Static and Dynamic Stereochemistry and Solvation of 2,2-Ditipylethenols

Article abstract of DOI:10.1021/jo971049z

The effect of increased bulk of the β-aryl group in enols Ar₂C=C(OH)R from Ar = mesityl = Mes (1) to Ar = 2,4,6-i-Pr₃C₆H₂ = Tip (2) was investigated. The solid-state structure when R = H (a) does not significant

Full text of DOI:10.1021/jo971049z

8372
J . Org. Chem. 1997, 62, 8372-8386
Sta tic a n d Dyn a m ic Ster eoch em istr y a n d Solva tion of
2,2-Ditip yleth en ols1
J oseph Frey and Zvi Rappoport*
Department of Organic Chemistry and the Minerva Center for Computational Quantum Chemistry,
The Hebrew University, J erusalem 91904, Israel
Received J une 11, 1997X
The effect of increased bulk of the â-aryl group in enols Ar2CdC(OH)R from Ar ) mesityl ) Mes
(1) to Ar ) 2,4,6-i-Pr3C6H2 ) Tip (2) was investigated. The solid-state structure when R ) H (a )
does not significantly differ for 1a and 2a . The dynamic behavior of 2a resembles that of 1a , but
the rotational barriers for both the threshold one-ring flip process and the two-ring flip process are
higher for 2a . The one-ring flip barrier for 2a is solvent-dependent. The threshold two-ring flip
barriers when R ) Me (2b) and t-Bu (2c) are higher than for the mesityl analogues, but that for
2c is higher than predicted. Solvations of 2a and 1a and their associations with DMSO are similar.
The CdCOH conformation is syn-planar with an OH‚‚‚π(Tip) association in non-hydrogen-bond-
accepting solvents and is anti-clinal with OH‚‚‚solvent association in hydrogen-bond-accepting
solvents. In summary, the increased bulk associated with the change Mes f Tip changes the
q
structure and behavior in the expected direction but, except for the ∆Gc values, not to a large
extent.
2,2-Dimesityl-1-R-ethenols (1) are stable enols whose
1b-1e,9 and the two-ring flip barriers decrease with the
Mes2CdC(OH)R
structures were proven by chemical, spectroscopic, and
crystallographic methods.2 They exist in enantiomeric
vinyl propeller conformations differing in their helicities.
Their static and dynamic behaviors were extensively
investigated2 and recently reviewed.3 X-ray diffractions
corroborated the proposed propeller conformations,4 and
molecular mechanics (MM) computations confirmed that
they are actual energy minima.5
In analogy with Mislow’s analysis6 the correlated
helicity reversal rotation of the aryl rings can be conve-
niently analyzed in terms of “flip” mechanisms. The ring
that “flips” passes through a plane perpendicular to the
ideal double bond plane while the remaining (“nonflip-
ping”) ring(s) rotate concurrently and disrotatorily and
pass through the double-bond plane. The number of
flipping rings designates the different flip mechanisms
as zero-, one-, or two-ring flip;7 e.g., if both the â and â′
rings pass through the normal to the ideal double-bond
plane, the pathway is designated [â,â′]-two-ring flip (vide
infra). A rotation pathway not resulting in helicity
reversal involves a 180° rotation of one or more rings.
The threshold rotational enantiomerization mechanism
involves a correlated rotation around the C(sp2)-C(Ar)
bond.8 It is a one-ring flip for 1a and a two-ring flip for
1a : R ) H
b: R ) Me
c: R ) Et
Mes ) mesityl
d : R ) i-Pr
e: R ) t-Bu
f: R ) Mes
increased bulk of R from 14.2 kcal mol-1 for R ) H (1a )
to 10.4 kcal mol-1 for R ) t-Bu (1e). Similar behavior
occurs with the tetra-m-Me and tetra-m-Br derivatives,10
and it was ascribed in all systems to steric effects. A
vicinal in-plane H/Mes (R/Mes) interaction is lower
(larger) than the perpendicular one in the rotational
transition state when R () alkyl, silyl) is bulkier than
H. Increased bulk of R increases the Mes-CdC torsional
angles and the ground-state energy and lowers the
rotational barrier. The proposed enantiomerization mech-
anism is also supported by a structural correlation
analysis11 of molecules Ar1Ar2CdCR1R.4b
X Abstract published in Advance ACS Abstracts, November 1, 1997.
(1) Presented in part at the 57th Meeting of the Israel Chemical
Society, Haifa, Feb 12-13, 1992, Abstract L-32 and at the 11th IUPAC
Conference on Physical Organic Chemistry, Ithaca College, Aug 2-7,
1992, Abstract C-13.
(2) For reviews up to 1990 see: (a) Rappoport, Z.; Biali, S. E. Acc.
Chem. Res. 1988, 21, 442. (b) Hart, H.; Rappoport, Z.; Biali, S. E. In
The Chemistry of Enols; Rappoport, Z. Ed. 1990, Chapter 8, pp 481-
590.
The preferred solution conformation was determined
primarily by NMR methods under slow exchange, when
separate signals for each hydrogen and carbon in the
1
molecule were displayed in the H and 13C NMR spec-
tra.5,8 The chiral propeller conformation is supported by
the NMR spectra of 1d or of the isopropyl ether
Mes2CdCHOPr-i5 where the i-Pr group serves as a
(3) Rappoport, Z.; Biali, S. E. Acc. Chem. Res 1997, 30, 307.
(4) Kaftory, M.; Nugiel, D. A.; Biali, S. E.; Rappoport, Z. J . Am.
Chem. Soc. 1989, 111, 8181.
(5) Biali, S. E.; Nugiel, D. A.; Rappoport, Z. J . Am. Chem. Soc. 1989,
111, 846.
(6) Mislow, K. Acc. Chem. Res. 1976 , 9, 26.
(7) (a) Kurland, R. J .; Schuster, I. I.; Colter, A. K. J . Am. Chem.
Soc. 1965, 87, 2279. (b) Gust, D.; Mislow, K. J . Am. Chem. Soc. 1973,
95, 1535.
(9) (a) Nugiel, D. A.; Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc.
1984, 106, 3357. (b) Nugiel, D. A.; Rappoport, Z. J . Am. Chem. Soc.
1985, 107, 3669.
(10) Eventova, I.; Nadler, E. B.; Rochlin, E.; Frey, J .; Rappoport, Z.
J . Am. Chem. Soc. 1993, 115, 1290.
(8) Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc. 1984, 106, 477.
(11) Bu¨rgi, H.-B.; Dunitz, J . D. Acc. Chem. Res. 1983, 16, 153.
S0022-3263(97)01049-9 CCC: $14.00 © 1997 American Chemical Society

Stereochemistry and Solvation of 2,2-Ditipylethenols
J . Org. Chem., Vol. 62, No. 24, 1997 8373
Sch em e 1
were substituted by two m-Me or m-Br groups the
buttressing effect on the conformation and rotational
behavior was not large.10 Consequently, a larger change
in the bulk of the â-aryl group is required for observing
a significant change. Since no substrate carrying gemi-
nal 2,4,6-t-Bu3C6H2 (supermesityl)15 groups on carbon is
known, a 2,2-ditipyl (tipyl ) Tip ) 2,4,6-triisopropylphe-
nyl) derivative seems an attractive system.
The stabilities of 2,2-diarylethenols toward ketoniza-
tion and oxidation increased by replacing the mesityl by
tipyl groups. This is ascribed to an “increased [steric]
hindrance” by the bulkier tipyl relative to mesityl.16
The tipyl (or 2,6-i-Pr2C6H3) group acts in stereodynamic
studies both as a bulky substituent and as a chirality
probe. In the chiral conformation where the plane
bisecting the Me-CH-Me angle is not the symmetry
prochiral “chiral probe”, and at slow exchange separate
signals are displayed for the diastereotopic i-Pr methyls.
The preferred CdCOH conformations of enols 1 in
solution were deduced from NMR and IR data.12 Eight
types of conformations exist that differ in the value of
the dihedral angle θ of the CdCOH moiety (Scheme 1):
two planar, two periplanar (deviating from planarity up
to 30°), and four “gauche”-type forms that belong to two
pairs of cisoid and transoid nonplanar “clinal” conformers.
1H NMR spectra of 1a -g in various solvents show a
strong solvent dependence of δ(OH)12 and also of 3J HCOH
for 1a .12 The preferred CdCHOH conformation of 1a was
determined by using the Frazer equation,13 which relates
molecular plane, the two i-Pr-methyls are diastereotopic
1
and display anisochronous H or 13C NMR signals.17
A
higher enantiomerization barrier (∆Gcq) for 2-tipylacenaph-
thenone compared with the mesityl analogue indicates
that the former is more sterically congested in the ground
state.18
In ketones TipCOR, the Tip-CdO rotational barriers
increase with the increased bulk of the R, from 12 kcal
mol-1 for R ) Me to 24 kcal mol-1 for R ) t-Bu. The
barriers were higher than for the corresponding MesCOR
analogues,19 thus reflecting the larger steric interaction
between the alkyl R and the o-i-Pr (vs o-Me) in the
3
the J HCOH values to the H-C-O-H dihedral angle. In
poor hydrogen-bond-accepting solvents, e.g., CCl4, the
CdC-O-H moiety is almost syn planar. The IR spectra
of 1a in CCl4 indicate the presence of two different
species. It was suggested that in the major one the OH
group is intramolecularly hydrogen bonded to the π
system of the mesityl ring cis to it.12a In good hydrogen-
bond-accepting solvents (e.g., DMSO), the CdCOH con-
formation is anti-clinal in which the OH group is hydro-
gen bonded to one solvent molecule. From the linear
correlation between δ(OH)[1a ] and those of other enols
in the same solvents, similar conformations were as-
sumed for other enols 1.
q
transition state. In ketones TipCOC6H4X-p ∆Gc cor-
+
relates linearly with σX values due to the effect of the
conformational change during the Ar-CdO rotation on
the ARCO/X resonance interaction.20
Akkerman and Coops studied the racemization of
diarylacetic acids (3-Me-2,4,6-R3C6H)2CHCOOH.21 Opti-
cally active acids could not be obtained when R ) Me,
Et, but for R ) i-Pr, measurement of the racemization
rate was possible, showing again a higher rotational
barrier for the Tip than for the Mes derivative.
2,2-Ditipylethenols 2a -c were therefore prepared in
order to investigate how the â-Mes f â-Tip change affects
(a) the ground-state conformation, (b) the threshold
rotational barrier, and (c) the values of the barriers and
their order as a function of the R-substituent.
3
Both δ(OH) and J HCOH strongly depend on the DMSO-
d6 molar fraction in CCl4-DMSO-d6 mixtures.12 Analysis
of the equilibrium between an internally hydrogen-
bonded species and a conformer intermolecularly hydro-
gen-bonded to the solvent gave enol‚DMSO association
constants.
Tip2CdC(OH)R
2a : R ) H
This hydrogen-bonding effect is supported by X-ray
diffraction data.14 The conformation of 2b‚EtOH is anti-
periplanar due to hydrogen bonding to the cocrystallized
EtOH molecule,4 whereas enols 1a ,4 MesC(Ph)dCHOH,14
and (Br2Mes)2CdCHOH10 crystallize as hydrogen-bonded
tetramers.
b: R ) Me
c: R ) t-Bu
The remarkable stability, the chemical properties, the
static and dynamic stereochemistry, and the OH confor-
mation of the crowded mesityl-substituted enols result
from the steric bulk and conjugation ability of the mesityl
groups. However, whereas the effect of R-R’s on these
phenomena was extensively investigated, the effect of
â-substituents was only slightly investigated. When the
hydrogens of the two mesityl groups of 1a , 1b, and 1e
Resu lts a n d Discu ssion
Syn th esis of th e En ols. Enol 2a was prepared by
Fuson and co-workers by the acid-catalyzed pinacol
(15) Weidenbuch, M.; Schaefers, K.; Pohl, S.; Saak, W. J . Organomet.
Chem. 1988, 346, 171.
(16) Fuson, R. C.; Chadwick, D. H.; Ward, M. L. J . Am. Chem. Soc.
1946, 68, 389.
(17) Mislow, K.; Raban, M. Top. Stereochem. 1967, 1, 1.
(18) Miller, A. R. J . Org. Chem. 1976, 41, 3599.
(19) Casarini, D.; Lunazzi, L.; Verbeck, R. Tetrahedron 1996, 52,
2471.
(20) Ito, T.; Umehara Y.; Nakamura, K.; Yamada, Y.; Matsuura, T.;
Imashiro, F. J . Org. Chem. 1981, 46, 4359.
(21) (a) Akkerman, O. S.: Coops, J . Recl. Trav. Chim. Pays-Bas
1967, 86, 755. (b) Akkerman, O. S. Recl. Trav. Chim. Pays-Bas 1970,
89, 673.
(12) (a) Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc. 1984, 106, 5641.
(b) Rappoport, Z.; Nugiel, D. A.; Biali, S. E. J . Org. Chem. 1988, 53,
4814. (c) Nadler, E. B.; Rappoport, Z. J . Am. Chem. Soc. 1989, 111,
213.
(13) Fraser, R. R.; Kaufman, M.; Morand, P.; Govil, G. Can. J . Chem.
1969, 47, 403.
(14) Nadler, E. B.; Ro¨ck, M.; Schmittel, M.; Rappoport, Z. J . Phys.
Org. Chem. 1993, 6, 233.

