Article Cite This: Chem. Mater. 2018, 30, 1808−1814
pubs.acs.org/cm
Cation Ordering of Zr-Doped LiNiO2 Cathode for Lithium-Ion Batteries Chong S. Yoon,† Min-Jae Choi,‡ Do-Wook Jun,‡ Qian Zhang,§ Payam Kaghazchi,§ Kwang-Ho Kim,∥ and Yang-Kook Sun*,‡ †
Department of Materials Science and Engineering, Hanyang University, Seoul 04763, South Korea Department of Energy Engineering, Hanyang University, Seoul 04763, South Korea § Physikalische und Theoretische Chemie, Freie Universität, Berlin D-14195, Germany ∥ School of Materials Science and Engineering, Pusan National University, Busan 46241, South Korea ‡
S Supporting Information *
ABSTRACT: Ordered occupation of Ni ions in the Li ion layer (and vice versa) was observed in 0.4 mol % Zr-doped LiNiO2 with R3̅m structure. Although cation mixing is prevalent in LiNiO2 and in other Ni-rich layered cathodes, cation ordering (Li and Ni) has not been previously reported in the as-prepared or fully discharged states. Firstprinciples calculations verified that low-level doping of LiNiO2 with Zr can energetically promote the observed cation ordering. Contrary to previous beliefs, antisite defects (or cation mixing), whose presence is unequivocally justified by the cation ordering, had hardly any negative effect on the electrochemical performance of LiNiO2; both pristine and Zr-doped LiNiO2 delivered 247.5 and 246.5 mAh g−1, respectively, with a Coulombic efficiency of 97%. The capacity retention after 100 cycles was improved by increasing Zr doping to 81% from 74%. The improved cycling stability was attributed to the particle morphology being conducive to Li migration and relieving the deeply charged LixNiO2 of its inherent structural instability.
■
antisite defects for LiNiO2, especially at high temperatures.9 It is presumed that such a structural disorder, even in small amounts, can deteriorate the electrochemical properties of LiNiO2.6 Arai et al. showed that ∼10% Ni substitution of Li sites reduced the discharge capacity of LiNiO2 by 40%; this was ascribed to displaced Ni ions disturbing the Li+ migration.7 However, the detrimental effect of cation mixing has not been explicitly demonstrated, partially due to the uncertainty in accurately estimating the level of cation mixing; in fact, Rietveld refinement has revealed that Li[Ni1/3Co1/3Mn1/3]O2 and Li[Ni0.5Mn0.5]O2, whose electrochemical properties have been well established over the years, contain antisite defects of 6%10 and 12%,11 respectively, at the Li sites. Hence, contrary to earlier propositions, it is speculated that antisite defects may not be solely responsible for the loss of the reversible capacity of LiNiO2. Here, it is shown that low-level doping of LiNiO2 with Zr can induce ordered interchange of Li and Ni ions, which unequivocally proves the occupation of Li sites by Ni ions, and that the cation ordering does not adversely affect the reversible capacity of LiNiO2. Although Li/vacancy ordering was reported in partially delithiated LixNiO2 (x < 0.75),12 cation ordering has not been previously observed in
INTRODUCTION Lithium-ion batteries (LIBs) have become ubiquitous power sources for a variety of portable electronic devices and small home appliances but are struggling to achieve the same level of success in electrification of personal transportation. Electric vehicles (EVs) powered by LIBs have a relatively low cost to drive range ratio, but further improvements in this aspect are primarily limited by the energy density and the cost of the cathodes. This has been the main obstacle to a wider consumer acceptance of EVs. LiNiO2, which offers a material cost advantage and higher theoretical capacity in comparison to LiCoO2, was actively studied during the 1990s for its potential as a high-density cathode for LIBs;1 however, its intrinsic cycling instability prevented commercial applicability. To address the EV’s requirement for a high energy density LIB, the Ni fraction in Li[Ni1−x−yCoxAly]O2 or Li[Ni1−x−yCoxMny]O2 was progressively increased to gain extra discharge capacity.2−4 In the process, LiNiO2 resurfaced as a likely candidate for the cathode with a potential to fulfill a drive range of 300 miles between charges, which is considered the threshold for a successful commercialization of EVs.5 LiNiO2 with a space group of R3̅m is known to be susceptible to antisite defects, i.e., occupation of Ni ions at the Li sites,6 as high temperature treatment typically decomposes LiNiO2 to Li1−xNi1+xO2 with excess Ni ions substituted at Li sites.7,8 Firstprinciples calculations have also predicted a high density of © 2018 American Chemical Society
Received: February 9, 2018 Revised: February 14, 2018 Published: February 14, 2018 1808
DOI: 10.1021/acs.chemmater.8b00619 Chem. Mater. 2018, 30, 1808−1814
Article
Chemistry of Materials stoichiometric LiNiO2. Using electron diffraction and density functional calculations, we demonstrate that doping LiNiO2 with Zr leads to cation ordering and, moreover, does not deteriorate its electrochemical properties.
