Each size-exclusion chromatography (SEC) curve generally exhibits two peaks when a solvent mixture is employed as the eluent.The first peak is the solvated-polymer peak and the second is the freesolvent peak.The area of the solvated-polymer peak A eff3 can be regarded as the sum of contributions from the naked-polymer A3 and from the extra polymer-bound solvent A1.The free-solvent peak

however

is caused by deficiency of the same solvent that enriches itself around polymer chains.Theoretical analysis shows that the size of free-solvent peak A2 and that of the extra polymer-bound solvent A1 are equal in magnitude but opposite in sign.This relationship suggests that the preferential adsorption coefficient λV can also be determined from the nalysis of the polymer peak. Polystyrene(3)-benzene(1)-methanol(2) is the system taken to test the correctness of the theoretical analysis. The polystyrene (PS) sample was purchased from Krauer Company

with Mw=9.26×104

and Mw/Mn=1.1. Solvents (analytical grade purity) were purified before use.Benzene was purified by repeatedly shaking with concentrated sulphuric acid (15 vol% benzene) in a separating funnel until the acid phase remained colorless.The benzene phase was then washed with water

aqueous sodium hydroxide (10 wt%)

again with water

dried with anhydrous calcium chloride

and finally distilled. Methanol was fractionally distilled

and the middle cut was used.The eluents were benzene-methanol mixtures of different compositions at a flow rate of 0.4 cm3/min.The mixtures were made by volume and equilibrated in the reservoir of the GPC system.The injected sample solutions were prepared using solvent mixtures from the solvent reservoir.The range of volume fraction of benzene in the eluent U10 was from 1.00 to 0.80.All measurements were conducted at room temperature with a Waters Associates model 244DS liquid chromatograph

equipped with a model R401 differential refractometer and a model 730 microcomputer.A single column packed with Styragel (nominal porosity 50 nm

length 30 cm) was employed. When a solvent mixture is used as eluent

two measurements can be carried out

with and without column in place.When there is a column in place

the polymer-bound solvent will get separate from the free solvent.The selective binding of solvent to the polymer causes polymer-bound solvent signal and the polymer signal to merge.The measured peak area is Aeff3.In the experiment without column in place

there is a direct connection between the outlet of the injector and the detector. In this situation

the chromatographic system cannot separate the components.The free solvent

polymer-bound solvent

and the solute will elute at the same time.The composition deviations of polymer-bound solvent and free solvent to the original solvent mixture caused by preferential solvation can generate signals

but

since they elute at the same time

the contributions from them to the refractive index will cancel because of their opposite response signals.In this case the measured peak area is A3.When Aeff3 and A3 are available

the preferential adsorption coefficient λV can be obtained. For the system consisting polystyrene (3) plus benzene (1)-methanol (2) solvent mixtures

the values of λV from the present work are plotted against （1-u10），together with the results for the same system reported by Hert from measurements of dialysis equilibrium and light scattering technique.The agreement of the present results with those from previous work shows that the proposed method is valid.The advantages of studying the solvated polymer peak are discussed. This method provides an alternate way to determine specific refractive index increment at constant potential (dn/dc)μ for polymer/mixed solvent systems.