Configurationally stable longitudinally twisted polycyclic aromatic compounds

Article abstract of DOI:10.1021/ja806958x

Two strategies for the synthesis of configurationally stable twisted polycyclic aromatic compounds (PACs) were pursued. The first approach employed dissymmetrically positioned 1-naphthyl substituents to bias the direction of twist in highly substituted PACs. 2,3-Bis(1-naphthyl)-1,4-diphenyltriphenylene (7) was prepared, and its meso cis-dinaphthyl and enantiomeric frans-dinaphthyl isomers were resolved by preparative supercritical fluid chromatography (SFC) on chiral supports. Similarly, several naphthyl-substituted derivatives of the more highly twisted 9,10,11,12,13,14-hexaphenylbenzo[b]triphenylene (2) were prepared. Of these, 10-(1-naphthyl)-9,11,12,14-tetraphenylbenzo[b]triphenylene (13) was resolved by SFC on a chiral support. The pure enantiomers of trans-7 showed moderately large specific rotations ([α]_D²⁵ = -330 and +320°), but the specific rotations for the enantiomers of 13 were unexpectedly small ([α]_D²⁵ = -23 and +23°). Computational studies suggest that the latter result is due to presence of a minor conformation of 13 possessing a larger rotation of opposite sign than the major conformation. Both 7 and 13 showed strong circular dichroism and moderately strong circularly polarized luminescence. A byproduct of these syntheses was 9,10,19,21-tetraphenyldiphenanthro[9,10-b:9,10-h]carbazole (15), a very crowded carbazole that exhibits an 81° end-to-end twist but is not resolvable. In the second approach, the large, twisted, polycyclic aromatic ligand 9,10,11,12,13,14-hexaphenylbenzo[h]naphtho[2,3-f]quinoline (21, an aza-2) was used to prepare the chiral, cyclometallated iridium(III) complex 4. The ligand 21 was prepared via an unusually stable benzannulated norbornadienone, for which the free energy of activation for decarbonylation was a remarkable 33.5 kcal/mol. The iridium complex 4 proved to be configurationally stable and resolvable by analytical HPLC on chiral supports, but the low solubility of 4 prevented its resolution on a preparative scale. A much more soluble dibutyl analogue of 4 (complex 28) was then prepared, but it was not resolvable on any of the available media.

Full text of DOI:10.1021/ja806958x

Published on Web 11/08/2008
Configurationally Stable Longitudinally Twisted Polycyclic
Aromatic Compounds
Robert S. Walters,† Christina M. Kraml,‡ Neal Byrne,‡ Douglas M. Ho,†,§ Qian Qin,†
Frederick J. Coughlin,† Stefan Bernhard,† and Robert A. Pascal, Jr.*,†
Department of Chemistry, Princeton UniVersity, Princeton, New Jersey 08544, and Lotus
Separations, Princeton, New Jersey 08544
Received September 9, 2008; E-mail: snake@chemvax.princeton.edu
Abstract: Two strategies for the synthesis of configurationally stable twisted polycyclic aromatic compounds
(PACs) were pursued. The first approach employed dissymmetrically positioned 1-naphthyl substituents to
bias the direction of twist in highly substituted PACs. 2,3-Bis(1-naphthyl)-1,4-diphenyltriphenylene (7) was
prepared, and its meso cis-dinaphthyl and enantiomeric trans-dinaphthyl isomers were resolved by
preparative supercritical fluid chromatography (SFC) on chiral supports. Similarly, several naphthyl-
substituted derivatives of the more highly twisted 9,10,11,12,13,14-hexaphenylbenzo[b]triphenylene (2)
were prepared. Of these, 10-(1-naphthyl)-9,11,12,14-tetraphenylbenzo[b]triphenylene (13) was resolved
by SFC on a chiral support. The pure enantiomers of trans-7 showed moderately large specific rotations
([R]2D5 ) -330 and +320°), but the specific rotations for the enantiomers of 13 were unexpectedly small
([R]2D5 ) -23 and +23°). Computational studies suggest that the latter result is due to presence of a minor
conformation of 13 possessing a larger rotation of opposite sign than the major conformation. Both 7 and
13 showed strong circular dichroism and moderately strong circularly polarized luminescence. A byproduct
of these syntheses was 9,10,19,21-tetraphenyldiphenanthro[9,10-b:9,10-h]carbazole (15), a very crowded
carbazole that exhibits an 81° end-to-end twist but is not resolvable. In the second approach, the large,
twisted, polycyclic aromatic ligand 9,10,11,12,13,14-hexaphenylbenzo[h]naphtho[2,3-f]quinoline (21, an aza-
2) was used to prepare the chiral, cyclometallated iridium(III) complex 4. The ligand 21 was prepared via
an unusually stable benzannulated norbornadienone, for which the free energy of activation for decarbo-
nylation was a remarkable 33.5 kcal/mol. The iridium complex 4 proved to be configurationally stable and
resolvable by analytical HPLC on chiral supports, but the low solubility of 4 prevented its resolution on a
preparative scale. A much more soluble dibutyl analogue of 4 (complex 28) was then prepared, but it was
not resolvable on any of the available media.
