SEARCH

SEARCH BY CITATION

Keywords:

  • degradation mechanisms;
  • layered structures;
  • lithium ion battery;
  • cathode materials

Abstract

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Results and Discussion
  5. 3 Conclusion
  6. 4 Experimental Section
  7. Acknowledgements
  8. Supporting Information

LiNixCoyMnzO2 (NCM, 0 ≤ x,y,z < 1) has become one of the most important cathode materials for next-generation lithium (Li) ion batteries due to its high capacity and cost effectiveness compared with LiCoO2. However, the high-voltage operation of NCM (>4.3 V) required for high capacity is inevitably accompanied by a more rapid capacity fade over numerous cycles. Here, the degradation mechanisms of LiNi0.5Co0.2Mn0.3O2 are investigated during cycling under various cutoff voltage conditions. The surface lattice structures of LiNi0.5Co0.2Mn0.3O2 are observed to suffer from an irreversible transformation; the type of transformation depends on the cutoff voltage conditions. The surface of the pristine rhombohedral phase tends to transform into a mixture of spinel and rock salt phases. Moreover, the formation of the rock salt phase is more dominant under a higher voltage operation (≈4.8 V), which is attributable to the highly oxidative environment that triggers the oxygen loss from the surface of the material. The presence of the ionically insulating rock salt phase may result in sluggish kinetics, thus deteriorating the capacity retention. This implies that the prevention of surface structural degradation can provide the means to produce and retain high capacity, as well as stabilize the cycle life of LiNi0.5Co0.2Mn0.3O2 during high-voltage operations.

1 Introduction

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Results and Discussion
  5. 3 Conclusion
  6. 4 Experimental Section
  7. Acknowledgements
  8. Supporting Information

Demands for lithium (Li)-ion batteries currently reach from portable electronics to large-scale applications including energy storage systems and electric vehicles.[1, 2] Higher energy density combined with longer cycle life is one of the key requirements that need to be addressed in these new applications.[3, 4] Layered Li transition metal oxides have served as the most important cathode materials for high-energy batteries due to their large theoretical capacity (≈280 mAh g-1) compared with those of olivine (≈170 mAh g-1) or spinel (≈150 mAh g-1) materials.[5-7] LiCoO2 delivers a high working voltage (≈3.9 V) with excellent rate capability and has been the material most widely used for commercial batteries. However, the limitations of LiCoO2, such as a small practical capacity (≈150 mAh g-1), unreliable safety, and the high cost of cobalt (Co), have led researchers to focus on other layered materials with less Co. While the single-component systems, including LiMnO2 and LiNiO2, also suffer from intrinsic problems such as poor cyclability and rate capability, as well as complexity in preparation, three-component layered LiNixCoyMnzO2 (NCM, 0 ≤ x,y,z < 1) has exhibited promising electrochemical properties, which vary depending on the composition of Ni, Co and Mn in the structure.[7-12] Generally, high Ni content in the NCM contributes to a higher capacity at the expense of safety and complexity in preparation; high Mn content enhances the structural stability at the expense of capacity and high Co content improves the rate performance and processing ability for a greater cost.

To date, two representative NCM layered materials have been successfully adopted in commercial Li-ion batteries: LiNi1/3Co1/3Mn1/3O2 (NCM111) and LiNi0.5Co0.2Mn0.3O2 (NCM523). Less use of Co in these materials, compared with LiCoO2, has contributed to a reduction in cost and improved safety for the new battery system. Nevertheless, sufficient energy density has yet to be achieved, even though NCM layered materials are capable of delivering higher capacity due to a larger range of reversible Li insertion and extraction. To acquire high capacity from NCM layered materials, operation in the high-voltage range (≈4.5 V) is required, which inevitably results in rapid capacity fade over numerous cycles. Various surface coatings have improved the high-voltage cycling stability of NCM materials; however, a lack of understanding still exists regarding the origin of the accelerated degradation of cycle performance during high-voltage operation.[13-20]

In this study, we investigated the degradation mechanisms of a LiNi0.5Co0.2Mn0.3O2 (NCM523) cathode, one of the most important commercialized NCM materials, as a function of charge cutoff voltages, with detailed structural analyses of both the bulk and surface. Our results indicate that surface structural degradation is critically affected by the cutoff voltage, which results in the formation of an ionically insulating layer on the surface of NCM, thus reducing capacity retention.

