Physical Chemistry of Surfaces Sixth Edition ARTHUR W. ADAMSON Department of Chemistry, University of Southern California Los Angeles, California and ALICE P. GAST Department of Chemical Engineering, Stanford University Stanford, California ®) [A WILEY-INTERSCIENCE PUBLICATION John Wiley & Sons, Inc. NEW YORK / CHICHESTER / WEINHEIM / BRISBANE / SINGAPORE / TORONTOAbout the Front Cover Design ‘The fer-like shapes are two-dimensional dendritic crystals of the protein streptavidin bound to a lipid monolayer. These dendrites were produced by Andrew Ku with the help of Seth Darst, Roger Kornberg, and Channing Robertson and are described in Chapter XV; note Fig. XV-5. ‘The island pattern is from a marbled paper, for which we thank Mrs. Phoebe Easton. Marbling is an ancient process whereby colored oils with surfactant are spread on the surface of water stiffened by agar-agar and the pattern lifted onto the paper. The pattern could be that of domains ‘of 2 and of CO chemisorbed on PX(100); note Fig. XVILI-28. is book is printed on acid-free paper. © Copyright © 1997 by John Wiley & Sons, In. All rights reserved. Published simultaneously in Canada. [No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or ‘otherwise, except as permitted under Sections 107 or 108 of the 1976 United States Copyright ‘Act, without either the prior written permission of the Publisher, or authorization through ‘payment of the appropriate per-copy fee to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, (508) 750-8400, fax (508) 750-4744. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., (605 Third Avenue, New York. NY 10158-0012, (212) 850-6011, fax (212) 850-6008. E-Mail: PERMREQ @ WILEY.COM. Library of Congress Cataloging in Publication Data: ‘Adamson, Arthur W. Physical chemistry of surfaces / Arthur W. Adamson and Alice P. Gast. —6th ed. P. om, “A Wiley-Interscience Publication.” Includes bibliographical references and index. ISBN 0-471-14873-3 (cloth : alk. paper) 1. Surface chemistry. 2. Chemistry, Physical and theoretical L. Gast, Alice P. (Alice Petry), 1958- Il, Tile. QD506.A3 1997 541.3°3—de21 Printed in the United States of America 10987654321XVI The Solid-Gas Interface—General Considerations 1. XVII. Adsorption of Gases and Vapors on Solids 1 . The Adsorption Time .. |. Experimental Procedures . ). The BET and Related Isotherm . The Potential Theory. . The Characteristic Isotherm and Related Concepts |. Chemical Physics of Submonolayer Adsorption . Phase Transformations in the Multilayer Region . . Thermodynamics of Adsorption ... CONTENTS Introduction . ‘The Surface Area of Solids . The Meaning of Surface Area . Surfaces as Having a Fractal Geometry . 5 Methods Requiring Knowledge of the Surface Free Energy or Total Energy . Rate of Dissolving The Mercury Porosimeter.. Other Methods of Surface Area Determination mmo ADP . The Structure and Chemical Nature of Solid Surfaces. The Nature of the Solid-Adsorbate Complex .. A. Effect of Adsorption on Adsorbate Properties B. Effect of the Adsorbate on the Adsorbent C. The Adsorbate-Adsorbent Bond... Problems .. General References Textual References Introduction . The Langmuir Adsorption Isotherm .. A. Kinetic Derivation. .. B. Statistical Thermodynamic Derivation. C. Adsorption Entropies D. Lateral Interaction. E, Experimental Applications of the Langmuir Equation ‘A. Derivation of the BET Equation B. Properties of the BET Equation... C. Modifications of the BET Equation . . Isotherms Based on the Equation of State of the Adsorbed Film A. Film Pressure—Area Diagrams from Adsorption Isotherms ... B. Adsorption Isotherms from Two-Dimensional Equations of State... A. The Polanyi Treatment .. B. Isotherms Based on an Assumed Variation of Potential with Distance .. . The Polarization Model. . Comparison of the Surface Areas from the Various Multilayer Models . xv sm si 372 372 374 516 577 377 580 381 582 582 589 591 592 593 594 627 629 630 631 634 639 641xvi CONTENTS ‘A. Theoretical Considerations. B. Experimental Heats and Energies of Adsorption. 13. Critical Comparison of the Various Models for Adsorption A. The Langmuir-BET Model... B. Two-Dimensional Equation of Statc Treatments . C. The Potential Model 14. Physical Adsorption on Heterogeneous Surfaces .. A. Distribution of Site Energy Patches . B. Thermodynamics of Adsorption on Heterogeneous Surfaces. . : C. Point versus Patch Site Energy Distributions D. Geometric Heterogeneity 15. Rate of Adsorption. ... Msi asia ce Paw Soe ta ‘A. Molecular Sieves....... B. Capillary Condensation. C. Micopore Analysis 17. Problems .... General References . Textual References. XVIII. Chemisorption and Catalysis... 