Shanghai Engineering Research Center of Hierarchical Nanomaterials, Frontiers Science Center for Materiobiology and Dynamic Chemistry, School of Materials Science and Engineering, East China University of Science & Technology, Shanghai 200237, China

Keywords:LiNi0.9Co0.1O2 Co-doping Crystal stability Cycling life Li-ion batteries

ABSTRACT The Ni-ultrahigh cathode material is one of the best choices for further increasing energy-density of lithium-ion batteries(LIBs),but they generally suffer from the poor structure stability and rapid capacity fade.Herein, the tungsten and phosphate polyanion co-doped LiNi0.9Co0.1O2cathode materials are successfully fabricated in terms of Li(Ni0.9Co0.1)1-xWxO2-4y(PO4)y by the precursor modification and subsequent annealing.The higher bonding energy of W-O (672 kJ·mol-1) can extremely stabilize the lattice oxygen of Ni-rich oxides compared with Ni-O (391.6 kJ·mol-1) and Co-O (368 kJ·mol-1).Meanwhile,the stronger bonding of Ni-() vs.Ni-O could fix Ni cations in the transition metal layer, and hence suppressing the Li/Ni disorder during the charge/discharge process.Therefore, the optimized Li(Ni0.9Co0.1)0.99W0.01O1.96(PO4)0.01 delivers a remarkably extended cycling life with 95.1% retention of its initial capacity of 207.4 mA·h·g-1 at 0.2 C after 200 cycles.Meantime, the heteroatoms doping does not sacrifice the specific capacity even at different rates.

1.Introduction

The eco-friendly electric vehicles (EVs) have drawn notable attention due to the shortage of fossil fuels and serious environmental pollution derived from the extensive utilization of internal combustion engine vehicle [1-4].However, the popularization of EVs has been largely beset by the insufficient driving range per charge, which is definitively dominated by the power producer of Li-ion batteries (LIBs) [5,6].The current power LIBs based on LiCoO2, LiMn2O4, LiFePO4, NCM or NCA (Ni ≤0.8) are incapable of satisfying the pursuit of long-range EV sowing to the lower energy density[7].Therefore,the exploration of Ni-rich LiNixCo(1--x)O2(x ≥0.9) cathode materials with higher specific capacity is of great significance to the widespread deployment of EVs [8,9].Although increasing Ni content could improve the specific capacity of high-nickel layered oxide cathodes,the concomitantly increased cation disorder will impede the Li+transfer and deteriorate the thermal stability [10,11].Moreover, the grievous lattice contraction/expansion derived from the phase transition in the deeply charged state induces the serious mechanical strain inside the particles, further leading to intergranular micro-cracking and even shattering [12].Such intergranular cracks act as the channels for electrolyte penetration, aggravating the detrimental surface reaction and further accelerating structural deterioration with enlarged interfacial impedance [13-15].

To date,element doping has been confirmed as a viable strategy to solve the aforementioned issues of Ni-rich cathode materials[16-18].Various cations doping (Al, Mg, Mo, V, B, Ti, et al.) have been devoted to improving the cycling stability of Ni-rich cathode[19-24].In particular,the transition metal ions with a high valence state could form strong metal-Oxygen bonds, and thus stabilizing crystal structure and interface.For example,Zhang et al.have prepared a series of W-doped LiNi0.8Co0.1Mn0.1O2with excellent cyclic reversibility.The addition of tungsten could stabilize the structure and restrain the surface transformation to rock-salt phase,but the cation mixing became more serious due to incremental Ni2+[25].In addition, polyanion doping based on nonmetal elements(,,)have also been employed to improve the cyclic durability by stabilizing the oxygen close-packed structure and impeding the transfer of Ni2+into Li layer [26-29].Admittedly,these modifications are valid for solving the shortcomings of Nirich cathode materials, but the inevitable rise in cost associated with complicated multi-step process is pernicious for the commercialization.Thus,exploiting a simple method to achieve cation and anion co-doping is meaningful for the practical application of Ni-rich cathode materials.

