Part 1. Research on silicon-based half-cell simulating manganese environment.
Nano silicon particles, a kind of commercial silicon material with uniform structure, stable performance and low cost was chose as silicon source in this study to reduce the influence of material defects on battery performance and improve the accuracy of comparison. Figure 1 shows the characterization of nano silicon particles. Figure 1a is the result of TGA and DSC simultaneous analysis of nano silicon particles. The blue curve in the figure shows that the mass of silicon material remains stable during the process of increasing the temperature from room temperature to about 600 ℃. As temperature continuously rising, it can be found that the mass of nano silicon particles start increasing. At 1000 ℃, total mass has increased to 151%. At the same time, in the heat flow curve, it can be observed that a wide range of exothermic process appears. The highest exothermic point (about 50MW) was fixed at 600 ℃. Judging from the gas composition of thermal analysis, mass rise and exothermic process of nano silicon particles was caused by oxidation reaction. To analyze the change of surface element of nano silicon particles further, X-ray photoelectron (XPS) spectroscopy was carried out. The XPS spectra of Si 2p orbit is displayed in Fig. 1b. The main 3/2 − 1/2 doublet (the spin-orbit splitting is 0.6 eV and the intensity ratio is 3:1), located at 99.1–99.7 eV corresponds to Si0 (~ 58% content), the lower peak, located at 98.2-107.1 eV corresponds to silicon oxide (~ 37% content), indicating that surface of nano silicon is partly oxidized.
Figure 2 is a morphological study of nano silicon and the prepared silicon electrode. Figure 2a is an SEM image of nano silicon particles. It can be observed that the nano silicon particles with regular spherical shape in particle size of 70-100nm. Figure 2b is the SEM image of the prepared nano silicon electrode (Preparation method is detailed in the test section). The electrode was coated on one side, and the upper part of SEM image is active material, which is composed of nano silicon particles, conductive agent and binder with total thickness of about 11.2µm; the lower part of SEM image is copper current collector with thickness of about 12.2µm. The coating material was uniformly covered on electrode, closely connected with the current collector.
To study the effect of Mn2+ existing in electrochemical environment on the performance of silicon electrode, the coin cell with and without MnF2 in electrolyte was prepared, and galvanostatic charge-discharge cycle test at current density of 100mA/g was carried out. The charge-discharge capacity and coulomb efficiency of the silicon-based half-cell in 100 cycles are shown in Fig. 3. Obviously the discharge capacity of both batteries in first cycle is much higher than that in second cycle, because of the formation of SEI membrane on the particle surface during the first lithium insertion in nano silicon particles. This irreversible reaction will increase the discharge capacity in the first cycle and reduce the first coulomb efficiency. Specifically, the first cycle coulomb efficiency of cell without MnF2 is 36.8%, and that of cell with MnF2 is 29.7%. In the 2nd to 5th cycles, the discharge capacity of the two batteries increased slowly, because of the high volume expansion rate of nano silicon in electrochemical reaction. In the first few cycles, the electrode expanded gradually then reached the best electrochemical state, so that the capacity will rise slowly. From the 6th cycle, the battery capacity showed a slow downward trend. When it came to the 100th cycle, the discharge specific capacity ofcell without MnF2 is 758mAh/g, 62.1% of the 6th cycle, while the discharge specific capacity of cell with MnF2 is 514mAh / g, 49.4% of the 6th cycle. As a conclusion, when MnF2 was added to the nano silicon particle half-cell, the first cycle coulomb efficiency and cycle stability of the battery decreased significantly. It can be considered that, the decrease of coulomb efficiency in the first cycle is mainly because more SEI generated during the first lithium insertion process. On the other hand, even in the same state, the decrease of cycle stability is also due to the lower coulomb efficiency of each cycle, that’s to say, in the battery with MnF2, there will be more irreversible reactions consuming lithium source. It can be inferred that Mn2+ is the dominate factor affecting the cycle stability of silicon-based half-cell.
