Colloid and surface chemistry in industrial research

surface chemistry and physics in industrial research are many and varied; they .... been developed in our chemical engineering department by Sternling...
5 downloads 0 Views 8MB Size
J. N. Wilson Shell Development Company Emeryville. California

Colloid and Surface Chemistry in Industrial Research

This paper has been written with two purposes in miod. The first is to demonstrate some incentives for teaching the principles of surface chemistry and physics to students of chemistry, many of whom will make their careers in industry. The second is to illustrate the utility of idealized models or simple concepts that can be derived from these principles in the study of the freqnently complex problems that arise in industrial research. Opportunities to apply the disciplines of colloid and surface chemistry and physics in industrial research are many and varied; they occur in a wide variety of industries. One illustration of this fact is the frequent occurrence in C&E News of product advertisements for materials that are applicable to the control of surface phenomena. In my own experience of 18 years in industrial research, I cannot remember a time when there was not a t least one important project in progress in my own organization that involved surface chemistry in a significant way; usually there are several. It is interesting, and perhaps unfortunate, that many research chemists in industry, university, and government laboratories have had little formal training in surface chemistry. After some experience in industry, a man will often develop an interest in surface chemistry and teach himself by reading and by experience, but this kind of learning generally comes harder than that from formal instruction in an academic atmosphere, where the principles can be presented in a systematic way and integrated with the rest of physical science. This interest is illustrated by the enthusiastic reception that was given to a series of seven lectures on the subject presented recently to a group of his associates by Dr. F. M. Fowkes, one of our leading specialists on the subject. The men to whom these lectures were presented were working on problems associated with the improvement of fuels and lubricants; they were clearly eager to become more familiar with whatever general principles might be available to serve as conceptual guides in their attack on frequently complex problems. The lectures stimnlated so much lively discussion that Dr. Fowkes was invited to repeat the series, not only a t Emeryville but a t other research installations within the company, and to groups of chemists concerned with a broader range of problems. Surface Chemistry in Industry

One reason for the strong interest in surface chemistry Presented as part of the Symposium on the Teaching of Colloid and Surface Chemistrv before the Divisions of Colloid and Snrface Chemistry and "chemical Education at the 140th ACS Meeting, Chicago, September, 1961.

among industrial chemists 13 the fact that it is so difficult to avoid interfaces. Most of our operations are contained by surfaces, many of them involve the creation and destruction of interfaces, and many of them occur a t interfaces. Some of the fields in which surface chemistry plays an important role are listed in Table 1. Table 1.

Some Fields in which Surface Chemistry Ploys on Important Role

Sintehg (ceramics, c&lysts,ketals) Semiconductors (recombination and trapping of charge carriers) Heterogeneous catalysis Electrochemical kinetics Biochemical phenomena Evaporationiontrol Liquid-liquid extraction Sorptive drying and pumping Surface coatings and adhesives Corrosion and its control Lubrication and prevention of wear Stabilization and destabilization of emulsions, foams, sols, and cels.

The relevance of surface chemistry to most of the topics in the table is fairly obvious. I n the case of mechanical grinding of solids, for example, it is well known that wet grinding often requires less energy than dry grinding to reach a given average degree of subdivision; Russian investigators have shown that the energy requirements can be substantially lowered still more by the addition of a suitable surface-active agent. The driving force in the sintering of compacted mixtures of small solid particles, a process of importance in ceramic technology and in powder metallurgy, is the difference in specific surface free energy between small particles and larger ones; surface diffusion is often an important component of the material transport process in the initial phases of sintering. Marangoni Effects

Chemical engineers have been plagued for years by the problem of designing liquid-liquid extraction systems on a rational basis without. extensive pilot plant data on the actual systems under consideration. Even with a sophisticated consideration of the hydrodynamics of turbulent systems, and the distribution of eddy sizes and velocities in such systems, attempts a t design based only on the mechanical parameters of the system and on the physical properties (viscosities, solubilities, diffusivities, etc.) of the two immiscible components and of the solute to be partitioned between them have in general been quite unsuccessful. Sometimes even the qualitative behavior of the extraction Volume 39, Number 4, April 1962