2.Experimental

2.1.Materials synthesis

The Ni0.9Co0.1(OH)2precursor was synthesized by a simple coprecipitation method as described previously [21].Briefly,2.0 mol·L-1mixed solution of NiSO4·6H2O (Aladdin, China) and CoSO4·7H2O (Aladdin, Ni: Co = 9:1 in molar ratio), 4.0 mol·L-1NaOH (Aladdin) and 1.5 mol·L-1NH4OH (General-Reagent, China)were separately pumped into the reactor under inert atmosphere.The pH was controlled at 11.0 with a constant stirring speed of 600 r·min-1.The precursor powder was finally obtained after washing,filtering,and vacuum drying at 110°C for 24 h.To prepare Li(Ni0.9-Co0.1)1-xWxO2-4y(PO4)y(x = y = 0, 0.005, 0.01, 0.02, 0.05), the precursor was blended with ammonium dihydrogen phosphate and phosphotungstic acid in deionized water, followed by stirring and drying to gain the mixture.Then,the mixture was mixed with LiOH·H2O(Aladdin,5%rich)by grinding 20 min in a mortar.To gain the final samples, the powders were calcined at 730 °C for 10 h under flowing oxygen with a heating/cooling temperature rate of 5 °C·min-1.Theand W co-doped samples with 0%, 0.5%, 1%,2%, 5% dopant (molar ratio) are denoted as NC, NCPW0.5, NCPW1,NCPW2 and NCPW5, respectively.

2.2.Material characterization

Powder X-ray diffraction (XRD, D8 Advance, Bruker, Germany;Cu Kα radiation)was employed to analyze the crystalline structure of prepared samples with a scan rate of 2(°)·min-1.The morphology was observed by field emission scanning electron microscopy(FESEM, S-4800, Hitachi, Japan) and field emission transmission electron microscopy (FETEM, JEM-2100F, JEOL, Japan) with an Xray energy dispersive spectrometer (EDS).In order to observe the cross-section of the sample, the powders were firstly implanted into the epoxies and then polished by Ultra-thin slicer equipment(EM UC7, Leica, Germany).The surface element compositions and valence were characterized via X-ray photoelectron spectra (XPS,ESCALAB 250Xi, Thermo Scientific, USA).The inductively coupled plasma atomic emission spectrometer (ICP-AES) was carried out with Agilent 725 (USA).

2.3.Electrochemical measurements

The active materials,carbon black(Super-P,Timcal graphite Co.Ltd., Switzerland), and poly(vinylidene fluoride) (PVDF, Solvay Group, USA) at a mass ratio of 8:1:1 in N-methyl-2-pyrrolidone(NMP,Aladdin)were uniformly mixed.Then,the slurry was coated on aluminum foil with an active material loading of 1.75-2.50 mg·cm-2.The electrolyte was 1.0 mol·L-1LiPF6in a mixture of ethylene carbonate (EC)/ethyl methyl carbonate (EMC) (VEC:VEMC= 3:7) with 2% (mass) vinylene carbonate (VC).The 2016-coin cells were assembled in an argon-filled glove box (＜0.0001‰for H2O and O2) with the prepared cathode, electrolyte, Celgard 2400 separator and lithium metal counter electrode.The galvano-static charge and discharge measurements were performed by the LANDCT2001A battery test system at 25 °C (1 C=190 mA·h·g-1).Electrochemical impedance spectra(EIS)experiment was conducted in Autolab PGSTAT302N electrochemical workstation with the range from 100 kHz to 0.01 Hz.

3.Results and Discussion

The refined structural information of the as-prepared materials was analyzed by the powder XRD and corresponding Rietveld refinement.As shown in Figs.1(a) and S1 (see Supplementary Material), the diffraction peaks of all samples could be indexed to a hexagonal α-NaFeO2structure (R3-m space group) [22].The distinct peak splitting of (0 0 6)/(1 0 2) and (1 0 8)/(1 1 0) reflections indicates a well-ordered layered structure except for NCPW5 [30].For NCPW5,the splitting of(0 0 6)/(1 0 2)and(1 0 8)/(1 1 0)peaks disappears with the arising of Li3PO4impurity peak, implying that the layered structure has been destroyed by the superfluous introduction ofand W.Specifically, the generation of the Li3PO4impurity phase results in the decrease of reactive Li content in NCPW5,which will further cause the Li deficiency in Ni-rich structure and exacerbate the Li/Ni disorder.Furthermore, XPS was employed to detect the surface element composition and valence state after modification.Fig.1(b) depicts the XPS spectra of the W4f for all samples, and the main peaks at 35.1 and 37.2 eV are identified as W4f7/2and W4f5/2for W6+[31].The peak at 133.6 eV is related to P5+in phosphate groups in the Fig.1(c)[32].Specifically, no P 2p peak can be observed in the NCPW0.5 because P content is insufficient to reach the detection threshold.As plotted in Fig.S2, two main peaks at 855 and 873 eV are assigned to Ni 2p3/2and 2p1/2.The typical peaks at 855.8 and 854.6 eV in the Ni 2p3/2spectra correspond to Ni3+and Ni2+for all materials, respectively [33,34].It could be found that the proportion of Ni2+increases slightly with the addition of W and PO43-concentration attributed to the comprehensive charge compensation.However, the proportion of Ni2+has risen significantly in the NCPW5, which is consistent with the severe Li/Ni disorder obtained in XRD result.Meanwhile, the characteristic peaks located at 779.8 and 794.9 eV are related to Co 2p3/2and Co 2p1/2of Co3+with ignorable changes, implying thatand W doping has no effect on Co ions [35].

