DENSITY FUNCTIONAL THEORY FOR STUDIES OF MULTIPLE STATES OF INHOMOGENEOUS FLUIDS AT SOLID SURFACES AND IN PORES.

Similar documents
Sorption and Phase Transitions in Nanopores

Gauge cell method for simulation studies of phase transitions in confined systems

Statistical geometry of cavities in a metastable confined fluid

Bridging scales from molecular simulations to classical thermodynamics: density functional theory of capillary condensation in nanopores

Peter A. Monson. Department of Chemical Engineering, University of Massachusetts, Amherst, MA 01003, USA. Acknowledgements: John Edison (Utrecht)

STATE-OF-THE-ART ZEOLITE CHARACTERIZATION: ARGON ADSORPTION AT 87.3 K AND NONLOCAL DENSITY FUNCTIONAL THEORY (NLDFT)

Phase Transitions and Criticality in Small Systems: Vapor-Liquid Transition in Nanoscale Spherical Cavities

A simulation method for the calculation of chemical potentials in small, inhomogeneous, and dense systems

Characterization of nanoporous materials from adsorption and desorption isotherms

APPLICATION OF A NOVEL DENSITY FUNCTIONAL THEORY TO THE PORE SIZE ANALYSIS OF MICRO/MESOPOROUS CARBONS. Abstract. Introduction

Calculation of Pore Size Distributions in Low-k Films

Modeling Dynamics (and Thermodynamics) of Fluids Confined in Mesoporous Materials. Escuela Giorgio Zgrablich,

Adsorption of Lennard-Jones Fluids in Carbon Slit Pores of a Finite Length. AComputer Simulation Study

Vapor-to-Droplet Transition in a Lennard-Jones Fluid: Simulation Study of Nucleation Barriers Using the Ghost Field Method

of Nebraska - Lincoln

Accepted Manuscript. On the Microscopic Origin of the Hysteresis Loop in Closed End Pores - Adsorbate. Poomiwat Phadungbut, D.D. Do, D.

Evaluation of Pore Structure Parameters of MCM-41 Catalyst Supports and Catalysts by Means of Nitrogen and Argon Adsorption

Adsorption properties of a colloid-polymer mixture confined in a slit pore

Studies of a lattice model of water confined in a slit pore

3.10. Capillary Condensation and Adsorption Hysteresis

CARBON 2004 Providence, Rhode Island. Adsorption of Flexible n-butane and n-hexane on Graphitized Thermal Carbon Black and in Slit Pores

Binary Hard-Sphere Mixtures Within Spherical Pores

CE 530 Molecular Simulation

PORE SIZE DISTRIBUTION OF CARBON WITH DIFFERENT PROBE MOLECULES

The birth of a bubble: A molecular simulation study

Statistical Physics. Solutions Sheet 11.

Interpretation of Full Sorption-Desorption Isotherms as a Tool for Understanding Concrete Pore Structure

Density Functional Theory for the Electron Gas and for Jellium. Abstract

Specifics of freezing of Lennard-Jones fluid confined to molecularly thin layers

Entropic wetting and the fluid fluid interface of a model colloid polymer mixture

arxiv: v1 [cond-mat.soft] 31 Oct 2007

Principles of Equilibrium Statistical Mechanics

30000, Thailand Published online: 26 Sep 2014.

Line adsorption in a mean-field density functional model

Specific Surface Area and Porosity Measurements of Aluminosilicate Adsorbents

Adsorption Processes. Ali Ahmadpour Chemical Eng. Dept. Ferdowsi University of Mashhad

Module 5: "Adsoption" Lecture 25: The Lecture Contains: Definition. Applications. How does Adsorption occur? Physisorption Chemisorption.

