AAAS

Research
Volume 2019, Article ID 2198906, 10 pages
https://doi.org/10.34133/2019/2198906

Research Article
Facilitating Lithium-Ion Diffusion in Layered Cathode
Materials by Introducing Li+/Ni2+ Antisite Defects for High-Rate
Li-Ion Batteries

Zhongfeng Tang,1 Sen Wang,2 Jiaying Liao,1 Shuo Wang,1 Xiaodong He,1 Bicai Pan,2
Haiyan He,2 and Chunhua Chen 1
1
CAS Key Laboratory of Materials for Energy Conversions, Department of Materials Science and Engineering & Collaborative
Innovation Center of Suzhou Nano Science and Technology, University of Science and Technology of China,

Downloaded from https://spj.science.org on June 19, 2024

Anhui Hefei 230026, China
2
School of Physical Sciences, University of Science and Technology of China, Anhui Hefei 230026, China

Correspondence should be addressed to Haiyan He; [email protected] and Chunhua Chen; [email protected]

Received 21 April 2019; Accepted 28 July 2019; Published 15 September 2019

Copyright © 2019 Zhongfeng Tang et al. Exclusive Licensee Science and Technology Review Publishing House. Distributed under a
Creative Commons Attribution License (CC BY 4.0).

Li+/Ni2+ antisite defects mainly resulting from their similar ionic radii in the layered nickel-rich cathode materials belong to one of
cation disordering scenarios. They are commonly considered harmful to the electrochemical properties, so a minimum degree of
cation disordering is usually desired. However, this study indicates that LiNi0.8Co0.15Al0.05O2 as the key material for Tesla
batteries possesses the highest rate capability when there is a minor degree (2.3%) of Li+/Ni2+ antisite defects existing in its
layered structure. By combining a theoretical calculation, the improvement mechanism is attributed to two effects to decrease
the activation barrier for lithium migration: (1) the anchoring of a low fraction of high-valence Ni2+ ions in the Li slab pushes
uphill the nearest Li+ ions and (2) the same fraction of low-valence Li+ ions in the Ni slab weakens the repulsive interaction to
the Li+ ions at the saddle point.

1. Introduction because a high-temperature treatment leads to the phase

transition from LiNiO2 to Li1‐x Ni1+x O2 which has a partially
Lithium-ion batteries are experiencing the large applications disordered cation distribution at the lithium layer, while a
in mobile electronic devices and electric vehicles worldwide. low-temperature treatment cannot bring sufficient crystallin-
High-energy density and high-power density are the two ity [11]. Although the cosubstitution of Co and Mn or Al for
most important factors in a commercial lithium-ion bat- Ni can greatly inhibit the LiNiO2 phase change, there is still
tery. For the cathode materials, compared to the widely used some cation disordering in Ni-rich layered materials because
LiFePO4, LiMn2O4, and LiNi1/3Co1/3Mn1/3O2 [1–4], the Ni2+ ions cannot be completely oxidized to Ni3+ even under
layered nickel-rich materials with higher capacity (170- oxygen-rich atmospheres [12, 13]. If there are a lot of Ni2+
200 mAh g-1) and appropriate working voltage (~3.75 V), such ions in the lithium layer (i.e., 3b sites), the diffusion of lith-
as LiNi0.6Co0.2Mn0.2O2, LiNi0.8Co0.1Mn0.1O2 (NCM811), and ium ions is inevitably hindered in the two-dimensional chan-
LiNi0.8Co0.15Al0.05O2 (NCA) [5–8], have attracted more and nels, which is why the researchers desire to decrease the
more attention. cation disordering in layered materials [14–16].
One of the challenging problems of layered nickel-rich Nevertheless, the Li+/Ni2+ antisite defects are not always
cathode materials is the Li+/Ni2+ antisite defects which result detrimental to layered materials. Zheng et al. found that these
from their similar ionic radii (Li+: 0.072 nm, Ni2+: 0.069 nm) defects benefit the thermal stability for Ni-rich NCM mate-
[9, 10]. It is very difficult to synthesize a pure LiNiO2 phase rials, because the Ni in the Li layer would form 180° Ni−O
2 Research

−Ni super exchange chains [17]. Moreover, it can partially

O
relieve the magnetic frustration by forming a stable antiferro- Co
magnetic state in hexagonal sublattice with nonmagnetic
Li
ions located in centers of the hexagons [18]. Lee et al. have
proved that a little amount of Ni2+ ions occupying Li+ (3b) Ni
sites is not a major obstacle to the diffusion of Li+ ions in
the lithium layer [19]. In some cases, the substitution of inac- c
tive ions for Li+ or Na+ in the layered cathode materials can a b
even improve the electrochemical properties due to the inhi-
bition of phase transition and the enlargement of lattice Figure 1: The structure of a 120-atom supercell consisting of 2 ×
parameters [20–22]. On the other hand, for the ions at 3a 5 × 1 LiNi0.8Co0.2O2 primitive unit cells.
sites, Kang et al. indicated that [23] lower valence transition
metal ions (Cu2+ or Ni2+ vs. Co3+ or Mn4+) can result in a microscopy (SEM, JSM-6390LA, JEOL). The crystal struc-
weaker electrostatic interaction between Li+ in the activated tures of the prepared samples were measured by a theta/theta
state and the transition metal ions at 3a sites and thus sub- rotating anode X-ray diffractometer (XRD, Rigaku TTR-lll).
stantially decrease the activation barrier for the migration The X-ray Rietveld refinement was performed by using a
of Li+ ions to the nearest vacancies. Hence, it is reasonable GSAS/EXPGUI package [25].
to speculate that the Li+ ions at 3a sites can further decrease
the activation barrier for the diffusion of Li+ (3b) ions. 2.3. Electrochemical Analysis. Li-half 2032 coin cells were
Furthermore, a Ni2+ ion at 3b site in the lithium layer may used to test the electrochemical properties. NCA powders
also have a bigger electrostatic interaction (repulsion) with (84 wt%), carbon black (8 wt%), and polyvinylidene fluoride

