Thin film materials pervade our everyday life as transparent conductors in LCD watches and computer displays and in defrosters for automobiles... antireflection coatings for camera lenses… optical fibers for communication … architectural glass coatings for both color and energy efficiency… solar cells… decorative coatings on plastics such as for toys and automobiles parts… a whole host of electronic and optoelectronic devices… hard coatings for cutting tools, drill bits, and bearings … even metallic coatings inside potato chip bags to keep the chips crisp!

Without thin films our lifestyles would be drastically different. And this trend toward increased use of thin film technology will only continue.

The varied reasons for using thin films and the specific deposition processes for preparing them are often complex; but usually relate to function, cost, beauty, materials and energy efficiency, and performance. In addition to technological applications, scientists are finding thin films to be an invaluable tool for investigating new physical phenomena, even at the quantum level. For instance, two of the most important new materials—high temperature ceramic superconductors and diamond coatings — are currently being made by several thin film deposition processes in order to explore both their scientific and technological potential.

Just 25 years ago the variety of deposition processes for preparing thin films was quite limited. Thin film scientists and technologists had at their disposal electrodeposition, elementary chemical vapor deposition, evaporation, and dc sputtering. Commercial equipment for electron-beam evaporation, a mainstay in the optical coatings industry, was just being developed. Most of the deposition processes reviewed in this and next month's MRS BULLETIN were either not commercially available or were not even conceived of then.

The growth of thin films has been instrumental in the study of many areas of material science, physics, metallurgy, and chemistry and is an important ingredient in the development of many devices. Although experimental studies have been extensively pursued for many years, theoretical studies have only been performed using model calculations which rely on a number of unknown parameters a priori. Only recently have attempts been made to understand thin film growth using realtime numerical simulation. The main reason for the recent increase of such studies is the development of computers capable of tackling a problem of the magnitude required to understand thin film growth. The phenomena present in thin film growth occur for systems containing many particles (e.g., columnar growth) and long relaxation times, which strain the capabilities presently available in modern supercomputers. Further increases in computational power might bring a number of important problems within reach and improve our understanding of thin film growth at the microscopic level.

I will present a number of epitaxial growth studies we have performed using molecular dynamics (MD) techniques. I will show that a number of properties predicted by these calculations are in good agreement with experimental observations. These include the microcrystalline and epitaxial growth of metal films, the growth of amorphous films in mixtures of metals, and the vapor phase growth of silicon. Finally, I will outline several important studies yet to be implemented.

Molecular Beam Epitaxy (MBE) is a thin film deposition process in which thermal beams of atoms or molecules react on the clean surface of a single-crystalline substrate, held at high temperatures under ultrahigh vacuum conditions, to form an epitaxial film. Thus, contrary to the CVD processes described in the other articles, the MBE process is a physical method of thin film deposition.

The vacuum requirements for the MBE process are typically better than 10−10 torr. This makes it possible to grow epitaxial films with high purity and excellent crystal quality at relatively low substrate temperatures. Additionally, the ultrahigh vacuum environment allows the study of surface, interface, and bulk properties of the growing film in real time, by employing a variety of structural and analytical probes.

Although the MBE deposition process was first proposed by Günther in 1958, its implementation had to wait for the development of the ultrahigh vacuum technology. In 1968 Davey and Pankey successfully grew epitaxial GaAs films by the MBE process. At the same time Arthur's work on the kinetics of GaAs growth laid the groundwork for the growth of high quality MBE films of GaAs and other III-V compounds by Arthur and LePore and Cho.

Modern compound semiconductor devices are usually fabricated in one or more layers of single-crystal material, which are epitaxially grown on a crystalline substrate. Since most of these semiconductors decompose into their constituent components at high temperature, epitaxial growth in its simplest form can be accomplished by transporting individual components to a heated substrate, where they react to form the compound semiconductor. This is the basis of molecular beam epitaxy (MBE), where the process is carried out in an ultrahigh vacuum environment.

Growth of these materials in an atmospheric (or reduced pressure) environment is highly advantageous from a commercial point of view. However, this presents a problem, since one of the components usually has a very low vapor pressure. For III-V compounds such as GaAs, AlAs, InAs, and InP, this is the column III component. Consequently, a technique must be provided to transport these elements by means of volatile compounds. This process is known as chemical vapor epitaxy, with halogenic compounds often used as transport agents. One example of this approach is the growth of GaAs from GaCl and AsCl3. During growth, the AsCl3 is made to prereact with gallium to form volatile GaCl, which is stable at elevated temperatures and can be transported to the substrate. High quality GaAs can be made by this process, and it is widely used for making a number of GaAs devices.

Chemical vapor deposition (CVD) is one of the few deposition processes in which the deposited phase is produced in situ via chemical reaction(s). Thus the vapor source for CVD can consist of high vapor pressure species at moderate temperatures and yet deposit very high-melting phases. For example, pure TiB2, which melts at 3225°C, can be produced at 900°C from TiCl4, BC13, and H2.

Chemical vapor deposition and its variants such as low pressure CVD (LPCVD), plasma-assisted CVD (PACVD), and laser CVD (LCVD) have been active areas of research for many years. Recent review articles have contained extensive lists of the phases deposited by CVD, which include most of the metals and many carbides, nitrides, borides, silicides, and sulfides. The techniques have found increased acceptance as commercial methods for the fabrication of films and coatings which are fundamental to the semiconductor device and the high-performance tool bit industries. They have been used to prepare multiphase-multilayer coatings, stand-alone bodies, and fiber-reinforced composites. As the demand increases for more complex and sophisticated materials, it is expected that CVD will play a still larger role.

In CVD a solid material is deposited from gaseous precursors onto a substrate. The substrate is typically heated to promote the deposition reaction and/or provide sufficient mobility of the adatoms to form the desired structure. Chemical vapor deposition was performed for the first time when early humans inadvertently coated cooking utensils with soot from the campfire. In this CVD process, hydrocarbons generated by the heated wood pyrolyzed on the utensil surface, depositing carbon.