Phosphorus-doped lithium-and manganese-rich layered oxide cathode material for fast charging lithium-ion batteries

covalent bonds with oxygen atoms, resulting in an improved rate performance (capacity retention from 38% to 50% at 0.05 C to 5 C) and thermal stability (50% heat release compared with pristine material). First-principles calculations showed the P-doping makes the transfer of excited electrons from the valence band to conduction band easier and P can form a strong covalent bond helping to stabilize material structure. Furthermore, the solid-state electrolyte modiﬁed P 5+ doped LMR showed an improved cycle performance for up to 200 cycles with capacity retention of 90.5% and enhanced initial coulombic efﬁciency from 68.5% (pristine) or 81.7% (P-doped LMR) to 88.7%. (cid:1) 2021 The Authors. Published by ELSEVIER B.V. and Science Press on behalf of Science Press and Dalian Institute of Chemical Physics, Chinese Academy of Sciences. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
Lithium-ion batteries (LIBs) are widely used in electronic products, such as mobile phones and notebook computers, as well as in industries, such as electric vehicles and energy storage power stations [1].Their clean and efficient characteristics may help to alleviate the petrochemical crisis and improve the global environment.Thus, they have attracted the attention of many researchers [2][3][4][5].However, the large-scale application of electric vehicles is still limited by the energy density, cycling and rate performances, and safety of LIBs [6].
Cathode materials play an important role in LIBs.However, commercial cathode materials (e.g., LiFePO 4 (170 mAh g À1 ), LiCoO 2 (140 mAh g À1 )) cannot meet the growing demand for new energy battery applications [7].The currently developed high-capacity nickel-rich ternary material LiNi x Co y Mn z O 2 (NCMxyz) has a maximum capacity of about 200 mAh g À1 and contains high-cost and toxic Co elements [8].Following considerable development, the lithium-and manganese-rich layered oxide (LMR) cathode material xLi 2 MnO 3 Á(1-x)LiMO 2 (M = Mn, Ni, etc.) has been shown to have a high specific energy (about 900 Wh kg À1 ) and theoretical specific capacity (approximately 280 mAh g À1 ), large discharge range (2.0 V ~4.8 V) and low cost, making it the focus of much research [9][10][11].
The main obstacles against the commercialization of LMRs are their sluggish Li + ion diffusion and poor intrinsic electron conductivity and structural stability [9][10][11].The Li + transport involves two steps from the electrolyte to the material: transport from the electrolyte into the material interface and transport in the material.In our previous research, we showed an interface engineering strategy to improve the stability and interfacial Li + penetrability of LMR material at low current density [12].Nevertheless, the performance of the LMR showed the same rate performance at high current densities of more than 2 C.This is because enough lithium ions from material interface and electrons are limited in the bulk material transportation.Thus, the low intrinsic Li + and electron transport in the bulk material needs to be overcome for highperformance battery applications [13].Within the material, the distance and rate of Li + and electron transport determine the battery rate performance.Reduction of the particle size to nanoscale levels can improve the rate performance of the material by shortening the lithium-diffusion and electron transport pathways in the bulk material at the expense of the energy density [14][15][16][17].For example, Zhang et al. showed that a one-dimension nanowire LMR can increase the lithium-ion diffusion coefficients more than by eight times [18].However, improvement of the intrinsic diffusion rate of Li + and electrons remains a challenge [19].
In this paper, the Li 1.2 Mn 0.6 Ni 0.2 O 2 (LMR) was modified by phosphorus-doping to increase the Li + conductivity in the bulk material.The P doping was effective in increasing the interlayer spacing of the lithium layer, electron transport, and structural stability, which therefore resulting in significant improvement of the rate and safety performances (Fig. 1).First-principles calculations were used to theoretically explain the doping modification mechanism.In addition, the solid-state electrolyte Li 6.24 Al 0.12 La 3 Zr 1.8 -Mo 0.2 O 12 was used to further improve the stability of P 5+ doped LMR.

