Thermal CVD reactors

A typical configuration of a thermal CVD reactor is shown in figure 13.1. The applications of conventional CVD are far too extended to cover in this introduction. 

Fig. 28 Thermal CVD reactor to deposit SiOg-PPXC nanocomposites and parylene based copolymers.

The thermal CVD reactors mentioned in the introduction are divided into two classes— hot wall reactors and cold wall reactors. Hot wall reactors consist of... 

The MOCVD process does not involve a new type of CVD reactor. The reactor used is basically a thermal CVD rector. Metallo-organic compounds are used as precursors usually in conjunction with other reactants to reduce the operating temperature required in a thermal CVD reactor. The precursors and equipment costs are high for MOCDV reactors. As a result MOCVD is used when high-quality coating is required. These reactors are typically open-type reactors that operate in the Torr to atmospheric pressure range. 

FIGURE 10.11 Thermal CVD reactor the plasma chamber is equipped with a heating jacket for preheating of the substrate/initiator. (See insert for color representation of the figure.)... 

Plasmas can be used in CVD reactors to activate and partially decompose the precursor species and perhaps form new chemical species. This allows deposition at a temperature lower than thermal CVD. The process is called plasma-enhanced CVD (PECVD) (12). The plasmas are generated by direct-current, radio-frequency (r-f), or electron-cyclotron-resonance (ECR) techniques. Eigure 15 shows a parallel-plate CVD reactor that uses r-f power to generate the plasma. This type of PECVD reactor is in common use in the semiconductor industry to deposit siUcon nitride, Si N and glass (PSG) encapsulating layers a few micrometers-thick at deposition rates of 5—100 nm /min. 

Thermal CVD requires high temperature, generally from 800 to 2000°C, which can be generated by resistance heating, high-frequency induction, radiant heating, hot plate heating, or any combination of these. Thermal CVD can be divided into two basic systems known as hot-wall reactor and cold-wall reactor (these can be either horizontal or vertical). 

By comparison, all bonds other than Sn - C in the tin hydroxides are quite strong. In ClsSnOH the bond energies are 125 kcalmor 95 kcal mol and 87 kcal mol for the 0 - H, Sn - 0, and Sn - Cl bonds, respectively. Thus, it appears likely that the hydroxide ligand is quite stable and could survive transit through the thermal boundary layer in a CVD reactor and form tin oxide. 

There are many similarities between CVD reactors and processes and PVD reactors and processes. PVD reactors differ from CVD reactors in the way that they generate the deposition species and in the physical characteristics of the deposition species. In a PVD reactor, a plate of source material is used as a thermal source, rather than having a separate thermal source as in CVD reactors. 

Effect of Thermal Boundary Conditions. When the side walls are cooled instead of being insulated, there is no critical Rat number, and any transverse temperature gradient will lead to a buoyancy-driven secondary flow. Compared with the previous example (Figure 8b), the rolls are reversed and now rotate outward. These examples demonstrate the strong influence of the thermal boundary conditions on CVD reactor flows. 

The plot of growth rate in Figure 8a shows that even without buoyancy-driven secondary flows, a considerable variation in the growth rate in the transverse direction exists. The decrease in the axial velocity near the side walls leads to both a shorter thermal entrance length and a greater depletion near the walls compared with the behavior in the middle of the reactor. These perturbations from two-dimensional behavior induced by the side walls extend away from the side walls to a distance about equal to the reactor height. Thus, two-dimensional models may not be sufficient to predict CVD reactor performance even in the absence of buoyancy-driven rolls. 

The balance over the ith species (equation IV. 5) consists of contributions from diffusion, convection, and loss or production of the species in ng gas-phase reactions. The diffusion flux combines ordinary (concentration) and thermal diffusions according to the multicomponent diffusion equation (IV. 6) for an isobaric, ideal gas. Variations in the pressure induced by fluid mechanical forces are negligible in most CVD reactors therefore, pressure diffusion effects need not be considered. Forced diffusion of ions in an electrical field is important in plasma-enhanced CVD, as discussed by Hess and Graves (Chapter 8). 

