Compressed Ultra-Fast Event-Based Electron Counting

Advanced techniques in transmission electron microscopy (TEM) such as in situ imaging, high-speed continuous tilt tomography, and ptychography are hampered by the lack of an ultra-fast electron counting camera with sufficient size and time resolution for these techniques. Currently available systems are an order of magnitude too slow, require severely limited beam intensity, and present practical challenges in the massive amounts of data delivered to the userĺs computer. Addressing these issues requires a new strategy. In this project we have begun addressing these challenges by developing a new edge-computing electron counting camera based on our novel high-speed, sparse, binary-readout direct electron sensor. In Phase I, we successfully completed a prototype camera, proving our edge-computing strategy for electron counting. In Phase II, we significantly improved this prototype and commercialized an initial product called ôApollo,ö targeted primarily for the cryo-EM market. The resounding success of Apollo in the cryo-EM marketŚimpressive results and growing salesŚclearly demonstrates the market viability of our technology. In Phase IIB, we will expand and commercialize the advanced features initiated in Phase II to broaden the market for this camera to for a much wider range of TEM applications. First, the features for in situ TEM developed in firmware during Phase II will be fully integrated into user-friendly software and our open API, enabling widespread use of the new camera for in situ imaging. Second, the time resolution of recorded data from our camera will be dramatically increased by implementing event streaming and compression in the camera's FPGAs, so that centroided detection events are streamed to the computer with the full time-resolution of our novel direct detection sensor. This will enable the camera to exceed the high-time-resolution requirements of future breakthrough TEM techniques. For example, continuous rotation tomography will require a large area and a time resolution of <500 µs to render inter-frame rotation negligible (<0.001 degrees tilt range per frame). Additionally, the feasibility of cryo-ptychography for biological specimens will require not only an electron counting camera with exceptional signal-to-noise ratio (SNR), but also at least another order of magnitude higher frame rate, so drift is minimized during large scans. Third, in situ imaging features will be added to our software to improve useability for in situ TEM experiments. Finally, a hardware trigger output from the camera will be implemented, enabling synchronization of the sensor with third-party hardware (such as fast electrostatic beam blankers and scan generators). The combination of these new features will not only enable in situ imaging of beam sensitive specimens with fast dynamics, but will also, for the first time, enable two next-generation techniques in electron microscopy: high-resolution continuous-rotation tomography and cryo-ptychography for biology.