Thermochemistry of Aluminum Species for Combustion Modeling from Ab Initio Molecular Orbital Calculations MARK T. SWIHART Department of Chemical Engineering, University at Buffalo (SUNY), Buffalo, NY 14260-4200, USA and LAURENT CATOIRE Laboratoire de Combustion et Syste `mes Re ´actifs (LCSR), CNRS, and University of Orleans, 1C, av. de la Recherche Scientifique, 45071 Orleans Cedex 2, France High accuracy ab initio methods for computational thermochemistry have been applied to aluminum compounds expected to be present during combustion of aluminum particles. The computed enthalpies of formation at 298.15 K agree well with experimental values from the literature for AlCl, AlCl 3 , AlO, AlOAl, linear OAlO, planar Al 2 O 2 , AlOH, AlH, and AlN. The agreement is fair for AlCl 2 . Major revisions to the recommended thermochemistry must be considered for OAlCl, OAlH, OAlOH, and AlC. This is not surprising since the thermodynamic data for OAlCl, OAlH, OAlOH, and AlC are given in the literature as rough estimates. Calculated thermochemical data are also presented for several species never studied experimentally, including AlH 2 , AlH 3 , AlOO, cyclic-AlO 2 , linear AlOAlO, AlHCl, AlHCl 2 , and others. Based on the performance of the CBS-Q and G2 methods observed in other systems, the calculated enthalpies of formation would be expected to be accurate to within 1 to 2 kcal mol -1 . However, relatively large differences between the results from the CBS-Q and G2 methods for the aluminum oxides indicate that the uncertainties are slightly larger for these compounds. The thermochemistry proposed here is shown to predict substantially different equilibrium composition from the thermochemistry previously available in the literature. © 2000 by The Combustion Institute INTRODUCTION The addition of aluminum particles to solid propellant is principally used to increase motor- specific impulse. This is due to the high heat of combustion of Al with various oxidizers that are encountered in practical applications, including CO 2 ,H 2 O, and HCl. Numerous studies of alu- minum combustion have been published. Most of these studies have considered the ignition and global combustion of single particles, pow- ders, or wires in various controlled environ- ments [1–7]. Much of this work was reviewed by Price [8]. The first numerical models of the combustion of aluminum particles neglected the finite rate chemistry of this combustion. The chemical reaction rates were assumed to be infinitely large [9 –13]. However, more recent analyses of the process have provided models that include finite chemical reaction rates [14, 15]. The com- bustion process involved is complicated and therefore not easily modeled. Since aluminum burns as a vapor, the first step of this process involves gas-phase reactions between Al and the oxidizers, or between aluminum and reaction intermediates or reaction products. Reactions are also expected to occur on the surface of the aluminum particles, and this heterogeneous chemistry must also be included in a complete model. Under typical conditions, most of the reactions in the gas phase are in chemical equilibrium. For these conditions, the accuracy of the predictions of the gas-phase composition and temperature is determined by the accuracy of the thermochemical data for the gas-phase species. Existing thermochemical data for sev- eral species are highly uncertain, as underlined by Fontijn [16]. For some other species ex- pected to be present, no thermochemical data are available in the literature. As a conse- quence, some of these species have never been introduced into kinetic models. Such species must be considered in detailed kinetic models from which reduced models are to be derived for engineering purposes. *Corresponding author. E-mail: swihart@eng.buffalo.edu COMBUSTION AND FLAME 121:210 –222 (2000) 0010-2180/00/$–see front matter © 2000 by The Combustion Institute PII S0010-2180(99)00128-5 Published by Elsevier Science Inc.