Materials synthesis
Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 (x = 0, 0.01, 0.02, 0.03) was synthesized by a high temperature solid phase method.Li 2 CO 3 (excess 5%) and Ni 1/4 Mn 3/4 (OH) 2 were weighed out according to the stoichiometric ratio, ground thoroughly, placed in a crucible for compaction and then put into a muffle furnace for high-temperature calcination.The temperature of the mixed material was initially increased to 500 °C at a rate of 5°min À1 and kept at 500 °C to pre-sinter the sample for 6 h.Subsequently, the temperature was increased to 850 °C at a heating rate of 5°min À1 and kept at 850 °C for 12 h.Afterwards, the temperature was reduced to 400 °C at a cooling rate of 2℃ min À1 , then cooled to room temperature to obtain the Li 1.2 Ni 0.2 Mn 0.6 O 2 material.Next, P 2 O 5 was weighed on aluminum foil, mixed with Li 1.2 Ni 0.2 Mn 0.6 O 2 and thoroughly ground, and placed in a muffle furnace for high-temperature calcination.The temperature of the mixed material was initially raised to 500 °C at a heating rate of 5°min À1 and kept at 500 °C to induce calcina-tion for 6 h.The sample was then left to cool to room temperature under ambient conditions.The obtained material was ground and sieved to obtain a dark brown powder of Li O were weighed in a stoichiometric ratio of 2 wt% and an appropriate amount of distilled water was added at room temperature.A magnetic stirrer was used to fully dissolve the reagents, citric acid and ethylene glycol were added under mixing, and the mixture was placed in an oil bath at 60 °C and stirred magnetically for 1 h to form a transparent solution.The LPMNO material was added to the sol with constant stirring, the solvent was evaporated at this temperature for about 6 h, and then the sample was transferred to an oven at 120 °C for 1 h for drying.The obtained materials were transferred to a muffle furnace at 650 °C (heating rate: 5°min À1 ) for 6 h, then cooled under ambient conditions to yield the final LPMNO@LALZMO material coated with LALZMO.

Cell assembly
For preparing the electrodes, the ratio of active material: Super-P: PVDF was 8:1:1.A uniformly mixed slurry of the reagents was used to coat clean aluminum foil and the material was placed in a blast drying oven at 110 °C for 3 h.A coin cell (CR2032 model) was assembled in an Etelux Lab2000 glove box filled with argon.A lithium metal sheet was used as the anode, the separator comprised a polypropylene microporous membrane material, and the electrolyte was 1.0 mol L À1 LiPF 6 (the solvent was EC: DMC: DEC with a volume ratio of 1:1:1).
A Land battery test system (LAND CT-2001A) was to test the cycling and rate performance of the assembled battery at room temperature.The voltage window was 2.0 V ~4.8 V.The assembled battery was first activated at a rate of 0.05 C for 30 cycles, and then at a rate of 0.2 C, 0.5 C, 1 C, 2 C and 5 C (1 C = 250 mAh g À1 ).The cycle performance test was similar to the rate performance test.First, the battery was activated for 2 cycles at a rate of 0.05 C, and then a constant current charge and discharge cycle was performed at 0.5 C.
For the impedance test, the assembled cell was activated on the LAND test system at a rate of 0.05 C and then a ZAHNER IM6eX electrochemical workstation was used for testing.The test conditions were as follows: voltage range 2.0 V ~4.8 V, frequency range 100 mHz ~100 kHz, disturbance voltage 5 mV.

DFT calculations
Density functional theory (DFT) calculations were carried out using the VASP software package with the projector-augmented- wave (PAW) method and Perdew-Burke-Ernzerhof (PBE) exchange correlation energy function established pseudopotential.Regarding the polyatomic system as a multi-electron and atomic nucleus system, the Schrodinger equation was used to solve the energy eigenvalue and eigenfunction based on the electron power, electron mass, speed of light, Planck's constant and Boltzmann constant involved in the system.To further study the electronic structure and properties of the materials, the influence of different spin polarization on the formation energy was assumed negligible.Therefore, the calculations were based on an approximate theory.
To improve the calculation accuracy, the exchange correlation of electrons between transition metals (TMs) was approximated through the local density approximation method (LDA) and generalized gradient approximation method (GGA).However, as the electronic exchange correlation of the system directly affected the electronic structure and energy characteristics of the system, so we choose the LDA + U method to correct its accuracy.A 1 Â 2 Â 1 K-point grid was selected, and the optimized supercell was relaxed to obtain the unit cell parameters and unit cell energy.
The Fermi energy level used Gauss expansion, and the sigma value was 0.10 eV.A 3 Â 16 Â 3 K-point grid was selected to calculate the electronic structure of the system to analyze the energy band structure and electronic DOS.The accuracy converged to 1.0 Â 10 À4 eV atom À1 and the cut-off energy of the plane wave was set to 500 eV.The calculations were carried out using only the ferromagnetic configuration, the exchange correlation energy between transition metal atoms J = 1.0 eV, and DFT-U (LDAU TYPE = 2) values of Ni, Mn and Mo on the d orbital set to 5.2 eV, 7.7 eV and 7.3 eV, respectively.
Fig. 2(b) shows X-ray diffraction spectra of Li  003) and (104) peaks that as the P 5+ doping amount was increased, the strong diffraction peak of the sample shifted to a smaller angle (Fig. 2c and d), indicating that the material lattice constant had become larger.After P 5+ doping, a, c and c/a of the Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 material all increased (a and c represent the spacing of the transition metals and layers of the lattice, respectively).Therefore, P 5+ may occur in combination with O 2À (1.40 Å), forming a larger radius species, such as a P 2 O 7 4À polyanion (thermochemical reaction) [21,22], that expands the inter-layer spacing of the Li layer, thereby helping to expand the two-dimension migration channel of Li + [23].
The calculated lattice parameters and unit cell volume are shown in Table 1 together with the experimental values for comparison.It can be seen from the optimized results that, taking the calculated value of the unit cell parameters of Li 1.2 Mn 0.6 Ni 0.2 O 2 as an example (a = 2.8689 Å, c = 14.2910Å, and Vol = 101.61Å 3 ), there was a deviation between the experimental and calculated values.This was because the generalized gradient approximation method generally overestimates the bond length and underestimates the bond energy in the calculation process.Therefore, the calculated theoretical data were slightly larger than the experimental values, but the overall trend was the same.The theoretical calculations confirmed that doping of 0.02 P increases the values of the unit cell parameters a and c of Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 (x = 0, 0.02) and the volume increases accordingly.As the volume of the crystal lattice is increased, the Li + migration barrier in the bulk phase decreases with increasing unit cell parameters a and c.Therefore, this doping scheme is beneficial in promoting the migration of Li + .

