Focal spot morphology reconstruction and dimensional quantification of microfocus x-ray sources using combined edges rotation method
Abstract
Micro-computed tomography (micro-CT) is a fundamental technology for detecting micro-defects and characterizing three-dimensional internal structures. It is widely used in advanced semiconductor packaging, new material research, and micro-nano manufacturing. The size and two-dimensional morphology of x-ray focal spots are critical parameters that determine the spatial resolution of micro-CT systems. Existing measurement methods have inherent limitations: they cannot precisely characterize the two-dimensional morphology or quantify the anisotropic morphology of micron-scale focal spots, and they rely heavily on micro-nano precision machining, resulting in limited universality. To address these challenges, this work proposes the combined edges rotation method (CERM). CERM utilizes an orthogonally combined edges structure to overcome the inaccurate rotation center positioning and operational complexity associated with conventional single-edge methods. Supported by a dedicated measurement setup and a comprehensive image processing pipeline—including edge region extraction, angle correction, and back-projection reconstruction—CERM enables high-precision dimensional measurement and two-dimensional morphology reconstruction of microfocus spots. Experimental results demonstrate that CERM accurately characterizes focal spot anisotropy, achieving relative deviations below 5% for approximately 4 μm single-stage focused x-ray sources and below 1.5% for approximately 2 μm two-stage focused x-ray sources at 600× magnification. Unlike traditional methods, CERM eliminates the need for precision micro-nano components (e.g. pinholes, masks), offering superior measurement accuracy, operational simplicity, universality, and simultaneous two-dimensional morphological characterization. This technique provides robust technical support for micro-CT system performance evaluation and x-ray source optimization.