2 Results and Discussion

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Results and Discussion
  5. 3 Conclusion
  6. 4 Experimental Section
  7. Acknowledgements
  8. Supporting Information

The phase purity of NCM523 was first confirmed by X-ray diffraction (XRD; Supporting Information, Figure S1). XRD data showed a typical pattern for the α-NaFeO2 structure of the Rinline imagem space group. No impurities were detected. Clear peak separation of (006)/(012) and (108)/(110) indicated a well-ordered crystalline layered structure.[21] The lattice parameters were obtained by Rietveld refinement, as shown in Figure S1 (Supporting Information), which agreed well with previous reports.[14] The measured transition metal composition by ICP-AES was Ni:Co:Mn = 0.482:0.222:0.296, comparable to the target composition of NCM523. The primary particle was ≈500 nm in diameter and composed of a single crystal phase, as revealed by the TEM diffraction pattern (DP; inset figure of Supporting Information, Figure S1).

Three identical cells were fabricated using the NCM523 cathode and operated in various voltage ranges: 3.0–4.3 V, 3.0–4.5 V and 3.0–4.8 V. Figure 1a,b show that increasing the charge cutoff voltages led to higher reversible capacities for the first few cycles, followed by faster capacity degradation. While NCM523 could deliver a cycle retention of ≈95% after 50 cycles with 4.3 V operation, the discharge capacity decreased to 72% and 61% of the initial capacity under 4.5 V and 4.8 V operation, respectively.

image

Figure 1. a) Cycle performance and b) discharge capacity retention at various cutoff voltages during 50 cycles. c) Electrochemical impedance spectroscopy (EIS) profiles of electrodes in coin cells of the as prepared sample and samples after 50 cycles. d) Equivalent circuit for electrodes. Rs: solution resistance; Rf: surface film resistance; and Rct: charge transfer resistance.

Download figure to PowerPoint

To confirm the effect of transition metals dissolution in electrolyte on capacity retention, the amounts of Ni, Co, and Mn ions in electrolyte were measured. Transition metal dissolution could contribute to capacity decay, if NCM cathode was exposed for a long time at high voltage (>4.5 V).[22] However, the amounts of transition metal ions in electrolyte of 4.3 V, 4.5 V, and 4.8 V cycled samples were similar, which indicates that the main cause of NCM523 degradation at high voltage was not related with transition metal dissolution.

Figure 1c shows EIS measurements, which clearly reveal the relationship between impedance change and the degree of cycle degradation. The as-prepared electrode exhibited one semicircle (Figure 1c), corresponding to the charge transfer resistance (Rct). The other electrodes exhibited two semicircles, attributed to Rct, and an additional surface film resistance (Rf). The Rf resistance is generally caused by the formation of an insulating solid electrolyte interphase (SEI) layer or the deposition of organic compounds on the surface of the electrode by electrolytic decomposition.[23, 24] For the detailed analysis, the equivalent circuit in Figure 1d was used. From the inset of Figure 1c, the Rf of the 4.8 V cycled sample exhibited a resistance (84.82 Ω) three times larger than those of the 4.3 V (28.12 Ω) or 4.5 V (29.38 Ω) cycled samples. This may indicate the formation of a thicker surface film during cycling at 4.8 V. In contrast to the relatively small variation in Rf among samples, Rct underwent a substantially larger increase during cycling at high voltages when the Rct increased by 3–13 times at 4.5 V (465.5 Ω) and 4.8 V (2305.2 Ω), respectively, while the 4.3 V cycled sample exhibited a nearly identical Rct (171.2 Ω) to that of the as-prepared electrode (175.7 Ω). All the values of resistance are listed in the Supporting Information (Table S2). The supplementary electrochemical cycling of samples after 50 cycles at a slower current rate revealed that the capacity nearly recovered to its initial capacity, which indicated that the electrode material remained electrochemically active even after cycling (Supporting Information, Figure S2). Thus, the cycle degradation at high voltage is closely correlated with the increase in the charge transfer resistance of the electrodes. In the following, we discuss the origin of the increased charge transfer resistance of the high-voltage cycled samples through analyses that consider both the bulk and surface structure.