P doping analysis
Fig. 3 shows the scanning electron microscope (SEM) images of the Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 materials.Fig. 3a and b reveals that all the elements O, Mn, Ni and P were uniformly distributed in the Li 1.2-5 - x P x Ni 0.2 Mn 0.6 O 2 .The distinct lattice fringes between the (010) crystal planes increased from ~0.474 nm to 0.488 nm upon P doping, confirming that P may enter the lattice of the LMR material, expanding the interlayer spacing of the lithium layer (Fig. 3c-f), in agreement with the XRD data and previous reports [23,24].When x = 0.02, the discharge platform of the doped material was most gentle, the discharge capacity reached as high as 316.5 mAh g À1 and the coulombic efficiency of the initial charge and discharge was increased to 81.7%.When x = 0.03, the charge and discharge capacity of the doped material was the lowest, indicating that excessive doping destroys the microstructure of the material.Therefore, the optimal P doping amount was 0.02.Fig. 4(c) shows Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 with different P 5+ doping amounts cycled at 0.2 C, 0.5 C, 1 C, 2 C, 5 C and 0.2 C after being activated at 0.05 C for three cycles.The results show that under different discharge rates, the discharge capacity of the doped sample Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 (x = 0.01, 0.02, 0.03) was higher than that of the undoped sample Li 1.2 Mn 0.6 Ni 0.2 O 2 , confirming that P 5+ doping can effectively improve the rate performance of the LMR.The doping of P 5+ increased the interlayer spacing of the Li layer, stabilized the layered structure of the Li layer and promoted the diffusion of Li + .Therefore, when x = 0.01, 0.02 and 0.03, the discharge capacity of the material showed enhanced performance at different rates.When x was increased from 0, 0.01 to 0.02, the 0.02 doped material exhibited the best discharge capacity (0.05 C, 316.5 mAh g À1 ) and a high-rate discharge capacity (5 C, 158.0 mAh g À1 ), improving the capacity retention from 38% (x = 0), 47% (x = 0.01) to 50% (x = 0.02) (Fig. 4d).

