. Battaglia, S.D. Xun,
R.S. Wang, High performance LiNi0.5Mn1.5O4 cathode material with a bi-
functional coating for lithium ion batteries, RSC Adv. 6 (2016) 19245e19251.
[29] H.P. Liu, G.W. Wen, S.F. Bi, P. Gao, Enhanced rate performance of nanosized
Li4Ti5O12/graphene composites as anode material by a solid state-assembly
method, Electrochim. Acta 171 (2015) 114e120.
[30] Q.Y. Zhang, M.G. Verde, J.K. Seo, X. Li, Y.S. Meng, Structural and electro-
chemical properties of Gd-doped Li4Ti5O12 as anode material with improved
rate capability for lithium-ion batteries, J. Power Sources 280 (2015) 355e362.
[31] J. Liu, X.F. Li, M. Cai, R.Y. Li, X.L. Sun, Ultrathin atomic layer deposited ZrO2
coating to enhance the electrochemical performance of Li4Ti5O12 as an anode
material, Electrochim. Acta 93 (2013) 195e201.
[13] S.J. Shi, J.P. Tu, Y.J. Mai, Y.Q. Zhang, Y.Y. Tang, X.L. Wang, Structure and elec-
trochemical performance of CaF2 coated LiMn1/3Ni1/3Co1/3O2 cathode material
for Li-ion batteries, Electrochim. Acta 83 (2012) 105e112.
[14] S.J. Shi, J.P. Tu, Y.Y. Tang, Y.Q. Zhang, X.Y. Liu, X.L. Wang, C.D. Gu, Enhanced
electrochemical performance of LiF-modified LiNi1/3Co1/3Mn1/3O2 cathode
materials for Li-ion batteries, J. Power Sources 225 (2013) 338e346.
[15] S.H. Yun, K.S. Park, Y.J. Park, The electrochemical property of ZrFx-coated Li
[Ni1/3Co1/3Mn1/3]O2 cathode material, J. Power Sources 195 (2010)
6108e6115.
[16] Z.X. Yang, Q.D. Qiao, W.S. Yang, Improvement of structural and electro-
chemical properties of commercial LiCoO2 by coating with LaF3, Electrochim.
Acta 56 (2011) 4791e4796.
[17] W. Xu, X.L. Chen, W. Wang, D.W. Choi, F. Ding, J.M. Zheng, Z.M. Nie, Y.J. Choi,
J.G. Zhang, Z.G. Yang, Simply AlF3-treated Li4Ti5O12 composite anode materials
for stable and ultrahigh power lithium-ion batteries, J. Power Sources 236
(2013) 169e174.
[18] W. Li, X. Li, M.Z. Chen, Z.W. Xie, J.X. Zhang, S.Q. Dong, M.Z. Qu, AlF3 modifi-
cation to suppress the gas generation of Li4Ti5O12 anode battery, Electrochim.
Acta 139 (2014) 104e110.
[32] Y.S. Jung, A.S. Cavanagh, L.A. Riley, S.H. Kang, A.C. Dillon, M.D. Groner,
S.M. George, S.H. Lee, Ultrathin direct atomic layer deposition on composite
electrodes for highly durable and safe Li-ion batteries, Adv. Mater. 22 (2010)
2172e2176.
[33] L.A. Riley, A.S. Cavanagh, S.M. Geroge, S.H. Lee, A.C. Dillon, Improved me-
chanical integrity of ALD-coated composite electrodes for Li-ion batteries,
Electrochem. Solid State Lett. 14 (2011) A29eA31.
[19] W. Li, H. Wang, M.Z. Chen, J.J. Gao, X. Li, W.J. Ge, M.Z. Qu, A.J. Wei, L.H. Zhang,
Z.F. Liu, The reaction mechanism of the Mg2þ and F- co-modification and its
influence on the electrochemical performance of the Li4Ti5O12 anode material,
Electrochim. Acta 188 (2016) 499e511.
[20] A.J. Wei, W. Li, L.H. Zhang, X.H. Li, X. Bai, Z.F. Liu, Effect of Zn2þ and F- co-