Methyl torsions

Methyl rotors pose relatively simple, fundamental questions about the nature of noncovalent interactions within molecules. The discovery in the late 1930s1 of the 1025 cm-1 potential energy barrier to internal rotation in ethane was surprising, since no covalent chemical bonds are formed or broken as methyl rotates. By now it is clear that the methyl torsional potential depends sensitively on the local chemical environment. The barrier is 690 cm-1 in propene,2 comparable to ethane,... 

To provide further experimental clues to the nature of substituent effects, we decided to test the additivity of orf/io-substituent effects. Specifically, we set out to compare methyl torsional potentials in S0, Sp and D0 for the sequence of molecules o-fluorotoluene (studied by Ito and coworkers), o-chlorotoluene, and 2-fluoro-6-chlorotoluene. 

In this section, we present a unified picture of the different electronic effects that combine to determine methyl rotor potentials in the S0, Sp and D0 electronic states of different substituted toluenes. Our approach is based on analysis of ab initio wavefunctions using the natural bond orbitals (NBOs)33 of Weinhold and cowork-ers. We will attempt to decompose the methyl torsional potential into two dominant contributions. The first is repulsive steric interactions, which are important only when an ortho substituent is present. The second is attractive donor-acceptor interactions between CH bond pairs and empty antibonding orbitals vicinal to the CH bonds. In the NBO basis, these attractive interactions dominate the barrier in ethane (1025 cm-1) and in 2-methylpropene (1010 cm-1) see Figure 3. By analogy, donor-acceptor attractions are important in toluenes whenever there is a substantial difference in bond order between the two ring CC bonds adjacent to the C-CH3 bond. Viewed the other way around, we can use the measured methyl rotor potential as a sensitive probe of local ring geometry. 

As a simple example of the influence of H-bonding on torsions, let us consider the methyl torsions in the acetamide molecule (5.95b). By bringing various H-bonding species into complexation with the carbonyl oxygen and/or amine group of the amide moiety, we can alter the relative weightings of covalent (wcov) versus dipolar (wdip) amide resonance forms,... 

The intimate connection between methyl torsional stiffening and the variation in amide CO/CN bond orders is illustrated in Fig. 5.65. This plot shows that the methyl rotation barrier A /s,b varies roughly linearly with the difference Ab in CO/CN bond orders,... 

Figure 8. Methyl-orientation results for the three methyl groups in N-acetyl-L-leucine. A, x-ray determined p(r) charge density map. B, graph showing relation between the H-C-C-H methyl torsion angles determined from the x-ray results and those determined by using a quantum chemical geometry optimization. The rms error is 8°.

Parameters describing the hydrogen position are sometimes included in the least squares calculations. But both methyl twist angles and related vibrational amplitudes thus obtained are particularly sensitive to the assumptions made concerning the methyl torsional potential because of the low scattering power of hydrogen. 

Raman circular intensity differentials (c.i.d.), which are observed in methyl asymmetric deformations and methyl torsions, may be valuable in probing chirality in monoterpenoids (—)-limonene and (+)-carvone each show a broad, weak depolarized Raman band at 250 cm-1 with a large c.i.d. The origin of these bands is not yet certain.23... 

Fig, 17. Band strvcturea corresponding to the methyl torsion and aldehy-die hydrogen wagging modes in the a A" — spectrum of thioacetalde ... 

As Figs. 8.3 and 8.4 show, the methyl torsion is a rare example of a reliable group frequency mode in ESTS spectroscopy. Most organic compounds that contain a methyl group bonded to an sp hybridized carbon atom have a strong band at 250 10 cm. This is often the most... 

The most obvious discrepancies are the methyl torsions. These are always observed at about 250 10 cm, irrespective of the alkane chain length or the crystallographic enviromnent of the end group. This is considerably above the maximum value for other out-of-plane torsional modes and shows them to be imlike the other TAMs. 