Downloaded from https://spj.science.org on June 19, 2024

the nearest six Li+ ions, which can increase the energy of (8 wt%) that dispersed in N-methyl-2-pyrrolidone were
the activated lithium ions and thus decrease the activation thoroughly mixed to make slurries. Then, the slurry was
barrier for lithium-ion migration. coated on an aluminum foil and dried at 80°C in an oven
Herein, taking NCA as an example of the Ni-rich cathode overnight. The laminates were punched into round disks
materials, we investigated the relationships between the rate with a diameter of 12 mm. A typical electrode disk contained
capability and the Li+/Ni2+ antisite defects by combining 4 mg NCA. The coin cells were assembled in an Ar-filled
experimental and theoretical calculations. Our experimental glove box (MBraun Labmaster 130) with an electrolyte of
results indicate that the NCA sample with 2.3% Li+/Ni2+ 1 M LiPF6 in ethylene carbonate (EC) and diethyl carbonate
exchange has the highest Li+ diffusion coefficient and exhibits (DEC) (EC DEC = 1 1 V/V) and a separator of Celgard
the best rate capability, and our calculation results reveal that 2400 porous membrane.
the activation barrier for the diffusion of Li+ ions significantly The cells were charged/discharged on a Neware BTS
decreases when there is a minor degree of Li+/Ni2+ antisite 610 multichannel battery cycler at a constant current
defect in the Ni-rich layered oxides. (0.1C-10C, 1C = 180 mAg−1 ) in the voltage range of 2.75-
4.3 V for five cycles, respectively, to investigate the rate per-
2. Materials and Methods formance. In order to calculate the lithium-ion diffusion
coefficient, the cyclic voltammograms (CV) were conducted
2.1. Synthesis of NCA Materials. The NCA samples were on a CHI640 Electrochemical Workstation with a scan rate
synthesized by a thermal polymerization method as of 0.1, 0.3, 0.5, 0.8, and 1.0 mV s-1, respectively, between 2.5
described in our previous studies [24]. Typically, lithium and 4.5 V. And the impedance measurement was also con-
nitrate (LiNO3), nickel nitrate (Ni(NO3)2·6H2O), cobalt ducted, at a fully discharged state, using electrochemical
nitrate (Co(NO3)2·6H2O), and aluminum nitrate (Al(NO3)3· impedance spectroscopy (EIS) in the frequency range of
9H2O) were dissolved in deionized water to prepare the pre- 0.01 Hz-100 kHz. The amplitude of the alternating current
cursor solutions. The molar ratio of Li, Ni, Co, and Al was (AC) signal was 5 mV.
1.00 : 0.8 : 0.15 : 0.05, and the metal (Li+Ni+Co+Al) concentra-
tion was 1.0 mol L-1. Then, acrylic acid (AA, CH2=CHCOOH) 2.4. Computational Method. All calculations were performed
was added into the mixture as a complexing agent, and the by using a Vienna ab initio simulation package (VASP 5.3.3)
volume ratio of AA and H2O is 1 : 2. The mixed solution based on the local density approximation [26]. The interac-
was kept at 160°C for 6 h to form xerogel, followed by thor- tion between ions and valence electrons was described using
oughly grinding and sintering at 500°C in air to remove the projector augmented wave (PAW) potentials [27], and the on-
organics. The obtained precursor powders were again sin- site electron-electron interactions were taken into account by
tered at a tube furnace with flowing oxygen atmosphere for performing GGA+U level calculations with U‐J = 4 91 eV
12 h, and the sintering temperatures were set as 720°C, for Co-3d electrons and U‐J = 6 eV for Ni-3d electrons taken
735°C, 750°C, 765°C, and 780°C, respectively. For conve- from the literatures [28–33]. A 120-atom supercell consisting
nience, the prepared samples were denoted as NCA-720, of 2 × 5 × 1 primitive unit cells was employed, as shown in
NCA-735, NCA-750, NCA-765, and NCA-780, respectively. Figure 1, and all calculations were performed with a plane
wave cutoff of 500 eV. In addition, we found a k-point sam-
2.2. Structure and Morphology Characterizations. Morpho- pling of 5 × 2 × 2 within the Monkhorst-Pack special k-point
logical studies were performed using scanning electron scheme in the Brillouin zone sufficiently dense for the bulk
Research 3

(003)
(003)

(003)
(104)

(104)
(104)
(101)

(101)

(101)
(108/110)

(108/110)
(108/110)
(006/104)

(006/104)
(006/104)
(107)
(015)

(113)

(107)

(107)
(015)

(015)
(113)

(113)
10 20 30 40 50 60 70 10 20 30 40 50 60 70 10 20 30 40 50 60 70
2 theta (degree) 2 theta (degree) 2 theta (degree)
obs obs obs
cal cal cal
diff diff diff
(a) (b) (c)
4.0
(003)

(003)

3.5
3.0

(104)
(104)

2.5
2.0
(101)
(101)

(108/110)
(108/110)

1.45
(006/104)

Downloaded from https://spj.science.org on June 19, 2024

(006/104)

(015)

(107)

(113)
(015)

(107)

(113)

1.40
1.35
1.30
1.25
10 20 30 40 50 60 70 10 20 30 40 50 60 70 720 735 750 765 780
2 theta (degree) 2 theta (degree) Sintering temperature (°C)

obs obs Li-Ni exchange (%)

cal cal I(003)/I(104)
diff diff
(d) (e) (f)

