The increasingly demand on secondary batteries with higher specific energy densities requires the replacement of the actual electrode materials. With a very high theoretical capacity (4200 mAh g−1) at low voltage, silicon is presented as a very interesting potential candidate as negative electrode for lithium-ion microbatteries. For the first time, the electrochemical lithium alloying/de-alloying process is proven to occur, respectively, at 0.15 V/0.45 V vs. Li+/Li wit. The increasingly demand on secondary batteries with higher specific energy densities requires the replacement of the actual electrode materials. With a very high theoretical capacity (4200 mAh g−1) at low voltage, silicon is presented as a very interesting potential candidate as negative electrode for lithium-ion microbatteries. For the first time, the electrochemical lithium alloying/de-alloying process is proven to occur, respectively, at 0.15 V/0.45 V vs. Li+/Li with Si nanowires (SiNWs, 200–300 nm in diameter) synthesized by chemical vapour deposition. This new three-dimensional architecture material is well suited to accommodate the expected large volume expansion due to the reversible formation of Li–Si alloys. At present, stable capacity over ten to twenty cycles is demonstrated. The storage capacity is shown to increase with the growth temperature by a factor 3 as the temperature varies from 525 to 575 °C. These results, showing an attractive working potential and large storage capacities, open up a new promising field of research.••Lithium batteriesSiliconSi nanowiresThin filmsThe current commercial lithium-ion secondary batteries are the most widely used because of their higher energy density, their higher operating voltages and their lower self-discharge,. They are based on an anode made of graphitic carbon or other carbonaceous materials that present on the one hand the advantage to be cheap and on the other hand interesting electrochemical properties such as a low and flat working voltage and a good cycleability. However, the maximal insertion of one lithium ion for six carbon atoms leads to a theoretical capacity limited to 372 mAh g−1, which is relatively low. In order to satisfy to the increasingly demand for new compact and modern portable electronic devices, both the active materials in the cathode and the anode should be replaced by new materials.Concerning the anodic materials, metals, metalloids and semiconductors such as Sn, Al, Sb and Si which can make alloys with lithium, are attractive alternatives to graphite due to their low cost and high storage energy density. Theoretical specific capacities more than ten times higher than with carbon can be obtained. However, the structural changes and a large volume expansion associated with lithium insertion, which can rise to more than 300% (whereas it is below 10% for LiC6), limit the ability of these materials to cycle with high efficiency. Indeed, successive charge–discharge cycles lead to mechanical str. SiNWs are synthesized by chemical vapour deposition (CVD) according to the vapour–liquid–solid (VLS) growth mechanism which has been extensively studied ∼40 years ago by Wagner for the growth of whiskers. This mechanism can be explained with the help of Fig. 1. A Si bearing gas (e.g. SiH4) is allowed to flow over catalyst clusters (e.g. Au) deposited on a chemically inert substrate which is heated above the catalyst-Si eutectic melting temperature. Some SiH4 molecules decompose by pyrolysis and some of the released Si atoms adsorb on the catalyst dots and alloy with the surface atoms. As more Si is brought to the catalyst surface, the alloy composition evolves towards the eutectic composition. Melting starts to occur (Fig. 1a), rapidly consuming the whole catalyst dot. Continued feed of Si makes the alloy composition evolve towards saturation, inducing Si precipitation. An equilibrium is rapidly reached, where the flux of Si atoms incorporated at the surface of the liquid alloy is balanced by the flux of Si atoms precipitating at the liquid–solid interface. As the liquid surface behaves as an ideally rough surface, the sticking coefficient of the gaseous SiH4 molecules is close to 1, resulting in a highly anisotropic growth that explains the whisker shape obtained (Fig. 1c). The Au–Si liquid drop rises on top of the crystal and appears spherical in shape due to surface tension effects. The feasibility of Si nanowire growth by the VLS mechanism has been demonstrated recently using gold clusters.A SEM image of the thin SiNWs grown at 550 °C is displayed in Fig. 2. It can be seen that SiNWs with diameters up to 300 nm are uniformly obtained on the surface of the TiN-covered substrate. The small Au droplets used as catalyst for the growth of Si wires and whose size mainly controls their diameter are still present and they can be observed at the top of the nanowires.The Raman spectrum of the as-deposited SiNWs is shown in Fig. 3. A well defined and symmetric peak at 520.1 cm−1 corresponding to the first order Raman line for crystalline silicon is observed which ascertains very well crystallized SiNWs are prepared.Before testing the SiNWs as active electrode material against Li alloying, we have evaluated the possible electrochemical contribution of the Au droplets still present at the top of the silicon wires after synthesis. Indeed even when only a little is known on the electrochemical properties of Li–Au alloys at room temperature, the Au electrode has been recently proved to accommodate Li in the 0.02–0.5 V potential range with a poor reversibility,. The discharge–charge curves obtained for a 50 nm thick gold film are shown in Fig. 4. Two voltage plateaus corresponding to the Li alloying process into two different phases are observed at 0.2 and 0.1 V while Li removal from the alloy.
