Report by Prof

Report by Prof. Ivo Leito

29.07.2008

University of Bremen

Institute of Inorganic and Physical Chemistry

Goals of the Work

1. To computationally (at DFT B3LYP 6-311+G** level) investigate the thermodynamic stability and acidity of the polyfluorinated and polytrifluoromethylated isomers of C6H5OH and C6H5H (derivatives of benzvalene,1 prismane,2 and Dewar benzene3, Scheme 1) in comparison with the corresponding derivatives of phenol and benzene and the dependence of their acidity on the structure.
2. To elucidate and compare the acidifying effect and its mode of action in the case of polytrifluoromethylation with that of polyfluorination.
3. To prepare material for publication.

Scheme 1. Structures of benzene (0), prismane (1), Dewar Benzene (2) and benzvalene (3).

Results of the Work

1. Investigation of stability and acidity of the compounds

The results of stability and acidity investigations are presented in Table 1.All neutrals calculated were stable with respect to the geometry optimization of the computational method used. However, in some of the anions significant bond elongations or bond ruptures took place. All anions of 2 with acidity center in position 1 underwent rearrangement (rupture of the 1-4 bond) to give the corresponding phenyl derivatives. Several anions of the monosubstituted hydroxy-derivatives of 3 also rearranged into monosubstituted phenolate ions. This is evident from the very large apparent acidifying effects, e.g. with compounds 1-OH-F5-2, 1-OH-(CF3)5-2, 5-OH-1-CF3-3 or 5-OH-2-CF3-3. However, on protonation of these anions a neutral different from the original one will form. Thus, the process is not a reversible Brønsted acid-base equilibrium and such acidities can only be called apparent. Due to this, the anions with disrupted bonds were left out of discussion.

Table 1. Results of computations (all values in kcalmol-1).

Stability of the derivatives of 1-3

The unsubstituted hydrocarbons 1-3 are significantly less stable than benzene. Prismane as the most strained of them is by 122.6 kcalmol-1 less stable than benzene. For Dewar's benzene and benzvalene the relative (in)stabilities are 82.8 and 80.4 kcalmol-1, respectively. Relative stabilities of the alcohols are very similar to that of the hydrocarbons.

Relative stability of the pentafluorinated 1 to 3 with respect to pentafluorobenzene is still worse: 131.0 kcalmol-1 in the case of 1-H-F5-1.

At the same time, introducing multiple CF3 substituents into the hydrocarbons changes the stability differences considerably. Thus (CF3)5-Prismane is by 77.8 kcalmol-1 less stable than (CF3)5-0. 1-H-(CF3)5-2 is by 41.1 kcalmol-1 less stable than (CF3)5-0. For the different isomers of (CF3)5-3 the differences range from 47.7 to 52.0 kcalmol-1. The reason to this is that introduction of the substituents into the already strained and partly globular hydrocarbons does not markedly increase steric strain, while introducing the same substituents into benzene ring leads to significant additional steric strain.4 The same relative stability energies for the respective hydroxy-derivatives C6(CF3)5OH of 1 to 3 are even by couple of kcalmol-1 less positive. This can be due to the destabilizing interaction energy of 3.8 kcalmol-1 in the neutral 1-OH-(CF3)5-0.4

Relative stabilities of the anions are in some cases significantly different from those of neutrals. The anions of hydroxy-derivatives of 1-3 are in most cases more destabilized (compared to phenolate anion), than the corresponding neutrals (compared to phenol). The reason is obviously the missing resonance between the –O– center and the aromatic ring in 1-3.

The anions of fluorinated hydroxy-derivatives of 1-3 have significantly higher relative stability than their non-fluorinated parents. It has been demonstrated that polyfluorination stabilizes the anions of strained alcohols.5 First of all this occurs via the partial transfer of electron density from the nonbonding orbitals of the –O– center to the * orbitals of the adjacent C-C bonds of the hydrocarbon skeleton. This effect causes elongation of the C-C bonds and partial release of the steric strain.

