This study was to investigate the internal friction of ultra-thin Cu film (nanometer scale) tested under different temperature (room temperature up to 120 C) under high vacuum conditions. Energy loss was measured according to the decay in the oscillation amplitude of a vibrating structure following resonant excitation. Our aim was to identify the temperature effect on the internal friction of ultra thin Cu films which has the greatest influence on its applications in IC and MEMS applications. 8.2 Experimental Detail The sample comprised a silicon cantilever beam under uniform stress supporting a relatively large paddle plate. Sample dimensions were as follows: frame (20 mm 20 mm), paddle plate (5 mm 5 mm), and length of tapered beam (3 mm), as shown in Fig. 8.1. Figure 8.2 shows the manufacturing process of paddle sample. High conductivity single crystal silicon was used for fabrication using the standard Si IC patterning and etching techniques reported in [5–7]. A cantilever beam with uniform stress distribution was used to support the extremely thin Cu films, which would otherwise be unable to support themselves. A thin (40 μm—the thickness of the tapered cantilever beam) section (uniform stress beam) was fabricated between the frame of the chip (250 μm thick—the thickness of the Si wafer from which it was fabricated) and the thick paddle plate (also 250μm). Due to the difference in stiffness among the paddle plate, tapered beam, and frame, all bending in the assembly occurred in the thin tapered cantilever beam. High conductivity Si wafers were used in place of conventional wafers used in the semiconductor industry. Metal films were deposited on the upper surface of the entire Si chip and differences in the behavior of the paddle with and without the metal film provided information about the metal film. The metal films were far thinner than the thickness of the cantilever beam; therefore, the strain experienced by the films due to beam deflection remained uniform across the thickness of the film and approximately equal to the strain in the top surface of the silicon beam. The test system presented in this work comprised a temperature controller, thermocouple, heater, PC, Labview software, power amplifier, charge sensitive preamplifier, lock-in amplifier and function generators. A Photograph of the system is presented in Fig. 8.3. The desktop computer sends a sine wave signal through the BNC connecter box to the power amplifier during testing. When the paddle is bent due to a change in voltage at the deflection electrode, the displacement current through the paddle capacitor as well as the sum of displacement current are altered. Measurement results were transformed from the time domain into the frequency domain using the Fast Fourier Transform (FFT) method. Bending force was generated through excitation, in which a waveform generator was used to drive the electrode beneath the paddle plate. Sweep frequency excitation was used to measure the resonance of the paddles, while constant frequency excitation was used to establish stable amplitude, before attempting to measure the free vibration of the paddle sample after excitation was ceased. The changes in displacement current were detected by the lock-in amplifier and recorded and stored using Labview. Fast Fourier Transform was then used to analyze the frequency component of the response. Environmental Fig. 8.1 Dimensions of the paddle sample 68 Y.-T. Wang et al.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==