To figure out the growth of SEI membrane of silicon-based negative electrode after 100th cycles, the cycled battery was adjusted to delithiation state, and then being disassembled in the argon glove box. After washing the residual electrolyte with dimethyl carbonate (DMC), the electrode was observed by SEM. Figures 4a and 4b are images of silicon negative electrode without MnF2, and Figs. 4c and 4d are electrode with MnF2. Overall, larger cracks appeared in the negative electrode with MnF2 after cycling. The generation of cracks will make part of the electrode materials lose their electrochemical activity as dropping off from the collector, which to some extent will impede nano silicon particles to become fully effective. Judging from the surface morphological pattern of silicon particles, floccules layer can be observed on the surface of silicon particles of the two electrodes, which is the SEI membrane generated on the surface of silicon particles. The growth of SEI with MnF2 is more serious than that without MnF2.
To obtain the direct evidence of the growth of SEI membrane on silicon-based NE, differential scanning calorimetry (DSC) was used to test the cycled battery in delithiation state, and fresh silicon NE as comparison. The results of DSC test are shown in Fig. 5. It can be seen from the figure that a significant exothermic peak between 100–200 ℃ representing decomposition reaction of SEI membrane on the negative electrode surface, according to previous research, appeared in cycled battery while not in new battery. The peak value0.54 W/g of exothermic peak is at about 162℃.
The difference of SEI thickness between with and without Mn2+ environment is shown in Fig. 6, then the damage mechanism of Mn2+ on the silicon anode in lithium-ion batteries is also explained. The SEI excessive growth is mainly caused by the large volume change of the silicon particles during charge /discharge cycling. By the way, Mn2+ makes SEI more fragile, lets SEI more easily break, results silicon particles more electrochemical reaction interface, and then accelerates the SEI thickening.
Part 2 Research on capacity fading mechanism on silicon/LMO full cell.
The experiments of simulating the electrochemical environment of Mn2+ show that Mn2+ has a serious impact on the coulomb efficiency and cycle performance of the negative electrode. In order to deeply explore the mechanism and influence of such effect in practical application of silicon-based batteries, full cell with LMO as positive electrode (PE) and nano silicon particle as NE were assembled to analyze their electrochemical behavior and capacity degradation trend.
Figure 7 shows SEM images of positive and negative electrodes used to prepare the full cell. Figure 7a shows the surface morphological image of LiMn2O4 electrode. It is clear that the coating is uniform and there are no obvious defects. Figure 7b is the detailed morphological pattern of LiMn2O4 particles whose particle size is about 500nm. The particles with size lower than 100 nm are uniformly distributed around the LiMn2O4 particles, which are the binders and conductive agents of the electrode. Figure 7c and 7d are the surface morphological image of nano silicon particle electrode. The distribution of silicon particles in the electrode is as uniform as half-cell.
Figure 8 shows galvanostatic charge and discharge test results of LMO/Si full cell, the results of silicon-based half-cell are also shown in the figure for comparison. The charge capacity of the 1st cycle and the 2nd cycle of full cell are 2820mAh / g and 3010mAh/g, respectively, and then reduced gradually in every next cycles. At 100th cycle, the capacity has decayed to 170mAh/g, 6% of original capacity. The battery can be considered as being failure at this time. In aspect of coulomb efficiency, the first and second cycle coulomb efficiency is 30.2% and 36.4%, respectively. As cycles increasing, the coulomb efficiency went down first and then went up inversely. At 100th cycle, the coulomb efficiency is 41.1%. Compared with the silicon based half-cell, full cell shows a serious downward trend of capacity: the coulomb efficiency of half-cell can be stably maintained at more than 90%, while the full cell was less than 40% in the first 90 cycles. Judging from the experiment of the electrochemical performance of full cell, it can be concluded that the capacity of LMO/Si full cell has a serious decline in the process of cyclic charging and discharging, and the low coulomb efficiency is the main factor for this rapid decline. In principle, low coulomb efficiency will lead to more irreversible reactions then generate more SEI membrane.
To study the growth of SEI membrane of silicon NE in full cell, the batteries after 100 cycles were disassembled in an argon atmosphere glove box, and then washed the negative electrode with DMC to remove the residual electrolyte, finally observed by SEM, as shown in Fig. 9. From Fig. 9a, it can be seen that not only serious cracks appeared on the surface of electrode, but white mottled traces came out. Figure 9b provides more details of silicon particles: silicon particles are almost invisible, instead by lots of floccule. This phenomenon is led by the excessive growth of SEI membrane coating on silicon particles. Compared with the SEM images of 100 cycles half-cell in Fig. 4, more SEI film grew on silicon NE in the full cell indirectly explains the faster decline of capacity than that of half-cell.