SYMPOSIUM COMMITTEE President: Sir ERIC K. RIDEAL, F.R.S. Chairman: Professor D. H. EVERETT Secretary: Dr. R. H. OTTEWILL Professor S. BRUNAUER Professor J. TH. G. OVERBEEK Professor J. H. DE BOER Professor M. PRETTRE Mr. A. S. JOY Mr. M. K. SCHWITZER Dr. H. VAN OLPHEN INTERNATIONAL UNION OF PURE AND APPLIED CHEMISTRY in conjunction with THE SOCIETY OF CHEMICAL INDUSTRY SURFACE AREA DETERMINATION Proceedings of the INTERNATIONAL SYMPOSIUM ON SURFACE AREA DETERMINATION held at the School of Chemistry, University of Bristol, U.K. \Q—\SJuly, 1969 Symposium Editors D. H. EVERETT and R. H. OTTEWILL LONDON BUTTERWORTHS ENGLAND: BUTTER WORTH & GO. (PUBLISHERS) LTD. LONDON: 88 Kingsway, W.C.2 AUSTRALIA: BUTTER WORTH & CO. (AUSTRALIA) LTD. SYDNEY: 20 Loftus Street MELBOURNE: 343 Little Collins Street BRISBANE: 240 Queen Street CANADA: BUTTERWORTH & CO. (CANADA) LTD. TORONTO: 14 Curity Avenue, 374 NEW ZEALAND: BUTTERWORTH & CO. (NEW ZEALAND) LTD. WELLINGTON: 49/51 Ballance Street AUCKLAND: 35 High Street SOUTH AFRICA: BUTTERWORTH & CO. (SOUTH AFRICA) LTD. DURBAN: 33-35 Beach Grove © International Union of Pure and Applied Chemistry 1970 Published as a supplement to Pure and Applied Chemistry Suggested U.D.C. number: 541-18:531-72 IUPAC Publications Chairman, Editorial Advisory Board: H. W. THOMPSON Scientific Editor: B. C. L. WEEDON Assistant Scientific Editor: C. F. CULLIS Assistant Editor: P. D. GUJRAL Standard Book Number: 0 408 70077 7 Printed in Great Britain by Page Bros. (Norwich) Ltd., Norwich PREFACE The proposal to hold the Symposium whose proceedings are reported in this volume was made, on the initiative of Sir Eric Rideal, by the IUPAG Commission on Colloid and Surface Chemistry at its meeting in Prague in August 1967. The objective was to take a first step towards reaching inter- national agreement on methods of determining the surface areas of solids of different kinds, acceptance of which would make feasible the subsequent goal of the establishment of a set of Standards for Surface Area Determina- tion. The project was supported by the Committee of the Colloid and Surface Chemistry Group of the Society of Chemical Industry who collaborated with IUPAC in the organisation of this Symposium. It would have been over-optimistic to have expected full and complete agreement in such a widely representative Symposium attended by some 289 delegates from 19 different countries. Nevertheless, as these proceedings show, many of the methods proposed in the past for surface area determina- tion are now well established, at least on an empirical basis, and the inter- relationships between them are becoming clearer. That the fundamental theoretical basis for these methods is not yet securely established is not very surprising: but there is now a wider appreciation of the nature of the out- standing difficulties facing a more complete general theoretical study. On the practical side, this Symposium, and the discussions which have arisen from it, have provided a basis for the planning of a project for the establishment of a set of Standards for Surface Area Determination. This is expected to begin as a pilot scheme initiated by the Society of Chemical Industry and then to expand into an international programme under the aegis of IUPAC. Bristol D. H. EVERETT March 1970 R. H. OTTEWILL v OPENING REMARKS Sir ERIC K. RIDEAL Physical Chemistry Laboratories, Chemistry Department, Imperial College, London, S.W.I., U.K. This joint meeting of the Colloid and Surface Chemistry Group of the Society of Chemical Industry and the Commission on Colloid and Surface Chemistry of the International Union of Pure and Applied Chemistry (IUPAC) comes at a very opportune time. The Group has now been in existence for eleven years having previously functioned as a panel for two years, whilst this IUPAC Commission was formed in Montreal in 1961. The same title was chosen for both Group and Commission to reflect the national and international aspects respectively. Whilst the Group has furthered the subject by means of many joint meetings with other groups of the Society and by the holding of symposia on subjects of topical impor- tance, the Commission had definite terms of reference laid down for it by IUPAC at the Montreal meeting. Since 1961 the Commission has been actively engaged in investigations as to the state of our knowledge, permitting us to make positive recommenda- tions for international usage. Two of the terms of reference were, briefly, (i) Definitions, terminology and symbols; (ii) Conventions in naming colloidal systems. The fruit of many years of labour requiring the most detailed enquiries and critical examination was presented in the form of a report by the Chair- man of our Commission, Professor J. Th. G. Overbeek, to the Council of IUPAC at the meeting at Cortina D'Ampezzo in Italy in July 1969. The Commission in consequence is in a position to commence work on other aspects of the problem laid down in the Montreal terms of reference. Two of these were (i) Definition of pore volumes, bulk density, particle density, types of isotherms, size and sign of phase boundary potential; (ii) Agreement covering the description of finely divided solids and of