A Model of the Gas-Phase Chemistry of Boron Nitride CVD From BCl 3 and NH 3
- PDF / 514,127 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 79 Downloads / 190 Views
459 Mat. Res. Soc. Symp. Proc. Vol. 410 0 1996 Materials Research Society
chemical processes in various potential CVD operating regimes, we attempt to provide upper limits for the extent of gas-phase reaction by using upper limits for the rates of unknown reactions. Selected reactions used in the mechanism are listed in Table I with the corresponding heats of reaction [9] and rate coefficients. Reactions 1 and 2 were not used in the simulations (see discussion below). Reactions that describe the decomposition of NH 3 [11], H2 [12], and HC1 [13], which are not listed in Table I with the exception of Reaction 11, were obtained from the literature. Rate constants for Reactions 3-8 are unknown and were estimated by applying highpressure rate theory [14] or by comparisons with analogous reactions. Thermochemistry used here to determine bond dissociation energies (BDE), heats of reaction, and estimate reaction rate constants was obtained from ab initio calculations (performed at the level of fourth-order MollerPlesset perturbation theory with empirical bond-additivity corrections, referred to here as the BACMP4 method) reported elsewhere [9]; these calculations are typically accurate within ± 3 kcal/mol. Plug-flow calculations were performed using the SENKIN program developed at Sandia [15]. The rates of unimolecular reactions shown in Table I (Reactions 1, 2, 4, 7, and 8) typically depend on pressure. Predicting this dependence requires knowledge of intermolecular energy
transfer rates, which are difficult to determine accurately without experimental measurements of rate constants. However, unimolecular reaction rates in the high-pressure (pressure-independent) limit can be predicted by RRKM and other theoretical methods with reasonable accuracy. Thus, the rate constants used for Reactions 2-4, 7, and 8 are those predicted by RRKM or estimated for the high-pressure limit of these reactions, providing an upper limit for the rate. Table 1: Gas-phase mechanism for BN CVD: selected reactions. All energies are in kcal mo1- 1. Reaction
A i-reacf
1. C13 BNH 3 BN(bulk) + 2HCI
463
(12) (13) (14) (15)
Alternatively, the surface reactivity of the various gas-phase species may be sufficiently similar that no variation in mechanism is observed as a function of temperature, even though the gas-phase composition varies considerably. In either case, the mechanism must be consistent with the observation [1, 2] that BN deposition is first-order in the BCI 3 concentration and independent of the NH 3 concentration. Since NH 3 was present in excess in the experiments of both P&L and Lee et al., the gas-phase reactions should be pseudo-first order. The deposition rate derived from Reactions 13-15 is also independent of the NH 3 concentration, if the concentrations of BC13 (surf) and Cl 2BNH 2 (surf) are at steady-state (note that Reactions 13-15 are not unique; other mechanisms can be imagined that give the same reaction orders). Thus, the two hypotheses appear to be equally plausible. CONCLUSIONS We conclude that gas-phase chemical
Data Loading...