Figure 2: Rietveld refinement of XRD. (a–e) Rietveld analysis of NCA-720, NCA-735, NCA-750, NCA-765, and NCA-780, respectively. (f)
The refinement results of the ratio of Li/Ni exchange and the peak intensity ratio of (003)/(104).

unit cell [34]. The energy relaxation iterates until the forces For the five samples, the average valences of transition
acting on all the atoms were less than 10-2 eV/Å. The climb- metals (Ni and Co) are almost the same according to the
ing image nudged elastic band (CI-NEB) method was used semiquantitative analysis of X-ray photoelectron spectroscopy
to determine the minimum energy paths for diffusion of Li (XPS) (Figures S2 and S3). For the cation disordering, only
atoms in LiNi0.8Co0.2O2 [35–37]. With these settings, we Li+ and Ni2+ occupations are considered to be exchanged
obtained the optimal crystallographic parameters of LiNi0.8- because Co3+ (0.061 nm) and Al3+ (0.054 nm) in the lithium
Co0.2O2, a = 2 79 Å, b = 2 95 Å, and c = 14 09 Å, which are layer would cause a drastically increased system energy due
in good agreement with the experimental values to the big differences in ionic radii and valences with Li+.
(a = 2 87 Å, b = 2 87 Å, and c = 14 13 Å) [38]. The refinement results are given in Table 1 and Figure 2(f),
according to which two crystal diagrams of the TM layer
and the Li layer (seen from c-axis direction) for NCA-750
3. Results
are given in Figure S4 to show the atom arrangements. It is
3.1. Morphology and Structure Characterization. Figure S1 quite clear that the ratio of Li+/Ni2+ exchange is highly
shows the SEM images of the NCA samples sintered at related to the intensity ratio of I 003 /I 104 that has been
different temperatures, which consist of irregular particles used to qualitatively analyze the degree of cation mixing in
around 700 nm-1 μm and without significant differences layered materials. It is widely believed that the lower the
in the morphology and particle size. As indicated in degree of cation disordering, the better the electrochemical
Figures 2(a)–2(e), all their diffraction peaks can be well properties [12, 14–16, 42]. NCA-720 has a relatively high
indexed in a hexagonal structure of α-NaFeO2 type without degree of Li/Ni mixing because the temperature is not high
any impurity [39, 40]. In order to reveal the detailed crystal enough to remove all of the lattice defects. The rest of the
information, the XRD data have been refined by the samples show an approximate linear relationship between
Rietveld method using GSAS software. In the refinement, the ratio of Li/Ni exchange and sintering temperature,
Li1/Ni2 are set at 3b sites (0, 0, 0.5), Li2/Ni1/Co1/Al1 at 3a which may be attributed to the phase transition of LiNiO2
sites (0, 0, 0), and O at 6c sites (0, 0, z) where z = 0 25 [41]. with increasing sintering temperature. It seems that NCA-
4 Research

Table 1: Summary of the Rietveld result of the five NCA samples based on the space group of R-3m.

Samples a (Å) c (Å) c/a χ2 Rwp% Rp% I 003 /I 104 Li/Ni exchange
NCA-720 2.8615 (4) 14.1738 (6) 4.9533 6.11 9.6 7.6 1.37131 3.19%
NCA-735 2.8614 (8) 14.1756 (4) 4.9541 7.629 10.3 7.6 1.45642 1.69%
NCA-750 2.8619 (3) 14.1776 (5) 4.9539 7.394 10.4 7.9 1.44492 2.39%
NCA-765 2.8628 (3) 14.1776 (3) 4.9539 9.188 11.4 8.5 1.41527 3.20%
NCA-780 2.8632 (1) 14.1781 (2) 4.9518 8.351 11 8.1 1.31637 3.82%

1.0
4.4 NCA-720
0.8 NCA-735
4.2 NCA-765
4.0 0.6
NCA-780

Current (mA)
3.8
Voltage (V)

0.4 NCA-750
3.6
0.2
3.4
3.2 0.0
3.0 −0.2
3.8
−0.4
2.6

Downloaded from https://spj.science.org on June 19, 2024

0 40 80 120 160 200 2.5 3.0 3.5 4.0 4.5
Specific capacity (mAhg−1) Voltage (V)
NCA-720 NCA-765
NCA-735 NCA-780
NCA-750
(a) (b)
220
200 0.1C 180
Specific capacity (mAh/g)

0.5C 0.1C
180 1C 150
2C
160 5C
120
−Z″ (𝛺)

140 10C

120 90

100 60
80 30
60
0
40
0 5 10 15 20 25 30 35 40 0 30 60 90 100 150 180
Cycle number Z′ (Ω)
NCA-720 °C NCA-765 °C NCA-720 °C NCA-765 °C
NCA-735 °C NCA-780 °C NCA-735 °C NCA-780 °C
NCA-750 °C NCA-750 °C
(c) (d)

Figure 3: Electrochemical measurements and characterization: (a) the initial charge-discharge profiles; (b) the initial CV curves; (c) rate
capabilities; and (d) Nyquist plots of Li/NCA half cells for the five samples.

