Adhesion Energy and Bonding Wave Velocity Measurements

Abstract

Direct bonding is now widely used in microelectronics for SOI elaboration or 3D integration. One of its main characterizations is the adherence energy, also called bonding energy, which estimates the strength of the bonded structure and its evolution with the different process parameters (annealing temperature, plasma activation…). The double cantilever beam (DCB) technique described by Maszara et al. [1] is the most common technic to measure the adherence energy. During the bonding another key characterization, the adhesion energy, deserves to be studied. This quantity gives information about the attractive forces involved when contacting the two surfaces and helps to understand the bonding wave velocity. Rieutord et al. [2] suggested an analytical model based on air flow dynamics between the two wafers and showed an analytical relation between the adhesion energy and the bonding wave velocity. In this work, a simple measurement apparatus derived from the DCB technique has been developed to measure the adhesion energy and the bonding wave velocity. Beams cut in <001> bonded silicon wafers are opened at the bonding interface by a blade insertion; then the blade is suddenly partially removed allowing the structure to be bonded again. Recording multiple infra-red pictures one can follow the rebonding wave propagation and measure its velocity. The blade allows an easy and long debonded area in order to be sure to detect a stable rebonding wave speed. As a part of the blade is not totally removed, one can also determine the remaining debonded length when the wave propagation stops due to the known blade thickness. The energy associated to this length and calculated using El Zein [3] equation is the adhesion energy. In this experiment, upper and lower substrates of the bonded structure are allowed to be deformed. We also check that measurements performed on 2 cm wide beams are equivalent to larger beams in order to prevent any fluid edge effect as described by Radisson et al. [4]. So according to these considerations, we assume that the apparatus well describe the Rieutord’s hypothesis of infinite wide beams free to deform in the perpendicular direction and we can then compare our measures to his analytical predictions. We will show in this paper a very good correlation between the adhesion energy and the bonding wave speed. The surface energy has been varied by changing the surface roughness. Surprisingly, an interesting evolution of the adhesion energy after the blade stop will be presented. During 90 minutes, a slow increase of the bonding energy is observed at room temperature and 50% Relative Humidity. A mechanism will be proposed to explain this long time evolution mainly based on a slow kinetics of capillary condensation between the rough surfaces [5]. We will then discuss the choice of the energy adhesion value and show that long time values combined to measured velocities well agree with Rieutord’s model (cf. figure). REFERENCES [1] W. P. Maszara, G. Goetz, A. Caviglia and J. B. McKitterick, “Bonding of silicon wafers for silicon on insulator”, J. Appl. Phys., 64 p 4943 (1988) [2] F. Rieutord, B. Bataillou, and H. Moriceau, “Dynamics of a bonding front”, Physical Review Letters, vol. 94, pp. 236101, (2005). [3] M. S. El-Zein and K. L. Reijsnider, “Evaluation of GIC of a DCB Specimen Using an Anisotropic Solution”, Journal of Compo. Tech. & Research, 10(4), pp.151-155, (1988) [4] D. Radisson, “Direct bonding on patterned surfaces”, PhD thesis, Grenoble Univ., France (2014) [5] F. Fournel, C. Martin-Cocher, D. Radisson, V. Larrey, E. Beche, C. Morales, P. A. Delean, F. Rieutord, and H. Moriceau, “Water Stress Corrosion in Bonded Structures”, ECS Journal of Solid State Science and Technology, 4 (5) P124-P130 (2015) Figure 1