Conformational studies by dynamic NMR. 94. Cogwheel pathway for the stereomutations of durene derivatives containing the mesityl ring

Article abstract of DOI:10.1021/jo034737t

The low-temperature NMR spectra of 1,4-bis(mesitoyl)durene, 1, and of 1,4-bis(mesitylethenyl)durene, 2, reveal the presence of syn and anti rotamers at the equilibrium, their relative proportions depending on the dielectric constant of the solvent. In sol

Full text of DOI:10.1021/jo034737t

Con for m a tion a l Stu d ies by Dyn a m ic NMR. 94.1 Cogw h eel P a th w a y
for th e Ster eom u ta tion s of Du r en e Der iva tives Con ta in in g th e
Mesityl Rin g
Carmine Coluccini,2 Stefano Grilli, Lodovico Lunazzi, and Andrea Mazzanti*
Department of Organic Chemistry “A. Mangini”, University of Bologna, Viale Risorgimento,
4 Bologna 40136, Italy
mazzand@ms.fci.unibo.it
Received May 30, 2003
The low-temperature NMR spectra of 1,4-bis(mesitoyl)durene, 1, and of 1,4-bis(mesitylethenyl)-
durene, 2, reveal the presence of syn and anti rotamers at the equilibrium, their relative proportions
depending on the dielectric constant of the solvent. In solution the more stable rotamer of 1 is the
anti whereas, in the case of 2, the more stable is the syn. Depending on the crystallization solvent
employed the more (anti) and the less stable (syn) rotamers were both observed (X-ray diffraction)
in the solid state of 1. On the other hand, only the less stable rotamer (anti) was found to be present
in the solid state of 2. As shown by MM calculations, the syn-to-anti interconversion occurs via a
correlated process (cogwheel pathway) involving the mesityl-C and durene-C bond rotations: the
dynamic NMR technique yields an experimental barrier of 8.2 kcal mol-1 for 1 and 13.1 kcal mol-1
for 2. In the case of derivative 2 a second barrier, due to a second type of correlated rotation process
(torsion), was also determined (8.6 kcal mol-1). As a consequence of the restriction of this second
torsional motion the anti rotamer of 2 displays two distinguishable NMR spectra at -133 °C,
corresponding to a pair of conformers with different symmetry (anti Ci and anti C2).
In tr od u ction
the 1,4-bis(mesitoyl)durene, 1, was synthesized. This
Hindered aromatic ketones have the plane of the
carbonyl group nearly orthogonal to that of the aromatic
ring: the barriers for the corresponding aryl-CO bond
rotation could be measured by variable-temperature
NMR spectroscopy.3 Durenes (2,3,5,6-tetramethylben-
zene) substituted by two equal acyl moieties in positions
1,4 were shown, as a consequence, to display two con-
formational isomers (rotamers), depending on whether
the two carbonyl groups are in an anti or in a syn
relationship.4 When the two acyl substituents are suf-
ficiently bulky (e.g. in the case of the t-BuCO moieties)
it even has been possible to achieve a physical separation
of the two rotamers.4 Recently we have reported that
dimesityl ketones and related compounds adopt a propel-
ler-like conformation that allows for the existence of
conformational enantiomers interconverting into each
other with a quite low barrier.5
compound should in fact comprise a rotamer syn and a
rotamer anti, each of which might also comprise a
number of diastereomeric or enantiomeric helical con-
formers.
Resu lts a n d Discu ssion
1,4-Bis(m esit oyl)d u r en e (1). Molecular Mechanics
calculations (MMFF force field)6 predict the existence of
an energy minimum corresponding to the anti rotamer
that is 0.36 kcal mol-1 lower than that corresponding to
the minimum for the syn rotamer (Scheme 1, X ) O). As
shown in the scheme, the calculations also predict that
in both rotamers the CO groups are symmetrically
twisted with respect to the plane of durene (55° and 54°
in the anti and syn rotamers, respectively) and likewise
are twisted with respect to the mesityl rings (44° and
45° in the anti and syn rotamers, respectively) thus
generating a propeller-like shape with a Ci symmetry for
the anti and a C2 symmetry for the syn. According to the
nomenclature proposed by Biali et al.7 they are referred
to as Au and Su, respectively.
We thus considered the possibility of combining these
two features into the same molecule and for this purpose
(1) Part 93: Casarini, D.; Rosini, C.; Grilli, S.; Lunazzi, L.; Mazzanti,
A. J . Org. Chem. 2003, 68, 1815.
(2) In partial fulfillment of the requirements for the PhD in
Chemical Sciences, University of Bologna.
(3) Bonini, B. F.; Grossi, L.; Lunazzi, L.; Macciantelli, D. J . Org.
Chem. 1986, 51, 517. Casarini, D.; Lunazzi, L.; Pasquali, F.; Gaspar-
rini, F.; Villani, C. J . Am. Chem. Soc. 1992, 114, 6521. Casarini, D.;
Lunazzi, L.; Verbeek, R. Tetrahedron 1996, 52, 2471. Leardini, R.;
Lunazzi, L.; Mazzanti, A.; Nanni, D. J . Org. Chem. 2001, 66, 7879.