8374 J . Org. Chem., Vol. 62, No. 24, 1997
Frey and Rappoport
F igu r e 1. Tetramer hydrogen-bonded arrangement of 2a held by hydrogen bonds.
F igu r e 2. ORTEP drawing and numbering scheme (a) and stereoview (b) of one of the structures of 2a .
rearrangement of 1,2-ditipylethanediol.16 It is interesting
that a stable enol is apparently thermodynamically more
stable than the diol precursor. We prepared 2a by this
method. However, the recent preparation of ditipylketene
322 enabled a more general synthesis of 2a -2c. The R-H
enol 2a was prepared by LiAlH4 reduction of 3 and
reaction with MeLi or t-BuLi gave, respectively, 2b and
2c (eq 1), which were identified by conventional methods.
2. Table 1 summarizes selected structural data for the
four symmetry-independent structures of 2a and com-
parative data for 2,2-dimesitylethenol (1a ).23 The four
structures in the unit cell differ only slightly in their bond
lengths and angles (Table 1). The enol exists in a chiral
propeller conformation in which all the rings are twisted
in the same direction. The solid-state structures of 1a
and 2a are partially similar, despite the increased bulk
of the isopropyl groups compared with methyl. This is
mainly due to the fact that the four o-isopropyl groups
are arranged so that the methine C-H bonds and the
rings are nearly coplanar and the methine hydrogens
point toward Câ. However, the i-Pr-methyls of 2a radiate
from the ring plane and could sterically interact (in
contrast with 1a ) with a neighboring ring.
The average bond lengths of the four symmetry-
independent molecules of 2a are close to those of 1a
whereas the average bond angles somewhat differ: R1,
R2, and R6 are smaller but R3 is larger by e2.5° in 2a
than in 1a . In the aryl torsional angles of 2a φ1 is 2.3°
smaller than in 1a but φ2 is larger by 8.3°, whereas in
Solid -Sta te Str u ctu r e of 2,2-Ditip yleth en ol (2a ).
Enol 2a crystallized from AcOH in the P21/c space group
as tetramers of four crystallographically independent
molecules held by hydrogen bonds between the hydroxyl
groups (Figure 1). The ORTEP drawing and the stereo-
view of one of the structures of 2a are shown in Figure
(23) Biali, S. E.; Gozin, M.; Rappoport, Z. J . Phys. Org. Chem. 1989,
2, 271.
(22) Frey, J .; Rappoport, Z. J . Am. Chem. Soc. 1996, 118, 5169.

Stereochemistry and Solvation of 2,2-Ditipylethenols
J . Org. Chem., Vol. 62, No. 24, 1997 8375
Ta ble 1. Selected Bon d Len gth s a n d Bon d (r) a n d Tor sion (O) An gles for En ols 2a a n d 1a
2a a
structure
1
2
3
4
average
1a b
bond lengths (Å)
CdC
C-O
dC-Ar(â)
dC-Ar(â′)
1.32(1)
1.383(7)
1.515(9)
1.50(1)
1.32(1)
1.380(9)
1.53(1)
1.509(8)
1.32(1)
1.34(1)
1.357(8)
1.50(1)
1.49(1)
1.325
1.376
1.509
1.506
1.326
1.372
1.502
1.504
1.385(8)
1.492(9)
1.524(8)
bond angles (deg)
R1
R2
R3
R6
115.8(6)
116.9(6)
116.4(7)
117.1(5)
116.6
120.6
122.8
122.4
118.1
121.4(6)
122.8(7)
122.5(6)
119.6(6)
123.4(5)
121.4(6)
120.4(6)
123.2(7)
122.4(7)
121.1(6)
121.8(5)
123.4(5)
121.0
120.8
123.7
torsion angles (deg)
φ1
φ2
55.72
58.67
54.52
59.20
53.17
58.57
54.24
59.03
54.41
58.87
56.7
50.2
nonbonded distances (Å)
i-Pr-C(A)-Câ
2.964
3.071
2.956
3.065
4.670
3.989
4.172
2.952
5.602
3.401
5.576
4.827
4.572
3.525
3.698
6.074
2.984
3.075
2.981
3.044
4.685
4.027
4.246
2.913
5.411
3.648
5.522
4.396
4.560
3.543
3.693
6.065
2.981
3.065
2.965
3.044
4.677
4.026
4.242
2.890
5.584
3.499
5.637
4.673
4.513
3.747
3.533
6.043
2.993
3.080
2.983
3.073
4.690
4.020
4.257
2.920
5.560
3.561
5.628
4.626
4.590
3.734
3.568
6.087
2.981
3.232
2.971
3.056
4.680
4.016
4.229
2.919
5.539
3.526
5.591
4.631
4.559
3.637
3.623
6.067
3.029c
i-Pr-C(B)-Câ
3.002c
2.991c
2.971c
i-Pr-C(C)-Câ
i-Pr-C(D)-Câ
i-Pr-C(A)dCH
i-Pr-C(B)dCH
i-Pr-C(C)dCH
i-Pr-C(D)dCH
i-Pr-C(A)dC-OH
i-Pr-C(B)dC-OH
i-Pr-C(C)dC-OH
i-Pr-C(D)dC-OH
i-Pr-C(A)-i-Pr-C(C)
i-Pr-C(A)-i-Pr-C(D)
i-Pr-C(B)-i-Pr-C(C)
i-Pr-C(B)-i-Pr-C(D)
3.766c
3.640c
a
b
1, 2, 3 and 4 stand for the four symmetry-independent crystallographic conformations. Average value of four crystallographically
independent molecules, ref 4. c The labeling of the methyl groups (ref 4) parallels the labeling of the isopropyl methines in 2a .
1a φ1 > φ2; i.e., the ring cis to the OH is more twisted
relative to the double-bond plane than the ring trans to
the OH, in 2a φ2 > φ1. The average difference between
φ1 and φ2 is 6.5° in 1a and only about 4.5° in 2a in the
opposite direction.
carbons and Câ, (ii) between methine carbons and vinylic
H and enolic OH, and (iii) between the o-methines located
on the same side of the double bond. These values reflect
the steric crowding of the molecule and assist in the
assignment of the 1H NMR spectrum at low tempera-
tures. Interestingly, both the o-i-Pr methine‚‚‚Câ and â
o-i-Pr methine‚‚‚â′ o-i-Pr methine distances are almost
similar to those of the o-Me‚‚‚Câ and â o-Me‚‚‚â′ o-Me
found in 1a , thus corroborating the similarity of the
ground-state structures of 2a and 1a .
Both 2a and 1a 4 crystallize as hydrogen-bonded tet-
ramers of symmetry-independent conformations. (Z)-2-
Mesityl-2-phenylethenol14 and 2,2-bis(pentamethylphenyl)-
ethenol10 also form solid cyclic hydrogen-bonded tet-
ramers: in the former it is composed of two symmetry-
independent conformers, and in the latter all four mol-
ecules are identical. In all these cases the CdCOH
conformation is anti. The tetrameric arrangement is
unique for the R-H-â,â′-diarylethenols since it does not
appear in the R-R derivatives, probably for steric reasons.
Solu tion Con for m a tion of 2a . 1H a n d 13C NMR
Sp ectr a a t Slow Exch a n ge. 1H NMR. The 400 MHz
The influence of the bulk of the R-alkyl substituent R
in 1a -e on the Ar-CdC torsional angle is known, but
less is known on the influence of the relative bulk of the
â-rings on the torsional angles. These angles reflect a
compromise between their tendency to decrease in order
to increase the Ar-CdC conjugation energy and to
increase in order to reduce repulsive steric interactions
between substituents on neighboring rings. The differ-
ences in the torsional angles of 2a and 1a may be due to
the interaction of the o-i-Pr methyls with the geminal
tipyl ring. A comparison of the solid-state structures of
the 2,2-ditipylethenols 2a -c could evaluate the influence
of the bulk of the R-R and â-Ar ring on the ground-state
conformation of 2a -c, but we could not obtain single
crystals of 2b and 2c.
Three types of relevant nonbonded distances found in
2a are displayed in Table 1: (i) between the methine

