[Nicholas Lee (OP)]

If you have two gamma knife beams, and the beams could be adjusted in width.
How thin could the gamma knife beams become, to ionize a cubic area as small as 20 microns.
Is there ball lens technology that allows you to collimate gamma waves, or beams  to do this.
If it cannot be done with gamma waves, what about X- rays, or other EM waves that ionize brain cells.
I am grateful for your help, anything helps even a few words.

[Atomic-S]

To focus on an area of 20 microns width, a good wavelength would be 2 microns or shorter. That would put it in the infrared band. Potentially therefore you can use any radiation shorter than that, including visible, ultraviolet, X, and gamma. If you want to ionize stuff using reasonable power levels, the quantum energy has to be sufficient, and ultraviolet may be sufficient.  It is probably possible to do what you want using presently-available lasers. The choice of wavelength would also be dictated by possible effects on the surrounding environment.  Lenses to handle radiations that are in the longer ultraviolet band and longer are no problem.  However,the nature of the source must also be taken into account. Not even the best lens can focus the light of an overcast sky into an intense pinpoint. Laser light is extremely focusable. But lasers tend to be available mainly in the visible and longer wavelengths. I don't know if ultraviolet lasers are available, but ultraviolet LEDs are readily available, but whether the power would be adequate is another question.  For X-rays, as far as I know we are still limited to vacuum tubes, but these certainly exist and are capable of producing a small emission source compact enough for imaging, which may well be workable with additional optics to give you the tiny impact zone you require.   With shorter wavelengths, things get more difficult, although X-rays have been successfully focused using grazing-incidence mirrors. I don't know any way to do pinpoint application of gamma rays except by direct insertion of a tiny radioactive source at the point of application.  I  think you should look first at visible lasers for brain surgery. However, if the point of application is inaccessibly within tissue, then X-rays may have to be considered. Here, the geometries could get complicated because the farther the point of application is from the lens or mirror structure, the harder it will be to get a narrow spot. This could be a problem if the source is an X-ray tube and the optics consist of grazing-incidence mirrors. (But of course, under such conditions, would you even be able to tell, to within an accuracy of 20 microns, exactly where to position it?)