It has been reported earlier that in the case of boron diffusion into silicon, the value of sheet resistance observed on plain check slices is almost always considerably smaller than the sheet resistance calculated from diffused resistors. The actual value of the discrepancy depends upon the size of the window, the surrounding masking oxide geometry, and the process parameters. These observations add another important factor in 2-D simulation for diffusion processes. Further investigations have therefore been made in an attempt to present a theoretical model to explain the experimental observations. The results of a careful series of experiments using multilayered mask structures consisting of thermal oxide, silicon nitride, and silox (SiO2 deposited using chemical vapor deposition process with silane) have been presented. On the basis of these experiments, it is suggested that most of the experimental observations may be explained with the help of a theoretical model based on "surface diffusion" of boron over silicon and silicon dioxide. The model assumes a high solubility of boron in oxide and a high surface diffusion coefficient. Numerical calculations demonstrate the capability of the model to explain most of the experimental observations.