/

187

process was not adequately predicted by the calculations; for example, in many cases there is a big difference between the rate of transfer of a given solute from one phase to another and the rate of the reverse process. The reason for these differences was until recently a mystery; it now seems clear that this mystery has its roots in well-known snrface phenomena that are called the Marangoni effects (I), and whose origin was first explained about a centurv ago (2). The first Marangoni effect is the motion of material along an interface in response to a gradient of interfacial tension along that interface. The flow is not restricted to the interface; it is propagated into adjacent bulk phases by viscous interactions. The second effect is the conjugate of the first: extension or contraction of an interface will cause the interfacial tension to depart from its st.atic equilibrium value. The first Marangoni effect is responsible for the phenomenon known as "tears of strong wine"; preferential evaporation of alcohol from the upper portions of the surface of a glass containing a liqueur will raise the local snrface tension of the thin film of liquid on the glass. As a consequence, mass flow of liquid up the sides of the glass will occur and drops of liquid will form and grow on the glass some distance above the snrface of the liquid. When a drop becomes large enough so that the downward force of gravity upon it is greater than the upward force of the flowing thin film of liquid, the drop will slide down the wall toward the bulk of the liquid, while new droplets will form and grow in its former place. When a solute is diffusing from one liquid phase to another, fluctuations in t,he concentration of solute along the interface will give rise to flows along the int,erface; under the right circumstances, this flow will give rise to a local small-scale eddy which reinforces the original fluctuation. I n this case, a localized interfacial turbulence, which can he quite intense, will develop ~pont~aneously on one side of the interface. The stirring produced by such a turbnlence can clearly increase the rate of transfer of solute across the interface substantially. Interfacial turbulence has been observed in many systems and is believed to be responsible for the wellknown phenomenon of spontaneous emulsification. A theory which clarifies the conditions under which interfacial turbulence will occur spontaneously has been developed in our chemical engineering department by Sternling and Scriven (3) on the basis of a simplified model. The theory shows that t,he stability of the interface when solute is diffusing from one liquid to another depends not only on the effect of the solute on the int,erfacial tension, but also on the relative magnitudes of the diffusion coefficient of the solute in the two phases and on the relative magnitudes of kinematic viscosity in the two phases. The same authors present a review of some of the strange and sometimes dramatic effects that are observed when two uneqnilibrated liquid phases are brought into contact (1). It has been recognized in recent years that another practical application of the Marangoni effects occurs in the drying of paint. If evaporation of the more vokile components from the paint film results in a n increase in the snrface tension of the film, the film may deform itself in such a way as t,o exaggerate brush 188 / Journal of Chemical Education

marks and other imperfections of the surface. Additives called leveling agents are frequently used to stop this kind of thing from happening. I have dwelt a t some length on these practical manifestations of the Marangoni effects for two reasons: to show that though many of the macroscopic principles of surface chemistry have been known for a long time, new applications of these principles to the understanding and control of practical processes are still being developed; and to illustrate the power of a general principle of this kind once its applicability has been recognized. Surface-Active Agents

Another way to illustrate the great variety of ways in which surface chemical problems arise in industrial research is to consider the variety and range of application of commercially available surface-active agents. The variety of known surface-active agents is very great: Schwartz, Perry, and Berch list over 2000 tradenames among which are included over 200 important specific compounds (4). These are normally classified into the categories anionic, cationic, and non-ionic; each category contains 6-12 different chemical types and each type has a great many specific variants. Some of these variants are polymers, and this means that the range of variation in molecular weight and composition is almost infinite. The manufacture and sale of surface active agents in this country is a $250 million per year business. This may be compared in magnitude to the manufacture of elastomers, which is about a $400 million per year business. There are clearly opportunities for research in the development of new surface active agents and the improvement of old ones. Volume 2 of Schwartz, Perry, and Berch (4), covering only developments for the period 1947-1956, lists 700 references that were judged to he sufficiently important for inclusion. That represents about 70 important papers and patents a year, and there are undoubtedly many others of lesser importance. Even t,hough the prospects for development of new surface-active agents are considerably more restricted than they were 25 years ago, there are broad opportunities for research in the application of these materials. The enormous variety of available agents means that the problem of selecting the best ones for a particular application is far from trivial. Some of the common applicat,ionsare listed in Table 2. Table 2 .

