Heinz Oberhammer

Structures and conformations of non-rigid molecules

From the beginnings of modern chemistry, molecular structure has been a lively area of research and speculation. For more than half a century spectroscopy and other methods have been available to characterize the structures and shapes of molecules, particularly those that are rigid. However, most molecules are at least to some degree non-rigid and this non-rigidity plays an important role in such diverse areas as biological activity, energy transfer, and chemical reactivity. In addition, the large-amplitude vibrations present in non-rigid molecules give rise to unusual low-energy vibrational level patterns which have a dramatic effect on the thermodynamic properties of these systems. Only in recent years has a coherent picture of the energetics and dynamics of the conformational changes inherent in non-rigid (and semi-rigid) molecules begun to emerge. Advances have been made in a number of different experimental areas: vibrational (infrared and Raman) spectroscopy, rotational (microwave) spectroscopy, electron diffraction, and, most recently, laser techniques probing both the ground and excited electronic states. Theoretically, the proliferation of powerful computers coupled with scientific insight has allowed both empirical and ab initio methods to increase our understanding of the forces responsible for the structures and energies of non-rigid systems. The development of theory (group theoretical methods and potential energy surfaces) to understand the unique characteristics of the spectra of these floppy molecules has also been necessary to reach our present level of understanding. The thirty chapters in this volume contributed by the key speakers at the Workshop are divided over the various areas. Both vibrational and rotational spectroscopy have been effective at determining the potential energy surfaces for non-rigid molecules, often in a complementary manner. Recent laser fluorescence work has extended these types of studies to electronic excited states. Electronic diffraction methods provide radial distribution functions from which both molecular structures and compositions of conformational mixtures can be found. Ab initio calculations have progressed substantially over the past few years, and, when carried out at a sufficiently high level, can accurately reproduce (or predict ahead of time) experimental findings. Much of the controversy of the ARW related to the question of when an ab initio is reliable. Since the computer programs are readily available, many poor calculations have been carried out. However, excellent results can be obtained from computations when properly done. A similar situation exists for experimental analyses. The complexities of non-rigid molecules are many, but major strides have been taken to understand their structures and conformational processes.

On the geometry and internal rotation of XSCF3 compounds (X = F, Cl and CF3)

Abstract The molecular structures ( r α 0 values) for XSCF 3 with X = F, Cl and CF 3 have been determined by electron diffraction of gases. While the geometry (C-F bond length and FCF angle) of the CF 3 groups and the bond angle at the sulfur atom depend very little on the substituent X, the S-C bond length increases with decreasing electronegativity of X from 1.805 (3) A for X = F to 1.824 (6) A for X = Cl. Torsional force constants for the CF 3 groups were derived from vibrational amplitudes. A strong increase of this force constant is observed between FSCF 3 ( f τ = 0.09 (2) mdyn A) and CISCF 3 ( f τ = 0.18 (5) mdyn A). The torsional frequencies derived from the electron diffraction experiment agree very well with the values observed in the far IR spectra for CISCF 3 , and CF 3 SCF 3 . A force field for CF 3 SCF 3 has been derived from IR and Raman data.

Peroxo compounds: Part XVI. Electron diffraction investigation of the molecular structures of di-t-butyl peroxide Me3COOCMe3 and bis(trimethylsilyl) peroxide Me3SiOOSiMe3

Abstract The molecular geometries of Me 3 COOCMe 3 and Me 3 SiOOSiMe 3 , in the gas phase were determined by electron diffraction. For the skeleton of di- t -butyl peroxide the following geometric parameters ( r a -values) were obtained: r (O—O) = 1.480 A (assumed), . This dihedral angle is compared with the results of IR and Ram an spectroscopy, dipole moment measurements and photoelectron spectroscopy. The main geometric parameters for bis(trimethylsilyl) peroxide are . For both peroxides SCF-MO calculations in the CNDO/2 approximation do not reproduce the experimental results.

