CVD Processes and Equipment

Like all chemical reactions, CVD reactions require activation energy to proceed. This energy can be provided, in practice, by several methods. Thermal activation is the original process, and it is still the major method for the chemical vapor deposition of metals and ceramics.

In thermal CVD, the reaction is activated by high temperature, generally above 900 °C (1650 °F) (Ref 6). A typical thermal CVD apparatus consists of three interrelated components: the reactant-gas supply system; the deposition chamber, or reactor; and the exhaust system (Fig. 1). A fourth component that is often used is a closed-loop process-control monitor, which is now available in a PC-based design.

Fig. 1 Thermal CVD reactor

Plasma CVD is a method that operates at lower temperatures than thermal CVD. The reaction is activated by a plasma at temperatures between 300 and 700 °C (570 to 1290 °F). The process was developed because the high deposition temperature of thermal CVD precludes the use of many substrates, such as low-melting-point metals; materials that undergo solid-state phase transformation over the range of deposition temperatures; polymers; and others. In addition, large mismatches in the thermal expansion of a substrate and a coating will generate stresses that can lead to cracking and delamination or spalling during cooling (Ref 7, 8). In the plasma CVD process, the stress that is due to thermal-expansion mismatch is reduced, and temperature-sensitive substrates can be more readily coated. Table 1 compares the deposition temperatures for thermal and plasma CVD for several commercially important coatings.

Table 1 Typical deposition temperatures for thermal and plasma chemical vapor deposition

Thermal CVD

Plasma CVD

°C

°F

°C

°F

Silicon nitride

900

1650

300

570

Silicon dioxide

800-1100

1470-2010

300

570

Titanium carbide

900-1100

1650-2010

500

930

Titanium nitride

900-1100

1650-2010

500

930

Tungsten carbide

1000

1830

325-525

615-975

Plasma CVD was initially developed in the 1960s for use in the semiconductor industry, but applications have been expanding ever since and are now common in the nonsemiconductor applications discussed in this article. Most plasma CVD systems use radio frequency (RF) with operating frequencies of 450 KHz or 113.56 MHz. A typical RF reactor with parallel electrodes is shown in Fig. 2. Microwave glow discharge is also used at a standard frequency of 2.45 GHz.

Fig. 2 Radio-frequency plasma CVD reactor configured for deposition on silicon wafers

A recent and promising development in the production of plasma is based on electron cyclotron resonance (ECR) and the proper combination of an electric field and a magnetic field (Ref 9). Cyclotron resonance is achieved when the frequency of the alternating electric field matches the natural frequency of the electrons orbiting the lines of force of the magnetic field. An ECR plasma reactor is shown schematically in Fig. 3. ECR and other plasma techniques are used extensively in semiconductor production but so far have remained mostly experimental in other areas of application (Ref 10).

Ecr Plasma
Fig. 3 Microwave/electron cyclotron resonance (ECR) plasma CVD reactor

Laser CVD. Two other activation methods based on a laser have recently been developed (Ref 8, 11). As of the mid-1990s, the thermal-laser and photo-laser CVD methods are still essentially in the experimental stage, but have great potential, at least in specialized areas. The materials that can be deposited include oxides, nitrides, tungsten, aluminum, and others.

Thermal-laser CVD (Ref 12), or laser pyrolysis, occurs when the laser thermal energy contacts and, thereby, heats an absorbing substrate. The wavelength of the laser can be such that little or no energy is absorbed by gas molecules. Because the substrate is locally heated, deposition is restricted to the heated area. Figure 4 illustrates the deposition of a thin stripe by moving a laser beam linearly across the substrate.

Cvd Grwoth
Fig. 4 Thermal-laser CVD growth mechanism

In photo-laser CVD, the chemical reaction is induced by the action of light, specifically ultraviolet (UV) radiation, which has sufficient photon energy to break the chemical bonds in the reactant molecules. In many cases, these molecules have a broad electronic absorption band and are readily excited by UV radiation. Although UV lamps have been used, more energy can be obtained from UV lasers, such as the excimer (e.g., excited dimer) lasers with photon energies ranging from 3.4 eV (XeFlaser) to 6.4 eV (ArFlaser). A typical photo-laser CVD system is shown in Fig. 5.