Some Functions of Surface-Active Aaents

Detergents Emulsifiem l>emulsifiers Foamers Deiosrnem Leveling Agents Evaporation Barriers Wetting Agents Rust Inhibitors

Anti-static Agents Anti-icing Agents Friction and Wear Reducers Flotation Agents Textile Finishes Paper Sieee Plating Aids Germicides

The function of detergents is to promote the detachment of part,icles of dirt from a snrface that is to be cleaned, and to suspend the dirt in the cleaning fluid either in the form of colloidal particles or by soluhilizing it within t,he interior of micelles. By far the largest

volume of detergents is used in aqueous systems, but there are important applications in non-aqueous systems. For example, some dry-cleaning fluids contain surface-active agents, and most lubricating oils used in internal combustion engines contain detergents whose function is to prevent the deposition of so-called varnish and carbon on the workimg surfaces. A good deal remains to be learned about the detailed mechanisms whereby detergents function, and about the ways in which mixtures of detergents either reinforce or neutralize the effectiveness of the single components. The same thing can be said of emulsifying agents. One of their functions, obviously, is to lower the interfacial tension between two liquids and thus to minimize the free energy required to extend the interface between them. Another function is to generate a n interfacial charge and thus to raise a free-energy barrier against the flocculation and coalescence of dispersed droplets. But there are considerable areas of ignorance still about the detailed requirements for a good emulsifying agent and about why, for example, one material promotes the formation of water-in-oil emulsions while another promotes the formation of oil-inwater emulsions. Similar things can be said about most of the functions on the list. I n most cases some general requirements for useful agents are known, but the theory is not snfficiently developed to permit one to predict the utility of a given amphipathic material accurately from its structure. This means that there is still plenty of room for both basic and applied research on these agents and the way they work. It is still possible, however, to usc some general principles as a guide in the application of these materials. For example, one sometimes wants to make an emulsion of a water-insoluble liquid or solution such that the suspended particles will plate out rapidly on a given surface: on a metal that is to be coated with an organic rust inhibitor, or on the fibers of a mop that is to be impregnated with an emulsion of a waxy polish. I n this case the principle is to choose the surface-active agent (anionic or cationic) in such a way that the emulsion particles will carry an electric charge that is opposite in sign to the charge that develops on the surfaces to be covered under the conditions of application. A second example is provided by the foampromoting additives that are sometimes added to synthetic detergents. If the additive is to affect the foaming properties of an aqueous solution of the detergent, it must compete effectively with the detergent for position a t the air-water interface. This means that it must be more effective than the detergent in lowering the surface tension of water (5). A foam promoter must meet some additional requirements, of course, but it must a t least meet the requirement ontlined above. This criterion provides a simple laboratory screening test for selecting candidates that are worthy of further consideration from those that are not; it also provides a guide for the modification of agents that contain a promising-looking arrangement of polar groups but do not meet the requirement of competitive adsorption a t the air-water interface. Industrially Important Colloidal Systems

Let us turn now to a consideration of some colloidal systems that are of industrial importance.

Table 3 shows some such systems that are based on interactions among anisometric particles. Paper, felts, and soapbase lubricating greases involve fibrous particles whose length is much greater than the other two dimensions. Drilling muds and clay ceramics involve Table 3.

Suspensions and Gels of Anisometric Particles

Lubricating Greases: Soap base, clay base, and dye base Paper manufacture Drilling muds Felted meterials Clay ceramm

platelike particles whose thickness is much less than the other two dimensions. Some of the principles underlying the behavior of drilling muds have recently been clarified by the work of H. van Olphen and his colleagues a t our Exploration and Production Research Laboratories in Houston (6). He has studied the properties of dilute suspensions of purified bentonite clays in water and in aqueous salt solutions and has arrived a t the following picture (Fig. 1): the ultimate particles of bentonite clay are very thin sheets, a few tens of angstroms thick and a few microns wide. The flat faces

Figure 1. Schematic reprerentmtion of pyrophillite cloy strudure. The open circler represent oxygen atoms. In the Montmorilloniter, negatively charged ion-exchange rite, me formed where silicon doms are replaced by aluminum.