8376 J . Org. Chem., Vol. 62, No. 24, 1997
Frey and Rappoport
The NOESY spectrum of 2a at 245 K in CDCl3 (Figure
5) enabled a full signal assignment of the two tipyl
groups. First, NOESY cross peaks were found between
heptet CD and all of the four aromatic protons I-IV,
reinforcing the assignment of the signals CD and efgh to
the p-isopropyl groups. Second, NOESY cross peaks were
found between methyl doublets a, d, and efgh and
aromatic proton I, between methyls b, c, and efgh and
proton II, between methyl k and proton III and between
methyl l and proton IV. The correlations between
methyls efgh, i, and j and protons III and IV could not
be resolved due to the very small differences in δ’s in both
sets of signals. Nevertheless, since the methyl doublet
pairs i/l and j/k are known from the COSY spectrum to
belong each to a given isopropyl, we assume that the
unresolved NOESY cross peak includes correlations for
methyls i and j (as well as for methyls efgh) and aromatic
protons IV and III, respectively. Consequently, we
conclude that A-bc, B-ad, E-il, and F-jk are the o-
isopropyls and that isopropyls A-bc and E-il together with
tipyl-H II and IV belong to one ring, whereas isopropyls
B-ad and F-jk and tipyl-H I and III are located on the
other ring. Since the p-isopropyls are not resolved, CD-
efgh is assigned to both rings.
The cis or trans location of each ring relative to the
OH was also obtained from the NOESY spectrum (Figure
5b). Clear cross peaks were observed between the
methine heptet E and the OH doublet and between the
methine heptet F and the vinylic doublet. Examination
of the nonbonded distances in the X-ray structure of 2a
indicates that the shortest distance between an isopropyl
methine carbon and the OH group is found for the “down”
o-isopropyl on the tipyl ring cis to the OH (i-Pr-C(B) in
Table 1) and therefore the signals E-il are assigned to
the â′-Tip ring. The shortest i-Pr-methine-vinylic-H
distance was found for the “up” o-isopropyl group on the
tipyl ring trans to the OH group (i-Pr-C(D) in Table 1),
and consequently, the signals F-jk were assigned to the
â-Tip ring. The full peak assignment is summarized in
Figure 6.
Additional evidence for the proposed structure fol-
lows: (i) Two NOESY cross peaks in 2a were found to
correlate methines A and F and B and E. The nonbonded
X-ray distances between i-Pr methines on different rings
are the shortest between the pair of isopropyls located
on the same size of the CdC plane (Table 1). Hence, the
pairs of signals A and F and B and E are assigned to
isopropyls located on the same side of the CdC plane.
(ii) Four of the o-i-Pr methyl doublets (a-d) are shifted
upfield and four (i-l) resonate downfield relative to the
p-i-Pr signal. The latter signal has a δ (1.22 ppm) close
to that of the methyls of 1,3,5-i-Pr3C6H3 itself (1.28 ppm),
which is not surprising since the p-substituents are the
furthest from the enol moiety and will sense less the
electronic and anisotropic influences of the double bond
and of the neighboring tipyl ring.
Similar upfield and downfield shifts were observed for
the o- relative to the p-i-Pr methines. Inspection of the
X-ray structure and of space-filling models of 2a confirm
that in the frozen propeller conformation an “up” o-i-Pr
on the â′ ring and a “down” o-i-Pr on the â ring reside in
the shielding region of the adjacent tipyl rings. They
were therefore assigned to A-bc and B-ad, respectively.
The two remaining o-i-Pr reside in the deshielding region
of the neighboring tipyl rings and consequently resonate
at a lower field relative to the p-isopropyls. Using the
same reasoning, within a given i-Pr group, the more
F igu r e 3. 1H NMR spectrum of 2a in CDCl3 at 245 K.
Labeling: lower case letters: isopropyl-Me’s; upper case
letters: isopropyl-CH; Roman numerals: Ar-H signals.
1H NMR spectrum of 2a in CDCl3 at 245 K (Figure 3 and
Table 2) displays distinct signals for almost all protons
and isopropyl groups: nine i-Pr-methyl doublets (8 × 3H
and 1 × 12H), five i-Pr-methine heptets (4 × 1H and 1
× 2H), one OH and one vinylic CH doublet (J HCOH ) 13.7
Hz), and four distinct aromatic doublets (J meta ) 1.8 Hz).
Because of the prochiral character of the isopropyl group,
the appearance of eight separate i-Pr-methyl doublets
strongly suggests that in solution at 245 K 2a adopts a
chiral conformation with restricted rotations around the
2
CAr-Csp bonds in which the o-isopropyls and the aro-
matic m-protons are diastereotopic. This conformation
is in accord with the X-ray solid-state structure of 2a .
We draw the three-dimensional propeller in the previ-
ously described two-dimensional projections.8 The mol-
ecule is viewed from the π(CdC) bond plane along an axis
that dissects this bond in a way that the OH is always
oriented at the top right. The edge pointing to the
observer is thickened and substituents on it that are
above and below the molecular CRCâOH plane are des-
ignated as “up” and “down”, respectively.
For conformational analysis and DNMR study of 2a
1
in solution, all the H NMR signals should be unequivo-
cally assigned. This requires us to distinguish the â
(trans to OH) from the â′ (cis to OH) tipyl rings, to
identify the isopropyl moieties and their location on the
tipyl rings, and to determine which aromatic proton and
isopropyl groups are vicinal. A combination of COSY and
1
NOESY experiments enabled a full H NMR character-
ization of 2a in solution at slow exchange.
The assignment of the isopropyl signals was obtained
from the COSY spectrum of 2a in CDCl3 at 245 K. Figure
4 shows the aliphatic region of the spectrum in which
the methyl doublets are pairwise correlated to a methine
heptet, thus identifying pairs of methyl signals corre-
sponding to the same isopropyl group. Hence, methyl
doublets a and d, b and c, i and l, and j and k (for labeling,
see Figure 3) are coupled to methine heptets B, A, E, and
F, respectively. The larger doublet efgh displays a cross
peak with the double-intensity heptet CD and was
intuitively assigned to two accidentally isochronous p-
isopropyls. This was corroborated by a NOESY spectrum
(vide infra). The J meta couplings of the aromatic protons
observed in the 1D spectrum correlate pairwise protons
I and III and II and IV (not shown). Since no 4J
crosspeaks were detected between the aromatic protons
I-IV and the i-Pr-methines A-F we could not determine,
on the basis of the COSY spectrum alone, which isopropyl
is located next to which aromatic proton.

Stereochemistry and Solvation of 2,2-Ditipylethenols
J . Org. Chem., Vol. 62, No. 24, 1997 8377
Ta ble 2. 1H NMR Ch em ica l Sh ifts a n d P ea k Assign m en t of 2a in Sever a l Solven tsa
solvent
(CD3)2COf
CD3CNf
CD3C6D5
c
d
c
e
f
f
d
assignment
b
CDCl3
CD2Cl2
C2D2Cl4
DMSO-d6
DMF-d7
THF-d8
â-o-i-Pr(Me)
(a)
0.21
1.04
2.72
1.25
1.34
3.70
1.22m
2.85
0.33
1.02
2.62
1.24
1.37
3.32
6.54
6.86
7.11
6.91
7.12
4.83
0.19
0.96h
2.75
1.20
1.27
3.69
1.14
2.84
0.29
0.96h
2.64
1.19
1.30
3.32
6.45
6.78
7.03
6.86
7.07
4.75
0.10
0.93h
2.58
1.14h
1.22
3.58
1.11
2.75
0.20
0.93h
2.49
1.14h
1.25
3.18
6.39
6.70
6.94
6.77
6.98
4.74
0.20
0.96
2.79i
1.23
1.27
3.65
1.16m
2.79i
0.27
0.95
2.65
1.19
1.32
3.30j
6.59
6.78k
7.05
6.78k
7.01
8.92
0.28
1.01h
2.86i
1.31h
1.31h
3.77
1.21m
2.86i
0.36
1.01h
2.75l
1.28
1.40
3.41
6.76
6.90g
7.16
6.90g
7.14
9.32
0.26
1.01g
2.88
1.29
1.30
3.77
1.19m
2.80
0.37
0.99g
2.74
1.26
1.37
3.40
6.61
6.82
7.07g
6.83
7.07g
7.76
0.28
1.01h
2.93
1.30h
1.30h
3.78
1.19m
2.85
0.38
1.01h
2.76
1.28
1.40
3.41
6.73
6.88k
7.13
6.88k
7.11
7.97
0.29
0.99h
2.83i
1.22
1.32g
3.68
1.20m
2.83i
0.37
0.99h
2.68
1.27
1.30g
3.30
6.42
6.87k
7.12
6.87k
7.10
6.59
0.52
1.20h
3.05
1.33
1.36
3.82
1.24g,m
2.79
0.60
1.20h
2.95
1.25
1.41
3.55
6.55
7.00
7.19
7.06
7.22
4.44
(d)
(B)
(j)
â-o-i-Pr(CH)
â-o-i-Pr(Me)
(k)
â-o-i-Pr(CH)
p-i-Pr(Me)
p-i-Pr(CH)
â′-o-i-Pr(Me)
(F)
(efgh)
(CD)
(b)
(c)
(A)
(i)
â′-o-i-Pr(CH)
â′-o-i-Pr(Me)
(l)
(E)
â′-o-i-Pr(CH)
dCHn
â-Tip-H
â-Tip-H
â′-Tip-H
â′-Tip-H
OHn
(I)
(III)
(II)
(IV)
a
b
d
e
All values in δ (ppm) relative to internal TMS standard. The labels refer to Figure 3. c At 240 K. At 250 K. At 298 K. f At 273 K.
g
h
i
j
k
l
dd. One 2 Me doublet. One 3H septet. Approximate value since the signal overlaps the water signal. One 2H singlet. Overlaps
the DMF septet. Integrates for 12H. For the coupling constants, see Table 4.
m
n
F igu r e 4. COSY spectrum of the aliphatic and vinylic region
of 2a in CDCl3 at 245 K.
shielded doublet is assigned to the Me pointing toward
the neighboring ring, as deduced from the X-ray structure
and space-filling models. Within the distinct isopropyl
groups, the ∆δ (ppm) values between the two diaste-
reotopic Me doublets are significantly larger for the i-Pr
possesing a shielded Me group (∆δ (ppm) A-bc ) 0.69;
B-ad ) 0.83) than for the deshielded ones (∆δ (ppm) E-il
) 0.13; F-jk ) 0.09).
The nonequivalence of all o-i-Pr methyls provides
F igu r e 5. (a) NOESY spectrum of 2a at 245 K in CDCl3. (b)
Expansion of the methine and vinyl CH and OH regions. (c)
Indication of the important NOE interactions.
bond in C6H5Pr-i is 0.25 kcal mol-1 (low resolution
MW),25b 2.0 kcal mol-1 (NMR),25c and 2.3 kcal mol-1
(calculations).25d Because of the low barrier, NMR stud-
ies have been unable to observe and identify specific i-Pr
conformations. The preferred orientation of the tipyl-
information related to the orientation of the different i-Pr
groups in the molecule. Spectroscopic and computational
studies on C6H5Pr-i have shown that out of the three
possible conformations of the i-Pr group relative to the
ring the planar conformation (Chart 1), in which the i-Pr-
methine hydrogen eclipses the ring plane, is the most
stable.24,25a The rotational barrier about the Csp -Csp
2
3

8378 J . Org. Chem., Vol. 62, No. 24, 1997
Frey and Rappoport
F igu r e 7. 400 MHz 1H NMR spectrum of 2a at room
temperature in CDCl3.
point toward the enol moiety, in an opposite direction to
the aromatic protons. The NOESY correlations found
between the Me doublets and the m-H’s (vide supra)
reinforce this suggestion.
F igu r e 6. Syn-periplanar propeller conformation and full
peak assignment of 2a . The labels correspond to those in
Figure 3 and in Table 2.
Ch a r t 1
Concerning the conformation of the i-Pr groups in
2
3
relation to the rotation around the Csp (Ar)-Csp (CH) bond
relative to the NMR time scale Oki emphasized that
“...NMR data do not always correspond to a single
structure but sometimes correspond to an average struc-
ture”.27 Since the average structure might be that of two
or more unequally populated conformations with a larger
contribution of the conformation found in the solid, we
conclude that the predominant solution conformation of
2a is identical with that found in the crystal.
Additional support to the proposed assignment comes
from the rt spectrum of 2a in CDCl3 (Figure 7), which
displays broad signals for heptets A and E, aromatic
protons II and IV, and all but two methyl doublets. This
broadening results from an enantiomerization process in
which only one ring undergoes site exchange of the
aromatic and methyl protons (vide infra). Hence, o-
isopropyls A-bc, E-il and m-H’s II and IV belong to one
ring and o-isopropyls B-ad, F-jk and m-H’s I and III
belong to the other one. The appearance of the 12H
p-isopropyl methyl doublet as two 6H doublets (C-ef and
D-gh) at rt results from a temperature-dependent chemi-
cal upfield shift of one of them.
isopropyl groups of 2a in solution was deduced from the
COSY and NOESY spectra, and it considers the non-
equivalence of the i-Pr methyls on the NMR time scale.
The COSY spectrum of 2a displayed no cross-peaks
between the methine and the aromatic protons, although
pairs of m-H were found to be mutually coupled (J meta
)
1.8 Hz). Since long-range H-Me couplings (4J HMe ∼ 0.75
Hz26a) were reported for 1g,8 the lack of observed coupling
between the methine and aromatic protons in 2a suggests
4
that the i-Pr groups assume a conformation in which J
is zero. 4J values in allylic systems strongly depend on
the torsional angle φ26b (cf. 4). They are large for φ ≈ 0
or 180° and small for φ ≈ 90 or 270°.
A full peak assignment of 2a at slow exchange was
achieved in CDCl3 where the best resolution was achieved.
The NMR spectra of 2a were recorded in nine solvents
(Table 2), and signals were assigned by analogy to the
spectrum in CDCl3. In poor hydrogen-bond-accepting
solvents (CDCl3, CD2Cl2, C2D2Cl4, C6D5CD3) the rt spec-
trum is broad, and a sharp and well-resolved spectrum
was obtained only at ca. 250 K. In contrast, in good
hydrogen-bond accepting solvents (DMSO-d6, DMF-d7)
the rt spectrum is sharp and in solvents with moderate
hydrogen-bond accepting ability (CD3CN, (CD3)2CO, THF-
d8) the spectrum is sharp at 273 K (vide infra).
Introduction of large alkyl residues on C-3 leads to a
preferred conformation with φ ≈ 270°, i.e., in which the
methine C-H bond and the tipyl ring plane are nearly
coplanar. This is consistent with the orientation of the
o-isopropyls in the solid (Figure 1). The two alternative
planar orientations of the i-Pr groups that should be
considered are with the methine C-H pointing toward
or away from the enolic moiety. Two NOESY cross peaks
correlating the p-CH multiplet and two pairs of Ar-H
suggest the presence of two coplanar conformations. No
such correlation was found between the o-CH protons and
their neighboring m-H’s, thus indicating that the former
Two points deserve attention. First, δ(OH) (and J HCOH
,
cf. Table 4) is strongly solvent dependent, as was
observed for di- and trimesitylethenols.12 Second, the
assignment of the Tip-H protons in the good hydrogen-
bond-accepting solvents differs somewhat from that in
the weak hydrogen-bond acceptors. The aromatic protons
were labeled (I-IV in Figure 3) following their δ(CDCl3)
from high to low field and were then assigned pairwise
to the â-Tip (I and III) and to the â′-Tip (II and IV). In
good hydrogen-bond-accepting solvents the assignment
(24) Eliel, E. L.; Willen, S. H. Stereochemistry of Organic Com-
pounds; Wiley: New York, 1994; p 626.
(25) (a) Seeman, J . I.; Secor, H. V.; Breen, P. J .; Grassian, V. H.;
Bernstein, E. R. J . Am. Chem. Soc. 1989, 111, 3140. (b) True, N. S.;
Farag, M. S.; Bohn, R. K.; MacGregor, M. A.; Radakrishnana, J . J .
Phys. Chem. 1983, 87, 7, 4622. (c) Schaefer, T.; Parr, W. J . E;
Danchura, W. J . Magn. Reson. 1977, 25, 167. (d) Kao, J . J . Am. Chem.
Soc. 1987, 109, 3817.
(26) Gu¨nther, H. NMR Spectroscopy; Wiley: Chichester, 1980; (a)
pp 384-385; (b) pp 117-8.
(27) Oki, M. Applications of Dynamic NMR Spectroscopy to Organic
Chemistry; VCH: Deerfield Beach, FL, 1985; p 199.

