Hot-wall effect

Note, in life expectancy calculations of condenser tube materials, that the test data on a specimen do not ordinarily reflect the hot-wall effect. Economic evaluation of any equipment should not be based on any single portion of this equipment but on a balanced investigation of the whole system. 

The temperature distribution has a characteristic maximum within the liquid domain, which is located in the vicinity of the evaporation front. Such a maximum results from two opposite factors (1) heat transfer from the hot wall to the liquid, and (2) heat removal due to the liquid evaporation at the evaporation front. The pressure drops monotonically in both domains and there is a pressure jump at the evaporation front due to the surface tension and phase change effect on the liquid-vapor interface. 

From an examination of Equation 8.1, it can be seen that several things can be done to improve the heat-transfer rate. Quite often the simplest approach is to increase the temperature differential, by using higher-pressure steam or a hot oil supply. In some cases this may have adverse effects, for example a very hot wall temperature may lead to fouling, or, worse, initiate unwanted reactions. This is likely to be more pronounced in cases where mass transfer is poor. In some instances this practice may... 

Closely related to the superheating effect under atmospheric pressure are wall effects, more specifically the elimination of wall effects caused by inverted temperature gradients (Fig. 2.6). With microwave heating, the surface of the wall is generally not heated since the energy is dissipated inside the bulk liquid. Therefore, the temperature at the inner surface of the reactor wall is lower than that of the bulk liquid. It can be assumed that while in a conventional oil-bath experiment (hot vessel surface, Fig. 2.6) temperature-sensitive species, for example catalysts, may decompose at the hot reactor surface (wall effects), the elimination of such a hot surface will increase the lifetime of the catalyst and therefore will lead to better conversions in a microwave-heated as compared to a conventionally heated process. 

The radicals were generated in the photolysis of the appropriate alkyl iodide in the presence of excess C02 to minimize hot radical and wall effects. The analysis was identical to that of Christie s work with CH3I, previously described (Sect. VII-B). The ratios found for k2i/ (k22 + k23) were 7, 11, and 22, respectively, for C2H5, n-C3H7, and z-C3H7 radicals. Lower limits for reaction (24) are known, and thus minimum values for k22 + k23 could be estimated (Table 7-3). 

Numerous modeling studies of CVD reactors have been made and are summarized in recent review papers (I, 212). Table 3 in reference 212 lists major examples of CVD models up to mid-1986. Therefore, rather than giving an exhaustive list of previous work, Table V presents a summary of the major modeling approaches and forms the basis for the ensuing discussion, which is most appropriately handled in terms of two groups (1) hot-wall LPCVD systems and (2) cold-wall, near-atmospheric-pressure reactors. In LPCVD reactors, diffusion and surface reaction effects dominate, whereas in cold-wall reactors operated at near-atmospheric pressures, fluid flow and gas-phase reactions are important in predicting performance, as discussed earlier in relation to transport phenomena. 

Another approach to this problem involves heating the wafer at 750 F at very low pressures (<10 10 Torr prior to deposition.28 This has the effect of removing the native oxide by evaporation of SiO. Depositions were achieved in the temperature range of 750° to 850°C in SiH4 + H2. Since the authors were developing a hot-wall system with many wafers stacked close to each other, the deposition was carried out at 2 mTorr. Deposition rates of 20 to 45 A/min were achieved. As expected, dopant transition widths were very narrow, several hundred angstroms. Again, device studies on such a system have not yet been done. 

Analyze the arrangement to assess the type(s) of heat transfer involved. The distance separating the hot and cold surfaces is small compared with the size of the surfaces. The approximation can thus be made that the furnace wall, the dense network of cryogenic piping, and the radiation shields are all infinitely extended parallel planes. This is a conservative assumption, since the effect of proximity to an edge is to introduce a source of moderate temperature, thus allowing the hot wall to cool off. Convection is omitted with the same justification. So the problem can be treated as pure radiation. 

There appear to have been few modelling efforts for hot-wall LPCVD reactors. Gieske et al. (28) and Hitchman et al. (29) present experimental data and discuss flow fields, mass transfer effects, and possible kinetics in rather general terms. A recent model by Kuiper et al. (30) cannot account for diffusion in the spaces between the wafers and the significant volume expansion commonly associated with LPCVD processes. Furthermore, it is restricted to isothermal conditions and plug flow in the main flow region in spite of the large diffusivities associated with LPCVD. 

There have been attempts to explain this low activation energy at low temperatures in terms of diffusion of radicals to the wall and heterogeneous reaction of CHg -4- acetone, the subseejuent photolysis of accumulated biacetyl (which Noyes has calculated to be insufficient), the reaction 6, and possible hot radical effects. None of these has yielded to quantitative analysis, although A. J. Nicholson, J. Avi. Chem. Soc.y 73, 3981 (1951), has shown that diffusion of CH3 radicals to the walls may become important at low intensities, low acetone pressures, or low temperatures. The fa(5t that I2 does not completely quench formation of CH4 is an indication that hot radical effects may be important. 

The effect of U on the retention of polystyrene in ethylbenzene is typical of most polymer-solvent systems. Therefore, fluctuations in T, of only a couple degrees are generally not a problem. Larger fluctuations can be a significant problem, however, especially when retention is used to monitor small batch-to-batch variations in a quality control situation. The magnitude of U depends on several factors. Heat, which is transferred from the hot wall, is typically removed by heat exchange with water flowing beneath the cold wall. 

---