Optically opaque materials such as metals and amorphous carbon are ubiquitous in modern nanoelectronic memory and logic devices. As their growth is usually not selective, i.e. they grow as blanket layers covering full wafers uniformly, optical techniques quickly become blind to any underlying material. As such, this leads e.g. to alignment and overlay challenges in the semi-damascene process flow or after the processing of the magnetic tunnel junction (MTJ) of a Magnetic Random-Access Memory. To mitigate this problem, mask alignment either relies on the topography transferred to the surface of the metal layer, leading to obvious accuracy issues, or requires extra expensive and time-consuming etch steps. In this context, Picosecond ultrasonic (PU) measurement is an ideal alternative since the generated and measured acoustic waves do propagate through these optically opaque materials. Use of PU to generate and detect high frequency surface acoustic waves (SAWs) in the GHz range on a periodically patterned nanostructures has shown a strong correlation between the frequency of the acoustic mode and the pitch of the sample.
In this paper, we evaluate the sensitivity of PU to objects with a simple linear geometry. We focused this work on metal line arrays of various geometries and show that the frequency profile of generated acoustics is dependent on the sample geometry including the width of the metal lines exposed or embedded under a blanket layer of the same metal.