Theoretical Study of the Heats of Formation of Small Silicon-Containing Compounds

Abstract

Heats of formation for nine small silicon-containing molecules were obtained from large basis set ab initio calculations using coupled cluster theory with a perturbative treatment of triple excitations. After adjusting the atomization energies for the finite basis set truncation error, core/valence correlation, scalar relativistic, higher order correlation, and atomic spin-orbit effects, the theoretical and experimental 0 K values of ∆H f values were in good agreement. Using 106.6 kcal/mol as the heat of formation of silicon, we obtain ∆H f values of SiH ) 87.7 ( 0.4 vs 89.5 ( 0.7 (expt); SiH 2 ( 1 A 1 ) ) 64.1 ( 0.4 vs 65.5 ( 0.7 (expt); SiH 2 ( 3 B 1 ) ) 85.4 ( 0.4 vs 86.5 ( 0.7 (expt); SiH 3 ) 47.3 ( 0.5 vs 47.7 ( 1.2 (expt); SiH 4 ) 8.7 ( 0.6 vs 9.5 ( 0.5 (expt); Si 2 ) 138.8 ( 0.4 vs 139.2 (expt); Si 2 H 6 ) 19.7 ( 0.5 vs 20.9 ( 0.3 (expt); SiF ) -14.8 ( 0.4 vs -5.2 ( 3 (expt); SiF 2 ) -151.7 ( 0.5 vs -140.3 ( 3 (expt); and SiF 4 ) -384.5 ( 0.9 vs -384.9 ( 0.2 (expt). Based on the present work, we suggest a number of revisions in the interpretation of the experimental data. Although a revision in ∆H f°( Si) to 107.4 ( 0.6 kcal/mol at 0 K leads to improved agreement between theory and experiment for the Si x H y compounds, it worsens agreement for SiF 4 . Given the remaining uncertainties in the theoretical approach, more definitive conclusions do not appear to be warranted.

The correlation consistent composite approach (ccCA): An alternative to the Gaussian-n methods

“…Due to the persistent discrepancies over the ⌬H f value of silicon gas, 34,[78][79][80][81][82] seven molecules from the G2-1 test set containing silicon atoms have been removed from our initial test set. Work is currently being carried out upon small silicon-containing compounds in order to assess and improve the errors of the ccCA-CBS methods, and more quantitatively address controversy in the gas phase enthalpy of formation of silicon.…”

“…Some of these, such as the complete basis set ͑CBS-n͒ model chemistries of Petersson and co-workers, [11][12][13][14] the focal point method of Allen and co-workers [15][16][17][18][19][20] and Császár et al, 21,22 the W1 and W2 methods of Parthiban and Martin 23 and Martin and de Oliveira, 24 and the High Accuracy Extrapolated ab initio Thermochemistry ͑HEAT͒ method of Stanton and co-workers, 25,26 attempt to approach the complete basis set ͑CBS͒/full-configuration interaction ͑FCI͒ limit of smaller systems, consistently obtaining accuracy better than within 0.5 kcal mol −1 of experimental data. Dixon, Feller, and co-workers [27][28][29][30][31][32][33][34][35] used large basis set coupled cluster equilibrium geometries and total energies while using smaller basis sets to perform further electron correlation and scalar relativistic corrections, in order to achieve at least chemical accuracy ͑±1 kcal mol −1 ͒ for enthalpies of formation. Feller and Peterson 36,37 have also examined the performance of G2 methods compared to coupled cluster and perturbation theories.…”

“…52 With respect to basis set size, the poor scaling of ab initio theories that can account for high levels of electron correlation remains the limiting factor in the maximum molecular size that can be addressed by a composite method. Dixon and co-workers 34,35 have demonstrated that employing quadruple-zeta ͑the minimum "zeta level" typically necessary for satisfactory CBS extrapolations͒ coupled cluster computations on molecules of modest size ͑e.g., n-octane͒ require extraordinary computational resources. Additionally, accounting for smaller, but significant energy corrections such as higher-order electron correlation, relativistic effects, spin-orbit coupling, and anharmonic zero-point energies can be time consuming and require more extensive expertise and software.…”

Quantitative Computational Thermochemistry of Transition Metal Species

“…Ab initio techniques 1 for the calculation of chemically accurate thermodynamics (i.e., (1-2 kcal mol -1 ) for main group species have been widely investigated, for example, Gaussian-n, 2-5 the complete basis set (CBS-n) methods, [6][7][8][9] the focal point method, [10][11][12][13][14][15][16][17] Wn methods, [18][19][20][21] the HEAT method, 22,23 the composite method of Dixon, Feller, Peterson, and coworkers, [24][25][26][27][28][29][30][31][32] and the newly developed correlation consistent composite approach (ccCA). [33][34][35][36] Yet to our knowledge, no methodical thermochemical analysis of ab initio model chemistries for a series of chemically diverse TM complexes on the scale of the G2 or G3 test sets 37 has been reported, though the most ambitious to date may be represented by the study of 62 TM-containing molecules by Furche and Perdew 38 and used by Truhlar and co-workers.…”

The Molecular Structures and Energetics of Cl2CO, ClCO, Br2CO, and BrCO

“…(All bond energies calculated at 0 K. [2][3][4] Our laboratory has recently been developing a composite theoretical approach that is intended to reliably predict a variety of thermodynamic quantities, including heats of formation, without recourse to empirical parameters. 1,[5][6][7][8][9][10][11][12][13] As described below, our approach starts with existing, reliable thermodynamic values (obtained from either experiment or theory). Missing pieces of information are then computed by using high-level ab initio electronic structure methods, such as coupled cluster methods including single, double, and connected triple excitations, with the latter being handled perturbatively.…”