Computerized analytical electron microscope for elemental imaging

Abstract

A computer system has been interfaced to an analytical scanning transmission electron microscope (STEM) to form an integrated system for high-resolution mapping of the elemental constituents of a specimen. The system controls the electron beam position, acquires data from electron energy-loss spectroscopy (EELS) and energy dispersive x-ray spectroscopy (EDS) detectors, and constructs elemental images by analyzing EELS and EDS spectra taken at each pixel. Data also are acquired and digitized from conventional STEM bright-field and dark-field detectors. Since image registration errors are eliminated by acquiring data from all detectors concurrently, elemental distribution images obtained from energy-loss and x-ray detectors can be correlated with morphological images taken from bright-field and dark-field detectors. Energy-loss and x-ray spectra of user-defined target areas can also be obtained. Data can be acquired, processed, and displayed at the same time because a satellite microcomputer interfaced to the microscope does much of the data acquisition, freeing the host computer to subtract the spectral background from the electron energy-loss and x-ray data ‘‘on the fly,’’ and also to display dynamically the background corrected energy-loss spectrum at each image pixel. Such a display is important for correct operation of the instrument and interpretation of the results. Images are displayed on a color display system equipped with a digital video array processor, where they can be enhanced, compared, measured, annotated, and photographed. Operation of the system is simplified by using menus for function selection and by filling out forms displayed on a video terminal to enter data-acquisition and processing parameters. The computer-controlled analytical electron microscope is used to provide elemental distributions from thin specimens in biology and materials science. Results show that concentrations of a few atomic percent can be mapped at a resolution of 10 to 20 nm with our system. The importance of real-time data processing is demonstrated in the case of EELS imaging by accurate subtraction of the spectral background at each pixel, and by the correction for spectrum drift during the course of data acquisition.