Polymorphism versus thermochromism: Interrelation of color and conformation in overcrowded bistricyclic aromatic enes

Article abstract of DOI:10.1002/chem.200501118

The nature of the thermochromic form of overcrowded bistricyclic aromatic enes (BAEs) has been controversial for a century. We report the single-crystal X-ray structure analysis of the deep-purple and yellow polymorphs of 9-(2,7-dimethyl-9H-fluoren-9-ylidene)-9H-xanthene (11), which revealed the molecules in a twisted and a folded conformation, respectively. Therefore, the deeply colored thermochromic form B of BAEs is identified as having a twisted conformation and the ambient-temperature form A as having a folded conformation. This re lationship between the color and the conformation is further supported by the X-ray structures of the deep-purple crystals of the twisted 9-(9H-fluoren-9-ylidene)-9H-xanthene (10), and of the yellow crystals of the folded 9-(11H-benzo[b]fluoren-11-ylidene)-9H-xanthene (12). Based on this conclusive crystallographic evidence, eleven previ ously proposed rationales of thermochromism in BAEs are refuted. In the twisted structures, the tricyclic moieties are nearly planar and the central double bond is elongated to 1.40 A and twisted by 42°. In the folded structures, the xanthylidene moieties are folded by 45° and the fluorenylidene moieties by 18-20°. Factors stabilizing the twisted and folded conformations are discussed, including twisting of formal single or double bonds, intramolecular overcrowding, and the significance of a dipolar aromatic "xanthenylium-fluorenide" push-pull structure.

Full text of DOI:10.1002/chem.200501118

FULL PAPER
DOI: 10.1002/chem.200501118
Polymorphism Versus Thermochromism: Interrelation of Color and
Conformation in Overcrowded Bistricyclic Aromatic Enes
[a]
P. Ulrich Biedermann,[a, b] John J. Stezowski,[c] andIsrael Agranat*
Abstract: The nature of the thermo-
chromic form of overcrowded bistricy-
clic aromatic enes (BAEs) has been
controversial for a century. We report
the single-crystal X-ray structure analy-
sis of the deep-purple and yellow poly-
morphs of 9-(2,7-dimethyl-9H-fluoren-
9-ylidene)-9H-xanthene (11), which re-
vealed the molecules in a twisted and a
lationship between the color and the
conformation is further supported by
the X-ray structures of the deep-purple
crystals of the twisted 9-(9H-fluoren-9-
ylidene)-9H-xanthene (10), and of the
yellow crystals of the folded 9-(11H-
benzo[b]fluoren-11-ylidene)-9H-xan-
thene (12). Based on this conclusive
crystallographic evidence, eleven previ-
ously proposed rationales of thermo-
chromism in BAEs are refuted. In the
twisted structures, the tricyclic moieties
are nearly planar and the central
double bond is elongated to 1.40 Š and
twisted by 428. In the folded structures,
the xanthylidene moieties are folded
by 458 and the fluorenylidene moieties
by 18–208. Factors stabilizing the twist-
ed and folded conformations are dis-
cussed, including twisting of formal
single or double bonds, intramolecular
overcrowding, and the significance of a
dipolar aromatic “xanthenylium-fluore-
nide” push–pull structure.
folded
conformation,
respectively.
Therefore, the deeply colored thermo-
chromic form B of BAEs is identified
as having a twisted conformation and
the ambient-temperature form A as
having a folded conformation. This re-
Keywords: aromaticity · conforma-
tion analysis
· polymorphism ·
strained molecules · thermochrom-
ism
Introduction
turn dark green upon heating. This change of color may also
be triggered by pressure (piezochromism)[4,5] or by UV irra-
diation at low temperatures (photochromism).[2,3,6–9]
The phenomenon of thermochromism[1–3] in overcrowded
bistricyclic aromatic enes (BAEs, 1; thermochromic ethyl-
enes[3a]) was discovered almost one hundred years ago.[4]
Meyer reported that the yellow solutions of 9,9’-bi-9(10H)-
anthracenylidene-10,10’-dione (bianthrone, 2) reversibly
[a] Dr. P. U. Biedermann, Prof. Dr. I. Agranat
Department of Organic Chemistry
The Hebrew University of Jerusalem
Philadelphia Building 201/205, Jerusalem 91904 (Israel)
Fax : (+972)2-561-1907
E-mail: isria@vms.huji.ac.il
[b] Dr. P. U. Biedermann
Institut für Organische Chemie, Biochemie und Isotopenforschung
Universität Stuttgart, Pfaffenwaldring 55, 70569 Stuttgart (Germany)
[c] Prof. Dr. J. J. Stezowski
Department of Chemistry, University of Nebraska-Lincoln
Lincoln, Nebraska 68588–0304 (USA)
BAEs have been developed as molecular switches trig-
gered by light.[10,11] Furthermore, BAEs have been incorpo-
rated into calixarenes,[12] crown ethers,[13,14] ligands for transi-
tion-metal complexation,[15–17] donor–acceptor systems,[18]
polymers,[19] liquid crystals,[11,20] Langmuir films,[21] and mo-
lecular rotors,[22] with the intention of adding a switchable
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: ORTEP draw-
ings, cell plots, bond lengths, intramolecular nonbonding distances,
bond angles, torsion angles, and intermolecular contact distances of
1
the X-ray crystal structures of 10, purple 11, yellow 11, and 12; H
and 13C NMR spectra of 11 and 18.
