This study concerns the design of protein-resistant polymer adsorbed layers for the control of surface binding of biospecific recognition entities. Polymer surface layers were prepared using the adsorption of poly(allylamine hydrochloride) (PAH), poly(L-lysine) (PL), and branched and linear polyethyleneimine (PEI) and further modified by the covalent attachment of biotin for specific avidin attachment. The adsorption of PAH, PL, and PEI on silicon substrates was studied as a function of pH and ionic strength using ellipsometry. Average dry layer thicknesses of similar to 10, similar to 5, similar to 9, and similar to 3 angstrom (+/- 1 angstrom) were obtained when polymer adsorption occurred from solutions at pH 9.5 that contained 0.5 M NaCl for PAH, PL, branched PEI, and linear PEI, respectively. These polymers showed significant differences in their efficiency to suppress nonspecific avidin adsorption. At low ionic strength, avidin adsorption occurred on all polymer-coated surfaces at basic pH values, despite the same positive electrostatic charge for protein globules and the surface. Though the net electrostatic repulsion between avidin molecules and branched PEI was efficiently screened in a protein solution of pH 7 and 0.15 M NaCl, branched-PEI coatings of high molecular weight were unique in their ability to provide avidin-resistant surfaces as a result of steric hindrance from the branched architecture of adsorbed polymer chains. All polymers studied were effective in suppressing avidin adsorption at pH 3 as a result of protonation of the avidin surface functional groups at this pH. Branched-PEI-coated surfaces were also effective for the suppression of smaller positively charged proteins such as lysozyme and ribonuclease A at pH 7 and 0.15 M NaCl. They were also resistant to the adsorption of negatively charged proteins such as BSA and fibrinogen at pH 7 and 0.75 M NaCl. Furthermore, by using PEI-modified protein-repellent surfaces, selective binding of avidin was achieved to surface-bound silver nanoparticles, which should provide a promising application for the label-free detection of biological species using surface-enhanced Raman scattering (SERS).