To gain a detailed analysis of structural evolution caused byand W co-doping,XRD Rietveld refinement results are plotted in Fig.2(a) and (b).The ratio of I003/I104represents the cation disorder, and a higher ratio corresponds to a lower Li/Ni mixing [22].Among the five samples,the ratio of I003/I104initially rises and then decreases with the augment of doping content, and NCPW1 exhibits the largest value (2.04) due to the opposite impact ofand W doping.On the one hand, the doping of W6+with high valence causes the reduction of Ni3+to Ni2+with more severe Li/Ni mixing according to the charge compensation effect.On the other hand, the introduction ofpolyanions with stronger bonding to Ni cations can markedly minimize Li/Ni mixing.Considering the competition of two doping elements,the optimal NCPW1 has the greatest layered structure with the minimum Li/Ni disorder.Meanwhile, the evolution of c/a value is similar, further confirming the above conclusion [36].The lattice parameters of all materials were also calculated by X-ray Rietveld refinement.It is noticeable that the lattice parameters a and c enlarge accompanied by the increase ofand W content, which is ascribed to the larger ionic radius of W6+(0.060 nm)and the occupied tetrahedral interstitial sites by P5+(0.034 nm).Additionally,the thicknesses of transition metal slab increase from 0.21258 to 0.21442 nm and the thicknesses of lithium layer also increase from 0.25736 to 0.25903 nm for the NCPW1 with NC as a contral, as shown in Table S1.It is reasonably speculated that the larger lattice parameters can enlarge Li+diffusion channels and the lower cation disorder could stabilize the crystal structure, justly resulting in enhanced electrochemical performance.Additionally, ICP-AES(Fig.2(c)) was carried to investigate the element compositions,the detective elemental ratio nearly agrees with the purported value for all materials.

Fig.2.(a) Ratio of I003/I104 and c/a, (b) lattice parameters a and calculated from X-ray Rietveld refinement, and (c) ICP results of the NC, NCPW0.5, NCPW1, NCPW2 and NCPW5.

Fig.3.(a)-(b) SEM images of the NC and NCPW1; (c) cross-sectional SEM image, (d) HRTEM and (e)-(j) cross-sectional maps of the NCPW1.

To further explore the influence ofand W co-doping on particle morphology, SEM images of all materials are presented in Figs.3(a) and (b) and S3.It is conspicuous that all particles are spherical in shape with an average diameter of 9 μm.But the morphology of the primary particles is obviously different before and after doping.The primary particles of doped materials are much shorter in width and longer in length.The cross-sectional SEM images(in Figs.3(c)and S4)ulteriorly exhibit the differences about the internal particles for the NCPW1 and NC.The NC comprises disorderly distributed block primary particles, whereas the NCPW1 consists of elongated primary particles with radially oriented from center to surface.This might be due to the changed growth surface energy on different crystal planes with the presence of W ions[37].The radial microstructure can effectively avoid local stress even in the highly delithiated state,and hence maintaining high structural integrity during cycles.Fig.3(d) depicts the HRTEM image of the NCPW1.The apparent lattice fringes of 0.47 nm can be wellindexed to (0 0 3) plane in layered structure (R3-m), which is similar to the HRTEM image of NC (Fig.S4) [38].The well-maintained lattice fringe indicates thatand W co-doping has no adverse impact on crystal structure, which is matched with the results of XRD.Furthermore, cross-sectional SEM with EDS mapping is employed to observe the spatial distribution of Ni, Co, P, W, O in the NCPW1 particles.As displayed in Fig.3(e)-(j),elements involving Ni, Co, P, W and O are all uniformly spread over the particle,suggesting the uniform doping of P and W.This conclusion is also verified by the HRTEM image and corresponding EDS elemental maps of the NCPW1 (Fig.S5).

