MODELING OXIDATION AND HYDROLYSIS REACTIONS IN SUPERCRITICAL WATER—FREE RADICAL ELEMENTARY REACTION NETWORKS AND THEIR APPLICATIONS

ABSTRACT

From the beginning of supercritical water oxidation (SCWO) research in the early 1980s, mathematical models have been used to correlate, predict, and explain experimental reaction kinetic data. Initially, these were simple global rate laws, involving only a single overall reaction or a few reactions, each with its own arbitrary rate law. As computational power increased and the library of elementary reactions from the combustion literature grew, it became feasible to construct elementary reaction rate models, which are a more accurate means of representing the SCWO process. Early efforts to construct elementary reaction rate models resulted in very poor agreement with experimental data unless model parameters were adjusted to optimize the fit. However, today with considerably more computing power and a more robust collection of elementary rate parameters from the combustion literature, rate predictions in the relatively low-temperature, high-pressure SCWO environment are more effective and accurate. These enhancements make it possible to build models that hold the predictive capacity to help guide experimental design and gain a greater mechanistic understanding. This paper details current best practices for the construction of these elementary reaction rate models, including selection of a base model from the combustion literature, identification of possible intermediate compounds, and estimation of unknown rate constants by ab initio calculations or analogy to known chemistry. A set of model compounds was selected to illustrate rate modeling approaches for both oxidation and hydrolysis pathways.

Keywords:

• kinetics
• modeling
• supercritical water oxidation
• elementary reaction mechanisms

We gratefully acknowledge financial support of this research by the U.S. Army Research Office (Grant Nos. DAAL03-92-G-0177 and DAAD19-99-1-0211); Shell, Inc.; the National Aeronautics and Space Administration; The Martin Foundation; and the National Science Foundation. We thank Dr. Robert Shaw of the ARO for his interest in this research. We thank Dr. Steven Rice of Sandia National Laboratories and Prof. Joseph Bozzelli of the New Jersey Institute of Technology for their continue collaboration in our modeling efforts. We thank Prof. Ken Smith, Prof. Jack B. Howard, Prof. Ronald Latanision, Prof. Greg McRae, Dr. Bill Peters, Dr. Michael Modell, Dr. Menner Tatang, Dr. Brian Phenix, Prof. Paul Webley, Dr. Rick Holgate, Dr. Phil Marrone, Dr. Fred Vogel, Prof. Jeff Steinfeld, and other present and former members of the MIT supercritical fluids research group for their advice and interest in our research.

Notes

At 250 bar, where k = A * exp(− E _a /RT) * [Organic]^a * [O₂]^b * [H₂O]^c, in units of kJ, mol, L, s. Confidence intervals are at the 95% level.

At 250 bar, where k = A * exp(− E _a /RT) * [Organic]^a * [H₂O]^c, in units of kJ, mol, L, s. Confidence intervals are at the 95% level.

^a Mechanism of Yetter et al. (1991) adapted to high-pressure.

^b k = AT ⁿ exp(− E _a /RT) with units of cm³, mol, s, K.

^c UF = multiplicative uncertainty factor (see text for definition).

^d (Cobos et al., 1985).

^e (Baulch et al., 1992).

^f (Atkinson et al., 1989).

The net flux for all reactions shown are in the forward direction.