Thickness measurements of supported thin polymer films on different substrates: Intercomparison between analytical techniques and their limitations

On the aggressive roadmap of nanotechnology stands the demand for reliable techniques able to perform direct and accurate characterization of materials. This basic requisite in modern materials science is crucial for understanding the emerging nanoscale properties of these materials as well as to optimize their performance in diverse applications, including biological and technological studies. Particularly, film thickness of polymers may control their mechanical (such as density, elastic moduli, stretchability), chemical (such as permeability, wettability) and thermal (such as glass temperature) properties. In this scenario, a comprehensive, critical comparison of film thicknesses determined from different analytical techniques is presented.

The material under study is polymethyl methacrylate (PMMA). Thin films supported on three different substrates (silicon, silicon dioxide and gold) were prepared by spin coating using anisole solutions. Film thicknesses varied between 1 up to 800 nm by changing the concentration of the solution. Three analytical techniques are compared, namely Atomic Force Microscopy (AFM), Nuclear Reaction Analysis (NRA) and Spectral Ellipsometry. They probe the film in very distinct ways, i.e. mechanically, collisionally and optically, respectively. The agreement is quite good for silicon and silicon dioxide. For gold, there seems to be a larger dispersion, which may be caused by the higher roughness observed in these films. Additionally, the techniques have their own limitations. AFM scans a limited area and the scalpel usage is not adequate for soft substrate materials as they may be removed along with the film. NRA is an ion beam technique, consequently, it is subjected to radiation effects as beam contamination, sputtering and damage. Furthermore, carbon adsorption may be critical for the evaluation of ultrathin films. Spectral Ellipsometry analysis relies on optically identifying the interface between two layers, consequently, the difference in their refractive indexes cannot be too small. Moreover, extracting two fitting parameters from data recorded for ultrathin films is also challenging.