Acidity of the substituted hydrocarbons 1-3

As seen in Table 1 similarly to benzene the unsubstituted hydrocarbons 1 to 3 are weak acids. Nevertheless, for example, 1-H-2is by ca 10 kcalmol-1 more acidic than benzene.

Pentafluorosubstitution increases the acidity of benzene by 46.8 kcalmol-1. In most of the pentafluorinated hydrocarbons 1-3 the acidifying effect is higher.

Substitution of 1-3 by five CF3 groups leads, somewhat surprisingly, to similar acidity increase as in the case of five fluorine substituents. Although CF3 is a reasonably strong electronacceptor group due to its hyperconjugation effect (R = 0.096) its field-inductive effect (F = 0.466) is weaker than that of the –F substituent. As opposed to –F this set of properties makes -CF3 a very efficient group in acidifying compounds with aromatic ring. At the same time in the cage-type structures of 1 to 3 it is less than optimal due to the loss of the resonance acceptor effect.

Acidity of the hydroxy-derivatives of 1-3

Hydroxy derivatives of 1 to 3 are distinctly more acidic than the hydrocarbons itself and have acidities similar to that of phenol. Because the strong resonance between the aromatic ring and the –O– center that stabilizes the phenolate anion is (nearly) absent in the (predominantly) aliphatic hydroxy-derivatives of unsubstituted 1 to 3 one would expect lower acidities of their hydroxy derivatives. Table 1 reveals that this is indeed the case with most of them (considering only those where bonds are not disrupted in anions). Nevertheless, for example 2-OH-2is more acidic than phenol by 3 kcalmol-1. in 2-OH-2 OH is attached to a double bond. The conventional explanation would be that this is due to the resonance delocalization of the charge in 2-O–-2. It would, however, be surprising that this rather modest level of delocalization can compete with the aromatic ring. Therefore it is most likely that the delocalization is supported (and perhaps dominated) by the partial release of steric strain on deprotonation. Indeed, the CO bond is significantly shortened in the anion while the bonds C2-C2 and C2-C3 are elongated, thus allowing a less strained arrangement.

Substituting all five hydrogens by fluorine atoms increases the acidity of phenol by 23.1 kcalmol-1. Significantly larger acidity increase is observed in the case of the pentafluoro-derivatives of practically all investigated alcohols. The acidifying effects are often around double of that in phenol. It is of interest, whether this effect is mainly due to a destabilizing effect in the neutral or stabilizing effect in the anion (relative to pentafluorophenol). Table 1 shows that in the case of 5-OH-F5-3 both effects are strongly operational. The neutral is destabilized by almost 10 kcalmol-1 (relative to 5-OH-3 vs phenol), while the anion is stabilized by ca 9 kcalmol-1. Similar picture is observed with other investigated alcohols.

The acidifying effect of pentakis-CF3-substitution in phenol is by more that two times larger than the effect of pentafluorosubstitution. Similar picture is not observed with 1 to 3. The effects of F5- and (CF3)5-substitution are more similar and in some cases pentafluorosubstitution outperforms pentakis-CF3-substitution, e.g. in the case of 1-OH-3.

1. Comparison of the effects of polyfluorination and poly-trifluoromethylation

In aromatic systems (especially if the acidity center is conjugated to the aromatic ring) CF3 groups are significantly more efficient in enhancing acidity than F substituents. At the same time in aliphatic cage-like structures F substituents sometimes outperform CF3 substituents by their acidifying effect. This is similar both for the investigated hydrocarbons and their hydroxy-derivatives.

