Extensive studies of the Emitter Dip Effect (EDE) in npn silicon structures are reported. Typically, 1 ohm‐cm n‐type silicon starting material was used. Boron diffusion into this material resulted in a 0.7μ base depth and surface concentrations of . The emitter regions had surface concentrations of and 0.4μ junction depths. The dip was typically 0.3μ, i.e., the base penetrated 0.3μ deeper under the emitter than elsewhere. Important experimental findings are: (A) Crucible grown, zone refined, and epitaxial silicon give the same results. (B) The EDE occurs only if phosphorus doping is sufficient to overcompensate the base doping. (C) Multiple emitter diffusion causes multiple dips. (D) Slow cooling from diffusion temperature enhances the dip. (E) An enhanced boron diffusion constant in the base layer apparently causes the dip rather than the rejection of boron by the region of high phosphorus concentration. (F) The dip is not present for lightly doped (1017 cm−3) base layers.