The molecular structure of cyclic methylsiloxanes

Abstract A series of methylcyclosiloxanes from the trimer to the hexamer has been investigated by gas phase electron diffraction. The Si-O bond distance varies between 1.635 and 1.620 A, the SiOSi angle has values from 132° to 149° within the series. The trimer is substantially planar, the best agreement for the tetramer was obtained with a model of S 4 . symmetry. As a consequence of large amplitude vibrations, the larger rings do not possess a well-defined conformation. The results are discussed in relation to previous spectroscopic investigations.

Conformational Properties of 1-Fluoro-1-silacyclohexane, C5H10SiHF: Gas Electron Diffraction, Low-Temperature NMR, Temperature-Dependent Raman Spectroscopy, and Quantum Chemical Calculations†

The molecular structures of axial and equatorial conformers of 1-fluorosilacyclohexane, C5H10SiHF, as well as the thermodynamic equilibrium between these species were investigated by means of gas electron diffraction (GED), dynamic nuclear magnetic resonance, temperature-dependent Raman spectroscopy, and quantum chemical calculations (MP2, DFT, and composite methods). According to GED, the compound exists in the gas phase as a mixture of two conformers possessing the chair conformation of the six-membered ring and Cs symmetry and differing in the axial or equatorial position of the Si-F bond (axial ) 63(8) mol %/equatorial ) 37(8) mol %) at T ) 293 K, corresponding to an A value of –0.31(20) kcal mol. Density functional theory (DFT) calculations were employed to obtain the minimal energy path of the conformational inversion. The MP2, G3B3, and CBS-QB3 methods were also employed to calculate the equilibrium geometries and energies of the local minima in the gas phase and in solution. The gas-phase results are in good agreement with the experiment, whereas a combined PCM/IPCM(B3LYP/6-311G(d)) approach overestimates the stabilization of the axial conformer by 0.3-0.4 kcal mol in solution at 112 K. Temperature-dependent Raman spectroscopy in the temperature ranges of 210–300 K (neat liquid), 120–300 K (pentane solution), and 200–293 K (dichloromethane solution) also indicates that the axial conformer is favored over the equatorial one by 0.25(5), 0.22(5), and 0.28(5) kcal mol (∆H values), respectively.

Vibrational spectra and normal coordinate analysis of CF3 compounds

Abstract Perfluorotrimethylamine, N(CF 3 ) 3 , has been studied in the liquid and gas state by Raman spectroscopy, and by infrared spectroscopy in the gas state and in a matrix. The assignment of the spectra supports C 3 rather than C 3h molecular symmetry, but the NC 3 pyramid is suggested to be flat. A normal coordinate analysis has been performed. The vibrations of the NC 3 skeleton are strongly mixed with vibrations of the CF 3 ligands and are entirely noncharacteristic. The structure of N(CF 3 ) 3 has been reinvestigated by electron diffraction. The following molecular parameters have been determined: r (CF) 1.323(4) A, r (CN) 1.426(6) A, ∠FCF 108.3°(4), ∠CNC 117.9°(4), τ(CF) 3 26.4° (10). The torsional fundamentals were located near 50 cm −1 both by the diffraction and the spectroscopic results. The CN bond in N(CF 3 ) 3 is shorter and stronger than that in N(CH 3 ) 3 .

Structural determination of a recalcitrant molecule (S2F4)

Abstract The structure determination of S 2 F 4 (or SF 3 SF), the dimer of SF 2 , is made difficult by the large variety of possible conformers and the instability of the compound. An electron diffraction and microwave study succeeded only with the help of a molecular model derived from ab initio calculations, after initial experimental attempts had failed. The force field required for a joint electron diffraction and microwave spectroscopy analysis was calculated by ab initio methods and adjusted to the experimental vibrational frequencies. The structure of S 2 F 4 is a trigonal bipyramid with the electron lone pair, the SF group and one fluorine atom in equatorial positions. The SF 3 group is strongly distorted with the two axial SF bonds differing by 0.10 A and bond angles between axial bonds and equatorial plane of about 77° and 92°, respectively. The geometric structure of S 2 F 4 in combination with the ab initio calculations allows one to visualize the dissociation process SF 3 SF → 2 SF 2 more clearly.