■
EXPERIMENTAL SECTION
Synthesis of Pristine LiNiO2. Spherical Ni(OH)2 precursors were synthesized via the coprecipitation method with NiSO4·6H2O (SAMCHUN Chemical, Korea) aqueous solution as a starting material. Detailed synthesis procedures and conditions are provided in a previous paper.13 Synthesis of Zr-Doped LiNiO2. Spherical Zr-doped Ni(OH)2 precursors were synthesized via the coprecipitation method with mixed NiSO4·6H2O and Zr(SO4)2·4H2O (Alfa Aesar, >98%) aqueous solution in the molar ratio Ni:Zr = 99.5:0.5 as a starting material. Detailed synthesis procedures and conditions are provided in a previous paper.13 The Ni(OH)2 and Zr-doped LiNiO2 precursors were mixed with LiOH·H2O (Sigma-Aldrich, >98.5%) (Li:(Ni + Co + Mn) = 1.01:1 molar ratio) and calcined under an oxygen atmosphere for 10 h at 650 °C. Analytical Techniques. The chemical compositions of the prepared powders were determined by inductively coupled plasma (ICP-OES, OPIMA 8300, PerkinElmer). The crystalline phase was analyzed by powder X-ray diffraction (XRD) (Rigaku, Rint-2000) using Cu Kα radiation. XRD data were obtained at 2θ between 10 and 110° with 0.03° step/min. The structural refinement was performed by Rietveld analysis using the Fullprof suite.14 The morphologies and structures of the prepared particles were observed by scanning electron microscopy (SEM, JSM 6400, JEOL) and transmission electron microscopy (TEM, JEOL 2010, JEOL). TEM samples were prepared by focused ion beam (FIB, NOVA 200/FEI), and element mapping was carried out using a JEOL Model JEM 2100F instrument. Electrochemical Testing. For the fabrication of positive electrodes, the synthesized powders were mixed with carbon black (superP:KS-6 = 4:6 wt %) and polyvinylidene fluoride (PVdF, Solef 5130) (85:10:5 wt %) in N-methylpyrrolidone (NMP, Deajungchem, Korea). The obtained slurry was coated on to Al foil (UACJ, Japan), rollpressed, and vacuum-dried. The electrolyte solution was 1.2 M LiPF6 in ethylene carbonate−ethyl methyl carbonate (EC:EMC = 3:7 by vol %) with 2 wt % vinylene carbonate (VC) (PanaX Etec, Korea). Preliminary cell tests were performed with a 2032 coin-type half-cell using lithium metal (Honjo metal, Japan) as the anode. The cells were charged and discharged by applying a constant specific current density of 90 mA g−1 (0.5 C rate) at 30 °C between 2.7 and 4.3 V versus Li+/ Li. Electrochemical impedance spectroscopy was investigated after charging to 4.3 V versus Li+/Li as a function of cycles with a multichannel potentiostat (Bio-Logic, VMP3) from 1 MHz to 1 mHz. Theoretical Modeling. Density functional theory (DFT) calculations have been carried out using the projector-augmented plane-wave (PAW) method as implemented in the Vienna ab initio simulation package (VASP).15,16 We used Perdew−Burke−Ernzerhof17 + Hubbard U parameter exchange-correlation functional18 with U − J = 3.0 and 4.0 eV for Ni and Zr, respectively.17 A 8 × 8 × 2 Monkhorst−Pack k-point mesh for a 1 × 1 × 1 unit cell and an energy cutoff of 450 eV were used. The convergence criteria for total energy were set at 10−4 eV.