Chem. Eur. J. 2006, 12, 3345 – 3354
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3345

functionality and/or an easily observable response. Recently,
light-driven unidirectional motors have been discovered that
are based on closely related molecules.[23]
Thermochromism is a reversible, temperature-dependent
change of color.[1] In BAEs, the phenomenon is based on a
unimolecular[24] equilibrium[25,26] between
a colorless or
yellow ambient-temperature form A and a deep-blue or
deep-green high-temperature form B:[27]
kT
ƒ!
A
B
ƒ
that showed that thermochromism does not lead to in-
creased susceptibility.[52,53] The ESR spectrum[42] was attrib-
uted to decomposition or photochemical side reactions.[7,54,55]
The 1,1’-cyclized valence isomer 6 has been identified as the
Although A absorbs in the UV region (and close to it),
the thermochromic form has a new absorption band at l=
600–700 nm. In thermochromic BAEs, the enthalpy differ-
ence DH of the two forms is 1–7 kcalmolÀ1.[25,26,28] DH is in-
dependent of the solvent.[26] The thermochromic, photochro-
mic, and piezochromic B forms are identical.[8,28–31] The re-
versible formation of color raises the question of the nature
of the colored species.[2,27,32] In fact, the enigma of thermo-
chromism in BAEs has played an important role in the de-
velopment of the theory of organic chromophores.[33–35]
photochromic
C form, which is distinct from the B
form.[2,31,56] The photocyclization of BAEs is analogous to
that of (Z)-stilbene to give 4a,4b-dihydrophenanthrene.[8,57]
Substitution with bulky substituents at the 1-and 1 ’-posi-
tions prevents thermochromism.[2,35,53,58,59] Because substitu-
ents in the fjord region would cause prohibitive strain in
planar conformations, this has been interpreted as indicating
a planar thermochromic B form.[35,59] However, even unsub-
stituted BAEs have overcrowded fjord regions that would
cause prohibitively high strain in a planar conformation,
precluding thermal population.[60–63]
Various rationales have been suggested to explain the
thermochromic phenomenon in BAEs:
1) An aggregation–disaggregation equilibrium in solu-
tion.[36]
2) Zwitterionic forms of the molecule with ionoid carbonyl
groups (3),[33,37] in accord with Diltheyꢁs theory of or-
ganic chromophores.[38]
Several nonplanar conformations of BAEs have been
characterized by X-ray crystal structure analysis.[61,64] Repre-
sentative examples are shown in Figure 1.
3) Mesomeric betaine structures that make small contribu-
tions to the resonance hybrid at ambient temperatures,
but make large contributions at high temperatures.[39]
4) A diradical state in which the central bond linking the
planar tricyclic moieties is twisted by 908 and is “not a
true double bond”.[24,25,40]
5) A “biradikalet”: planar molecular halves twisted by less
than 908 about the central double bond, the p electrons
have parallel spins with a detectable contribution from
the triplet state.[25,41]
6) A thermally populated triplet excited state of the mole-
cule.[26,32,42]
7) A coplanar (or nearly planar) conformation with en-
hanced p delocalization and a stretched central double
bond.[35,43]
8) A twisted conformation (singlet state).[7,29,44]
9) A doubly electrocyclized diradical (4), as suggested by
Woodward and Wassermann.[45]
Figure 1. Molecular structures of 9,9’-bi-(9H-fluoren-9-ylidene) (7), 9,9’-
bi-(9(10H)-anthracenylidene)-10,10’-dione (2), and 5,5’-bi-(5H-dibenzo-
[a,d]cyclohepten-5-ylidene) (9), determined by X-ray crystallography.
10) Solid-state effects.[46]
11) 1,1’-Cyclized valence isomers 5, and/or 6.[47]
12) A double-chair conformation.[2,48,49]
The smallest BAE, 9,9’-bi-9H-fluoren-9-ylidene (bifluor-
enylidene, 7), with central five-membered rings E and F, is
bright red and adopts a twisted conformation in the crys-
tal.[65] The two tricyclic moieties are nearly planar, and the
central double bond is twisted by 328.
Zwitterions and polar mesomeric structures should be sta-
bilized by polar solvents; however, no such solvent depend-
ence of the equilibrium constant and/or the absorption spec-
trum was observed.[2,26,27,37,50] Initial reports of “paramagnet-
ic absorption” in connection with thermochromism[42,51]
were later refuted by more careful magnetic measurements
BAEs with central six-membered rings, including thermo-
chromic BAEs, such as bianthrone (2)[66] and 9,9’-bi-9H-
3346
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3345 – 3354

Polymorphism Versus Thermochromism
FULL PAPER
Here, we report the single-crystal X-ray structure analyses
of the purple and the yellow polymorphs of 9-(2,7-dimethyl-
9H-fluoren-9-ylidene)-9H-xanthene (11). This example of a
single BAE with two polymorphs of such widely differing
colors reveals the nature of the thermochromic form. Fur-
thermore, we report the X-ray structures of the purple crys-
tals of the parent BAE 10, and of the yellow crystals of 9-
(11H-benzo[b]fluoren-11-ylidene)-9H-xanthene (12).
xanthen-9-ylidene (dixanthylene, 8),[46] adopt anti-folded
conformations.[61,64] The central double bonds are not twist-
ed, but slightly pyramidalized. The rings E and F have boat
conformations leading to tricyclic moieties that are folded in
opposite directions (anti). The dihedral angles of the least-
square planes, A–B (and C–D), are 36–558. Based on these
X-ray crystal structures, the ambient-temperature form A of
BAEs was characterized as the anti-folded conformation. In
5,5’-bis-5H-dibenzo[a,d]cyclohepten-5-ylidene (9), which has
central seven-membered rings, a metastable syn-folded con-
formation has been found in addition to an anti-folded con-
formation.[67] In 9, the degrees of folding are very high: A–
B=55.78 in anti-9 and A–B/C–D=62.68/55.58 in syn-9.