Impedance analysis
To further explore the improvement of electrochemical performance of the Li 1.2 Mn 0.6 Ni 0.2 O 2 material after P 5+ doping, an AC impedance test was conducted after activating the material at 0.05 C for 3 cycles.The equivalent circuit was established by ZView fitting software and is shown in Fig. 5(a).The AC impedance curve comprised two parts.The high frequency area (semi-circular arc    Z where the constants F and R are 96,500 C mol À1 and 8.314 J K À1 mol À1 , respectively, T = 298 K, c is the concentration of Li + in the material, A is the area of the electrode, and the Warburg coefficient r can be calculated according to Eq. ( 2).The low-frequency zone impedance and angular frequency are plotted in Fig. 5(b).Values of D Li + calculated by Eq. ( 1) are 3.56 Â 10 À16 cm 2 s À1 , 1.06 Â 10 À15 cm 2 s À1 , 1.43 Â 10 À15 cm 2 s À1 and 6.00 Â 10 À16 cm 2 s À1 for x = 0, 0.01, 0.02 and 0.03, respectively.When x = 0.02, the value of r was the smallest and D Li + was the largest, consistent with the capacity results showing that this material had the optimal rate performance.Fig. 5c-f shows the energy band of Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 (x = 0.02) and Li 1.2 Mn 0.6 Ni 0.2 O 2 .These two materials are indirect band gap semiconductors since the top of their valence band and bottom of the conduction band are not at the same point in K space.Therefore, when the valence band electrons are excited by the conduction band transition, they experience band gaps of different widths.The width of the forbidden band influences the difficulty of the electronic transition and also affects the conductivity of the material during battery charging and discharging.It can be seen from Fig. 5c and d that the energies at the top of the valence band and bottom of the conduction band of the undoped material Li 1.2 Mn 0.6 Ni 0.2 O 2 were À0.24 eV and 0.308 eV, respectively, and the band gap was 0.548 eV.In contrast, the energies at the top of the valence band and bottom of the conduction band of Li 1.2-5x P x Ni 0.2 -Mn 0.6 O 2 (x = 0.02) were À0.159 eV and 0.252 eV, respectively, and the band gap was 0.411 eV, which is 0.137 eV smaller than that of the undoped material.Theoretical calculations showed that the conduction band energy before and after doping Li 1.2 Mn 0.6 -Ni 0.2 O 2 did not change much but the valence band height after doping was increased.Therefore, the forbidden band width was reduced, making the transfer of excited electrons from the valence band to conduction band easier, thereby improving the conductivity.In addition, compared with the undoped material, the gap between the adjacent energy bands of Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 (x = 0.02) was smaller and the overlap and intersection were significantly increased.Thus, the bulk of the doped material had more covalent character, which was due to the incorporation of P to form P-O. Therefore, the electronic conductivity of the material when the P doping amount was 0.02 was greatly improved, as reflected in the decreased impedance of Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 (x = 0.02) and improved charging and discharging performance.We also demonstrated experimentally that the electronic conductivity increased when the doping amount was increased.The resistance of Li 1.2-5x P x Ni 0.2 Mn 0.6 O 2 measured using a four-probe resistivity tester decreased from 1.23 X.cm À1 to 1.02 X.cm À1 when x was increased from 0 to 0.02.Fig. 5e and f shows that the total density of states (DOS) was near the Fermi level in Li 1.2 Mn 0.6 Ni 0.2 O 2 and with P doping (x = 0.02).The electron DOS and energy band structure had a high similar form, and the density of the energy band was reflected in the strength of the DOS peak.The peak position of the Li 1.2 Ni 0.2 Mn 0.6 O 2 total DOS before and after P doping ranged from À105 eV to 80 eV (Fig. 5e and f).There was very little change in the peak shape before and after doping, except P doping affected the peak intensity of the total DOS, indicating that the electronic DOS inside the material was significantly larger after P doping, enhancing the electronic conductivity of the material.