Introduction
All of the other acene derivatives that are known to possess
end-to-end twists of 60° or more5-9 undergo rapid racemization
Longitudinally twisted polycyclic aromatic compounds (PACs)
have been objects of our attention for over 20 years,1,2 but only
recently have any of these highly strained, strongly helical
molecules been resolved into pure enantiomers. Pentacene 1,
which has a 144° end-to-end twist, has been resolved by
chromatography on a chiral support, and its pure enantiomers
display exceptionally high specific rotations ([R]2D5 ) 7400°).3,4
However, 1 racemizes with a half-life of 9.3 h at 25 °C (∆Grqac
) 23.8 kcal/mol)3,4 and thus is less than ideal for extensive
studies of its chiroptical properties and certainly unsuitable for
any materials applications that depend upon enantiomeric purity.
at room temperature. These stand in contrast to the helicenes,
which have comparable specific rotations (hexahelicene, [R]D24
) 3640°; [13]helicene, [R]2D4 ) 8840°) but generally high
barriers to racemization (∆Gqrac ) 36-44 kcal/mol).10
How might longitudinally twisted PACs be made configu-
rationally stable?11 The most direct approach is to extend the
structure of 1 to give even more highly twisted molecules, for
which the transition state(s) for racemization would be extremely
(5) Pascal, R. A., Jr.; McMillan, W. D.; Van Engen, D.; Eason, R. G.
J. Am. Chem. Soc. 1987, 109, 4660–4665.
† Princeton University.
‡ Lotus Separations.
(6) Fagan, P. J.; Ward, M. D.; Caspar, J. V.; Calabrese, J. C.; Krusic,
P. J. J. Am. Chem. Soc. 1988, 110, 2981–2983.
§ Present address: Department of Chemistry and Chemical Biology,
Harvard University.
(7) Smyth, N.; Van Engen, D.; Pascal, R. A., Jr. J. Org. Chem. 1990, 55,
1937–1940.
(1) For a review, see: Pascal, R. A., Jr. Chem. ReV. 2006, 106, 4809–
4819.
(8) Qiao, X.; Padula, M. A.; Ho, D. M.; Vogelaar, N. J.; Schutt, C. E.;
Pascal, R. A., Jr. J. Am. Chem. Soc. 1996, 118, 741–745.
(9) Qiao, X.; Ho, D. M.; Pascal, R. A., Jr. Angew. Chem., Int. Ed. Engl.
1997, 36, 1531–1532.
(2) Pascal, R. A., Jr.; McMillan, W. D.; Van Engen, D. J. Am. Chem.
Soc. 1986, 108, 5652–5653.
(3) Lu, J.; Ho, D. M.; Vogelaar, N. J.; Kraml, C. M.; Pascal, R. A., Jr.
J. Am. Chem. Soc. 2004, 126, 11168–11169.
(4) Lu, J.; Ho, D. M.; Vogelaar, N. J.; Kraml, C. M.; Bernhard, S.; Byrne,
N.; Kim, L. R.; Pascal, R. A., Jr. J. Am. Chem. Soc. 2006, 128, 17043–
17050.
(10) For a review, see: Meurer, K. P.; Vogtle, F. Top. Curr. Chem. 1985,
127, 1–76.
(11) This and related questions are without a doubt the most frequent queries
after lectures on twisted acenes by R.A.P.
9
10.1021/ja806958x CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 16435–16441 16435

A R T I C L E S
Walters et al.
Scheme 1a
crowded and likely ensure a high barrier.12 However, the
synthesis of such molecules is most challenging. Perhaps a better
question is, “Can easily prepared twisted PACs be made
configurationally stable?” Compound 2 provides an excellent
framework from which to build such molecules: it possesses a
60° end-to-end twist7 and is easily prepared by the reaction of
a bulky aryne with a cyclopentadienone,7 and its precursors
contain easily modified structural elements; however, 2 does
racemize rapidly at room temperature. In this paper, we explore
two methods for conferring configurational stability to twisted
PACs related to 2.
The first strategy is to attach bulky groups to a twisted PAC
in a dissymmetric orientation. For example, the benzo[b]triph-
enylene core of compound 3, like 2, has a low barrier for twist
inversion, but the two “twistomers” are diastereomeric because
the naphthyl-group rotations are impossible at room temperature.
A significant difference in the energies of the diastereomers
would yield a preferred twist direction. The second approach
is to use twisted PACs as ligands for chiral metal complexes.
