3D SEM for structural biology

3D SEM for structural biology

SUPERVISOR: Trevor Sewell

The feasibility of high resolution, three-dimensional scanning electron microscopy and its application in structural biology

Three dimensional scanning electron microscopy is more of a concept than a reality. The purpose of this project is to explore the concept and analyse each of the underlying technologies to see whether there is a class of problems in structural biology that could be solved by a combination of the technologies.

The key enabling technology is the algorithm (in the process of development) by Dr Fred Nicolls of the department of Electrical Engineering, UCT. The project will explore the possibility of creating a machine which could usefully exploit this algorithm. The student will also use the algorithm to make a number of three-dimensional images of biological objects using the available SEM (Leica/ Zeiss S440 cryo-SEM).

The algorithm

Reconstructions are made from a series of images taken of the specimen at different angles. At the present time we have only utilised a single geometry - rotation of the tilted specimen by 360 degrees around the tilt axis. Other geometries are feasible and this is not really a limitation.

All that can be done at present is the extraction of a low density pointset from a low noise image sequence. You can see it at here with the original images in this directory.

You'll need a VRML viewer if you don't already have one. Use Cortona: http://www.parallelgraphics.com/products/cortona

After loading the file press "Fit" (bottom right) and "Study" (middle left) to view the reconstruction.

The lines in the image are the estimates of the principal rays of the camera views. This file just shows the point set without the lines.

It is anticipated that progress will be rapid in the next few months - progressing through generating a high density pointset, mesh and texture. Problems that will be addressed will be the use of noisy (low dose) images and structure refinement.

The class of problems

Surface images have made a contribution to our understanding of cells and macromolecules. Work in this area has a long history but recent examples can be seen in the papers from the Heuser group in New York, Martin Goldberg in Durham and the Hoenger group at Heidelberg among many others. Heuser has adressed problems of cell interior structure, Goldberg has studied the nuclear pore basket and Hoenger has focussed on fibrous structures, mostly in isolated states or with a controlled number of components. (Note that neither Heuser or Hoenger use the SEM to obtain their images.) The task would be to argue specifically that a certain class of projects would benefit from the use of the combination of technologies suggested.

The instrument

The components of the instrument that need to be analysed would include the electron source, the cryo transfer system, the stage, the shadowing system, the detector system and the possibility of using low vacuum techiques.

It is certain that the highest resolutions are only obtainable with a field emission electron source. The project would entail a detailed benefit analysis for the class of problems selected and would consider the benefits in particular of low accelerating voltages. The benefits would have to be weighed against cost and complexity.

The instrument would have to include a cryo-transfer system which also has apparatus for specimen coating. This is certainly the key component of the instrument. Coating technologies are based on metal beam devices using rotary shadowing or on sputter coating at higher pressures. The work of Martin Mueller may be of use in assessing what could potentially be done in this area. A key difficulty with coating is obviously that a layer of some metal (be it Pt-C or W) imposes structure of its own. It is usually necessary to have this in order to conduct away the charge deposited by the electrons. But there are other ways of preventing charging - the most obvious being to allow gas molecules to do this by raising the pressure in the sample area. The downside of this is that the resolution is compromised by interaction of the gas and the incoming electron beam. An important part of the project would be to explore the possibilities and find the optimum mix of technologies.

There are three manufacturers of SEMs that sell in the South African Market - FEI, JEOL and Zeiss. There are also third party companies that manufacture add-ons for specific purposes - the most important of these is possibly Gatan. The student should look at the available products from these companies and see whether any real-world machine is a close approximation to the "ideal" instrument and also ascertain the cost of this machine and of the modifications necessary to make it "fit to task".


The student will be required to make a number of three dimensional reconstructions of interesting biological objects using a variety of different instruments and experimental techniques. The primary instrument in the early stages of experimentation will be the Leica S440 in the EMU at UCT. However access to the FEGSEM in Pretoria and the ESEM in Pietermaritzburg will be negotiated so that the student can gain experience with these technologies. If good progress is made it may be possible to do experiments on the cryo-FEGSEM at the University of Durham and present the results at the 3DEM Gordon conference in Barga, Italy.