Optimizing a Microlectronic Device Response

Miguel,

A PB design is certainly the way to go for a screening DOE, but it depends on the number of main effects that you want to study.  If you have 6 or 7 factors that you want to study, then a 8-run PB is the way to go.  Remember a 8-run PB is a resolution III design with a complex alias structure.  To increase your design resolution (IV), it is a good idea to "fold" your design as a follow up experiment (another 8-run PB).  This foldover will clear up the picture as to how main effects are affected by your design (i.e separating main effects from 2 factor interactions).  If you have only have 3-5 factors to study, you may get a lot more value out of doing a full factorial or fractional factorial depending on how much it costs to setup and conduct your experiment.

I hope that helps a little.

Scott