Figure 1. SEM images of the (a) pristine and (b) Zr-doped LiNiO2 cathode particles.
primary particles that were compactly packed into a spherical geometry. Lattice parameters calculated from the XRD spectrum of the pristine LiNiO2 with a space group of R3m ̅ were a = 2.8737 Å and c = 14.1952 Å, which matched well the literature values reported for LiNiO2 with excess Li.7,19 The unit cell was slightly enlarged by Zr doping, as the lattice parameters for Zr-doped LiNiO2 were a = 2.8771 Å and c = 14.2038 Å (Figure 2). The
Figure 2. Rietveld refinement results for XRD data of (a) LiNiO2 and (b) Zr-LiNiO2 cathodes.
cation mixing estimated from the Rietveld refinement was approximately 2% for both cathodes. Dark field STEM images of the Zr-doped LiNiO2 in Figure 3a,b disclose the internal
Figure 3. (a, b) STEM images of a single particle from the Zr-doped LiNiO2 cathode in the as-prepared state. (c) Bright field TEM image of an elongated primary particle from the Zr-doped LiNiO2 cathode.
structure. The pore-free particle core was composed of equiaxed crystals, while thin elongated primary particles aligned along the radial directions were densely packed near the surface (Figure 3b). The elongated primary particle was ∼200 nm in width, and its length extended up to 1−2 μm, as can be seen from a conventional bright field TEM image in Figure 3c. Bright field TEM (Figure 4a) and the corresponding [010] zone electron diffraction pattern (Figure 4b) demonstrate a strong crystallographic texture of the needle-shaped primary particle. Its growth direction was along the a axis, and the particle was oriented parallel to the radial direction such that the layer planes of LiNiO2 are aligned in the radial direction. A similar crystallographic texture, observed in the compositionally graded NCM cathodes, was found to be beneficial to Li
■
RESULTS AND DISCUSSION The chemical composition of the as-prepared LiNiO2, determined by ICP, was Li1.04Ni0.96O2, which had a slight excess of Li, as did the Zr-doped LiNiO2, whose estimated composition was Li1.06Ni0.937Zr0.04O2 with a Zr composition of 0.4 mol %, slightly less than the designed composition of 0.5 mol %. SEM images of the pristine and Zr-doped LiNiO2 particles are shown in Figure 1. Both cathodes had similar spherical morphology with an average particle diameter of 10− 11 μm. Each cathode particle was composed of nanosized 1809
DOI: 10.1021/acs.chemmater.8b00619 Chem. Mater. 2018, 30, 1808−1814
Article
Chemistry of Materials
consistent with the superlattice peaks observed in Figure 4b. Incidentally, the superlattice peaks (namely (1/201̅) and (1/ 202)) exactly matched the extra peaks observed by Peres et al. in the monoclinic Li0.63Ni1.02O2 due to the Li/vacancy ordering if their diffraction patterns were to be interpreted as an R3̅m structure.12 In addition to the diffraction patterns shown in Figure 4, the superlattice diffraction patterns were observed from nearly all of the primary particles near the surface of the Zr-doped LiNiO2, while such extra spots were not observed in the pristine LiNiO2 (see Figures S1 and S2). Therefore, the cation ordering in the Zr-doped LiNiO2 was not localized but was prevalent in the particles at least near the surface. With no appreciable concentration of cation vacancies in the as-prepared Zr-doped LiNiO2 (as verified by ICP and XRD refinement), the coincidence of the extra peaks with those observed in the Li/ vacancy ordering substantiates the Li/Ni ordering in the Zrdoped LiNiO2. Although ordering of cations within the transition-metal layer has been reported in a layered cathode (Li[Ni0.5Mn0.5]O2),21 there has been no previous report on interchange of Li ions with transition-metal ions in an ordered manner. Similar to Li0.63Ni1.02O2 in which Li/vacancy ordering was prevalent, many twinned crystals were observed in the Zrdoped LiNiO2. A [010] zone electron diffraction pattern from the twinned crystals is shown in Figure 5a. Paired diffraction spots in the pattern can be explained by the [010] zone spots mirror-reflected about the (001) plane to create a set of [010̅ ] zone diffraction pattern as illustrated. Pairs of extra peaks resulting from the cation ordering are also visible in the twinned diffraction pattern. A schematic diagram of the twinned crystals showing mirror imaging of (102̅) planes in Figure 5b clarifies the twinned structure having the (001) plane as its twin boundary. The appearance of twinned crystals in both Zr-doped LiNiO2 and Li0.63Ni1.02O2 may not be purely coincidental, as twinning was rarely reported in LiNiO2 or related compounds. Hence, it is speculated that twin formation may have been facilitated by the cation ordering ,as was the case for the Li/vacancy ordered Li0.63Ni1.02O2. Since the cation ordering was not observed in pristine LiNiO2, it appears that the presence of Zr ions in LiNiO2 promoted the cation ordering. To support this experimental finding, density functional theory calculations were used to estimate the formation energy needed for interchanging Li and Ni ions, assuming each Li and Ni layer formed an equivalent supercell structure consisting of [Li0.75Ni0.25]3a[Ni0.75Li0.25]3bO2 (Figure
Figure 4. (a) Bright field TEM image of a primary particle from the Zr-doped LiNiO2 cathode (b) Corresponding selected area electron diffraction pattern in the [010] zone from the marked region in (a) (yellow circles indicate the extra peaks that should have been extinct from the normal layered structure) (c) Schematic drawing of the Li plane in which Ni ions are substituted to form a superlattice. (d) [22̅1] zone electron diffraction from a different primary particle in the Zrdoped LiNiO2 cathode, also showing superlattice peaks.
