Originally posted by fluffy
Ah. It sounds like it was an experimental device, then, and not something which would scale up to the physical size needed for modern ICs, much less TFT displays (which are basically a gigantic IC in many respects). Not to mention that modern ICs have *much* smaller gates. 
*********************
Well, he was having success at the time.
It's been over a decade.
It only took a few minutes ( under 30 ) for the average repair.
If they can manipulate individual atoms and spell words with them, I don't think a sub-pixel that I can image with the loupe in my Swiss army knife should be such a challenge... punch out the pixel as a cylinder, from behind the glass face. Pop a pre-fab micro cookie pixel in its place, a little laser work, a little epoxy or acrylic, and you have replaced a whole pixel... in the center of an Apple HD 23" display.
Not a bad deal.
I can even imagine LCD's manufactured by popping pre-fab pixels, as little cylinders, that are "pushed" into the holes of a black mask similar to that used in CRT's. The necessary contacts already etched onto the vertical walls of the cylinders to mate with their correlatives along the walls of the holes of the mask. The lit pixels would bloom against the glass face and obliterate the black mask. Dark pixels would not and the mask would yield a very high quality "black" for dark scenes.
Repair would consist of "pushing" a bad pixel plug through and out with a good pixel plug.
Of course one would not want to do a display with half the screen gone bad.
I don't think LCD's are nearly as complex as a modern IC.
The circuitry of each pixel ass'y is identical to all others and very simple in its topology.
And the transistors are not as small.
The repair is virtually identical for each case.
The geometry too.
This cries out for pre-fabricated pixel plugs
... and automation.
When you figure $2,300 for a display trashed vs. some five minutes to repair -- sounds like a good deal to me.
But this is all academic, I expect OLEDS will take care of the problem in economic terms: a roll of FOLED can be tested, on the fly, and CUT where the flaws are. This will certainly work statistically for far fewer trashed screens. The current method loses an entire piece of mother glass due to a flaw. The roll-edit way, you get a smaller screen (say, 15" instead of 17") coming out the other end to be grouped with its like, followed by several perfect 17" screens. The rectangular 2" ribbon edit could be further trimmed for use in cell phones, etc. In the end, only a tiny square with a bad pixel at it's center would remain.
Fun, huh ?
---gooddog
*****************************
