How to treat C–F stretching vibrations? A vibrational CD study on chiral fluorinated molecules

Nora M. Kreienborg and Christian Merten *
Ruhr-Universität Bochum, Fakultät für Chemie und Biochemie, Organische Chemie 2, Universitätstraße 150, 44801 Bochum, Germany. E-mail: christian.merten@ruhr-uni-bochum.de; Web: http://www.rub.de/chirality

Received 14th April 2018 , Accepted 25th May 2018

First published on 25th May 2018


The analysis of vibrational circular dichroism and infrared spectra typically requires the prediction of spectra on the density functional theory level. In particular for absolute configuration determinations, for which the comparison between experiment and theory is often supported by similarity analysis algorithms, it is important that frequencies, relative band intensities and VCD signs are predicted correctly. Due to the poor prediction of harmonic frequencies, carbon–fluorine stretching vibrations are often strongly misplaced by common hybrid functionals such as B3LYP. Herein we show that the M06-2X functional provides harmonic C–F stretching frequencies with an accuracy sufficient for a reliable spectra analysis. We briefly discuss the origin of this exceptional performance and show that it is likely to be related to a cancellation of errors.


Introduction

Vibrational circular dichroism (VCD) spectroscopy has become an important method for the determination of absolute configurations (AC) of chiral molecules, with applications ranging from natural products1 to fully synthetic samples,2–4 from small molecules to large polymers and biomolecules, and from solution phase5 to cryogenic matrices.6,7 The interpretation of VCD spectra relies heavily on the accurate theoretical prediction of spectral signatures. A comparison of experimental and calculated IR and VCD spectra is thus an indispensable step in every VCD spectral analysis, and it is typically carried out either visually and thus qualitatively or quantitatively using similarity analysis algorithms.8–10 Therefore, the correct prediction of vibrational frequencies and of the corresponding IR and VCD intensities (dipole and rotational strength) is extremely important and the key for the characterization of the stereochemistry and the conformational preferences of the chiral target molecules. Since the introduction of Stephens’ original theoretical treatment of the vibrational rotational strength, which coincided with the development of the B3LYP functional,11–13 the VCD community uses such hybrid functionals (B3LYP, B3PW91) in combination with Pople or correlation-consistent basis sets (e.g. 6-311++G(2d,p) or aug-cc-pVTZ) as they yield very good results. Often, solely a shift of the vibrational frequencies by a scaling factor in the range of 0.96–0.99 needs to be applied in order to account for errors due to the harmonic approximation and to obtain a very good agreement with the experimental spectra. The spectral region around 3000 cm−1 usually needs stronger corrections, i.e. lower values of smaller scaling factors, as the CH-stretching modes observed in this region are more anharmonic than in the fingerprint region, where mostly deformation modes are present. More recently, theoretical approaches to account for anharmonicities in IR and VCD spectra have also been reported, which make the use of scaling factors obsolete.14–17 However, as the agreement of solution phase spectra with predicted harmonic spectra is already sufficient for most applications, the added computational cost is often not justified and they have thus not become frequently used yet.

During our VCD spectroscopic studies on intermolecular interactions of chiral molecules,5 we encountered problems with the prediction of carbon–fluorine stretching frequencies. These C–F-stretching vibrations are usually among the strongest signals in IR spectra of organic molecules, and CF3-groups typically feature two strong bands in the range of 1200–1100 cm−1, which arise from the asymmetric and symmetric stretching vibrations. Using one of the aforementioned common hybrid functionals, the computed (harmonic) IR and VCD signatures of bands which do not feature any contributions of C–F stretching motions are found to be in good agreement with the experimentally observed frequencies and intensities. However, in the same set of calculations, C–F stretching frequencies are often found to be predicted about 20–50 cm−1 too low! To the best of our knowledge and to our surprise, this critical failure of B3LYP and B3PW91 in harmonic vibrational spectra prediction of fluorinated compounds has not been addressed in literature. In fact, most studies dealing with small fluorinated molecules often do not require a perfect overlap of vibrational modes, as there are only very few spectral features which can in turn be assigned without any doubt. Alternatively, it is often referred to frequency shifts induced by intermolecular interactions instead of exact positions, and these shifts themselves are actually mostly predicted quite well.18–20

As pointed out above, the VCD spectral analysis relies heavily on a good agreement of the predicted spectral patterns with the experimental data. Therefore, a frequency mismatch of several tens of wavenumbers will completely alter the overall appearance of the VCD signatures. Consequently, due to the dominance of the bands in the IR spectra, every quantitative similarity analysis21 will be confused and routine AC determinations can become very troublesome.

