Abstract

Thin films prepared by physical vapor deposition (PVD) generally exhibit large residual stresses which are mainly related to the deposition process [1]. Due to the large elastic energy stored in such thin layers, stress relaxation may occur from thin film decohesion i.e. development of microcracks at the interface [2]. According to the stress sign, we observe film peeling (tensile) [3] or buckling (compressive) [4]. In this last case, circular and/or “cord telephone” blisters may appear depending on stress magnitude (large for cord telephone). Although these buckling phenomena are undesirable for technological applications, they may be useful to estimate intrinsic mechanical properties such as internal stresses or Young modulus and adhesion energy from blister geometry and particularly from its maximum deflection. This method may be used as an alternative one to the Bulge-Blister test [5] which is extremely difficult to achieve because of the extreme care required in realizing a fine circular orifice through the back face of the substrate (generally silicon). Furthermore, the analysis may be done for different substrates (glass, polycarbonate, silicon, metals...) which allows one to vary the film/substrate adhesion energy. In this letter, we would like to underline the interest of AFM investigations for small blister size (first bulking stage). In this way, we present first results concerning in-plane residual stresses obtained on 200 nm thick Ni and Mo films deposited by ion beam sputtering on polycarbonate and thick silicon substrates respectively. Results are compared to the average in-plane stress deduced in films adherent to a thin silicon substrate (200 μm) from the curvature of the substrate according to Stoney formula [6, 7] and also, in the case of Mo film, to the average in-grain stress obtained by X-ray diffraction measurements using sin2ψ method [7–9]. Thin Ni and Mo films were deposited at room temperature using an Ar ion beam sputtering technique. The sputtering system was already described in details in reference [10], so we just give the deposition conditions here. The starting pressure was 2× 10−8 Torr and during deposition, the pressure was maintained at about 10−4 Torr. Both deposition rate (0.08 nm/s for Ni and 0.05 nm/s for Mo) and film thickness (200 nm) were controlled by a quartz oscillator and the actual thickness of the as-deposited layer was determined ex-situ with a Dektak profilometer from a step height. For curvature and X-ray diffraction measurements, the films were deposited on thin silicon wafers (200 μm thick) while polycarbonate and thick silicon (650 μm) substrates were used for AFM observations. All silicon substrates were cleaned in acetone with ultrasonic treatment and then dropped in alcohol. Thus, the silicon substrates are covered with native oxide. The AFM images presented in the following were taken in contact mode along the trace direction and no significant effects have been noticed in relation to the scanning direction. V-shaped cantilevers with a theoretical normal spring constant of 0.1 N/m and sharpened pyramidal silicon nitride tips were used in this study. Fig. 1 shows an example of a circular blister observed by AFM on Ni/Polycarbonate set. The scan size is about 1.73× 1.73μm2 with a height extension of 27 nm. This

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call