1. Introduction 2. Chemisorption: A. LEED Structures....... B. Surface Microscopies..... = C. Spectroscopy of Chemisorbed Species . D. Work Function and Related Measurements. E. Programmed Desorption sues 3. Chemisorption Isotherms. A. Variable Heat of Adsorption. B. Effect of Site and Adsorbate Coordination Number. C. Adsorption Thermodynamics seve 4. Kinetics of Chemisorption A. Activation Energies. B. Rates of Adsorption C. Rates of Desorption .. 5. Surface Mobility 6. The Chemisorption ‘A. Some General Aspects. B. Metals....... C. Semiconductors . D. Acid-Base Systems 7. Mechanisms of Heterogeneous Catalysis. ‘A. Adsorption or Desorption as the Rate-Determining Step... B, Reaction within the Adsorbed Film as the Rate-Determining Step. secsveeeee ee 641 647 652 652 653 654 655 65s 703 705 707 709 n2 12 ns 7 m8 70 720 72CONTENTS 8. Influence of the Adsorption Isotherm on the Kinetics of Heterogeneous Catalysis ‘A. Unimolecular Surface Reactions B. Bimolecular Surface Reactions..... 9. Mechanisms of a Few Catalyzed Reactions ‘A. Ammonia Synthesis .... B. Fischer-Tropsch Reactions. C. Hydrogenation of Ethylene D, Catalytic Cracking of Hydrocarbons and Related Reactions. E. Oxidation of CO... F. Photochemical and Photoassisted Processes at Surfaces. 10, Problems ...... General References . Textual References... Index... xviiCHAPTER XVI The Solid-Gas Interface—General Considerations 1. Introduction ‘These concluding chapters deal with various aspects of a very important type of situation, namely, that in which some adsorbate species is distributed between a solid phase and a gaseous one. From the phenomenological point of view, cone observes, on mechanically separating the solid and gas phases, that there is a certain distribution of the adsorbate between them. This may be expressed, for example, as n,, the moles adsorbed per gram of solid versus the pressure P. The distribution, in general, is temperature dependent, so the complete empirical description would be in terms of an adsorption function n, = f(P. 7). While a thermodynamic treatment can be developed entirely in terms of J(P,T), to apply adsorption models, it is highly desirable to know n, on a per square centimeter basis rather than a per gram basis or, alternatively, to know 0, the fraction of surface covered. In both the physical chemistry and the applied chemistry of the solid~gas interface, the specific surface area L is thus of extreme importance. All gases below their critical temperature tend to adsorb as a result of general van der Waals interactions with the solid surface. In this case of Physical adsorption, as it is called, interest centers on the size and nature of adsorbent-adsorbate interactions and on those between adsorbate molecules. There is concem about the degree of heterogeneity of the surface and with the ‘extent to which adsorbed molecules possess translational and internal degrees of freedom. If the adsorption energy is large enough to be comparable to chemical bond energies, we now speak of chemisorption. The adsorbate tends to be localized at particular sites (although some surface diffusion or mobility may still be present), and the equilibrium gas pressure may be so low that the adsorbent- adsorbate system can be studied under high-vacuum conditions. This allows the many diffraction and spectroscopic techniques described in Chapter VIII to be used to determine what actual species are present on the surface and their packing and chemical state.) This is also true for physisorption systems if the surface is well defined and the temperature low enough that the equilibrium Pressure is very low, see Fig. XVII-17 for example. It has become increasingly appreciated in recent years that the surface struc ture of the adsorbent may be altered in the adsorption process. Qualitatively, 571572 XVI THE SOLID-GAS INTERFACE. TABLE XVI-1 ‘Types of Adsorption Systems ‘Type of Adsorbent ‘Type of Adsorbate- Adsorbent Interaction Molecular Refractory Van der Waals (physical Surface restructuring on 'No change in surface adsorption) adsorption structure on adsorption Chemical Chemical reaction Surface restructuring (chemisorption) modifying the adsorbent on adsorption such structural perturbation is apt to occur if the adsorption energy is compa- rable to the surface energy of the adsorbent (on a per molecular unit basis). As summarized in Table XVI-1, physical adsorption (sometimes called physisorp- tion) will likely alter the surface structure of a molecular solid adsorbent (such as ice, paraffin, and polymers) but not that of high-surface-energy, refractory solids (such as the usual metals and metal oxides and carbon black). Chemisorp- tion may alter the surface structure of refractory solids. ‘The function of this chapter is to summarize some of the general approaches to the determination of the physical and chemical state in both of the types of adsorption systems described. 