The adsorption of different proteins in a single biospecific and hydrophobic adsorbent particle for preparative protein chromatography has been observed directly by confocal laser scanning microscopy as a function of time at a constant bulk concentration c(b). The bulk concentration was in the non-linear part of the adsorption isotherm. At all times the concentration of free protein at the particle surface was almost equal to the bulk content indicating that external mass transfer resistance is not rate limiting for the adsorption under these conditions. Inside the particles a distinct maximum in adsorbed and free protein concentration that moved inside to a distance of approximately 0.2 R (R particle radius) from the particle surface, was observed. This is due to a decreasing solid-phase density and adsorptive capacity in the particle between 0.8 R and R indicating that the fraction of macropores (or void space) is larger in the outer than in the inner part of the adsorbent particles. By increasing the bulk concentration by a factor of 10 the equilibration time was reduced by about the same magnitude. This is in agreement with the concentration dependence of the effective pore diffusion coefficient D(p,eff)=D(p)/[epsilon(p)[1+nK/(K +c)(2)]] derived from the mass conservation relations describing the adsorption process. The time dependence protein adsorption up to approximately 90% of the equilibration value q* could be described by a bilinear free driving force model. The rapid equilibration in the outer part of the particle with a half-life time of approximately 100 s in the studied systems accounted for 0.3-0.4 q*. The slower equilibration with a up to ten times longer half-life time, was the adsorption in the inner part of the particle that outside 0.5 R accounts for 0.5-0.6 q*. These data were compared with literature data for batch adsorption of proteins in biospecific, hydrophobic and ion-exchange adsorbents. They could also be described by a bilinear free driving force model, with about the same quantitative results as obtained for similar conditions in the single particle experiments. The static adsorption parameters, maximum binding site concentration n, and dissociation constant for the protein binding to a binding site K, were determined from Scatchard plots. For the same protein-adsorbent system the plots changed from linear to non-linear with increasing n. This change occurred when the average distance between adjacent binding sites become of the same order of magnitude as the size of the binding site or adsorbed protein. This causes a shielding of free binding sites increasing with n and the concentration of adsorbed protein, yielding a concentration dependence in K. These results show that for a high throughput and rapid adsorption in preparative chromatography, the adsorption step should be carried out in the non-linear part of the adsorption isotherm with concentrations up to c(b) where q*/c(b)>/=10 to obtain high protein recoveries. To avoid tailing due to the flow of adsorbed proteins in the inner part of the particles further into the particles at the start of the desorption, and to speed up desorption rates, protein adsorption in the particle within 0.5 R from the particle center should be avoided. This requires the further development of suitable pellicular particles for preparative protein chromatography that meet this requirement.