In situ Scanning electron microscope study and microstructural evolution of nano silicon anode for high energy Li-ion batteries
Graphical abstract
Introduction
Lithium-ion batteries have become the most popular technology for electric energy storage, with applications as power source for portable electronics, and more recently for hybrid and electric vehicles. This success is due to the development and optimization of active materials as positive electrodes (see Refs. [1], [2], [3] for a review), and negative electrodes (see Ref. [4] for a review). Many efforts are currently made to increase the energy density, the power density and the intrinsic safety of the battery. The commercialized anode is usually graphite carbon, although Li4Ti5O12 is presently used for applications that demand a lot of power [5], [6]. The capacity of carbon graphite is 370 mAh g−1. This is much smaller than silicon that has a gravimetric capacity 4200 mAh g−1 when lithiated to Li4.4Si, and volumetric density 9786 mAh cm−3 based on the initial volume of Si, respectively [7], [8], [9]. These are the highest capacities among all the anode elements for Li-ion batteries, except Li metal itself. For this reason, Si has been considered as the promising element to increase the energy density of the Li-ion batteries since many years [10], [11], [12]. It is, however, difficult to believe it, because the cathode limits the energy density of a Li-ion cell, not by the anode element. The capacity of LiFePO4 olivine is 170 mAh g−1, that of lamellar compounds slightly larger, but still much smaller than that of graphite, and at the price of thermal instability that reduces the intrinsic safety of the batteries [13]. Nevertheless, the investigation on the Si anode is of interest for another reason, as the performance of the graphite anode depends strongly on the stabilization and control of the solid–electrolyte interface that limits the performance and calendar life of the battery. So far however, Si-based anodes suffer from numerous problems that prevent them being commercialized. In particular, the performance degrades during the first cycles due to the large variations of the volume during the charging/discharging process [14], [15], [16], [17], [18], [19], [20]: when transforming from Si to Li4.4Si, the volume expansion is 420% [21], [22], [23], [24], [25]. This large volume expansion/contraction during lithium insertion/extraction is responsible for the cracking of the Si particles, observed by atomic force microscopy during the Li-extraction [26], [27], [28]. In an attempt to overcome this problem, many efforts have been made to reduce the size of the Si particles to the nanoscale, which have been recently reviewed [29], in order to reduce the internal stress, Indeed, the electrochemical properties have been improved when Si is under the from of clamped hollow structures, such as double-walled nanotubes and yolk-shell nanoparticles. Such a structure, with Si nanotubes coated with a SiOx layer showed a remarkable capacity of 1000 mAh g−1 at rate 12C, with long cycling life (6000 cycles with 88% capacity retention) [30]. Such devices, however, are still too expensive to be commercialized. In this context, the size dependence of the Si-properties and the nature of the solid–electrolyte interface need to be studied and are identified as two areas where research is needed [29]. For this purpose, in-situ electron microscopy measurements are a very useful tool, since they allow real-time observation of the charging/discharging behavior of individual particles. This strategy has already been used to investigate the effect of metallic coating on Si expansion [20], and the behavior of individual Si nanowire electrodes [31], [32], [33], [34].
In the present work, we report both in-situ and ex-situ SEM experiments to observe the charging/discharging behavior of bigger particles to investigate the micro-structural evolution of nano Si particles during electrochemical cycling as a function of their size, completed by an analysis of SiO-based anode, aiming to determine the critical size above which cracking of bare particles cannot be avoided. We also investigate dynamically the microstructural change of morphology of the whole Si and SiO electrodes to investigate the effect of the changes of the volumes of the particles at the level of the entire electrode (active material, binder, conductive carbon).
Section snippets
Materials
Different sources of Si were selected to cover a broader range of particle sizes: nano-Si particles (average particle size ca. 100 nm); bigger SiOx particles (2–8 μ). 0.8 cm × 0.8 cm cells were assembled and prepared for SEM cross-section observation using a cryo-microtome. The electrodes were prepared by mixing Si and graphite (1:1 weight ratio) with the binder dissolved in N-methyl-2-pyrrolidinone (NMP) in the ratio 10%. The graphite (OMAC1S, 15-μm average particle size) was obtained from
Results on Si electrodes
The SEM image of the nano Si electrode before cycling in Fig. 1 shows that Si particles are poly dispersed with sizes in the range 70–200 nm (average particle size ca. 100 nm). The cycling results for this electrode are illustrated in Fig. 2. We have checked that the capacity 1160 mAh g−1 during the first cycle is almost equal to the theoretical one and does not depend on the cycle rate at the low C-rates up to C/15. Then we have reported in Fig. 2 the voltage of the cell that is the only
Results on SiOx anodes
The SiOx (x ∼ 0.95) particles used in this work are the same as in our earlier work [35]. SiO is thermodynamically unstable. The commercial SiO is then constituted of nanoclusters of amorphous Si clusters and amorphous SiO2 clusters, surrounded by Si-suboxide [50], [51]. The suboxide part is expected to be inactive in the chemical process. Advantage has been taken this property by using SiO as a buffer to constrain the volume change of the active part [52]. Therefore, the mechanical properties
Conclusion
The in-situ electron microscopy of Si and SiO-based electrodes with different sizes of particles between 0.1 and 20 μm has provided a new insight into the micro-structural evolution of the particles. When the particles are larger than 10 μm, the cell is out of thermodynamic equilibrium even at low C/24 rate, because the lithium ions do not have time enough to exit from the particles. The consequence is an inhomogeneous distribution of Li in the particles that can increase the internal stress
Acknowledgments
BATT-DOE (US) and HydroQuébec and are gratefully acknowledged for the financial support.
References (62)
- et al.
J. Power Sources
(2013) - et al.
J. Power Sources
(2011) - et al.
J. Power Sources
(2012) - et al.
J. Power Sources
(2009) - et al.
J. Power Sources
(2004) - et al.
J. Power Sources
(1997) - et al.
J. Power Sources
(2004) - et al.
Nano Energy
(2012) - et al.
J. Power Sources
(2004) - et al.
Electrochim. Acta
(2006)