(4) Casarini, D.; Lunazzi, L. J . Org. Chem. 1996, 61, 6240.
(5) (a) Grilli, S.; Lunazzi, L.; Mazzanti, A.; Casarini, D.; Femoni, C.
J . Org. Chem. 2001, 66, 488. (b) Casarini, D.; Grilli, S.; Lunazzi, L.;
Mazzanti, A. J . Org. Chem. 2001, 66, 2757. (c) Grilli, S.; Lunazzi, L.;
Mazzanti, A. J . Org. Chem. 2001, 66, 4444.
(6) MMFF force field, as implemented in the computer package PC-
Model 7.5 (Serena Software: Bloomington, IN).
(7) Lindner, A. B.; Grynszpan, F.; Biali, S. E. J . Org. Chem. 1993,
58, 6662.
10.1021/jo034737t CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/23/2003
7266
J . Org. Chem. 2003, 68, 7266-7273

Conformational Studies by Dynamic NMR. 94
TABLE 1. Exp er im en ta l a n d Com p u ted (in P a r en th esis)
Ba r r ier s for th e Ster eom u ta tion P r ocesses in 1 a n d 2a
SCHEME 1. Rep r esen ta tion of th e Cor r ela ted
P a th w a y Com p u ted for th e An ti-to-Syn
In ter con ver sion of 1 (X ) O) a n d 2 (X ) CH2)a
syn/anti
interconversion
torsional
process
compd
1 (X ) O)
2 (X ) CH2)
8.2 (10.8)
13.1 (14.1)
<5 (4.4)
8.65 (11.1)b
8.5 (10.5)c
a
The values are in kcal mol-1 and refer to the exchange of the
more into the less stable conformer. The errors are about (0.15
b
kcal mol-1
.
These values refer to the torsional process in the
major conformer syn. c These values refer to the torsional process
in the minor conformer anti.
according to the so-called “cogwheel” mechanism.5c,9,10 As
a consequence there is only a single transition state
(shown in Scheme 1) and thus a single barrier for
interconverting the anti (Au) into the syn (Su) rotamer.
This pathway is conceptually the same as the correlated
motion referred to as the “one-ring flip” process,7,10,11 and
the corresponding barrier is computed to be 10.8 kcal
mol-1, as in Table 1.
This value is high enough as to allow one to obtain an
experimental verification, and in fact the variable-tem-
perature NMR spectra of 1 show that the three 1H methyl
signals observed at ambient temperature broaden to a
different extent when the temperature is lowered (Figure
1, top) and eventually split into a number of lines at -138
°C (Figure 1, bottom). Two closely spaced lines, having a
different intensity, are observed for the methyl groups
of durene as well as for those in the para position of the
mesitylene ring. On the other hand the signal of the
o-methyl groups of mesitylene12 displays two pairs of
lines, with the same unequal intensity as above: within
each pair, however, these lines exhibit a 1:1 intensity
ratio (Figure 1, bottom). Such features prove that the
durene-CO bond rotation has become slow enough as
to yield distinguishable spectra for the anti and syn
rotamers, the intensity of the corresponding lines reflect-
ing their different population. Furthermore the 1:1 ratio
of the o- and o′-methyl lines of mesitylene also indicates
that the mesityl-CO bond rotation has been frozen.
If the freezing of the latter motion had generated the
Au and Su conformers of Scheme 1, the methyl signals of
the durene moiety would also have appeared anisochro-
nous. To reconcile the shape computed for these conform-
ers with the experimental observation of a single line for
the durene methyl groups in both the syn and anti
rotamers, it is required that the mesityl rings still
undergo a rapid rotation, with a less than 180° ampli-
tude, about the mesityl-CO bond. For the sole purpose
of distinguishing this lower rotation barrier from the
previous one (i.e. the one interchanging the syn and anti
rotamers) we dubbed it “torsion”. Because of this rapid
a
The hydrogen atoms are omitted for convenience and the black
atoms can be either oxygen (1) or carbon (2). The computed relative
energies are in kcal mol-1
.
To observe experimentally two different spectra for the
anti and syn rotamers it is necessary to render the
durene-CO bond rotation sufficiently slow in the NMR
time scale. The barrier for this process would correspond
to the energy difference between the transition state,
having the carbonyl coplanar8 with the durene ring, and
the ground state of the more stable rotamer (i.e. Au with
Ci symmetry). If this barrier is high enough as to be
amenable to NMR observation at an accessible low
temperature, all the signals of the anti would become
distinguishable from those of the syn: the two sets of
signals would obviously have a different intensity. An
analogous barrier should also occur for the mesityl-CO
bond rotation. The shape of the latter transition state
would differ, however, from the previous one in that the
carbonyl should be orthogonal to the mesityl ring.
