The accuracy of STEM mass measurements is limited by counting statistics. Theexpected performance of STEM can be calculated easily from probe size, specimen geometry, detector geometry, atomic scattering cross-sections, and dose. The bright field, small angle (SA) dark field, and large angle (LA) dark field detectors each have counting efficiency close to unity. For typical specimens, we can calculate the expected mass accuracy at 10 electronS/A2 : Fab fi-agments (50 kDa q 12%), glutamine synthetase (620 kDa i: 2Yo), earthworm hemoglobin (3.6 MDa q 0.9%), and tailess T7 particles (50 MDa 0.2%). Mass accuracy agrees well above 2%, but does not improve as expected for largzer particles. We have adopted tailess T7 vinis (40 MDa) as a test specimen because of its well-defined structure and stability. With this we are investigating factors affecting the accuracy of mass measurements such as: wash buffers, substrate parameters, f1reezing conditions and fi-eeze drying parameters. A PCMass program views the mass profile of particles in the STEM image in comparison to models in various orientations to determine distortion or locate defects. With the availablfity of STEM3, we have begun parallel studies of the same specimens in both microscopes. Those giving clean backgrounds and homogereous particles in STEM I are transferred under vacuum to STEM3 and remeasured. In STEM3 we have more control of beam dose, specimen temperature (down to LHe temperature) and data acquisition parameters. Frozen specimens can also be loaded into STEM3 and freeze dried in the stage while being observed.
Showing the most recent 10 out of 101 publications
