Fast Mechanical Shutter in Hitachi HF 3300, a 60 kV to 300 kV TEM Marek Malac * , Robert A. McLeod * , Peng Li * , Wilbur C. Bigelow ● , Jane Howeº, Larry Allardº, Yoshifumi Taniguchi ▲ , M. Sergio Moreno ▀ . * National Institute of Nanotechnology, 11421 Saskatchewan Drive, Edmonton, Canada. ● Materials Engineering, University of Michigan, Ann Arbor, MI 48109-2136. º Oak Ridge National Laboratory, Oak Ridge, Tennessee. ▲ Hitachi High Technologies Corporation, Hitachinaka-shi, Ibaraki-ken, 312-8504, Japan. ▀ Centro Atomico Bariloche, 8400 – San Carlos de Bariloche, Argentina. The detection of electrons in a TEM is typically performed by means of a slow scan CCD camera (SSCCD), film, or image plates. In many advanced applications the ability to reliably and quickly shutter the beam incident onto a detector determines whether usable data can be obtained or not. Various approaches were explored over the years to shutter the beam either above or below the sample. Here we report on the performance of a mechanical shutter in an HF 3300 at NINT. This type shutter was originally designed by one of the authors (WB) for the Hitachi HF 2000 cold field emission microscope in 1997. The simple and reliable design is a drop-in replacement for the factory installed shutter, either above or below sample, or for both. As shown in Fig 1, the shutter. consists of a titanium barrel with four slits mounted in the beam path with the long axis perpendicular to the electron beam such that it intercepts the beam at a suitable position along the electron beam path. The barrel is rotated 45 degrees to change between “open” and “closed” positions. The rotation is achieved by a stepper motor mounted outside the microscope and triggered by the usual signal for the SSCCD camera (Gatan Ultrascan 1000 in case of the HF 3300 at NINT). Fig. 2 illustrates the shutter performance. The x-axis shows the nominal acquisition time on the SSCCD while the y-axis shows the number of counts summed across the entire camera. The shutter has linear response down to about 13 ms acquisition time (the readout time of the SSCCD is not affected). For shorter exposures the time required for the shutter blades to cut off the beam becomes sufficient to allow a constant residual exposure. It is likely that the acquisition time can be further decreased by tuning the program of the stepper motor controller; however, this seems unnecessary because about 20 ms acquisition time is entirely adequate for all experiments we attempted so far. Fig. 3 a) shows is a diffraction pattern (DP) acquired from a thick (as witnessed by the Kikuchi-lines) Si <110> sample using the shortest reasonable exposure time for the factory- installed shutter. The streaking of the diffraction spots and saturation of the main beam spot prevent reliable interpretation of this pattern. Fig 3b) shows a DP acquired from gold nanorods with less than 20 nm thickness on amorphous Si substrate about 5 nm thick using the new shutter and about one fifth the exposure time. Here the 000 beam is not saturated and the diffraction spots are sharply defined, allowing this DP to be reliably interpreted Fig 4a) shows an angular-resolved EELS spectrum acquired from a carbon support film via the Gatan Tridiem spectrometer using the spectrometer’s internal shutter with a 100 ms acquisition time. The streak artifact (marked by an arrow) indicates that sweeping of the beam away from the SSCCD occurred for a significant fraction of the exposure. The streak artifact is in the energy-loss region of interest for valence loss EELS such as band-gap determination. Indeed, when a vertically summed spectrum is acquired (as done when using Digital Micrograph Turbo View mode) the shutter streak can be easily misinterpreted as the tail of zero-loss, or as 338 doi:10.1017/S1431927610055777 Microsc. Microanal. 16 (Suppl 2), 2010 © Microscopy Society of America 2010 https://doi.org/10.1017/S1431927610055777 Published online by Cambridge University Press