of theso sheets are negatively charged because some of the tetrapositive silicon atoms in the clay lattice have been replaced by tripositive aluminum. I n aqueous suspensions, these negatively charged surfaces will be surrounded by a diffuse layer of positively charged gegenions. If the concmtration of electrolyte is very low, the double layer will be very diffuse, and the clectrical repulsion between the faces of two particles will be more than sufficient to overcome the London attractive forces between them. The edges of the particles, on the other hand, are positively charged, because a t the edges some of the aluminum ions are exposed and are no longer completely neutralized; some of the oxide ions which balance their charge in the intact lattice are absent a t the edge. These positively charged edges will attract the negatively charged faces. I n the absence of excess electrolyte, then, conditions are ideal for the development of a gel a t quite low concentrations of clay, Volume 39, Number 4, April 1962

/

189

less thau one per cent by weight,. The prevalence of edge-to-face contacts makes it possible to build a scaffold structure rather like a house of cards (Fig. 2). I t is easy to show that the minimum concentration of clay required to fill space with such a structure is of the order of magnitude pt/w, where t is the thickness, p is the densky, and w is the width of the platelets. One can therefore form a very dilut,e gel in this way only wit,h highly anisometric part.icles.

Figure 2. Schematic representotion of different stater of aggregation of Montmorillinoid clay plateleb.

If one now adds salt to this system, the major effect is to decrease the thickness of the double layer around the particles, in accordance with the type of theory that has been elaborated in recent years by Verwey, Overbeek, Derjaguin, and others (7). At a certain salt concentration, the range of the electrical repulsion bet,ween the faces of the clay particles will be sufficiently reduced that the clay platelets will be pulled together by the London attractive forces between them and will form aggregates similar in structure to a poorly arranged pack of cards. These will be much less anisomebric than were the original isolated platelets, and much higher concentrations of clay will now be required t,o form a gel. This model is a good example of a combination of several concepts that are important in modern colloid chemistry: the effects of anisometry in the interactions among groups of part,icles, the electrical attractions and repnlsions between particles, and the effect of electrolyt,es on the range of these electrical forces. The model has been elaborated by van Olphen to give a qualitative and in some cases even a semiquantitative t,reatmmt of the rheology, gel-strength, and thiiotropy of bentonite suspensions as a function of the pH, electrolyte concentration and type, and clay concentrat,ion. Of course, actual drilling muds are much more complex, since they contain a variety of other ingredients which are added to fuifill special functions. The practical systems are largely developed by empirical trial. Nevertheless, van Olphen's model provides a conceptual framework which is useful as a guide in attempts to develop cheaper and more effective drilling fluids. This type of model is frequent.ly useful in industrial research. Soap-Based Greases

Lubricating greases are colloidal systems which present a variety of interesting problems. The soapbased greases contain from 5-20% by weight of a soap, or mixture of soaps, dispersed in a moderately viscous oil. They are made by either of two methods: the more common method is to heat a mixture of soap and 190

/

Journol o f Chemical Education

oil to a temperature a t which t,he system becomes a homogeneous fluid, chill the mixture to thicken it, and subject the gelled mass to mechanical shear to soften the gel to the final consistency. The second method is to dissolve fatty acid or a mixture of fats and fatty acids in oil and react the mixture with lime or soda to form the soap in situ. The process is completed by heating the grease to drive out the water formed in t,he react,ion, cooling the mixture, and shearing it to homogenize it and to stabilize the structure. I t is now known, as a result of investigations carried out in these Laboratories (8)aud elsewhere, that greases based on the soaps of the alkdies or alkaline earths have a tangled brush-heap structure of long thin crystalline soap fibers, with the oil held by capillarity within its irregular meshes. It is easy to show by simple calculations that to obtain the desired consistency in the grease with the minimum concentration of soap one shodd maximize the length-to-thickness ratio of the soap fibers. In principle one should be able to measure this ratio by direct measurement,^ of the images of the soap fibers on electron micrographs such as those shown in liigure 3; in practice t,his is not only tedious but difficult. One can obtaiu a measure of this ratio from the intrinsic viscosity of dilute suspensions of the soap fibers in a suitable hydrocarbon (7). The ability of the struct,ure to hold oil by capillarity (hence the ahility of the grease to resist "bleeding") is detcrmined by the average size of t,he interstices bet,ween the soap fibrrs; this in turn can be related to the concentration of the soap and the dimensions of its fibers. , : .. . .

Figure 3. Eiectmn micrograph d crystdline loop flberr from a lubricating grease thickened with sodium roapr: left, chilled rapidly from the melt; right, cooled slowly from the melt.