measuring the adsorption per gramme and per square centimeter of adsorbent. This Symposium on surface area determination will give the Commission not only a good start but also encouragement to pursue this important phase of its activity. We are further encouraged by the fact that Professor V. N. Kondratiev, the President of IUPAC, in his Presidential address in 1968 expressed the view that the phenomenon of catalysis was now becoming so important both industrially and from the academic point of view that IUPAC should pay particular attention to this field and in his opinion the Commission on Colloid and Surface Chemistry, possibly with modification, seemed to be the proper vehicle for carrying out this work. Whilst Professor Kondratiev included both homogeneous and hetero- geneous catalysis in his address, the Commission was already aware through the lengthy report of Professor J. Horiuti of the complexities involved in any really adequate treatment of heterogeneous catalysis without completion 1 SIR ERIC K. RIDEAL of the remit in respect to adsorption and surface structure laid down for us at Montreal, and this I regard as an essential priority. I would like to say a few words about this present meeting. A number of countries in the world have bureaux which are responsible for standards especially those in which commerce is involved. For standards to become international in character we must have a uniform and consistent experi- mental procedure as well as an agreed method of representing or treating the experimental results. Materials usedfor adsorption are no exception. Perhaps the B.E.T. method of treatment is the nearest approach to this aim. We are honoured in having one of the originators, Dr. S. Brunauer, with us to-day. If we are to adopt this pragmatic approach, we could ask the question, how far should we extend this idea? Should we develop inter- nationally agreed methods for the evaluation of Lewis and Bronsted acidities of alumino-silicates, for the efficiencies of catalysts for various chemical reactions, e.g., SO3, NH3, hydrogenation, cyclisation and the like or develop standard methods for the evaluation of supported catalysts? Apart from this utilitarian objective we must attempt to arrive at agreements on concepts regarding surface area since widely different values are obtained when different methods of evaluation are employed. Adsorbents and catalysts are complex systems and in any ultimate analysis of total and available surface we should be able to evaluate the respective contributions of lattice defects, crystal planes and edges as well as Kelvin and sub-Kelvin capillaries and their shapes to the total sorption. We should also seek agreement on the concept of a close-packed monolayer formed on a crystal plane and what P, 0 relationship permits us to evaluate it. Whilst it has long been customary to consider that the effective diameter of a Kelvin pore is two molecular thicknesses less than the true diameter, this arbitrary correction must be reconsidered in the light of the possibilities of anomalous values of surface energies and molecular volumes of vapours condensed in what Hardy termed the boundary state. These possible anomalies may likewise affect the validity of computation of the areas determined by methods involving heats of wetting. Finally, I might mention that within the compass of our task we require to know more about the state of an adsorbed monolayer on a crystal facet. The simplest form, namely, an epitactic configuration, would be revealed by a Langmuir isotherm in which complete atomic coverage could be achieved. In the future the ligand structure and bond strength should be evaluated. In general, lateral interation between the adsorbate molecules causes the growth either of a two phase surface structure, i.e., a patchwise interfacial growth, or repulsive forces may impose regular patterns not coincident with the sub- strate on the complete monolayer. Dispersion or physisorption is frequently termed van der Waals sorption. We might ask whether critical phenomena, e.g., pressure and temperature, have an objective reality or whether a Boltzmann distribution between bound and free molecules represents the state of affairs more closely, whether the spreading pressure is analogous to osmotic rather than gas pressure and whether two-dimensional diffusion constants could not be evaluated for such systems. If but a few of these observations have any validity it is evident that there is plenty of work ahead of us. The contributions to this meeting show that advance is possible. 