Charge-Coupled Devices (CCD)

A Charge-Coupled Device, or CCD, is basically an array of closely-spaced metal-oxide-semiconductor (MOS) diodes that can store and transfer information using packets of electric charge, or charge packets. Applying the proper sequence of voltage pulses (clock signals) to a CCD biases the array of MOS diodes into the deep depletion region where the charge packets may be moved in a controlled manner across the semiconductor substrate. Some people also refer to a CCD's MOS diodes as 'MOS capacitors.'

There are two basic types of CCD, namely, the surface channel CCD (SCCD) and the buried channel CCD (BCCD). Charge is stored and transferred at the semiconductor surface in the SCCD, while in the BCCD, the charge packets are stored and transferred in the bulk semiconductor below the surface.

The semiconductor substrate of a CCD may be n- or p-type. Over this semiconductor substrate, silicon dioxide is grown as a dielectric or insulating layer. An array of very closely-spaced metal electrodes is then formed over this dielectric layer. This is why the CCD is considered a MOS device, i.e., its top to bottom layers are metal, oxide, and semiconductor.

In an n-type CCD, grounding the substrate and applying a negative voltage -V1 to all the closely-spaced metal electrodes will create a depletion region in the substrate right beneath the oxide layer. This depletion region is devoid of majority carriers (electrons), since these have been repelled by the negative voltage applied at the electrodes. On the other hand, some of the minority carriers (holes) present in the substrate will be attracted towards this depletion region.

Applying a significantly more negative voltage -V2 at one of the electrodes while maintaining the other electrodes at -V1 will cause the depletion region beneath the more negative electrode (let's call it 'Ek') to extend more deeply (see Fig. 1). This deeper depletion region beneath Ek creates a 'potential well' that extends from the edge of the electrode to its left to the edge of the electrode to its right. This 'potential well' is the only region within the depletion region wherein a positive charge may move about. Thus, placing a positive charge into this potential well traps it there, since it can not move outside the well. This, in essence, is how a CCD stores a charge, and how it can be used as a memory device.

Figure 1. A simplified cross-section of a CCD

The stored charge under the more negative electrode Ek (and the one which has an extended depletion region) may actually be laterally transferred to the next adjacent electrode, Ek+1. This is done by also putting Ek+1 at the more negative voltage -V2, while allowing Ek to adjust back to -V1. While this is happening, all other electrodes (esp. Ek-1 and Ek+2) must be at -V1. As Ek ramps up to -V1, its potential well becomes shallower, and all the trapped charges in it gets transferred to the deeper potential well of Ek+1. Eventually the charges previously stored by Ek will get stored in the potential well Ek+1.

Moving the charge packets from one electrode to the other in this manner requires at least three electrodes to move one bit of information. In any transfer cycle, one electrode is at -V2, another electrode is ramping up from -V2 to -V1, and the third electrode is at -V1. The lateral movement of charge packets in a CCD from one electrode to the next is very similar to how digital data move in a shift register.

Although the CCD was invented as a memory device, its extreme sensitivity to light soon made it a popular choice as an image sensor. In an image-sensing CCD chip, each MOS diode or 'capacitor' represents one pixel. The charge packets are generated when light excites electrons in the valence band into the conduction band. The light-generated charge packets that carry the image information are stored and transferred from one potential well to another until they are eventually shifted out of an output register. Most video cameras today use a CCD for its image sensing requirements.

HOME