Thus, a configurationally stable iridium complex might have
cyclometalating ligands elaborated into highly twisted PACs,
as in 4. The twistomers of 4 are diastereomeric, and once again,
any energy difference leads to a preferred twist orientation. In
this paper, we report examples of both approaches.
a Conditions: (a) Ph2O, 300 °C, 26%; (b) isoamyl nitrite, 1,2-dichloro-
ethane, reflux. Yields in (b): 12, trace; 13, 0.6%; 14, 0.1%.
Scheme 2
that enforces the twist, and more complex substituents have been
avoided because of problems that arise when slow aryl rotation
results in conformational diastereomerism. However, if the
rotation of aryl substituents is frozen at room temperature, then
the resulting stereoisomerism can be used to advantage.
Consider the case of 2,3-bis(1-naphthyl)-1,4-diphenyltriph-
enylene (7) (Scheme 1). The triphenylene core of the corre-
sponding all-phenyl-substituted compound, 1,2,3,4-tetraphenyl-
triphenylene, is twisted by ∼30°,13 and it exists as a rapidly
interconverting set of two enantiomeric conformations. How-
ever, with two naphthyl groups in place of phenyls, compound
7 possesses six conformations, whose B3LYP/6-31G(d)-
optimized structures and relative energies are illustrated in
Scheme 2.
The rotation of 1-naphthyl groups in crowded polyarylben-
zenes is extremely slowsat least as slow as the rotation barriers
for o-tolyl groups in similar compounds, which range from 33
to 38 kcal/mol.14-16 For this reason, compound 7 under normal
conditions exists as a mixture of cis and trans isomers with
respect to the orientation of the two naphthyl groups, and the
Results and Discussion
Twist Bias by Dissymmetrically Disposed Napthyl Substitu-
ents. Approximately 10 acene derivatives have been prepared
with end-to-end twists in excess of 60°,2-9 but all of them
undergo twist inversion (enantiomerization) on time scales
ranging from microseconds to hours at room temperature. In
most of these compounds, phenyl groups provide the steric bulk
(13) Pascal, R. A., Jr.; Van Engen, D.; Kahr, B.; McMillan, W. D. J. Org.
Chem. 1988, 53, 1687–1689.
(14) Gust, D. J. Am. Chem. Soc. 1977, 99, 6980–6982.
(15) Gust, D.; Patton, A. J. Am. Chem. Soc. 1978, 100, 8175–8181.
(16) Qiao, X.; Pelczer, I.; Pascal, R. A., Jr. Chirality 1998, 10, 154–158.
(12) Pascal, R. A., Jr.; Qin, Q. Tetrahedron 2008, 64, 8630–8637.
9
16436 J. AM. CHEM. SOC. VOL. 130, NO. 48, 2008

Configurationally Stable Longitudinally Twisted PACs
A R T I C L E S
C2-symmetric trans isomer possesses enantiomeric M and P
forms. These three isomers cannot interconvert at room tem-
perature, but the inversion of the twist of their triphenylene cores
remains rapid. In the case of the cis isomer, twist inversion
rapidly interconverts enantiomeric conformations of equal
energy. At room temperature, the time-averaged structure is Cs-
symmetric; thus, the cis isomer is a meso compound. However,
twist inversion in the trans isomers interconverts diastereomeric
conformations of unequal energy; thus, the orientation of the
naphthyl groups determines the principal direction of the twist
of the triphenylene core. In the case of compound 7, the
Boltzmann distribution of the various conformations indicates
that the compound should exist at equilibrium at 298 K as a
17:3 mixture of trans and cis isomers, and the direction of the
twists in the trans isomers should be biased by roughly 4:1 in
favor of the more stable conformation (Scheme 2).
A similar analysis can be performed for 10,13-bis(1-naph-
thyl)-9,11,12,14-tetraphenylbenzo[b]triphenylene (12). B3LYP/
6-31G(d) calculations again find six conformations for com-
pound 12. In this case, the energy differences between the
conformations are greater than for 7, so at equilibrium at 298
K, compound 12 should be an 8:1 mixture of trans and cis
isomers, and the direction of the twists in the trans isomers
should be biased by more than 80:1 in favor of the more stable
conformation. The related compounds 13 and 14 contain only
one naphthyl group each, and thus, the twist bias is somewhat
smaller than for 12, but there is also no possibility of cis/trans
isomerism, which is a distinct advantage.
described in the open literature but only outlined in a review.19-21
Its synthesis from 1,3-bis(1-naphthyl)acetone22 is reported in
Experimental Procedures (see the Supporting Information).
Diazotization of the anthranilic acid 87 (to give a crowded aryne)
in the presence of 9 provided only traces of compound 12, as
judged by mass spectrometry, and all of the samples were
contaminated with the giant carbazole 15. The low yield of 12
was discouraging; furthermore, only the cis-dinaphthyl isomer
was likely to have been formed under the mild conditions of
this reaction, because both naphthyl groups of cyclopentadienone
9 must be oriented away from the approaching aryne for the
cycloaddition and the temperature is not high enough to
equilibrate the naphthyl orientations. The necessity to perform
an extra thermolysis and purification to obtain the trans-
dinaphthyl isomer from an already problematic reaction led us
to abandon 12 and turn to the synthesis of the mononaphthyl
compounds 13 and 14, for which there is less steric hindrance
to the cycloaddition and no cis/trans isomerism.