Thermal stability analysis of samples
Results from charging the battery with Li 1.2 Ni 0.2 Mn 0.6 O 2 or Li 1.1 -P 0.02 Ni 0.2 Mn 0.6 O 2 as electrode materials to a 4.8 V delithiation state at a current of 0.05 C and using differential scanning calorimetetry (DSC) to measure its thermal behavior are shown in Fig. 6(a).It can be seen that the heat release of the Li 1.2 Ni 0.2 Mn 0.6 O 2 electrode material was concentrated at 258 °C.In contrast, the exothermic reaction of the Li 1.1 P 0.02 Ni 0.2 Mn 0.6 O 2 electrode material was dispersed at two different temperatures, i.e., 264 °C and 315 °C.It is possible that the exothermic reaction at 264 °C occurred because Ni 4+ is unstable in the high oxidation state, whereas the exothermic peak at 315 °C was likely due to strong P-O bonds inhibiting oxygen in the crystal structure (Fig. 6a).
Fig. 6b and c shows the total DOS (up-spin) of Li 1.2 Ni 0.2 Mn 0.6 O 2 before and after P doping (x = 0, 0.02) compared to the atomic subwave density (up-spin).Firstly, the total DOS after doping was mainly concentrated in the energy range À8 eV to 5 eV and included s, p and d states.The band overlap in this range was stronger, indicating that the covalent strength between atoms was enhanced.Secondly, in the range À8 eV to 5 eV, the total DOS was mainly composed of s (i.e., Li 1s, P 2s, Mn 4s and O 2s orbitals), p (i.e., part of Mn 2p, P 2p and O 2p orbitals) and d (small amount of Mn 3d and Ni 3d orbitals) states.In this energy range, the overlap between the p and d states was strong, indicating that the overlap of the electron orbitals generated strong covalent bonds, enhancing the stability of the crystal structure.In addition to the presence of enhanced Mn-O, the 2p orbital of P and that of O resonated and the peak shape was sharp, indicating that O and P formed a strong covalent bond after doping, which is helpful to bind oxygen atoms.Moreover, the DOS peak of O near the Fermi energy level shifted slightly to both ends and the electron mobility of O became weaker.Therefore, P doping effectively suppressed the charge compensation of O and precipitation of lattice oxygen and weakened O 2À electron migration at the Fermi energy level.Although the P doped LMR can improve the rate performance, it is difficult to strongly enhance the surface stability of the materials.Therefore, solid state electrolyte, a fast ionic conductor, was used to engineer the LMR for long service life of battery.Fig. 7(a) shows X-ray diffraction (XRD) patterns of 0 wt% and 2 wt% Li 6.24 Al 0.12 La 3 Zr 1.8 Mo 0.2 O 12 (LALZMO) coated Li 1.1 P 0.02 Mn 0.6 -Ni 0.2 O 2 (LPMNO) materials.It can be seen from the strong diffraction peaks of the sample that the material belonged to the a-NaFeO 2 layered structure of the R À 3m space group.The coated sample also showed weak peaks at around 30°, which were attributed to characteristic peaks of the composite compound Zr 1-x M x O 2 .The peaks of (006)/(102) and (108)/(110) were weaker but clearly split, indicating that the coated material still had a good layered structure.EDS further confirmed a homogeneous LALZMO coating on the surface of LPMNO, as shown in Fig. 7(b).TEM image showed that the LALZMO coating on the surface of LPMNO was up to 15 nm thick (Fig. 7c).

Surface engineering of Li
The experiments showed that the cycle stability of the 2 wt% LALZMO-coated modified LPMNO material was significantly improved.After 200 cycles, the capacity retention rate of the coated sample decreased to 90.5%, whereas after about 125 cycles, the capacity of the uncoated sample dropped sharply because of the electrolyte consumption, demonstrating the aggravated side reactions at the cathode interface (Fig. 7d).Fig. 7(e) shows the rate performances of 2 wt% LALZMO coated modified LPMNO material and uncoated material at 0.2 C, 0.5 C, 1 C, 2 C and 5 C. The discharge capacity of the LPMNO material showed good performance at different rates.The initial discharge capacity was 326.9 mAh g À1 at 0.05 C and 169.3 mAh g À1 at 5 C.
The 2 wt% LALZMO coating clearly reduced the interface impedance from 51.8 X to 46.5 X with an improvement of rate performance from 83.7% to 85.1% at 0.05 C and 0.2 C, respectively.This was because LALZMO is a fast ion conductor.On the one hand, the coating layer stabilizes the structure of the cathode material, whereas on the other hand, it inhibits the oxidative decomposition of the electrolyte and erosion of the matrix material, reducing the deposition of reaction products on the surface of the material that prevent Li + transport, and thereby promoting the migration of Li + (Fig. 7f).The Li + diffusion coefficient of the 2 wt% LALZMO coated modified LPMNO material was further improved, reaching 2.18 Â 10 À15 cm 2 s À1 .

Conclusion
We demonstrated that phosphorus doping of lithium-and manganese-rich layered oxide (LMR) cathode materials can increase the intrinsic Li + conductivity (by increasing the interlayer spacing of the lithium layer), electron transport (by increasing the electronic density) and structural stability in the bulk material, resulting in significant improvements of the rate and safety performances.The experimental results were consistent with firstprinciples calculations.We believe this work opens new avenues for further adjusting fast charge cathode or anode materials.Future research will focus on element doping for lithium layers or other metal layers (e.g., sodium, potassium, magnesium) for advanced batteries.

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.

Fig. 3 .
Fig. 3. SEM and TEM images of Li 1.1 P 0.02 Ni 0.2 Mn 0.6 O 2 samples.(a and b) EDS elemental mapping images of Li 1.1 P 0.02 Ni 0.2 Mn 0.6 O 2 .High resolution TEM images layered structure of (c) Li 1.2 Mn 0.6 Ni 0.2 O 2 and (e) Li 1.1 P 0.02 Ni 0.2 Mn 0.6 O 2 with line scans of (d) and (f) along with the (010) direction, respectively.