735 which has the least Li/Ni exchange in the layered on the sintering temperature. NCA-750 achieves the highest
structure should have exhibited the best electrochemical initial discharge capacity of 187.9 mAh g-1 as well as the high-
properties. However, our electrochemical measurements est initial Coulombic efficiency of 89.5%. The CV curves with
give us some anomalous but interesting results as shown in a scan rate of 0.1 mV s-1 in Figure 3(b) show the redox behav-
the following part. iors of the transition metals as well as the phase transitions
during the charging/discharging process. All the samples
3.2. Electrochemistry. The electrochemical properties of NCA show the similar oxidation of Ni3+ to Ni4+ at around 3.8 V
electrode materials are shown in Figure 3. In Figure 3(a), the but with slight difference in the oxidation peaks which may
charge/discharge profiles are obtained at 0.1C in the range of be caused by the larger polarization for NCA-720 and
2.75-4.3 V. Obviously, the discharge capacity depends highly NCA-765. It should also be noted that NCA-780 has another
Research 5

4 5
Oxidation Oxidation 5 Oxidation
3 4
4
3

Current (mA)

Current (mA)
Current (mA)

2 3
2 2
1 1 1
0 0 0
−1 −1 −1
−2 −2 −2
Reduction −3 Reduction −3 Reduction
2.5 3.0 3.5 4.0 4.5 2.5 3.0 3.5 4.0 4.5 2.5 3.0 3.5 4.0 4.5
Voltage (V) Voltage (V) Voltage (V)
0.1 mVs−1 0.8 mVs−1 0.1 mVs−1 0.8 mVs−1 0.1 mVs−1 0.8 mVs−1
0.3 mVs−1 1.0 mVs−1 0.3 mVs−1 1.0 mVs−1 0.3 mVs−1 1.0 mVs−1
0.5 mVs−1 0.5 mVs−1 0.5 mVs−1
(a) (b) (c)
5
4.0 4 4
3.2 Oxidation Oxidation
3 3
2.4
Current (mA)
Current (mA)

2 2

Ip (mA)
1.6
0.8 1 1 Linear fit of oxidation peak
0.0 0 0
Linear fit of reduction peak
−0.8 −1

Downloaded from https://spj.science.org on June 19, 2024

−1
−1.6 −2
−2.4 Reduction −2 Reduction −3
2.5 3.0 3.5 4.0 4.5 2.5 3.0 3.5 4.0 4.5 0.010 0.015 0.020 0.025 0.030 0.035
Voltage (V) Voltage (V) v1/2 (V1/2 s−1/2)
−1 −1 −1 −1
0.1 mVs 0.8 mVs 0.1 mVs 0.8 mVs NCA-720 NCA-765
0.3 mVs−1 1.0 mVs−1 0.3 mVs−1 1.0 mVs−1 NCA-735 NCA-780
0.5 mVs−1 0.5 mVs−1 NCA-750
(d) (e) (f)

Figure 4: CV measurements. (a–e) CV curves of NCA-720, NCA-735, NCA-750, NCA-765, and NCA-780, respectively; (f) linear fit of the
peak currents and the square root of scanning rates for the five samples.

Table 2: The calculated diffusion coefficient of Li+ for antisite defects in the layered structure is beneficial to the fast
deintercalation and intercalation processes. diffusion of Li+ in the lithium layer.
The lithium-ion diffusion coefficient was calculated based
Samples O‐DLi (cm2 s-1) R‐DLi (cm2 s-1)
on the CV curves with different scan rates (Figure 4) [44]. In
NCA-720 5 79 ∗ 10‐11 2 49 ∗ 10‐11 a diffusion controlled process, the peak current is propor-
NCA-735 1 08 ∗ 10‐10 3 63 ∗ 10‐11 tional to the square root of the scanning rates, which is in line
‐10 with the Randles-Sevcik equation:
NCA-750 1 45 ∗ 10 3 66 ∗ 10‐11
NCA-765 6 54 ∗ 10‐11 3 26 ∗ 10‐11 ip = 2 69 ∗ 105 ∗ n3/2 CLi Av1/2 DLi 1/2 , 1
‐11 ‐11
NCA-780 5 51 ∗ 10 2 16 ∗ 10
where ip is the peak current, n is the number of the electrons
transferred, C Li is the concentration of Li+, A is the surface
area of the electrode, v is the scan rate, and DLi is the diffusion
oxidation peak at around 4.5 V which mainly is attributed to coefficient of lithium ions. As shown in Figures 4(a)–4(e), the
the phase transition from H1 to O1 [43]. As mentioned deintercalation O‐DLi and intercalation R‐DLi are calculated
before, the LiNiO2-based materials decompose at high tem- from the oxidation peak and reduction peak, respectively,
peratures and form the nonstoichiometric Li1‐x NiO2 phase. and the data are summarized in Table 2. The slope of the lin-
And that is the main reason for its lowest capacity in the ear fit of each peak, as shown in Figure 4(f), is positively pro-
voltage range of 2.75-4.3 V. As shown in Figure 3(c), the rate portional to the diffusion coefficient of lithium ions. Clearly,
performances differ significantly with different sintering the diffusion coefficient of the deintercalation process has a
temperatures. NCA-750 shows the best rate capability even higher dependence on temperature, which is consistent with
with a capacity of 129.8 mAh g-1 at 10C rate while it is only the rate performance, than that of the intercalation process.
119.5 mAh g-1 for NCA-735. And the EIS also shows the
results consistent with the rate performance (Figure 3(d)). 3.3. Calculation Results. In order to understand the anoma-
These results and the diffusion coefficient of lithium ions lous experimental result that NCA-750 with higher Li/Ni
(see below) indicate that an appropriate amount of Li+/Ni2+ antisite defects show better electrochemical properties,
6 Research

(b)

(c)

Downloaded from https://spj.science.org on June 19, 2024

(a)

(d)

(e)
O
TM ions

Li+

Figure 5: Structure and diffusion path of Li+ ions of the Ni-rich cathode materials. (a) Layered structure of the Ni-rich cathode materials; the
diffusion of Li+ ions under different conditions: (b) without Li/Ni antisite defects, (c) Li ion occupies TM ion sites, (d) Ni2+ occupies Li+ sites;
(e) simultaneously with two defects near the diffusion pathway.