Results
Synthesis: The fluorenylidene-xanthenes 10, 11, and 12 have
been synthesized by Bartonꢁs diazothione addition/twofold
extrusion method.[69] Synthesis and characterization of the
parent compound 10[68,70–74] and 12[68,73] has already been re-
ported. BAE 11 was synthesized by starting from the tricy-
clic ketones 2,7-dimethyl-9H-fluoren-9-one (13)[75] and 9H-
xanthen-9-one (14). Ketone (13) was reacted with hydrazine
hydrate in ethanol to give 2,7-dimethyl-9H-fluoren-9-one
hydrazone (15), which was oxidized with yellow mercury
oxide to 9-diazo-2,7-dimethyl-9H-fluorene (16).[76,77] Ketone
(14) was converted to 9H-xanthene-9-thione (17)[78] with
Lawessonꢁs reagent, and reacted with (16) in boiling ben-
zene to give the intermediate thiirane 2,7-dimethyl-dispiro-
(9H-fluorene-9,2’-thiirane-3’,9’’-[9H]-xanthene) (18). The
thiirane sulfur was eliminated with triphenylphosphine[71] in
boiling benzene to give the purple 9-(2,7-dimethyl-9H-fluo-
ren-9-ylidene)-9H-xanthene (11) (Scheme 1), which was pu-
In 1963, Mills and Nyburg reported crystal structures of
two polymorphs of dixanthylene (8): a yellow b form ob-
tained by slowly cooling an m-xylene solution of 8 and a
blue–green a form obtained by sublimation at 2508C.[46]
Both polymorphs contain molecules in very similar anti-
folded conformations.[46] The authors found that the confor-
mations are identical within the high experimental error
(bond lengths Æ0.02 Š, angles Æ1.58, R=20%).[46] They
concluded: “We believe the differences in packing cause the
differences in color of the crystals, but we are unable to
specify what features of the packing are responsible.”[46]
Kortüm and co-workers argued that, at the temperature of
sublimation, only 1% of the dixanthylene molecules would
be in the thermochromic B form, sufficient for the green
color of the crystals, but not detectable by X-ray diffrac-
tion.[32,48] As the B form has not yet been crystallized and
subjected to X-ray structure analysis, one has to rely on indi-
rect and inconclusive arguments regarding its structure.[2]
An analysis of the previously reported X-ray structures in-
dicated that BAEs with central five-membered rings E and
F have twisted conformations, whereas BAEs with central
six-membered rings have anti-folded conformations.[61] In
heteromerous BAEs that combine tricyclic moieties with
one central five-membered ring and one central six-mem-
bered ring, twisted and folded conformations may be ex-
pected to have similar energies. This introduces the possibil-
ity of crystallizing both conformations in the same BAE or
in closely related derivatives. The present study is focused
on the heteromerous BAE 9-(9H-fluoren-9-ylidene)-9H-xan-
thene (10) and its derivatives, whereby a subtle equilibrium
of twisted and folded conformations was found in solu-
tion.[68] A comparison of the structures of twisted and folded
conformations of closely related BAEs may reveal the fac-
tors governing the relative stability of the conformations.
rified by chromatography with
a yield of 67%. The
1H NMR spectrum revealed signals for the fjord region hy-
drogen atoms at d=8.16 ppm for H1 and H8 and at
7.71 ppm for H1’ and H8’. The carbon atoms of the central
double bond appear at d=131.2 (C9’), 130.2 ppm (C9).
Scheme 1. Synthesis of 9-(2,7-dimethyl-9H-fluoren-9-ylidene)-9H-xan-
thene (11).
Crystallization: Single crystals for X-ray crystallography
were grown by sublimation of 11 at 1808C under argon
(2 Torr). After 2 d, purple crystals formed on the walls of
Chem. Eur. J. 2006, 12, 3345 – 3354
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3347

I. Agranat et al.
the tube; m.p. 193–1958C. In addition, a small amount of
yellow crystals formed in the cooler regions; m.p. 1458C
(purple melt). The yellow crystals gave purple solutions. The
NMR spectrum of the yellow crystals dissolved in CDCl3
was identical to that of solutions of the purple material.
