The study of electrochemical cycle for LiCoO₂ by dual-mode EPR

Abstract

Dual-mode electron paramagnetic resonance (EPR) spectroscopy was employed to analyze redox mechanisms in lithium cobalt oxide LiCoO₂ (LCO) cathode material during delithiation and lithiation. It was found that the O3-II could not fully convert back to the pristine O3–I phase while oxygen vacancies quickly generate and accumulate during the cycling. Our study paves the way for better understanding the doping effects of different elements on LiCoO₂ in the future.

Graphical abstract

Dual-mode EPR was employed to study the full deintercalation/intercalation process for LiCoO₂.

1. Introduction

LiCoO₂ (LCO) is the first commercialized and currently the most common cathode material in commercial lithium-ion batteries for portable electronics, primarily due to its high volume energy density and stable cycling [1,2]. LiCoO₂ is usually cycled with a cutoff voltage of <4.35 V (vs. Li/Li⁺), leading to a limited capacity of <165 mAh g⁻¹ to keep the industrial-level cycling stability, which is far from its theoretical capacity of 274 mAh g⁻¹ [2]. Therefore, it still provides large potential to increase the energy density of LiCoO₂ by increasing the charging voltage. However, deep cycling causes lattice distortions, which triggers severe electrochemical performance fading with cycling.

The deintercalation/intercalation process has been studied by various techniques, such as ⁷Li nuclear magnetic resonance (NMR) [3], ¹⁷O NMR [4], X-ray diffraction (XRD) [[5], [6], [7]], first-principles calculations [8,9], and magnetic analysis [[10], [11], [12]], etc. The pristine LiCoO₂ is O3–I phase, and after delithiation the O3–I phase will convert to O3-II phase for Li_xCoO₂ with a two-phase domain of O3–I and O3-II phase in the region of 0.75 < x < 0.98 with the voltage U < 3.96 V (Fig. 1). For the region of 0.5< x < 0.75, the material becomes a single component with only O3-II phase with the voltage of 3.96 V < U < 4.2 V. Afterwards, LCO will convert to O3M phase when 0.2 < x < 0.5 and the voltage U > 4.2 V. After charging to the voltage U > 4.5 V and x < 0.2, LiCoO₂ undergoes gradual phase transition from O3M to H1-3 phase (a hybrid structure of O1 and O3). Such phase transition causes large volume expansion, mechanical fractures due to inhomogeneous stress, and irreversible structure collapse. When the process of Li⁺ ion reinsertion begins, H1-3 phase would convert back to O3M phase, then O3-II phase, and finally O3–I phase.

Fig. 1.

The charge-discharge profiles of LiCoO₂ electrodes with the voltage window of 3.0–4.6 V (vs. Li/Li+) at a current density of 100 mA g⁻¹. The circles denote the different states of charge (SoCs) for the EPR experiments. The pristine sample has the voltage of ∼3.33 V.

Electron paramagnetic resonance (EPR) has been also employed to study LiCoO₂. Most of the work only analyzed the pristine sample and did not analyze the delithiated sample Li_xCoO₂ (x < 1) [13,14]. Ni³⁺ or Fe³⁺ doping was also used as spin probes to study LCO [15,16]. In particular, the very recent work by Niemöller [17] has performed the conventional ex-situ perpendicular-mode EPR analysis with two voltages, one at ∼3.0 V and the other at ∼5.0 V, but they did not monitor the full deintercalation/intercalation process. They employed in-situ perpendicular-mode EPR to monitor the full process at room temperature but the EPR signal after U > 2.3V is very weak and difficult to analyze [17].

In this paper, we employ dual-mode EPR at low temperature of 2 K to study the full deintercalation/intercalation process for LiCoO₂, and try to assign all the resonances. Dual-mode EPR is a valuable technique for studying both half-integer and integer electron spin systems, and is particularly useful for studying transition metals in lithium-ion batteries. The dual-mode EPR technique encompasses both perpendicular and parallel mode EPR. This study will pave the way for understanding the doping effects of different elements on LiCoO₂ in the future.

2. Experimental

Material synthesis. LiCoO₂ was synthesized by solid-state reaction using Li₂CO₃ (Shenzhen Kejing) and Co₃O₄ (Aladdin, 99.99%) as precursors. The stoichiometric amounts of raw materials with an excess of 5 mol% Li₂CO₃ were ground via Agate mortar to mix thoroughly, and then the mixture was pretreated at 600 °C for 3 h first, followed by calcination at 900 °C for 12 h.