To investigate the origin of the accelerated capacity fading above the 4.5 V cutoff, we first examined the bulk structure after cycling. Each sample at the discharged state after 50 cycles was collected and analyzed. Figure 2 compares the XRD patterns of cycled samples. Regardless of the cutoff voltage, all of the samples retained their layered structure; no indication of significant structural degradation or formation of any other new phases was noted. Our results showed a slight difference in the lattice parameters (Supporting Information, Figure S3), but this was attributable to the differing amounts of remaining Li contents in each discharged sample (Supporting Information, Figure S4). The Li deficiency in the discharged (to 3.0 V) samples implied that the electrode polarization had been built up over numerous cycles, preventing full lithiation of the cathode even at 3.0 V. Note that a higher Li deficiency was observed for samples that were operated at higher voltages. This indicates that the electrode polarization becomes more severe when cycled at high voltages, which is consistent with the higher impedance value observed in Figure 1c.

image

Figure 2. XRD patterns of the as-prepared electrode before and after 50 cycles.

Download figure to PowerPoint

To probe the local structural change of the samples, transmission electron microscopy (TEM) analyses were performed. We acquired selected area diffraction (SAED) patterns from both the bulk and surface regions of a 4.8 V cycled sample (Figure 3). Figure 3a shows the hexagonal diffraction pattern from the bulk region, without any trace of a secondary phase, consistent with the XRD results. However, additional diffraction spots were detected for the surface region, as shown in Figure 3b. For clarity, the peak intensity was integrated following the bright lines, which indicated the presence of additional peaks shown in the inset of the figure. The additional diffraction spots could be indexed as (–220)s and (11–1)s in Figure 3b, corresponding to a spinel structure. The partial phase transformation from the layered structure to a spinel structure in this material is reminiscent of the commonly observed transformation of layered LiMnO2 into a spinel structure during electrochemical cycling.[8, 25] In general, layered materials are thermodynamically prone to transform into the spinel phase when the Li contents are reduced to half that of the original layered structure.[26] While the ease of layered-spinel phase transformation of Li0.5MO2 (M = transition metal) is dependent on the octahedral site stabilization energy (OSSE) of the transition metal, Ni, Co, and Mn4+ ions in NCM materials have been regarded as stable against phase transformation.[27] Thus, the partial phase transformation in the surface region was attributed to the prolonged Li deficiency near the surface of the NCM523, which is generally observed in layered cathode material, even in the discharged state.[28, 29] The surface lithium deficient state can be easly formed during electrochemcial cycing as was shown in LiNixCoyAl1-x-yO2 layered cathode material.[29]

image

Figure 3. Diffraction patterns at the zone axis [–1–21]R after 50 cycles under 3.0–4.8 V conditions for the a) bulk and b) surface regions. The figure insets show bright intensity maps of the spots along yellow and blue lines. R represents a peak from the layered phase; S represents peaks from the spinel phase.

