2022 Volume 90 Issue 3 Pages 037004
Analysis of the differential capacity profile (dQ/dV vs. V), which can capture the changes in the electrode structure inside a battery, is an effective electrochemical method for investigating the degradation behavior and mechanism of Li-ion cells. However, the electrode reactions corresponding to the peaks in a dQ/dV vs. V curve are undefined. Hence, it is difficult to qualitatively analyze the mechanism of cell degradation using this curve. In this work, we propose an original method for attributing the peaks in a dQ/dV vs. V curve. Peak attribution is implemented using a three-electrode Li-ion laminate cell with a LiCoO2 cathode, graphite anode, and lithium reference electrode. During charge and discharge, the potential differences (dE) of the LiCoO2-Li and graphite-Li sides and voltage difference (dV) of the LiCoO2-graphite full cell are measured simultaneously. Meanwhile, the Coulomb amounts (dQ) of the LiCoO2-Li and graphite-Li sides are equivalent to that of the LiCoO2-graphite full cell at a certain time. By comparing the dQ/dV vs. V curve of the full cell to the dQ/dE vs. E curves of the LiCoO2-Li and graphite-Li sides, the peaks in the differential capacity curve can be attributed to specific structural changes in the electrodes. Importantly, this information is acquired without disassembling the cell, making our proposed analytical method a convenient means of studying cell degradation under various conditions, such as low temperature, high temperature, or high-rate charging.
Lithium-ion batteries (LIBs) have been widely used as reliable energy sources for portable devices such as laptops, cellular phones, and camcorders. In addition, they are also considered promising candidates for use in electric vehicles.1–3 From the perspective of building a low-carbon society and achieving energy security, the applications of LIBs are expected to further expand in the future. Hence, it is very important to prolong the service life of LIBs, which, with long-term use, are known to show decreased voltage and capacity.4–9 Therefore, to improve the performance of LIBs and broaden their applications, it is necessary to interpret the mechanisms of cell degradation.
In general, cell degradation is a very complex process owing to the combination of cathode/anode materials and electrolytes. The degradation of LIBs during charge/discharge operation has been investigated intensively in several studies through analysis of the differential capacity profile (dQ/dV vs. V).4,10–15 The differential capacity profile introduces a constant-current analytical method, wherein the behavior of the redox system is analyzed by monitoring the electrode potential resulting from the application of current.10 Equation 1 shows that dQ/dV can be regarded as the current (I) divided by the sweep rate (dV/dt).
| \begin{equation} \frac{\mathrm{d}Q}{\mathrm{d}V} = \frac{\mathrm{d}Q/\mathrm{d}t}{\mathrm{d}V/\mathrm{d}t} = \frac{\mathrm{d}(It)/\mathrm{d}t}{\mathrm{d}V/\mathrm{d}t} = \frac{I}{\mathrm{d}V/\mathrm{d}t} \end{equation} | (1) |
However, because commercial 18650-type lithium-ion cells are two-electrode cells composed of one positive and one negative electrode, the peaks in the dQ/dV vs. V curve are affected by the cathode and anode concurrently. Hence, it is difficult to evaluate the degradation arising from specific structural changes in one electrode by using the dQ/dV vs. V curve directly if the electrode reactions corresponding to the peaks are undefined. Therefore, to analyze the degradation mechanisms of 18650-type lithium-ion cells by using the dQ/dV vs. V curve, it is necessary to attribute the peaks to specific electrode reactions.
In this work, peak attribution of the dQ/dV vs. V curve of a LiCoO2-based lithium-ion cell was implemented using a three-electrode Li-ion laminate cell with a LiCoO2 cathode, graphite anode, and lithium reference electrode. During charge and discharge, the potential differences (dE) of the LiCoO2-Li and graphite-Li sides and voltage difference (dV) of the LiCoO2-graphite full cell were measured simultaneously. By importing the Coulomb amount (dQ) of the LiCoO2-graphite full cell, the dQ/dE vs. E curves of the LiCoO2-Li and graphite-Li sides and dQ/dV vs. V curve of the LiCoO2-graphite full cell were obtained. This allowed us to not only capture the structural changes in the LiCoO2 cathode and graphite anode during charge and discharge, but also attribute the peaks of the full cell to specific electrode reactions.