removal/insertion.20 It appears that Zr doping may have altered the surface energy of LiNiO2 in a way that caused a highly anisotropic morphology to develop. Even more peculiar was the appearance of extra peaks (circled in yellow in Figure 4b) that cannot be indexed in the framework of the R3̅m space group. Two extra diffraction spots, marked by red arrows in Figure 4b and indexed as (1/201̅) and (1/202), are only possible from a superlattice as described in Figure 4c. A Li plane in the R3̅m structure composed of alternating fully Li occupied and half Li occupied rows will generate a cation-ordered supercell whose dimensions are twice the original lattice parameters of LiNiO2. To verify the superlattice peaks, a [22̅1] zone electron diffraction pattern was obtained from another primary particle (Figure 4d). This pattern also contained well-defined extra peaks. The marked spots indexed as (1̅/201) and (01̅/21̅) were
Figure 5. (a) Electron diffraction pattern from a twinned primary particle from the Zr-doped LiNiO2 with magnified pattern explaining the paired diffraction spots due to the mirror reflection of the [010] zone pattern about the [100] twin plane. (b) Schematic drawing of the atomic arrangement in the twinned region that produced the twin diffraction pattern. 1810
DOI: 10.1021/acs.chemmater.8b00619 Chem. Mater. 2018, 30, 1808−1814
Article
Chemistry of Materials 6). We have considered two possible distributions for Zr dopants. The corresponding atomic structures are presented in
Figure 6. Atomic structures and formation energies of [Li0.75Ni0.25]3a[Ni0.75Li0.25]3bO2 in the absence (left) and presence (right) of Zr dopant.
Figure S3. The interchange energy per Ni(Li) ion for pristine [Li0.75Ni0.25]3a[Ni0.75Li0.25]3bO 2 is 0.20 eV, whereas the corresponding interchange energies for the doped case with Zr placed on the Li site were 0.14 and 0.12 eV for dopant concentrations of 2.08% and 6.25%, respectively. The Zr ion was placed on the Li site because of the similar ionic radii of Li+ (79 pm) and Zr4+ (80 pm). The interchange energy was, however, still lower for the doped structure in comparison to the pristine case even if the Zr ion was placed at the Ni site. That the Zr-induced Li−Ni interchange leads to the cationordered structure was further verified by calculating the total energies of several possible configurations of [Li0.75Ni0.25]3a[Ni0.75Li0.25]3bO2 with disordered arrangements of Ni and Li cations (see Figure S4). The total energies calculated for almost all of the disordered arrangements were higher than that for the ordered state. Only one configuration is 16 meV more favorable than the ordered structure, and this is due to energy gain from interchanging a nearby Ni3+ cation by a Li+ cation of smaller charge. This short-range effect is limited to cations close to the dopant and therefore cannot change the long-range ordering of the doped structure. Calculated lattice parameters for different percentages of Li−Ni interchange (Figure 7) indicate that both a and c parameters shrink with higher degrees of Li−Ni interchange in pristine LiNiO2. In the case of the a parameter, however, the value plateaus beyond 12.5%. Meanwhile, the c parameter decreases linearly for all the studied values of Li−Ni interchange concentration. Moreover, we find that the relative contraction of the a parameter in percentage value is larger than that of c. Figure 7 also shows that increasing the concentration of Zr (on the Li sites) expands the lattice dimensions. We have considered two possible distributions of dopants: (i) the Zr−Zr separation is large in the a−b plane but small in the c direction (Figure 7a,b) and (ii) Zr−Zr separation is small in the a−b plane but large in the c direction (Figure 7c,d). The experimental XRD data indicate that the lattice constants of both pristine and Zr-doped LiNiO2 were similar and that the concentration of the Li−Ni interchange determined from the Rietveld refinement was relatively small (∼2%). To satisfy these two conditions, we have chosen [Li0.98Ni0.02]3a[Ni0.98Li0.02]3bO2 and [Zr0.0625Li0.6875Ni0.25]3a[Ni0.75Li0.25]3bO2 for calculating respective delithiation energies. The Zr concentration should be much higher than the studied composition to match the lattice
Figure 7. Calculated lattice parameters for nondoped [Li1−xNix]3a[Ni1−xLix]3bO2 with different Ni−Li interchange concentrations (x in %) labeled by Pristine LNO and Zr-doped [ZrxLi0.75‑xNi0.25]3a[Ni0.75Li0.25]3bO2 with different Zr dopant concentrations (x in %) labeled by Zr-doped LNO.
