Deposition Variables

The requirements of CVD films for electronic device applications have become increasingly more stringent as device size is continually reduced. Film thickness must be uniform (ideally better than 1%) across an entire wafer whose dimensions are currently 20.3 cm (8 in) for Si, and 12.7 cm (5 in) for GaAs. The structure of the film and its composition must be controlled and reproducible, both on a single wafer, as well as between wafer samples. It is also desirable that the process is safe, inexpensive, and easily automated. 

For any individual material, a number of variables determine the quality and rate of film growth. In general, the deposition rate increases with increased temperature and follows the Arrhenius equation (Eq. 5 2), where R is the deposition rate, is the activation energy, Tis the temperature (K), A is the frequency factor, and k is Boltzmann s constant (1.381 x 10 JK ). 

The necessary incorporation of dopants (e.g., phosphorus and boron doped in silicate glasses) usually lowers deposition rates, due to competitive surface binding. 

---

Until recently, most CVD operations were relatively simple and could be readily optimized experimentally by changing the reach on chemistry, the activation method, or the deposition variables until a satisfactory deposit was achieved. It is still possible to do just that and in some cases it is the most efficient way to proceed. However, many of the CVD processes are becoming increasingly complicated with much more exacting requirements, which would make the empirical approach too cumbersome. 

The deposition variables are the process parameters most suited to regulate the particle composite content within the limits set by the particle properties and plating bath composition. Particle bath concentration is the most obvious process variable to control particle codeposition. Within the limits set by the metal plating process and the practical feasibility also current density, bath agitation and temperature can be used to obtain a particular composite. Consequently the deposition process variables are the most extensively investigated parameters in composite plating. The models and mechanisms discussed in Section IV almost exclusively try to explain and model the relation between these process parameters and the particle codeposition rate. 

Before describing individual important CVD systems, it is worth outlining general considerations to be taken into account with regard to the growth by CVD of any insulating film the type of CVD method, deposition variables, and limitations of the precursor. 

TABLE 4.7.7 Influence of Depositing Variables on Diaphragm Characteristics [87,88]... 

The CVD of pyrolytic graphite can be optimized by experimentation. The carbon source (hydrocarbon gas), the method of activating the decomposition reaction (thermal, plasma, laser, etc.), and the deposition variables (temperature, pressure, gas flow, etc.) can be changed until a satisfeictory deposit is achieved. However, this empirical eipproach may be too cumbersome and, for more accurate results, it should be combined with a theoretical analysis. 

III. Individual Deposition Variability (Nose, Mouth, and Lung)... 

Quantifying the biological variability in deposition by the coefficient of variation reveals that, in healthy human subjects, coefficients of 0.38 and 0.19 were observed for 1 and 3- im-particle deposition under spontaneous-breathing conditions. Under these conditions, coefficients of variation for minnte ventilation, tidal volume, and breathing frequency were found to be about 0.2 (176,177). For controUed-breathing conditions, total deposition variability was diminished, and the respective coefficients of variation reduced to 0.27 and 0.13 (71). 

Deposition of particles of the thermodynamic size range has not been studied extensively in humans. Cheng and co-workers (185) measured deposition efficiencies for 4-, 8-, 20-, and 150-nm particles in ten healthy adult, male volunteers. For nose-in mouth-out breathing at flow rates of 333 cm s, deposition fractions for these particle sizes were 36.7 10.6 (mean SD), 21.2 8.9, 11.1 7.7, and 5.2 3.8%. However, because expiratory deposition in the mouth is not negligible for particles in the thermodynamic domain, these values overestimate nasal deposition. The large SD values indicate that intersubject deposition variability is notable in the thermodynamic domain. Mean nasal deposition, as de-... 

---