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.

LiFePO 4 /Li 4/3 Ti 5/3 O 4 Li-ion cells have been investigated by many groups and their behavior in standard electrolytes such as 1 M LiPF 6 ethylene carbonate: diethyl carbonate ͑EC:DEC͒ is well known. Here we report on the behavior of these cells with 2,5-ditertbutyl-1,4-dimethoxybenzene added to the electrolyte as a redox shuttle additive to prevent overcharge and overdischarge. We explore methods to increase the current-carrying capacity of the shuttle and explore the heating of practical cells during extended overcharge. The solubility of 2,5-ditertbutyl-1,4-dimethoxybenzene was determined as a function of salt concentration in lithium bis-oxolatoborate-͑LiBOB͒ and LiPF 6 -containing electrolytes based on propylene carbonate ͑PC͒, EC, DEC, and dimethyl carbonate ͑DMC͒ solvents. Concentrations of 2,5-ditertbutyl-1,4-dimethoxybenzene up to 0.4 M can be obtained in 0.5 M LiBOB PC:DEC ͑1:2 by volume͒. Coin-type test cells were tested for extended overcharge protection using an electrolyte containing 0.2 M 2,5-ditertbutyl-1,4-dimethoxybenzene in 0.5 M LiBOB PC:DEC. Sustained overcharge protection at a current density of 2.3 mA/cm 2 was possible and hundreds of 100% shuttle-protected overcharge cycles were achieved at current densities of about 1 mA/cm 2 . The diffusion coefficient of the shuttle molecule in this electrolyte was determined to be 1.6 ϫ 10 −6 cm 2 /s from cyclic voltammetry and also from measurements of the shuttle potential vs. current density. The power produced during overcharge was measured using isothermal microcalorimetry and found to be IV as expected, where I is the charging current and V is the cell terminal voltage during shuttle-protected overcharge. Calculations of the temperature of 18650-sized Li-ion cells as a function of time during extended shuttle-protected overcharge at various C-rates are presented. These show that Li-ion cells need external cooling during extended shuttle-protected overcharge if currents exceed about C/5 rates.

A method for measuring the energy produced from parasitic cell reactions in lithium ion cells by electrochemical calorimetry is described. Negative electrode symmetric cells were charged and discharged by high precision current sources in an isothermal heat flow calorimeter while the cell voltage was accurately measured. Two sources of graphite of different BET surface areas were investigated. Symmetric cells of Li 4 Ti 5 O 12 and lithium/graphite half cells were also measured by this method. The measured parasitic energy was well correlated to the loss of active Li, or coulombic efficiency, confirming the source of the parasitic energy as the heat of reaction occurring between the lithiated electrodes and the electrolyte. The effect of electrode formulation was also explored. Electrochemical calorimetry of symmetric cells is an excellent method to study new material sets to determine which will lead to extended cell lifetime.

The electrochemical performance of negative electrodes based on commercially available micrometer-sized ␣-Fe 2 O 3 powder and four different binders was investigated. ␣-Fe 2 O 3 electrodes made using sodium carboxymethyl cellulose binder and two proprietary binders show better cycling performance than electrodes made from the conventional binder, polyvinylidene fluoride ͑PVDF͒. Heat-treating the PVDF electrodes to 300°C improves electrode performance dramatically. A specific capacity of over 800 mAh/g for over 100 cycles has been achieved for Li/␣-Fe 2 O 3 cells with micrometer-sized ␣-Fe 2 O 3 , by contrast to many teachings in the literature where it is claimed nanometer sized particles are required to obtain such performance. As is the case for composite electrodes made from powders of metal alloys, we suggest that binder choice has a great impact on the performance of metal oxide electrodes demonstrating large volume expansion during lithiation, such as ␣-Fe 2 O 3 .