Poly to Diffusion

At the time that the 6561 was made, the NMOS process usually used "buried contacts" for poly to diffusion contacts. An older approach called the "butting contact" used a small metal section to connect the poly to diffusion, but this took up more area. The "buried contact" was smaller and was subject to fewer design rules and therefore became more widely used in NMOS than the butting contact.

In many places the 6561 uses both polysilicon and diffusion lines as wires or paths. Quite often these "wires" alternate between polysilicon and diffusion, i.e. pink, green, pink, green, etc. when under metal, or grey, brown, grey, brown, etc. when not under metal, and they'll also be very close to each other. This is because the design rules allow for the gap between poly and diffusion lines to be smaller than between two poly lines or two diffusion lines. This is discussed in more detail in the Lines section. It is of relevance here though because a line might come in towards a transistor gate as a diffusion line (due to it having come some distance as a group of alternating poly and diffusion lines), but obviously in order to form a transistor, that gate line needs to be polysilicon. So you will see many examples of a connection between diffusion and polysilicon so that a transistor can be formed. You'll also see poly lines changing to diffusion lines so that a pass transistor can control that line's entry in to a particular part of the chip.

Another common example of a poly to diffusion contact is in depletion mode pull up transistors. Such transistors have their gate connected to their source, and given the gate is always polysilicon, and the source always diffusion, it requires a polysilicon to diffusion contact.

The above diagram shows a number of these different examples of poly to diffusion contacts (all buried contacts):

• • In the upper left, there is a brown poly line coming in. It needs to switch to being a diffusion line so that the pass transistor can control its entry.

  • Once past that pass transistor, the diffusion line then switches back to being a poly line so that it can be the gate of the two pull down transistors.

  • Although difficult to spot, there are two large depletion mode transistors under the the wide metal line. Both of these have a poly to diffusion contact at one end.

  • Just to the left of the second metal line from the right, there is a pass transistor under which a diffusion line passes. This diffusion line then connects to a polysilicon gate.

The easiest way to tell whether a poly to diffusion contact is indeed a contact and not a transistor is by looking for the absence of the polysilicon overlap that is typical of a MOS transistor. The design rules state that the polysilicon when passing over diffusion to form a transistor must continue a short distance beyond the diffusion. A poly to diffusion contact will not have this overlap.