Characterization of nanopores by standard enthalpy and entropy of adsorption of probe molecules

An Extended van der Waals Equation of State Based on Molecular Dynamics Simulation

ADSORPTION IN MICROPOROUS MATERIALS: ANALYTICAL EQUATIONS FOR TYPE I ISOTHERMS AT HIGH PRESSURE

THERMODYNAMICS THERMOSTATISTICS AND AN INTRODUCTION TO SECOND EDITION. University of Pennsylvania

MOLECULAR MODELING OF EQUILIBRIUM OF SIMPLE FLUIDS IN CARBONS: THE SLIT PORE MODEL REVISITED

(# = %(& )(* +,(- Closed system, well-defined energy (or e.g. E± E/2): Microcanonical ensemble

Statistical physics. Emmanuel Trizac LPTMS / University Paris-Sud

Introduction Statistical Thermodynamics. Monday, January 6, 14

Author's personal copy

CHARACTERIZATION OF FLUID-ROCK INTERACTION BY ADSORPTION CALORIMETRY

2. Thermodynamics. Introduction. Understanding Molecular Simulation

Crystal nucleation for a model of globular proteins

A Study of the Thermal Properties of a One. Dimensional Lennard-Jones System

Adsorption in noninterconnected pores open at one or at both ends: A reconsideration of the origin of the hysteresis phenomenon

Capillary Contact Angle in a Completely Wet Groove

Gibbs ensemble simulation of phase equilibrium in the hard core two-yukawa fluid model for the Lennard-Jones fluid

Contents. 1 Introduction and guide for this text 1. 2 Equilibrium and entropy 6. 3 Energy and how the microscopic world works 21

REPORT DOCUMENTATION PAGE

Phenomenological Theories of Nucleation

Thermodynamics of Three-phase Equilibrium in Lennard Jones System with a Simplified Equation of State

(i) T, p, N Gibbs free energy G (ii) T, p, µ no thermodynamic potential, since T, p, µ are not independent of each other (iii) S, p, N Enthalpy H

(Crystal) Nucleation: The language

Density Functional Theory Study on the Structure and Capillary Phase Transition of a Polymer Melt in a Slitlike Pore: Effect of Attraction

Density functional theory for inhomogeneous mixtures of polymeric fluids

Phase Transitions in Condensed Matter Spontaneous Symmetry Breaking and Universality. Hans-Henning Klauss. Institut für Festkörperphysik TU Dresden

Chemical Potential, Helmholtz Free Energy and Entropy of Argon with Kinetic Monte Carlo Simulation

Phase Equilibria and Molecular Solutions Jan G. Korvink and Evgenii Rudnyi IMTEK Albert Ludwig University Freiburg, Germany

Thermodynamic integration

Aspects of Physical Adsorption Characterization of Nanoporous Materials. Dr. Matthias Thommes Quantachrome Instruments Boynton Beach, FL, USA

Liquid-gas phase transition in canonical and micro-canonical ensembles of strange hadronic matter

Received September 21, In Final Form: November 30, 2004

Van der Waals equation: event-by-event fluctuations, quantum statistics and nuclear matter

MSE 201A Thermodynamics and Phase Transformations Fall, 2008 Problem Set No. 8. Given Gibbs generals condition of equilibrium as derived in the notes,

Nanoscale simulation lectures Statistical Mechanics

Pore Size Heterogeneity and the Carbon Slit Pore: A Density Functional Theory Model

Physics and Chemistry of Interfaces

Weighted density functional theory of the solvophobic effect

Density Functional Theory Model of Adsorption on Amorphous and Microporous Silica Materials

The Second Virial Coefficient & van der Waals Equation

Supporting Information to

NOTICE CONCERNING COPYRIGHT RESTRICTIONS

Importance of pore length and geometry in the adsorption/desorption process: a molecular simulation study

G. P. Androutsopoulos* and C. E. Salmas

Linear Theory of Evolution to an Unstable State

Gas content evaluation in unconventional reservoir

Supplementary information for

V.E Mean Field Theory of Condensation

Accurate Determination of Pore Size Distributions

Density functional theory for molecular orientation of hard rod fluids in hard slits