Single crystals of the parent compound 10 were grown by
recrystallization from a toluene solution;[74] purple crystals,
m.p. 2448C. The same purple-crystal modification was ob-
tained by sublimation at 2008C under 1 Torr argon. Single
crystals of 9-(11H-benzo[b]fluoren-11-ylidene)-9H-xanthene
(12) were grown by sublimation in a sealed glass tube at
180–1908C and 0.05 Torr to afford yellow crystals, m.p. 221–
2228C.[68] In solution, 12 is purple.[68]
molecules adopt a twisted conformation with nearly planar
tricyclic moieties forming
a dihedral angle of 52.38
(Figure 3). The central double bond is twisted by w=
42.3(4)8. Notably, the molecules of 10 are orientationally
disordered about a crystallographic C2 axis that passes
through the fjord regions.
Crystal structure of purple 9-(2,7-dimethyl-9H-fluoren-9-yli-
dene)-9H-xanthene (11): In the purple polymorph, purple
11, the molecule adopts a twisted conformation (Figure 2a).
The two tricyclic moieties are nearly planar and form a di-
hedral angle of 50.18. The ethylenic twist w of the central
double bond, defined as the average of the torsion angles
t(C8a-C9-C9’-C8a’) and t(C9a-C9-C9’-C9a’),[63,64] is 42.1(3)8.
Figure 3. ORTEP drawing of the X-ray crystal structure of 9-(9H-fluo-
ren-9-ylidene)-9H-xanthene (10). Anisotropic displacement parameters
are drawn at the 50% probability level (120 K).
Crystal structure of 9-(11H-benzo[b]fluoren-11-ylidene)-9H-
xanthene (12): In the yellow crystal of 12, the molecule
adopts the anti-folded conformation shown in Figure 4. The
xanthylidene moiety is folded by A–B=45.5(1)8 and the flu-
orenylidene moiety is folded by C–D=19.7(2)8. The two
six-membered rings of the naphthalene ring system are
almost coplanar: D–G=2.2(2)8. The ethylenic twist w of the
central double bond is 1.0(5)8. The two carbon atoms of the
central double bond are anti-pyramidalized: c(C9)=
À5.2(5)8 and c(C11’)=4.1(6)8.
Figure 2. ORTEP drawings of the molecular structures of 9-(2,7-dimeth-
yl-9H-fluoren-9-ylidene)-9H-xanthene (11) in the (a) purple and (b)
yellow polymorphs. Anisotropic displacement parameters are drawn at
the 50% probability level.
Crystal structure of yellow 9-(2,7-dimethyl-9H-fluoren-9-yli-
dene)-9H-xanthene (11): In the yellow polymorph, yellow
11, the conformation is characterized by folded xanthylidene
and fluorenylidene moieties (Figure 2b). The dihedral angle
of the least-squares planes A–B, defined by the ring carbon
atoms, is 44.5(2)8. The fluorenylidene moiety has a smaller
folding dihedral angle, C–D=18.3(1)8, and is folded in the
opposite direction (anti-folding). The twist of the central
double bond is very small, w=À1.9(9)8; however, there is a
substantial anti-pyramidalization, c(C9)=À8(1)8 and c-
(C9’)=5(1)8. The pyramidalization angles are defined as
c(C9)=[t(C9a-C9-C9’-C8a) MOD3608]À1808, and c(C9’)=
[t(C9a’-C9-C9’-C8a’) MOD 3608]À1808.[63,64]
Figure 4. ORTEP drawing of the molecular structure of 9-(11H-benzo[b]-
fluoren-11-ylidene)-9H-xanthene (12) derived from the X-ray structure
of the yellow crystals. Anisotropic displacement parameters are drawn at
the 50% probability level.
Crystal structure of 9-(9H-fluoren-9-ylidene)-9H-xanthene
(10): In the purple crystals of the parent compound 10, the
3348
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3345 – 3354

Polymorphism Versus Thermochromism
FULL PAPER
Discussion
almost planar (RMS deviation from plane <0.07 Š for
heavy atoms) and the molecules are twisted about the cen-
tral double bond. The dihedral angle of the least-squares
planes AEB–CFD is 50.18 in 11 and 52.38 in 10. The main
burden of the out-of-plane deformation is on the central
double bonds that are highly twisted, w=42.1(3)8 (purple
11), 42.3(4)8 (10), and elongated to C9=C9’=1.400(2) Š
(purple 11) and 1.401(6) Š (10).
Similarly, the folded conformations found in yellow 11
and 12 closely resemble each other. The folding dihedrals
angles of the least-squares planes A–B and C–D are 44.5(2)
and 18.3(1)8 (yellow 11) versus 45.5(1) and 19.7(2)8 (12). In
both folded conformations, the central double bond is short,
1.353(5) Š in yellow 11 and 1.360(4) Š in 12. In the folded
structures, the main out-of-plane deformations are in the
formal single bonds connecting the central double bond
with the peripheral aromatic rings, and in the bridges X and
Both twisted and folded conformations of the same bona
fide BAE (11) were characterized for the first time by
single-crystal X-ray analysis. Moreover, we report the first
twisted conformations of heteromerous BAEs (purple 11
and 10) that are based on crystal structure determinations.