parameters and the low level of cation mixing determined experimentally. The local Zr concentration could be high due to the segregation of Zr near the particle surface, where the cation ordering was mainly observed. The calculated lattice parameters for [Li 0 . 9 8 Ni 0 . 0 2 ] 3 a [Ni 0 . 9 8 Li 0 . 0 2 ] 3 b O 2 and [Zr0.0625Li0.6875Ni0.25]3a[Ni0.75Li0.25]3bO2 are a = 2.92 Å, c = 14.19 Å and a = 2.91 Å, c = 14.20 Å, respectively. The difference between the calculated and experimental values of 1.38%, −0.07% and 1.04%, 0.07% is due to the approximate nature of corresponding experimental values of the exchangecorrelation functionals in density functional theory and computational errors. Subsequently, we calculated delithiation energies for Li1−xNiO2, [Li0.98−xNi0.02]3a[Ni0.98Li0.02]3bO2 and [Zr0.0625Li0.6875−xNi0.25]3a[Ni0.75Li0.25]3bO2 with x = 0.0208, 0.1250, 0.1875 (see Table 1). It is found that the delithiation Table 1. Calculated Formation Energies (eV) for Li Vacancies as a Function of State of Charge for Li1−xNiO2, [Li0.02Ni0.98]3a[Ni0.02Li0.98‑x]3bO2, and Zr0.0625[Li0.75‑xNi0.25]3a[Ni0.75Li0.25]3bO2 state of charge, x perfect Li1−xNiO2 [Li0.98−xNi0.02]3a[Ni0.98Li0.02]3bO2 [Zr0.0625Li0.6875−xNi0.25]3a[Ni0.75Li0.25]3bO2
0.0208
0.125
0.1875
3.48 3.11 3.14
3.31 3.19 3.12
3.18 3.18 3.12
energy (per Li) decreases with increasing level of delithiation in the first case but does not change much in the second and third cases. Although the pristine case had higher delithiation energy than the interchanged and Zr-doped cases, all three cases exhibited similar delithiation energies at the highest level of delithiation studied in this work. This result shows that the Zrdoped structure has a similar stability in comparison to the nondoped case with or without 2% Li−Ni interchange. To further validate the structural model for the calculation, the influence of the Hubbard U parameter was examined by 1811
DOI: 10.1021/acs.chemmater.8b00619 Chem. Mater. 2018, 30, 1808−1814
Article
Chemistry of Materials recalculating the a, b, and c lattice parameters of [Zr 0.0625 Li 0.6875 Ni 0.25 ] 3a [Ni 0.75 Li 0.25 ]3b O 2 (hereafter called model I) for U = 3.0 by using U = 4.0 for Ni while U = 4.0 was kept for Zr. As can be seen in Figure 8, changing the U
Electrochemical properties of the pristine and Zr-doped LiNiO2 cathodes were evaluated using 2032 coin-type half-cells employing Li metal as the counter electrode. Figure 9a shows the initial charge−discharge curves of the pristine and Zr-doped LiNiO2 cathodes cycled between 2.7 and 4.3 V versus Li+/Li at 0.1 C (18 mA g−1) and 30 °C. The initial discharge capacities of the pristine and Zr-doped LiNiO2 were nearly identical (247.5 mAh g−1 for pristine and 246.5 mAh g−1 for Zr-doped LiNiO2) with a Coulombic efficiency of 97% for both cathodes. The discharge capacity delivered by the cathodes amounts to 90% of the theoretical capacity of LiNiO2 and represents one of the highest values ever reported for LiNiO2 at 4.3 V versus Li+/Li. The subsequent voltage profiles indicate that while the discharge capacity for the pristine LiNiO2 cathode quickly dropped during the initial cycles, that of the Zr-doped LiNiO2 cathode remained rather stable (see Figure S5). Figure 9a clearly demonstrates that the cation ordering (i.e., interchange of Li and Ni ions) had hardly any effect on Li removal/ insertion. We believe that the outstanding discharge capacity of LiNiO2 stemmed from the unique morphology of the cathode particles, unlike in previous cases, wherein LiNiO2 particles were typically synthesized via a solid-state reaction or sol−gel. It is conjectured that Li migration was expedited by the particle morphology in which nanosized particles favorably oriented for Li migration into a micron-sized