Van der Waals and Capillary Adhesion of Polycrystalline Silicon Micromachined Surfaces

Abstract

Microelectromechanical systems are especially sensitive to adhesion as a result of their large surface area-to-volume ratios, small surface separations, and compliant components. Interfacial forces that can contribute to the overall adhesion between micromachined surfaces include van der Waals, capillary meniscus, electrostatic, and solid bridging forces. In this chapter, we focus on van der Waals and capillary meniscus forces between polycrystalline silicon micromachined surfaces and describe a joint experimental-modeling technique that examines in depth when these forces are active and how they change with different processing and environmental conditions. In the experiments, microcantilever test structures were brought into contact with a landing pad in an environmental chamber. Adhesion energies were extracted from measured deflection profiles using finite element analysis. As roughness increased, the adhesion at a given relative humidity (RH) decreased, while the RH at which adhesion abruptly jumped, or the threshold RH, increased. Once the jump occurred, the adhesion increased toward the upper limit of \(2\gamma \cos \theta \), where γ is the liquid-vapor surface energy and θ is the contact angle. A detailed model based on the topography of the polysilicon surfaces as measured by atomic force microscopy was developed. Below the threshold RH, the adhesion could be modeled with only van der Waals forces active. Above the threshold RH, the adhesion was modeled by assuming that capillary menisci had nucleated. It was found that the effect of asperity plasticity was small while the effects of topographic surface correlations and disjoining pressure were important. Several possible mechanisms that might explain the threshold RH are examined.KeywordsAdhesion EnergyDisjoin PressureWater Adsorption IsothermCapillary AdhesionThreshold Relative HumidityThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.