Accordingly, if this rotation process is likewise frozen
at an appropriate low temperature, the two o-methyl
groups (positions 2,6) of the mesityl ring would become
diastereotopic and should consequently display two aniso-
chronous lines with the same intensity. Such a situation
should be observable in both the anti and syn rotamers.
These two rotation processes, however, are unlikely to
be independent of each other and are thus expected to
undergo a correlated pathway,7 sharing a common tran-
sition state, due to the crowding of the methyl substit-
uents. Actually, when the three-dimensional energy
surface is computed (MMFF force field6) as a function of
these angles, it appears that driving the durene-CO
dihedral angle forces also the mesityl-CO dihedral angle
to be simultaneously driven in a disrotatory manner,
(9) (a) Mislow, K. Acc. Chem. Res. 1976, 9, 26, (b) Mislow, K.
Chemtracts Org. Chem. 1989, 2, 151. (c) Glaser, R. In Acyclic Orga-
nonitrogen Stereochemistry; Lambert, J . B., Takeuchi, Y., Eds.; VCH:
New York, 1992; Chapter 4, p 123.
(10) Biali, S. E.; Nugiel, D. A.; Rappoport, Z. J . Am. Chem. Soc. 1989,
111, 846.
(11) Grilli, S.; Lunazzi, L.; Mazzanti, A.; Mazzanti, G. J . Org. Chem.
2001, 66, 748 and references quoted therein.
(12) This signal was unambiguously assigned by a 2D heteronuclear
correlation (gHMBC sequence) showing the presence of a three-bond
long-range coupling between these methyl hydrogens and the aromatic
CH carbon of mesitylene. This coupling was not observed in the case
of the methyl hydrogens of durene.
(8) This barrier is sometimes referred to as the π-barrier to indicate
the possibility of conjugation, due to the planarity of the transition
state. In the present case conjugation can occur in the transition state
between the CdO substituent and the coplanar aromatic ring. See:
Anderson, J . E.; Casarini, D.; Lunazzi, L. Tetrahedron Lett. 1988, 29,
3141.
J . Org. Chem, Vol. 68, No. 19, 2003 7267

Coluccini et al.
uncertainty, that the two processes are not independent
but share a common transition state and thus exhibit a
unique barrier, as expected on the basis of the correlated
pathway predicted by computations. The experimental
∆Gq value for this process (Table 1) was found to be 8.2
kcal mol-1 15
in the Eyring equation a transmission
:
1
coefficient of /2 was used to account for the fact that
rotation of either of the two durene-CO bonds allows
interconversion13,16 (although the computations overes-
timate this barrier by 2.6 kcal mol-1, the difference is
quite acceptable given the approximations involved in the
MM approach). It should also be mentioned that the ∆Gq
value measured for 1 is lower than that reported for the
same type of anti-to-syn interconversion occurring in less
hindered durene derivatives:4 for instance in the 1,4-
diacetyl durene the barrier is 10.9 kcal mol-1. This
further supports the existence of correlated rotations in
1 since this type of pathway renders more facile the
interconversion processes.
The low-temperature NMR spectra of 1 do not show
evidence of the propeller shape that, according to calcula-
tions, is adopted by the two rotamers of 1. To observe
this situation both the lower barrier motion (torsion)
about the mesityl-CO bond and that about the durene-
CO bond should be frozen. The first type of motion has a
transition state with the mesityl and carbonyl planes
parallel with each other whereas the second has a
transition state with the plane of durene orthogonal to
that of carbonyl. If the torsional process about the
durene-CO bond is frozen, the methyl groups of durene
in position 2,6 will become diastereotopic in both the syn
and anti rotamers (the same will also occur for the methyl
in positions 3,5). As a consequence the corresponding
NMR signals would split into a 1:1 pair, providing
evidence for the helical shape. If also the torsional process
about the mesityl-CO bond is simultaneously frozen,
additional conformers are expected to be obtained.
According to MM calculations,6 these so-called torsional
processes are also correlated and, as a consequence, two
possible energy minima for the anti and two for the syn
rotamer are expected. In other words, in addition to the
meso Au rotamer (Ci) of Scheme 1, the freezing of the two
correlated torsional processes would yield a second anti
conformer (Al, C2 racemic),7 which is 3.4 kcal mol-1 less
stable. Likewise, in addition to the racemic Su rotamer
(C2) it is also possible to have a second syn conformer
(Sl, Cs meso)7 that is 3.5 kcal mol-1 less stable.
F IGURE 1. 1H NMR spectrum (400 MHz in CHF2Cl/acetone-
d6) of the methyl region of 1. The three methyl lines observed
at -37 °C (top trace) split at -138 °C (bottom trace), showing
the presence of the anti (a) and syn (s) rotamers in a 55:45
proportion (the signal of acetone at about 2.06 ppm has been
omitted for the sake of clarity).
torsional process the dynamic symmetry for the anti
rotamer should be actually considered of C2h type and
that for the syn of C2v type.