High H2 Adsorption by Coordination Framework Materials

Phases of matter and phase diagrams

Simulation of Alkane Adsorption in the Aluminophosphate Molecular Sieve AlPO 4-5

CHARCOAL SORPTION STUDIES I. THE PORE DISTRIBUTION IN ACTIVATED CHARCOALS1

MOLECULAR DYNAMICS SIMULATION ON TOTAL THERMAL RESISTANCE OF NANO TRIANGULAR PIPE

LIQUID VAPOR ISOTHERMS IN NANO-POROUS MEDIA UNDER NMR OBSERVATION

Molecular Driving Forces

BEHAVIOR OF THE CONFINED HARD-SPHERE FLUID WITHIN NANOSLITS: A FUNDAMENTAL-MEASURE DENSITY-FUNCTIONAL THEORY STUDY

V.C The Second Virial Coefficient & van der Waals Equation

On the experimental values of the water surface tension used in some textbooks

Nanoporous Carbons: Porous Characterization and Electrical Performance in Electrochemical Double Layer Capacitors

Supplemental Material for Temperature-sensitive colloidal phase behavior induced by critical Casimir forces

arxiv: v2 [nucl-th] 4 Feb 2015

Chapter 6. Phase transitions. 6.1 Concept of phase

Transcription:

J. Smith, D. Stroud, MRS Symposium Proceedings Series, v.49, p.7-33, 998. DENSITY FUNCTIONAL THEORY FOR STUDIES OF MULTIPLE STATES OF INHOMOGENEOUS FLUIDS AT SOLID SURFACES AND IN PORES. A.. NEIMARK, and P.I. RAIKOITCH TRI/Princeton, 6 Prospect Av., Princeton, NJ, 854-65, aneimark@triprinceton.org Chemical Engineering/Yale University, New Haven, CT, 65-886 ABSTRACT Two new versions of the density functional theory (DFT), which correspond to the canonical ensemble and the Gibbs ensemble respectively, are applied to study multiple equilibrium states and associated hysteresis on examples of capillary condensation of argon in nanopores of MCM-4 mesoporous molecular sieves and Kr adsorption on carbon. The canonical ensemble DFT (CEDFT) allows us to trace not only the stable and the metastable states along the hysteresis loop, but also the unstable states inside the hysteresis loop. The Gibbs ensemble DFT (GEDFT) allows us to determine precisely the points of equilibrium phase transitions in confined fluids. INTRODUCTION Multiple states of confined inhomogeneous fluids are well documented in surface science. They manifest themselves in step-wise phase transitions and hysteresis behavior, which are inherent to thin films between colloidal particles, soap films, wetting and adsorption films on solid surfaces, fluids in pores, etc. Some of the prominent examples are the capillary condensation phenomena in nanopores and the adsorption-desorption transitions in mono- and multilayers. Better understanding of the specifics of confined fluid transformations is needed in a variety of practical applications such as stability of colloids, fabrication of nanostructures, separations, lubrication, fluid flow and sorption in porous media. The density functional theory (DFT) has been used by many authors for analyses of inhomogeneous fluids at solid surfaces and in pores (for a comprehensive review, see []). In particular in our earlier papers [-5], we employed the non-local density approximation due to Tarazona [6] to model hysteresis phenomena such as capillary condensation/evaporation in MCM-4 mesoporous molecular sieves and phase transitions in adsorption multilayers on carbon. Conventional versions of the DFT imply minimization of the grand thermodynamic potential with respect to the fluid density within fixed solid boundaries at the given temperature T and chemical potential μ. They correspond to the grand canonical (μ--t) ensemble and are below referred to as GCE DFT. In this paper, we apply two new versions of the DFT, which correspond to the canonical ensemble (CE DFT) and the Gibbs ensemble (GE DFT) respectively. We study capillary condensation of Ar in nanopores of MCM-4 mesoporous molecular sieves and Kr adsorption on carbon. The GCE DFT studies of these systems have been reported in the previous MRS meeting [5]. It was shown that GCE DFT qualitatively describes these phenomena and is in a satisfactory agreement with experimental adsorption/desorption. CE DFT allowed us to trace not only the stable and the metastable states along the hysteresis loop, which can be done with GCE DFT, but also the unstable states inside the hysteresis loop. The obtained isotherms are alike the van-der-waals S-shaped isotherms. Thermodynamic foundations for this type of isotherms have been discussed in the literature