Stereochemistry and Solvation of 2,2-Ditipylethenols
J . Org. Chem., Vol. 62, No. 24, 1997 8379
Ta ble 3. 13C NMR Ch em ica l Sh ifts a n d Assign m en t of 2a
in Tw o Solven ts
solvent
solvent
a
b
a
b
assignment CDCl3 DMSO-d6 assignment CDCl3 DMSO-d6
i-Pr-Me
22.60
22.75
23.89
23.93c
24.05
24.07
24.13
24.54
24.64
26.29
27.21
29.79
29.89
30.16c
22.39
22.48
23.75
23.81
23.88c
23.91
24.00
24.11
24.53
25.70
26.83
29.06
29.18
29.75
30.22
p-i-Pr-CH
33.84
33.94
33.14
33.21
Câ
m-Tip-C
110.48
121.28
122.15
122.43
122.75
129.69
134.05
143.69
146.85
147.28
147.54
148.03
148.33
149.46
108.47
120.20
120.64
121.27
121.66
133.51
135.84
144.65
145.69
145.98
146.79
147.07
147.38
147.69
ipso-Tip-C
CR
o,p-Tip-C
o-i-Pr-CH
a
b
At 245 K. At 298 K. c Double intensity.
Ta ble 4. Coa lescen ce Da ta for Tip 2CdCHOH (2a ) a t 400
MHz
q
∆Gc
,
solvent
process
∆ν, Hz Tc, K kcal mol-1
DMSO-d6 â′-o-i-Pr-Me h â′-o-i-Pr-Me 15.2c
319.4
16.5
16.6
15.9a
16.5
16.4
16.6
18.5
18.3
15.1
14.8
14.8
15.1
14.7
14.8
18.4
18.4
15.0
14.8
14.8
14.9
â′-o-i-Pr-Me h â′-o-i-Pr-Me 303.6d 362e
â′-o-i-Pr-CH h â′-o-i-Pr-CH 263.0a 345.3a
â′-m-Tip-H h â′-m-Tip-H
â-o-i-Pr-Me h â-o-i-Pr-Me
â-o-i-Pr-Me h â-o-i-Pr-Me
â-o-i-CH h â-o-i-Pr-CH
â-m-Tip-H h â-m-Tip-H
89.9
341.5
360e
369.3f
146.2g 350e
341.9
108.3
403.6
381.5
F igu r e 8. 1H NMR spectra of the (a) i-Pr-methine and (b)
i-Pr-methyl regions of 2a in DMSO-d6 (a) and CD2Cl2 and C2D2-
Cl4 (b) at various temperatures.
C2D2Cl4
â′-o-i-Pr-Me h â′-o-i-Pr-Me 28.0b,c 300.6
â′-o-i-Pr-Me h â′-o-i-Pr-Me 308.1b,d 325e
â′-o-i-Pr-CH h â′-o-i-Pr-CH 280.9b 322.0
â′-m-Tip-H h â′-m-Tip-H
â-o-i-Pr-Me h â-o-i-Pr-Me
â-o-i-Pr-Me h â-o-i-Pr-Me
82.4b
312.7
A comparison of δ(CR) and δ(Câ) for 2a and 1a (δ(Câ)
) 112.77, δ(CR) ) 143.89 in CDCl3 at 75 MHz at 223 K)9b
shows that the change Mes f Tip is somewhat stronger
on Câ than on CR. This might reflect the higher steric
congestion at Câ since at CR both molecules are rather
similar.
360.1b,f 325e
136.0b,g 325e
â-o-i-Pr-CH h â-o-i-Pr-CH 392.7a 405.6
â-m-Tip-H h â-m-Tip-H
C6D5CD3 â′-o-i-Pr-Me h â′-o-i-Pr-Me 13.4c
98.0a
381.2
291.0
â′-o-i-Pr-Me h â′-o-i-Pr-Me 273.3d 322e
â′-o-i-Pr-CH h â′-o-i-Pr-CH 237.2
â′-m-Tip-H h â′-m-Tip-H
320.0
305.6
1
In contrast to the H NMR spectrum (Table 2) the 13C
64.0
spectrum (Table 3) is only slightly affected by the
solvent’s hydrogen-bond-accepting ability. Mostly af-
fected are CR and Câ. In DMSO-d6 Câ is shifted upfield
by 2 ppm and CR is shifted downfield by 1 ppm relative
to CDCl3. This is consistent with computations showing
an upfield shift of Câ and a downfield shift of CR on
association with a hydrogen-bond-accepting solvent.28 It
was shown computationally that hydrogen bonding po-
larizes the O-H bond in the sense Oδ-‚‚‚Hδ+. Hence, the
more negatively charged oxygen (in 2a ‚DMSO vs 2a ) will
delocalize the charge better to Câ with its consequent shift
to a higher field and of CR to a lower field.
a
b
See text. Measured in CD2Cl2 (see text). c i-Pr-Me’s “j” and
“k” in Table 2. i-Pr-Me’s “a” and “d” in Table 2. e (5 K, see text.
d
f i-Pr-Me’s “b” and “i” in Table 2. i-Pr-Me’s “c” and “l” in Table 2.
g
differs from the labeling used in CDCl3 as shown by
comparing the rt spectra of 2a in the various solvents.
Whereas in CDCl3 and other weak hydrogen-bond ac-
ceptors the second and fourth protons (II and IV in Figure
6) are broadened, in the good hydrogen-bond-accepting
solvents the second and third Ar-H are broad. (Al-
though in DMSO-d6, DMF-d7, CD3CN, and (CD3)2CO the
two most upfield Tip-H protons are accidentally isoch-
ronous, in THF-d8 they are resolved and the second
proton is broad; we assume analogy of the spectrum in
THF-d8 to that in the other solvents). Since the ring
undergoing site exchange at rt is â′, the assignment was
determined accordingly (vide infra).
Dyn a m ic Ster eoch em istr y of 2a . The rt 1H NMR
spectra of 2a in poor hydrogen-bond-accepting solvents
displayed several broad signals for all the i-Pr-Me, the
â′o-i-Pr-CH’s and the â′-Tip-H’s. On raising the tem-
perature, additional signals broadened and several coal-
escence processes were observed. Figure 8 shows the
13C NMR. The 100 MHz 13C NMR spectrum of 2a in
CDCl3 at 245 K displayed distinct signals for all but two
of its 32 carbons (Table 3). One Me- and one CH signal
showed double intensity due to accidental isochrony. At
rt most signals were broad. A similar spectrum in
DMSO-d6 at rt shows only one accidentally isochronous
Me signal. The peak assignment was by gated decoupled
13C NMR spectra. However, we could not distinguish the
o- from the p-isopropyl methyls or the various Tip-C’s. A
frozen conformation on the NMR time scale is again
indicated.
1
temperature dependence of the H NMR spectra of the
i-Pr-methyl (in DMSO-d6) and i-Pr-methine (in CD2Cl2
and C2D2Cl4) region of 2a , and Figure 9 displays experi-
1
mental and calculated H NMR spectra in the aromatic
region in CD2Cl2 and C2D2Cl4 at various temperatures
(experimental) and rate constants (simulated). The
coalescence data are given in Table 4.
(28) Frey, J .; Eventova, I.; Rappoport, Z.; Muller, T.; Takai, Y.;
Sawada, M. J . Chem. Soc., Perkin Trans. 2 1995, 621.

8380 J . Org. Chem., Vol. 62, No. 24, 1997
Frey and Rappoport
F igu r e 10. Idealized transition states for the flip mechanisms
of a 2,2-diarylethenol propeller. The rectangular projection
represents a ring that is pependicular to the CdC plane.
(present in the DMSO-d6) with a â′-o-i-Pr-CH multiplet
that hampers exact determination of ∆ν and therefore
of kc; (ii) in C2D2Cl4 the ∆ν value used was the value
measured in CD2Cl2 at 240 K, which could result in a
different kc value; (iii) the o-i-Pr-Me and o-i-Pr-CH probes
should be less accurate since additional, usually broad,
peaks appear between the coalescing signals making the
accurate determination of Tc difficult. Because of the
uncertainty of both the ∆ν of the i-Pr-Me signals involved
in the high energy process (see below) and their Tc value,
F igu r e 9. Experimental (a) and simulated (b) 1H NMR
spectra of the aromatic region of 2a in CD2Cl2 and C2D2Cl4 at
various temperatures.
q
we calculated ∆Gc for this process only from the i-Pr-
CH and Tip-H signals. Since the dynamic processes took
place at Tc > 400 K, due to instrument limitation we were
not always able to detect a sharp average post-exchange
peak that usually sharpens at ca. 20° > Tc.
The dynamic processes undergone by 2a at ca. 240 to
>400 K require the use of low-melting and high-boiling
solvents. In spite of its high mp, DMSO-d6 was found
suitable since a sharp and well-resolved slow exchange
spectrum was obtained at rt. In C6D5CD3 a sharp
spectrum was obtained at ca. 250 K, but its boiling point
precluded measurements in the fast exchange regime. In
C2D4Cl2, at 245 K, the high viscosity of the solution
resulted in a broad spectrum. We therefore measured
the low-temperature spectra in CD2Cl2 and assumed that
the similarity of the two solvents permits neglect of
solvent effects on the stereodynamic properties of 2a .
The exchange rates (kc) at the coalescence temperature
(Tc) were calculated by using the Gutowsky-Holm ap-
proximation29 (kc ) π∆ν/x2) for the isopropyl methines
and methyls and by using the equation kc ) (π/x2)(x∆ν2
+ J 6) for the mutually coupled m-tipyl protons.30 The
Three interesting points arise from Table 4. (i) Two
different processes are clearly observed, each involving
site exchange of o-i-Pr-CH and m-Tip-H signals at a
different tipyl ring: a low energy process for the â′-ring
and a high energy process for the â-ring. Total line shape
simulations of the m-Tip-H coalescence processes agree
with this observation (Figure 9) and yield k’s from which
∆Gcq values similar to those in Table 4 are obtained. (ii)
Whereas only the â′-o-i-Pr-CH and m-Tip-H signals are
involved in the low energy process, both â and â′-i-Pr-
Me’s coalesce in this process. (iii) The low energy barrier
in DMSO-d6 is 1-1.5 kcal mol-1 higher than those in
C2D2Cl4 and C6D5CD3. A similar solvent effect, albeit on
both the low and high energy barriers, was observed for
1a in CD2Cl2/CS2 vs (CD3)2CO (∆∆Gc ) 1.7 kcal mol-1).10
q
q
rotational barriers (∆Gc ) were obtained from the Eyring
equation assuming a transmission coefficient of unity.31
The high energy barrier for 2a is solvent-independent.
Four different flip mechanisms of correlated rotation
can be envisioned for 1,1-diarylvinyl propellers. Their
idealized transition states are schematically depicted in
Figure 10. Rotations of none, one, or two rings by 180°
while the nonrotating rings remain fixed, which do not
involve helicity reversal, should also be considered.
In analyzing the dynamic processes the different
magnetic sites of one enantiomer of 2a were labeled by
lower case letters for i-Pr-methyls, by upper case letters
for i-Pr-methines, and by Arabic numerals for the meta-
Ar-H (Figure 11). The orientation of the diastereotopic
i-Pr-methyls relative to the tipyl ring plane is represented
q
We calculated ∆Gc from the coalescence data of o-i-Pr-
Me pairs when possible, and of o-i-Pr-CH and aromatic
q
protons pairs. The ∆Gc values obtained from either
process were mostly similar. Discrepancies of 0.1-0.6
kcal mol-1 between the barriers calculated by the three
probes are ascribed to experimental errors caused by the
following: (i) accidental overlap of the water signal
(29) Gutowski, H. S.; Holm, C. H. J . Chem. Phys. 1964, 40, 2426.
(30) Kurland, R. J .; Rubin, M. B.; Wise, W. B. J . Chem. Phys. 1964,
40, 2426.
(31) Sandstro¨m, J . Dynamic NMR Spectroscopy; Academic Press:
London, 1982; Chapter 6, p 77.

