Quantum mechanically determined electrostatic potentials for isosurfaces of electron density of a variety of CHNO explosive molecules are analyzed to identify features that are indicative of sensitivity to impact. This paper describes the development of models for prediction of impact sensitivity of CHNO explosives using approximations to the electrostatic potentials at bond midpoints, statistical parameters of these surface potentials, and the generalized interaction properties function [J. S. Murray, T. Brinck, P. Lane, K. Paulsen and P. Politzer, J. Mol. Struct (THEOCHEM) 1994, 307, 55] or calculated heats of detonation. The models are parametrized using a set of 34 polynitroaromatic and benzofuroxan explosives for which impact sensitivity measurements exist. The models are then applied to a test set of 15 CHNO explosives from a variety of chemical families in order to assess the predictive capability of the models. Patterns of the surface potentials of the molecules examined in this study suggest that the level of sensitivity to impact is related to the degree of positive charge buildup over covalent bonds within the inner framework of these explosives. The highly sensitive explosives show large positive charge buildup localized over covalent bonding regions of the molecular structures, whereas the insensitive explosives do not exhibit this feature. For the nitroaromatic and benzofuroxan compounds, sensitivity appears to be related to the degree and distribution of positive charge build-up localized over the aromatic ring or over the C−NO2 bonds.

We present simple atom and group-equivalent methods that will convert quantum mechanical energies of molecules to gas phase heats of formation of CHNO systems. In addition, we predict heats of sublimation and vaporization derived from information obtained from the quantum-mechanically calculated electrostatic potential of each isolated molecule. The heats of sublimation and vaporization are combined with the aforementioned gas phase heats of formation to produce completely predicted condensed phase heats of formation. These semiempirical computational methods, calibrated using experimental information, were applied to a series of CHNO molecules for which no experimental information was used in the development of the methods. These methods improve upon an earlier effort of Rice et al. [Rice, B. M.; Pai, S. V.; Hare, J. Combust. Flame 1999, 118, 445] through the use of a larger basis set and the application of group equivalents. The root-mean-square deviation (rms) from experiment for the predicted group-equivalent gas phase heats of formation is 3.2 kcal/mol with a maximum deviation of 6.5 kcal/mol. The rms and maximum deviation of the predicted liquid heats of formation are 3.2 and 7.4 kcal/mol, respectively. Finally, the rms and maximum deviation of predicted solid heats of formation are 5.6 and 12.2 kcal/mol, respectively, an improvement in the rms of approximately 40% compared to the earlier Rice et al. predictions using atom equivalents and a smaller basis set (B3LYP/6-31G*).

No abstract

A four-dimensional intermolecular potential energy surface for the carbon dioxide dimer has been computed using the many-body symmetry-adapted perturbation theory (SAPT) and a large 5s3p2d1f basis set including bond functions. The SAPT level applied is approximately equivalent to the supermolecular many-body perturbation theory at the second-order level. An accurate fit to the computed data has been obtained in a form of an angular expansion incorporating the asymptotic coefficients computed ab initio at the level consistent with the applied SAPT theory. A simpler site-site fit has also been developed to facilitate the use of the potential in molecular dynamics and Monte Carlo simulations. The quality of the new potential has been tested by computing the values of the second virial coefficient which agree very well with the experimental data over a wide range of temperatures. Our potential energy surface turns out to be substantially deeper than previous ab initio potentials. The minimum of −484 cm−1 has been found for the slipped parallel geometry at the intermolecular separation R=3.54 Å and a saddle point at −412 cm−1 for the T-shaped configuration and R=4.14 Å. Three minima and two first-order saddle points have been located on the pairwise-additive potential energy surface of the CO2 trimer. The nonplanar structure of C2 symmetry has been found to be 48.8 cm−1 more stable than the cyclic planar form of C3h symmetry, in disagreement with experimental observation. It is suggested that the relative stability of the two isomers cannot be reliably determined by pairwise-additive potential and inclusion of three-body forces is necessary for this purpose.