J. Smith, D. Stroud, MRS Symposium Proceedings Series, v.49, p.7-33, 998. earlier in mid-8s [7]. However, they were constructed for the first time only recently [8] by means of the canonical ensemble and the Gibbs ensemble MC simulations. The proposed DFT methods provide a better understanding of hysteresis phenomena in nanoconfinements. In systems with several consecutive phase transitions, the CE DFT isotherms exhibit multiple zigzags. The backward regions of the isotherms correspond to the unstable states inside the hysteresis loop(s). These unstable states cannot be realized in a physical or a numerical experiment unless density fluctuations, which drive the system to one of the metastable or stable states, are completely suppressed. In the developed DFT, the density fluctuations are suppressed by fixing the total amount of fluid molecules in the confinement. CANONICAL ENSEMBLE DFT FOR INHOMOGENEOUS FLUIDS Consider a closed system consisting of N molecules of fluid confined in a volume surrounded by solid walls. The system is embedded in a bath of constant temperature T. The fluid is modeled as a hard sphere fluid with a given pairwise attractive potential, Φ ff ( r,r ). The fluid-solid interactions are modeled by a given spatially varying external potential, Φ sf. Equilibrium states of the closed system are defined by minimization of the Helmholtz free energy functional at constant N,, and T. The DFT implies that the intrinsic Helmholtz free energy, F int, depends entirely upon given fluid-fluid interaction potential(s), Φ ff ( r,r ), and is represented as a functional of the spatially varying fluid density ρ. The external solid-fluid interaction potential, Φ sf () r, does not affect the expression for intrinsic Helmholtz free energy functional, Fint[ ρ () r, Φ ff ], and makes an additive contribution, F ext [ ρ,φ sf ], into the total Helmholtz free energy, F: F ρ r = F ρ r, Φ + ρ r, Φ ( ) [ ( )] ( ) int [ ] [ ( ) ] ff F ext Equilibrium states in the canonical ensemble are obtained by minimization of the Helmholtz free energy functional () with respect to the fluid density ρ, provided the total number of molecules is a given constant N: ρ (r) d r = N () This is a conditional extremum problem. To find a solution, one has to solve the Euler equation for the functional F ρ r μ ρ dr (3) Ω[ ρ () r ] = [ ( )] Here, an additional unknown parameter, μ, is introduced. The Euler equation for the functional (3) reads δf μ = (4) δρ δ Here, denotes the functional derivative. The right hand side of the Eq. (4) depends on an δρ employed representation of the intrinsic Helmholtz free energy functional, F int [ ρ () r ]. Equations (4) and () represent a closed system for determination of two unknowns, a function ( ) r ρ and a constant μ. This system implies a unique solution at given N. The solution gives the equilibrium sf