With this contribution, we want to draw attention to the surprising performance of the M06-2X functional,22,23 which is usually neglected by the vibrational spectroscopy community as the mean error in vibrational frequencies is often larger than for B3LYP.24–26 Using two examples from our studies which feature CF3-groups in chemically different environments (Scheme 1), we show that M06-2X outperforms the classical functionals in the prediction of harmonic IR and VCD spectra for fluorinated compounds. However, this advantage cannot be extended to anharmonic simulations, which are recommended with the M06-2X functional.27


image file: c8cp02395f-s1.tif
Scheme 1 Chemical structures of the two fluorinated target molecules.

Experimental and computational details

Synthesis of FTUP

A solution of 3,5-bis(trifluoromethyl)phenyl isothiocyanate (0.14 ml, 0.76 mmol) and (R)-α-phenylethylamine (0.1 ml, 0.75 mmol) in pentane was stirred overnight at room temperature. The resulting white solid was filtered off, washed with pentane and dried under oil pump vacuum to obtain 283 mg (0.72 mmol, 95%) of (R)-FTUP. 1H-NMR (200 MHz, chloroform-d) δ = 7.66 (s, 3H, CF3-ArH), 7.39 (m, 5H, ArH), 5.32 (s, 1H, CH–CH3), 1.62 (d, 3H, CH–CH3) ppm.

IR/VCD spectroscopy

The vibrational spectra were recorded on a Bruker Vertex 70 equipped with a PMA 50 unit for polarization modulated measurements. Solutions of the samples were held in a transmission cell with BaF2 windows with 100 μm path length. Both IR and VCD spectra were recorded at 4 cm−1 spectral resolution by accumulating 32 respectively ∼20[thin space (1/6-em)]000 scans (8 h accumulation time). Baseline correction of the VCD spectra was done by subtraction of the spectra of the racemic mixture recorded under identical conditions.

Computational details

Geometry optimizations and frequency calculations were performed using the Gaussian 09 E.01 software package with tight convergence criteria and ultrafine integration grids.28 Solvent effects were taken into account either implicitly by using the integral equation formalism of the polarizable continuum model (IEFPCM).29 Vibrational line broadening was simulated by assigning a Lorentzian band shape with half-width at half-height of 6 cm−1 to the calculated dipole and rotational strength.

Results and discussion

As first case study, we selected the chiral 2,2,2-trifluoro-1-(9-anthryl)ethanol (TFAE, Scheme 1), which is a common chiral shift reagent in NMR spectroscopy.30 It interacts through hydrogen bonds with, for instance, carbonyl groups of chiral esters and amides, which results in a stereo-discriminating deshielding of protons in the vicinity of the interaction site. Similar to the procedure for Mosher esters,31 the chemical shift differences observed for the two enantiomers of TFAE can be used to determine the AC and enantiopurity of the interacting chiral molecules.

In context of a study investigating these chiral interactions of TFAE with a dipeptide, we have recorded its IR and VCD spectra and performed vibrational spectra calculations.32 We found the monomeric TFAE to feature two almost equally populated conformations, which only differ in the orientation of the hydroxy-proton (cf. Table S1, ESI). Based on the single conformer spectra obtained at B3LYP-level, we simulated the Boltzmann weighted spectra shown in Fig. 1. We applied a frequency scaling factor of 0.98, which is common for the given level of theory. Our initial band assignments, which are given in the figure, took into account that modes involving the CF3-group, such as the Cα–CF3 and CF-stretching vibrations, were predicted at significantly too low frequencies. For instance, band 11 of the Cα–CF3 stretching vibration is experimentally observed at 1230 cm−1 and predicted at 1190 cm−1 and band 14 of the CF3 stretching mode is found experimentally at 1095 cm−1 and at 1060 cm−1 in the calculation. This is in contrast to vibrational modes not affected by CF-contributions, such as bands 1–8, which belong to CAr–H in-plane deformation and Cα–H deformation modes. Band 6, for instance, is found at 1372 cm−1 and predicted to occur at 1370 cm−1. A better agreement of the strong spectral signatures would have required a frequency scaling factor of about 1.05, which had resulted in most other bands being significantly too high in frequency. The theoretical spectra of TFAE clearly show that vibrational modes of non-fluorine containing molecules or remote parts of the molecules, which are not influenced by any CF-vibrational modes, are predicted very nicely with B3LYP and scaling factors around 0.98, while vibrations with contributions of the CF-bonds are systematically underestimated.