2. The Surface Area of Solids The specific surface area of a solid is one of the first things that must be determined if any detailed physical chemical interpretation of its behavior as an adsorbent is to be possible. Such a determination can be made through adsorp- tion studies themselves, and this aspect is taken up in the next chapter; there are a number of other methods, however, that are summarized in the following material. Space does not permit a full discussion, and, in particular, the meth- ods that really amount to a particle or pore size determination, such as optical and electron microscopy, x-ray or neutron diffraction, and permeability studies are largely omitted. A. The Meaning of Surface Area It will be seen that each method for surface area determination involves the ‘measurement of some property that is observed qualitatively to depend on the extent of surface development and that can be related by means of theory to the actual surface area. It is important to realize that the results obtained by different methods differ, and that one should in general expect them to differ. ‘The problem is that the concept of surface area tums out to be a rather elusive ‘one as soon examined in detail. The difficulty in stating just what is meant by the “surface area” of a solid, inXVI-2 THE SURFACE AREA OF SOLIDS 373 (o) (b) (e) Fig. XVI-1. Successive enlargements of a hypothetical coastline. (From Ref. 1.) ‘one aspect, can be illustrated by considering the somewhat analogous question of what is meant by the “length” of a particular section of coastline. Super- ficially, it seems easy to say that the length is simply the distance along the shore between two points. Geography books, however, are prone to make such statements as “because of its many indentations, the coast of Maine is actually some 2000 miles long.” It seems reasonable that the coast of Maine should be considered as being greater than that of a straight line drawn from border to border and that the added length due to harbors, bays, points, and so on should be included. However, once one begins to consider irregularities, the problem is when to stop. Figure XVI-1 shows how successive magnifications of a hypo- thetical stretch of coastline might look. Quite obviously there are indentations on indentations, and with each successive magnification one increases the com- puted coastline by a more or less constant factor. There seems to be no logical stopping point to the sequence of successive enlargements if one is after some “absolute” value for the length of the shore- line. One thus arrives at the point of considering irregularities due to individual stones on the beach, due to the individual grains of sand, and due to the sur- face roughness of each grain and, finally, of considering the surface area of individual atoms and molecules. In our particular example, the tides and waves would prevent the exact carrying out of the sequence, but apart from this, the Heisenberg principle comes in eventually to inform us that there is considerable ‘uncertainty in the location of electrons and that the “surface area” of an atom is a somewhat philosophical concept. Quite obviously, the geographer solves this logical difficulty by arbitrarily deciding that he will not consider irregularities smaller than some specified size. A similar and equally arbitrary choice is frequently made in the estimation of surface areas. The minimum size of irregularities to be considered may be574 XVI THE SOLID-GAS INTERFACE defined in terms of the resolving power of a microscope or the size of adsorbate molecules.+ ‘Another approach is the iterative one. In the case of the Maine coastline, one might count the number of coastal villages or landing jetties and take this number as a measure of the length of a stretch of coast. In the case of a solid surface, one might apply a method that attempts to count the number of atoms or molecules in the surface. The result obtained is undoubtedly useful, but again, it does not have the quality of being a unique measure of absolute surface area. Perhaps the closest approach to such a measure of surface area is the geometric area of a plane liquid surface or of a perfect crystalline plane. Finally, in the case of solids, there is the difficulty that surface atoms and molecules differ in their properties from one location to another. The discussion in Section VII-4 made clear the variety of surface heterogeneities possible in the case of a solid. Those measurements that depend on the state of surface atoms or