J. Smith, D. Stroud, MRS Symposium Proceedings Series, v.49, p.7-33, 998. density profile ρ ( r, N,, T ) and the value of μ(n,,t), which turns out to be equal to the chemical potential of the corresponding N,,T state. Therewith, the functional Ω[ ρ ( r, N,, T ), μ(n,,t)] is equal to the grand potential. While the obtained chemical potential, μ, is a singlevalued function of N, there is no one-to-one correspondence. On the contrary, the function N(μ) can be non-monotone and exhibit S-shaped swings in the regions of phase transitions. The points of equilibrium transitions, μ e, can be defined from the Maxwell s rule of equal areas by integrating along the S-shaped isotherm between the coexisting states () and (): N ( μe ) N ( μe ) GIBBS ENSEMBLE DFT FOR EQUILIBRIUM TRANSITIONS μ dn = μ ( N N ) (5) The Gibbs ensemble [9], is the canonical (N--T) ensemble for a two-phase system. Unrestricted exchange of volume and mass between the phases provides mechanical and chemical equilibrium. Consider N Σ molecules of a fluid, which may be distributed between similar pores. Suppose the pores may be in two different states, which are characterized by distinct fluid density profiles, ρ and ρ. Let and, and N and N be the volumes of pores in the states and and the numbers of molecules in these pores respectively. The total volume of pores, Σ = +, and the total number of molecules, N Σ = N + N, are fixed. The phase volumes and the numbers of molecules in phases are variable. The problem is to find an equilibrium phase distribution at given Σ and N Σ, and also the equilibrium conditions in terms of chemical potential μ. The equilibrium states in the GE are found by minimization of the total Helmholtz free energy of the two-phase system, F ( ρ ) = αf ρ ( ) e [ ] + ( α ) F[ ] Σ r ρ min, (6) at the condition ( α ) ρ ( r dr ρ α ρ ( r ) dr + ) =. (7) Here, we introduced the unknown portion of pores in the state, α = / Σ, and the fixed average density <ρ>= N Σ / Σ. The solution of this problem gives the density profiles ρ and ρ in the equilibrium states and the chemical potential μ of equilibrium transition. Therewith, ρ and ρ fulfill the equation (6), and μ is determined form the condition of equality of the grand potentials in the equilibrium states, ( ρ ()) F[ ρ ]- μ ρ dr = ( ρ ( )) F[ ] G r = G = - μ ρ dr. (8) r ρ In order to demonstrate capabilities of the proposed CE DFT and GE DFT methods, we present below several prominent examples. Detail description of the techniques will be given elsewhere.

J. Smith, D. Stroud, MRS Symposium Proceedings Series, v.49, p.7-33, 998. EXAMPLES. In Figs. -4, we present two prominent examples of multilayer adsorption and capillary condensation hysteresis: ) Ar adsorption on MCM-4 mesoporous molecular sieve; calculations were performed at Ar boiling temperature 87 K (T * =.739) in a cylindrical pore of 4. σ ff (internal diameter 4.4 nm), and ) Kr adsorption on carbon; calculations were performed at O boiling temperature 9 K (T * =.547) in a slit-shaped pore of 4 σ ff (internal diameter 4.7 nm). We used the Tarasona s representation of the Helmholtz free energy [6] with the parameters for fluid-fluid and solid-fluid interaction potentials, which were employed in our previous paper [5]. We employed the Broyden's method for solution of nonlinear set of equations []. On these examples, we observe formation of consecutive adsorption layers prior to spontaneous (spinodal) capillary condensation. Equilibrium capillary condensation transition links a liquid-like state on the desorption branch with a two-layer adsorption film on the adsorption branch. Multilayer adsorption is also associated with spinodal transitions (most prominent for Kr adsorption) and equilibrium coexistence of distinct states. We observed up to five states at the same chemical potential. CONCLUSIONS We studied multiple equilibrium states of inhomogeneous confined fluids and associated hysteresis phenomena by means of new versions of the density functional theory, which correspond to the canonical ensemble and the Gibbs ensemble respectively. On the examples of capillary condensation of Ar in nanopores of MCM-4 mesoporous molecular sieves and Kr adsorption on carbon, it was shown that the proposed CE DFT and GE DFT methods provide a better understanding of hysteresis phenomena in nanoconfinements. They allow us to generate continuous adsorption isotherms, including unstable states inside the hysteresis loop, and to determine precisely the points of equilibrium and spinodal phase transitions in confined fluids. ACKNOWLEDGEMENT. Work has been supported by the TRI exploratory research program. REFERENCES. R. Evans, in Fundamentals of Inhomogeneous Fluids; Henderson, D., Ed.; Marcel Dekker: New York, 99.. A.. Neimark, Langmuir,, 483 (995). 3. P.I. Ravikovitch, A.. Neimark, S.C.Ó Domhnaill, F. Schüth, K.K. Unger, Langmuir,, 4765 (995). 4. A.. Neimark, P.I. Ravikovitch, Langmuir, 3, 548 (997). 5. A.. Neimark, P.I. Ravikovitch, in Dynamics in Small Confining Systems III, Eds. J.M. Drake, J. Klafter, R. Kopelman, MRS Symposium Proceedings olume 464, 997, p.65. 6. P. Tarazona, Phys.Rev.A, 3, 67 (985). 7. R. Evans, U. Marini Bettolo Marconi, and P. Tarazona, J. Chem. Phys., 84, 376 (985). 8.. Yu. Gusev, A.. Neimark, Paper L7, TRI/Princeton Workshop Characterization of Porous Materials: from Angstroms to Millimeters, June 8-, 997, Princeton, NJ;.Yu. Gusev, A.. Neimark, Paper S.6, 997 MRS Fall Meeting, Dec. -5, 997, Boston, MA. 9. A.Z. Panagiotopoulos, Molecular Physics, 6, 83 (987); 6, 7 (987).. W.H. Press, B.P. Flannery, S.A. Teukolsky, W.T. etterling. Numerical Recipes in C: The Art of Scientific Computing, Second Edition, Cambridge University Press: Cambridge, 99.