Soap-base greases are mauufactured by dissolving a suit,able mixture of soaps in hot oil, cooling the solution at a suitable rate, and snhject,ing the grease either during or after cooling, or both, to a process of mechanical working to obtain the desired consistency. The control of this process involves the control of the nucleat,ion and growth of soap crystals; these processes are strongly inflvenced by the composition of the soaps used and by surface active additives, as well as by the degree of supersaturation and hence the rate of cooling of the saturated soap solution (see Fig. 3). The function of the final mechanical shearing process is to break up weak crystals and aggregates of crystals so that the consistency of the grease will change less rapidly with further shearing during use. Indeed, in modern greases, this is done so successfully that a very long useful life can be achieved.

Such a grease is thermodynamically unstable, since the soap phase has a very high specific surface area. There should be a tendency for the thicker fibers to grow a t the expense of the thinner ones. There will be a greater tendency for soap to deposit a t the junction of two soap fibers, if those fibers are in actual contact, since deposition in the acute angle of such a junction will cause an even greater reduction in surface free energy. Deposition a t such a junction should result in a cementing of the fibers together, and this should reflect itself in a hardening of the grease. I t is desirable for practical reasons to minimize this hardening, which does occur in some grease formulations. One mechanism through which such a deposition could occur is solution from the thinner crystals, diffusion through the oil, and deposition a t the junctions. This suggests the desirability of using a soap whose solubility is very low a t storage temperatures. Lithium soaps and calcium soaps are widely used for the manufacture of stable greases. The small radius of the monovalent lithium ion and the double charge of the calcium ion should make for low solubility, high melting point, and high strength of the soap fibers. Poper Monufochrre

There are some interesting parallels between the colloid chemistry involved in the manufacture of paper and the colloid chemistry of soap-base grease. Paper is made by depositing a suspension of cellulose fibers on a wire screen, draining the water from the deposit, drying it, and pressing it flat. Fillers incorporated in the initial suspension modify the poresize distrihution, the bonding between the fibers, and the mechanical properties of the finished paper. Sizing agents incorporated in the initial suspension or added to the surface of the sheet after it is formed also modify these properties as well as the wettahility of the paper. The dimensions of the fibers used are obviously important in several stages of the paper-making process: they control the rheological properties of the initial slurry, the rate of drainage of water from the deposit on the screen, and the strength and pore-size distrihution in the finished sheet. They will also control the extent to which the sheet is compacted by capillary forces as it is dried. The fiber dimensions are controlled by the paper maker by means of a cutter or breaker, which cuts the fibers of the starting material into shorter lengths, and by a device called a beater, which churns the slurry and splits the fibers and frays them to increase the effective length-to-diameter ratio. The sizing agents incorporated in the beaten slurry may modify the structural changes which occur in the beating process; they certainly modify the wettability of the paper fibers by water or by ink. The pressure to which the paper is subjected during the finishing operations will also modify the pore-size distribution, and this will affect the rate a t which liquids can penetrate the paper.

control of inter-particle interactions have a great deal in common. Let us consider a few examples: The chemist who is developing new surface-coatings such as paints must concern himself with the adequate dispersion of pigment part,icles and with control of their flocculation. In general it is difficult to prevent the sedimentation of pigment particles completely; when sediment does form, it should be loose rather than compact in order to facilitate re-dispersion. This implies that the attractive forces between the pigment particles should not be completely neutralized by elcctrostatic repulsive forces. I t is also desirable that a paint wet the surface it is to cover so that it will flow out well and not gather into drops on the surface. I t is also desirable that the paint stay flat while it dries. This will recall our previous discussion of the Marangoni effects. The relatively new latex paints involve additional surface and colloid chemical problems. They are normally made by latex polymerization, a process in which micelle formation by surface-active agents and the solubilization of water-insoluble monomers by micelles are important. The stability of the latex must be controlled so that the paint has adequate storage stability over a wide range of temperatures and is compatible with a variety of dispersed pigments; yet when the paint dries, the particles in tho latex must wet the surface to which the paint is applied and must coalesce with each other t,o form a coherent coating. Recently latex paints have appeared that are thixotropic, so that the paint will not run down the handle of the brush and into the painter's hand, but will flow smoothly from the brush on to the surface to be covered. The balancing of all of these factors is a complicated art, and it is often impossible or a t least prohibitively difficult to underst,and all of the effects in quantitative detail, but a knowledge of the general principles that control these effects can still be a very effective guide to applied research. One factor that can be measured in these systems is the average size of the latex particles, and for this purpose the techniques of light-scattering, turbidimetry, and ultracentrifugation that have long been used in colloid chemistry provide useful methods. Rubber latex is used in substantial volume for the manufacture of paints but it is also used for manufacture of dipped goods, elastic thread, and foam rubber. I n all of these applications the colloidal properties of the latex play an important role. The art is frequently complicated and many of its details are not understood quantitatively, but some of the qualitative principles are clear. Silica gel is an important material commercially as an adsorbent, a gelling agent, a filler for waxes and polymeric materials, and a carrier for catalysts. It is now Table 4.

Emulsions: Food and medicinal emulsions Cosmetics Pesticide sprays Sols:

Isometric Portkle Systems

Table 4 lists some industrially important colloidal systems that involve isometric or nearly isometric particles. A wide range of areas of application is repre seuted here, but the scientific problems involving the

Suspensions and Gels of Isometric Particles

~~~~& } Emulsion polymerization

Pigments

Gels:

Silica gel Alumina-silica cracking catalyst Carbon blacks Silicebase erease Volume 39, Number 4, April 1962

/

191

well established that the elementary particles in silica gel are very small spheres. Since gels can be made in water containing as little as one per cent by weight of SiOn, which corresponds to a volume fraction of SiOn around 0.005, it is interesting to inquire into the mechanism of gelation. It is worth recalling in this case that the fraction of void space in a simple cubic packing of spheres is 47.6%, whereas in dilute silica gels the volume fraction of fluid is 99.5%. Gelation in this case must involve an extremely loose or open packing of spheres, with strong short-range attractive forces but long-range repulsive forces between the spheres so that most of the spheres are bonded to only two others. I n dried silica gels one seldom finds the volume fraction void higher than 70%; a considerable collapse of the very open initial structure occnrs when the gel is dried. The example of silica gel shows that the occurrence of gelation a t high dilutions does not necessarily imply that the elementary particles from which the gel is formed are highly anisometric. Studies of the silicagel type of gelation mechanism have arisen in other industrial applications. Semiconductor Surfaces

The physicists, chemists, and electrical engineers who are interested in semiconductor devices are concerned with so-called ambient effects, the effect of the ambient atmosphere on performance. These effects arise because chemisorption changes the nature of the surface, and thus changes the distribution of electronic energy levels. This is reflected in a change in the populat,ion and mobility of charge carriers near the surface, and in the work function. Electrical phenomena a t the interface between two solids are import,ant in some electronic devices. I n addition, etching technioues arc oft,en used to obtain information about solids, e.g., the number and d k i b u t i o n of dislocations. The incentive to learn more about semiconductor surfaces has resulted in a considerable revival of research on the surface physics of solids in general. Heterogeneous Catalysis

Heterogeneous catalysis is obviously a field of great industrial importance and just as obviously it involves surface chemistry. Despite a great deal of research in both academic and industrial laboratories all over the world, we are still unable to predict a pion' what catalysts will be active and selective for a given chemical reaction or t o explain why, for example, silver is a uniquely good catalyst for the reaction of ethylene with oxygen to yield ethylene oxide. This is hardly surprising, since we are not yet able t o make very good predictions of the activation energies and frcquency factors even for homogeneous uncatalyzed reactions. Even though the present state of knowledge in surface chemistry is inadequate for predicting catalytic behavior, however, there are still possibilities for using a knowledge of colloid and surface chemistry in the improvement of known catalysts, or for studying their behavior in use. For example, both the activity and the selectivity of a given catalyst may be influenced by diffusional effects that are controlled by the pore-size distribution of the catalyst. It is important to eliminate or to correct for these effects when one is testing 192

/

Journol of Chemical Education

