Physical vapor deposition reactors

The epitaxy reactor is a specialized variant of the tubular reactor in which gas-phase precursors are produced and transported to a heated surface where thin crystalline films and gaseous by-products are produced by further reaction on the surface. Similar to this chemical vapor deposition (CVE)) are physical vapor depositions (PVE)) and molecular beam generated deposits. Reactor details are critical to assuring uniform, impurity-free deposits and numerous designs have evolved (Fig. 22) (89). 

For many applications, like chemical-vapor-deposition reactors, the semi-infinite outer flow is not an appropriate model. Reactors are often designed so that the incoming flow issues through a physical manifold that is parallel to the stagnation surface and separated by a fixed distance. Typically the manifolds (also called showerheads) are designed so that the axial velocity u is uniform, that is, independent of the radial position. Moreover, since the manifold is a solid material, the radial velocity at the manifold face is zero, due to the no-slip condition. One way to fabricate a showerhead manifold is to drill many small holes in a plate, thus causing a large pressure drop across the manifold relative to the pressure variations in the plenum upstream of the manifold and the reactor downstream of the manifold. A porous metal or ceramic plate would provide another way to fabricate the manifold. 

Physical vapor deposition (PVD) is used for the deposition of semiconductor, insulator, and metal layers in the fabrication of a variety of electronic devices. Reactors for PVD are characterized by direct line-of-sight transport of molecular species from the gas phase to the desired substrate, where they react to form solid films with the desired properties. A reactor-and-reaction analysis of PVD quantitatively examines the generation of gas-phase species, spatial distribution of species arriving on the substrate, and surface reactions leading to film growth. 

The SEMICONDUCTOR, insulator, or conductor layers in microscale or larger scale electronic devices such as a photovoltaic cell are created in a reactor. The reactor needs to be designed and operated to produce materials that have the desired optical and electronic properties. The design of reactors is a nontrivial research and design problem. In this chapter, some of the theoretical and experimental framework for this research and for more-effective designs of physical-vapor-deposition-type reactors will be developed. 

Physical vapor deposition (PVD) is a direct line of sight impingement deposition technique. At the low pressures employed in a PVD reactor, the vaporized material encounters few intermolecular collisions while traveling to the substrate, and modeling of deposition rates is a relatively straightforward exercise in geometry. 

The catalytic wall reactor with channel diameters in the range 50-1000 pm and a length dependent on the reaction time required circumvents these shortcomings. However, in most cases, the catalytic surface area provided by the wall alone is insufficient for the chemical transformation and therefore the specific surface area has to be increased by chemical treatment of channel walls or by coating them with highly porous support layers. This can be done by using a variety of techniques such as sol-gel, electrophoretic and chemical or physical vapor deposition [8, 9]. 

In the case of ex situ application of an elemental catalyst, a number of physical vapor deposition techniques can be utilized, including sputtering and molecular beam epitaxy. In this way, a few to tens of monolayers of the catalyst metal are deposited onto the substrate, which is then inserted into the CVD reactor. For both of these choices the rationale for depositing a few monolayers of the catalyst is twofold. First, an extremely low coverage of the catalyst material will facilitate islanding, which is critical to nanowire formation. Second, the amount of catalyst on the substrate will dictate the size of the metal islands, which in turn will dictate the diameter of the nanowires. The primary drawback to these two approaches is that the size distribution of the islands caimot be readily controlled. 

Since no synthetic chemistiy infrastructure was available at the Department (or, indeed, the Institute) before 2008, polyciystalline samples of catalysts had to be obtained from external, often industrial, partners. In order to produce model systems in house, researchers in the Department of Inorganic Chemistry developed a suite of instruments allowing the synthesis of metal oxides by physical vapor deposition of elements and by annealing procedures at ambient pressure. They chose the dehydrogenation of ethylbenzene to styrene on iron oxides as the subject of their first major study. Figure 6.6 summarizes the main results. The technical catalyst (A) is a complex convolution of phases, with the active sites located at the solid-solid interface. It was possible to synthesize well-ordered thin films (D) of the relevant ternary potassium iron oxide and to determine their chemical structure and reactivity. In parallel. Department members developed a micro-reactor device (B) allowing them to measure kinetic data (C) on such thin films. In this way, they were able to obtain experimental data needed for kinetic modeling under well-defined reaction conditions, which they could use to prove that the model reaction occurs in the same way as the reaction in the real-life system. Thin oxide... 

Various nanoparticle preparation methods, such as physical vapor deposition, chemical vapor deposition," reactive precipitation, sol-gel,° microemulsion, sonochemical processing and supercritical chemical processing, have been developed and reported in the literature. Among these methods, reactive precipitation is of high industrial interest because of its convenience in operation, low cost and suitability for massive production. The conventional precipitation process is, however, often carried out in a stirred tank or column reactor, and moreover the quality of the product is difficult to control and the morphology and size distribution of the nanoparticles usually change from one batch to another during production. 

---