especially the rate capability, than that of NCA-735 which [23]. And that is the main reason for the activation barrier
contains the lowest antisite defects, we perform the calcu- in layered electrode materials. If the high-valence TM ions
lations from the perspective of activation barrier which are replaced by the Li+ ions, the activation barrier for Li+ dif-
strongly affects the diffusion of lithium ions in electrode fusion decreases due to the weaker electrostatic repulsion
materials. However, due to the limited calculation abilities, (Figure 5(c)). On the other hand, when a high-valence TM
we can only build a 120-atom supercell consisting of 2 × ion occupies Li+ sites, i.e., Li/Ni exchange (Figures 5(d)
5 × 1 LiNi0.8Co0.2O2 primitive unit cells as mentioned and 5(e)), it should also have some effect on the activation
before but not a LiNi0.8Co0.15Al0.05O2 supercell in which barrier for the six nearest lithium ions’ diffusion, which will
the five percent aluminum ions would need a huge amount be discussed in the following part in detail.
of computations. We then calculate the activation barriers in lithium
In a typical layered structure as shown in Figure 5(a), the vacancy environments, i.e., a lithium ion migrates to occupy
lithium ions and transition metal ions alternately occupy a its nearest lithium vacancy in Li slab, as shown in Figure 6.
layer of octahedron sites, which means that there is a two- This is similar to the Li diffusion in LiCoO2 [45]. Firstly,
dimensional transfer pathway for lithium ions. When a lith- we explore Li diffusion in LiNi0.8Co0.2O2 without Li-Ni
ium ion migrates to the nearest lithium vacancy, it has to exchange and find that the activation barriers are approxi-
pass through the adjacent tetrahedron sites, in which the mately 0.69 eV, as shown in Figure 6(a). Then, we consider
neighboring transition metal ion hinders the Li+ diffusion the Li diffusion in LiNi0.8Co0.2O2 with 4.17% Li-Ni exchange,
because of the strong electrostatic repulsion (Figure 5(b)) i.e., one Ni for one Li in the 120-atom supercell. The initial
Research 7

0.80

Relative energy (eV)

0.60 Li vacancy
0.69 eV 0.64 eV
0.40

0.20 a
c
0.00 b
0.00 1.00 2.00 3.00 4.00
Reaction coordinate
(a)

0.80 Ni Li
Relative energy (eV)

0.60 Li vacancy
0.45 eV
0.40 0.64 eV

c
0.20

Downloaded from https://spj.science.org on June 19, 2024

0.00 a b
0.00 1.00 2.00 3.00 4.00
Reaction coordinate
(b)

Li
0.80 Li vacancy Ni
Relative energy (eV)

0.60
0.25 eV 0.60 eV
0.40
c
0.20

b
0.00 a
0.00 1.00 2.00 3.00
Reaction coordinate
(c)

Figure 6: Calculation results of the activation barriers in lithium vacancy environments. (a) Energy profiles for Li diffusion and diffusion path
in LiNi0.8Co0.2O2, (b) energy profiles for Li diffusion in LiNi0.8Co0.2O2 with 4.17% Li-Ni exchange and diffusion path under Ni ion, (c) energy
profiles for Li diffusion in LiNi0.8Co0.2O2 with 4.17% Li-Ni exchange and diffusion path under Li ion. Squares are lithium vacancies, black
arrows represent diffusion path, and the blue double-sided arrows represent Li-Ni exchange.

site of a Li ion is adjacent to a Ni ion in the Li slab, and the Then, we focus on Li diffusion in LiNi0.8Co0.2O2 with dif-
diffusion path is under the Ni ion in Ni(Co) slab, as shown ferent ratios of Li-Ni exchange. Our calculations reveal that
in Figure 6(b). The calculated activation barriers decrease to the activation barrier further decreases with the increase of
be 0.45 eV, indicating that Li-Ni exchange favors the diffu- the ratio of Li-Ni exchange, as depicted in Figure 7(a). From
sion of Li. For this ratio of Li-Ni exchange, we also concern Figure 7(a), we can see that for each concerned ratio of Li-Ni
another case of Li diffusion path; that is, a diffusing Li ion exchange, the activation barrier decreases by 0.2 eV at least
is under an exchanged Li ion in the Ni(Co) slab, as shown when the diffusion path is under a Ni ion, as compared
in Figure 6(c). Surprisingly, the activation barrier decreases to that under a Li ion. Therefore, we can conclude that
to 0.25 eV. Moreover, we further investigate diffusion of Li Li-Ni exchange promotes the diffusion possibilities of Li
at the next neighboring Ni ion in the Li slab and find that in LiNi0.8Co0.2O2.
the activation barriers are almost the same as those in LiNi0.8- To understand this interesting phenomenon above, we
Co0.2O2 without Li-Ni exchange. This means that the Ni ions investigate the electronic properties of the system. The charge
in the Li slab just influence the diffusion of Li ions within the distribution for the states at the saddle points of the diffusion
scope of the neighboring region. paths for the typical cases is displayed in Figures 7(b)–7(d).
8 Research

0.70

Activation barrier (eV) 0.60

0.50

0.40

0.30
Ni Li
0.20

0.10

0.00
0.00% 4.17% 8.33% 12.50% 16.67%
Li-Ni exchange
Path under Ni ion Saddle point
Path under Li ion
(a) (b)

Downloaded from https://spj.science.org on June 19, 2024

Ni Li Ni Li Ni
Ni

Saddle point Saddle point

(c) (d)

Figure 7: Calculation results of the activation barriers in Li-Ni exchange environments. (a) Calculated activation barrier for Li migration with
different Li-Ni exchanges under Li or Ni ion, (b) charge distribution for Li diffusion in LiNi0.8Co0.2O2 without Li-Ni exchange, (c) charge
distribution for Li diffusion in LiNi0.8Co0.2O2 with 4.17% Li-Ni exchange and diffusion path under Ni ion, and (d) charge distribution for
Li diffusion in LiNi0.8Co0.2O2 with 4.17% Li-Ni exchange and diffusion path under Li ion.

