"What I am trying to determine is why Elmer feels AMD would have more of a particulate contamination problem than Intel."
The whole particle contamination thing came about because Elmer ran across an article that claims the equipment sheds a similar number of particles for a 200mm wafer as a 300mm wafer. If true, and assuming the equipment is the source of most of the particles, then that would give a big advantage to 300mm as far as defect density. I am not sure this is totally valid, I suspect that the particles shed is a function of time, and for many steps a 300mm wafer will take longer to process than a 200mm one. But...
Any way, assuming that AMD does have a lower yield than Intel, it is more likely a function of other factors. For one, they may be more aggressive on design rules. Having more metal layers would be a DD problem, since there are more steps, there are more possibilities for particle defects. In the past, Intel tuned their processes more for yield than performance, AMD often did the reverse. This also can impact yield, what would normally be the slower chips of the batch just don't work because they have transistors than can't be turned on instead of just working slower. Or AMD's yield could be fine, but to get the bin splits they want, they trash can the slower ones instead of selling them. These problems are quite as easy to correct. |