image file: c8cp02395f-f1.tif
Fig. 1 Comparison of the experimental IR and VCD spectra of (S)-TFAE (0.18 M, 100 μm path length, CDCl3) with calculated spectra obtained with the B3LYP and M06-2X functionals, the 6-31+G(2d,p) Pople basis set and the IEFPCM of chloroform. Numbers indicate bands assignments.

This apparent mismatch between experimental and scaled harmonic frequencies could not be improved by inclusion of dispersion interactions using the Grimme-D333 scheme or a change of the functional to B3PW91. Even the use of the double-hybrid functional B2PLYP34 or pure second-order Møller–Plesset perturbation theory (MP2) does not improve of the harmonic frequencies (cf. comparison in ESI). It should also be mentioned that VCD spectra are not available on B2PLYP and MP2 levels, which reduces the scope of applications of these methods.

After Boltzmann-weighting and frequency scaling by a factor of 0.975, the predicted IR and VCD spectral signatures of TFAE obtained with the M06-2X functional show a significant improvement compared to any of the aforementioned methods. As indicated by the band assignments in Fig. 1, all experimentally observed IR and VCD bands can easily be correlated with bands in the predicted spectra. With the exception of the bands 13 and 14, which are difficult to disentangle, the frequencies and relative intensities match quite well and an AC determination would not be obscured by misplaced bands.

In order to increase the complexity of the fluorine-bearing structural motif, the second case study focuses on the 3,5-bis(trifluoromethyl)phenyl group. It is a privileged building block in the design of hydrogen-bonding organocatalysts, as the trifluoromethyl-substituents increase the acidity of the ortho-CH bonds and thereby involve them in binding events with Lewis basic centers. For this study, we have selected the simple chiral thiourea FTUP (Scheme 1) as model substance in order to also highlight the implications of our observation for the characterization of catalytically active molecules. Fig. 2 shows the IR and VCD spectra of (R)-FTUP. The strongest bands in the IR spectrum are associated with the asymmetric and symmetric CF3-stretching vibrations (1185 and 1145 cm−1), and the strong band at 1280 cm−1 arising from coupled CAr–CF3 stretching and CAr–H in-plane bending motions. The broad and less defined spectral signatures in the range from 1600–1400 cm−1 are due to vibrational modes localized at the thiourea moiety, the α-phenylethyl side group and C[double bond, length as m-dash]C stretching vibrations of the aromatic rings. The intensities of the VCD signatures are again quite evenly distributed and the VCD bands corresponding to strong IR signatures are not found to be particularly more intense than other bands. Furthermore, it should be noted that the strong negative VCD signature at 1450 cm−1 is characteristic for the α-phenylethyl stereocenter.35,36


image file: c8cp02395f-f2.tif
Fig. 2 Comparison of the experimental IR and VCD spectra of (R)-FTUP (76 mM, 50 and 100 μm path length, CDCl3) with calculated spectra at the B3LYP and M06-2X level of theory with the 6-311++G(2d,p) basis set and the IEFPCM of chloroform. Frequency scaling factors of 0.985 and 0.975 were used for B3LYP and M06-2X. VCD spectra were recorded at two different path lengths in order to capture the signatures of both the strong and the weaker absorbance bands.

In comparison to TFAE, the conformational space of FTUP is significantly larger, as the thiourea moiety can adopt different conformations with respect to the relative orientation of the N–H bonds. In addition, the α-phenylethyl substituent may also rotate, generating further degrees of conformational freedom. In total we found 22 stable conformations, which can be classified in three families according to their thiourea conformation (cf.Scheme 2). The theoretically predicted IR and VCD obtained based on the conformational analysis at B3LYP and M06-2X level of DFT (Tables S2–S4, ESI) are provided alongside the experimental spectra in Fig. 2. From the direct comparison of the B3LYP results with the experimental IR spectrum, strong shifts of the vibrations with contributions of the CF3 groups, similar to those observed for TFAE, can immediately be notice. On the other hand, the spectral range of 1600–1400 cm−1 for which the underlying vibrational motions do not feature contributions of the CF-stretching motions is predicted quite well. Overall, the similarity between the B3LYP-predicted IR spectrum and the experimental spectrum is apparent and, despite the strong band shifts, band assignments may still be made.


image file: c8cp02395f-s2.tif
Scheme 2 Conformer families of FTUP.

