I.A.1.

Summary of the CCCBDB

The NIST Computational Chemistry Comparison and Benchmark Database is a collection of experimental and ab initio thermochemical properties for a selected set of molecules.

The goals are:

1) Provide a benchmark set of molecules for the evaluation of ab initio computational methods.

2) Allow the comparison between different ab initio computational methods for the prediction of thermochemical properties.

The thermochemical values with error bars included in the CCCBDB are:

1. Enthalpies of formation.

2. Entropies, heat corrections (integrated heat capacity).

3. Data needed to compute thermochemical properties, such as geometries, rotational constants, vibrational frequencies, barriers to internal rotation, and electronic energy levels.

4. Additional computed properties such as atomic charges, electric dipole moments, quadrupole moments, polarizabilities, and HOMO-LUMO gaps. See the index (Section XIX. Index of Properties) for more properties.

The calculations are ongoing. Over 250000 have been completed. Model chemistries included so far:

Hartree Fock methods	HF, ROHF
Moller Plesset perturbation methods	MP2, MP2=FULL, MP3, MP4
partial configuration interaction	CID, CISD
quadratic configuration interaction	QCISD, QCISD(T)
coupled cluster methods	CCD, CCSD, CCSD(T), CCSD(T)=FULL
Density functional methods	BLYP, B1B95, B3LYP, B3PW91, MPW1PW91, PBE
semiempirical methods	AM1, PM3
composite methods	G1, G2, G2MP2, G3, G3B3, CBS-Q
Basis Sets
STO-3G, 3-21G, 3-21G*, 6-31G, 6-31G*, 6-311G*, 6-311G**, 6-31G**, 6-31+G**, 6-31G(2df,p), 6-311+G(3df,2p)
cc-pVDZ, cc-pVTZ, cc-pVQZ, aug-cc-pVDZ, aug-cc-pVTZ, aug-cc-pVQZ, cc-pCVTZ, cc-pV(T+d)Z
CEP-31G, CEP-31G*, CEP-121G, CEP-121G*, LANL2SZ, SDD

The molecules to be included have been selected from previous collections, such as the G2 data set [Larry A. Curtiss, Krishnan Raghavachari, Gary W. Trucks et al., JCP 94 (11), 7221 (1991)] and the G2 extended data set [Larry A. Curtiss, Krishnan Raghavachari, Paul C. Redfern et al., J. Chem. Phys. 106 (3), 1063 - 1079 (1997)]. This set has been augmented from the NIST Chemistry Webbook using the following criteria:

The error bars on the Enthalpy of formation are less than or equal to 8 kJ/mol.

The species contains no atoms with atomic number greater than 17 (chlorine).

The species contains six or fewer heavy atoms and twenty or fewer total atoms. This allows species as large as hexane.

The above constraints have been loosened to include some larger species (substituted benzenes, cyclooctatetraene, S₈, naphthalene, adamantane) and some Br-containing and Se-containing molecules, as well as a few Cu-containing molecules. See sections I.B.1.b or I.B.1.c for lists of the molecules sorted by number of atoms or by number of heavy atoms.

The experimental thermochemical data have been obtained primarily from the following sources:

1. CODATA [J. D. Cox, D. D. Wagman, and V. A. Medvedev, CODATA Key values for Thermodynamics (Hemisphere, New York, 1989)] This is a short list of species with internationally accepted values for enthalpies of formation, entropies and heat corrections. Provides error bars for all three properties.
2. JANAF [M. W. Chase, Jr., C. A. Davies, J. R. Downey, Jr. et al., Journal of Physical and Chemical Reference Data 14 Supplement No. 1 (1985)] This is an evaluated list of mostly inorganic species. Provides error bars for enthalpies and entropies.
3. Gurvich [L. V. Gurvich, I. V. Veyts, and C. B. Alcock, Thermodynamic Properties of Individual Substances (Hemisphere Publishing Corporation, New York, New York, 1989)] Evaluated. Small organics. Error bars for enthalpy.
4. TRC [Michael Frenkel, G. J. Kabo, K. N. Marsh et al., Thermodynamics of Organic Compounds in the Gas State (Thermodynamics Research Center, College Station, Texas, 1994)] Evaluated. No error bars.
5. Other sources on individual species where noted.

General discrepancy between measured and calculated data.

1. Enthalpies of formation are not measured. The properties determined experimentally are usually Enthalpies of reaction at some temperature above 0 K. The ab initio methods provide an absolute energy (relative to compete ionization of all atoms) at 0 K. To convert from one temperature to another, one uses the integrated heat capacity (heat correction). Ideally the procedure and supporting data used to convert from the experimental measurement to a Enthalpy of formation would be available in the database. This is true for the ab initio determinations as well, where the simplest property to calculate is the atomization energy of a molecule at 0 K.

Interconnectedness of data.

1. Geometries are needed to determine rotational constants, which are needed for heat corrections and entropies. Geometries are also useful as starting input for ab initio calculations. Experimental geometries can be compared with ab initio geometries to evaluate ab initio methods.

2. Vibrational frequencies are needed for heat corrections and entropies.

3. Some Enthalpies of formation are derived from measured equilibrium, this derivation requires an entropy. If the vibrational frequencies come from ab initio methods, the Enthalpy of formation is not purely an experimental measurement.

Lack of experimental data

1. The integrated heat capacity, or heat correction, is used to convert the enthalpy of formation from one temperature to another. It is calculated using standard statistical mechanics, using information such as mass, rotational constants, vibrational energy levels, and electronic energy levels. These have been experimentally determined in some cases, but not all. Ab initio methods can provide rotational constants and vibrational energy levels. Many of the errors for the integrated heat capacity obtained from ab initio calculations come from ignoring spin orbit splitting in open-shell linear molecules or ignoring low lying electronic states and errors in the vibrational energy levels for non-harmonic vibrations. They can amount to ≈1kJ/mol. The errors in entropies obtained from calculations arise from the same sources and can amount to ≈10 J/mol K (so @ 300 K ≈3kJ/mol)

Send comments, corrections, additions and suggestions to cccbdb@nist.gov