Surface profile 3D visualization from thickness map reconstruction of thin films using scattering of surface plasmon polaritons

We present a method to obtain wide (~mm2) three-dimensional (3D) thickness maps of dielectric thin films, with sub-nanometer precision, using the scattering produced by surface plasmon polaritons (SPPs). SPPs are... [ view full abstract ]

We present a method to obtain wide (~mm²) three-dimensional (3D) thickness maps of dielectric thin films, with sub-nanometer precision, using the scattering produced by surface plasmon polaritons (SPPs). SPPs are electromagnetic modes coupled to electron plasma oscillations confined to the interface between a metal and a dielectric. We use the phenomenon of SP resonance to determine the thickness of a thin film. The experimental setup (Kretschmann configuration) with an imaging system was used to take images with a CCD camera of the same area of interest as the prism is rotated [Figure 1]. Optical images of the surface at different angles of excitation were achieved. Each pixel of the CCD as individual power detectors was used to register the intensity of the scattered light. By using an automated post-processing algorithm that runs a MATLAB script we find the thickness of the film at each point of the surface, corresponding to every pixel of the camera, using a semi-analytical numerical fit (Figure 2). Figure 3 shown scattering measurements obtained with the CCD camera and fitting (solid lines) for the four fabricated samples with flat surfaces. Figure 4 shown the full 3D reconstruction of the dielectric film thickness. The details of the logo are clearly visible. The inset shown an atomic force microscope topography measurement of a transverse section x along the edge of the structure. The samples consisted of thin magnesium fluoride (MgF₂) coatings (< 100 nm), deposited via thermal evaporation on top of a 45-nm-thick gold film on a glass substrate. The intensity of the scattered light was detected by SPPs decoupled by the local roughness of the films. We used the correlation coefficient R² to fit the results as a figure of merit to determine the minimum amount of square pixels that would be necessary to average, in order to get a reliable experimental curve (R² > 0.9). A value of R² = 0.9279 was found for the 3x3 array case. The resonances were analyzed through numerical fitting to find the thickness of dielectric films. This technique can be applied to investigate macroscopic inhomogeneities in dielectric thin films.