Two-step diffusion: It represents practical representation of pre-deposition and drive-in processes. It is most common method used for silicon devices and micro circuits. In this silicon dioxide (mask) is grown simultaneously with drive in phase to provide immediate protection for silicon surface.

The process is initiated by first conducting constant source diffusion at low temperature for short time.

The impurity supply is shut off and the drive in phase initiated in an oxidizing ambient. This results in formation of a surface oxide of silicon preventing further diffusion of impurities.

This process is preferred since it avoids additional masking step and also minimizes the time for which slice is maintained at elevated temperature, reducing undesired moment of already diffused impurity.

Two step processes can be used to approximate an erfc (error function) diffusion profile without the accumulation of large amount of impurities on surface of silicon. Thus it avoids the possibility of surface damage. Dry oxidation is used for this purpose.

It can be also used to approximate a Gaussian profile while avoiding the necessity of exposing the surface of slice at any time. Wet oxidation is used for this purpose.

Successive diffusion:

It is required to calculate total effect of diffusion during series of temperature cycles.

For example, in BJT emitter diffusion takes place after base drive-in. Thus impurities in base are subjected to one set of temperature and time values, during base drive-in process and to a second set during emitter diffusion phase.

To compute total effort of these cycles, it is necessary to obtain Dt product

$(Dt)_{eff}=\sum D_1 t_1 +D_2 t_2+D_3 t_3+ ..…........$

Where $t_1,t_2,t_3$ are different diffusion times and $D_1,D_2,D_3$ are the appropriate diffusivities in effect during these times.