secondary particle, and in contrast to previous propositions, the specific locations of Li ions in LiNiO2 hardly affected Li migration or the structural stability in the delithiated LixNiO2. In fact, Zr doping improved the cycling stability, as can be seen from the cycling performance in Figure 9b. The Zr-doped LiNiO2 cathode maintained 81% of its initial discharge capacity after 100 cycles at 0.5 C (90 mA g−1), whereas the pristine cathode retained 74% of its initial capacity during the same cycling period. The capacity retention for both cathodes easily outperforms previously reported retention data for LiNiO2, which cannot be reversibly cycled while it delivers the high discharge capacity reported in Figure 9b.6,7,22 Again, the favorable particle morphology combined with a strong crystallographic texture in radial direction likely helped relieve the large internal strain induced by multiple phase transitions in LiNiO2. With regard to rate capability, Figure 9c shows the discharge capacity of both LiNiO2 cathodes as a function of C rate ranging from 0.2 to 5 C (charge rate fixed at 0.2 C) in 2032 coin-type half-cells at 30 °C between 2.7 and 4.3 V versus Li+/Li. The rate capability of the cathodes hardly differed from each other. The pristine cathode had a capacity retention of 82.0% at 5 C in comparison to 81%
Figure 8. Calculated lattice parameters and top views of the layer with Zr substitution in the various models studied in this work. Deviation of calculated lattice parameters from the experimental values (given in the last row) is given in parentheses.
parameter from 3 to 4 does not affect the lattice parameters much. Furthermore, the possibility of Li vacancy formation in model I was studied by removing 3 Li from the layer where Zr is located ([Zr0.0625Li0.50Ni0.25]3a[Ni0.75Li0.25]3bO2, hereafter called model II). In model II, the charge of the Zr4+ dopant is compensated by the replaced Li+ as well as the three Li+ vacancies. It is found that the a and b values for model II match our experimental results (see Figure 8) better than those from model I do. On the other hand, the value of c in model II is larger than that in model I, which was in good agreement with the experimental value. However, deviation of the calculated lattice parameters in models I and II from the experimental values is less than 1.04%. We have also investigated a structure with substitution of 25% Li by Zr without any Li−Ni interchange ([Zr0.25Li0.75]3a[Ni1.00]3bO2, hereafter called model III). A comparison between lattice parameters of model III with those of experimental measurements indicates that model III is unlikely. Therefore, the structure used for the theoretical calculations provides a model that is consistent with the experimental result.
Figure 9. (a) Initial charge/discharge curves of the pristine and Zr-doped LiNiO2 cathodes cycled between 2.7 and 4.3 V versus Li+/Li at 0.1 C rate at 30 °C using a Li metal anode. (b) Corresponding cycle performance of the two cathodes at 0.5 C rate for 100 cycles. (c) Comparison of rate capability of the pristine and Zr-doped LiNiO2 cathodes. 1812
DOI: 10.1021/acs.chemmater.8b00619 Chem. Mater. 2018, 30, 1808−1814
Chemistry of Materials at 0.1 C for the Zr-doped LiNiO2 cathode. The rate capability data clearly disagree with the proposition that Li diffusion is adversely affected by antisite defects. Considering that the amount of Zr doped into the bulk was 0.4 mol %, which is too small to have directly altered structural stability or surface chemistry with the electrolyte, it can be concluded that antisite defects, whose presence is unequivocally confirmed by cation ordering, may not be the main reason for the initial irreversible capacity and rapid capacity fading often observed from LiNiO2 