3.0
cies, and it reduces the activation barriers of Li diffusion
2.5 around the Ni ion. This is why the Ni ions in Li slab can only
Relative energy (eV)

2.0 alter the diffusion of Li ions within the scope of the neighbor-
ing region. For the same reason, when the Li diffusion path is
1.5
under a Li ion, the above Li ion will reduce the interactions
1.0 with surrounding O ions and then the distance between O
0.5 ions and Li diffusion path, which results in the further
decrease of activation barriers.
0.0
However, as shown in Figure 8, if there are too many Li-
0.00% 4.17% 8.33% 12.50% 16.67% Ni antisite defects in Ni-rich layered materials, the total sys-
Ratio of Li/Ni exchange tem energy drastically increases, which means that the lay-
ered structure becomes quite unstable with the increasing
Figure 8: Change of total energy with different amounts of Li-Ni Li/Ni antisite defects. Therefore, it seems impossible to syn-
exchange in the supercell.
thesize the layered structure materials with so many defects
under the mentioned experimental conditions. On the other
We can see clearly from Figure 7(b) that the interactions hand, the excessive inactive TM ions in Li slab would inevita-
between Li and O ions are weak while the interactions bly cause a longer diffusion path for lithium ions. Therefore,
between Ni and O ions are strong in LiNi0.8Co0.2O2. When the highest rate of Li+ diffusion can be achieved when the
Li-Ni exchange occurs with a Ni ion appearing in the Li slab, benefit from the decrease of activation barriers and hin-
stronger interactions between Ni ion in the Li slab and O ions drance reach a balance, which corresponds to the experimen-
drive Li ions near this Ni ion diffuse to its neighboring vacan- tal result of 2.39% for NCA-750 in our study.
Research 9

4. Discussion based batteries at elevated temperatures,” Journal of Materials

Chemistry A, vol. 3, no. 8, pp. 4092–4123, 2015.
In this work, our experimental results show that, unlike [3] C. V. Rao, A. L. M. Reddy, Y. Ishikawa, and P. M. Ajayan,
the widely assumed, the layered Ni-rich material with an “LiNi1/3Co1/3Mn1/3O2–graphene composite as a promising
appropriate amount of Li/Ni antisite defects exhibits the best cathode for lithium-ion batteries,” ACS Applied Materials &
electrochemical properties. By building a 120-atom supercell Interfaces, vol. 3, no. 8, pp. 2966–2972, 2011.
consisting of 2 × 5 × 1 LiNi0.8Co0.2O2 primitive unit cells, we [4] J. Li, R. Yao, and C. Cao, “LiNi1/3Co1/3Mn1/3O2 nanoplates with
have investigated the relationships between the Li/Ni antisite {010} active planes exposing prepared in polyol medium as a
defects and activation barrier for lithium ions’ diffusion. high-performance cathode for Li-ion battery,” ACS Applied
When a Ni2+ ion occupies a Li+ site, the stronger interactions Materials & Interfaces, vol. 6, no. 7, pp. 5075–5082, 2014.
between Ni and O drive the nearest Li+ ion diffuse to the [5] H. Kim, M. G. Kim, H. Y. Jeong, H. Nam, and J. Cho, “A new
neighboring vacancies. In other words, high-valence TM ions coating method for alleviating surface degradation of LiNi0.6-
can decrease the activation barrier for the diffusion of lithium Co0.2Mn0.2O2 cathode material: nanoscale surface treatment
ions, resulting in the improved rate capabilities. And the of primary particles,” Nano Letters, vol. 15, no. 3, pp. 2111–
activation barrier can be further decreased when there is a 2119, 2015.
synergy between Ni (3b sites) and Li (3a sites), i.e., Li/Ni [6] J. H. Lee, C. S. Yoon, J. Y. Hwang et al., “High-energy-density
exchange. However, a large amount of Li/Ni antisite defects lithium-ion battery using a carbon-nanotube–Si composite
cause the drastic increase of system energy and make the anode and a compositionally graded Li[Ni0.85Co0.05Mn0.10]O2
layered structure unstable. We expect that these interesting cathode,” Energy & Environmental Science, vol. 9, no. 6,
pp. 2152–2158, 2016.
findings offer some opportunities to design the layered
cathode materials for advanced lithium-ion batteries. [7] P. Zhou, H. Meng, Z. Zhang et al., “Stable layered Ni-rich
LiNi0.9Co0.07Al0.03O2 microspheres assembled with nanoparti-