2 INTRODUCTORY LECTURE J. TH. G. OVERBEEK van't Hqff Laboratory, University of Utrecht, Utrecht, The Netherlands It is my task now to give a brief introduction to this Symposium and perhaps to announce a few subjects that we might want to emphasize in the discussions. The first idea for this Symposium was to obtain a survey of the different methods for surface area determination, to compare their results, and, if possible, to formulate recommendations for methods of measurement or calculation, that could be internationally accepted. I do not need to stress that area or specific area (= area per unit mass) is an important parameter of a finely divided solid (or liquid) phase. But why is it difficult to measure? The answer is: because it is not well defined. Determining the surface area of a solid is affected with the same difficulties as determining the surface area of England, or Holland or of the Italian Dolomites. On a small scale map the answer is simple, but it is not very accurate and it neglects the structure of the surface completely. So then we have to decide which part of the surface roughness is to be taken into account. Only those features that can be read from the map with elevation contours ? Or the actual roughness of the rocks and soil? Or the roughness of the sandgrains and the individual pebbles? There is no unambiguous answer; only an arbitrary choice is possible. The same is true for the surface area of a finely divided solid. And the choice will depend on the use we are going to make of the value of this surface area. Looking at the problem from this point of view, we want to know the complete geography of the surface, accurate to the kind of, and number of, the surface atoms, and including the way in which they are bound to the next lower layer. Such information would be good enough to explain all possible interactions with substances that can be adsorbed, move along the surface, react at the surface, or even move in the gas or liquid phase in the neighbourhood of the surface. However, this detailed answer is as good as and about as useless as saying that all the properties of a glass of good wine are contained in the Schrodin- ger equation for its nuclei and electrons. In the papers for this Symposium, these aspects are mostly not mentioned explicitly, but it will be well to keep them in mind in the discussions. Cases, where indeed this 'molecular' or 'atomic' way to look at the surface becomes preponderant are found when crystallographic data are used for surface area determination (assuming that the total surface consists of flat crystallographic surfaces) or also, when instead of asking how many molecules of nitrogen or of another adsorbate are packed on a certain area, one considers how many surface atoms of a given kind of molecule of a certain adsorbate is attached. But in the main we will discuss surface area in these days from a somewhat simplified point of view, i.e., we form a kind of envelope through the centres 3 J. TH.G. OVERBEEK of all the surface atoms, and consider the area of this envelope. In a great many cases we are not content to know the area of this envelope, but we want to know also, how it is folded in slits, pores, and other surface irregularities. In discussing pores, most people will first think of cylindrical pores, but on second thoughts, these are rather improbable. Slits either between two parallel crystal faces, or perhaps wedge shaped slits are already more probable, and in the papers, the old ink-bottle image is used several times to describe a cavity with a narrow entrance. Remarkably enough the image of interstices between a mass of packed and possibly somewhat sintered particles, which might be applicable in several practical cases, is only rarely referred to. Passing now to the papers presented at this Symposium, we find great emphasis on surface area determination by adsorption methods in different variants. a\. The methods used most frequently are based on the determination of the monolayer capacity of a given adsorbent by the BET method, by the ^-method, which is based on it, or by adsorption from solution, using either simple molecules such as benzene or toluene, or large and complicated ones such as dyes. In order to convert the number of molecules in a complete monolayer to a surface area, one needs the area per adsorbed molecule and this requires a calibration based on the adsorption on a non-porous adsorbent, the surface area of which is known from other data. A very important question is then, whether the surface area per molecule is independent of the chemical nature of the adsorbent. If the adsorbent is porous, application of the Kelvin equation leads in prin- ciple to the determination of the volume contained in pores of given sizes and so, with the aid of a pore model, to the distribution of the surface area amongst the pores. However, the method needs several refinements, one of these being the division of pores into macropores (too big for capillary con- densation), intermediate pores ovmesopores (typical for capillary condensation) and micropores (too small for application of the Kelvin equation). a%. Instead of using the adsorption isotherm at about the complete monolayer and higher coverages one may use the low (linear) end of the adsorption isotherm and when the Henry's law constant for the particular combination of adsorbent and adsorbate is known, derive a surface area from that part of the isotherm. Obviously, surface heterogeneity and the need of calibration with an adsorbent with known area are weak points of this otherwise attractive method. 