Adsorption of fluids in confined disordered media from inhomogeneous replica Ornstein-Zernike equations Orest Pizio 1 and Stefan Sokolowski 2 1 Instituto de Quı ´mica de la Universidad Nacional Auto ´noma de Me ´xico, Circuito Exterior, Coyoaca ´n 04510, Distrito Federal, Mexico 2 Department for the Modelling of Physico-Chemical Processes, Faculty of Chemistry, Marie Sklodowska Curie University, 200-31 Lublin, Poland Received 12 February 1997 The density distribution and pair distribution function for a fluid adsorbed in a slitlike pore filled with a quenched hard-sphere fluid have been investigated using the inhomogeneous replica Ornstein-Zernike equa- tions. The one-particle and two-particle functions are related via Born-Green-Yvon equation and the inhomo- geneous Percus-Yevick approximation is used. The effect of the matrix is to lower the amount of adsorbed fluid in the entire pore at low chemical potentials. For high values of the fluid chemical potential, layering of adsorbed fluid is observed. The solution of the problem is important for gel-exclusion chromatography and in separation science. S1063-651X9752007-1 PACS numbers: 68.45.-v, 61.20.Gy The problem of describing the adsorption of fluids in ran- dom porous media has received much interest recently. Im- portant experimental observations 1–4, and relevant theo- retical developments 5–15have been concerned with homgeneous systems. Adsorption of fluids in confined inho- mogeneous disordered media, however, is of much interest in practice for gel-exclusion chromatography and in separation science. This problem is qualitatively more difficult to deal with than the homogeneous case. To our best knowledge, there was only one attempt to consider inhomogeneous fluids adsorbed in disordered porous media 14. The inhomoge- neous replica Ornstein-Zernike IROZequations, supple- mented by either Born-Green-Yvon BGY, or Lovett-Mou- Buff-Wertheim LMBWequation for density profiles DPs, have been proposed to study adsorption of a fluid near a plane boundary of a disordered matrix. The theory has not been followed by any numerical solution. In this paper our main goal is to propose a simple model for adsorption of fluids in confined disordered porous media and to solve it. Let us consider the adsorption of a fluid consisting of particles m , in a slitlike pore of the width H . The pore walls are normal to the z axis and the pore is centered at z =0. The fluid m , i.e., the matrix, in the pore is in an equlibrium with the bulk fluid with the chemical potential m . The fluid is then characterized by the DP m ( z ), and by the inhomoge- neous pair correlation function h mm (1,2). Due to external factors, the structure of the fluid becomes quenched at a state determined by m , and a confined porous medium is formed. Now, we would like to investigate adsorption of another fluid, f , in the pore filled by the matrix. The fluid f , in the bulk has the chemical potential f ; at equilibrium the ad- sorbed fluid f reaches the DP f ( z ) and is characterized by the inhomogeneous correlation function h ff (1,2). The matrix and fluid species are denoted by the subscripts 0 and 1 5. We assume the interactions between particles and between particles and pore walls in a simple form choosing both spe- cies as hard spheres of unit diameter: U ij r = , r 1 0, r 1 ; Uz = , z 0.5| H -1 | 0, otherwise , 1 where i , j =0,1. The evaluation of the matrix structure is irrelevant to the procedure below. The structure of a quenched inhomoge- neous matrix is obtained from the inhomogeneous Ornstein- Zernike OZ2equation 14,16 h 00 1,2-c 00 1,2= d 3 0 z 3 c 00 1,3h 0 3,2, 2 supplemented by the LMBW equation for the DP, ln 1 z 1 z 1 + Uz 1 z 1 = d 2 c 00 1,2  0 z 2 z 2 , 3 and the second order Percus-Yevick PY2closure, y 00 1,2=1 +h 00 1,2-c 00 1,2. 4 In Eq. 4, y 00 (1,2) is the inhomogeneous cavity distribution function CDF. The solution of Eqs. 24yields 0 ( z ) and h 00 (1,2) such that the one-particle CDF y 0 ( z ), y 0 ( z ) = 0 ( z )exp U(z), outside the pore tends to its limiting value, determined by the configurational chemical potential, y 0 ( z ) =exp( 0 ). In fact, the structure of a matrix may be intentionally prepared experimentally given some functions 0 ( z ) and h 00 (1,2). The IROZ equations, representing the essence of the pro- cedure, are 14 h 10 1,2-c 10 1,2= d 3 0 z 3 c 10 1,3h 00 3,2 + d 3 1 z 3 c c ,11 1,3h 10 3,2, 5 RAPID COMMUNICATIONS PHYSICAL REVIEW E JULY 1997 VOLUME 56, NUMBER 1 56 1063-651X/97/561/634/$10.00 R63 © 1997 The American Physical Society