Electrochemical characterization. The working electrodes were fabricated using 80% active material, 10% carbon nanotubes or super P as the conductive additive, and 10 wt% polyvinylidenefluoride (PVDF) as the binder with N-methyl-2-pyrrolidone (NMP) as the solvent. The active mass loading of LCO electrodes was ∼3 mg/cm². The electrochemical performance was assessed by 2032-type coin cells using lithium metal as the anode, Celgard 2500 polypropylene (PP) membrane as the separator, 1.0 M LiPF6 in a mixture of EC-DMC-EMC (1:1:1 by volume, EC: ethylene carbonate, DMC: dimethyl carbonate, EMC: ethylmethyl carbonate) with 5% fluoroethylene carbonate (FEC) as the electrolyte. The cells were assembled in an argon-filled glovebox. The galvanostatic charge/discharge measurements were carried out on a LAND CT2001A battery test system (Wuhan, China). For all ex-situ characterizations, the samples at different states of charge (SoCs) were prepared by disassembling the cycled batteries inside an argon-filled glovebox, washing thoroughly with DMC and then drying for at least 12 h.

EPR experiments. Continuous-wave EPR spectra were recorded on a Bruker EMX plus 10/12 spectrometer in X-band (9.4 GHz) at 2 K, with a dual-mode cavity and an Oxford ESR910 liquid helium cryostat. The microwave frequency was 9.647 G. The modulation was set up as the amplitude of 2G at a modulation frequency of 100 kHz. The microwave power was set to 0.1 mW.

3. Results and discussion

The charge-discharge profiles of LiCoO₂ electrodes with the voltage window of 3.0–4.6 V at a current density of 100 mA g⁻¹ are shown in Fig. 1. With this current density, the LCO could achieve the capacity of ca. 210 mAh g⁻¹. Several states of charge (SoCs) were selected for further dual-mode EPR analysis, which was circled in Fig. 1. The voltage of 4.2 V at charging was selected because it is an important point for the phase transition from O3-II to O3M, while the voltage of 3.8 V at discharging was selected because the discharging process is almost done. For convenience, the traditional perpendicular-mode EPR will be referred to as ⊥-EPR, while the parallel-mode EPR will be referred to as //-EPR.

The LCO ⊥-EPR of pristine LCO (Fig. 2(a1)) shows a very complicated EPR spectrum, while the charged and discharged ⊥-EPR spectra are relatively simple (Fig. 2(a2) to Fig. 2(a6)). The LCO ⊥-EPR spectrum of the pristine sample (Fig. 2(a1)) is similar to the one reported by Delmas and therefore we follow the assignments of Delmas [13]. The resonances at g∼24.6, g∼2.00 and g∼1.908 in Fig. 2(a1) were ascribed to the impurity of Co₃O₄, and the line at g∼2.12 can be attributed to paramagnetic defects in Co₃O₄. The features in the magnetic field region 500–2800 G in Fig. 2(a1) correspond to the hyperfine structure associated with high-spin Co²⁺ in a distorted octahedron and these signals possibly arise from the creation of Co²⁺ surface centers in LiCoO₂. The line at g∼24.6 at //-EPR (Fig. 2(b1)) should come from Co₃O₄. The line at g∼4.84 at //-EPR (Fig. 2(b1)) should also come from high spin Co²⁺ in Co₃O₄. Here it should be pointed out that for some high-spin Co(II) systems, //-EPR always shows the signal for Co(II) at g∼4 to g∼5 because the spin-orbit coupling (SOC) mixes excited state character into the ground state [18]. This corresponds to a classic spin S = 3/2 paramagnet with an M_S = ±1/2 ground state with $g_{\perp }^{\text{eff}}$ ∼4.5, but the resonance of $g_{//}^{\text{eff}}$ ∼2 is too small to be observable [19]. It should be mentioned that we have analyzed many LiCoO₂ samples from many companies, and all these samples contain Co₃O₄.

Fig. 2.

The perpendicular-mode EPR spectra (a) and parallel-mode EPR spectra (b) collected at various charge states during the initial cycle. C4.2 denotes charging to 4.2 V and D4.2 denotes discharging to 4.2 V, and so on.

