According to Smith's Nobel Prize lecture, the driving force behind the invention was an inter-departmental competition for resources inside Bell Labs (a purely bureaucratic affair.)
Remarkably, ATT didn't benefit commercially from this and many other inventions developed at Bell Labs.
Bill Boyle was Executive Director of the semiconductor part and I was a Department Head under him. Jack Morton was anxious to speed up the development of magnetic bubbles as a major memory technology, and there was talk of transferring resources from Bill’s division to the other where the bubble work was being done. For this not to happen, Morton demanded that Bill’s division come up with a semiconductor device to compete with bubbles. To address this demand, on October 17, 1969, Bill and I got together in his office. In a discussion lasting not much more than an hour, the basic structure of the CCD was sketched out on the blackboard, the principles of operation defined, and some preliminary ideas concerning applications were developed.
The figure from US Patent 3,792,322 (below) shows an improved version of the sensor: Buried Channel CCD.
Here's how Smith describes the train of thought they used to come up with the invention:
- First, the semiconductor analogy of the magnetic bubble is needed. The electric dual is a packet of charge.
- The next problem is how to store this charge in a confined region. The structure which came to mind, of course, was the simple MOS capacitor shown in Fig. 3. Charge can be introduced into this depletion region with the amount of charge stored being the magnitude of the signal. To understand this better, a plot of electron energy vs distance into the structure is shown in Fig. 4. As charge is introduced into the depletion region, the potential at the surface rises until the maximum allowable charge is reached. Any further charge added will flow into the substrate.
- The last problem was to shift the charge from one site to another, thereby allowing manipulation of the information. This is solved by simply placing the MOS capacitors very close together as shown in Fig. 5, one with charge and the second empty. In order to pass the charge from one to the next, one simply applies a more attractive voltage to the second, causing its depletion region to overlap the first and the charge to flow along the surface to the silicon-silicon dioxide interface of the second capacitor.