Stereochemistry and Solvation of 2,2-Ditipylethenols
J . Org. Chem., Vol. 62, No. 24, 1997 8381
rotation. Hence, no lower energy process could be
responsible for the enantiomerization.
An unequivocal identification of the high energy pro-
cess undergone by 2a is more difficult. The coalescence
data for the â-Tip-H and â-i-Pr-CH probes are consistent
with either a [â,â′]-two-ring flip or a sequential [â′]- and
[â]-one-ring flip processes (Table 5). The “ideal spectrum”
of the i-Pr-Me region at fast [â′]-one-ring flip should
display four average peaks for the exchanging pairs:
(bhi,ibh), (chl,lcj), (adh,daj), and (khj,jkh). In a [â,â′]-two-ring flip
process magnetic site exchange between o-i-Pr-Me’s on
different â-ring isopropyls should yield 3 average o-i-Pr-
Me signals after coalescence [(bhi,ibh), (chl,lcj), (adh,daj), and
(khj,jkh)]. In contrast, in the sequential two one-ring flip
processes only two average post-coalescence peaks for
protons (bhichl,ibhlcj) and (adhkhj,dajjkh) will appear, since the
nonflip process of the â′-Tip ring should exchange mag-
netic sites on a certain i-Pr as well. From Figure 8a this
ambiguity cannot be unequivocally resolved since at ca.
440 K two broad o-i-Pr-Me peaks (and 2 sharp p-i-Pr-Me
doublets) appear and it is unclear whether one of them
will split into two after sharpening at a higher unattain-
able (on the NMR spectrometer) temperature. A similar
uncertainty concerning the higher energy process in 1a
was solved by MM calculations of the transition state
energies of the various possible routes:5 the calculated
[â,â′]-two-ring flip barrier gave the best fit with experi-
F igu r e 11. Relationship between the two enantiomers of 2a
and their enantiotopic sites.
by solid and dashed wedges for methyls pointing above
and below the ring plane, respectively. The enantiotopic
sites (by external comparison) are assigned the same
letters with an overbar, i.e., methyls a and aj are enan-
tiotopic. To facilitate the analysis, 2a is reoriented at
the right of Figure 11. The sites exchanged by the
different flip and 180° rotation processes are summarized
in Table 5.
An enantiomerization via a flip of a given ring will be
accompanied by coalescence of three sets of NMR signals,
i.e., o-i-Pr-CH, m-Tip-H, and o-i-Pr-Me signals. Helicity
reversal via a nonflipping route involves only the coal-
escence of i-Pr-Me signals, while the Tip-H and i-Pr-CH
remain unchanged. In the 180° rotations, all the mag-
netic sites of the rotating ring(s) are exchanged. The sites
exchanged in the Tip-H and i-Pr-CH coalescences are
identical with those described for the o-Me and m-Mes-H
signals in 1a , and the information derived from these
processes should suffice to determine the rotational
mechanisms of 2a . However, the coalescence data for the
i-Pr-Me signals adds valuable direct information that
completes the unequivocal identification of the enanti-
omerization mechanism. The two types of diastereotopic
i-Pr-Me site exchange involve two methyls on either one
isopropyl or on different isopropyls. Whereas the former
process defines the process unambiguously as an enan-
tiomerization, the latter does not necessarily imply
enantiomerization. Site exchange of two geminal o-Pr-
methyls will occur only when, along the rotational
process, the isopropyl is bisected by the molecular plane
of symmetry. In contrast, the two p-i-Pr-methyls will
coalesce when the tipyl ring rotation involves a confor-
mation in which the ring is either coplanar with or
perpendicular to the CdC double bond plane, assuming
that rotation about the Ar-i-Pr bonds is faster than Ar-
CdC bond rotation. Site exchange of two o-i-Pr-Me’s
located on different isopropyls will occur when during the
rotation the ring passes through a plane perpendicular
to the CdC bond (e.g., in a ring flip mechanism).
q
ment. Furthermore, ∆Gc for the [â,â′]-two-ring flip
processes of 1a -e correlate linearly with Taft’s steric
parameter (Es), thus reinforcing the identification of the
second process as a two-ring flip (vide infra). In our case,
calculations were unsuccessful (see below) and we can
only assume that the similar behavior of 2a and 1a for
the low energy process will also hold for the high energy
process.
The tipyl groups of 2a served as intrinsic enantiomer-
ization probes for a relatively unambiguous study of the
stereodynamic processes. For 1a , which lacks such
probes, the isopropyl ether 5 was applied for further
delineating the rotational mechanism, and it was con-
cluded that the [â′]-one-ring flip is the threshold mech-
anism. A low energy nonhelicity reversal process was
q
ruled out by the identity of the rotational barriers (∆Gc
q
[o-Me h o-Me]) and the enantiomerization barrier (∆Gc
[i-Pr-Me h i-Pr-Me]) of 5.
Mes2CdCHOPr-i
5
The relatively large solvent effect on the [â′]-one-ring
flip barrier and its absence on the [â,â′]-two-ring flip
(Table 4) are noteworthy. In poor hydrogen-bond-accept-
ing solvents the OH of 1a hydrogen bonds with the ring
cis to it (Ar)12 and this interaction should stabilize the
syn periplanar CdCOH conformation. The fact that in
the threshold mechanism the flipping ring is the â′-Tip,
rather than the â-Tip, suggests that the OH‚‚‚π(Ar)
interaction modifies the ground-state conformation of 2a
as to facilitate the flip process. Since increasing the Ar-
Three conclusions regarding the low energy process
undergone by 2a arise from the coalescence data. (i)
According to the â′-Tip-H and â′-i-Pr-CH coalescence
processes (and their absence in the â-Tip counterpart)
and from the â- and â′-i-Pr-Me exchanging sites the
mechanism is a [â′]-one-ring flip. (ii) The flip of the â′-
Tip and the nonflip rotation of the â-Tip rings are
q
CdC torsional angle in 1a -e reduces ∆Gc , we propose
q
correlated since the ∆Gc values for all the probes,
that the syn-periplanar OH increases the â′-Tip-CdC
torsional angle of 2a compared with the anti-solvent
hydrogen-bonded conformation and consequently lowers
the activation barrier for enantiomerization. Due to the
lack of such OH‚‚‚π(â-Tip) interaction the solvent change
should not affect the high energy barrier. Similarly, the
rotation barrier for 2-hydroxybiphenyls differed by up to
including the â-i-Pr-Me are identical within experimental
error. This conclusion requires additional computational
support.5 (iii) The [â′]-one-ring flip is the threshold
enantiomerization mechanism of 2a since the ∆Gcq value
for the averaging of the diastereotopic geminal â-i-Pr-
Me’s is identical with those measured for the â′-Tip-ring

