We have characterized the e-beam proximity effect as it applies to the write pole break-point angle of magnetic recording heads. These narrow isolated negative resist lines have been measured using, an automated CD-SEM. The CD data allows us to quantify the e-beam proximity effect on silicon wafers with thin metallic films of varying thickness. Nickel and tantalum have atomic numbers of 28 and 73, respectively, and this difference is quantified by the increase in the CD of the Ta films compared to Ni. The CD was found to change at a rate of 0.17 nm per degree of break-point angle for the Ni films, and 0.25 nm per degree for Ta. We have analyzed the experimental data. by comparing it to two relevant models. First, we compare the data to the traditional expression used to describe e-beam exposure, a double Gaussian. From both the CD data and the double Gaussian, we calculate a proximity effect term we refer to as the dose fraction. This dose fraction has a linear relationship with the "eta" parameter, which also relates the contribution of forward to back scattered electrons in the final exposure profile. We determine both a dose fraction, and the "eta" parameter, for each substrate material and thickness. Second, we compare this dose fraction term to a simple Rutherford elastic scattering model. The final outcome of this work is a quantifiable measure of how the break-point angle contributes to the final CD of the write pole when employing e-beam lithography. This work also demonstrates a practical way to quantify the e-beam proximity effect by the calculation of dose fraction and "eta" as a metric of metal layer material and thickness. (C) 2004 American Vacuum Society.