The electrochemical performance of all samples were tested to explore the effect of co-modification with a voltage window (2.7-4.3 V vs.Li/Li+) at 25 °C.Fig.4(a)depicts the cycling stability of all samples at 1 C.Among the five samples,the NCPW1 possesses the most outstanding capacity retention of 98.7% after 100 cycles,greatly out performing that of NC (82.5%), NCPW0.5 (92.7%),NCPW2(96.4%)and NCPW5(81.3%).Additionally,the NCPW1 displays the highest initial coulombic efficiency of 90.7% compared with that of NC(89.8%),NCPW0.5(90.4%),NCPW2(89.2%),NCPW5(87.5%), as plotted in Fig.S6.The optimized performance is attributed to the minimum Li/Ni disorder, which is already certified by XRD and XPS.Furthermore, cycling performance at 0.2 C of the optimal NCPW1 is shown in Fig.4(b), with NC as a reference.The NCPW1 demonstrates a notably improved cycling stability with a capacity retention of 95.1% after 200 cycles.In sharp contrast,the discharge capacity of NC suffers from a severe fading after 200 cycles (66.1%).Furthermore, voltage hysteresis evolution of both samples over 200 cycles are depicted in Fig.4(c).Compared to NC (0.17 V), NCPW1 maintains much lower voltage hysteresis(0.05 V).Additionally, the charge and discharge curves, evolution of the average charge/discharge voltage over 200 cycles at 0.2 C and corresponding calculated dQ/dV profiles of the NCPW1 and NC are shown in Fig.S8.The difference between the charge and discharge voltages only increases from 0.02 to 0.05 V after 200 cycles for the NCPW1, much lower than that of NC (from 0.02 to 0.17 V).Meanwhile, the voltage plateaus (4.2 V) referred to H2-H3 phase transition is well-retained during cycling for the NCPW1,while that for NC is completely vanished.Similarly, the redox peaks in the dQ/dV profiles for the NCPW1 show negligible changes in intensity, whereas that of NC significantly decline in prolonged cycling with severe irreversibility of the redox reactions.Such phenomena reveal that the NCPW1 has pimping electrochemical polarization and excellent reversibility of the H2-H3 phase transitions benefiting from the synergetic modification of, W codoping.Besides, the ultra-long cycling stability (Fig.4(d)) can also be achieved at 2 C for NCPW1(93.4%after 400 cycles),much better than that of NC (52.4%).More impressively,NCPW1 still delivers a high discharge capacity of 141.6 mA·h·g-1at 10 C(Fig.4(e)),which is attributed to the shorter Li+diffusion channels resulting from the radially aligned primary particles and faster Li+transfer attributed from inhibited Li/Ni disorder.

Fig.4.(a)Cycling performance of the five samples at 1 C within 2.7-4.3 V;(b)over 200 cycles at 0.2 C,(c)the corresponding voltage hysteresis,and(d)400 cycles at 2 C of the NCPW1 and the NC; (e) the capacity values at different rates for the NCPW1.

Fig.5.Self-discharge test of the NCPW1 and NC at 55°C:(a)the charge curve at 0.5 C,(b)OCV evolution during resting for 96 h,(c)the discharge curve after resting,(d)O 1s XPS spectra, (e) the amount of Ni deposited on the lithium anode and (f) XRD patterns of the NCPW1 and NC after 200 cycles at 0.2 C.

To obtain a deep understanding of the co-modification effect on the surface stability during electrochemical process, the selfdischarge experiments of the NCPW1 were carried out in contrast to NC, as presented in Fig.5(a)-(c).For the self-discharge test at 55°C,the Li|NCPW1 and Li|NC cells were charged to 4.5 V and then rested for 96 h,followed by discharging to 2.7 V at 0.5 C.Both samples deliver a semblable charge capacity during the charging process.Nonetheless, the evolution of open circuit voltages (OCV)for the NCPW1 and NC during resting are entirely different.The OCV of NCPW1 delivers a slightly decline and ends up at 4.21 V.However, NC under goes a continuous OCV decline and ends up at 4.15 V, suggesting a more severe self-discharge.After hightemperature storage, the NCPW1 possesses a discharge capacity of 222.6 mA·h·g-1, much higher than that of NC (119.5 mA·h·g-1).Such phenomena suggest the NCPW1 has a more stable and robust interface accompanied by the decreased detrimental electrode/-electrolyte reactions.Furthermore, XPS characterizations were employed to investigate elemental composition and surface chemistry for both cathodes after 200 cycles at 0.2 C.As shown in Figs.5(d) and S9, the increased M-O bonding and reductive C=O bonding in the XPS spectra of the NCPW1 reveal a less interfacial side reaction,which is due to the less escape of lattice oxygen resulting from the strong W-O bonds [39,40].The F element roots in LiF/MFxcompounds in CEI films,so the decreased content of F element(from 24.25% to 12.97% (atom)) can reflect the mitigated dissolution of transition metal ions.Meanwhile, it is also validated by ICP-AES analysis on Li anodes after 200 cycles in Fig.5(e).The mass of Ni in Li anode after 200 cycles for NCPW1 is 1.63 μg,much lower than NC(3.00 μg),further indicating that the dissolution of transition metal ions is suppressed byand W co-doping.Fig.5(f)compares the structural changes of the NCPW1 and NC after 200 cycles at 0.2 C.The NCPW1 still retains well layered structure without detected impurity after long-term cycling.On the contrary,NC suffers from severe interfacial side reactions and irreversible phase transformation with the appearance of NiO impurity phase.These results reveal that NCPW1 exhibits superior interfacial and structural stability owing to the W andco-doping, which is vital for extending the ultra-long cycling life of Ni-rich cathode materials.