In this section we will review the various types of CVD reactors scientists and engineers have used for the development of thermal CVD processes. This will be distinct from the commercial reactors used for production which will be covered in a later chapter. A similar review of reactors for development of plasma-enhanced CVD processes will be made at the end of the next chapter. We will cover the so-called cold wall systems for either single or multiple wafers first, followed by a discussion of continuous belt systems. Finally, we will review the hot wall reactor approach. 

Having covered some of the elements of plasma behavior and how it relates to reactors, it is appropriate to consider the plasma-enhanced CVD reactor specifically. This is where the plasma is created in an appropriate gas mixture so that a suitable thin film will grow on a chosen substrate. As discussed in the previous chapter, a gas mixture can react thermally, both in the gas phase and on the surface, to grow films. A similar process occurs with plasma enhancement, except that the gas mixture presented to the surface has many more species due to decomposition of the starting gas by high-energy electron impact, and there can be a high density of such species. 

In Chapters 1 and 2, we not only covered the basics of thermal and plasma-enhanced CVD, but we described the general reactor configurations that researchers have explored over the years. From these concepts have come a few production CVD reactors that satisfy the commercial needs of the integrated circuit manufacturing process. 

All of the remaining new CVD reactor systems are cold wall reactors carrying out thermal or plasma-enhanced CVD processes. They are being developed to deposit a variety of films, but each system is initially targeting a particular material. 

As there are no suitable organometallic precursors commercially available, initial work dealt with the synthesis of such a precursor [4]. 2,5-Bis(rbutyl)-2,5-diaza-l-germa-cyclopentane is a monomeric solid with a melting point of 45 °C and a sufficient vapour pressure of 0.40 mbar at 40 °C to allow its introduction into the CVD reactor. For the details about the synthesis and properties of this precursor we refer to a recent paper [4]. The present work deals with the investigation of the thermal decomposition of the precursor, the deposition of amorphous germanium (a-Ge) and the characterization of the deposited thin films. Finally some data should try to give some understanding about the deposition mechanism. 

Many different embodiments of the CVD techniqne are available for the growth of thin films, and the readers is referred to excellent recent monographs for more detail. " The last review on metal-organic precursors that appeared in this encyclopedia contained a detailed overview of thermal CVD processes, schematics of several reactors, as well as a listing of common variants of thermal CVD processes. In addition to these well-known techniques, several new CVD techniqnes have gained importance in recent years. 

In chemical vapor deposition (CVD) reactive vapor precursors react to produce solid materials in the gas phase or at the solid-gas interface on the substrate surface at appropriate temperatures. Typical precursors used in the CVD process are metal hydrides, metal chlorides, and metal organic compounds. In the case that the precursor species are metal organic compounds, the process is called metal-organic chemical vapor deposition (MOCVD). The precursor molecules are introduced into a reactor sometimes with a carrier gas and decompose by means of heat, irradiation of UV light, or electrical plasma formed in the gas. Thermal CVD is the most commonly used method. This technique has an advantage that refractory materials can be vapour-deposited at relatively low temperatures,... 

Besides the MPCVD reactors, other CVD reactors are also used for diamond deposition. They are hot filament, DC plasma, radio-frequency (rf) plasma, thermal rf plasma, plasma jet, and combustion CVD reactors. In the following, hot filament and DC plasma CVD reactors will be described, because they have been used for oriented growth of diamond. 

In Ref. [421], an HOD film of about 6-pm thickness were deposited on p-type Si by the two-step process using a custom-made CVD reactor under the conditions given in Table H.4. Then, B ions were implanted at —76°C, which was followed by a rapid thermal annealing. The B concentration was 10 /cm, the hole mobility was 80cm /V s, and the hole concentration was lO /cm at 20°C. The gauge factor K was as high as 1200 at lOOp strain at room temperature, which is far superior to 6 for polycrystalline diamond films and 550 for homoepitaxial film. 

Many of the first papers which discussed the use of (selective) CVD of tungsten for IC applications used conventional hot wall tube CVD reactors [Broadbent et al.44, Pauleau et al.45, Cheung47]. This type of reactor was and still is the workhorse in IC fabs. Excellent films such as TEOS based oxides, thermal silicon-nitride and poly-silicon can be grown in such equipment. Hot wall tube reactors are suitable for these films because such materials stick very well to quartz tubes and are quite transparent to IR radiation of the heating elements. Thus neither particle nor temperature control is a problem. One other major advantage is that high throughputs are typically obtained. 

---