To assign the more intense of the two low-temperature
NMR spectra to the anti or to the syn structure we
considered that the proportion of the latter would in-
crease with the dielectric constant of the environment
because the dipole moment of the syn rotamer is expected
to be larger than that of its anti companion. Indeed we
observed that in a CHF2Cl/C6D6 solution the rotamer
ratio is 1.8:1 at -138 °C, whereas in the more polar CHF2-
Cl/acetone-d6 solution the ratio becomes 1.2:1 at the same
temperature (Figure 1). Having increased its population
in a more polar medium, the minor rotamer should
correspond to the syn structure:13 this conclusion agrees
with the prediction of the MM calculations for Au and Su
of Scheme 1.
Satisfactory line shape simulation of the spectra of 1
could be achieved,14 in the temperature range -80 to
-110 °C, by using the same value for the rate constant
that exchanges the signals involved in the mesityl-CO
rotation (i.e. the 1:1 lines for the o- and o′-methyl groups
of mesitylene) as well as the signals involved in the anti-
to-syn exchange (i.e. all other lines having a different
intensity ratio). This confirms, within the experimental
(15) As was often observed in conformational processes the ∆Gq
value is independent of temperature, indicating a negligible ∆Sq value.
See: Hoogosian, S.; Bushweller, C. H.; Anderson, W. G.; Kigsley, G.
J . Phys. Chem. 1976, 80, 643. Lunazzi, L.; Cerioni, G.; Ingold, K. U. J .
Am. Chem. Soc. 1976, 98, 7484. Bernardi, F.; Lunazzi, L.; Zanirato,
P.; Cerioni, G. Tetrahedron 1977, 33, 1337. Lunazzi, L.; Magagnoli,
C.; Guerra, M.; Macciantelli, D. Tetrahedron Lett. 1979, 3031. Ander-
son, J . E.; Tocher, D. A.; Casarini, D.; Lunazzi, L. J . Org. Chem. 1991,
56, 1731. Cremonini, M. A.; Lunazzi, L.; Placucci, G.; Okazaki, R.;
Yamamoto, G. J . Am. Chem. Soc. 1992, 114, 6521. Borghi, R.; Lunazzi,
L.; Placucci, G.; Cerioni, G.; Foresti, E.; Plumitallo, A. J . Org. Chem.
1997, 62, 4923.
(16) Casarini, D.; Lunazzi, L.; Mazzanti, A. J . Org. Chem. 1997, 62,
7592. Casarini, D.; Lunazzi, L.; Mazzanti, A. J . Org. Chem. 1998, 63,
4991. Bonini, B. F.; Fochi, M.; Lunazzi, L.; Mazzanti, A. J . Org. Chem.
2000, 65, 2596. Grilli, S.; Lunazzi, L.; Mazzanti, A. J . Org. Chem. 2000,
65, 3563. Dell’Erba, C.; Gasparrini, F.; Grilli, S.; Lunazzi, L.; Mazzanti,
A.; Novi, M.; Pierini, M.; Tavani, C.; Villani, C. J . Org. Chem. 2000,
65, 1663. Grilli, S.; Lunazzi, L.; Mazzanti, A.; Pinamonti, M. J . Org.
Chem. 2002, 67, 5733.
(13) Casarini, D.; Lunazzi, L. J . Org. Chem. 1996, 61, 6240.
(14) QCPE program no. 633, Indiana University: Bloomington, IN.
7268 J . Org. Chem., Vol. 68, No. 19, 2003

Conformational Studies by Dynamic NMR. 94
SCHEME 2. Sch em a tic Rep r esen ta tion of th e
Cor r ela ted Tor sion a l P r ocesses Occu r r in g in th e
An ti a n d Syn F or m s of 1 (X ) O) a n d 2 (X ) CH2)a
F IGURE 2. X-ray diffraction (single crystals obtained from
ethanol) shows that the crystal cell of 1 comprises two slightly
different syn helical structures (Su type), identified as A and
B (see text).
a
The second mesitylene ring has been indicated as Mes for
simplicity. The computed relative energies are in kcal mol-1
.
These conformers, and the computed transition states
through which their exchange takes place, are displayed
in Scheme 2. The large energy difference between the
ground states predicted by calculations (3.4 kcal mol-1
)
suggests, however, that only the more stable rotamers
Au and Su are likely to be populated in the case of 1.
Accordingly, the only detectable feature due to the
freezing of the torsional process would be the 1:1 splitting
of the durene methyl line.
Computations predict, however, that the energy of the
transition state for the correlated torsion pathway is only
4.4 kcal mol-1 higher than that of the ground state (Table
1), a value conceivably too low to produce any detectable
effect on the NMR spectra. Thus the absence of 1:1
splitting of the methyl durene lines in the NMR spectrum
of 1, even at -180 °C, suggests an experimental barrier
of less than 5 kcal mol-1 (Table 1).17
F IGURE 3. X-ray diffraction structure of 1 (single crystals
obtained from chloroform) showing the anti (Ci meso) helical
structure (Au type).