molecules will generally be influenced differently by such heterogeneities. Different methods of measuring surface area will thus often not only give dif- ferent absolute values, but may also give different relative values for a series of solids. ‘The preceding discussion serves to emphasize the point that, although the various topics to be discussed in this and succeeding chapters relate to the common property, surface area, the actual meaning of the answer arrived at is generally characteristic of the method and associated theory. Fortunately, one is very commonly more interested in relative surface areas than in accurate “abso- lute” values. Any given method will usually be valuable in determining ratios of areas between various materials (subject to the reservation made in the preced- ing paragraph). The comparisons, moreover, will be particularly valid when they are to be applied to the prediction or correlation of properties closely related to the one used in the measurement. Thus surface areas estimated by adsorption studies are more likely to lead to valid predictions of how additional, similar adsorbates will behave than to accurate predictions of, say, how the relative heats of immersion or the relative microscopic areas should vary. B. Surfaces as Having a Fractal Geometry Figure XVI-1 and the related discussion first appeared in 1960 [1], and since then a very useful mathematical approach to irregular surfaces has been applied to the matter of surface area measurement. Figure XVI-1 suggests that a coast- line might appear similar under successive magnifications, and one now pro- ceeds to assume that this similarity is exact. The result, as discussed in Section VII-4C and illustrated in Fig. VIL-6, is a self-s ilar line, or in the present case, a self-similar surface. Equation VII-21 now applies and may be written in the form ‘Reference 9 shows a needlepoint embroidery (by V. L. Adamson) illustrating the point that a large adsorbate molecule would "see" less of a rough surface than would a small oneXVI-2 THE SURFACE AREA OF SOLIDS 575 B= No) = Co“) (XVE-L) where N(a) is the number of adsorbate molecules of area ¢ needed to “cover” cone gram of the surface. For a perfect plane surface, D = 2 and E = C’, that is, does not depend on the area of the adsorbate molecules. In the case of a self- similar surface such that 2 < D < 3, E will increase with decreasing « and D may be obtained from a plot of In E versus In a. To the extent that Eq. XVI-1 is obeyed, C’ rather than © now characterizes the adsorbent. The specific surface area E will also vary with the particle size, and a more general equation is iD Cagumoa Rua (XVI-2) where R is the particle radius. If the particle surface is smooth, D = 2 and E varies inversely as R, as expected. In the limit of very porous particles, D—> 3, however, and L becomes independent of the particle size. A plot of In E versus In R thus provides an alternative method for determining D. The question of what @ to use is discussed in Ref. 2. The above methods for obtaining D, as well as other ones, are reviewed in Refs. 3-12, and Refs. 7-9 give tables of D values for various adsorbents. For example, D is close to 3 for the highly porous silica gels and close to 2 for nonporous fumed silica and for graphitized carbon black; coconut charcoal and alumina were found to have D values of 2.67 and 2.79, respectively [7]. ‘The chemical reactivity of a self-similar surface should vary with its frac- tional dimension. Consider a reactive molecule that is approaching a surface to make a “hit.” Taking Fig. VII-6d as an illustration, it is evident that such a molecule can “see” only a fraction of the surface. The rate of dissolving of quartz in HF, for example, is proportional to R-3), where D,, the reactive dimension, is about the same as D determined from adsorption studies (see Refs. 7-12). Values of D have also been obtained from bimolecular energy transfer reactions where acceptor and donor molecules must make an encounter by dif- fusing along the surface and also from the variation of x-ray scattering with scattering angle (see Refs. 4-15). The common factor in all of these approaches is that an analysis of the change in some measurable quantity as the yardstick size (adsorbent size, adsorbate size, diffusion distance, scattering distance) is varied. It might be thought that self-similarity is too demanding a mathematical property to be likely to occur in nature, and some reservations are, indeed, appropriate. If the range is small, data may only appear to fit the linear log-log plot used to obtain a D value. Other explanations, including surface chemical heterogeneity, may be important. Yet it does appear