The geometric structures of the disulphur difluoride isomers: an experimental and ab initio study

The molecular structures of FSSF and SSF2 have been studied by gas-phase electron diffraction, in combination with literature data on rotational constants from microwave spectroscopy. The following rav parameters were obtained, with 3σ uncertainties in parameters: for FSSF, r(SS) = 1.890(2) A, r(SF) = 1.635(2) A, ∠ SSF = 108.3(2)° and τFSSF = 87.7(4)°; for SSF2, r(SS) = 1.856(2) A, r(SF) = 1.608(2) A, s SSF = 108.1(2)° and < FSF = 91.7(3)°. These results are consistent with, and more precise than, those originally reported from microwave spectroscopy (R.L. Kuczkowski, J. Am. Chem. Soc., 86 (1964) 3617), and also consistent with, but slightly less precise than, very recent microwave results for SSF2 only (R.W. Davis, J. Mol. Spectrosc., 116 (1986) 371). Ab initio SCF calculations predict the structure of SSF2 satisfactorily, but do not account completely for the extraordinarily short SS bond in FSSF, which is shown to have a substantial (p-p)π component. Correlated calculations at the MP2 level with a DZP basis provide reasonably accurate geometrical predictions for FSSF. At the SCF level, FSSF is more stable than SSF2, but this error is removed with the inclusion of MP2 energies. Both cis and trans barriers to internal roation in FSSF are predicted to be very high, at 135 and 117 kJ mol−1, consistent with the observed torsional frequency. The SS distance in planar cis and trans forms of FSSF is much longer than in the equilibrium skew form, due to the elimination of π overlap.

How reliable are ab initio calculations? Experimental and theoretical investigation of the structure and conformation of chlorocarbonyl isocyanate, ClC(O)NCO

Abstract A gas electron diffraction study of ClC(O)NCO results in a mixture of trans ( anti ) and cis ( syn ) isomers in a ratio of ca. 3:1 (Δ G = 0.7(3)kcal mol −1 ), and confirms the interpretation of vibrational spectra. The following geometric parameters ( r a , values with 3σ uncertainties) have been derived for the trans structure CO (isocyanate) = 1.139(16)A, CO (carbonyl) = 1.201(16) A, NC = 1.218(23) A, N-C = 1.384(6) A, C-Cl = 1.757(5) A, C-NC = 127.1(16)°, N-CO = 124.8(15)°, N-C-Cl = 115.8(8)° and NCO = 173.4(23)°. Calculated (HF/6-31G ∗ ) differences between geometric parameters of cis and trans isomers have been incorporated in the experimental structure determination. These ab initio calculations reproduce the experimental trans structure reasonably well, but the relative stability of the two isomers is at variance with the experiment. Calculations with different basis sets (4-31G ∗ , 6-31G ∗ and D95 ∗ ) and at various levels of theory (HF, MP2 and MP4SDTQ) predict the cis form to be more stable by ΔE = 0.40 to 1.58 kcal mol −1 , whereas experiments demonstrate the trans isomer to be more stable.

Ab initio study of some peroxides. HOOH, CH3OOH and, CH3OOCH3

Abstract The geometries of HOOH, CH3OOH, and CH3OOCH3, were optimized with different basis sets (3-21G, 6-31G*(*) and D95**) at different levels of theory (HF, MP2, MP4, and CI). HF/3-21G optimizations result in planar trans conformations for all three peroxides. HF/6-31G** calculations predict skew conformations for HOOH and CH3OOH, but a planar trans struture for CH3OOCH3. For the larger basis set the calculated bond lengths, especially the O-O bonds, are too short. Optimizations for HOOH including electron correlation at the MP2, MP3, MP4, CI, and CCD level improve the agreement for bond lengths and the OOH angle, but result in dihedral angles Which are too large by 3– 8°. In the case of CH3OOCH3, similar calculations at the MP2 and CI level predict planar trans structures instead of the experimentally observed skew conformation. On the other hand, MP4 single point calculations at MP2 optimized parameters result in a correct skew structure. For all three peroxides a computationally “economic” method, i.e., single point calculations at MP2 or MP4 level with HF/3-21G optimized parameters, result in close agreement between calculated and experimental structures.