2.5.1 Analytical Ion Implantation - Monte Carlo Ion Implantation



next up previous contents
Next: 2.5.2 Knock-in Implantation - Up: 2.5 Applications Previous: 2.5 Applications

2.5.1 Analytical Ion Implantation - Monte Carlo Ion Implantation

Here, we will sketch results obtained by the Analytical Ion Implantation Module and compare them with Monte Carlo simulations obtained by the Monte Carlo Implant Module of PROMIS. Using a test structure (Figure 2.5-1) we will elaborate three important factors to be taken into account for accurate treatment of two-dimensional ion profiles - tapered mask layers, lateral moments vary with depth and the effect of the beam tilt angle.

 

 

Masking layers are often tapered at the edge rather than perfectly abrupt, so that ions are gradually prevented from entering the silicon. Note that the ions penetrate through the thinner region of the sloping mask and increase the doping near the corner (Figure 2.5-1 ). The concentration profile in Figure 2.5-1 was calculated using constant lateral moments and, therefore, the maximum lateral extent of the profile under the nitride mask is at the depth of the maximum concentration of the vertical profile (Figure 2.5-1 ).

Using the more sophisticated model including depth dependent lateral moments (top in Figure 2.5-2) the lateral shape of the profile is predicted more accurately (Figure 2.5-2 ) and the lateral extent underneath the tapered mask is not overestimated (Figure 2.5-2 ).

The agreement of these results with Monte Carlo simulations is quite good. One detail is not (and cannot be) reproduced by the analytical model. Ions scattered into free space travel without scattering until they reenter the target, e.g. nitride silicon and vice versa (Figure 2.5-2 ). For the Monte Carlo simulation we used distinct trajectories which gives a total number of simulated ions for this geometry applying Hobler's superposition method [Hob89]. Nevertheless, all these 16 million ion positions are constructed from only 40,000 different trajectories. Considering the peak concentration of one simulated trajectory end point gives a concentration of at least and, therefore, the contour line is statistically insignificant (Figure 2.5-2 ).

 

Finally, we take a closer look at the effect of a beam tilt angle. This tilt reduces the profile peak depth less than of but shifts the profile laterally by of the mask height.

In modern VLSI processes the layers above the substrate have not been scaled vertically with decreasing line width, on the contrary, with the development of multilayer metallization the tendency is towards thicker covering layers. Self aligned processes use the gate as an implantation mask, so a tall gate stack can cause a significant shadow at the edge of the source or drain. In our example we notice a lateral shift of the profile of Å away from the nitride mask (Figure 2.5-3 ). Again, the agreement of the analytical results with the Monte Carlo simulations is quite good, except for the known back-scattering effect (Figure 2.5-3 ).



next up previous contents
Next: 2.5.2 Knock-in Implantation - Up: 2.5 Applications Previous: 2.5 Applications



Martin Stiftinger
Wed Oct 19 13:03:34 MET 1994