When charged to 4.2 V in the delithiation process with Co³⁺→Co⁴⁺, most peaks disappear in ⊥-EPR spectrum (Fig. 2(a2)), and two resonances including one broad peak and one narrow peak could be observed. The broad peak at g∼2.2 with the full width at half height (FWHH) of ca. 2000 G in ⊥-EPR spectra is consistent with a spin S = 1/2 species, as expected for Co⁴⁺. The narrow peak at g∼2.00 in ⊥-EPR spectrum could be ascribed to the oxygen vacancies (V_O•). It should be noted that the carbon nanotubes were used here as conductive additives for EPR experiments and carbon nanotubes have no EPR signal. In //-EPR spectrum (Fig. 2(b2)), the signal from Co₃O₄ at g∼24.6 could be still observed, but showing the ferromagnetic feature. In //-EPR spectrum (Fig. 2(b2)), the line at g∼4.08 could be ascribed to the high-spin Co²⁺ species on the surface. When charged to 4.6 V and then discharged to 3.0 V, the ⊥-EPR and //-EPR spectra are almost the same (Fig. 2(a3) to Fig. 2(a6), Fig.2(b3) to Fig. 2(b6)). During this period, Co₃O₄ signal is eliminated, suggesting that the Co²⁺ in Co₃O₄ has been also converted to Co⁴⁺. Overall, after being charged, most of these paramagnetic centers of Co²⁺ in Co₃O₄ are converted to the Co⁴⁺, which is indicated by the broad peak with almost flat features except for the oxygen vacancy.

Here it should be pointed out that g∼4.84 was ascribed to Co²⁺ in Co₃O₄ while g∼4.08 was ascribed to Co²⁺ on the surface, because the line g∼4.08 existed at high voltage and g∼4.84 disappeared which could be safely ascribed to the disappearance of Co₃O₄ species. Furthermore, for 1% Mg-doped LCO samples, three peaks of g∼24.6, g∼4.84, and g∼4.08 will co-exist at 3.8 V discharging and only one peak g∼4.08 exists at 4.2 V discharging, suggesting that g∼24.6 and g∼4.84 come from Co₃O₄. Furthermore, it is anticipated that the surface Co²⁺ has changed after charging, resulting in the change of EPR feature from ca. 500–2800 G in ⊥-EPR to ca. 1690 G in //-EPR.

It should be mentioned here that Alcántara et al. suggested that the resonance at g∼2.14 could be ascribed to Ni³⁺ impurities in LCO, while the resonances at g∼2.03 and g∼1.95 could be ascribed to Fe³⁺ impurities [14]. However, our EPR spectra do not have the resonances at g∼2.03 and g∼1.95. Furthermore, the resonance at g∼2.12 in our EPR spectra disappears after charging to 4.2 V and never shows up again. If this resonance at g∼2.12 comes from Ni³⁺ impurities, this signal will disappear after charging to 4.2 V due to Ni³⁺→Ni⁴⁺ but will show up when the discharging to 3.8 V. Therefore, it is suggested here that the resonance at g∼2.12 comes from the paramagnetic defects of Co₃O₄. We have also used inductively coupled plasma-mass spectrometry (ICP-MS) to measure the Ni or Fe content. The Ni or Fe content is ca. 34.5 or 37.6 mg/kg respectively. However, the observable Ni³⁺ or Fe³⁺ signal reported by Alcántara et al. corresponds to 0.5 mol% (Ni/Co + Ni, Fe/Co + Fe) content, which corresponds to ca. 3000 mg/kg or 2800 mg/kg respectively. Their contents are much larger than the Ni or Fe contents in our LCO samples. Therefore, we believe that these signals should not come from Ni or Fe impurities.