Download figure to PowerPoint

Extensive high-resolution TEM (HR-TEM) and fast Fourier transformation (FFT) studies were performed to estimate the range where the phase transformation occurred. To avoid any confusion, various regions in the sample were examined, with consistent results. Figure 4a shows the representative structural change in the NCM cycled at a 4.8 V cutoff voltage. Our results showed that the change in the structure was mainly localized on the surface (Figure 4c), while the bulk region remained in the rhombohedral phase (Figure 4b). Diffraction spots from the spinel phase were detected in Region 2, along with the layered phase, which was consistent with the SAED pattern shown in Figure 3a,3b. The spinel phase was continuously observed over distances 12–15 nm from the surface. Note that in the narrow region ≈2–3 nm from the surface, the cubic phase began to appear as determined by the lattice images and FFTs in Figure 4d,e. In Regions 3 and 4, only two strong diffraction spots were detected. The reduction in the number of diffraction spots was indicative of the formation of the high symmetry rock salt phase. The d-spacing of the rock salt phase, acquired from the lattice image and FFT, was ≈2.45 Å, which was comparable to those of the (11–1) planes of the NiO (2.42 Å) and CoO (2.46 Å) rock salt phase. At this time, however, whether the observed rock salt phase was NiO or CoO was not clear. From the lattice parameter (2.45 Å) that lies between those of NiO (2.42 Å) and CoO (2.46 Å), we speculated that it is likely a solid solution between NiO and CoO.

image

Figure 4. HR-TEM images and FFTs after 50 cycles under 3.0–4.8 V conditions. a) Lattice image of the surface region where (b–e) correspond to the FFTs of Regions 1–4, respectively. (11–1)c is the diffraction spot of the rock salt phase of the metal monoxide.

Download figure to PowerPoint

The probing of local environments from energy electron loss spectroscopy (EELS) measurements further indicated the formation of the rock salt phase. EELS spectra of the O–K edge of the as-prepared and tested electrodes were collected, as shown in Figure S5 (Supporting Information). The oxygen EELS peaks in the Supporting Information show that two different oxygen local environments arose when cycled at 4.8 V. The main oxygen EELS peak was observed broadly at 547 eV in both the as-prepared and 4.5 V cycled samples, having similar shapes. However, the oxygen peak was clearly split into two peaks of 547 eV and 543 eV after 4.8 V cycling. The peak near 543 eV corresponded to the characteristic oxygen peak of NiO. This observation agreed well with the formation of the rock salt phase suggested by TEM results. The formation of the rock salt phase in NCM523 during 4.8 V operation is similar to the general characteristics of Ni-rich layered systems, such as LiNi0.73Co0.17Al0.10O2, which tend to form NiO on the surface after cycle testing.[28] In layered cathode materials, the rock salt phase can form through oxygen evolution when Co3+/4+ t2g or the Ni3+/4+ eg orbital substantially overlaps the oxygen 2p orbital.[30, 31] Oxygen evolution in NCM523 is also likely to occur during high-voltage operation, when Ni and Co are in a highly oxidative environment.

The rock salt and spinel phases were also detected at the surface of the 4.5 V cycled sample shown in Figure 5. However, from extensive sample examination, we found that the range of cubic and spinel phases differed from that of the 4.8 V cycled sample. In the 4.5 V sample, the cubic phase was only partially observed on the surface, while the spinel phase was dominant for the outermost surfaces. In contrast, the cubic phase continuously encircled the spinel phase in the 4.8 V cycled sample, which implies that the higher cutoff voltage promotes the formation of the cubic phase during cycling. An exceptionally oxidative environment imposed by the high-voltage charge is likely to trigger a severe oxygen evolution from the electrode and induce phase transformation. Note that the 4.3 V cycled electrode did not show any trace of phase transformation, contrary to the 4.5 V and 4.8 V cycled electrodes (Supporting Information, Figure S6,S7).

image

Figure 5. HR-TEM images and FFTs after 50 cycles under 3.0–4.5 V conditions. a) Lattice image of the surface region where (b–e) correspond to the FFTs of Regions 1–4, respectively.