cathodes. This finding has far-reaching implications, as one of the extreme manifestations of antisite defects or cation mixing is the formation of a rock salt phase (50% random mixing of Li and transition metals ions at 3a and 3b sites) on the NCM and NCA cathode surface during cycling.23,24 The surface rock salt phase is supposedly inactive and raises the charge transfer resistance, leading to fading of the cathodes, especially Nienriched layered cathodes.25 Our result suggests that the rock salt phase may not be detrimental to cycling stability, especially for the Ni-rich layered cathodes, which is supported by the TEM observation that the several nanometers thick rock salt formed in the initial stage did not grow appreciably in thickness and remained stable during extended cycling.26
ACKNOWLEDGMENTS
■
REFERENCES
(1) Ohzuku, T.; Ueda, A.; Nagayama, M.; Iwakoshi, Y.; Komori, H. Comparative study of LiCoO2, LiNi1/2Co1/2O2 and LiNiO2 for 4 V secondary lithium cells. Electrochim. Acta 1993, 38, 1159−1167. (2) Noh, H.-J.; Youn, S.; Yoon, C. S.; Sun, Y.-K. Comparison of the structural and electrochemical properties of layered Li[Ni x Co y Mn z ]O 2 (x = 1/3, 0.5, 0.6, 0.7, 0.8 and 0.85) cathode material for lithiumion batteries. J. Power Sources 2013, 233, 121−130. (3) Lee, K.-S.; Myung, S.-T.; Amine, K.; Yashiro, H.; Sun, Y.-K. Structural and Electrochemical Properties of Layered Li[Ni 1−2x Co x Mn x ]O 2 (x = 0.1−0.3) Positive Electrode Materials for Li-Ion Batteries. J. Electrochem. Soc. 2007, 154, A971−A977. (4) Yoon, C. S.; Choi, M. H.; Lim, B.-B.; Lee, E.-J.; Sun, Y.-K. ReviewHigh-Capacity Li[Ni 1-x Co x/2 Mn x/2 ]O 2 (x = 0.1, 0.05, 0) Cathodes for Next-Generation Li-Ion Battery. J. Electrochem. Soc. 2015, 162, A2483−A2489. (5) Andre, D.; Kim, S.-J.; Lamp, P.; Franz Lux, S.; Maglia, F.; Paschosa, O.; Sitasznya, B. Future generations of cathode materials: an automotive industry perspective. J. Mater. Chem. A 2015, 3, 6709− 6732. (6) Ohzuku, T.; Ueda, A.; Nagayama, M. Electrochemistry and Structural Chemistry of LiNiO2 (R3m ̅ ) for 4 V Secondary Lithium Cells. J. Electrochem. Soc. 1993, 140, 1862−1870. (7) Arai, H.; Okada, S.; Ohtsuka, H.; Ichimura, M.; Yamaki, J. Characterization and cathode performance of Li1‑xNi1+xO2 prepared with the excess lithium method. Solid State Ionics 1995, 80, 261−269. (8) Kanno, R.; Kubo, H.; Kamiyama, T.; Izumi, F.; Takeda, Y.; Takano, M. Phase Relationship and Lithium deintercalation in lithium Nickel Oxides. J. Solid State Chem. 1994, 110, 216−225. (9) Koyama, Y.; Arai, H.; Tanaka, I.; Uchimoto, Y.; Ogumi, Z. Defect Chemistry in Layered LiMO2 (M = Co, Ni, Mn, and Li1/3Mn2/3) by First-Principles Calculations. Chem. Mater. 2012, 24, 3886−3894. (10) Kim, J.-M.; Chung, H.-T. The first cycle characteristics of Li[Ni1/3Co1/3Mn1/3]O2 charged up to 4.7 V. Electrochim. Acta 2004, 49, 937−944. (11) Venkatraman, S. Manthiram. Structural and Chemical Characterization of Layered Li1‑xNi1‑yMnyO2‑δ (y= 0.25 and 0.5, and 0 ≤ (1 - x) ≤ 1) Oxides. Chem. Mater. 2003, 15, 5003−5009. (12) Peres, J. P.; Weill, F.; Delmas, C. Lithium/vacancy ordering in the monoclinic LixNiO2 (0.50 ≤ x ≤ 0.75) solid solution. Solid State Ionics 1999, 116, 19−27. (13) Lee, M.-H.; Kang, Y.-J.; Myung, S.-T.; Sun, Y.-K. Synthetic optimization of Li[Ni1/3Co1/3Mn1/3]O2 via co-precipitation. Electrochim. Acta 2004, 50, 939−948. (14) Roisnel, T.; Rodriguez-Carvajal, J. Fullprof Manual; Institut Laue-Langevin, Grenoble, France, 2001. (15) Kresse, G.; Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (16) Chen, Y.; Peng, F.; Yan, Y.; Wang, Z.; Sun, C.; Ma, Y. Exploring High-Pressure Lithium Beryllium Hydrides: A New Chemical Perspective. J. Phys. Chem. C 2013, 117, 13879−13886. (17) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868.
