Experimental and theoretical studies were carried out to investigate the performance of a batch, fluidized/expanded ion-exchange system. Exchange of phenylalanine on cation exchange resin was studied in the linear velocity range 5.5-39.8 cm/min. The breakthrough curves were obtained at various design and operating conditions with changes of column size, distributor, bed type, bed height, flow rate, and number of stages. The breakthrough data were analyzed by two theoretical models. A CSTR-type model which we called a lumped model fitted the data well when bed aspect ratio (bed height/bed diameter) was less than 1.5 and column diameter was less than 4 cm. We considered the fluidized bed as a bed that can be well simulated by the lumped model (uniform mixing), and the expanded bed as a bed where the lumped model fails because of reduced backmixing. Another model, which is adapted from a packed bed model (VERSE) with axial dispersion, film diffusion, and intraparticle diffusion, fitted the data for both fluidized beds and expanded beds successfully. Experimental and theoretical results showed the bed type changed from a fluidized bed to an expanded bed as the bed aspect ratio increases. The breakthrough curves for the expanded bed and a packed bed were almost identical except at very short times. When sieve plates were used as liquid distributors, the plate with more holes of a smaller size was most efficient. Mass transfer at different flow rates and the effects of bed type and multiple stages are discussed.