The two methyl torsions of the shorter alkane chains couple to produce an in-phase and out-of-phase pair of bands. The longer the chain the weaker this coupling and so the less the frequency splitting of the pair. In the limit, the splitting should completely collapse. The methyl torsion of one end of a chain would then be unaware of the nature, or even the presence, of the other end of the chain and both methyl torsions oscillate independently. Only one accidentally degenerate band is observed. In the alkanes this collapse has already occurred in undecane. Naively then, all methyl torsions, n > 11, will occur at the same frequency. Unfortunately this is not observed. 

In the DFT calculations the degenerate methyl torsions occur at either 251 cm" or 247 cm", dependent upon whether the n-alkane has n-even (n > 14), or n-odd n > 15). This difference also manifests itself in the observed INS results (12 > n> 20) for n-even these have methyl-torsion frequencies at 255 cm" and, for n-odd, at 250 cm. Further, the average frequency values for the shorter alkanes, with coupled methyl torsions, fall close to the appropriate n-even, or n-odd, frequency. There is an obvious difficulty with the suggestion that many (n > 11) alkanes are so long that their end of chain vibrations are independent but, nonetheless, that the nature (n-odd, or -even) of a chain s ends determines these dynamics. This apparent inconsistency is resolved if we recognise that, for the purpose of their physical properties, the n-odd and n-even alkanes are quite different and should not be naively compared. Thus the methyl... 

In Fig. 10.15 we show the dispersion curves [25] for isotactic polypropylene, the S(Q,o ) derived from them and the INS spectrum recorded on TOSCA. The dispersion curves were based on an erroneous assignment of 200 cm for the methyl torsion so there is a marked discrepancy at that point. For the remainder of the spectrum, there is qualitative agreement but the detail in the INS spectrum is not reproduced. This probably stems from the neglect of the site symmetry and the intermolecular interactions. Clearly this is an area that is ripe for re-investigation with modem INS spectrometers and ab initio calculations. 

In this section we will consider polydimethylsiloxane (PDMS) as an example of the type of work that is possible with amorphous polymers. The structure and INS spectrum of PDMS are shown in Fig. 10.21a [40]. The repeat unit shown in Fig. 10.21b was used to model the spectrum using the Wilson GF matrix method [41]. The major features are reproduced skeletal bending modes below 100 cm", the methyl torsion and its overtone at 180 and 360 cm respectively, the coupled methyl rocking modes and Si-0 and Si-C stretches at 700-1000 cm and the unresolved methyl deformation modes 1250-1500 cm. The last are not clearly seen because the intensity of the methyl torsion results in a large Debye-Waller factor, so above 1000 em or so, most of the intensity occurs in the phonon wings. 

The limitations are also clear the integrated intensity of the methyl torsion is seriously underestimated and the relative intensities of the modes at 680 and 744 cm are not correct. All of these problems stem from the simplicity of the model the methyl group sees many different environments and this results in the large width (70 cm ) of the methyl torsion. This has the potential to be used as a probe of the local environment as has been done for the ester methyl of poly(methylmethacrylate) [42],... 

Acetanilide, and some of its isotopomers, have been studied by INS spectroscopy [56-58]. The dispersion curves of the fully deuterated material have been measured by coherent INS [59]. A comprehensive analysis of acetanilide in the solid state was carried out with molecular dynamics simulations [57]. This includes all the lattice modes, as shown in Fig. 10.27 The simulations suggested that the barrier to the methyl torsion was enhanced when the peptide group is hydrogen-bonded and that this was a through-bond polarization effect. The methyl torsion was... 

H. Takeuchi, J.S. Higgins, A. Hill, A. Maconnachie, G. Allen G.C. Stirling (1982). Polymer, 23, 499-504. Investigation of the methyl torsion in isotactic polypropylene—comparison between neutron inelastic scattering spectra and normal coordinate calculations. 

In the present discussion we shall not consider the special problems introduced by large amplitude vibrations such as methyl torsions (see H. Dreizler, this volume) or ring puckering (see W. J. Lafferty, this volume). Also, it is assumed that perturbations, such as nuclear hyperfine structure, have been corrected. 

---