03. For surfaces carrying an electric layer negative adsorption of the co-ions can be used to obtain a surface area. This method, which was first proposed by Schofield for clay surfaces, leads to an absolute value of the surface area without the need of a calibration, but it is not applicable to porous solids and in its present form it is not accurate for specific surface areas below a few m2/g. b. A completely different approach, leading without calibration to an absolute value for the surface area is based upon the determination of the heat of immersion. If the adsorbent is precoated with a sufficiently thick layer of molecules the liquid in which it is to be immersed, immersion results in 4 INTRODUCTORY LECTURE the destruction of the surface areas containing a surface enthalpy per unit area equal to that of the free liquid and to the liberation of a corresponding amount of heat. With presently available calorimetric techniques the method requires a non-porous solid with a specific surface area not less than 10 m2/g. c. Another method or rather group of methods uses the flow of a gas or liquid through a plug or membrane to give information on its surface area. From the steady rate of flow, using the Kozeny-Carman method, a relation between porosity and surface area can be obtained. This method, however, does not take the area and porosity connected with blind pores into account. Moreover, a pore model and an estimate of the 'tortuosity' are needed. Non-steady state methods, in particular the determination of adsorption— and desorption—time lags do give more information and, if judiciously applied, some information on the blind pores. From the adsorption isotherm, the total area can be obtained as in the methods considered under a. d. In the case of expanding clays the surface area can be determined directly from x-ray crystallographic data, but this case is a rare exception. Small angle x-ray scattering, on the other hand, allows the determination of the total surface area in the sample without any separate calibration. Moreover, information can be obtained on pore size or particle size dis- tribution. A remarkable feature of this method is, that it also takes closed pores into account, which are inaccessible to all adsorption methods. e. Finally, mention should be made of the use of the electron microscope or of the ordinary microscope for determining particle size distributions and surface areas of non-porous particles and of the important case of carefully drawn glass fibres, where length, mass, and density are sufficient to allow calculation of the area. Summarizing, we find that there is quite a variety of methods available and, in favourable cases, more than one method may be applied to the same sample. The methods are of different reliability and of different accuracy; some are absolute, some need a calibration. It is one of the purposes of this Symposium to evaluate these methods critically and perhaps reach a conclusion as to which method or methods and which calibration parameters are to be recommended for particular purposes. 5 THE BET-METHOD J. H. DE BOER Department of Chemistry and Chemical Engineering, Technological University of Delft, Delft, The Netherlands 1. PHYSICAL AND CHEMICAL ADSORPTION The equation of Brunauer, Emmett and Teller1, the BET-equation, is based on van der Waals' attraction forces. These forces may all be considered to be of a physical nature and we may suggest that the conceptions of Van der Waals' adsorption' and of 'physical adsorption' cover the same pheno- mena. Langmuir's original equation2 may, under certain circumstances, be used to describe certain adsorption phenomena of physical nature, but also those of chemical nature, hence 'chemisorption' fits into this conception. Both the BET- and the Langmuir-equations can be, and are, used to estimate specific surface areas of adsorbents. In this contribution I shall make use of both conceptions. 2. THE NATURE OF PHYSICAL ADSORPTION The simplest definition of van der Waals' forces is perhaps the one given by Margeriau3 in 1939, namely that van der Waals' forces are those which give rise to the constant a in van der Waals' equation: (v - b) = RT (1) This includes: (i) attraction forces caused by the polarization of molecules by static dipole- or quadrupole fields of other molecules, hence Debye's4 induction effect of 1920; (ii) the forces between molecules possessing dipoles or quadrupoles, hence Keesom's5 alignment effect of 1921; and (iii) the non-polar van der Waals' forces, also called the dispersion forces since London6 discovered the close connection between their nature and the cause of optical dispersion. It is also logical that the forces between a polar surface and dipole mole- cules and the electrostatic induction of atoms or molecules by polar surfaces are included in the definition of van der Waals' adsorption forces. Doing this we follow closely Brunauer's7 concept of physical adsorption. We, therefore, include all cases in which neutral atoms or molecules interact with surfaces without sharing of electrons taking place, or without exchanging of electrons, thus preserving the individuality of the adsorbed neutral atoms or molecules. The nature of the thus defined van der Waals' forces does not exclude them from causing intra- or intermolecular bonds in cases which may be considered to belong to the domain of chemistry; I only need to draw the 7