Previously, twisted conformations have been reported for bi-
fluorenylidene and its derivatives only.[13,16,65,79] The purple
crystals of 9-(2,7-dimethyl-9H-fluoren-9-ylidene)-9H-xan-
thene (11) comprise twisted molecules and the yellow crys-
tals contain folded molecules, which demonstrates the inter-
relation of color and conformation. The parent compound
9-(9H-fluoren-9-ylidene)-9H-xanthene (10) forms purple
crystals with twisted molecules and 9-(11H-benzo[b]fluoren-
11-ylidene)-9H-xanthene (12) forms yellow crystals with
folded molecules. In view of the very different crystal-pack-
ing motifs found in the four X-ray structures and the lack of
any indication of unusual intermolecular interactions (Sup-
porting Information Figures S4–S7 and Tables S5–S8), it is
highly unlikely that the purple or yellow colors of the crys-
tals are caused by an intermolecular solid-state effect. This
questions the conclusion made by Mills and Nyburg.[46] In
each case, the deep purple crystals revealed a twisted con-
formation of the molecules, and the yellow crystals revealed
a folded conformation. This strongly indicates that color is
determined by molecular conformation. The large twist in
the central double bond reduces the p overlap, which causes
a substantial red shift of the UV-visible absorptions and
smaller HOMO–LUMO gaps in the twisted conforma-
tions.[80,81] Therefore, the deeply colored thermochromic B
form of BAEs may be assigned to a twisted conformation of
the molecule, whereas the yellow (or colorless) ambient-
temperature form A is confirmed to be the (anti-)folded
conformation.
À
À
Y. In particular, the C9 C9a and C9 C8a bonds of the xan-
thylidene moiety are highly twisted with, for example, t(C9’-
C9-C9a-C1)=53.0(8)8, and t(C9’-C9-C8a-C8)=À49.0(8)8 in
yellow 11. The formal single bonds of the fluorenylidene-
xanthenes are longer in the folded structures than in the
À
twisted structures: in the xanthylidene moieties, C9 C9a=
À
1.488 versus 1.465 Š, C4a O10=1.389 versus 1.369 Š, in
À
the fluorenylidene moieties, C9’ C9a’=1.497 versus 1.475 Š
(average of equivalent bonds in yellow 11 and 12, versus
average in purple 11). These changes in the bond lengths in-
dicate that, in the folded conformations, the p systems are
more localized in the aromatic rings and the double bond. A
priori, twisting of formal single bonds as in the folded con-
formations may be energetically more favorable than twist-
ing about the central double bond as in the twisted confor-
mation. However, delocalization of the p-electron system
because of the more planar tricyclic moiet-
ies may stabilize the twisted conformations.
Selected structural parameters of the X-ray crystal struc-
tures of 10, purple 11, as well as yellow 11 and 12 are sum-
marized in Table 1 (see also Supporting Information Ta-
bles S1–S4). The molecules in purple 11 and 10 have very
similar twisted conformations. The tricyclic moieties are
Moreover, enhanced contributions of the di-
polar aromatic “xanthenylium-fluorenide”
(19) push–pull structure may be a signifi-
cant factor in stabilizing the twisted confor-
mation of fluorenylidene-xanthenes.
In the fjord regions of yellow 11 and 12,
overcrowding is severe for the C1···H1’ and
Table 1. Selected structural parameters of the X-ray crystal structures.
C8···H8’ distances, which average ꢀ2.40 Š;
the overlap of the van der Waals radii (C 1.71, H 1.15 Š)[82]
is 16%. The carbon–carbon contacts are also very short: the
C1···C1’ and C8···C8’ distances are 2.99–3.05 Š, correspond-
ing to an overlap of ꢀ12%. On the other hand, in the twist-
ed conformation of purple 11, the above intramolecular con-
tacts of C1 and C8 with C1’, C8’, H1’, and H8’ are longer, in-
dicating considerably less overcrowding (5–9% overlap).
However, the very short H1···C9a’ and H8···C8a’ contacts in
twisted 11 (average 2.56 Š, 11% overlap) are not over-
crowded in the folded conformations.
Compound
10
11
11
12
Crystal color
Conformation
purple
twisted
purple
twisted
yellow
folded
yellow
folded
w[8][a]
42.3(4)
0.3(2)
5.0(2)
1.401(6)
0.6(6)
À1.3(6)
3.144(4)
3.125(3)
42.1(3)
6.01(3)
4.9(1)
1.400(2)
À1.2(3)
À2.6(3)
3.145(3)
3.113(3)
À1.9(9)
44.5(2)
18.3(1)
1.353(5)
À8(1)
1.0(5)
45.5(1)
19.7(2)
1.360(4)
À5.2(5)
4.1(6)
A–B[8][b]
C–D[8][b]
C9=C9’[Š]
c(C9)[8][c]
c(C9’)[8][c]
C1···C1’[Š]
C8···C8’[Š]
À5(1)
3.046(7)
2.990(7)
2.996(5)
3.051(5)
[a] Ethylenic twist w=1= [(C9a-C9-C9’-C9a’)
+ t(C8a-C9-C9’-C8a’)].
2
In summary, the enhanced delocalization of the p-electron
system, the push–pull effect, and the more effective dissipa-
tion of intramolecular overcrowding in the fjord regions
counterbalance the strain introduced by twisting about the
[b] Dihedral angles of the least-squares planes defined by the aromatic-
ring carbon atoms. [c] Pyramidalization c(C9)=[t(C9a-C9-C9’-C8a)
modulus (MOD) 3608]À1808, c(C9’)=[t(C9a’-C9-C9’-C8a’) MOD
3608]À1808.