Downloaded from https://spj.science.org on June 19, 2024

cles as high-performance cathode materials for lithium-ion
Conflicts of Interest batteries,” Journal of Materials Chemistry A, vol. 5, no. 6,
pp. 2724–2731, 2017.
The authors declare no competing financial interests. [8] Z. Tang, J. Bao, Q. Du et al., “Surface surgery of the nickel-rich
cathode material LiNi0.815Co0.15Al0.035O2: toward a complete
Authors’ Contributions and ordered surface layered structure and better electrochem-
ical properties,” ACS Applied Materials & Interfaces, vol. 8,
Zhongfeng Tang and Sen Wang contributed equally to this no. 50, pp. 34879–34887, 2016.
work. [9] F. Wu, J. Tian, Y. Su et al., “Effect of Ni2+ content on lithium/-
nickel disorder for Ni-rich cathode materials,” ACS Applied
Materials & Interfaces, vol. 7, no. 14, pp. 7702–7708, 2015.
Acknowledgments [10] C. Fu, G. Li, D. Luo, Q. Li, J. Fan, and L. Li, “Nickel-rich lay-
ered microspheres cathodes: lithium/nickel disordering and
This work was supported by the National Key R&D Program
electrochemical performance,” ACS Applied Materials & Inter-
of China (grant no. 2018YFB0905400) and the National Sci- faces, vol. 6, no. 18, pp. 15822–15831, 2014.
ence Foundation of China (NSAF U1630106, grant no.
[11] A. Hiranoa, R. Kannoa, Y. Kawamotoa et al., “Relationship
51577175). We are also grateful to Elementec Ltd. in Suzhou
between non-stoichiometry and physical properties in
for its technical support. LiNiO2,” Solid State Ionics, vol. 78, no. 1-2, pp. 123–131, 1995.
[12] G. Hu, W. Liu, Z. Peng, K. Du, and Y. Cao, “Synthesis and elec-
Supplementary Materials trochemical properties of LiNi0.8Co0.15Al0.05O2 prepared from
the precursor Ni0.8Co0.15Al0.05OOH,” Journal of Power
Figure S1: SEM images of (a) NCA-720, (b) NCA-735, (c) Sources, vol. 198, pp. 258–263, 2012.
NCA-750, (d) NCA-765, and (e) NCA-780, respectively. [13] V. Bianchi, S. Bach, C. Belhomme et al., “Electrochemical
Figure S2: XPS spectra of (a) survey spectrum and Ni 2p investigation of the Li insertion–extraction reaction as a func-
for (b) NCA-720, (c) NCA-735, (d) NCA-750, (e) NCA- tion of lithium deficiency in Li1−xNi1+xO2,” Electrochimica
765, and (f) NCA-780, respectively. Figure S3: XPS spectra Acta, vol. 46, no. 7, pp. 999–1011, 2001.
of Co 2p for (a) NCA-720, (b) NCA-735, (c) NCA-750, (d) [14] Y. Ding, D. Mu, B. Wu, R. Wang, Z. Zhao, and F. Wu, “Recent
NCA-765, and (e) NCA-780, respectively. Figure S4: crystal progresses on nickel-rich layered oxide positive electrode
diagrams of the TM layer (a) and Li layer (b) for NCA-750. materials used in lithium-ion batteries for electric vehicles,”
(Supplementary Materials) Applied Energy, vol. 195, pp. 586–599, 2017.
[15] A. Manthiram, J. C. Knight, S. T. Myung, S. M. Oh, and Y. K.
References Sun, “Nickel-rich and lithium-rich layered oxide cathodes:
progress and perspectives,” Advanced Energy Materials,
[1] R. Amisse, M. T. Sougrati, L. Stievano et al., “Singular struc- vol. 6, no. 1, article 1501010, 2016.
tural and electrochemical properties in highly defective [16] F. Schipper, E. M. Erickson, C. Erk, J. Y. Shin, F. F. Chesneau,
LiFePO4 powders,” Chemistry of Materials, vol. 27, no. 12, and D. Aurbach, “Review—Recent advances and remaining
pp. 4261–4273, 2015. challenges for lithium ion battery cathodes I. nickel-rich,
[2] G. Xu, Z. Liu, C. Zhang, G. Cui, and L. Chen, “Strategies for LiNixCoyMnzO2,” Journal of the Electrochemical Society,
improving the cyclability and thermo-stability of LiMn2O4- vol. 164, no. 1, pp. A6220–A6228, 2017.
10 Research

[17] J. Zheng, T. Liu, Z. Hu et al., “Tuning of thermal stability in [34] H. J. Monkhorst and J. D. Pack, “Special points for brillouin-
layered Li(NixMnyCoz)O2,” Journal of the American Chemical zone integrations,” Physical Review B, vol. 13, no. 12,
Society, vol. 138, no. 40, pp. 13326–13334, 2016. pp. 5188–5192, 1976.
[18] Y. Xiao, T. Liu, J. Liu et al., “Insight into the origin of lithium/- [35] G. Mills, H. Jónsson, and G. K. Schenter, “Reversible work
nickel ions exchange in layered Li(NixMnyCoz)O2 cathode transition state theory: application to dissociative adsorption
materials,” Nano Energy, vol. 49, pp. 77–85, 2018. of hydrogen,” Surface Science, vol. 324, no. 2-3, pp. 305–337,
[19] W. Lee, S. Muhammad, T. Kim et al., “New insight into Ni- 1995.
rich layered structure for next-generation Li rechargeable [36] G. Henkelman, B. P. Uberuaga, and H. Jonsson, “A climbing
batteries,” Advanced Energy Materials, vol. 8, no. 4, article image nudged elastic band method for finding saddle points
1701788, 2018. and minimum energy paths,” The Journal of Chemical Physics,
[20] H. Xie, K. Du, G. Hu, Z. Peng, and Y. Cao, “The role of sodium vol. 113, no. 22, pp. 9901–9904, 2000.
in LiNi0.8Co0.15Al0.05O2 cathode material and its electrochem- [37] G. Henkelman and H. Jonsson, “Improved tangent estimate in
ical behaviors,” The Journal of Physical Chemistry C, vol. 120, the nudged elastic band method for finding minimum energy
no. 6, pp. 3235–3241, 2016. paths and saddle points,” The Journal of Chemical Physics,
[21] Z. Y. Li, R. Gao, J. Zhang, X. Zhang, Z. Hu, and X. Liu, “New vol. 113, no. 22, pp. 9978–9985, 2000.
insights into designing high-rate performance cathode mate- [38] L. Xiao, Y. Yang, Y. Zhao, X. Ai, H. Yang, and Y. Cao, “Synthe-
rials for sodium ion batteries by enlarging the slab-spacing of sis and electrochemical properties of submicron LiNi0.8-
the Na-ion diffusion layer,” Journal of Materials Chemistry A, Co0.2O2 by a polymer-pyrolysis method,” Electrochimica
vol. 4, no. 9, pp. 3453–3461, 2016. Acta, vol. 53, no. 6, pp. 3007–3012, 2008.
[22] Z. Huang, Z. Wang, Q. Jing, H. Guo, X. Li, and Z. Yang, [39] Y. Makimura, T. Sasaki, T. Nonaka et al., “Factors affecting
“Investigation on the effect of Na doping on structure and cycling life of LiNi0.8Co0.15Al0.05O2 for lithium-ion batteries,”
Journal of Materials Chemistry A, vol. 4, no. 21, pp. 8350–