Fig. 5 Photo-laser CVD apparatus

Photo-laser CVD differs from thermal-laser CVD in that it does not require heat, because the reaction is photon-activated, and the deposition essentially occurs at room temperature. Moreover, there is no constraint on the type of substrate that can be used. It can be opaque, absorbent, transparent, or even temperature-sensitive.

A limitation of this method that has, to date, restricted its application is a slow deposition rate. If higher-power excimer lasers can be made more economical, then the process could compete with thermal CVD and thermal-laser CVD, particularly in critical applications where low temperature is essential.

Closed-Reactor CVD or Pack Cementation. The CVD systems described above use open reactors, in which reactants are introduced continuously and flow through the reactor (Ref 1). Another important system utilizes a closed reactor. The chemical vapor deposition in such a system is also known as pack cementation (Ref 13).

The entire process is carried out isothermally, because the driving mechanism for the reaction is not a difference in temperature, as in thermal CVD, but rather a difference in chemical activity between a metal in the free state and a metal in solution with another metal. A common reaction involves coating iron objects (such as turbine blades) with chromium, using chromium powder and ammonium iodide as reactants and aluminum oxide as an inert filler. Parts and chemicals are loaded in a molybdenum container that is then sealed, as shown schematically in Fig. 6. Pack cementation is a common industrial process with large-scale applications in chromizing, aluminizing, and siliconizing.

Fig. 6 Pack-cementation chromizing/siliconizing apparatus. Pack material composed of 3 wt% Cr, 11 wt% Si, 0.25 wt% NH4I, and balance, Al2O3

Chemical vapor infiltration (CVI) refers to the particular CVD process in which gaseous reactants infiltrate a porous structure, such as an inorganic open foam or a fiber array. Deposition occurs on the foam or fiber, and the structure is gradually densified to form a composite (Ref 14).

In a typical CVI system (Fig. 7), both the gas inlet and substrate are water cooled, and only the top of the substrate is heated. Under pressure, the gaseous precursors enter the cool side of the substrate and flow through it to reach the hot zone, where the deposition reaction occurs.

Carbon Vapor Infiltration
Fig. 7 Chemical vapor infiltration apparatus. Source: Ref 15

This process is used to produce high-strength silicon carbide and carbon-carbon composites, as well as other reinforced metal or ceramic composites. As contrasted with sintering, CVI does not require high pressure, and the processing temperatures are lower. As a result, mechanical and chemical damage to the substrate is minimized.

The major limitation of this method is the necessity for the interdiffusion of reactants and byproducts through relatively long and narrow channels. Chemical vapor infiltration is a slow process that can take several weeks. Full densification is nearly impossible to obtain because of the formation of closed porosity.

Metal-organic CVD (MOCVD) is a specialized process that utilizes organometallic compounds as precursors, usually in combination with hydrides or other reactants. Most MOCVD reactions occur at temperatures between 600 and 1000 °C (1110 and 1830 °F). When the most precise controls and high-purity gases are used, extremely thin deposits (<10 nm, or 0.4 pin.) with abrupt interfaces (<1 nm, or 0.04 pin.) can be produced. Because MOCVD equipment and chemicals are expensive and production costs are high, this method is considered most often when high performance is essential or when substrates are temperature sensitive.

The MOCVD method is being used extensively in microwave and optoelectronic applications, as discussed in detail in the article "Chemical Vapor Deposition of Semiconductor Materials" in this Section of the Volume. In addition, it is being introduced in wear and corrosion applications. An example of a nonsemiconductor application of MOCVD is the deposition of iridium via the decomposition of iridium acetylacetonate for oxidation protection (up to 2200 °C, or 3990 °F) of the maneuvering thruster nozzle of a spacecraft satellite (Ref 16). Additional information on the use of MOCVD in nonsemiconductor applications is provided in Ref 8.

References cited in this section

1. H.O. Pierson, Handbook of Chemical Vapor Deposition, Noyes Publications, 1992

6. T.M. Besmann et al., Chemical Vapor Deposition Techniques, MRS Bulletin, Nov 1988, p 45-50

7. P.K. Bachman, G. Gartner, and H. Lydtin, Plasma Assisted CVD Processes, MRS Bulletin, Dec 1988, p 5159

8. D. Bhat, Techniques of Chemical Vapor Deposition, Surface Modification Technologies--An Engineer's

0 0

Post a comment