Note: Artificial neural networks for the automated analysis of force map data in atomic force microscopy
Force curves recorded with the atomic force microscope on structured samples often show an irregular force versus indentation behavior. An analysis of such curves using standard contact models (e.g., the Sneddon model) would generate inaccurate Young's moduli. A critical inspection of the force curve shape is therefore necessary for estimating the reliability of the generated Young's modulus. We used a trained artificial neural network to automatically recognize curves of “good” and of “bad” quality. This is especially useful for improving the analysis of force maps that consist of a large number of force curves.