It is very interesting to observe that the broad peak at g∼2.2 for Co⁴⁺ species does not disappear during the whole lithiation process. It was anticipated that this signal will disappear at the voltage of 3.0 V because Co⁴⁺ should be reduced to Co³⁺ after the Li⁺ ion insertion. It was known that the pristine sample of O3-I phase will convert to O3-II phase then O3M phase, and finally H1-3 phase. When Li⁺ ion reinsertion begins, H1-3 phase would convert back to O3M phase, then O3-II phase, and finally O3-I phase. However, our previous NMR results suggested that when charging back to 3.0 V, O3-II phase of LCO could not convert back to O3-I phase [20]. This EPR result also confirms the existence of O3-II phase during the full cycling. The O3-II phase with the formula of Li_xCoO₂ (0.75< x < 0.98) has residual Co⁴⁺. Due to the large FWHM of this resonance, strong coupling between Co⁴⁺ is anticipated. Here it should be pointed out that, binary Ba and Ti-based surface modification of LCO could make the phase-transition reversible, that is, O3-II could convert to O3-I at 3 V discharging, possibly because the surface modification increases the electrical conductivity and facilitates the insertion of lithium at the end of discharge [20].

We also acquire the EPR spectra for different cycles, as shown in Fig. 3. It is obvious that the EPR spectra are quite similar. It is obvious that the signals for oxygen vacancies always increase from the 1^st cycle to the 100^th cycle implying the occurrence of oxygen redox and more oxygen vacancies will lead to structural instability. It should be pointed out that the broad signal in ⊥-EPR spectra (Fig. 3(a)) and the small peak in //-EPR (Fig. 3(b)) for the 100^th cycle are also slightly changed compared with that for the 1^st cycle.

Fig. 3.

The perpendicular-mode EPR spectra (a) and parallel-mode EPR spectra (b) collected at 3.0 V for the 1^st, 50^th and 100^th cycle.

At last, we propose a scheme for the structural change during the charge/discharge process, as shown in Fig. 4. For the pristine LiCoO₂, it has Co₃O₄ particles inside the LiCoO₂ matrix with O3-I phase and surface Co²⁺ species. After charging, the Co₃O₄ gradually disappears, especially at a high voltage of 4.6 V. Meanwhile, many oxygen vacancies generate and surface Co²⁺ species still exist but slightly change the environment. After discharging, the surface Co²⁺ and oxygen vacancies persist but Co₃O₄ does not reappear anymore. Moreover, O3-II will not convert back to O3–I state, leaving O3-II phases in the bulk samples with the formula of Li_xCoO₂ (0.75< x < 0.98) of residual Co⁴⁺. Since then, the surface Co²⁺, oxygen vacancies and O3-II phases will persist in all the cycles.

Fig. 4.

Schematic illustration of structural change during charge/discharge process.

4. Conclusion

Using dual-mode EPR, the EPR spectra of the pristine LCO and delithiated LCO were carefully assigned for the first time. Our results suggest that the impurity signal for Ni³⁺ or Fe³⁺ does not show up in the LCO system. The EPR spectra for the LCO samples at different SOC states suggest that the O3-II phase does not fully convert back to the original O3-I phase. Moreover, the oxygen vacancies easily generate at a high voltage and will increase when the cycle increases, hinting the participation of oxygen reactions. Therefore, it is anticipated that to improve the electrochemical performances, doping or surface modification is needed to make the phase transition fully reversible between O3-II and O3-I phases together with less oxygen vacancies. Our work lays the groundwork for a better understanding of the deintercalation/intercalation mechanism for future improvements in electrochemical performances.

CRediT authorship contribution statement

Bei Hu: Conceptualization, Experiments, Data analysis, Writing - original draft. Fushan Geng: Data analysis. Ming Shen: Data analysis. Bingwen Hu: Conceptualization, Data analysis, Writing - original draft, Writing - review & editing.

Declaration of competing interest

The authors declared that they have no commercial or associative interest that represents a conflict of interest in connection with the work submitted.

Biographies

Bei Hu received her B. S. degree in Department of Chemistry from University of Jinan and her Ph.D. degree in School of Physics and Electronic Science from East China Normal University. She is currently a post-doctoral scholar at Shanghai Jiaotong University. Her research interests include high-energy-density cathode and their mechanism study based on solid-state NMR/EPR, as well as application of low dimensional nano materials in electronic devices.

Bingwen Hu received the B.S. degree from Fudan University in 2003, the master degree from Fudan University in 2006 and Ph.D. degree from University of Lille-1 (France) in 2009. He is currently holding Zijiang Professorship in East China Normal University. His research interests are nuclear magnetic resonance (NMR), electron paramagnetic resonance (EPR) and their applications for lithium-ion batteries and other batteries-related materials.