Compound 13 was prepared in a manner analogous to that
used for 12. The requisite cyclopentadienone 10 was formed
by double aldol condensation of (1-naphthyl)acetone23 with
benzil followed by dehydration. Diazotization of 8 in the
presence of 10 gave compound 13 in 0.6% yield; thus, even
one naphthyl group appears to strongly inhibit the aryne
addition, as judged by comparison with the synthesis of the all-
phenyl-substituted 2 (24% yield). However, the low yield is
also due in part to the difficulty of separating 13 from carbazole
15; the two compounds coelute in many chromatographic
systems. Thus, when compound 8 was diazotized in the absence
of a cyclopentadienone, carbazole 15 was isolated relatively
easily in 8% yield. All such diazotization reactions, with or
without a diene, produce numerous byproducts.24 With 8, the
most abundant of these is 1,4-diphenyltriphenylene, which is
formed in 10-15% yield.
Though inefficient, the synthesis described above provided
sufficient compound 13 for its resolution by means of SFC on
a Chiralpak AD-H column. The pure (+)- and (-)-enantiomers,
although configurationally stable, exhibited only modest specific
rotations ([R]2D5) of +23 and -23°, respectively. Given that
compound 1 and its derivatives, with end-to-end twists of
138-144°, display [R]D values ranging from 5200 to 7400° 4
and compound 7, with a B3LYP/6-31G(d)-calculated end-to-
end twist of 32°, shows an [R]D value of 330°, the small rotations
of 13 were quite a surprise for a compound with a calculated
end-to-end twist of 58°. However, calculations of the specific
rotations of these molecules at the B3LYP/6-31G(d) level (using
the dipole electric field polarizabilities) were able to shed some
light on this result. In the case of compound 7, the two M-trans
twistomers (top row of Scheme 2) were calculated to have
Synthesis and Resolution of Triphenylene 7. The crowded
triphenylene 7 was chosen as the first “twist bias” target because
its synthesis (Scheme 1) was expected to be trivial. The two
starting materials, phencyclone13,17 (5) and dinaphthylacetylene
6,18 were in hand, and warming them in diphenyl ether at 300
°C gave compound 7 as a mixture of cis and trans isomers in
26% yield. Although inseparable by ordinary chromatography,
the cis (meso) and trans (chiral) isomers were cleanly separated
by supercritical fluid chromatography (SFC) on a Chiralcel
OD-H column (see the Supporting Information). A 7:3 trans-cis
ratio was observed, which is lower than the calculated 17:3
Boltzmann distribution for 298 K. However, the synthesis was
carried out at 573 K, where the calculated distribution would
be 77:23, which is much closer to the observed ratio (and
certainly well within reasonable errors for the B3LYP/6-31G(d)
calculations), and the conformational distribution would have
been frozen upon rapid cooling. A high barrier to naphthyl-
group rotation was demonstrated by heating pure cis-7 in DMSO
at 180 °C for 24 h without detectable isomerization to the trans
isomer. This establishes a lower limit of ∼39 kcal/mol for the
∆Gqrot of 7. The purified trans-7 was resolved by SFC on a
Chiralpak AD-H column to give (+)- and (-)-enantiomers with
specific rotations ([R]2D5) of +320 and -330°, respectively.
Synthesis of Naphthyl-Substituted Benzo[b]triphenylenes. The
syntheses of compounds 12-14 are direct but inefficient. The
preparation of the cyclopentadienone precursor of 12, 2,5-bis-
(1-naphthyl)-3,4-diphenylcyclopentadienone (9), is not fully
(17) Dilthey, W.; ter Horst, I.; Schommer, W. J. Prakt. Chem. 1935, 143,
189–210.
(22) King, J. A.; McMillan, F. H. J. Am. Chem. Soc. 1951, 73, 4911–
4915.
(18) Nakasuji, K.; Akiyama, S.; Akashi, K.; Nakagawa, M. Bull. Chem.
Soc. Jpn. 1970, 43, 3567–3576.
(23) Jozwiak, K.; Khalid, C.; Tanga, M. J.; Berzetei-Gurske, I.; Jimenez,
L.; Kozocas, J. A.; Woo, A.; Zhu, W.; Xiao, R.-P.; Abernethy, D. R.;
Wainer, I. W. J. Med. Chem. 2007, 50, 2903–2915.
(19) Ogliaruso, M. A.; Romanelli, M. G.; Becker, E. I. Chem. ReV. 1965,
65, 261–367.
