arXiv subjects

Jan M. L. Martin

Publications and source records attributed to Jan M. L. Martin.

95 records · Page 6Linked to original sources

arXiv1999

A definitive heat of vaporization of silicon through benchmark ab initio calculations on SiF_4

In order to resolve a significant uncertainty in the heat of vaporization of silicon -- a fundamental parameter in gas-phase thermochemistry -- $ΔH^\circ_{f,0}$[Si(g)] has been determined from a thermochemical cycle involving the precisely known experimental heats of formation of SiF_4(g) and F(g) and a benchmark calculation of the total atomization energy (TAE_0) of SiF_4 using coupled-cluster methods. Basis sets up to $[8s7p6d4f2g1h]$ on Si and $[7s6p5d4f3g2h]$ on F have been employed, and extrapolations for residual basis set incompleteness applied. The contributions of inner-shell correlation (-0.08 kcal/mol), scalar relativistic effects (-1.88 kcal/mol), atomic spin-orbit splitting (-1.97 kcal/mol), and anharmonicity in the zero-point energy (+0.04 kcal/mol) have all been explicitly accounted for. Our benchmark TAE_0=565.89 \pm 0.22 kcal/mol leads to $ΔH^\circ_{f,0}$[Si(g)]=107.15 \pm 0.38 kcal/mol ($ΔH^\circ_{f,298}$[Si(g)]=108.19 \pm 0.38 kcal/mol): between the JANAF/CODATA value of 106.5 \pm 1.9 kcal/mol and the revised value proposed by Grev and Schaefer [J. Chem. Phys. 97, 8389 (1992}], 108.1 \pm 0.5 kcal/mol. The revision will be relevant for future computational studies on heats of formation of silicon compounds.

physics.chem-ph↗

arXiv1999

The ground-state spectroscopic constants of Be_2 revisited

Extensive ab initio calibration calculations combined with extrapolations towards the infinite-basis limit lead to a ground-state dissociation energy of Be_2, D_e=944 \pm 25 1/cm, substantially higher than the accepted experimental value, and confirming recent theoretical findings. Our best computed spectroscopic observables (expt. values in parameters) are G(1)-G(0)=223.7 (223.8), G(2)-G(1)=173.8 (169 \pm 3), G(3)-G(2)=125.4 (122 \pm 3), and B_0=0.6086 (0.609) 1/cm; revised spectroscopic constants are proposed. Multireference calculations based on a full valence CAS(4/8) reference suffer from an unbalanced description of angular correlation; for the utmost accuracy, a CAS(4/16) reference including the $(3s,3p)$ orbitals is required, while for less accurate work a CAS(4/4) reference is recommended. The quality of computed coupled cluster results depends crucially on the description of connected triple excitations; the CC5SD(T) method yields unusually good results because of an error compensation.

physics.chem-ph↗

arXiv1998

Ab initio thermochemistry beyond chemical accuracy for first-and second-row compounds

By judicious use of extrapolations to the 1-particle basis set limit and $n$-particle calibration techniques, total atomization energies of molecules with up to four heavy atoms can be obtained with calibration accuracy (1 kJ/mol or better, on average) without any empirical correction. For the SCF energy a 3-point geometric extrapolation is the method of choice. For the MP2 correlation energy, a 2-point $A+B/(l+1/2)^3$ extrapolation is recommended, while for CCSD and CCSD(T) correlation energies we prefer the 3-point $A+B/(l+1/2)^C$ formula. Addition of high-exponent `inner polarization functions' to second-row atoms is essential for reliable results. For the highest accuracy, accounts are required of inner-shell correlation, atomic spin-orbit splitting, anharmonicity in the zero-point energy, and scalar relativistic effects.

physics.chem-ph↗

arXiv1998

A fully ab initio quartic force field of spectroscopic quality for SO_3

The quartic force field of SO$_3$ was computed fully ab initio using coupled cluster (CCSD(T)) methods and basis sets of up to $spdfgh$ quality. The effect of inner-shell correlation was taken into account. The addition of tight $d$ functions is found to be essential for accurate geometries and harmonic frequencies. The equilibrium geometry and vibrational fundamentals are reproduced to within 0.0003 Åand (on average) 1.15 cm^{-1}, respectively. We recommend the following revised values for the harmonic frequencies: $ω_1 = 1082.7, ω_2 = 502.6, ω_3 = 1415.4, ω_4 = 534.0 cm^{-1}$. In addition, we have shown that the addition of inner polarization functions to second-row elements is highly desirable even with more approximate methods like B3LYP, and greatly improves the quality of computed geometries and harmonic frequencies of second-row compounds at negligible extra computational cost. For larger such molecules, the B3LYP/VTZ+1 level of theory should be a very good compromise between accuracy and computational cost.

physics.chem-ph↗

arXiv1998

Benchmark ab initio thermochemistry of the isomers of diimide, N_2H_2, using accurate computed structures and anharmonic force fields

A benchmark ab initio study on the thermochemistry of the trans-HNNH, cis-HNNH, and H$_2$NN isomers of diazene has been carried out using the CCSD(T) coupled cluster method, basis sets as large as $[7s6p5d4f3g2h/5s4p3d2f1g]$, and extrapolations towards the 1-particle basis set limit. The effects on inner-shell correlation and of anharmonicity in the zero-point energy were taken into account: accurate geometries and anharmonic force fields were thus obtained as by-products. Our best computed $ΔH^\circ_{f,0}$ for trans-HNNH, 49.2 \pm 0.3 kcal/mol, is in very good agreement with a recent experimental lower limit of 48.8 \pm 0.5 kcal/mol. CCSD(T) basis set limit values for the isomerization energies at 0 K are 5.2 \pm 0.2 kcal/mol (cis-trans) and 24.1 \pm 0.2 kcal/mol (iso-trans). Our best computed geometry for trans-HNNH, $r_e$(NN)=1.2468 Å, $r_e$(NH)=1.0283 Å, and $θ_e$=106.17$^\circ$, reproduces the precisely known ground-state rotational constants of trans-HNNH to within better than 0.1 %. The rotation-vibration spectra of both cis-HNNH and H$_2$NN are dominated by very strong Coriolis and Fermi resonances. In addition, the NH stretches in H$_2$NN are so strongly anharmonic that vibrational perturbation theory breaks down, and the molecule appears to be an excellent test case for variational treatments of the vibrational Schrödinger equation.

physics.chem-ph↗