8382 J . Org. Chem., Vol. 62, No. 24, 1997
Ta ble 5. Sites Exch a n ged by F lip a n d 180° Rota tion Rou tes
Frey and Rappoport
site
ring flip
flipping
[none]
site
rotating
[none]
zero-ring
(AAh)(EEh)(DDh )(BBh)(FFh)(CCh)
(I hI)(III III)(II II)(IV IV)
(adh)(daj)(khj)(jkh)(ehf)(fej)
(AA)(EE)(DD)(BB)(FF)(CC)
(I I)(III III)(II II)(IV IV)
(aa)(dd)(kk)(jj)(ee)(ff)
(bcj)(cdh)(ihl)(lji)(hgj)(ghh)
(bb)(cc)(ii)(ll)(hh)(gg)
one-ring
one-ring
two-ring
[â]
(BFh)(FBh)(CCh)(AAh)(EEh)(DDh )
(I III)(III hI)(II II)(IV IV)
(akh)(kaj)(dhj)(jdh)(eej)(fhf)
[â]
(BF)(FB)(CC)(AA)(EE)(DD)
(I III)(III I)(II II)(IV IV)
(aj)(ja)(dk)(kd)(ef)(fe)
(bcj)(cbh)(ihl)(lji)(hgj)(ghh)
(bb)(cc)(ii)(ll)(hh)(gg)
[â′]
(AEh)(EAh)(DDh )(BBh)(FFh)(CCh)
(II IV)(IV II)(I hI)(III III)
(bji)(ibh)(chl)(lcj)(hhh)(ggj)
[â′]
(AE)(EA)(DD)(BB)(FF)(CC)
(II IV)(IV II)(I I)(III III)
(bc)(cb)(il)(li)(hg)(gh)
(adh)(daj)(khj)(jkh)(ehf)(fej)
(aa)(dd)(kk)(jj)(ee)(ff)
[â,â′]
(AEh)(EAh)(DDh )(FBh)(BFh)(CCh)
(I III)(III hI)(II IV)(IV II)
(akh)(kaj)(dhj)(jdh)(eej)(fhf)
[â,â′]
(AE)(EA)(DD)(BF)(FB)(CC)
(I III)(III I)(II IV)(IV II)
(aj)(ja)(dk)(kd)(ef)(fe)
(bji)(ibh)(chl)(lcj)(hhh)(ggh)
(bc)(cb)(il)(li)(hg)(gh)
a
E.g., (adh) designates a site exchange between the magnetic sites “a” and “dh” in Figure 11.
3
Ta ble 6. â(OH), δ(r-CH), a n d J HCOH va lu es for
Tip 2CdCHOH (2a ) a n d Mes2CdCHOH (1a ) in Sever a l
Solven ts a t Room Tem p er a tu r e
1.7 kcal mol-1 in different solvents according to their
hydrogen-bond-accepting ability.32
To corroborate the conclusions regarding the enanti-
omerization of 2a , we attempted to calculate some of the
possible transition-state structures for the processes by
molecular mechanics (MM2 as implemented in Macro-
model). Unfortunately, presumably due to the high steric
demands, the calculations did not converge to a chemi-
cally significant energy minimum, the tipyl rings were
highly distorted, and introduction of planarity constraints
of the rings precluded convergence.
2a a,b
1a a,c
3
3
no.
solvent
â
δ(OH) δ(CH) J HCOH δ(OH) δ(CH) J HCOH
1
2
3
4
5
6
7
8
9
DMSO-d6
DMF-d7
THF-d8
0.76 8.92
0.69 9.08
0.55 7.54
6.59
6.75
6.59
6.72
6.23
6.54
6.52
6.50
6.35
5.6
6.2
8.1
6.8
7.0
12.8
13.5
12.8
13.8
9.01
9.17
7.51
7.65
6.22
4.30
4.66
4.72
4.47
6.5
6.6
6.4
6.5
6.5
6.2
6.4
6.4
6.3
5.9
4.6
9.2
8.2
8.6
12.3
13.5
12.9
14.1
CD3COCD3 0.48 7.68
CD3CN
C6D5CD3
CDCl3
CD2Cl2
CCl4
0.31 6.59
0.10 4.32
0.00 4.72
0.00 4.77
0.00 4.42
The 6 and 5 kcal mol-1 higher rotational barriers for
2a compared with 1a for the one- and the two-ring flip,
respectively, in halogenated hydrocarbon solvents reflect
an increased steric hindrance. These large differences
are remarkable in view of the generally similar ground-
state structures of both enols. Akkerman21 had demon-
strated that ∆Gq for the racemization of bis(3-methyltipyl)-
acetic acid consisted of a large entropy and a relatively
low enthalpy contribution. The total line-shape analysis
used to simulate the spectra of the aromatic protons of
2a in CD2Cl2 and C2D2Cl4 at various temperatures
(Figure 9) yielded k’s from which for the one-ring flip ∆Hq
) 13.5 kcal mol-1, ∆Sq ) -5.4 eu (R ) 0.995, SD ) 0.13)
and for the two-ring flip ∆Hq ) 14.2 kcal mol-1, ∆Sq )
-11.0 eu (R ) 0.987, SD ) 0.14). Unfortunately, a
similar simulation for 1a did not produce an acceptable
fit between experimental and calculated spectra, prima-
rily due to extensive temperature-dependent chemical
shifts and line broadening. Since 1a reaches the slow
exchange regime only at <150 K the spectra measured
at 150 K are broad due to the solvent’s viscosity.
a
b
All chemical shifts are relative to internal TMS standard.
At 298 K. c At 296 K, from ref 12a,b.
methyl substituents, may induce less entropy changes
between the ground and transition states than tipyl
whose isopropyls lack this symmetry.
Con for m a tion of th e CdCOH Moiety. δ(OH) a n d
3J HCOH in P u r e Solven ts. The rt H NMR spectrum of
1
2a was measured in nine solvents differing in their
polarity and hydrogen-bond-accepting ability. Table 6
3
summarizes the δ(OH), δ(R-CH), and J HCOH values for
2a and those previously measured for 1a 12a and the
solvents’ Kamlet-Taft solvatochromic parameter â.34
3
Both δ(OH) and J HCOH are strongly solvent and temper-
ature dependent. A plot of δ(OH) [2a ] against δ(OH) [1a ]
(Figure 12) gave a linear relationship with a slope of 1.02
(R ) 0.998, SD ) 0.13), indicating that both enols interact
similarly with the solvent. A plot of the Kamlet-Taft â
vs δ(OH) for 2a was also linear (R ) 0.981, SD ) 0.35)
(Figure S1, Supporting Information).
We conclude that, as for 1a , (i) the large δ(OH) shifts
on changing the solvent result from OH association with
the hydrogen-bond-accepting solvents, that (ii) 2a adopts
different CdCOH arrangements in different solvents
according to their hydrogen-bond-accepting abilities, and
that (iii) the different conformers interconvert rapidly on
the NMR time scale (eq 2). In good hydrogen-bond-
The ∆Hq values for 2a are very similar for both
processes, and the ∆Gq differences result mainly from
different ∆Sq terms. The more negative ∆Sq for the two-
ring flip compared with the one-ring flip is consistent
with Akkerman’s observation21,33 of the high ∆Sq for the
racemization of the bis(3-methyltipyl)acetic acid by a two-
ring flip process. The close ∆Hq values calculated for the
two transition states may be fortuitous. The ∆Hq (2a )
value for the two-ring flip process resembles the ∆Gq (1a )
in a CS2/CD2Cl2 mixture (14.2 vs 13.5 kcal mol-1). This
could suggest that ∆Gq for rotation in 1a is mainly
enthalpic, although we have no evidence for the low ∆Sq
contribution to ∆Gcq in 1a . The mesityl, with its conical
(32) See references in ref 27, p 144.
(33) Finocchiaro, P. Gazz. Chim. Ital. 1975, 105, 149.
accepting solvents an anti-clinal conformer hydrogen

Stereochemistry and Solvation of 2,2-Ditipylethenols
J . Org. Chem., Vol. 62, No. 24, 1997 8383
3
Ta ble 7. δ(OH) a n d J HCOH Va lu es a n d Der ived
Equ ilibr iu m Con sta n ts for 2a in Bin a r y CCl4-DMSO-d 6
Mixtu r es a t 298 Ka
3
DMSO-d6,b [DMSO-d6] J HCOH
,
δ(OH),
ppm
%
M
Hz
Fanti-DMSO
Kd
Kassoc
c
0
0
13.8
4.42
5.18
6.66
7.28
7.78
8.22
8.53
8.75
8.83
8.92
0
0
0
0
0e
0f
0.2
0.028
0.084
0.28
0.56
1.4
12.3
9.1
7.7
6.8
6.1
5.7
5.6
5.6
5.6
0.18
0.17
0.57
0.50
0.74
0.64
0.85
0.75
0.94
0.84
1.5
0.91
1.0
0.96
1.0
0.98
1.0
0.22 10.0e
0.20
8.9f
0.6
2
1.34 20.7e
0.99 14.7f
2.90 11.4e
1.74
6.7f
4
5.83 11.0e
2.95
15.4
5.43
81.0
10.5
∞
26.8
∞
49.0
∞
5.5f
11.2e
4.0f
10
30
60
4.2
19.4e
2.5f
F igu r e 12. Plot of δ(OH) values of 1a vs that of Tip2CdCHOH
2a in nine solvents at room temperature. The numbering is
according to Table 6.
8.4
3.2f
4.4f
80
11.2
14.0
100
bonded to a solvent, 6, predominates. In poor hydrogen-
bonding solvents 2a adopts a syn-(peri)planar conforma-
tion 7 that is intramolecularly OH‚‚‚π hydrogen bonded
to the tipyl ring cis to it.
1.0
∞
a
b
[2a ] ) 0.0335 M. Volume % of DMSO in a total of 10 mL
solution. c Based on K(J ). Based on K(δ). e Calculated from eq 6a.
d
f Calculated from eq 6b in the Supporting Information.
The syn-planar CdCHOH conformation in non-hydro-
gen-bonding solvents was deduced for 1a from the 3J HCOH
their binary mixtures at 298 K. The data (Table 7) indi-
cate a similar enol-solvent interaction for both enols. For
2a the plots of 3J HCOH and δ(OH) vs [DMSO] (Figure S3,
Supporting Information) are very steep at low [DMSO]
and approach a plateau at [DMSO] ≈ 1.5 M. Addition
3
values.12a A plot of J HCOH(2a ) (DMF excluded) against
those of 1a (Figure S2, Supporting Information) gave
fairly linear correlation with a slope of 0.943 (R ) 0.959,
SD ) 0.35), strengthening the suggestion of similar
hydrogen-bonded conformations for both enols. However,
there are also differences: (i) The 3J HCOH (2a ) in both CCl4
and DMSO-d6 are lower than those of 1a by 0.3 Hz. A
larger difference (1.6 Hz) in the opposite direction was
3
of 0.6% (v/v) of DMSO in CCl4 changes J HCOH from 13.8
to 9.1 Hz and δ(OH) from 4.42 to 6.66 ppm, which is ca.
50% of the overall change between CCl4 and DMSO. As
for 1a 12a the changes in J HCOH and δ(OH) of 2a are
3
3
observed in DMF-d7. While small J HCOH differences
linearly correlated (Figure S4, Supporting Information).
From the δ(OH) and J HCOH values in the various
CCl4-DMSO-d6 mixtures we calculated Kassoc for 2a
according to eq 3 (6a ) 6 ) solvent ) DMSO).
might result from different electronegativities of the
substituents, it is more likely that they result from the
slightly different conformations of 2a and 1a that differ
in their bulk. (ii) The syn-planar conformation of 1a in
non-hydrogen-bond-accepting solvents is stabilized by
intramolecular OH‚‚‚π(cis-Ar) hydrogen bonding and
displays a νOH(CCl4) stretching at 3528 cm-1; another
stretching with 2% intensity at 3628 cm-1 was ascribed
to a free OH, presumably in an anti-type conformation.
3
Kassoc
7 + DMSO y z 6a
(3)
The methods used previously12 for calculating Kassoc
values for 1a assume that the associated/nonassociated
enol ratio can be calculated from the δ(OH) and J HCOH
values, if the solvent effect can be neglected and the two
conformers interconvert rapidly on the NMR time scale.
The observed δ(OH) and J HCOH values in the various
3
In contrast, 2a displayed only one νOH(CCl4) at 3522 cm-1
,
indicating a lower concentration (<1%) of the free OH
conformer (if any) in CCl4 than in 1a . (iii) 1a associates
in solution via hydrogen bonding to di- or tetrameric
supramolecular assemblies at low temperature, but 2a
did not show a similar association, presumably due to a
stronger OH‚‚‚π interaction.35
3
CCl4-DMSO-d6 mixtures will then be the weighted
averages of the δ and J values of the pure species. The
3
equations used for calculating Kassoc and the iterative
procedures are similar to those published previously.12
The procedure is described, together with Figures S5-
S7, in the Supporting Information. The Kassoc value for
2a ‚DMSO is 6.42 L M-1 and the δ(OH) for the anti-
conformation is 8.91. These values resemble Kassoc ) 5.25
and δ(OH) (anti-conformation) of 9.01 for 1a .12
Consequently, Kassoc values for 2a and 1a with DMSO
are very similar, and from the data in pure DMSO only
the anti-type conformer 6a , which is associated with one
DMSO molecule, is present. The 20% higher Kassoc for
2a compared to 1a is within the expected accuracy of the
determination of Kassoc considering the known mutually
δ(OH) in Bin a r y CCl4-DMSO-d 6 Mixtu r es. It is
possible that the geometry of the anti-clinal conformer
(observed for 1a )12a may differ in the different solvents
3
and affect the J HCOH and δ(OH) values used for deter-
mining the conformation.12c The spectra of 1a -e were
therefore measured in binary CCl4-DMSO-d6 mixtures,
and from the results it was concluded that only one
DMSO molecule is hydrogen bonded to one enol molecule
and enols/solvent association constants (Kassoc) were
calculated.12
For comparing Kassoc values for 1a and 2a , we measured
1
the H NMR spectra of 2a in CCl4, in DMSO-d6, and in
dependent errors resulting from the simultaneous deter-
36
mination of Kassoc and δanti-DMSO
.
(34) Kamlet, M. J .; Doherty, R. M.; Abraham, M. M.; Carr, P. W.;
Doherty, R. F.; Taft, R. W. J . Phys. Chem. 1987, 91, 1996. Kamlet, M.
J .; Taft, R. W. Acta Chem. Scand. 1986, B40, 619.
(35) Eventova, I.; Nadler, E. B.; Frey, J .; Rappoport, Z. J . Phys. Org.
Chem. 1994, 7, 28.
(36) (a) Benesi, H. A.; Hildebrand, J . H. J . Am. Chem. Soc. 1949,
71, 2703. (b) Mulliken, R. S.; Person, W. B. Molecular Complexes; Wiley:
New York, 1969; Chapter 7, pp 81-90.