In view of the inconclusive result of the latter NMR
experiment in solution, the proof that this molecule
actually adopts a propeller-like shape had to be searched
by X-ray diffraction in the solid state. When the com-
pound was crystallized from ethanol, which has a high
dielectric constant, the structure observed is that of the
more polar conformer Su (C2 racemic). The diffraction
data also indicate that there are two independent syn
(racemic) structures having slightly different angles and
bond lengths (Figure 2).
like shape, which consequently entails the existence of
two nonsuperimposable antipodes: in the crystal cell of
1 there are in fact two pairs of enantiomerically related
molecules (see Supporting Information).
On the other hand, when the crystals of 1 are obtained
from chloroform, which has a much lower dielectric
constant, the X-ray diffraction indicates that the struc-
ture is that of a meso anti (Ci symmetry), also displaying
a propeller-like shape (Figure 3), very similar to the
computed one. The possibility of detecting the structures
of both rotamers by using different crystallization sol-
vents is probably a consequence of the fact that in more
polar solutions the rotamer syn is definitely more abun-
dant that in the less polar solutions, whereas the opposite
occurs for the rotamer anti.
Both structures have dihedral angles extremely similar
to the computed ones, thus giving evidence of a propeller-
(17) Obviously it cannot be excluded that the negative results might
be a consequence of a separation of the corresponding lines being
smaller than the line width, which is quite large at -180 °C (about
40-50 Hz in a CHF2Cl/CHFCl2/C6D6 solution) owing to the increased
viscosity.
J . Org. Chem, Vol. 68, No. 19, 2003 7269

Coluccini et al.
F IGURE 4. MM computed energy (kcal mol-1) surface of 2
as function of the angles ϑ and æ. The four minima, corre-
sponding to the four possible conformers, are indicated. The
three bent arrows refer to the three conformers whose spectra
were experimentally observed.
1,4-Bis(m esityleth en yl)d u r en e (2). It is conceivable
to predict that substitution of the CdO with the bulkier
CdCH2 moiety should greatly increase the barrier for all
the stereomutation processes, and for this purpose we
synthesized 1,4-bis(mesitylethenyl)durene, 2.
F IGURE 5. (a) 1H NMR (400 MHz) methyl region spectra of
2 in C2Cl4 at +100 °C. The o- and p-methyl signals are those
of the mesitylene moiety. (b-d) Temperature dependence of
the same region in CHF2Cl/CHFCl2/C6D6. At -60 °C (trace b)
the o-methyl signal of mesitylene splits into a pair of more
intense and of less intense lines (rotamers syn and anti,
respectively). At -104 °C (trace c) the signals of the four
methyl groups in positions 2,3,5,6 of durene split into a pair
of lines arbitrarily identified as 3,5 and 2,6. At -133 °C (trace
d) the lines with the same symbol refer to the same type of
methyl group: the major lines are singlet (syn C2) and the
minor lines doublets of slightly unequal intensity (anti C2 and
anti Ci). Impurities are identified by the letter i.
As shown in Schemes 1 and 2, the MM computed6
barrier for the syn-to-anti interconversion is 14.1 kcal
mol-1 and those for the so-called torsional process are
11.1 and 10.5 kcal mol-1 for the syn and anti rotamers,
respectively. As it appears from the computed tridimen-
sional energy surface reported in Figure 4, all these are
correlated processes (cogwheel pathway7,9) because they
follow a diagonal pathway. In particular both the higher
energy syn-to-anti interconversion and the lower energy
correlated torsional processes are predicted to follow a
disrotatory one-ring-flip pathway,5c,9-11 since the latter
appears to have a computed barrier that is lower than
the barrier of the other possible process (i.e. the two-ring-
flip mechanism).5c, 9-11
That the correlated torsional process of compounds 1
and 2 corresponds to a one-ring-flip pathway, as predicted
by the theory, also can be deduced from experimental
evidence. In fact the observation of anisochronous lines
for the o-methyl groups of the mesitylene ring is compat-
ible only with the dynamic symmetry of the one-ring-flip
process: in the case of a two-ring-flip process these lines
would have appeared isochronous. Although these com-
putations might have the tendency of overestimating the
barriers, the computed values for 2 are predicted to be
high enough as to be amenable to an experimental
verification. It is thus expected that also the correlated
torsion, which had escaped detection in the case of 1,
should become NMR visible in the more hindered com-
pound 2.
1
Even at ambient temperature the H NMR spectrum
of 2 shows a number of exchange-broadened signals. On
cooling to about -60 °C the single signal of the four
o-methyl groups of mesitylene12 splits into a 1:1 pair of
more intense lines and a 1:1 pair of less intense lines
(Figure 5). This indicates that the correlated rotation
about the durene-CCH2 and mesityl-CCH2 single bonds
has been frozen, generating the anti (Ci symmetry) and
syn (C2 symmetry) rotamers displayed in Scheme 1 for
X ) CH2 (however, the occurrence at this temperature
of a rapid torsional process would make the dynamic
symmetry of the anti be considered of C2h type and that
of the syn rotamer of C2v type). This feature is the same
as observed in the case of 1, but it occurs at a higher
temperature, in agreement with the expectation of a
larger syn-anti interconversion barrier in the case of 2.
