In situelectron microscopy is an investigatory technique where an electron microscope is used to watch a sample's response to a stimulus in real time. Due to the nature of the high-energy beam of electrons used to image a sample in an electron microscope, microscopists have long observed that specimens are routinely changed or damaged by the electron beam. Starting in the 1960s, and using Transmission Electron Microscopes, scientists made deliberate attempts to modify materials while the sample was in the specimen chamber, and to capture images through time of the induced damages. Also in the 1960s, materials scientists using TEMs began to study the response of electron-transparent metal samples to irradiation by the electron beam. This was in order to understand more about metal fatigue during aviation and space flight. The experiments were performed on instruments with high accelerating voltages; the image resolution was low compared to the sub-nanometer resolution available with modern TEMs. Improvements in electron microscopy from the 1960s onwards focused on increasing the spatial resolution. This required increased stability for the entire imaging platform, but particularly for the area around the specimen stage. Improved image-capture systems using charge-coupled device cameras and advances in specimen stages coupled with the higher resolution led to creating systems devoted to applying stimuli to samples in specialized holders, and capturing multiple frames or videos of the samples' responses. In addition to materials samples, in situ electron microscopy is performed on biological specimens, and is used to conduct experiments involving mechanical, chemical, thermal, and electrical responses. Early experiments mostly used TEMs, because the image is captured in a single frame, whereas the Scanning Electron Microscope must move or scan across the sample while the stimuli is being applied, altering the sample. Early problems that limited in situ electron microscopy included mechanical vibration at all scales, and thermal and electrical interference, particularly at the specimen holder. These problems all required fast capture times. However a fast capture time creates an image with a low signal-to-noise ratio, limits the resolution of the image, and also limits the amount of time available for conducting the experiment.