CONCLUSION Using electron diffraction, the ordering of Li and Ni ions in Zrdoped LiNiO2, which has not been previously reported in stoichiometric LiNiO2 or any other Ni-rich NCM or NCA cathodes, was observed for the first time. First-principles calculations suggested that low-level doping of LiNiO2 with Zr can indeed energetically promote the observed cation ordering. Contrary to previous beliefs, antisite defects, whose presence is unequivocally justified by the cation ordering, hardly showed any negative effects on the electrochemical performance of LiNiO2. Furthermore, it was shown that reduction of the initial irreversible capacity and improved stability during subsequent cycles can be better achieved by controlling the particle morphology to make it conducive to Li migration and relieving the inherent structural instability of the deeply charged LixNiO2. ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemmater.8b00619. Electron diffraction patterns for pristine and Zr-doped LNiO2, schematic drawings of the calculated structure models, and charge/discharge curves for the pristine and Zr-doped LiNiO2 cathodes during cycling (PDF)
■
■
This work was mainly supported by the Global Frontier R&D Program (2013M3A6B1078875) on Center for Hybrid Interface Materials (HIM) funded by the Ministry of Science, Information & Communication Technology (ICT) and the Human Resources Development program (No. 20154010200840) of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korea government Ministry of Trade, Industry and Energy. Q.Z. and P.K. acknowledge the North-German Supercomputing Alliance (HLRN) for providing HPC resources.
■
■
Article
AUTHOR INFORMATION
Corresponding Author
*E-mail for Y.-K.S.:
[email protected]. ORCID
Chong S. Yoon: 0000-0001-6164-3331 Yang-Kook Sun: 0000-0002-0117-0170 Notes
The authors declare no competing financial interest. 1813
DOI: 10.1021/acs.chemmater.8b00619 Chem. Mater. 2018, 30, 1808−1814
Article
Chemistry of Materials (18) Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.; Sutton, A. P. Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA+U study. Phys. Rev. B: Condens. Matter Mater. Phys. 1998, 57, 1505−1509. (19) Li, W.; Reimers, J. N.; Dahn, J. R. Crystal structure of LixNi2‑xO2 and a lattice-gas model for the order-disorder transition. Phys. Rev. B: Condens. Matter Mater. Phys. 1992, 46, 3236−3246. (20) Noh, H.-J.; Chen, Z.; Yoon, C. S.; Lu, J.; Amine, K.; Sun, Y.-K. Cathode Material with Nanorod Structure - An Application for Advanced High-Energy and Safe Lithium Batteries. Chem. Mater. 2013, 25, 2109−2115. (21) Meng, Y. S.; Ceder, G.; Grey, C. P.; Yoon, W.-S.; Jiang, M.; Bréger, J.; Shao-Horn, Y. Cation Ordering in Layered O3 Li[NixLi1/3−2x/3Mn2/3‑x/3]O2 (0 ≤ x ≤ 1/2) Compounds. Chem. Mater. 2005, 17, 2386−2394. (22) Kubo, K.; Fujiwara, M.; Yamada, S.; Arai, S.; Kanda, M. Synthesis and electrochemical properties for LiNiO2 substituted by other elements. J. Power Sources 1997, 68, 553−557. (23) Lin, F.; Markus, I. M.; Nordlund, D.; Weng, T.-C.; Asta, M. D.; Xin, H. L.; Doeff, M. M. Surface reconstruction and chemical evolution of stoichiometric layered cathode materials for lithium-ion batteries. Nat. Commun. 2014, 5, 3529. (24) Watanabe, S.; Kinoshita, M.; Hosokawa, T.; Morigaki, K.; Nakura, K. Capacity fade of LiAlyNi1‑x‑yCoxO2 cathode for lithium-ion batteries during accelerated calendar and cycle life tests (surface analysis of LiAlyNi1‑x‑yCoxO2 cathode after cycle tests in restricted depth of discharge ranges). J. Power Sources 2014, 258, 210−217. (25) Liu, W.; Oh, P.; Liu, X.; Lee, M.-J.; Cho, W.; Chae, S. J.; Kim, Y. J.; Cho, J. P. Nickel-Rich Layered Lithium Transition-Metal Oxide for High-Energy Lithium-Ion Batteries. Angew. Chem., Int. Ed. 2015, 54, 4440−4457. (26) Kim, U.-H.; Myung, S.-T.; Yoon, C. S.; Sun, Y.-K. Extending the Battery Life Using an Al-Doped Li[Ni0.76Co0.09Mn0.15]O2 Cathode with Concentration Gradients for Lithium Ion Batteries. ACS Energy Lett. 2017, 2, 1848−1854.
1814
DOI: 10.1021/acs.chemmater.8b00619 Chem. Mater. 2018, 30, 1808−1814