Characterization of Hysteresis of Surface Energy in Room-Temperature Direct Bonding Processes David S. Grierson and Kevin T. Turner Department of Mechanical Engineering, University of Wisconsin Madison, WI 53706 USA Room temperature direct wafer bonding is driven by surface forces and the formation of hydrogen bonds that pull the wafers being bonded into intimate contact. The magnitude of these forces is often represented as a work of adhesion, and is critical as it determines the level of non-flatness and patterning that may be tolerated in bonding. Traditionally, it has been assumed that the work of adhesion that drives the formation of the bond is equal to the work of separation that is measured through fracture tests, such as the familiar razor-blade test. In the present work, we use a novel test structure to measure both the work of separation and the work of adhesion of room temperature hydrophilic Si-Si bonds as a function of relative humidity (~15 – 50% RH). The results show that across a range of values of RH, the work of adhesion is significantly lower than the work of separation. Introduction Direct wafer bonding is an important process that enables the fabrication of a wide range of micro- and nanosystems (1). This process is commonly used in the fabrication of SOI substrates (2) and complex multi-wafer MEMS devices (3). In direct bonding, initial adhesion between the wafers is typically achieved at room temperature via van der Waals forces and hydrogen bonding. The strength of the adhesion is commonly described in terms of a single parameter, the work of adhesion (W adh ). The work of adhesion of bonded wafer pairs is most commonly measured with the razor-blade test (4), which is essentially a fracture test. The value of work of adhesion measured in the test is truly a work of separation value, rather than a work of adhesion value. Here, we measure both the work of adhesion and work of separation to establish if they are equal to one another in room temperature silicon direct bonding. Previous researchers have used microfabrication techniques to develop measurement schemes for studying stiction between polycrystalline silicon beams and silicon substrates with various surface coatings as a function of humidity [(5, 6), for example]. These techniques have proven to be beneficial in understanding adhesion between micromachined surfaces, but are limited in their ability to fully characterize the adhesion and separation behavior bonded interfaces directly, and the range of material pairs is also limited. The novel measurement approach detailed here is similar to the previous micromachined beam studies reported previously, but is distinct in that the beams are not ECS Transactions, 33 (4) 573-580 (2010) 10.1149/1.3483549 ©The Electrochemical Society 573 Downloaded 29 Nov 2010 to 146.151.127.221. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp