Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Report of the Research Briefing Panel on Science of Interfaces and Thin Films
Research Briefing Panel on Science of Interfaces and Thin Films John A. Armstrong (Cochairman), IBM Corporation, Yorktown Heights, N.Y. George M. Whitesides (Cochairman), Harvard University, Cambridge, Mass. John H. Birely, Los Alamos National Laboratory, Los Alamos, N.Mex. Peter R. Bridenbaugh, ALCOA Tech Center, Alcoa Center, Pa. Norman Gjostein, Ford Motor Corporation, Dearborn, Mich. Arthur C. Gossard, AT&T Bell Laboratories, Murray Hill, N.~. Franz I. Himpsel, IBM Corporation, Yorktown Heights, N.Y. N. l. Johnston, NASA Langley Research Center, Hampton, Va. Robert W. Mann, Massachusetts Institute of Technology, Cambridge, Mass. Thomas McGill, California Institute of Technology, Pasadena, Calif. Calvin G. Quate, Stanford University, Stanford, Calif. 2 Dotsevi Y. Sogah, E. I. du Pont de Nemours & Company, Inc., Wilmington, Del. Mark Wrighton, Massachusetts Institute of Technology, Cambridge, Mass. Staff William Spindel, Project Director, Commission on Physical Sciences, Mathematics, and Resources Robert M. Simon, Staff Officer, Commission on Physical Sciences, Mathematics, and Resources Alfred B. Bortz, Consultant Sandra Nolte, Senior Secretary Allan R. Hoffman, Executive Director, Committee on Science, Engineering, and Public Policy
Report of the Research Briefing Panel on Science of Interfaces and Thin Films DEFINITIONS, PROPERTIES, AND BEHAVIORS A thin film is matter of microscopic thick- ness typically, only a few atoms to a few thousand atoms. Its extent in its other two dimensions is macroscopic. An interface is the junchon of two different substances or two phases of the same substance. The properties of these quasi-two-dimensional entities are often remarkably different from the properties of bulk matter of the same · . composition. Thin films and interfaces are associated concepts. Thin films (for example, the- iri- descent film that forms when oil floats on water) are familiar. Interfaces are less fa- miliar, although ubiquitous; they occur wherever different homogeneous phases meet. For example, the oil film on water has two interfaces: one with water and one with air. As the thickness of a film decreases, its properties are increasingly determined bv its interfaces. At the limit of thinness, thin films and interfaces merge. An interface can be thought of as a film so thin it has no homogeneous interior; a thin film is a sys- tem whose interior is strongly influenced by the close proximity of its interfaces. it_ J The current interest in thin films and in terfaces reflects both tamely opportunities in basic science and importance and perva- siveness in technology. · A broadly important problem in science is that of the basic relations between the macroscopic properties of matter (e.g., wettabilit,v, electrical and thermal conduc- hvity, hardness, reflectivity) and its atomic- level structure. Because the detailed struc- tures of interfaces and thin films can be ma- n~pulated with greater control than can those of bulk solids or liquids, they provide par- ticularly attractive systems for studying these basic relations. · A number of interesting phenomena- especially manifestations of quantum be- havior only appear in small systems. Thin films and interfaces can be of a size that displays these phenomena. Thus, for ex- ample, the rate of transport of electrons across thin films by quantum tunneling may be very high when the thickness of the films is comparable to the extension of electronic wave functions (10 to 100 angstroms). · The structure, properties, and reactivity of matter at an interface can be very differ- ent from those of matter in bulk because of 3
Figure 1 Schematic cross section of a drop of water spreading on a solid support. The exis- tence of a "precursor film," a thin lip of liquid (not drawn to scale) extending beyond the drop, is due to long-range interactions between the solid and the liquid. SOURCE: Based on P. G. deGennes, Reviews of Modern Physics, Vol. 57, No. 3 (1985):827-863. the close proximity of the interracial matter to matter of different composition, or of the interfacial matter to vacuum. Thus, gold atoms at a gold-silicon interface do not be- have like gold atoms in bulk because many of their nearest neighbors are silicon atoms; gold atoms at a gold-vacuum interface be- have differently from those in bulk because they lack the complement of near neighbors present in a solid. · The connections between science and technology are particularly close and useful in matters concerning thin films and inter- faces. Technologically important physical properties-strength, corrosion resistance, biocompatibility- are often determined by the characteristics of thin films and inter- faces. An interface can be the exposed atoms on the exterior of a metal single crystal in vacuum, the junction between a silicon substrate and a silicon dioxide overIayer, the boundary between phase-separated block copolymers, the junction between a fiber and polymer matrix in a composite aircraft part, or the region of contact be- tween blood and an implanted prosthesis. A thin film can be the membrane enclosing a living cell; the thin layers of conductors, semiconductors, and insulators that con- stitute a microelectronic device; the lay- ered media of a magnetic information storage disk; the layers used for lubrica- t~on, a~nes~on, and passivation; or the coatings of surfactants that are used to sta- bilize suspensions. The properties of these .. .. . ~. . . thin films depend strongly on their con- stituent interfaces. The examples that follow illustrate several interesting and important characteristics of matter in two dimensions. · The presence of a solid in contact with liquids water has a profound influence on the character of the water. Figure ~ is a diagram of a smaD drop of water spreading on a solid. The cur- vature of the major part of the surface of the drop is determined by a balance of ener- gies at the liquid-vapor, liquid-solid, and solid-vapor interfaces. A feature of great current interest is the so-called "precursor film," a lip of liquid a few hundred ang- stroms thick extending for microns beyond the edge of the drop. Current explanations of this precursor him attribute its existence, at least in part, to long-range interactions between the liquic! and the solid surface. The strength of these interactions is suffi- cient to pull a thin film at the edge of the drop flat against the surface. Electrochem- ical evidence supports a mode! for water next to an interface that is qualitatively dif- ferent from bulk water: interfacial water may have a dielectric constant as low as 30 (the dielectric constant of bulk water is 78~. Un- derstanding the behavior of liquid-solid in- terfaces is critical to understanding wetting, and thus to such technologies as adhesion and corrosion protection. · Matter in thin films may exhibit phase be- havior that is quite different from the phase be- havior the same matter exhibits in bulk. A system composed of krypton that is adsorbed on 4
SCIENCE OF INTERFACES AND THIN FILMS graphite at a pressure and temperature sufficient to give a coverage of one to two monolayers shows remarkably complex phase behavior. The krypton in the first monolayer is a fluid at 130K, the high end of the temperature range. As the temper- ature is lowered, the krypton first freezes Into a two-dimensional solid, then melts into a new fluid, and finally freezes again into a new solid. The structures of the solict phases exhibited by this system have been characterized (see Figure 2~. In the higher- temperature solid phase, the krypton atoms position themselves in a low-density crys- tal in register with the underlying graphite lattice. This structure is dominated by krypton-carbon interactions; and in it, the krypton atoms are spaced slightly beyond hard-sphere contact. In the lower-temper- ature solid, krypton condenses to a higher . ~ ~ W~ , ~ :~ ~ ;~/~¢ ~ ~a ~ Low ~ ~ ~ -43~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . ~ ~ ~ /~ Y ~ ~ ~;~ ~ Y 5 density in the first monolayer anc! crys- tallizes in a structure dominated by kryp- ton-krypton interactions; in this structure, the registration with the graphite lattice is destroyed. This system illustrates the del- icate balance between forces among the krypton atoms in an adsorbed monolayer and between these atoms and the graphite substrate on which they are adsorbed. Studies such as this of krypton on graphite establish the ability of current instrumen- tal techniques to characterize the struc- tures of monolayers; they also provide a starting point for studying the thermo- dynamics of thin films. Although inert gas systems are of limited practical interest, they are the simplest systems to analyze theoretically, and they provide conceptual models for systems dominated by weak atom-atom interactions. Figure 2 density crystalline phase (upper diagram) is dominated by k~ypton-graphite interactions: the krypton atoms are placed slightly beyond hard- sphere contact. The high-density crystalline phase (lower diagram) brings the krypton atoms into contact but destroys the registration with the underlying graphite lattice. SOURCE: Based on R. J. Birgeneau and P. M. Home, Science 232 (1986):329. Krypton adsorbed on graphite. A low
- · BuZk and interfacial electrical and magnetic properties of materials may be strikingly differ- ent. Bulk crystalline silicon is a semicon- ductor; the interface of a silicon crystal cut to expose a particular crystal plane, the Sigh face, exhibits metallic behavior. This phe- nomenon is not yet understood theoreti- catly, but a critical first step establishing the structure of the Sigh interface has been taken recently through the use of scan- ning tunneling microscopy. The observed structure shows significant changes in atomic positions relative to bulk; some bond angles are different. · Far from being passive containers for the contents of the cell, the membranes covering cells are highly organized, dynamic, structurally com- plex biological systems that regulate communi- cation between matter lying inside and outside of the cells. One important constituent of cell membranes is a class of molecules, the phos- pholipids, that spontaneously forms bilayer films in a number of geometries. Many of the important physical properties of cell membranes, such as two-dimensional dif- fusion and differentiation between the "in- side" and "outside" of biological entities shaped like a tube or sphere, can be studied using these spontaneously formed struc- tures. · The characteristic chemical reactivities of metal atoms at the exposed interface of bulk metal and of small metal clusters provide the basis for het- erogeneous catalysis. Heterogeneous catalysis (that is, catalysis using solid catalysts) is an important technology for the production of fuels and chemicals. The metal atoms used in a heterogeneous catalyst are chosen for their high reactivity. As a result of their lo- cation in a solid interface, they are simul- taneously accessible to reactants in a contacting vapor or liquid and isolated from reaction with one another. · Microelectronic devices are assemblies of thin films; many of the properties of these devices de- rive from the special properties of electrons in the films and the transport of electrons in and across the interfaces joining them. Figure 3 is a sche 6 matic cross section through one part of such a device: a multilayer system connecting a transistor to make ohmic contact to a solder pad. The conductors in this system are 0.! to 5 micrometers (1,000 to 50,000 angstroms) in cross section; in operation, they can carry current densities of up to )05 amps/cm2. The force of this "electron wind" is sufficient to cause electrom~gration, or migration of atoms in the conductors from their normal lattice sites. Remarkably, some of these atoms mi- grate with the wind, and some migrate against it. The basis of this phenomenon is incompletely understood, but its origin clearly lies in the small size of the conduc- tors; its control is important in reducing the size of integrated circuit chips. · Insertion of a monolayer, 20 angstroms thick, of a simple organic substance hexadecylamine, CH3(CH24~5NH2 between two steel surfaces in slidling contact reduces the friction between them by a factor of 10. This reduction in friction reflects spontaneous formation of an or- dered film of the organic species: the amine (NH2) groups bond to the metal, and the hydrocarbon chains orient roughly perpen- cticular to it. The film is thus a thin hydro- carbon liquid or liquid crystal bonded to the metal, a material that resists the transitory adhesion between the metal pieces that con- tributes strongly to friction. Understanding the details of the relations between the structure of the adsorbing organic mole- cules, the solid phase on which they adsorb, and the structure and properties of these types of spontaneously self-organizing monolayer films promises to stimulate the design of thin-fiIm systems to control fric- tion, wear, corrosion, and adhesion. In short, matter present in thin films or at interfaces can exhibit unique properties. Understanding and controlling these prop- erties have been difficult because of the small quantities of material present in most inter- faces and thin films relative to the bulk, and because many of the interfaces of greatest interest are "buried" inside solids or under
SCIENCE OF INTERFACES AND THIN FILMS P~Sn solder pad ~ Cu-Sn interrnetallic phased Cr-Cu I": =_Cr Char 2.8 Am SiO2 '-3-8 Em SiO2 2.3 ,um Al-4~o Cu 0.85 mn Al-4% Cu ~ / / -~.4 lam b'~2 race ~1.4 rim A14~ Cur//////////// ~ it_ ~ _ Si3N4 Thermal Sip PtSi' Silicon liquids. New instrumental techniques com- bined with theory and computer modeling, however, enable us to define the structures of many interfaces. With sound structural information ant! new methods of prepara- tion, it is now possible to explore the rich phenomenology of interfaces and thin films. RESEARCH ISSUES CHARACTERIZATION OF INTERFACES AND THIN Firms The bonding characteristics and electronic structure of most interfaces (even the "sim- ple" solid-vacuum interfaces of crystalline elements) are still poorly understood, and no technique for establishing these struc- tures is universally applicable. (The struc- ture of an interface cannot necessarily be extrapolated from that of the underlying bulk solid.) Much of the information about the struc- tures of interfaces has come from forms of spectroscopy that are limited to solid-vac- uum interfaces, but emerging techniques can now characterize solid-gas and solid-liquid interfaces. The most exciting of these tech · . . . .. . Pique s IS scans lug tu nnel lug microscopy, or STM (Figure 4~. STM measures the very small current that flows when a potential is ap O. 15 Am Cr-Cr Joy 7 Figure 3 Sectional drawing of multilevel in- terconnections for advanced bipolar devices. SOURCE: L. T. Fried et al., IBM journal of Re- search and Development 26 (3 May 1982~. plied between a conducting interface and a probe tip (only a few atoms in size) scanned across the interface at a distance of ang- stroms. The current is caused by quantum tunneling of electrons between individual atoms on the interface and on the probe, and it is extremely sensitive to the distance .. - between the atoms. This remarkable device makes it possible to observe individual atoms on irregular interfaces. STM is being used to study interfaces in contact with insulating liquids; it is applicable to noncrystalline sol- ids; and it can be used to examine dynamic processes occurring at interfaces. It should be particularly useful in examining the structures of individual defects on inter- faces. A second instrument relying on the ability to position interfaces with angstrom-scale control is the interface force balance. This de- vice holds two flat solids (for example, sheets of mica) at accurately known separations of from 3 to 500 angstroms, and measures the attraction or repulsion between them. The measurements can be carried out with the solids separated by vacuum, gas, or liquid, or with the solids carrying monolayers of other materials bonded or adsorbed to their interfaces. The measurements make possi- ble direct analysis of the forces responsible for interactions between interfaces, modi- fication of these forces by intervening con
Figure 4 Scanning tunneling microscopy spec- troscopy a topograph of an Si(111) 7 x 7 sur- face. By changing the voltage, one is able to measure electronic states within an area of atomic dimensions. Defects (some of which show up as extra dark spots in the topography can also be probed by this method. SOURCE: l. E. De- muth et al., IBM Corporation. densed phases, and, by inference, interaction A number of techniques are used for prep of the interfaces with the condensed phases. aration of electronic materials, ranging from A selection of the wide range of other in- simple vacuum evaporation to molecular beam strumental techniques now being used to epitaxy (MBE) and metal-organic vapor phase characterize interfaces is presented in Table 1. epitaxy (MOVPE). (Epitaxy is the growth of The sheer number of available techniques a crystalline film of one material on a crystal presents interface science with both an op- face of a second in such a way that the crys portunity and a problem. The techniques talline orientation of the deposited material provide many useful and complementary follows that of the substrate.) Current sci types of information; but because no single entific understanding of the processes un technique uniquely characterizes any sys- derlying aD these techniques-processes that tem, it is necessary for an effective labora- include the movement of species on the in tory to have access to several instruments. terraces, mechanisms of annealing and re Their expense and complexity in turn raise fief of stress, incorporation of impurities, a substantial problem in management for nucleation and growth of defects is lim small research groups. ited; an improvement in that understanding Current objectives of research in struc- would be immensely valuable in technol tural aspects of interface science are the de- ogy. velopment of techniques for characterizing Epitaxial growth techniques seem certain buried interfaces (for example, grain bound- to be particularly useful. They can be used aries in metals and ceramics, the fiber-ma- to make very small structures such as quan trix interface, and defects in composites); turn wells (structures whose composition is and for examining electrically insulating in- tailored at angstrom scales to control elec terfaces such as those on organic polymers. ironic energy levels); films with extraordi narily high electron and hole mobilities; and transistors, lasers, and magnetic materials that are capable of record-setting perfor mances. The most interesting strategies for prep aration of thin films of organic constituents PREPARATIVE TECHNIQUES One objective of current research is the development of techniques for producing highly perfect, smooth, single-crystal films. 8
SCIENCE OF INTERFACES AND THIN FILMS are based on the spontaneous self-assembly of low molecular weight molecules at inter- faces. For example, so-called Langmuir- Blodgett monolayers are formed by spread- ing an organic substance such as stearic acid (CH3tCH2~6CO2H) at the interface between air and water. The hydrophilic CO2H groups are attracted to the water, while the hydro- phobic hydrocarbon chains are excluded from it. When the surface film is compressed, the organic molecules pack as a two-dimen- sional crystal or liquid crystal that can be transferred intact to a solid support. Similar films can be formed in many cases by simply aDowing the organic substance to adsorb from solution onto a support: the desired order- ing occurs spontaneously. These techniques make it possible to prepare macroscopic, two- dimensional, organic monolayer films while maintaining a high degree of control over the nature of the exposed surface functional groups, the order of the films, and (to a more limited extent) their physical and me- chanical properties. These systems are ex- ceptionally attractive as substrates for the study of relationships between interface structure and such properties as wettability, biocompatibility, electrical resistivity, and nonlinear optical response. PROPERTIES OF MATTER IN TWO DIMENSIONS The quasi-two-dimensional geometry of interfaces and thin films, and the high gra- dients in properties across them, can result in unique properties for matter in these sys- tems. Of the wide range of topics that might be used to illustrate the characteristic prop- erties of interfaces and thin films, device physics offers a particularly clear demon- stration of the interplay between science and technology. The use of thin films is impor- tant in the construction of devices for two reasons: (~) small size permits a high den- sity of devices, thus minimizing power con- sumption and maximizing speed; and (2) small size is required for many devices that exploit quantum effects. A two-dimensional electron gas exists at the interface between the silicon channel and gate insulator in metal oxide semiconductor (MOS) devices; similar electron gases are TABLE 1 Selected Spectroscopic Techniques Applicable to Interfaces Technique (Acronym) Scanning tunneling microscopy (STM) Neutron scattering Low-angle x-ray scattering Surface-enhanced Raman sp ectroscopy (SERS) X-ray photoelectron spectroscopy (XPS); Auger spectroscopy High-resolution electron microscopy Rutherford backscattering (RBS) Secondary ion mass spectroscopy (SIMS) Reflectance infrared spectroscopy Acoustic microscopy Nuclear magnetic resonance spectroscopy (NMR) Electron paramagnetic resonance spectroscopy (EPR) 9 Application Individual atomic positions on surfaces Structures of crystalline surfaces; surface morphology Vibrational spectroscopy of adsorbates on small metal particles Electronic structure of atoms in the top 10 angstroms of an interface A wide variety of information concerning structure, composition, and morphology; single atom . . . . Imaging In spectra cases Atomic composition as a function of depth with resolution of hundreds of angstroms Molecular-sized fragments of interfaces Vibrational and structural analysis of organic thin films Interface structures at the 10,000 A level Interface structures at the 10,000 A level Paramagnetic centers in interfaces
important in many other types of devices. The transport properties of these two-di- mensional electron gases exhibit a number of new phenomena: for example, negative resistance, in which current decreases with increases in applied voltage because of elec- tron tunneling into energy sub-bands with lower momentum; and ballistic transport, in which electrons move without phonon scat- tering in structures with very small dimen- sions. Heterojunctions (structures involving an interface between two different semicon- ductors) are the basis for a number of de- velopments in device physics. New light- emitting structures involving multiple het- erojunctions (so-called quantum-well light emitters) provide light at wavelengths pre- viously unattainable. Superiattices formed by producing a series of periodically spaced heterojunctions are materials with new non- linear optical and magnetic properties. INTERFACE REACTIVITY Atoms and molecules present at an inter- face can experience a highly anisotropic en- vironment with characteristics that are different from surrounding bulk phases). The chemical reactivity of a species present at an interface may, in consequence, be dif- ferent from the reactivity of the same species present in an isotropic phase. As one example, aggregates of platinum atoms supported on alumina react with hy- drocarbons in ways that depend strongly on the size and shape of the metal!aggregate, on the acidity of the underlying alumina support, and on the nature of the interaction between the aggregate and the support. Platinum atoms are intrinsically highly re- active toward hydrocarbons. But platinum atoms in bulk platinum are not accessible; hence, they are not active catalytically. Plat- inum atoms in solution tend to react indis- criminately with one another, with species used to increase their solubility, and with the intended reactants. Small aggregates of supported platinum thus provide systems 10 do, in which a high proportion of the platinum is stably isolated and exposed at an inter- face, available for reaction. In these sys- tems, the reactivity of the platinum can also be tailored to favor useful reactions by changing the underlying support. The reac- tivity toward hydrocarbons of platinum supported on alumina forms the basis for one of the critical steps in petroleum refin- ~ng. A second example is the unexpected ac- idities of organic functional groups present at the interface between low dielectric poly- mers, such as polyethylene, and water. The apparent acidity of a carboxylic acid (CO2H) group at such an interface that is, the con- centration of protons in solution at which these groups are half-ionized to carboxylate ions (CO2-) is shifted by lot from that characterizing the same groups in homo- geneous aqueous solution. The shift in ap- parent acidity reflects three factors: (~) the low local dielectric constant at the polymer- water interface; (2) the anomalously low po- larity of the water present at this interface; and (3) electrostatic interactions at the sur- face. The unexpected reactivity of functional groups present at polymer-fluid interfaces is clearly relevant to wetting of and adhe- sion to polymers, and to other processes involving the reaction and salvation of func- tional groups present in polymer interfaces. It is also relevant to the characteristics of functional groups present at many other in- terfaces, especially those between water and suspensions, micelles, and proteins. BIOCOMPATIBEE SURFACES AND INTERFACES A biomaterial is any substance or device whose function depends on contact with a biological medium. Thin films and inter- faces play an essential role in the design and function of the numerous implants and de- vices that are now being used clinically. Their surfaces induce deposition of proteins, platelets, and other cellular elements, and
SCIENCE OF INTERFACES AND THIN FILMS often induce platelet aggregation and blood clot (thrombus) formation. Although functional aspects of the per- formance of artificial materials in the human body can be predicted with some reliability, forecasting their biological performance is difficult. Fundamental information on the correlation between the in viva and in vitro responses is limited. An understanding of the dynamic biological changes occurring at the material-tissue interface is necessary to predict biological performance. For practical application, factors influencing interactions of blood with materials and biological sur- faces are particularly important. We must learn how these interactions are affected by surface morphology or by specific chemical groups on the surface, and what influence is exerted by the mechanical properties of the interface. APPLICATIONS MICROELECTRONICS IN COMMUNICATIONS AND COMPUTERS Microelectronic technologies are founded largely on structures composed of thin films and interfaces. The silicon-based semicon- ductor industry depends critically on the still incompletely understood interface between silicon and insulators such as silicon dioxide and silicon nitride. Lasers, detectors, and new high-speed devices use heterojunc- tions in which the interface is the key active component. The conduction of electricity in microfabricated devices requires reproduc- ible interfaces between metallic conductors and the active semiconducting layer, the so- called ohmic contact. Such contacts become increasingly difficult to establish as dimen- sions decrease. Mass storage technologies (magnetic disks, optical and magneto-opti- cal storage devices, magnetic bubble mem- ories) depend on films that are only hundreds of angstroms in thickness but that are uni- form over many square centimeters. 11 Packaging and interconnect technologies are as important as chip technologies to fu- ture developments in high-performance computing systems. Processing speeds are often limited by the time required to prop- agate information from one chip or subsys- tem to another, and the fabrication and assembly of very small systems is crucial in building high-speed computers. Of neces- sity, the components in these systems are densely packed, and they may generate large quantities of heat whose removal is con- trolled by interracial thermal conductivity. Long-term objectives include the devel- opment of systems that allow direct con- nectioh of electronic devices to biologically based sensors or to nerves. Both require the solution of substantial problems in interface science. Another challenging interracial problem is the development of practical techniques that interconvert electrons (the currency used by digital devices) and neu- rotransmitters (the currency used by nerves) to permit the nondestructive stimulation or sensing of nerve impulses. CONTROL OF CORROSION, FRICTION, AND WEAR Corrosion is the result of electrochemical processes occurring at interfaces between metal and water and air. Control of corro- sion is usually achieved by the application of thin protective films to the metal inter- face. Although techniques for corrosion control are highly developed empincally, too often they are effective only for short peri- ods, and their fundamental basis is often obscure. Studies of the formation, thermo- dynamic and kinetic stability, and barrier properties of thin films on metals are now possible at levels of detail that will be in- creasingly useful in developing new strat- egies for control of corrosion. The practice of lubrication has developed adequate engineering models for elastohy- drodynamic lubrication (EHL), provided the lubricant is present as a thick film and be
haves as a Newtonian fluid. As the thick- ness of the lubricant film approaches molecular dimensions, as in asperity con- tact, pressures become high, approaching ~ gigapascal; shear forces become very high; and lubricants solidify, crystallize, and de- grade. Many lubricated surfaces are covered with softer adherent films derived at least in part from the lubricant; these films mit- igate asperity contact when forces exceed the EHL limit. Understanding thin lubricant and adherent films, especially under the ex- treme conditions of asperity contact, is now conceivable, although still difficult. Controlling wear at the head-medium in- terface in a high-density magnetic storage device provides an example of a current problem in this technology. The thin-fiIm magnetic recording medium on a disk is part of a multilayer structure. It is protected by another thin film a wear-resistant over- coat that may, in the future, be a dia- mon~ike carbon. The magnetic medium may be bonded to its substrate by yet another thin film, and there may be a magnetic un- derIayer. In Winchester hard disk technol- ogy, a m~crofabricated head literary flies over the disk, supported by an air bearing whose width is comparable to the average distance between collisions of inctividual molecules in ambient air (~900 angstroms). Efforts to increase the density of data storage require closer head/disk spacings; they also raise formidable problems in lubrication, "stic- tion" (adhesive effects of the lubricant dur- ing static head/disk contact), wear, adhesion, and delamination. STRUCTURAL MATERIAES Interfaces play a central role in determin- ing the mechanical properties of a range of structural materials, among which are fiber- reinforced composites, metals, and ceram- ics. In broad terms, the failure of structural elements almost always involves inelastic deformation and fracture. The atomistics the details of kinetics and thermodynam 12 ics of the processes involved in initiation and propagation of cracks in a material un- der strain, and the processes that protect against fracture (dissipation of energy through the creation of dislocations, for- mation of cracks and voids, and other mech- anisms) are various and incompletely understood. But they all have as a common result the formation or alteration of inter- faces. Fiber-reinforced composites consist of an ordered dispersion of fibers in a polymer matrix. The fibers provide stiffness and strength; the matrix distributes the load among the fibers and protects them from damage. The large area of interface con- necting fiber and matrix makes a critical con- tribution to the mechanical properties of the material. Failure often involves these inter- faces (Figure 5), but the relationship be- tween their structure and mechanical performance is poorly unclerstood. How tightly should the fiber and the matrix ad- here to achieve optimal performance? What mechanical properties in the interface best dissipate energy during incipient failure? How should the optimum interface be pre- pared? What is the proper surface chemistry for the fiber, and how should the fiber be brought into contact with the math? An- swers to these questions would provide the basis for rational optimization of composite systems and would be of substantial value to users of composites, especially in the au- tomotive and aircraft industries. ENERGY PRODUCTION Chemically active interfaces are required as heterogeneous catalysts in the refining of petroleum. Corrosion-resistant interfaces are needed in the heat transfer systems of power plants. Fuel cells and certain kinds of batteries can be substantially improved by new cata- lytic electrodes that will allow more efficient use of atmospheric oxygen. The reduction of oxygen by electrons to water is slow at
SCIENCE OF INTERFACES AND THIN Figure 5 Fractured surface of a polymeric composite: unsized chopped carbon fibers in a polycarbonate matrix. SOURCE: NASA Jet Propulsion Laboratory. conventional electrodes, and even the best platinum catalysts have rates that give out- put voltages that are only one-half of what is theoretically possible. The inefficiency of these catalysts is not intrinsic to all sys- tems that reduce oxygen; in biological sys- tems, through molecular mechanisms that are still incompletely understood, the re- duction of oxygen to water takes place rap- idly. Recent experiments with synthetic catalysts have demonstratec! the reduction of oxygen to water without the use of plat- inum, but substantial improvements in the lifetime and cost of electrode-confined ca- talysts must be achieved before the sys- tems are economical. 13 The production of petroleum illustrates another important set of interracial prob- lems. One of the approaches to increasing the production of crude oil from partially depleted fields is to "launder" the oil-bear- ing rock using detergents similar to those used in cleaning oil-stained clothes. Water containing the detergents is pumped into the reservoir; the detergent separates the petroleum from the rock and permits it to pass through fine pores in the rock. Currently, the economic feasibility of de- tergent-based methods for of! recovery re- mains unclear for most oil fields; it has proved difficult to find detergent mixtures that are inexpensive, effective, and stable in the res
ervoir. Designing successful detergents will require a detailed knowledge of the rela- tions between their structures and the ways in which they modify the properties of oil- water interfaces. NATIONAE SECURITY Many of the applications of interface and thin-film technology in military and civilian systems are similar, with the important dif- ference that military systems must function in extremely demanding environments with very high standards of performance. A fi- ber-reinforced composite wing for a military aircraft is subjected to greater stress than a similar wing on a civilian aircraft; therefore, optimization of the fiber-matrix interface is correspondingly more important. Because an engine for a tank operates much closer to its limits of failure than a comparable en- g~ne for a civilian truck, the design of critical components such as bearings requires bet- ter control of interfaces to minimize friction and wear. There are also, however, certain areas of technology in which military requirements are unique. Interfaces are critical wherever high electromagnetic fields interact with matter. Mirrors for high-powered lasers must be immune to optical damage, and interfa- cial phenomena play a dominant role in de- termining their damage thresholds . Multilayer structures composed of alternat- ing thin films of refractory metals and di- electrics are important for x-ray opt*s. High- powered accelerators operate with very high surface fields in their resonant cavities; the composition and morphology of the cavity interfaces contribute both to the sharpness of the resonant frequency and to the rate of surface heating during use. Certain military electronics devices must resist damage ~ ra- diation; high speeds and low power con- sumption are also important. Promising technologies to meet these requirements are based on thin-fiIm heterostructure devices. In addition, new thin-film coatings are 14 needed that prevent the reaction of lithium hydride and uranium with water vapor over intervals of decades. The stabilities of nuclear devices, an important element in the design of test ban treaties, are greatly influenced by interracial reactivities. MANAGEMENT ISSUES Increased U.S. investment in the science of thin films and interfaces will produce a large return in improved technology. What are the goals of an appropriate investment strategy? What are the management issues raised by its execution? 1. Investment should build strong, two- way interactions between basic science and advanced technology. Inefficient two-way communication between these two spheres is a major hindrance in the cycle of product development; because technological appli- cation of thin films and interfaces is only a small step beyond basic research, research results will be invaluable to technologists. Unexplained and uncontrolled phenomena in technology, in turn, will stimulate basic research. 2. Investment should encourage interdis- ciplinary collaboration and join and exploit the strengths of academic, industrial, and government research and development in- stitutions. 3. An effective strategy should provide selected groups and/or institutions with a sufficiently complete subset of the sophis- ticated analytical and preparative tools re- quired to conduct effective research in thin films and interface science. One possible in- stitutional structure would be research groups of about four faculty members that would focus on a coherent theme (for ex- ample, microelectronic interfaces, electroac- tive surfaces, or biocompatible materials). Each such group would manage a reason- able subset of preparative and analytical tools (see Table 1) that while possibly incomplete should still be sufficient to carry out a major
SCIENCE OF INTERFACES AND THIN FILMS portion of the group's preparative and an- alytical work, supplemented by the use of equipment in other locations. 4. The scale of modern surface and in- terface research fats in between "small" and "big" science; this intermediate size com- plicates issues in management. For exam- ple, although individual researchers are not forced (as is the case in experimental high- energy physics) into large collaborative projects with explicit management struc- tures, the single principal investigator, with his specialized technique or knowledge, sel- dom has adequate resources to solve com- plex experimental problems. 5. Personnel Requirements. Too few trained students are being produced in some important areas of interface research. For 15 example, although there are many well- trained students in compound semicon- ductor interface science, there are compar- atively few in such areas as polymer-metal, polymer-carbon, and polymer-ceramic in- terfaces; silicon epitaxial growth; colloid sci- ence; and biocompatible surfaces. In summary, interface science is excep- tional in its close, two-way interaction with technology. Only within the past few years have the tools and techniques become avail- able for a concerted attack on the scientific problems of buried interfaces. Increased in- vestment is therefore both scientifically timely and certain to pay significant dividends for technology.