Although intensities are not reproduced with the same quality as in the IR, the VCD signatures in the range of 1600–1400 cm−1 are predicted reasonably well. Likewise, the VCD spectral pattern within 1400–1300 cm−1 can be recognized in the predicted spectrum. Most noteworthy is the predicted VCD signature centered at around 1100 cm−1 and consisting of a positive band above and several negative bands below 1100 cm−1. Although the CF3-stretching modes are falsely predicted to occur in this spectral range, this predicted signature nicely resembles the experimentally observed one. As IR and VCD intensities are not directly related,37 we initially assumed that the strong IR modes of the CF3-stretching vibrations may simply not feature strong VCD bands and that the observed VCD signatures arise form underlying weak IR bands. However, a detailed analysis of the vibrational energy distribution showed that the predicted VCD pattern also features strong contributions of the CF3-stretching modes (see ESI). Consequently, both IR and VCD spectra are only of little use for the confirmation of conformational preferences of FTUP as the agreement is likely to be purely coincidental.38

The IR spectrum predicted with the M06-2X functional again provides a striking match with the experimental spectrum. The frequencies of most bands appear to be predicted correctly (after scaling with 0.975), and solely the intensities of very few bands do not exactly match. Similarly, the predicted VCD spectrum shows a very good agreement with the experimental spectrum and, as indicated by the band assignments in Fig. 2, almost all experimental VCD bands are given with correct sign and relative intensity. The most noteworthy and important improvement in terms of pattern comparison is certainly found for the range from 1500–1400 cm−1, where the strong negative band and the many small positive features on its high wavenumber flank are nicely reproduced. Moreover, almost all VCD bands observed in the “problematic” spectral range within 1300–1100 cm−1 are reasonably to very well reproduced in terms of relative intensities. Sole exception is the band at about 1350 cm−1, which is not very pronounced in the experiment but predicted to feature very high intensity.

In order to elucidate the influence of anharmonicity on the frequencies of the problematic vibrational motions, we selected iodotrifluoroethene (C2F3I) as an example for a smaller fluorine-containing molecule, which can easily be computed on VPT2 level.15–17 A comparison of the frequencies calculated using B3LYP and M06-2X with the experimental data recorded in fingerprint region using liquid xenon as solvent are provided in Table 1.18,19 In case of the B3LYP functional, the unscaled harmonic frequencies of modes containing C–X stretching motions (X = F, I) are found to be slightly smaller than the experimental values, while the frequency of the C[double bond, length as m-dash]C stretching vibration is, as expected, too high. Consequently, applying a frequency scaling factor comparable to those employed above only causes an even stronger disagreement. At VPT2-level, anharmonic corrections to the C–X stretching modes are rather small and only the C[double bond, length as m-dash]C stretching mode is adjusted. Although the predicted frequencies of the fundamental modes match quite well with the experimental band positions, an underestimation of the anharmonicity of the vibrations is still apparent. On the other hand, good accuracy of C–F stretching vibrations was reported for iodopentafluorobenzene39 using B3LYP anharmonic corrections with more accurate harmonic frequencies computed at coupled cluster level. Hence, the poor performance of the B3LYP functional can be associated with an underestimation of the harmonic frequencies of vibrations involving contributions of fluorinated groups as well as their strong anharmonicity. In contrast, the harmonic frequencies obtained with the M06-2X functional are somewhat too large after scaling by 0.975, while a smaller scaling factor of 0.95 yields numbers comparable to those obtained from VPT2-calculations at B3LYP level and thus a good agreement with the experimental band positions. Interestingly, the VPT2 calculations consistently predict the fundamental modes at significantly too high frequencies, suggesting that the M06-2X functional systematically overestimates the force constants for carbon–fluorine bond stretches.27