8384 J . Org. Chem., Vol. 62, No. 24, 1997
Frey and Rappoport
Ta ble 8. 400 MHz 1H NMR Ch em ica l Sh ifts a n d
Assign m en t of En ols 2b a n d 2ca
Ta ble 10. Coa lescen ce Da ta for Tip 2CdC(OH)R a t 400
MHz
q
∆Gc
2b
2c
Tc,
K
kcal
c
d
d
assignment
(b)
(a)
(d)
(B)
(j)
CDCl3
DMSO-d6
DMSO-d6
R
solvent
process
∆ν, Hz
mol-1
2b Me DMSO-d6 â-o-i-Pr-Me h â-o-i-Pr-Me 480.0a 355e 16.0
116.0b 331e 15.8
â-o-i-Pr(Me)
0.03
0.93e
2.57f
1.22g
1.30
3.31
1.15i
2.79f
0.08
0.93d
2.57f
1.20g
1.21
3.11
1.93j
6.74
6.99
6.80
7.05
5.10
0.09
0.94
2.74
1.23
1.29
3.20h
1.14i
2.79f
0.16
0.91
2.65
1.20
1.32
3.10
1.80j
6.72
7.01
6.75
6.98
8.14
-0.07
0.97
2.65f
1.29d
1.29d
3.42f
1.12i
2.76f
-0.03
0.87
â-o-i-Pr-CH h â-o-i-Pr-CH 184.1 345e 16.2
â-o-i-Pr(CH)
â-o-i-Pr(Me)
â-m-Tip-H h â-m-Tip-H
117.0 338.0 16.1
â′-o-i-Pr-Me h â′-o-i-Pr-Me 416.1c 355e 16.1
164.0d 331e 15.5
(k)
â-o-i-Pr(CH)
p-i-Pr(Me)
p-i-Pr(CH)
â′-o-i-Pr(Me)
(F)
(efgh)
(CD)
(b)
(c)
(A)
(i)
â′-o-i-Pr-CH h â′-o-i-Pr-CH 180.0 345e 16.2
â′-m-Tip-H h â′-m-Tip-H
2c t-Bu DMSO-d6 â-m-Tip-H h â-m-Tip-H
â′-m-Tip-H h â′-m-Tip-H
90.3 338.0 16.3
124.0 345.2 16.4
100.0 345.2 16.6
â′-o-i-Pr(CH)
â′-o-i-Pr(Me)
2.65f
1.20
1.33
a
i-Pr-Me’s “a” and “k” in Table 8. b i-Pr-Me’s “d” and “ j” in Table
8. c i-Pr-Me’s “b” and “i” in Table 8. i-Pr-Me’s “c” and “l” in Table
d
(l)
(E)
8. e (5 K, see text.
â′-o-i-Pr(CH)
dCR
â-Tip-H
3.10
1.10k
6.68
reaching negative δ values in 2c (Table 8). This remark-
able upfield shift is attributed to the proximity of these
methyls to the shielding zone of the adjacent tipyl rings.
The increased bulk of the R-alkyl group apparently forces
the already shielded methyls into an even more shielded
environment. This may result from an increase in the
Tip-CdC torsional angle analogous to that found in
ethenols 1a -e, but evidence for this is lacking. Second,
as for 1a -e, the δ(OH) values in DMSO-d6 decrease with
the increased bulk of the R-alkyl group. Consequently,
in DMSO 2b and 2c adopt an anti-type conformation in
which the OH is hydrogen bonded to one DMSO molecule.
The 13C NMR spectra of 2b and 2c at rt displayed
distinct signals for almost all their carbons (excluding
accidental overlap). Assignment by analogy to 2a was
corroborated by gated decoupling. In both spectra only
some of the peaks were slightly broad due to a dynamic
process, and this was used to distinguish the o- from the
p-i-Pr-CH and Tip-C’s since the latter remain sharp
during the dynamic process. A comparison of δ(CR) and
δ(Câ) values of 2a -c shows that, as for 1a -e,12b increasing
the bulk of the R-alkyl group deshields CR and shields
Câ. This is ascribed to the contribution of the canonical
structure 8b in which Câ is negatively charged, while CR
is positively charged, and hence they are shifted to a
lower and a higher field, respectively. Elongation of the
(I)
(III)
(II)
(IV)
6.99
6.73
6.98
7.49
â′-Tip-H
OH
a
b
All values in δ (ppm) relative to internal TMS standard. The
labeling parallels that in Figure 6. c At 245 K. At 298 K. e One
d
2Me doublet. f One 2H septet. g dd. Approximate value since the
h
signal overlaps the water signal. Integrates for 12H. j Me signal.
i
k
C-Me3 signal.
Ta ble 9. 13C NMR Da ta for 2b a n d 2c in CDCl3 a t 298 Ka
assignment
2b
2c
assignment
2b
2c
i-Pr-Me
22.42 22.54
22.49 22.99
22.83 23.78
p-i-Pr-CH
33.87
33.97
33.78
33.88
Câ
107.15 105.80
121.28 121.65
122.33 122.31
122.59 122.44
123.51 123.39
132.17 134.46
134.97 136.59
148.23 147.37
148.42 148.48
148.80 149.48
150.03 150.47
147.07 146.81
148.16 147.90
151.04 159.43
19.39b
23.89 23.95b m-Tip-C
23.96 24.17c
23.99 24.87
24.60 25.45
24.69 26.11
24.84
25.01
26.03
26.29
29.48 28.06
30.03 28.90
30.27 29.89
30.53 29.93
ipso-Tip-C
o-Tip-C
o-i-Pr-CH
p-Tip-C
CR
R
29.53b
37.77c
-
+
Tip CdC(OH)R
TMS as internal standard. Me signal. c C-Me3 signal.
a
b
T Tip2C -C(dO H)R
2
8a
8b
1
Due to the stereodynamic process, the H NMR spec-
trum of 2a at rt in weak and in good hydrogen-bond-
accepting solvents is broad and sharp, respectively.
Addition of small amounts of DMSO-d6 to a sample of
2a in CCl4 causes sharpening, and at ca. 10% DMSO-d6
in CCl4 the spectrum is as sharp as in pure DMSO-d6.
This sharpening parallels the steep increase of δ(OH) and
CR-Câ bond reduces the alkyl and the â-Tip group
interaction, increasing the contribution of 8b for the
bulkier R-t-Bu group. Despite the increased bulk of the
tipyl group, the differences in ∆δ values for the R-H f
R-t-Bu change between dimesityl and ditipyl systems are
relatively small: 0.6 ppm for Câ and 3.44 for CR.
Dyn a m ic NMR Sp ectr a of 2b a n d 2c. When solu-
tions of 2b and 2c in DMSO-d6 were heated, all the o-i-
Pr-Me, o-i-Pr-CH, and Tip-H signals first broadened and
then coalesced. It was convenient to follow the dynamic
process in the aromatic region where a lower number of
signals interfere with the coalescence process (Table 10).
Coalescence of the i-Pr-Me and -CH signals of 2b was
also followed, but the calculated rotational barriers
should be treated with caution for four reasons. First,
the measured Tc values are less accurate due to difficul-
ties in distinguishing separate coalescence temperatures,
due to the vicinity of pairs of coalescing signals. Second,
since one i-Pr-CH signal overlaps the water signal
present in the solvent, ∆ν is very approximate. Third,
3
decrease of J HCOH observed upon addition of DMSO-d6
to CCl4, suggesting that the 2a /solvent association alters
both the CdCOH and the Tip2CdC conformations.
Sta tic a n d Dyn a m ic Ster eoch em istr y of 1-Me-
1
a n d 1-t-Bu -Su bstitu ted 2,2-Ditip yleth en ols. Th e H
a n d 13C NMR Sp ectr a of En ols 2b a n d 2c a t Slow
1
Exch a n ge. The H and 13C NMR spectra of 1-methyl
(2b) and 1-tert-butyl-2,2-ditipylethenol (2c) measured
under slow exchange conditions resemble that of 2a and
are consistent with a frozen propeller conformation on
the NMR time scale. Consequently, the signals were
assigned (Tables 8 and 9) by analogy to 2a .
The two unusually high-field i-Pr-methyl doublets
found in 2a are even more shielded in 2b and 2c,

Stereochemistry and Solvation of 2,2-Ditipylethenols
J . Org. Chem., Vol. 62, No. 24, 1997 8385
chemical shift may be responsible but over a 120 °C range
δ(p-i-Pr-Me) shifted by only e0.1 ppm. Moreover, anal-
ogy between 2b and 1b suggests that the two-ring flip is
the threshold enantiomerization mechanism. The coa-
lescence data for the p-i-Pr-Me signals should enable us
to distinguish these two possibilities (Table 5), but at 400
MHz these signals are not resolved. The coalescence data
of the o-i-Pr-Me exclude the occurrence of an NMR-
invisible, lower energy enantiomerization process such
as a zero-ring flip. The average peaks (adh ) 0.52, khj )
1.27, bjc ) 0.53, ihl ) 1.26 ppm) to be generated in this
mechanism were not observed during the dynamic pro-
cess, and hence, the proposed two-ring flip route is also
the threshold enantiomerization mechanism of 2b.
The threshold mechanism of 2c was deduced by anal-
ogy to 2b. From the coalescence of the Tip-H signals,
∆Gc ) 16.5 ( 0.1 kcal mol-1
. The similar rotation
q
F igu r e 13. 1H NMR spectrum of the i-Pr-Me region of 2b at
420 K in DMSO-d6. The impurity peaks (x) were truncated
for clarity and belong to an oxidation product.
barriers for the two-ring flip of 2b and 2c is in contrast
with the 2 kcal mol-1 decrease in ∆Gcq found for R-methyl-
(1b) and R-tert-butyl-2,2-dimesitylethenols (1e). The
linear correlation of the ∆Gcq values for the two-ring flip
process of 1a -e with the torsional angle φ2 suggests that
the main effect of changing the R-alkyl group is steric. It
is therefore puzzling why the decrease in ∆Gcq values for
the two-ring flip observed on replacing the R-H of 2a by
a methyl (2b) stops when the Me is replaced by a t-Bu
group (2c), especially since the steric hindrance of the
Tip2C moiety exceeds that of Me2C. It may be speculated
that at a certain level of steric hindrance, which was
for the i-Pr-Me signals, identification of pairs of coalesc-
ing signals is based on the assumption that the process
measured is a two-ring flip and not on an unequivocal
experimental observation. Fourth, enols 2b and 2c
undergo facile air oxidation upon heating to form poly-
alkylbenzofurans.37 The 1H NMR signals of the latter
overlap the enols’ signals and remain sharp over the
whole temperature range, thus hampering an accurate
determination of Tc. Oxidation of 2c is so extensive that
the coalescence was followed only in the aromatic region
where the benzofuran signals are the least disturbing.
q
reached in 2c, ∆Gc start to increase again.
Su m m a r y. 2,2-Ditipylethenols 2a -c exist in chiral
propeller conformations in solution, and 2a also in the
∆Gc (2b) ) 16.0 ( 0.5 kcal mol-1 and ∆Gc (2c) ) 16.5
( 0.1 kcal mol-1 (Table 10). These values are higher than
those of the dimesityl analogues 1b and 1e as expected
for the bulkier tipyl group. For both 2b and 2c the
barriers measured for the exchange of magnetic sites at
each ring are identical, suggesting a correlated motion
of the two rings. However, supporting MM calculations
are required to exclude occurrence of intermediate con-
formations in which the two rings rotate consecutively
with identical barriers.5
q
q
1
solid. The peak assignment in solution and the H and
13C NMR spectra are consistent with the proposed
conformation. 2a -c undergo an enantiomerization pro-
cesses on the NMR time scale involving a correlated
rotation of the rings. The threshold mechanism for 2a
is a one-ring flip. The higher energy process is a two-
ring flip mechanism. The rotation barrier for the one-
ring flip is solvent-dependent. The rotational barriers
q
(∆Gc ) of 2a -c are higher than those of 1a , 1b, and 1e
due to the increased bulk of the tipyl group. The
threshold mechanism for 2b and 2c is a two-ring flip with
If we consider only the i-Pr-CH and Tip-H signals,
Table 5 shows that the observed site exchange can be
attributed either to a [â,â′]-two-ring flip, two succesive
[â]- and [â′]-one ring flip or to a simultaneous or succesive
rotation of 180° by both rings. The observation of four
averaged post-coalescence o-i-Pr-Me peaks in the 1H
NMR spectrum of 2b at 420 K (Figure 13) excludes the
two successive [â]- and [â′]-one ring flip mechanism since,
as discussed for the high energy rotational mechanism
of 2a , it should yield only three new peaks. In the two
remaining possible mechanisms four post-coalescence
peaks will appear but the average peaks generated in a
two-ring flip (akh ) 0.69, dhj ) 1.09, bhi ) 0.68, chl ) 1.11
ppm) fit the experimental data (0.73, 0.77, 1.11, 1.14
ppm) better than those resulting from the nonhelicity
reversal [â] and [â′] 180° rotation (aj ) 0.67, dk ) 1.13,
bc ) 0.53, il ) 1.26 ppm). A temperature-dependent
∆Gcq values of 16.0 and 16.5 kcal mol-1, respectively. ∆Gc
q
of 2c is higher than expected from the analogy with 2e.
The conformation of the CdCOH moiety of 2a is
solvent dependent. In good hydrogen-bond-accepting
solvents (DMSO, DMF) it is anti-(peri)planar and the OH
is hydrogen bonded to a solvent molecule, and in poor
hydrogen-bond-accepting solvents it is syn-clinal having
an OH‚‚‚π stabilizing interaction with the ring cis to it.
The anti-type conformer in DMSO-CCl4 mixtures is
associated with one DMSO molecule with Kassoc of 6.4,
similar to the value for 1a .
In summary, the changes in structure and behavior
associated with the change â,â′-Mes2 f â,â′-Tip2 are in
the expected direction, but are not very significant, except
q
for the ∆Gc values.
Exp er im en ta l Section
(37) This is reminiscent of the oxidation of 2,2-dimesitylethenols to
the corresponding benzofurans, which were extensively investigated
by Schmittel and co-workers (for represetative papers see: Schmittel,
M.; Baumann, U. Angew. Chem., Int. Ed. Engl. 1990, 29, 541.
Schmittel, M.; Ro¨ck, M. Chem. Ber. 1992, 125, 1611. Schmittel, M.;
Heinze, J .; Trenkle, H. J . Org. Chem. 1995, 60 , 2726) and of the
oxidation of 2,2-ditipylethene-1,1-diol (Frey, J .; Rappoport, Z. J . Am.
Chem. Soc. 1996, 118, 5182). These oxidations will be reported
elsewhere.
Gen er a l Met h od s. Melting points were measured on a
Thomas-Hoover apparatus and are uncorrected. UV spectra
were measured on a Uvikon 930 spectrophotometer and FT
infrared spectra on a Nicolet Impact 400 spectrometer. NMR
spectra were recorded on Bruker WP 200 SY and Bruker AMX
400 pulsed FT spectrometers operating at 200.13 and 400.13
MHz and H and 50.32 and 100.62 MHz for 13C, respectively.
1