In this work, the rate-determining step theory is the theoretical basis for attributing the peaks in the dQ/dV vs. V curve of the full cell. Here, the methodology for peak attribution is expounded. The dQ/dE vs. E curves of the cathode and anode sides and dQ/dV vs. V curve of the full cell were measured simultaneously. To achieve time consistency and thus attribute the peaks precisely, the electrode potential (E) in the dQ/dE vs. E curves and cell voltage (V) in the dQ/dV vs. V curve were converted to time values (t).
During charge/discharge cycling, the Coulomb amounts (dQ) of the cathode and anode reactions are equivalent at a certain time (dt) because the current (I) is constant (refer to Eq. 2).
| \begin{equation} Q = I\times t \end{equation} | (2) |
The dQ/dV vs. t curve of the full cell with a reference electrode and dQ/dE vs. t curves of the cathode and anode sides are schematically shown in Fig. 1. During charging, the deintercalation of Li ions from the cathode and intercalation of Li ions to the anode occur simultaneously. However, for the total reaction of the full cell, the reaction rate is always limited by the relatively slow deintercalation reaction of the cathode or intercalation reaction of the anode.
Schematic diagram of peak attribution based on the rate-determining step theory in Fig. 4. In situation 1, the anode reaction is regarded as the rate-determining step, which is attributed as the main reaction of the full cell. In situation 2, the anode reaction and the cathode reaction are regarded as the rate-determining steps before and after t1, which are attributed as the main reaction of the full cell before and after t1 respectively.
In situation 1, the dQ/dE values of the anode reaction are smaller than those of the cathode; thus, the intercalation reaction of the anode is regarded as the rate-determining step. Additionally, the shape and site of this anode peak is the same as that of the full cell peak. Therefore, the main reaction of the full cell is attributed to the intercalation of Li ions in the anode.
In situation 2, the dQ/dE values of the anode are smaller than those of the cathode before t1, and the dQ/dE values of the cathode are smaller than those of the anode after t1. Thus, the intercalation reaction of the anode and deintercalation reaction of the cathode are regarded as the rate-determining steps before and after t1, respectively. Additionally, the intercalation reaction peak of the anode and deintercalation reaction peak of the cathode are uphill and downhill of the full cell peak, respectively. Therefore, the main reaction of the full cell is attributed to the anode reaction before t1 and cathode reaction after t1. The test and data analysis were then carried out according to the above description.
A square three-electrode Li-ion laminate cell with a nominal capacity of 13.5 mAh was fabricated with a LiCoO2 cathode (30 mm × 30 mm, HS-LIB-P-Co-003, Hohsen Corp.), graphite anode (29 mm × 29 mm, HS-LIB-N-Gr-001, Hohsen Corp.) and lithium reference electrode. The capacity ratio of LiCoO2 cathode and graphite anode is 1 : 1. A polypropylene film (Celgard 2400) was used as the separator (the separators were installed on both sides of lithium reference electrode to ensure that the electrode reactions cannot be affected by lithium reference electrode16,17), and a solution of 1 mol dm−3 LiPF6 dissolved in a mixture of ethylene carbonate and dimethyl carbonate (3 : 7 v/v%, Kishida Chemical) was used as the electrolyte. The components of the three-electrode Li-ion laminate cell are shown in Fig. 2. The cells were assembled in an argon-filled glove box and charged/discharged at room temperature.
Schematic diagrams of the (a) general appearance and (b) individual components of a three-electrode Li-ion laminate cell (Gas barrier film: 52 mm × 52 mm, Laminate film: 47 mm × 47 mm).