The reasons for the high efficiency of polyfluorination in aliphatic systems compared to aromatic systems are the following:

1. Fluorine as a substituent has a strong field-inductive effect (F = 0.576), but is at the same time a resonance donor group (R = -0.336). In the aromatic ring both effects are operational, have opposite direction and partially cancel. In the predominantly aliphatic structures 1-3 mostly the field-inductive effect works and is not counterbalanced by resonance-donative effect.
2. The geometric arrangement of the F substituents in pentafluorobenzene and pentafluorophenol is not the best possible from the point of view of acidity. The efficiency of the field-inductive effect decreases rapidly as the distance from the reaction center increases thereby reducing the activity of 3- and 4-F substituents. At the same time the 2-F substituents that are near to the reaction center, are due to the geometry of the benzene ring bent quite near to the acidity center. This causes significant charge-charge repulsion in the respective anion (especially true in –OH derivatives). Contrary to this, the geometry of some of the derivatives of 1 to 3 permits the F substituents to be arranged in a very efficient way. In particular, in both 1-OH-F5-1 and 5-OH-F5-3 there are three F substituents in the alpha-position to the reaction center and all of them are bent away from the OH group, which minimizes steric repulsion in the anions.
3. In the substituted hydrocarbons: At first sight in the carbanions studied here the charge from the anionic centre cannot be delocalized by the hyperconjugation effect. This is because the geometry of the acids is non-planar and cannot be planarized in the anions because of the Bredt's rule and because in the cage-like structures achieving a planar geometry would require major rearrangements. In a pyramidal carbanionic centre hyperconjugation is normally not effective, because of geometric impossibility of overlapping between the lone pair electrons of the anionic centre and the * orbitals of the -CF bonds.7 At the same time, comparing e.g. in pentafluoroprismane the lengths of the bonds C1-C2, C1-C5 and C1-C6, indicates slight shortening of the C1-C5 and C1-C6 bonds on deprotonation and slight deformation of the C1carbon towards planarity. Thus some contribution from hyperconjugation is not impossible.
4. In alcohols, the shortening of the C-O bonds in the anions, relative to neutrals and the lengthening of the C-C bonds and thus release of steric strain. At molecular orbital level this effect is enabled by the partial transfer of electron density from a lone pair of the –O– center to the * orbitals of the C-C bonds and this effect is absent in pentafluorophenol.

Although CF3 is a reasonably strong electronacceptor group due to its hyperconjugation effect (R = 0.096) its field-inductive effect (F = 0.466) is weaker than that of the –F substituent. As opposed to –F this set of properties makes -CF3 a very efficient group in acidifying compounds with -systems, such as aromatic ring. At the same time in the cage-type structures of 1 to 3 it is less than optimal.

3. Preparation of Material for Publication

The manuscript for the first article is nearing completion and will be submitted in the coming months to some established journal of physical or computational chemistry.

References

1. (a) Katz, T. J.; Jang Wang, E.; Acton, N. J. Am. Chem. Soc.1971, 93, 3782-3783. (b) Katz, T. J.; Roth, R. J.; Acton, N.; Jang Carnahan, E. J. Org. Chem.1999, 64, 7663-7664.
2. Katz, T. J.; Acton, N. J. Am. Chem. Soc.1973, 95, 2738-2739.
3. Van Tamelen, E. E.; Pappas, S. P. J. Am. Chem. Soc.1963, 85, 3297-3298.
4. Kütt, A.; Movchun, V.; Rodima,T.; Dansauer, T.; Rusanov, E. B.; Leito, I.; Kaljurand, I.; Koppel, J.; Pihl, V.; Koppel, I.; Ovsjannikov, G.; Toom, L.; Mishima, M.; Medebielle, M.; Lork, E.; Röschenthaler,G.-V.; Koppel, I. A.; Kolomeitsev, A. A. J. Org. Chem. 2008, 73, 2607.
5. Herrero, R., Dávalos, J. Z.; Abboud, J.-L. M.; Alkorta, I.; Koppel, I.; Koppel, I. A.; Sonoda, T.; Mishima, M. Int. J. Mass Spec. 2007, 267, 302-307.
6. Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165-195.
7. Koppel, I. A.; Pihl, V.; Koppel, J.; Anvia, F.; Taft, R. W. J. Am. Chem. Soc.1994, 116, 8654-8657.

1