8386 J . Org. Chem., Vol. 62, No. 24, 1997
Frey and Rappoport
MS were recorded on a MAT-311 mass spectrometer, and
HRMS was conducted at the Mass Spectrometry Center at the
Technion, Haifa, on a Finnigan Mat 711 apparatus. The X-ray
diffraction data of the single crystals were measured with a
PW 1100/20 Philips Four-Circle computer-controlled diffrac-
tomer. Mo KR (λ ) 0.710 69 Å) radiation with a graphite
crystal monochromator in the incident beam was used. All
crystallographic computing used the TEXSAN structure analy-
sis software.
Solven ts a n d Ma ter ia ls. Tetrahydrofuran (THF) was
distilled from sodium benzophenone ketyl and ether from
LiAlH4. CCl4 was dried by standing over MgSO4 for 30 min
and was then filtered. All purchased reagents were the best
commercial samples and were used without purification.
Polydeuterated solvents for NMR spectroscopy (Aldrich, DMF-
d7 was from Ferak, Berlin, Germany) were dried over 4 Å
(DMSO-d6 and DMF-d7) or 3 Å (acetone-d6 and acetonitrile-
d3) molecular sieves. THF-d8 was kept in small portions in
separate sealed ampules.
2,2-Bis(2,4,6-tr iisop r op ylp h en yl)eth en ol (2a ). (a ) De-
h yd r a tion of 1,2-Bis(2,4,6-tr iisop r op ylp h en yl)eth a n ed iol.
1,2-Bis(2,4,6-triisopropylphenyl)ethanediol (50 mg, 1.1 mmol)
was dissolved in AcOH (3 mL), concentrated HCl (0.5 mL) was
added, and the mixture was refluxed for 3 h. Upon cooling,
white crystals started to precipitate. AcOH (3 mL) was added,
and the solution was heated until redissolution of the crystals.
The solution was cooled slowly to room temperature, and the
precipitated 2,2-bis(2,4,6-triisopropylphenyl)ethenol was col-
lected (32 mg, 67%): mp 115-117 °C (lit.16 mp 114-115 °C).
1H NMR and 13C NMR spectra are given in Tables 2 and 3.
FTIR νmax (Nujol): 3522 (s, OH) cm-1. MS (EI) m/z (relative
abundance, assignment): 448 (B, M). 405 (12, M - i-Pr), 387
(98, M - i-Pr - H2O), 372 (12, M - i-Pr - Me - H2O), 347
(17, M - 2 i-Pr - Me), 321 (14, M - i-Pr - 2 MeCHdCH2),
279 (11, M - i-Pr - 3 MeCHdCH2), 245 (15), 143 (15, C11H11),
131 (17, C10H11). HRMS: m/z 448.3729, requires 448.3705
calcd for C32H48O.
1H NMR and 13C NMR spectra are given in Tables 8 and 9.
IR νmax (Nujol): 3510 (s, OH), 1630 (m, CdC) cm-1. MS (EI)
m/z (relative abundance, assignment): 462 (B, M), 420 (14, M
- MeCHdCH2), 419 (13, M - i-Pr), 377 (37, M - i-Pr -
MeCHdCH2), 335 (20, M - i-Pr - 2 MeCHdCH2), 293 (10, M
- i-Pr - 3 MeCHdCH2), 251 (6, M - i-Pr - 4 MeCHdCH2),
209 (5, M - i-Pr - 5 MeCHdCH2), 131 (6, C10H11). HRMS:
m/z 462.3892, requires 462.3861 calcd for C33H50O.
Anal. Calcd for C32H48O: C, 85.65; H, 10.89. Found: C,
85.88; H, 10.75.
1,1-Bis(2,4,6-tr iisop r op ylp h en yl)-3,3-d im eth yl-1-bu ten -
2-ol (2c). To a solution of bis(2,4,6-triisopropylphenyl)ketene
(253 mg, 0.57 mmol) in dry ether (15 mL) was added t-BuLi
(0.6 mL of a 1 M solution in ether) at 0 °C in an argon
atmosphere. After the mixture was allowed to stand at rt for
1.5 h, 10% aqueous NH4Cl (30 mL) was added. Ether (30 mL)
was added, the organic phase was separated, washed with
water (30 mL), and dried (MgSO
4), and the solvent was
1
evaporated to dryness, leaving a crude orange-red oil. Its H
NMR spectrum is nearly identical with that of the pure enol
(2c), but four spots are apparent on silica TLC plates. The
oil was dissolved in MeOH (10 mL), and an orange-pink
precipitate (120 mg, 42%) was formed after 2 days at 4 °C.
1H NMR and 13C NMR spectra are given in Tables 8 and 9.
FTIR νmax (Nujol): 3566 (s, OH), 1584 (m, CdC) cm-1
.
Attempted isolation of an additional amount of 2c from the
mother liquor by silica gel chromatography (petroleum ether:
ether 4:1 as eluent) yielded 2-tert-butyl-3-(2,4,6-triisopropyl-
phenyl)-4,6-di(isopropyl)benzofuran (105 mg, 40%), mp 142 °C.
1H NMR (400 MHz, CDCl3) δ: 1.00 (6H, d, J ) 6.8 Hz, i-Pr-
Me), 1.05 (6H, d, J ) 6.8 Hz, i-Pr-Me), 1.15 (6H, d, J ) 6.8
Hz, i-Pr-Me), 1.16 (9H, s, t-Bu), 1.28 (6H, d, J ) 6.9 Hz, i-Pr-
Me), 1.30 (6H, d, J ) 6.9 Hz, i-Pr-Me), 2.59 (1H, m, J ) 6.8
Hz, i-Pr-CH), 2.86 (2H, m, J ) 6.8 Hz, i-Pr-CH), 2.92 (1H, m,
J ) 6.9 Hz, i-Pr-CH), 2.99 (1H, m, J ) 6.8 Hz, i-Pr-CH), 6.91
(1H, d, J ) 1.5 Hz, benzofuran-H), 6.98 (2H, s, Tip-H), 7.15
(1H, d, J ) 1.5 Hz, benzofuran-H). 13C NMR (100.64 MHz,
CDCl3) δ: 24.15, 24.22, 24.31, 24.81, 25.24 (5 i-Pr-Me), 26.22
(t-Bu-Me), 29.15 (t-Bu-CMe3), 30.24, 34.12, 34.18, 34.79 (4 i-Pr-
CH), 104.94 (CdC-t-Bu), 110.96, 118.59, 120.45, 125.41,
129.36, 143.62, 144.70, 147.78, 148.39, 153.86, 158.49 (10 Tip-C
+ CdC-t-Bu). FTIR νmax (Nujol): 1615, 1608 (m, CdC), 1417(s)
cm-1. MS (EI) m/z (relative abundance, assignment) 460 (B,
M), 445 (59, M - Me), 403 (8, M - t-Bu), 387 (11, M - H -
t-Bu - Me), 361 (6, M - t-Bu - MeCHdCH2), 345 (7, M - H
- 2t-Bu).
Crystal data: C32H48O. space group P21/c, a ) 23.703(4) Å,
b ) 31.709(6) Å, c ) 17.528(6) Å, â ) 110.50(3), V ) 12340(2)
Å3, Z ) 16, Fcalc ) 0.97 g cm-3, µ(CuKR) ) 3.59 cm-1, no. of
unique reflections ) 11201, no. of reflections with I g 3σI )
8134, R ) 0.099, Rω ) 0.130.
(b) Red u ction of Bis(2,4,6-tr iisop r op ylp h en yl)k eten e.
Bis(2,4,6-triisopropylphenyl)ketene22 (230 mg, 0.5 mmol) was
dissolved in dry THF (10 mL), and LiAlH4 (70 mg, 1.9 mmol)
was added. The resulting suspension was stirred at rt under
argon for 1 h and cooled to 0 °C, and water (5 drops) was added
to destroy unreacted LiAlH4. MgSO4 (50 mg) was added, the
suspension was filtered, the solvent was evaporated to dryness,
and the residual solid was recrystallized from hot AcOH,
affording white crystals of bis(2,4,6-triisopropylphenyl)ethenol
(137 mg, 59%), mp 115 °C. The MgSO4 was dissolved in 3%
aqueous HCl (10 mL) and the solution extracted with ether (4
× 25 mL). The combined organic phases were dried (MgSO4),
and the solvent was evaporated. Recrystallization of the
residual solid from AcOH yielded an additional 73 mg of
bis(2,4,6-triisopropylphenyl)ethenol (total yield 91%).
1,1-Bis(2,4,6-tr iisopr opylph en yl)-1-pr open -2-ol (2b). To
a solution of bis(2,4,6-triisopropylphenyl)ketene (158 mg, 0.35
mmol) in dry ether (40 mL) was added MeLi (0.3 mL of a 1.4
M solution in ether) at rt under argon. After 10 min, 10%
aqueous NH4Cl (50 mL) was added. The organic phase was
separated, washed with water (30 mL), and dried (MgSO4),
the solvent was evaporated, the residual light pink oil was
dissolved in high-boiling petroleum ether, and upon standing
at -18 °C very light pink crystals (90 mg, 56%), mp 141 °C,
were formed.
Anal. Calcd for C33H48O: C, 86.02; H, 10.50. Found: C,
86.18; H, 10.50.
Ack n ow led gm en t. We are indebted to the A. Gal
Memorial Fund for support and to Prof. Silvio Biali for
critical comments.
Su p p or tin g In for m a tion Ava ila ble: Solvent effects,
including the equations for calculating Kassoc, Figures S1-S7,
Table S1, and tables of bond lengths and angles and a listing
of positional and thermal parameters of 2a (40 pages). This
material is contained in libraries on microfiche, immediately
follows this article in the microfilm version of the journal, and
can be ordered from the ACS; see any current masthead page
for ordering information.
J O971049Z