Download figure to PowerPoint

Additionally, we investigated the effect of current rate on the degradation behavior and phase transformation in the surface region. The current rate was controlled as 1000 mA g-1 (3.63 C) which is about ten times larger than the previous condition (110 mA g-1, 0.4 C). At the 3.0–4.5 V voltage range, it showed rapid capacity decay during electrochemical cycling (Supporting Information, Figure S8). Furthermore, the phase transformation from rhombohedral to spinel on surface was more extensively detected on the surface. The spinel phase occupied larger area (30–35 nm) than the case of 110 mA g-1 current rate (15–20 nm) 4.5 V cycled electrode (Supproting Information, Figure S9). This indicates that high current rate at high voltage promotes the degradation and phase transformation on the surface.

The degradation mechanism of NCM523 appeared to be different from the previous reports on that of NCM111. Capacity fade in NCM111 was attributable to the phase transformation of O3 to the O1 phase for a highly charged state via the generation of stacking faults between the two phases.[30, 32, 33] However, NCM523 did not show any evidence of the O1 phase, after cycling or during its highly delithiated state. This is clearly shown in the TEM results in the Supporting Information (Figure S10). We speculate that the absence of the O1 phase in NCM523 was related to possibly more Li–Ni site exchange in the Ni-rich layered compounds, which would prohibit the sliding of the slabs. This implies that the degradation mechanisms of the NCM material can vary with the composition of Ni, Co, and Mn, and should be carefully considered. Based on our results, we propose the following degradation mechanism of LiNi0.5Co0.2Mn0.3O2 at high voltages depicted in Figure 6. When the NCM523 electrode is cycled with a 4.5 V cutoff, the surface of the electrode suffers from a phase transformation mainly to the spinel phase, with a trace of rock salt formation. If the cutoff voltage increases to 4.8 V, then the highly oxidative environment promotes the formation of the rock salt phase more universally than the 4.5 V cutoff. As a consequence, the rock salt phase encircles the spinel and rhombohedral phases. The range of the phase-transformed regions were nearly the same (15–20 nm) for both 4.8 V and 4.5 V cutoff conditions, but with varying fractions of the spinel and rock salt phase. We believe that these new phases are critical to the increase in charge transfer resistance. The presence of the ionically blocking rock salt phase inhibits the motion of Li ions. The variation of charge transfer resistance between 4.5 V and 4.8 V in the electrochemical experiments was attributable to the rock salt phase that existed more profoundly on the surface of the 4.8 V samples than the 4.5 V samples.

image

Figure 6. Degradation mechanisms of LiNi0.5Co0.2Mn0.3O2 and phase transformation after cycle tests under high-voltage conditions.

Download figure to PowerPoint

3 Conclusion

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Results and Discussion
  5. 3 Conclusion
  6. 4 Experimental Section
  7. Acknowledgements
  8. Supporting Information

The degradation mechanism of the NCM523 electrode under high-voltage battery operation was investigated by combined XRD and TEM structural analysis. Our results indicate that the structural degradation occurred mainly on the surface of the materials where the phase transformation from rhombohedral to spinel was dominant. However, depending on the cutoff voltage, the formation of the rock salt phase could also occur, which was suspected to increase the charge transfer resistance. The degradation mechanism of NCM523 differed from that of NCM111, which involves the formation of the O1 phase in the highly delithiated state. This indicated that the degradation mechanisms of the NCM material should be carefully considered for different compositions.

4 Experimental Section

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Results and Discussion
  5. 3 Conclusion
  6. 4 Experimental Section
  7. Acknowledgements
  8. Supporting Information

Characterization of LiNi0.5Co0.2Mn0.3O2: The LiNi0.5Co0.2Mn0.3O2 sample was supplied by Samsung Fine Chemicals (Daejeon, Korea). The powder sample was characterized using a D2 PHASER (Bruker, Bremen, Germany) equipped with Cu-Kα radiation (λ = 1.54178 Å) at a scanning speed 0.2° min–1 in the 2θ range of 10–70° . The stoichiometry of the compound was confirmed by inductively coupled plasma atomic emission spectroscopy (ICP-AES; Polyscan 60E; Thermo Jarrell Ash, Franklin, MA, USA). The shape and size of the primary particles were examined using a 200 kV field emission transmission electron microscope (FE-TEM; Techani-F20; FEI, Hillsboro, OR, USA).