(24) (a) Qiao, X.; Ho, D. M.; Pascal, R. A., Jr. J. Org. Chem. 1996, 61,
6748–6750. (b) Lu, J.; Zhang, J.; Shen, X.; Ho, D. M.; Pascal, R. A.,
Jr. J. Am. Chem. Soc. 2002, 124, 8035–8041.
(20) Idelson, M. B.S. Thesis, Polytechnic Institute of Brooklyn, 1952.
(21) Qiao, X. Ph.D. Thesis, Princeton University, 1997.
9
J. AM. CHEM. SOC. VOL. 130, NO. 48, 2008 16437

A R T I C L E S
Walters et al.
other criteria. Strong circularly polarized luminescence (CPL)
is one of the properties of interest for these compounds, so a
strongly luminescent metal complex is to be preferred. In
addition, preliminary studies showed that cationic chiral metal
complexes, such as derivatives of ruthenium(II) tris(phenan-
throline), were much more difficult to resolve chromatographi-
cally than neutral metal complexes such as iridium(III) bis-
(benzo[h]quinolinato-N,C10)acetylacetonate (16).25
Our prior experience with these and related cyclometalated
Figure 1. Molecular structure of compound 15. Thermal ellipsoids are
drawn at the 50% probability level.
iridium complexes26-28 led to the choice of compound 16 as
the parent for a family of configurationally stable metal
complexes bearing twisted acene ligands. For use as a reference,
a sample of the highly luminescent 16 was prepared and resolved
by preparative HPLC on a Chiralpak AD column, giving (-)-
and (+)-enantiomers with specific rotations ([R]2D5) of -340 and
+360°, respectively. The relative ease of this endeavor encour-
aged us to prepare similar complexes with highly twisted
ligands.
Synthesis of Iridium Complex 4. The initial synthetic target
was compound 4, a derivative of compound 16 in which the ben-
zo[h]quinoline ligands have been replaced by 9,10,11,12,13,14-
hexaphenylbenzo[h]naphtho[2,3-f]quinoline (21) (Scheme 3). As
noted above, the hydrocarbon 9,10,11,12,13,14-hexaphenylben-
zo[b]triphenylene (2) (Scheme 1), which is isosteric with 21,
possesses a 60° end-to-end twist.7 The synthesis of 21, like those
of many twisted hydrocarbons, requires the addition of a crowded
aryne to a crowded cyclopentadienone. First, aldol condensation
of benzo[h]quinoline-5,6-dione29 (17) and 1,3-diphenylacetone gave
cyclopentadienone 18. This material was used without purification
for the cycloaddition of 18 and tetraphenylbenzyne, generated by
diazotization of tetraphenylanthranilic acid (19),8 to form the adduct
20 in 18% yield from compound 17.
This “norbornadienone” adduct 20 was quite unexpected
because the decarbonylation of such compounds almost always
occurs rapidly at the temperature of the tetraphenylbenzyne
addition (80 °C). However, X-ray analysis of a crystal of
compound 20 unambiguously confirmed the result, and the
molecular structure of 20 is illustrated in Figure 2. Compound
20 possesses evenly distributed, moderate strain, as judged by
the observed bond distances in the norbornadiene core, all of
which were roughly 0.04 Å longer than the canonical sp3-sp2
C-C bond distance of 1.51 Å: C(9)-C(8B), 1.545(4) Å;
C(9)-C(9A), 1.552(4) Å; C(9)-C(15), 1.548(4) Å; C(14)-
C(13A), 1.550(4) Å; C(14)-C(14A), 1.545(4) Å; and C(14)-
C(15), 1.550(4) Å.
specific rotations of +526 and -342° for the major and minor
conformations, respectively. The Boltzmann-weighted sum of
the two rotations is approximately +344°, in remarkable
agreement with the experimental rotation measurements. How-
ever, for compound 13, the major and minor twistomers were
calculated to have specific rotations of +184 and -238°,
respectively; thus, the absolute magnitudes of the rotations are
smaller than for 7 and the minor component has the stronger
rotation. In this case, the calculated energy difference is 1.1
kcal/mol (for an 84:16 ratio of twistomers), yielding a Boltz-
mann-weighted sum of the rotations of 116°. This is higher than
the experimental value, but the results provide a reasonable
explanation for the observed low rotation.
The final member of this series, compound 14, was isolated
in a dismal 0.1% yield from the reaction of 8 and cyclopenta-
dienone 11. This inexplicably low yield, combined with the
disappointing results for 13, discouraged any attempts to resolve
compound 14 on a preparative scale.
Molecular Structure of Carbazole 15. We have frequently
observed octaphenylcarbazole24 as a byproduct of the diazoti-
zation of tetraphenylanthranilic acid (19); thus, the isolation of
the giant carbazole 15 upon diazotization of 8 was not a surprise.
However, this polycycle is quite crowded, and molecular
mechanics calculations indicated that it should have a very large
end-to-end twist. Fortunately, compound 15 crystallized easily,
so its X-ray structure could be determined; its molecular
structure is shown in Figure 1.