Cycling tests were carried out using a charge/discharge tester (HJ1020mSD8, Hokuto Denko), and the cells were subjected to 0.2 C-rate for two cycles and 0.1 C-rate for one cycle before testing at 25 °C. A current of 13.5 mA was defined as 1 C-rate. The differential capacity profiles of the three-electrode Li-ion laminate cell were measured at 25 °C and 0.05 C-rate between 2.7 and 4.2 V (the charge/discharge curves are shown in Figs. S1a–S1c). In this way, the differential capacity profiles of the LiCoO2-Li and graphite-Li sides (dQ/dE vs. E) and three-electrode full cell (dQ/dV vs. V) could be measured simultaneously, which enabled the attribution of the peaks of the full cell to specific electrode reactions.
Figure 3 shows the differential capacity profiles of the LiCoO2-Li and graphite-Li sides (dQ/dE vs. E curves, Figs. 3a and 3b, respectively) and full cell (dQ/dV vs. V curve, Fig. 3c). The original profiles of Figs. 3a–3c are shown in Figs. S2a–S2c. Characteristic peaks (peaks A, B, and C) were observed at 3.97, 4.10, and 4.22 V vs. Li/Li+ during charging, as well as the reverse reaction peaks (peaks A′, B′, and C′) during discharge (Fig. 3a). It suggests that these peaks attributable to not experimental error but some electrochemical processes. These peaks can be attributed to structural changes in LiCoO2 during Li ion intercalation/deintercalation.16,18 By comparing this result with the Raman spectroscopy study of lithium intercalation in LiCoO2, peaks A, B, and C can be explained by the structural phase transitions between hexagonal-I, hexagonal-II, and monoclinic.19 The structural changes at the characteristic peaks of LiCoO2 are summarized in Table 1.
Differential capacity profiles of (a) LiCoO2-Li, (b) graphite-Li, and (c) full cell surveyed at 25 °C and a constant current of 0.05 C using a three-electrode Li-ion laminate cell.
| Electrode | Characteristic peaks | Peak potential (vs. Li/Li+) | Structural changes16–21 |
|---|---|---|---|
| LiCoO2 | A (charge) | 3.97 V | Hexagonal-I to Hexagonal-I/II |
| B (charge) | 4.10 V | Hexagonal-I/II to Hexagonal-II | |
| C (charge) | 4.22 V | Hexagonal-II to Monoclinic | |
| A′ (discharge) | 3.81 V | Hexagonal-I/II to Hexagonal-I | |
| B′ (discharge) | 4.02 V | Hexagonal-II to Hexagonal-I/II | |
| C′ (discharge) | 4.15 V | Monoclinic to Hexagonal-II | |
| Graphite | a (charge) | 0.20 V | dilute stage-1 to stage-4 |
| b (charge) | 0.15 V | stage-4 to stage-3 | |
| c (charge) | 0.13 V | stage-4 to stage-3 | |
| d (charge) | 0.12 V | stage-3 to stage-2 | |
| e (charge) | 0.08 V | stage-2 to stage-1 | |
| a′ (discharge) | 0.22 V | stage-4 to dilute stage-1 | |
| b′ (discharge) | 0.19, 0.17 V | stage-3 to stage-4 | |
| c′ (discharge) | 0.15 V | stage-3 to stage-4 | |
| d′ (discharge) | 0.14 V | stage-2 to stage-3 | |
| e′ (discharge) | 0.10 V | stage-1 to stage-2 |
In the dQ/dE vs. E curve of the graphite-Li side (Fig. 3b), the five characteristic peaks (peaks a–e) during charging can be attributed to structural changes in graphite.17,20,21 The dQ/dE vs. E curve is similar to that of the reported cyclic voltammogram profile of graphitized particles.20 Considering the mechanism of Li insertion into graphite, the peaks can be explained by the staging process.21 During charging, in the electrode potential range of 0.3–0.18 V vs. Li/Li+, Li ions are randomly inserted between all layers of graphite, and the phase transition from dilute stage-1 to stage-4 occurs at peak a. From 0.18 to 0.13 V vs. Li/Li+, the phase transition from stage-4 to stage-3 occurs at peaks b and c. In the range of 0.13–0.085 V vs. Li/Li+, the phase transition from stage-3 to stage-2 occurs at peak d. Finally, in the more negative potential region, the phase transition from stage-2 to stage-1 occurs at peak e. On the other hand, the peaks (peaks a′–e′) identified during discharge were validated as concrete inverse reactions of phase transition. The structural changes at the characteristic peaks of graphite are summarized in Table 1. As shown in Fig. 3c, seven peaks (peaks 1–7), which are considered to be the combined effects of structural changes in LiCoO2 and graphite, were identified during charge and discharge.