Table 1 Comparison of the experimental vibrational frequencies of C2F3I with predicted values (given in cm−1; Δν refers to the difference between experiment and either the scaled harmonic or the anharmonic frequencies) obtained with the B3LYP and M06-2X functional and the 6-311++G(2d,p) basis for carbon and fluorine and the LanL2DZ pseudo potential for iodine
B3LYP M06-2X
ν exptl 18,19 Assignm. ν calcd ν scaled Δνscaleda ν anharm Δνanharm ν calcd ν scaled Δνscaleda ν scaled Δνscaledb ν anharm Δνanharm
a Scaling factor of 0.975 is used for both B3LYP and M06-2X. b Optimum scaling factor of 0.95 obtained by averaging over νexptl/νcalcd.
1758 C[double bond, length as m-dash]C str. 1803 1758 0 1774 16 1865 1818 60 1772 14 1825 67
1311 CF2asym str. 1299 1267 −44 1296 −15 1375 1341 30 1306 −5 1341 30
1174 CF str. 1165 1136 −38 1166 −8 1233 1202 28 1171 −3 1206 32
1001 CF2sym str./C–I str. 1003 978 −23 999 −2 1044 1018 17 992 −9 1030 29


Conclusions

In summary, the examples discussed in this study show that the M06-2X functional outperforms B3LYP and other hybrid functionals in the prediction of harmonic vibrational spectra of organic molecules with fluorinated groups. The analysis of the IR and VCD spectra and the detailed band assignments of vibrational modes suggest that there is a clear link between the quality of the harmonic frequency prediction and the contribution of C–F bond stretching motions to the overall vibrational motion. It becomes obvious from comparison with VPT2 calculations that the good performance of M06-2X is due a cancellation of errors introduced by the harmonic approximation on the one hand and by the overestimation of force constants on the other hand. In conclusion, the findings of this study lead to the following recommendations for vibrational spectra calculations of fluorinated compounds: (1) anharmonicity should be taken into account whenever VPT2 calculations are feasible with respect to the number of conformers and the size of the molecule. (2) For large and flexible molecules, the cancellation of errors occurring with the M06-2X enables harmonic frequency and spectra prediction with an accuracy sufficient for a reliable spectra analysis.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank Prof. Dr Julien Bloino for fruitful discussions. Furthermore, we thank the Fonds der Chemischen Industrie (FCI) for a Liebig fellowship (C. M.) and a PhD stipend (N. M. K.). Finally, financial support by the Deutsche Forschungsgemeinschaft (DFG) through the Cluster of Excellence RESOLV (“Ruhr Explores SOLVation”, EXC 1069) and a research grant (ME 4267/3-1) is gratefully acknowledged.