Carbazole 15 crystallized in the space group C2/c, and since
the molecule lies on a special position, it possesses crystal-
lographic C2 symmetry. The twisted polycyclic core of 15
resembles a two-bladed propeller with the N-H bond as the
“axle”. The C(10)-C(11)-C(11A)-C(10A) torsion angle is
80.8°, which is larger than the end-to-end twists of all but three
other twisted acenes.3,4,9 However, all attempts to resolve 15
by chromatography on chiral supports, even at 0 °C, were
unsuccessful, and thus, its barrier to racemization is likely to
be less than 20 kcal/mol.
Twist Bias by Fusion to a Configurationally Stable Metal
Complex. A second way to confer configurational stability to a
twisted acene is to attach it to a chiral metal complex. As in
the cases of compounds 7 and 13, the barrier to twist inversion
remains low, but the resulting twistomers are diastereomeric;
thus, one twistomer must be preferred to a greater or lesser
degree. The specific choice of metal complex hinged on two
(25) Lamansky, S.; Djurovich, P.; Murphy, D.; Feras, A.-R.; Kwong, R.;
Tsyba, I.; Bortz, M.; Mui, B.; Bau, R.; Thompson, M. E. Inorg. Chem.
2001, 40, 1704–1711.
(26) Lowry, M. S.; Hudson, W. R.; Pascal, R. A., Jr.; Bernhard, S. J. Am.
Chem. Soc. 2004, 126, 14129–14135.
(27) Lowry, M. S.; Goldsmith, J. I.; Slinker, J. D.; Rohl, R.; Pascal, R. A.,
Jr.; Malliaras, G. G.; Bernhard, S. Chem. Mater. 2005, 17, 5712–
5719.
(28) Coughlin, F. J.; Oyler, K. D.; Pascal, R. A., Jr.; Bernhard, S. Inorg.
Chem. 2008, 47, 974–979.
(29) Kloc, K.; Mlochowski, J.; Szulc, Z. J. Prakt. Chem. 1977, 319, 959–
967.
9
16438 J. AM. CHEM. SOC. VOL. 130, NO. 48, 2008

Configurationally Stable Longitudinally Twisted PACs
A R T I C L E S
Scheme 3a
a Conditions: (a) 1,3-diphenylacetone, KOH, MeOH; (b) isoamyl nitrite, DCE (yields from 17: 20, 18%; 26, 21%); (c) PhNO2, 200 °C (yields: 21, 77%;
27, 37%); (d) IrCl3, 2-methoxyethanol, H2O followed by 2,4-pentanedione, Na2CO3, 2-ethoxyethanol (yields: 4, 8%; 28, 4%); (e) 6-undecanone, KOH,
EtOH followed by Ac2O, H2SO4; (f) maleimide, PhNO2, 210 °C (yield: 36% from 22); (g) NaOCl, NaOH, MeOH followed by KOH, n-PrOH, reflux (yield:
90%).
Only a handful of even moderately stable norbornadienones
are known, and the Cambridge Structural Database30 contains
only three crystallographically characterized examples of such
compounds, all of which are crowded molecules.31-33 Com-
pound 20 is perhaps the most robust norbornadienone thus far
observed; it is completely stable at room temperature and
decarbonylates with a half-life of 40 min at 180 °C to give the
desired twisted ligand 21. The kinetics of decarbonylation were
between 283 and 303 K (see the Supporting Information).
Although the pairs of exchanging resonances cannot be con-
clusively identified, these data give a limiting range of free
energies of activation for the dynamic processes involved
(almost certainly phenyl rotations and probably the twist
inversion as well) of 12-15 kcal/mol.
With ligand 21 in hand, the preparation of complex 4 was
relatively straightforward. Initial reaction of 2 equiv of 21 with
1 equiv of IrCl3 gave an intermediate µ-chloro-bridged iridium
dimer, and treatment of this material with 2,4-pentanedione and
base gave complex 4 in 8% overall yield after purification by
preparative TLC. Although no X-ray structure of 4 was obtained,
its constitution was confirmed by the expected NMR resonances
for both the acetylacetonate and twisted ligands (including
broadening of the latter), the molecular ion in the MALDI mass
spectrum, and the characteristic UV spectrum of this brilliant-
orange compound. Compound 4 is soluble in chlorinated organic
solvents but quite insoluble in the alcohols and hydrocarbons
that are permissible solvents for the available chiral chromato-
graphic media. Thus, it proved possible to resolve complex 4
analytically on a Chiralcel OD column (see the Supporting
Information), but its preparatiVe resolution was out of the
question because of the low solubility.
1
determined by monitoring the H NMR spectrum of 20 in
DMSO-d6 at various temperatures; the relative concentrations
of 20 and 21 were followed by integration of their characteristic
aromatic resonances. The reaction was first-order with rate
constants k433K ) 3.47 × 10-5 s-1, k443K ) 1.00 × 10-4 s-1
,
and k453K ) 2.86 × 10-4 s-1 (see the Supporting Information).