4.2 Attribution of peaks in differential capacity profilesFor the convenience of peak attribution, the directions of charge and discharge in Figs. 3a–3c were adjusted uniformly, and the potential/voltage values shown in the horizontal axis were converted to time values during charge and discharge. Overlapping the adjusted figures shows that the characteristic peaks of LiCoO2/graphite correspond to the seven peaks of the full cell (Fig. 4, and the original profile is shown in Fig. S3). If the reactions corresponding to the full cell peaks are regarded as the total reactions in the charge/discharge system, every total reaction can be considered to be due to the reaction of the cathode, anode, or both. Hence, peak attribution of each total reaction is necessary. The reaction rate of the total reaction is considered to constantly depend on the slowest reaction in the system, which is the rate-determining step. On the basis of this theory, the peak intensities of the reaction peaks of LiCoO2 and graphite shown in Fig. 4 were compared, and the peaks of LiCoO2 or graphite that are closer to the full cell peaks were attributed to the main reaction (see Figs. 5a–5g, which are enlarged views of Fig. 4). The peak attribution results for the full cell are summarized as follows:
Enlarged views of Fig. 4 in the time ranges of (a) 0–4 h, (b) 4–15 h, (c) 20–25 h, and (d) 25–29 h of charging and (e) 0–2 h, (f) 3–6 h, and (g) 10–25 h of discharging.
The charging process:
The discharging process:
Thus far, peak attribution of the dQ/dV vs. V curve has been accomplished using a three-electrode Li-ion laminate cell. This can provide theoretical support for further analysis of the degradation of commercially available 18650-type LIBs. In this way, the degradation pattern of each electrode inside a cell during cell cycling can be captured by measuring and comparing the differential capacity curves at different periods, instead of disassembling the cell after cycling. Furthermore, we believe that this analytical method can be applied in the analysis of cell degradation under various conditions, such as low temperature, high temperature, or high-rate charging. Subsequent investigations of degradation mechanisms of 18650-type LIBs using the results of this study are ongoing in our group. Furthermore, in the future, we plan to carry out comparative experiments using different active materials such as LiNiO2 and LiFePO4.
The peaks in the dQ/dV vs. V curve of a three-electrode Li-ion laminate cell with a LiCoO2 cathode, graphite anode, and lithium reference electrode were attributed to specific structural changes in the electrodes. In this work, the methodology for peak attribution of the dQ/dV vs. V curve was expounded with an overview diagram, which makes it convenient to understand the different situations in peak attribution. The peaks in the dQ/dE vs. E curves of the LiCoO2-Li and graphite-Li sides during charge and discharge were determined on the basis of references. Finally, the peak attribution was completed by comparing these peaks with those in the dQ/dV vs. V curve of the LiCoO2-graphite full cell. This analysis of the peak attribution of the dQ/dV vs. V curve is an original method, and the results can be used to investigate different patterns of cell degradation and to interpret the mechanisms.
The data that support the findings of this study are openly available under the terms of the designated Creative Commons License in J-STAGE Data at https://doi.org/10.50892/data.electrochemistry.18737735.
Shuo Li: Conceptualization (Equal), Data curation (Lead), Formal analysis (Lead), Investigation (Lead), Methodology (Lead), Writing – original draft (Lead)
Sachiyo Tsutsumi: Data curation (Equal), Investigation (Equal)
Sayoko Shironita: Formal analysis (Equal), Investigation (Supporting), Supervision (Lead)
Minoru Umeda: Conceptualization (Lead), Project administration (Lead), Supervision (Lead)
The authors declare no conflict of interest in the manuscript.
S. Li and S. Tsutsumi: ECSJ Student Members
S. Shironita and M. Umeda: ECSJ Active Members