To further compare the extent of the surface damage for both samples, EIS measurement is carried out at every 50th cycle.The electrochemical-impedance data of NCPW1 and NC are shown as Nyquist plots in Fig.6(a) and (b), all containing two semicircles and a sloped line.The first and second semicircle in the high and medium frequency range represents surface film resistance (Rsf)and charge transfer impedance (Rct), while the sloped line is related to the Li+diffusion resistance in electrodes [41,42].As displayed in Table S2,the Rsfof NC significantly increases from 51.59 to 138.1 Ω after 200 cycles,while that of NCPW1 shows a negligible increase from 52.27 to 58.74 Ω.Similarly,the increase of Rctfor NCPW1 is effectively moderated and reaches a final value of 89.57 Ω after 200 cycles,far less than that of NC(626.4 Ω),implying the alleviative polarization and side reactions due to stronger cathode-electrolyte interphase afterand W co-doping.The stronger bonding ofpolyanions to Ni cations can effectively restrain the migration of Ni cations to reduce the Li/Ni disorder,further improving Li+transfer kinetics and suppressing detrimental phase transformation during cycling.Additionally, thepolyanions and W6+cations can stabilize the oxygen framework in crystal structure, which is advantageous to inhibit oxygen loss and corresponding side reactions with electrolyte.Additionally,SEM and cross-sectional SEM images of both samples after 200 cycles are shown in Figs.6(c), (d) and S10 to intuitively examine the change of morphology and structure.The NC undergoes severe structural deterioration with numerous wide microcracks after cycling.The microcracks would undermine the mechanical integrity of particles and expose a fresh surface with tempestuously electrolyte erosion,thus leading to fast capacity degradation.However, NCPW1 can maintain integrated spherical-like morphology without visible cracks, besides some hairline microcracks, ulteriorly suggesting that the dramatic volume change with H2-H3 phase transition is minimized due to the radially aligned primary particles.

4.Conclusions

Fig.6.Nyquist plots of the electrochemical impedances and SEM images of the NCPW1 (a, c) and NC (b, d) over 200 cycles at 0.2 C.

In this work, the tungsten and phosphate polyanion co-doped Ni-rich Li(Ni0.9Co0.1)0.99W0.01O1.96(PO4)0.01cathode material with ultra-stable cycling was successfully prepared via a simple precursor doping with subsequent annealing.The introduction of W-O bond(high bond energy of 672 kJ·mol-1)can strengthen the lattice oxygen framework and thus stabilize the electrode-electrolyte interface.Meantime, the markedly relieved internal stress due to the radial microstructure can minimize intergranular microcracking,thus guaranteeing the structural integrity during cycling.Additionally,the stronger bonding ofpolyanions to Ni cations can suppress Li/Ni mixing and dissolution of transition-metal ions.Consequently, NCPW1 possesses a superior capacity retention of 95.1% after 200 cycles at 0.2 C, much better than that of NC(66.1%).More impressively, it still affords 400 cycles with 93.4%capacity retention even at 2 C.This work provides a viable solution for extending the cycling life of Ni-rich cathode materials with high energy density.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (91834301), the Innovation Program of Shanghai Municipal Education Commission, the Shanghai Scientific and Technological Innovation Project (18JC1410500), and the Fundamental Research Funds for the Central Universities(222201718002).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.04.001.