Journal of Chemical Physics, Vol.117, No.24, 11175-11187, 2002
Heats of formation of phosphorus compounds determined by current methods of computational quantum chemistry
The heats of formation of a range of phosphorus containing molecules (P-2, P-4, PH, PH2, PH3, P2H2, P2H4, PO, PO2, PO3, P2O, P2O2, HPO, HPOH, H2POH, H3PO, HOPO, and HOPO2) have been determined by high level quantum chemical calculations. The equilibrium geometries and vibrational frequencies were computed via density functional theory, utilizing the B3LYP/6-31G(2df,p) functional and basis set. Atomization energies were obtained by the application of ab initio coupled cluster theory with single and double excitations from (spin)-restricted Hartree-Fock reference states with perturbative correction for triples [CCSD(T)], in conjunction with cc-pVnZ basis sets (n = T, Q, 5) which include an extra d function on the phosphorus atoms and diffuse functions on the oxygens, as recommended by Bauschlicher [J. Phys. Chem. A 103, 11126 (1999)]. The valence correlated atomization energies were extrapolated to the complete basis limit and corrected for core-valence (CV) correlation and scalar relativistic effects, as well as for basis set superposition errors (BSSE) in the CV terms. This methodology is effectively the same as the one adopted by Bauschlicher in his study of PO, PO2, PO3, HPO, HOPO, and HOPO2. Consequently, for these molecules the results of this work closely match Bauschlicher's computed values. The theoretical heats of formation, whose accuracy is estimated as ranging from +/-1.0 to +/-2.5 kcal mol(-1), are consistent with the available experimental data. The current set of theoretical data represent a convenient benchmark, against which the results of other computational procedures, such as G3, G3X, and G3X2, can be compared. Despite the fact that G3X2 [which is an approximation to the quadratic CI procedure QCISD(T,Full)/G3Xlarge] is a formally higher level theory than G3X, the heats of formation obtained by these two methods are found to be of comparable accuracy. Both reproduce the benchmark heats of formation on the average to within +/-2 kcal mol(-1) and, for these molecules at least, they are superior to the basic G3 method. The performance of G3X2 is further improved, however, by the incorporation of BSSE corrections in the CV component of the energies. All the G3n methods have difficulties, however, with molecules which have multiple or highly strained P-P bonds, such as P-2 and P-4.