that approximate self-sim- ilarity over a fair range of yardstick size is not uncommon. Nor is it actually unexpected. Computer simulations of particle growth by nonequilibrium accre- tion or aggregation processes do lead to self-similar surfaces, for example (see Refs. 11-18).576 XVL_ THE SOLID-GAS INTERFACE, C. Methods Requiring Knowledge of the Surface Free Energy or Total Energy ‘A number of methods have been described in earlier sections whereby the surface free energy or total energy could be estimated. Generally, it was nec- essary to assume that the surface area was known by some other means; con- versely, if some estimate of the specific thermodynamic quantity is available, the application may be reversed to give a surface area determination. This is true if the heat of solution of a powder (Section VII-5B), its heat of immersion (Section X-3A), or its solubility increase (Section X-2) are known, ‘Two approaches of this type, purporting to give absolute surface arcas, might be mentioned. Bartell and Flu [19] proposed that the heat of immersion of a powder in a given liquid Gimm = L(Esy — Es.) (XVI-3) can be combined with data on the temperature dependence of the contact angle for that liquid to obtain the area ¥ of the sample. Thus from Eq. X-39, (XVE-4) dos @ imm = 2 (& cos @- Ty Fr ) where ginm is on a per gram basis so that E is area per gram, The actual use of Eq. XVI-4 to obtain 5 values is far from easy (see Ref. 20). ‘The general type of approach, that is, the comparison of an experimental heat of immersion with the expected value per square centimeter, has been discussed and implemented by numerous authors [21,22]. It is possible, for example, to estimate Esy - Es. from adsorption data or from the so-called isosteric heat of adsorption (see Section XVII-12B). In many cases where approximate rel- ative areas only are desired, as with coals or other natural products, the heat of immersion method has much to recommend it. In the case of microporous adsorbents surface areas from heats of immersion can be larger than those from adsorption studies [23], but the former are the more correct [24]. A second method that has been suggested as giving absolute surface areas involves the following procedure. If the solid is first equilibrated with saturated vapor, the molar free energy of the adsorbed material must be equal to that of the pure liquid, and it is a reasonable (but nonthermodynamic) argument that therefore the interfacial energy £, for the adsorbed film-vapor interface will be the same as for the liquid-vapor interface. Note that the adsorbed film is thought of as being duplex in nature in that the film-vapor interface is regarded as distinct from the film-solid interface. If the solid is now immersed in pure liquid adsorbate, the film-vapor interface is destroyed, and the heat liberated should correspond to 4 E,, thus giving the area A. This assumption was invoked by Harkins and Jura [21] in applying the method, with some success, to a nonporous powder. A concem, however, isXVI-2_ THE SURFACE AREA OF SOLIDS 577 that if the adsorbed film is thick, interparticle and capillary condensation will occur to give too low an apparent area. Roquerol and co-workers, however, have shown that in the case of a ground quartz—water system, the heat of immersion became constant after only about 1.5 statistical monolayers of adsorbed water [25]; see also Ref. 24. D. Rate of Dissolving A rather different method from the preceding is that based on the rate of dissolving cof a soluble material. At any given temperature, one expects the initial dissolving rate to be proportional to the surface area, and an experimental verification of this expectation hhas been made in the case of rock salt (see Refs. 26, 27). Here, both forward and reverse rates are important, and the rate expressions are ANAC) _ a = ha (XVES) R ah = keane (KVI-6) where aysci denotes the activity of the dissolved salt. By employing a single cubic piece of rock salt, it was possible to confirm the above rate expressions. In such cases, the rate constants ki and k2 are found to depend on the degree of stirring, although their ratio at equilibrium, of course, does not. The rate-controlling step is therefore probably one of diffusion through a boundary liquid film, since the liquid layers next to the solid surface will possess very litle velocity normal to the surface. In dissolving, a concentration Co, equal to that of the saturated solution, is built up in the layer immediately adjacent to the solid, whereas at some distance 7 away the concentration is equal to that of the bulk solution C. The dissolving rate is determined by the rate of diffusion of solute across this concentration