Downloaded from https://spj.science.org on June 19, 2024

Li-ion kinetics of layered LiNi0.6Co0.2Mn0.2O2 cathode mate-
rial,” Electrochimica Acta, vol. 192, pp. 120–126, 2016. 8358, 2016.
[23] K. Kang, Y. S. Meng, J. Bréger, C. P. Grey, and G. Ceder, “Elec- [40] H. Z. Zhang, C. Liu, D. W. Song, L. Q. Zhang, and L. J. Bie, “A
trodes with high power and high capacity for rechargeable lith- new synthesis strategy towards enhancing the structure and
ium batteries,” Science, vol. 311, no. 5763, pp. 977–980, 2006. cycle stabilities of the LiNi0.80Co0.15Al0.05O2 cathode material,”
[24] Q.-X. Du, Z.-F. Tang, X.-H. Ma et al., “Improving the electro- Journal of Materials Chemistry A, vol. 5, no. 2, pp. 835–841,
chemical properties of high-energy cathode material LiNi0.5- 2017.
Co0.2Mn0.3O2 by Zr doping and sintering in oxygen,” Solid [41] F. Schipper, M. Dixit, D. Kovacheva et al., “Stabilizing nickel-
State Ionics, vol. 279, pp. 11–17, 2015. rich layered cathode materials by a high-charge cation doping
[25] B. H. Toby, “EXPGUI, a graphical user interface for GSAS,” strategy: zirconium-doped LiNi0.6Co0.2Mn0.2O2,” Journal of
Journal of Applied Crystallography, vol. 34, no. 2, pp. 210– Materials Chemistry A, vol. 4, no. 41, pp. 16073–16084, 2016.
213, 2001. [42] W. Liu, P. Oh, X. Liu et al., “Nickel-rich layered lithium
[26] G. Kresse and J. Furthmüller, “Efficient iterative schemes for transition-metal oxide for high-energy lithium-ion batteries,”
ab initio total-energy calculations using a plane-wave basis Angewandte Chemie International Edition, vol. 54, no. 15,
set,” Physical Review B, vol. 54, no. 16, pp. 11169–11186, 1996. pp. 4440–4457, 2015.
[27] P. E. Blöchl, “Projector augmented-wave method,” Physical [43] K. Chang, B. Hallstedt, and D. Music, “Thermodynamic
Review B, vol. 50, no. 24, pp. 17953–17979, 1994. description of the LiNiO2–NiO2 pseudo-binary system and
extrapolation to the Li(Co,Ni)O2–(Co,Ni)O2 system,” Cal-
[28] J. P. Perdew, K. Burke, and M. Ernzerhof, “Generalized gradi-
phad, vol. 37, pp. 100–107, 2012.
ent approximation made simple,” Physical Review Letters,
vol. 77, no. 18, pp. 3865–3868, 1996. [44] B. K. Zou, H. Y. Wang, Z. Y. Qiang et al., “Mixed-carbon-
coated LiMn0.4Fe0.6PO4 nanopowders with excellent high rate
[29] S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys,
and low temperature performances for lithium-ion batteries,”
and A. P. Sutton, “Electron-energy-loss spectra and the struc-
Electrochimica Acta, vol. 196, pp. 377–385, 2016.
tural stability of nickel oxide: an LSDA+U study,” Physical
Review B, vol. 57, no. 3, pp. 1505–1509, 1998. [45] A. van der Ven, G. Ceder, M. Asta, and P. D. Tepesch, “First-
principles theory of ionic diffusion with nondilute carriers,”
[30] D. Kramer and G. Ceder, “Tailoring the morphology of
Physical Review B, vol. 64, no. 18, article 184307, 2001.
LiCoO2: a first principles study,” Chemistry of Materials,
vol. 21, no. 16, pp. 3799–3809, 2009.
[31] F. Zhou, M. Cococcioni, C. A. Marianetti, D. Morgan, and
G. Ceder, “First-principles prediction of redox potentials in
transition-metal compounds with LDA+U,” Physical Review
B, vol. 70, no. 23, article 235121, 2004.
[32] H. Moriwake, A. Kuwabara, C. A. J. Fisher et al., “First-princi-
ples calculations of lithium-ion migration at a coherent grain
boundary in a cathode material, LiCoO2,” Advanced Materials,
vol. 25, no. 4, pp. 618–622, 2013.
[33] Y. Kim, H. Lee, and S. Kang, “First-principles and experimen-
tal investigation of the morphology of layer-structured LiNiO2
and LiCoO2,” Journal of Materials Chemistry, vol. 22, no. 25,
pp. 12874–12881, 2012.