Ultra Shallow Junctions

As technologists push the limits of how shallow a junction can be fabricated, terminology has emerged which calls these aggressive processes “Ultra Shallow Junctions”.

Transistor Requirements

As transistors are made smaller, the junctions that form the source and drain regions of the transistor must be made shallower to improve performance and provide adequate breakdown characteristics.  In particular, the region known as the Drain Extension must be extremely shallow to create a high performance device.  The Drain Extension requires the shallowest junction and therefore the lowest implant energy.  For current production technology (130nm), the drain extensions need to be around 30-40nm in depth.  The next generation (100nm, 2003) will require drain extensions of 20-30nm depth, and future technology will require even shallower junctions.  These requirements are very challenging.

P-type USJ

The P-type junction is more difficult to make shallow than the N-type, so the P-type USJ is the most challenging implant process.  The P-type implants use boron as the impurity and the characteristics of boron create the difficulty in making this junction shallow.  First, the boron atom is small and light (only 11amu) and therefore the boron atom penetrates more deeply into the silicon crystal upon implantation than any of the other typical impurities used.  This means that boron implants must use lower energies to attain an equivalent depth profile from the implant.  Next, all implants must be annealed (heat treated) to produce the proper electrical characteristics, and the boron atoms diffuse quickly during anneal.  Thus, a P-type USJ is the most challenging implant process required by today’s semiconductor fabrication schedules.

N-type USJ

The N-type shallow junctions are easier to form, due mainly to the higher mass of the arsenic atom (75amu) compared to the boron atom.  The high mass results in a much shallower implant profile for equivalent implant energy.  However, the NMOS transistor is generally much more aggressively designed since it is higher performance for the same geometry.  Thus, technology will soon scale to where the implantation process for N-type USJ becomes an issue.