Probe conditioning via convolution neural network for scanning probe microscopy automation
Zhuo Diao, Linfeng Hou, Masayuki Abe
Abstract
Abstract We present an automation system for conditioning a scanning probe microscopy (SPM) probe into different states on a Si(111)–(7 × 7) surface at room temperature. Topography images representing multiple surface states and probe condition states divided into 11 categories and trained by a convolution neural network with an accuracy of 87% were used to estimate the effectiveness of the probe with an accuracy of 98%. We demonstrate the responsiveness of the method by experimentally reforming a probe into different conditions defined by preset categories. This system will promote advancements in autonomous SPM experiments at atomic scale and room temperature.
Topics & Concepts
Convolution (computer science)Scanning probe microscopyAutomationConvolutional neural networkArtificial neural networkConditioningSurface (topology)Materials scienceArtificial intelligenceMicroscopyScale (ratio)Computer scienceOpticsNanotechnologyPhysicsMathematicsEngineeringMechanical engineeringStatisticsGeometryQuantum mechanicsForce Microscopy Techniques and ApplicationsSurface and Thin Film PhenomenaAdvanced Electron Microscopy Techniques and Applications