Electrochemical Analysis: Test electrodes were fabricated by the following sequence. A slurry of 70 wt% LiNi0.5Co0.2Mn0.3O2, 20 wt% carbon black (Super-P; Timcal, Bodio, Switzerland), and 10 wt% polyvinylidene fluoride (PVDF) binder, dissolved in N-methyl-1,2,-pyrrolidone (NMP, 99.5%; Sigma-Aldrich, St. Louis, MO, USA), was cast onto aluminum foils using a doctor blade method. NMP was evaporated overnight at 70 °C in a vacuum oven. Coin cells (CR2016; Hohsen, Osaka, Japan) were assembled with the LiNi0.5Co0.2Mn0.3O2 electrode, Li counter electrode, a separator (2400; Celgard, Tokyo, Japan), and a 1 M solution of LiPF6 in a mixture of ethyl carbonate/dimethyl carbonate (EC/DMC, 1:1 v/v) in an argon-filled glove box. The galvanostatic charge/discharge profile was measured over the voltage ranges of 3.0–4.3, 3.0–4.5, and 3.0–4.8 V by a potentiogalvanostat (WBCS 3000; WonA Tech, Seoul, Korea) under 110 mA g–1 (0.4 C) at room temperature for 50 cycles. After the 50 cycles were completed, electrochemical impedance spectroscopy (EIS) measurements were performed on the tested cells using an impedance analyzer (ZIVE SP2; WonA Tech) at room temperature in the frequency range of 1 MHz to 1 mHz.

Ex Situ Structural Analysis: After 50 cycles with different cutoff voltage conditions (3.0–4.3, 3.0–4.5, and 3.0–4.8 V), three electrodes were collected for structural analysis. The fully discharged electrodes were disassembled from coin cells, rinsed with DMC several times, and dried at room temperature in a vacuum oven for XRD measurements. The electrolyte soaked in glass fiber filter (GFF) was collected and diluted in dimethyl carbonate (DMC) for transition metal dissolution measurement by inductively coupled plasma mass spectroscopy (ICP-MS). Lattice parameters of electrodes were determined by the Rietveld method using Fullprof software.[34] For HR-TEM analysis, the active material was racked out and dispersed in ethanol using a sonicator before being transferred onto a lacey carbon-supported Cu grid. HR-TEM images and EELS spectra of the samples were recorded using a 200 kV FE-TEM. EELS spectra were acquired from the outermost surface region within a square area of ≈3 nm × 3 nm, with an acquisition time of 2 s. The collection angle and convergence angles were 40 mrad and 30 mrad, respectively. The exposure time of the sample to the electron beam was minimized. O–K edge spectra were collected.

Acknowledgements

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Results and Discussion
  5. 3 Conclusion
  6. 4 Experimental Section
  7. Acknowledgements
  8. Supporting Information

This work was supported by Samsung Fine Chemical and the Energy Efficiency & Resources of the Korea Institute of Energy Technology Evaluation and Planning(20112010100140) grant funded by the Korea government Ministry of Knowledge Economy. This work was also supported by the Research Center Program of IBS(Institute for Basic Science) in Korea and the National Research Foundation of Korea Grant funded by the Korean Government (MEST) (NRF-2009-0094219)

Supporting Information

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Results and Discussion
  5. 3 Conclusion
  6. 4 Experimental Section
  7. Acknowledgements
  8. Supporting Information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
aenm201300787-sup-0001-S1.pdf2964KSupplementary

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.