The upper image shows measurements of the 7x7 reconstruction of a silicon surface. The lower shows the model for the same surface. The similarity between measurement and model is striking. Both images were taken from Surface Science Letters, 157, G. Binnig, H. Rohrer, F. Salvan, Ch. Gerber, A. Baro, "Revisiting the 7x7 reconstruction of Si(111)", L373-L378, Copyright 1985, with permission from Elsevier Science.

It took 10 months for the next paper to be submitted, in collaboration with Salvan, Gerber, and Baro, to Surface Science Letters. The images show their findings of the Si(111) 7x7 reconstruction. They found three-fold symmetry for both n- and p-type with an occasional missing maximum, attributable to a surface boron.

We are talking about doped silicon here, and about work going on at IBM. Is there a computer chip connection? Was IBM was funding STM research precisely in order to get leverage in electronic chip research. As our interview reveals (members may click here), the initiative was entirely Heinrich Rohrer's. It did not hurt that IBM might have interest in tools with which to characterize silicon crystal surfaces, but the IBM scientists were free to choose their own research topics. The main reason, Binnig and Rohrer worked on silicon surfaces was that surface scientists were generally interested in it, and obviously their interest was prompted by the large field of application in the computer industry. So there is a computer chip connection, but it is a little roundabout.

Experimental details mentioned reveal the kind of resources required to perform the experiment: the 7x7 reconstruction was generated by annealing at 900-1200 degrees Celsius. The degree of contamination was not checked by an independent method. The actual results were checked against low energy electron diffraction and Auger electron spectroscopy, though. They at first could only get images with voltages above 2.5V and at 300 degrees Celsius, but found a mundane explanation.