gradient. Although the rate of dissolving measurements do thus give a quantity identified as the total surface area, this area must include that of a film whose thickness is on the order of a few micrometers but basically is rather indeterminate. Areas determined by this procedure thus will not include microscopic roughness (or fractal nature). The slow step in dissolving, alternatively, may be that of the chemical process of dissolving itself. This apparently was the case in a study (28] of the rate of dissolving of silica by hydrofluoric acid. The method was calibrated by determining the rate of dissolving of silica rods of geometrically determined area and was then applied to sam- ples of powdered material. The rate was not affected by stirring rate and, moreover, the specific surface area of the silica powder so obtained, about 0.5 m?/g, was found to agree well with that estimated from adsorption isotherms. As noted in Section XVI-2B, however, chemical dissolving rate constants may vary with particle size. E. The Mercury Porosimeter The method to be described determines the pore size distribution in a porous material or compacted powder; surface areas may be inferred from the results,578 XVI THE SOLID-GAS INTERFACE however. The method is widely used and, moreover, does make use of one of the fundamental equations of surface chemistry, the Laplace equation (Eq. 11-7). It will be recalled that Bartell and co-workers [29,30] determined an average pore radius from the entry pressure of a liquid into a porous plug (Section X-6A), ‘A procedure that is more suitable for obtaining the actual distribution of pore sizes involves the use of a nonwetting liquid such as mercury—the contact angle on glass being about 140° (Table X-2) (but note Ref. 31). If all pores are equally accessible, only those will be filled for which 2y\ cos 6} rete (XVE7) 80 that each increment of applied pressure causes the next smaller group of pores to be filled, with a concomitant increase in the total volume of mercury penetrated into the solid. See Refs. 32-35 for experimental details; commercial equipment is available (see Refs. 36, 37). The analysis of the direct data, namely, volume penetrated versus pressure, is as follows. Let dV be the volume of potes of radii between r and r - dr; dV will be related to r by some distribution function D(r): aV=-D(r) dr (XVI-8) For constant 9 and + (the contact angle was found not to be very dependent on pressure), one obtains from the Laplace equation, Pdr+rdP=0 (XVI-9) which, in combination with Eq. XVI-8, gives aV- Di) 5 P= DO) nee? ap (XVE-10) or _ @/nav . pn: Sa (XVI) Thus D(r) is given by the slope of the V versus P plot. The same distribution function can be calculated from an analysis of vapor adsorption data showing hysteresis due to capillary condensation (see Section XVII-16). Joyner and co- workers [38] found that the two methods gave very similar results in the case of charcoal, as illustrated in Fig. XVI-2. See Refs. 36 and 39 for more recent such comparisons. There can be some question as to what the local contact angle is [31,40]; an error here would shift the distribution curve.XVI-2 THE SURFACE AREA OF SOLIDS 579 rasan TS) nen emt arog water 58 | 0.002 = Porosineter #130") 645 Mo) 22 | Masaures (BET) a8 ese ° 700 200 300 Pore radius, + (A) Fig. XVI-2. Comparison of the pore volume distribution curves obtained from porosimeter data assuming contact angles of 140° and 130° with the distribution curve obtained by the isotherm method for a charcoal. (From Ref. 38.) ‘The foregoing analysis regards the porous solid as equivalent to a bundle of capillaries of various size, but apparently not very much error is introduced by the fact that such solids in reality consist of interconnected channels, provided that all pores are equally accessible (see Refs. 41, 42); by this it is meant that access to a given-sized pore must always be possible through pores that are as large or larger. An extreme example of the reverse situation occurs in an “ink bottle” pore, as illustrated in Fig. XVI-3. These are pores that are wider in the interior than at the exit, so that mercury cannot enter until the pressure has risen to the value corresponding to the radius of the entrance capillary. Once this pressure is realized, however, the entire space fills, thus giving an erroneously high apparent pore volume for capillaries of that size. Such a situation should also lead to a hysteresis effect, that is, on reducing the pressure, mercury would leave the entrance capillary at the appropriate pressure, but the mercury in the ink-bottle part would be trapped. Indeed, there is often some hysteresis in that Fig. XVI-3. An ink bottle pore.