Chem. Eur. J. 2006, 12, 3345 – 3354
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3349

I. Agranat et al.
central double bond. Therefore, in fluorenylidene-xanthenes,
the twisted and folded conformations may be expected to
have similar energies. Indeed, an analysis of the NMR
chemical shifts of the hydrogen atoms in the fjord regions of
fluorenylidene-xanthenes in solution indicates an equilibri-
um of twisted and folded conformations.[68] Furthermore, a
very low enantiomerization barrier of 6.5 kcalmolÀ1 was ob-
served in the isopropyl derivative of 10.[68] This process also
interconverts twisted and folded conformations in solu-
tion,[68] and explains the instantaneous change in color as
the yellow crystals of 11 and 12 are dissolved.
Recently, a criterion for thermochromism has been de-
rived from B3LYP/6–31G* DFT calculations of various ther-
mochromic and nonthermochromic BAEs.[80] B3LYP/6–
31G* conformational energies agree with experimental re-
sults to within Æ1 kcalmolÀ1.[80] Three types of conforma-
tional behavior have been distinguished. Thermochromism
appears in BAEs with Type 1 conformational behavior that
satisfy the following necessary conditions: i) an anti-folded
global minimum conformation (a) corresponding to the am-
bient-temperature form A, and a low-energy local-minimum
twisted conformation (t) corresponding to the thermochro-
mic form B; ii) the energy difference between these two
conformations has to be sufficiently small to allow thermal
population of t, E(t)ÀE(a)<8 kcalmolÀ1.[80] Type 3 confor-
mational behavior is displayed by 9-(9H-fluoren-9-ylidene)-
9H-xanthene (10): the folded conformation is 1.1 kcalmolÀ1
higher in energy than the twisted conformation.[80] Thus, the
colored twisted conformation is energetically more stable,
and, therefore, dominates in the equilibrium at all tempera-
tures. This is consistent with the observed deep-purple color
of the solutions of all fluorenylidene-xanthenes, which show
the thermochromic form B at room temperature. However,
the folded conformation may be substantially populated in
thermal equilibrium.[80] Moreover, owing to the very similar
energies, crystal-packing forces may freeze the twisted as
well as the folded conformation in the solid state.
ent in the 1-position in the folded conformation. H1 and H8
are tilted away from the opposing fluorenylidene moiety
(see Figure 2b and Figure 4).
The structures of the heteromerous fluorenylidene-xan-
thenes may be compared to those of the corresponding ho-
momerous BAEs bifluorenylidene (7) and dixanthylene (8).
The bright red bifluorenylidene has a twisted conformation
with w=32.08 and C9=C9’=1.36 Š (b modification),[65]
which is 88 less twisted and 0.036 Š shorter than the twisted
fluorenylidene-xanthenes. Nevertheless, the C···C distances
in the nonbonding fjord region are slightly longer in b-bi-
fluorenylidene, namely, 3.17 Š (7% overlap) versus 3.13 Š
(8% overlap) in purple 11 (averages). The C9a-C9-C8a
bond angle of 1158 in the xanthylidene moiety of twisted 11
is larger than the bond angle of ꢀ1058 imposed by the cen-
tral five-membered ring of a fluorenylidene moiety. This
larger angle pushes the carbon and hydrogen atoms in the 1-
and 8-positions deeper into the overcrowded fjord regions.
Therefore, a higher twist is required in fluorenylidene-xan-
thenes to alleviate the intramolecular overcrowding to a per-
missible level. In the reported (low-resolution) X-ray struc-
tures of Mills and Nyburg, dixanthylene (8) has an anti-
folded conformation, with A–B=40 and 438 in the b and a
polymorphs, respectively. This is comparable to the folding
of the xanthylidene moieties in yellow 11 (44.58) and 12
(45.58). The nonbonding C···C distances of ꢀ3.0 Š are also
comparable in the dixanthylene structures as well as in the
folded fluorenylidene-xanthenes. Apparently, the small 5-
ring bond angle in the fluorenylidene compensates for its
lower degree of folding. Several cases have been reported in
which overcrowded ethylenes crystallized as two distinctly
colored pseudopolymorphs, but with solvent molecules in
one of the two crystal structures.[83] In BAE 11, both poly-
morphs[84] are free of solvent, thus, excluding solvatochrom-
ism as a rationale for the different colors.
The X-ray crystal structures of two fluorenylidene-xan-
thene derivatives have already been reported: 12-(9H-fluo-
ren-9-ylidene)-2-methoxy-12H-benzo[a]xanthene (20), and
its 4,5-diazafluorene derivative 21.[17] The yellow crystals
contained molecules in folded conformations. The ben-
zo[a]xanthylidene moieties of the three independent mole-
cules of 20 and of 21 are folded by A–B=42–478, the fluore-
nylidene moieties by C–D=13–168. The central double
bonds are twisted by w=1–58. Evidently, little conforma-
tional change is required to accommodate a bulky substitu-
Conclusion
The yellow and the purple polymorphs of 9-(2,7-dimethyl-
9H-fluoren-9-ylidene)-9H-xanthene (11) are the first exam-
ples of a bona fide BAE 1 in which a folded as well as a
twisted conformation was characterized by single-crystal X-
ray structure analysis (Figure 2). In the yellow polymorph,
the molecules have a folded conformation with A–B=
44.5(2)8 (xanthylidene), and C–D=18.3(1)8 (fluorenyli-
dene). In the purple polymorph, the molecules have a twist-
ed conformation with w=42.1(3)8 and nearly planar tricyclic
moieties. The interrelation of color and conformation is un-
ambiguously demonstrated. Based on this conclusive crystal-
lographic evidence, the deeply colored purple (or green)
thermochromic form B of BAEs is identified as the twisted
conformation. The assignment of the yellow or colorless am-
bient-temperature form A as (anti-)folded conformation is
confirmed.
