A correlation-based optimization model to recover lost and distorted data from scanning tunneling microscopy images...

Scanning tunneling microscopy (STM) has significantly influenced the fields of nanoscience and nanotechnology. However, the tip effect and thermal drift cause loss and distortion of data in the STM images. Here, we propose a physics-guided optimization model for extracting STM imaging parameters, including tip shape, thermal drift, depth of field, current, and height. The model uses partial charge densities from density functional theory (DFT) simulation and works based on the mass comparison of experimental and simulated images using a two-dimensional Pearson correlation. Testing the model on Si(111)-7 × 7 reconstruction images provided higher than 96 % correlations for both biases. The fitting showed the highest correlation for only two bands in each image instead of the integration of all bands. Gaussian functions were used in the model to simulate the tip effect, which could recover 1.5-6 % of the lost data due to the blurring effect. Additionally, thermal drift was detected and corrected in the negative bias image, which could linearly distort the data by about 19 %. An important advantage of using this model is increasing the microscopy speed because there is no need to slow down the scanning process in microscopy experiments to evade thermal drift.