Working without accumulation membrane in flow field-flow fractionation.

Abstract

Nonideal interaction of sample with the separation device is a difficulty found in chromatographic methods as well as in field-flow fractionation. However, in field-flow fractionation (FFF), greater flexibility in the choice of carrier solution composition is possible, thus reducing the need of a wide choice of surface chemistry when nonideal sample interaction is to be minimized. The use of an ultrafiltration membrane as the surface for the accumulation wall is common practice in flow field-flow fractionation. Typical membranes in use are laminates of a skin membrane onto a backing material such as woven polyester. At this point, only a limited choice of membrane chemistries is available. Many membranes have been developed for protein applications as membranes are widely used in the pharmaceutical industries. While these membranes work well for protein applications, flow field-flow fractionation is applicable to polymeric particulate as well as protein samples. Thus, sample interaction with the membrane surface is possible with nonprotein applications and these interactions can induce significant secondary effects on retention ratio and affect instrumental reliability. Also, the woven texture of membranes may detrimentally affect the FFF separation. For these reasons, the study of flow field-flow fractionation using a flat, smooth surface of controlled chemistry is of relevance. We present here the results of a new, membraneless channel that uses a bare frit as the accumulation wall and that is intended for analysis of micrometer-sized particles only. Selectivity results are comparable to those obtained with the membrane, while relative sample recovery indicates that the best quantitative performance can be obtained without the membrane. Moreover, neither sample immobilization nor losses through the frit occur when operating membraneless. On the other hand, first experimental evidence of a certain level of frit surface activity suggests that optimization of experimental conditions is required.