A method is described in which crystalline silicon can be used as a practical anode material for lithium-ion batteries. Commercial lithium-ion cells are typically charged at a constant current to a fixed voltage and then are held by the charger at constant voltage until the current decreases to a certain value ͑also known as constant current/constant voltage or CCCV charging͒. It is first shown that CCCV charging can be used to reversibly cycle crystalline silicon and limit its capacity. A cycling method is then demonstrated in which crystalline silicon is first partially converted to amorphous silicon, in situ, during conditioning cycles. After the conditioning cycles the silicon can be cycled normally, using CCCV cycling limits, with good coulombic efficiency and little overlithiation during the first cycle.

No abstract

No abstract

Methods and criteria for assessing the commercial viability of Si-based materials are discussed and demonstrated with the 3M V6 alloy and 60 nm nano Si powder. These materials are firstly evaluated through the cycling of neat electrodes containing only alloy and binder to characterize the capacity, first cycle efficiency, binder compatibility, and microstructure stability of the material. The alloy displays higher first cycle efficiency, lower fade, and a more stable amorphous microstructure compared to the nano Si, which displays a variable microstructure with a rate dependent presence of crystalline Li 15 Si 4 . The materials are then evaluated in graphite-containing composite electrodes having high areal capacities (> 2 mAh/cm 2 ). In a well designed composite electrode including carbon nanotubes, 3M V6 material was found to cycle with little fade and high coulombic efficiency (∼99.8%) while maintaining a stable dQ/dV. A composite electrode of equivalent volumetric capacity with nano Si powder shows similar capacity retention over 50 cycles but an unacceptably low coulombic efficiency (∼99.2%). High precision coulometry and calorimetry results show surface area as the dominant factor in levels of parasitic reactions with Si based materials.

Si 1Ϫx Sn x samples for 0 Ͻ x Ͻ 0.5 were prepared by magnetron sputtering using a combinatorial materials science approach. The room-temperature resistivity and X-ray diffraction ͑XRD͒ patterns of the samples were used to select materials having both an amorphous structure and good conductivity for further study. The reaction of lithium with amorphous Si 0.66 Sn 0.34 was then studied by electrochemical methods and by in situ XRD. The electrode material apparently remains amorphous throughout all portions of the charge and discharge profile, in the range 0 Ͻ x Ͻ 4.4 in Li x Si 0.66 Sn 0.34. No crystalline phases are formed, unlike the situation when lithium reacts with tin. Using the Debye scattering formalism, we show that the XRD patterns of the a-Si 0.66 Sn 0.34 starting material and a-Li 4.4 Si 0.66 Sn 0.34 can be explained by the same local atomic arrangements as found in crystalline Si and Li 4.4 Si or Li 4.4 Sn, respectively. In fact, the in situ XRD patterns of a-Li x Si 0.66 Sn 0.34 , for any x, can be well approximated by a linear combination of the patterns for x ϭ 0 and x ϭ 4.4. This suggests that predominantly only two local environments for Si and Sn are found at any value of x in a-Li x Si 0.66 Sn 0.44. However, based on differential capacity vs. potential results for Li/a-Si 0.66 Sn 0.34 there is no evidence for two-phase regions during the charge and discharge profile. Thus, the two local environments must appear at random throughout the particles. We speculate that the charge-discharge hysteresis in the voltage-capacity profile of Li/ a-Li x Si 0.66 Sn 0.34 cells is caused by the energy dissipated during the changes in the local atomic environment around the host atoms.