This interrelation of color and molecular conformation is
supported further by the purple crystals of the unsubstituted
3350
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3345 – 3354

Polymorphism Versus Thermochromism
FULL PAPER
dimethyl-9H-fluoren-9-one hydrazone (15) (1.000 g, 4.49 mmol) were
carefully ground by using a mortar, transferred to a round-bottomed
flask equipped with magnetic stirrer and drying tube, and dry diethyl
ether (25 mL) was added. Upon dropwise addition of a freshly prepared
saturated solution of KOH in ethanol (0.5 mL), the reaction mixture
turned deep red. The formation of 16 was monitored by TLC (silica,
Rf(15)=0.07 yellow, Rf(16)=0.78 orange (petroleum ether/toluene 4:1)).
After 3 h of stirring at RT, the reaction was complete. The solution was
filtered, and the residue was washed with diethyl ether (3”5 mL). The
combined filtrates were evaporated to afford 0.924 g (4.19 mmol, 95%)
of 9-diazo-2,7-dimethyl-9H-fluorene (16) as orange-red powder, m.p.
parent compound 10, which has a twisted conformation with
w=42.3(4)8, and the yellow crystals of the benzoannelated
derivative 12, which revealed a folded conformation with
A–B=45.5(1)8 and C–D=19.7(2)8.
Enhanced p delocalization in the almost planar tricyclic
moieties, the push–pull effect, and reduced overcrowding
stabilizes the twisted conformation relative to the anti-
folded conformation, counterbalancing the high strain of the
twisted double bond. In solution, the twisted conformation
dominates in fluorenylidene-xanthenes at all temperatures,
leading to “thermochromism at room temperature”. Owing
to the small energy difference between the twisted and the
folded conformations, both conformations could be stabi-
lized as single crystals and, hence, elucidated unambiguous-
ly.
1
1198C (lit. 113.5–1158C[76]); H NMR (400 MHz, CDCl3, 295 K): d=7.782
(d, 7.9 Hz, 2H; H4, H5), 7.302 (quintet, 0.7 Hz, 2H; H1, H8), 7.128 (dq,
7.8, 0.7 Hz, 2H; H3, H6), 2.489 ppm (s, 6H; CH3); 13C NMR (100 MHz,
CDCl3, 295 K): d=135.9 (C), 133.2 (C), 129.3 (C), 125.8 (CH), 120.4
(CH), 119.8 (CH), 62.8 (CN2), 21.7 ppm (CH3).
9H-Xanthene-9-thione (17): Thione 17 was prepared according to a liter-
ature procedure.[78] 9H-Xanthen-9-one (14) (1.57 g, 8.00 mmol) and Law-
essonꢁs reagent (1.63 g, 4.04 mmol, 1% excess) were dissolved in dry ben-
zene (15 mL) and refluxed for 1 h in a round-bottomed flask with a mag-
netic stirrer, a reflux condenser, and a CaCl2 drying tube. The reaction
was monitored by TLC (silica, Rf(14)=0.09, Rf(17)=0.52 yellow (petro-
leum ether/diethyl ether 19:1)). Predried silica (10 g, 1208C/20 Torr) was
added to the reaction mixture and the benzene was carefully evaporated
under a vacuum. The dry powder was added to the top of a silica column
(l=12.5 cm, 1=4.5 cm) and eluted with petroleum ether/diethyl ether
(98:2). The blue-green fractions containing pure 17 (TLC) were com-
bined and evaporated under a vacuum to afford 1.36 g (6.39 mmol, 80%)
9H-xanthene-9-thione (17) as small black needles with a greenish luster,
m.p. 1558C (lit. 1578C[78]); 1H NMR (400 MHz, CDCl3, 295 K): d=8.762
(ddd, 8.2, 1.7, 0.5 Hz, 2H; H1, H8), 7.774 (td, 8.4, 7.0, 1.7 Hz, 2H; H3,
H6), 7.518 (ddd, 8.4, 1.2, 0.5 Hz, 2H; H4, H5), 7.389 ppm (td, 8.2, 7.0,
1.2 Hz, 2H; H2, H7); 13C NMR (100 MHz, CDCl3, 295 K): d=205.0 (C9),
150.5 (C4a, C10a), 134.9 (CH), 129.9 (CH), 129.1 (C8a, C9a), 124.7 (CH),
118.3 ppm (CH).