No abstract

Geometry optimizations and normal-mode analyses of three conformers of 1,3,5-trinitro-s-triazine (RDX) are performed using second-order Moller−Plesset (MP2) and nonlocal density functional theory (DFT) methods. The density functional used in this study is B3LYP. The three conformers of RDX are distinguished mainly by the arrangement of the nitro groups relative to the ring atoms of the RDX molecule. NO2 groups arranged in either pseudo-equatorial or axial positions are denoted with (E) or (A), respectively. The AAE conformer has C s symmetry and is the structure in the room-temperature stable crystal (α-RDX). The AAA and EEE conformers have C 3 v symmetry, a symmetry consistent with vapor and β-solid infrared spectra. The AAE and AAA conformers are studied at the MP2/6-31G*, B3LYP/6-31G*, and B3LYP/6-311+G** levels, and the EEE conformer is studied using the B3LYP density functional and the 6-31G* and 6-311+G** basis sets. The geometric parameters and infrared spectra of the AAA conformer are in good agreement with experimental gas-phase and β-solid data, supporting the hypotheses derived from experiment that the AAA structure is the most probable conformer in vapor-phase and β-solid RDX. The B3LYP/6-311+G** structures and simulated infrared spectra are in closest overall agreement with experimental data. The MP2/6-31G* structures and spectra are in poorest overall agreement with experiment.

A recently developed method, symmetry-adapted perturbation theory based on the density-functional description of monomers [SAPT(DFT)], is shown to be sufficiently accurate and numerically efficient to facilitate predictions of the structure of molecular crystals from first principles. In one application, a SAPT(DFT) potential was used to generate and order polymorphs of the cyclotrimethylene trinitramine crystal, resulting in the lowest-energy structure in excellent agreement with the experimental crystal. In a different application, a SAPT(DFT)-based calculation reproduced the lattice energy of the benzene crystal to within a few percent.

Theoretical predictions of the crystallographic properties of a series of 10 energetic molecular crystals have been done using a semiempirical correction to account for the van der Waals interactions in conventional density functional theory (termed DFT-D) as implemented in a pseudopotential plane-wave code. This series contains compounds representative for energetic materials applications, that is, hexahydro-1,3,5-trinitro-1,3,5-s triazine (R-and γ-RDX phases), 1,3,5,7-tetranitro-1,3,5,7-tetraaza-cyclooctane (β-, R-, and δ-HMX phases), 2,4,6,8,10,12-hexanitrohexaazaisowurtzitane (CL20) (ε-, β-, and γ-HNIW phases), nitromethane (NM), trans-1,2,-dinitrocyclopropane, 1,2,3,5,7-pentanitrocubane (PNC), pentaerythritol tetranitrate (PETN), 2,4,6-trinitro-1,3,5-benzenetriamine (TATB), 2,4,6-trinitrotoluene (TNT-I phase), and 1,1-diamino-2,2-dinitroethylene (FOX-7), systems belonging to diverse chemical classes that encompass nitramines, nitroalkanes, nitroaromatics, nitrocubanes, nitrate esters, and amino-nitro derivatives. At ambient pressure, we show that the DFT-D method is capable of providing an accurate description of the crystallographic lattice parameters with error bars significantly lower than those obtained using conventional DFT. Practically, for all crystals considered in this study the predicted lattice parameters are within 2% from the corresponding experimental data [R-RDX (1.58%), β-HMX (0.64%), ε-HNIW (1.42%), NM (0.75%), DNCP (1.99%), TATB (1.74%), TNT-I (0.92%), PNC(0.78%), PETN(1.35%), FOX-7(1.57%)], with the best level of agreement being found for systems where experimental data have been collected at low temperatures. A similar good agreement of the predicted and experimental crystallographic parameters was obtained under hydrostatic compression conditions as demonstrated for the cases of RDX, HMX, CL20, NM, TATB, and PETN crystals. These results indicate that the DFT-D method provides significant improvements for description of intermolecular interactions in molecular crystals at both ambient and high pressures relative to conventional DFT. In this last case, large errors of the predicted lattice parameters have been found at low pressures; theoretical values approach the experimental results only at pressures in excess of 6 GPa.