catalysts, so that meaningful values of the activity and selectivity can be obtained. It is also important to be able to control the porosity and the particle size of commercial catalysts in order to minimize the diffusional limitations. This may lead one to studies of the colloid chemistry involved in the process of catalyst preparation and to studies of the nature of the porosity in the cstalyst and the mechanisms whereby pores are formed. When solid catalysts are used a t high temperatures, or when they are regenerated by burning off carbonaceous deposits, the catalysts frequently lose both surface area and pore volume. These effects usually cause a loss in activity and sometimes in selectivity. For these reasons the student of heterogeneous catalysts is interested in the kinetics and mechanism of the sintering of high-area porous solids, an interest that is shared by the ceramist and the worker in powder metallurgy. Since the driving force in sintering is the tendency to lose surface free energy, this process also involves principles of surface chemistry. I n our own studies of sintering, we have found it helpful to distinguish between two types of material transport: surface processes, which lead t o a loss in surface area with little loss in pore volume; and volume processes which lead t o a loss of surface area and pore volume simnltaneously (9). Studies of chemisorption have hem a useful part of research on heterogeneous catalysis in several ways. First, the results of such studies have made it clear that the intermediates in heterogeneously catalyzed reactions are chemically bonded to the surface of the catalyst. Nowadays this is a widely accepted principle, but the principle has been established by many decades of research, some of it done in industrial laboratories. The detailed nature of this chemical bonding is still only imperfectly understood, however, and there is a need for better tools and for well-designed experiments to stndy it. Even in the absence of detailed understanding, chemisorption provides a useful tool for comparing different methods of preparation of a given catalyst, since the amount of chemisorption is often proportional to the amount of catalytically active surface in the catalyst sample under study. Conclusion

I have tried to give in this paper an impression of the wide range of problems in industrial research to which the principles of surface and colloid chemistry are relevant. I should mention also that in recent years there has been a growing awareness of the importance and utility of these principles among chemical engineers (10).

There are many opportunities for research in surface chemistry, not only in the applied fields but also in basic studies. Though there are well-established macroscopic principles available for the correlation and prediction of surface-chemical and colloidal behavior, our understanding of surface chemistry a t the molecular level, particularly with respect to the dynamics of surface phenomena, is still primitive. The growing sophistication in the applied fields mentioned above, the recent advances in biophysics and biochemistry, and the realization of the importance of surface phenomena

in the new applications of solid-state physics to instrumentation, communication, and computing have focused attention on these deficiencies in our knowledge. As a result there is a renewed interest in surface-chemical problems; certainly the number and quality of publications in the field have been increasing in recent years. There is no doubt in my mind that instruction in the principles of surface and colloid chemistry will be useful to most students of chemistry, whether they make their careers in industrial research, in chemical engineering, or in academic teaching and research. Literature Cited (1) SCRIVEN, L. E., AND STERNLING, C. V., Nature, 187, 186 (1960). This is an interesting account of the history of the Marrtngoni effects and their importance in a variety of practical problems.

J., Phil. Mag. Sw.4, 10, 330 (1855). (2) THOMSON, L. E., A. I. Ch. E. Journal, (3) STERNLING, C. V., AND SCRIYEN, 5,514-23 (Dec. 1959). (4) SCHWARTZ, A. M., PERRY,J. W., A N D BERCH,J., "Surface Active Agents and Detergents," Interscience Publishers, New York; Vol. 1, 1949; Vol. 2, 1958. W. M., A N D FOWKES,F. M., J . P h y ~Chem., (5) SAWYER, 62,159 (1958). (6) TAN OLPHEN,H., in "Clays and Clay Minerals," SWINEFORD. ADA.Editor. Nat. Acad. Sci.-Nat Res. Council Pub. 456, i956,'pp. 20ti224. (7) VERWEY,E. J. W., AND OVERBEEK, J. Th. G., "Theory of the Stability of Lyapbobio Colloids," Elsevier Publishing Co., Inc., New York, 1948. (8) BONDI,A. A,, ET AL., World Pdrol. Congr., Proc., b d Cong., The Hague, 1951, C. J. Brill, Leiden, 1951, Section 8, p. 373. (9) SCHLAFFER, W. G., MORGAN, C. Z., AND WILSON,J. N., J . Phys. Chem., 61, 714 (1957); ADAM%C. R., AND VOGE,H. H., J. Phys. Chem., 61, 722 (1957). (10) DAYIES,J. T., Trans. Inst. Chem. Eng. (London), 38, 289 (1960).

Volume 39, Number 4, April 1962

/

193