The methyl groups of durene and the methyl groups in
the para position of mesitylene yield single signals since
the corresponding shift differences for the two rotamers
are lower than the line width.18
7270 J . Org. Chem., Vol. 68, No. 19, 2003

Conformational Studies by Dynamic NMR. 94
SCHEME 3a
Contrary to the case of 1, calculations had predicted
that the Su rotamer of 2 is more stable (by 0.44 kcal
mol-1) than its Au companion (see Scheme 1, X ) CH2)
and for this reason the more intense pair of o-,o′-methyl
lines of mesitylene in Figure 5 were labeled syn and the
less intense were labeled anti. This assignment was
confirmed by the observation that the rotamer ratio is
1.8 in toluene-d8 and 1.75 in CS2 at -60 °C but 2.9 and
3.2 in CD2Cl2 and acetone-d6, respectively, at the same
temperature.
This trend is opposite to that found in the case of 1 in
that the proportion of the major rotamer increases with
an increase in the dielectric constant19 of the solvent, as
expected if the predominant rotamer has the syn struc-
ture.
a
The numbers refer to the experimental barriers (in kcal mol-1
)
for the interconversion of the more into the less stable conformers
of compound 2.
(lines at 2.78, 2.82 ppm arbitrary labeled as positions 3,5),
for the minor signal of the methyl groups of durene (lines
at 2.04, 2.05 ppm arbitrarily labeled as positions 2,6),
and for the minor signal of the o-methyl groups of
mesitylene (arrowed lines at 1.88, 1.95 ppm).20
Line shape analysis allowed us to obtain the rate
constants for the syn-to-anti interconversion at various
temperatures 15 that yielded an activation energy of 13.1
1
kcal mol-1 (again a
/ transmission coefficient was
2
employed13,16), in fair agreement with the theoretical
prediction of 14.1 kcal mol-1. As anticipated, this barrier
is definitely higher than that measured for the analogous
process in 1.
On the basis of the theoretical predictions concerning
the relative stability of the conformers of 2, as reported
in Scheme 2, the computed energy difference between the
two syn conformers Su, Sl is quite high (1.4 kcal mol-1),
indicating that only Su should be essentially populated.
As a consequence, when the correlated torsional process
is frozen, the major signals would correspond only to
those of Su, the population of Sl being negligible. In this
case evidence of the frozen torsional process is given
solely by the anisochronicity of the durene methyl lines.
On the other hand, the computed energy difference
between the two anti conformers Au, Al is significantly
When the spectrum of 2 is obtained at temperatures
lower than -60 °C, the signal of the four methyl groups
of durene broadens considerably and eventually splits
into a pair of 1:1 lines at -104 °C (Figure 5) due to the
torsional process about the durene-CCH2 bond becoming
slow in the NMR time scale (as discussed for the case of
1 this process is conceivably correlated with the analo-
gous torsional process occurring about the mesityl-CCH2
bond).
Such a feature is in keeping with the propeller-like
conformation predicted by theory, where the pair of
methyl groups in positions 2,6 of durene (and likewise
those in positions 3,5) are expected to be diastereotopic
in the ground state conformation.
lower (its value is almost halved, being 0.8 kcal mol-1
)
and this makes plausible that, in addition to the aniso-
chronicity of the durene methyl signals, the signals of
both species become visible when the correlated torsional
process is frozen at low temperature.
Line shape analysis provided a barrier of 8.65 kcal
mol-1 for the correlated torsional process involving the
major rotamer syn (Table 1) and 8.5 kcal mol-1 for the
same process involving the minor rotamer anti. Within
the experimental uncertainty these values are obviously
equal: the difference of 0.6 kcal mol-1 predicted by the
theory for these two barriers (i.e. 11.1 and 10.5 kcal
mol-1, as in Table 1) is thus slightly overestimated and
is a consequence of the approximations involved in the
computations. The values computed for this process,
however, are in the same range as those determined
experimentally and correctly predict that this barrier is
lower than that for the syn-to-anti interconversion, as
observed experimentally.
At about -133 °C a further splitting of some of the
minor signals also becomes visible. This is a consequence
of the same correlated torsional process mentioned above.
This effect is visible at a lower temperature, with respect
to the 1:1 separation of the methyl durene signals,
because the corresponding chemical shift differences are
much smaller. As shown in Scheme 2, when the cor-
related torsional process is frozen, not only the methyl
groups of durene become diastereotopic but also a pair
of conformers is expected to be generated for the anti (Ci
and C2 symmetry) and for the syn (C2 and Cs symmetry)
arrangement. In other words the dynamic anti symmetry
(C2h) and the dynamic syn symmetry (C2v) are broken at
low temperature so that the static symmetries of the
ground state conformers can be revealed.