From these data the following parameter values were calculated:
Ea ) 41.5 ( 1.5 kcal/mol, ∆Hq
) 40.6 ( 1.5 kcal/mol,
443K
∆Sq
) 16 ( 2 cal mol-1 K-1, and ∆Gq
) 33.5 ( 1.5
443K
443K
kcal/mol.
The preparative decarbonylation of 20 was conducted in
1
nitrobenzene at 200 °C, giving 21 in 77% yield. The H and
13C NMR spectra of compound 21 possess numerous highly
broadened resonances resulting from dynamic exchange pro-
cesses at room temperature, a feature that complicates the
interpretation of the spectra. Fortunately, a single crystal of
compound 21 was obtained, and X-ray analysis revealed the
molecular structure shown in Figure 2. As expected, compound
21 possesses an end-to-end twist of 56°. Proton NMR spectra
of 21 were recorded in toluene-d8 at a variety of temperatures
from 223 to 363 K, and several coalescence events occur
Synthesis of Butyl-Substituted Iridium Complex 28. The low
solubility of compound 4 led us to consider a variety of more
soluble alternatives, and we speculated that perhaps the addition
of alkyl groups to complex 4 would enhance its solubility. First,
the reaction of 21 with IrCl3 and 2,2,6,6-tetramethyl-3,5-
heptanedione (in order to place two tert-butyl groups on the
enolate ligand) was examined, but none of the desired complex
was formed in several attempts, perhaps because of steric
hindrance. Second, an analogue of compound 21 with tert-butyl
groups at the para positions of the C9 and C14 phenyl groups
was prepared (data not shown), but this ligand inexplicably
failed to give the desired iridium complex, even though the tert-
(30) Allen, F. H.; Kennard, O.; Taylor, R. Acc. Chem. Res. 1983, 16, 146–
153.
(31) Plummer, B. F.; Currey, J. A.; Russell, S. J.; Steffen, L. K.; Watson,
W. H.; Bourne, S. A. Struct. Chem. 1995, 6, 167–173.
(32) Zhang, J.; Ho, D. M.; Pascal, R. A., Jr. J. Am. Chem. Soc. 2001, 123,
10919–10926.
(33) Wooi, G. Y.; White, J. W. Org. Biomol. Chem. 2005, 3, 972–974.
9
J. AM. CHEM. SOC. VOL. 130, NO. 48, 2008 16439

A R T I C L E S
Walters et al.
Figure 2. Molecular structures of (top) 20 and (bottom) 21. Thermal
ellipsoids are drawn at the 50% probability level, and hydrogen atoms have
been omitted for clarity.
Figure 3. (top) CD and (bottom) CPL spectra of compounds 7 and 13.
All of the spectra were recorded in chloroform solution.
butyl groups are far from the metal. In a third, ultimately
successful approach, compound 27, an analogue of compound
21 having n-butyl groups in place of the C10 and C13 phenyls,
was prepared and then used to prepare complex 28 (Scheme
3).
Chiroptical Properties of Twisted Acenes 7 and 13. In
addition to the conventional specific rotation measurements for
compounds 7 and 13, their circular dichroism (CD) and CPL
spectra were measured (Figure 3).34 The CD results display the
mirror-image spectra expected for the two enantiomers of each
compound as well as large values of ∆ε, as high as 5% of ε
(∆ε ≈ 700 M-1 cm-1 vs ε ) 14 000 M-1 cm-1) in the case of
the 370 nm absorption band of compound 13.
The key intermediate is 3,6-dibutyl-4,5-diphenylanthranilic
acid (25), which was prepared in a manner analogous to that
for 19.8 The aldol condensation of benzil (22) and 6-undecanone
gave 2,5-dibutyl-3,4-diphenylcyclopentadienone (23) as a red
oil. The Diels-Alder reaction of 23 with maleimide and the
dehydogenation of the adduct were achieved by heating the two
starting materials in refluxing nitrobenzene; the product imide
24 was obtained in 36% overall yield from benzil. Hofmann
degradation of 24 followed by a stringent base hydrolysis then
gave the desired amino acid 25 in 90% yield. Diazotization of
25 in the presence of cyclopentadienone 18 formed adduct 26
(23%), and decarbonylation of this norbornadienone in hot
nitrobenzene gave the ligand 27 (37%). Finally, the two-step
reaction of 27 with IrCl3 and 2,4-pentanedione produced
complex 28 in a low 4% yield.