J. Smith, D. Stroud, MRS Symposium Proceedings Series, v.49, p.7-33, 998..8 ρ, σff -3.7.6.5.4.3.. Canonical DFT Grand Canonical DFT - - -3-4 Ω, kt σff - - -.5 - -.5 μ μ, kt -5 Fig.. Ar in 4. σ ff (4.4 nm) cylindrical pore at 87.3 K (T * =.739). In the stable and metastable states, the GCE DFT isotherm (line) coincides with the CE DFT isotherm (points). The GE DFT gives the vertical line of equilibrium capillary condensation transition. The point of equilibrium transition, Δμ=-.835 kt, corresponds to the intersection point of the Grand Potential (zigzag thin line). The leftmost and rightmost vertical lines, which bound the hysteresis loop, correspond to the spinodal transitions in the GCE DFT. 7 ρ, σ ff -3 6 5 4 3 -layers at P/Po=.46 unstable state at P/Po=.46 3-layers at P/Po=.46 unstable state at P/Po=.46 'liquid' at P/Po=.46 4 6 r, σ ff Fig.. Density profiles of Ar in 4. σ ff (4.4 nm) cylindrical pore at 87.3 K (T * =.739). The five profiles correspond to multiple states at P/Po=.46 (Δμ=-.76 kt). Three equilibrium states: the absolutely stable liquid-like state after the capillary condensation (circles); the metastable coexisting two-layer adsorption film and the three-layer adsorption film (lines). Two unstable states: the two-and-a-half-layer (dashed line) and the four-and-a half-layer (crosses).

J. Smith, D. Stroud, MRS Symposium Proceedings Series, v.49, p.7-33, 998..9.8.7.6 ρ, σ ff -3.5.4.3.. Canonical DFT Grand Canonical DFT Equilibrium transitions..4.6.8 P/P Fig. 3. Isotherm of Kr in a 4 σ ff (4.7 nm) slit-shaped pore of carbon at 9 K (T * =.547). In stable and metastable states the GCE DFT isotherm (line) coincides with the CE DFT isotherm (points). The GE DFT gives the vertical lines of equilibrium transitions between the coexisting phases. We observe four consecutive layering transitions and the capillary condensation transition. 8 7 6 5 layers, P/Po=.54 unstable state, P/Po=.54 3 layers, P/Po=.54 unstable state, P/Po=.54 liquid-like phase, P/Po=.54 liquid spinodal, P/Po=.4 ρ, σ ff -3 4 3 3 4 5 6 7 z, σ ff Fig. 4 Density profiles of krypton in 4 σ ff (4.7 nm) slit-shaped pore at 9 K (T * =.547). Five states correspond to the relative pressure of P/Po=.54. Three equilibrium states: the absolutely stable liquid-like state after the capillary condensation (circles); the metastable coexisting twolayer adsorption film (squares) and the three-layer adsorption film (triangles). Two intermediate unstable states (lines): the two-and-a-half-layer state and the six-layer state. A density profile near the liquid spinodal point, P/Po=.4, (dotted line) is presented for comparison.