In the spectrum of Figure 5 at -133 °C, however, the
more intense signals due to the syn rotamer do not
display such additional splitting, which is detectable only
for the minor signals of the rotamer anti. This splitting
is particularly evident for the minor signal of the o′-
methyl groups of mesitylene (starred lines at 2.83, 2.85
ppm), for the minor signal of the methyl group of durene
Borrowing from the theoretical scheme proposed by
Biali et al,7 the results obtained for 2 are represented
(see Scheme 3) as an experimental version of their graph,
where each vertex corresponds to a conformer and each
edge to a barrier (Sl is placed in square brackets to
indicate that it was not experimentally detected).
The single-crystal X-ray structure of 2 shows that the
rotamer observed in the solid state is Au, with a Ci
symmetry (see Figure 6). The structure present in the
(18) In the case of 1 these separations were barely detectable, being
only 0.005 ppm for the p-methyl signals of mesitylene and 0.006 ppm
for the methyl signals of durene (see Figure 1).
(19) Landolt-Bo¨rnstein, Band II, Teil 6; Springer-Verlag: Berlin,
Germany, 1959.
(20) The two minor lines of the p-methyl group of mesitylene (at
2.55 ppm) are so close as to make their separation barely distinguish-
able.
J . Org. Chem, Vol. 68, No. 19, 2003 7271

Coluccini et al.
MM calculations confirm that the structures of the more
and of the less stable rotamers correspond to those
determined by NMR in solution. The syn-anti intercon-
version barriers of rotamers of 1 and of 2 were experi-
mentally determined by line shape analysis of the
temperature-dependent NMR spectra, and the corre-
sponding pathways are believed to occur via a cogwheel
mechanism, identified as the one-ring-flip process. In the
case of the hydrocarbon 2 it was observed that at lower
temperatures another motion (dubbed correlated torsion)
takes place with the same mechanism, and a value for
this second barrier was experimentally determined.
Exp er im en ta l Section
Ma t er ia ls. 1,4-Bis(m esit oyl)-2,3,5,6-t et r a m et h ylb en -
zen e (1). 1,4-Dibromodurene (7 mmol in 40 mL of dry THF)
was first reacted with n-BuLi at -78 °C (7 mmol, solution 1.6
M in hexane), then with mesitylaldehyde (10 mmol in 5 mL of
dry THF). After the temperature was raised, the mixture was
treated with NH4Cl, extracted with ether, and dried (Na2SO4)
to give the 1-bromo-4-methylhydroxymesityl-2,3,5,6-tetram-
ethylbenzene. The intermediate was treated again with n-BuLi
at -78 °C and then with mesitylaldehyde to give, after workup,
1,4-bis(methylhydroxymesityl)-2,3,5,6-tetramethylbenzene. The
diol (4 mmol in 40 mL of CH2Cl2) was then treated with PCC
(10 mmol). After 1 h, Et2O was added to the mixture and the
solution was filtered on decalite and concentrated at reduced
pressure. The crude product was purified by use of silica gel
chromatography (petroleum ether/Et2O 30/1) to give 1 (2.5
mmol, overall yield over 1,4 dibromodurene 35%). Single
crystals suitable for X-ray diffraction were obtained by slow
evaporation from absolute ethanol or chloroform.
1H NMR (CDCl3, 400 MHz) δ 2.03 (12H, s), 2.19 (12H, s),
2.29 (6H, s), 6.87 (4H, s). 13C NMR (CDCl3, 100.6 MHz) δ 17.63
(CH3), 21.71 (CH3), 22.24 (CH3), 130.56 (CH), 132.19 (Cq),
137.23 (Cq), 138.2 (Cq), 141.13 (Cq), 144.22 (Cq), 202.6 (CO).
Anal. Calcd for C30H34O2: C, 83.79; H, 9.23. Found: C, 83.84;
H, 9.15. Mp 184-186°C (from ethanol).
1,4-Bis(m esitilvin yl)-2,3,5,6-tetr a m eth ylben zen e (2). A
solution of 1 (2.5 mmol in 15 mL of dry hexane) was treated
with MeLi (8 mmol in Et2O) and heated to reflux for 3 h.24
The mixture was subsequently quenched with water, extracted
F IGURE 6. Single-crystal X-ray structure of 2 showing the
propeller-like shape of the anti Ci rotamer (Au type).
crystal corresponds to that of the less abundant rotamer
observed at the equilibrium in solution.
The X-ray structural determination, combined with the
knowledge of the syn-to-anti interconversion barrier (13.1
kcal mol-1), offers the opportunity of testing in an
independent way the assignment of the syn structure to
the more abundant rotamer in solution. From the mea-
sured barrier, in fact, the half-lifetimes of the rotamers
are estimated to be close to 5 min at about -85 °C. Thus
the very same crystal of 2 used for X-ray diffraction
(which contains only the rotamer anti) was dissolved in
CD2Cl2 at this temperature21 and the sample transferred
as quickly as possible into the probe of the NMR
spectrometer, cooled to -90 °C. In this way the system
should not have sufficient time to reach a complete
equilibrium and a proportion of the anti rotamer, greater
than that expected at the equilibrium, would conse-
quently be observed.