Circularly polarized luminescence35,36 (emission having dif-
ferent intensities for the left and right circularly polarized
components) is of increasing interest because of its potential
application in stereoscopic displays37 if only a material with a
large enough emission dissymmetry could be prepared. In the
present examples, the luminescence maxima for compounds 7
and 13 occur at 400 and 445 nm, respectively, with correspond-
(34) Coughlin, F.; Westrol, M.; Oyler, K.; Byrne, N.; Kraml, C.; Zysman-
Colman, E.; Lowry, M.; Bernhard, S. Inorg. Chem. 2008, 47, 2039–
2048.
(35) Riehl, J. P.; Richardson, F. S. Chem. ReV. 1977, 77, 773–792.
(36) Riehl, J. P.; Richardson, F. S. Chem. ReV. 1986, 86, 1–16.
(37) Peeters, E.; Christiaans, M. P. T.; Janssen, R. A. J.; Schoo, H. F. M.;
Dekkers, H. P. J. M.; Meijer, E. W. J. Am. Chem. Soc. 1997, 119,
9909–9910.
Compound 28 proved to be quite soluble in many organic
solvents. However, although its solubility properties are nearly
ideal, all attempts to resolve it by chromatography (both HPLC
and SFC) on a chiral support proved unsuccessful.
9
16440 J. AM. CHEM. SOC. VOL. 130, NO. 48, 2008

Configurationally Stable Longitudinally Twisted PACs
A R T I C L E S
ing emission dissymmetries [gem ) 2(Ileft - Iright)/(Ileft + Iright)]
of 1.4 × 10-3 and 0.8 × 10-3. While it is easy to understand
that the longer-wavelength emission occurs in the larger π
system (13), the differences in emission dissymmetry cannot
be easily rationalized. However, the magnitudes of the CPL for
7 and 13 are comparable to those of other large polycyclic
aromatic compounds38 but not nearly so large as the CPL
sometimes observed in lanthanide-containing systems.39
twisted compound 7 (calculated end-to-end twist: 32°) exhibited
a reasonably large specific rotation ([R]D ) 330°) and strong
circular dichroism, but the more highly twisted 13 (58°) showed
a small specific rotation ([R]D ) 23°) but a strong CD spectrum.
Similarly, the emission dissymmetry of the less-twisted 7 was
significantly stronger than that of the more-twisted 13.
Clearly, a great deal of information is needed for the truly
effective design of materials with exceptional chiroptical
properties. Apart from the accurate prediction of the chiroptical
properties themselves, one must find stable molecules that can
be easily synthesized, easily dissolved, and easily resolved. At
present, no twisted polycyclic aromatic compound meets all of
these criteria, but the present studies show that some such
molecules should exist.
Conclusion
Two synthetic approaches were investigated that might confer
configurational stability to polyphenyl, polycyclic aromatic
compounds with large longitudinal twists. Both were successful:
two configurationally stable twisted hydrocarbons (7 and 13)
and two chiral metal complexes containing twisted polycyclic
aromatic ligands (4 and 28) were prepared. However, some
problems arose during chromatographic resolution of the
racemates. In each case, numerous chromatographic media and
solvent combinations were examined to find one (if any) that
gave satisfactory resolution. Furthermore, the resulting resolu-
tions were idiosyncratic: the successful resolution of a compound
by one method provided little insight into the best method for
the resolution of a closely related molecule. Ultimately, the two
twisted hydrocarbons (7 and 13) were resolved preparatively,
and their chiroptical properties were examined; however, the
preparative resolution of the chiral metal complex 4 was
confounded by solubility problems, and no conditions were
found for resolving its more highly soluble derivative 28.
Interestingly, the chiroptical properties of compounds 7 and
13 did not obviously parallel their structures. The modestly
Acknowledgment. This work was supported by National
Science Foundation Grants CHE-0614879 (to R.A.P.), CHE-
0449755, and DMR-0605621 (to S.B.), which are gratefully
acknowledged.
Supporting Information Available: Experimental Procedures,
containing full details of the syntheses of compounds 4, 7, 9,
13-15, 20, 21, and 24-28 as well as procedures for the
chromatographic resolution of 4, 7, 13, and 16, measurement
of the kinetics of decarbonylation of 20, and dynamic NMR
1
studies of 21; H and/or 13C NMR spectra of compounds 4, 7,
13, 15, 20, 21, and 24-28; CIF files for the structure
determinations of compounds 15, 20, and 21; chromatograms
illustrating the resolution of compounds 4, 7, and 13; kinetic
data for the decarbonylation of compound 20; and dynamic
NMR data for compound 21. This material is available free of
charge via the Internet at http://pubs.acs.org.
(38) Kawai, T.; Kawamura, K.; Tsumatori, H.; Ishikawa, M.; Naito, M.;
Fujiki, M.; Nakashima, T. ChemPhysChem 2007, 8, 1465–1468.
(39) Schippers, P. H.; van den Beukel, A.; Dekkers, H. P. J. M. J. Phys.
E: Sci. Instrum. 1982, 15, 945–950.
JA806958X
9
J. AM. CHEM. SOC. VOL. 130, NO. 48, 2008 16441