AFM investigations of the nanoscale roughness of polymer replicas

The polymer materials are widely used in replication technologies. Recently different methods such as polymer imprinting, casting, hot embossing, injection molding were successfully used for polymer replication of the different surface structures with nanometer scale elements [1-4]. However, most investigations are connected only with resolution studies and directed at fabrication of sub-micron elements with a high aspect ratio. The quality of a polymer replica surface on sub-nanometer scales was not investigated. The present work is devoted to SPM investigation of the polymer films surfaces on glass substrates to be used in X-ray mirror fabrication. Thin films on the base of acryl anaerobic adhesives and photopolymer compounds were investigated as possible replicate layers. Deep polished silicon wafers (thickness is 0.5 mm) and zerodur substrates (thickness is 5 mm) were used as the etalon surfaces for the replicas preparation. A possibility of smoothing the roughness of different glass substrates (rms roughness about 1.5 and 0.7 nm measured by x-ray reflection methods) was investigated. The polymer replicas were prepared in the following way. Thin layers of liquid prepolymer were spread evenly on a glass substrate surface. Afterwards, etalon plates and the glass substrate were pressed together. Liquid polymer remaining on the edges of samples was removed to protecting of the replica surface. The polymerization process was carried out in the absence of air or under light illumination in free state without any external pressure. As a result, thin polymer layers of near 5 μm thickness were formed between the two surfaces. Etalon plates were separated from polymer layer by flexural deformation for silicon plates and under thermoelastic stress in the etalon sample – glass substrate system for zerodur substrates. A scanning atomic force microscope “Solver P-47” designed by NT-MDT company (Zelenograd, Russia) was used for the surface roughness investigation. A set of frames of the same sizes was taken from different areas of the sample surface. The root-mean-square roughness was calculated as the average value for the set of AFM frames of certain scales. Our AFM measurements and other authors’ investigations showed that practically all surfaces have a multiscale relief structure. The main component of this structure is the low scale (Gauss) roughness. But there are a lot of various large-scale features increasing the root-mean-square roughness in the large AFM frames. Thus, as a result of AFM measurements we formed the scale dependences of the value on the frame size, which characterized surface roughness at different scales. σ