Indeed we measured in this experiment an initial 3.5:1
ratio, and observed a decreasing of the proportion of the
minor rotamer with time, until a 5.2:1 ratio22 was
obtained when the equilibrium conditions were finally
achieved. This allowed us to identify the minor rotamer
as that having the same structure (anti) observed in the
single crystal by X-ray crystallography, a result in
agreement with the previous assignment.
with Et2O, dried (Na2SO4), and purified on
a silica gel
chromatography column (petroleum ether/Et2O 10/1) to obtain
1,4-bis(1-hydroxyethylmesityl)-2,3,5,6-tetramethylbenzene. The
intermediate diol (0.35 mmol in 20 mL of CHCl3) was then
treated with P2O5 (5 mmol) for 15 h and the solution was
filtered and concentrated at reduced pressure. Crystallization
from acetone yielded 2 as a white solid. Single crystals suitable
for X-ray diffraction were obtained by slow evaporation from
acetone.
Con clu sion s
1H NMR (CDCl3, 400 MHz) δ 2.10 (12H, s), 2.22 (6H, s),
5.50 (2H, d, J ) 4.0 Hz), 5.60 (2H, d, J ) 4.0 Hz), 6.81 (4H, s).
13C NMR (CDCl3, 100.6 MHz) δ 18.95 (CH3), 20.83 (CH3), 22.0
(CH3), 125.37 (CH2), 129.61 (CH), 132.78 (Cq), 135.83 (Cq),
In both compounds 1 and 2 we have been able to
observe, by X-ray diffraction, the crystal structure of the
less stable of the two rotamers (syn and anti, respectively)
detected in solution by low-temperature NMR spectros-
copy. Although other examples of this type have been
reported,23 such a feature is not that usually observed.
136.33 (Cq), 141.28 (Cq), 144.91 (Cq). Anal. Calcd for C32H38
:
C, 89.80; H, 10.20. Found: C, 89.82; H, 10.18. Mp 190-192
°C.
NMR Mea su r em en ts. The assignment of the 13C spectra
was carried out by DEPT and 2D heteronuclear correlation
(gHMBC sequence). The samples for the low-temperature
measurements were prepared by connecting the NMR tubes
containing the compound and some deuterated solvent for
(21) Methylene chloride was selected because, even at temperatures
in the range -80 to - 85 °C, it is capable of dissolving compound 2
quite rapidly and is still fluid enough as to yield sufficiently resolved
spectra for distinguishing the signals of the syn from those of the anti
form. At lower temperatures the time required to dissolve the
compound was too long, thus allowing the equilibrium to be established
before the sample could be transferred into the NMR probe.
(22) The conformer ratio for the equilibrium in CD2Cl2 increases
from 2.9:1 to 5.2:1 by lowering of the temperature from -60 to -90
°C. The more stable species increases its proportion not only according
to the Boltzamann equation but also because of the larger value of
the dielectric constant at lower temperatures.19
(23) Taha, M.; Marks, V.; Gottlieb, H. E.; Biali, S. E. J . Org. Chem.
2000, 65, 8621. Marks, V.; Gottlieb, H. E.; Melman, A.; Byk, G.; Cohen,
S.; Biali, S. E. J . Org. Chem. 2001, 66, 6711. Marks, V.; Nahmany,
M.; Gottlieb, H. E.; Biali, S. E. J . Org. Chem. 2002, 67, 7898.
(24) Roberts, R. M.; El-Khawaga, A. M.; Roengsumram, S. J . Org.
Chem. 1984, 49, 1380.
7272 J . Org. Chem., Vol. 68, No. 19, 2003

Conformational Studies by Dynamic NMR. 94
locking purpose to a vacuum line and condensing therein the
gaseous solvents (CHF2Cl and CHFCl2) by means of liquid
nitrogen. The tubes were subsequently sealed in vacuo and
introduced into the precooled probe of the spectrometer. The
temperatures were calibrated by substituting the sample with
a precision Cu/Ni thermocouple before the measurements.
Complete fitting of the dynamic NMR line shape was carried
out with use of a PC version of the DNMR-6 program.14
the University of Bologna (Funds for selected research
topics 2001-2002 and RFO) and from MIUR-COFIN
2001, Rome (national project “Stereoselection in Organic
Synthesis”).
Su p p or tin g In for m a tion Ava ila ble: Crystallographic
data and ORTEP drawings for compounds 1 and 2. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Ack n ow led gm en t. The authors gratefully thank
Prof. S. E. Biali, Hebrew University, J erusalem for very
helpful comments. Financial support was received from
J O034737T
J . Org. Chem, Vol. 68, No. 19, 2003 7273