Notes and references

  1. J. M. Batista, Jr., E. W. Blanch and V. D. S. Bolzani, Nat. Prod. Rep., 2015, 32, 1280–1302 RSC.
  2. S. Murarka, Z.-J. Jia, C. Merten, C.-G. Daniliuc, A. P. Antonchick and H. Waldmann, Angew. Chem., Int. Ed., 2015, 54, 7653–7656 CrossRef CAS.
  3. G. Berkhan, C. Merten, C. Holec and F. Hahn, Angew. Chem., Int. Ed., 2016, 55, 13589–13592 CrossRef CAS PubMed.
  4. L. M. Schneider, V. M. Schmiedel, T. Pecchioli, D. Lentz, C. Merten and M. Christmann, Org. Lett., 2017, 19, 2310–2313 CrossRef CAS.
  5. C. Merten, Phys. Chem. Chem. Phys., 2017, 19, 18803–18812 RSC.
  6. C. H. Pollok and C. Merten, Phys. Chem. Chem. Phys., 2016, 18, 13496–13502 RSC.
  7. C. H. Pollok, T. Riesebeck and C. Merten, Angew. Chem., Int. Ed., 2017, 56, 1925–1928 CrossRef CAS PubMed.
  8. C. L. Covington and P. L. Polavarapu, J. Phys. Chem. A, 2013, 117, 3377–3386 CrossRef CAS.
  9. J. Shen, C. Zhu, S. Reiling and R. Vaz, Spectrochim. Acta, Part A, 2010, 76, 418–422 CrossRef PubMed.
  10. T. Kuppens, K. Vandyck, J. van der Eycken, W. Herrebout, B. van der Veken and P. Bultinck, Spectrochim. Acta, Part A, 2007, 67, 402–411 CrossRef PubMed.
  11. P. J. Stephens, J. Phys. Chem., 1985, 89, 748–752 CrossRef CAS.
  12. P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch, J. Phys. Chem., 1994, 98, 11623–11627 CrossRef CAS.
  13. P. J. Stephens, Theor. Chem. Acc., 2008, 119, 5–18 Search PubMed.
  14. J. Bloino and V. Barone, J. Chem. Phys., 2012, 136, 124108 CrossRef.
  15. J. Bloino, M. Biczysko and V. Barone, J. Phys. Chem. A, 2015, 119, 11862–11874 CrossRef CAS PubMed.
  16. C. Cappelli, J. Bloino, F. Lipparini and V. Barone, J. Phys. Chem. Lett., 2012, 3, 1766–1773 CrossRef CAS PubMed.
  17. C. Merten, J. Bloino, V. Barone and Y. Xu, J. Phys. Chem. Lett., 2013, 4, 3424–3428 CrossRef CAS.
  18. D. Hauchecorne and W. A. Herrebout, J. Phys. Chem. A, 2013, 117, 11548–11557 CrossRef CAS PubMed.
  19. Y. Geboes, F. De Proft and W. A. Herrebout, J. Phys. Chem. A, 2015, 119, 5597–5606 CrossRef CAS PubMed.
  20. A. Hermann, F. Trautner, K. Gholivand, S. von Ahsen, E. L. Varetti, C. O. Della Vedova, H. Willner and H. Oberhammer, Inorg. Chem., 2001, 40, 3979–3985 CrossRef CAS.
  21. P. L. Polavarapu and C. L. Covington, Chirality, 2014, 26, 539–552 CrossRef CAS PubMed.
  22. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241 Search PubMed.
  23. Y. Zhao and D. G. Truhlar, Acc. Chem. Res., 2008, 41, 157–167 CrossRef CAS PubMed.
  24. M. Biczysko, P. Panek, G. Scalmani, J. Bloino and V. Barone, J. Chem. Theory Comput., 2010, 6, 2115 CrossRef CAS PubMed.
  25. C. Puzzarini, M. Biczysko and V. Barone, J. Chem. Theory Comput., 2010, 6, 828–838 CrossRef CAS.
  26. C. Fábri, T. Szidarovszky, G. Magyarfalvi and G. Tarczay, J. Phys. Chem. A, 2011, 115, 4640–4649 CrossRef PubMed.
  27. J. Bloino, A. Baiardi and M. Biczysko, Int. J. Quantum Chem., 2016, 116, 1543–1574 CrossRef CAS.
  28. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09 (Revision E.01), Wallingford CT, USA, 2013 Search PubMed.
  29. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3094 CrossRef CAS.
  30. W. H. Pirkle, D. L. Sikkenga and M. S. Pavlin, J. Org. Chem., 1977, 42, 384–387 CrossRef CAS.
  31. T. R. Hoye, C. S. Jeffrey and F. Shao, Nat. Protoc., 2007, 2, 2451 CrossRef CAS PubMed.
  32. K. Bünnemann and C. Merten, J. Phys. Chem. B, 2016, 120, 9434–9442 CrossRef.
  33. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104 CrossRef PubMed.
  34. S. Grimme, J. Chem. Phys., 2006, 124, 034108 CrossRef.
  35. C. Merten, M. Amkreutz and A. Hartwig, Chirality, 2010, 22, 754–761 CAS.
  36. C. Merten, M. Amkreutz and A. Hartwig, Phys. Chem. Chem. Phys., 2010, 12, 11635–11641 RSC.
  37. L. A. Nafie, Vibrational Optical Actvity, John Wiley & Sons Ltd, UK, 2011 Search PubMed.
  38. N. M. Kreienborg, C. H. Pollok and C. Merten, Chem. – Eur. J., 2016, 22, 12455–12463 CrossRef CAS.
  39. M. H. Palmer, M. Biczysko, K. A. Peterson, C. S. Stapleton and S. P. Wells, J. Phys. Chem. A, 2017, 121, 7917–7924 CrossRef CAS.

Footnote

Electronic supplementary information (ESI) available: Conformational analysis, geometries, vibrational energy distribution analysis. See DOI: 10.1039/c8cp02395f

This journal is © the Owner Societies 2019