Experimental Section
General considerations: Diethyl ether and benzene were dried over
sodium wire. Petroleum ether was distilled before use (b.p.ꢀ608C). The
NMR spectra were measured in CDCl3 solution at 295 K by using an
AMX or
a DRX 400 MHz spectrometer (Bruker Analytical Instru-
ments). Chemical shifts d are reported relative to d(CHCl3)=7.275 ppm
1
for H NMR spectra, and relative to d(CDCl3)=77.008 ppm for 13C NMR
spectra. Multiplet shapes are abbreviated as: singlet (s), broad singlet
(brs), doublet (d), doublet of doublet of doublets (ddd), and triplet of
doublets (td). Assignment of the 1H and 13C NMR spectra was based on
2D DQF-COSY, 2D long-range 1H–13C correlation, and 2D inverse
13C–1H correlation spectra. The three-bond 1H–13C cross peaks of the
xanthylidene carbon atoms C4a, C10a at dꢀ155.0 ppm with the hydrogen
atoms H1, H8 and H3, H6 were used to identify the xanthylidene hydro-
gen atoms. The aromatic hydrogen atoms of the 2,7-dimethyl-fluorenyl-
idene moiety have characteristic multiplet shapes: H1’ and H8’ have only
small coupling constants (0.7 Hz) with the methyl and meta hydrogen
atoms; H3’ and H6’ have one ortho coupling constant (7.7 Hz) and small
coupling constants with methyl and meta hydrogen atoms; H4’ and H5’
have only one ortho coupling constant.
2,7-Dimethyl-dispiro(9H-fluorene-9,2’-thiirane-3’,9’’-[9H]-xanthene) (18):
Compound 18 was prepared analogously to the parent thiirane dispir-
o(9H-fluorene-9,2’-thiirane-3’,9’’-[9H]-xanthene).[70]
9H-Xanthene-9-
thione (17, 0.520 g, 2.35 mmol) was dissolved in dry benzene (10 mL),
and 9-diazo-2,7-dimethyl-9H-fluorene (16, 0.500 g, 2.36 mmol) dissolved
in benzene (5 mL) was added. Upon heating, nitrogen evolved and the
reaction mixture turned deep purple. After refluxing for 4 h, the reaction
was complete (TLC, silica, Rf(18)=0.37, Rf(11)=0.42, Rf(17)=0.60
yellow, Rf(16)=0.70 orange (petroleum ether/toluene 4:1)). The reaction
mixture was cooled to RT and evaporated. The black residue was dis-
solved in warm benzene (2 mL) and diluted with petroleum ether
(3 mL). Upon cooling and standing overnight, crystals of 2,7-dimethyl-
2,7-Dimethyl-9H-fluoren-9-one (13): The sample, prepared according to
a literature procedure,[75] was purified by column chromatography (silica,
toluene/petroleum ether (608C) 20:80–80:20). Orange crystals, m.p. 155–
1
1568C (lit. 1578C[75]); H NMR (400 MHz CDCl3, 295 K): d=7.456 (quin-
tet, 0.7 Hz, 2H; H1, H8), 7.369 (d, 7.5 Hz, 2H; H4, H5), 7.270 (dq, 7.6,
0.8 Hz, 2H; H3, H6), 2.378 ppm (s, 6H; CH3).
dispiro(9H-fluorene-9,2’-thiirane-3’,9’’-[9H]-xanthene)
(18,
0.181 g,
0.448 mmol, 19.0%) formed, which were filtered off and washed with pe-
troleum ether. Another crop of 18 (0.121 g, 0.30 mmol, 12.7%) was iso-
lated from the mother liquor by column chromatography (silica, petrole-
um ether/dichloromethane 95:5–80:20), m.p. 159–1628C; 1H NMR
(400 MHz, CDCl3, 295 K): d=7.977–7.953 (m, 2H; H4’’, H5’’), 7.381 (d,
7.6 Hz, 2H; H4, H5), 7.255–7.193 (m, 4H; H2’’, H3’’, H6’’, H7’’), 7.006–
6.983 (m, 2H; H1’’, H8’’), 6.950 (dq, 7.6, 0.7 Hz, 2H; H3, H6), 6.824
(quintet, 0.7 Hz, 2H; H1, H8), 2.121 ppm (s, 6H; CH3); 13C NMR
(100 MHz, CDCl3, 295 K): d=156.6 (C), 141.7 (C), 138.6 (C), 135.3 (C),
128.7 (CH), 128.5 (CH), 128.3 (CH), 125.7 (CH), 125.0 (C), 122.1 (CH),
118.9 (CH), 116.4 (CH), 58.2 (C), 54.4 (C), 21.3 ppm (CH3).
2,7-Dimethyl-9H-fluoren-9-one hydrazone (15): Hydrazone 15 was pre-
pared analogously to 9H-fluoren-9-one hydrazone.[85] Hydrazine hydrate,
(98%, d=1.032 gmLÀ1, 1.47 g, 29.3 mmol) was added to 2,7-dimethyl-
9H-fluoren-9-one (13; 1.222 g, 5.87 mmol) dissolved in hot ethanol
(20 mL), and refluxed for 2.5 h. The reaction was monitored by TLC
(silica, Rf(13)=0.56 yellow, Rf(15)=0.25 yellow (toluene/chloroform
4:1)). The reaction mixture was cooled in an ice bath, and the orange
precipitate was filtered off and washed with methanol to afford 2,7-di-
methyl-9H-fluoren-9-one hydrazone (15) as a yellow powder (0.421 mg
1.90 mmol, 32%), m.p. 165–1668C. Reduction of the filtrate volume to
ꢀ5 mL caused precipitation of additional 15 (0.687 mg, 3.09 mmol,
53%), yellow powder, m.p. 1648C (toluene) (lit. 166–1678C
(decomp)[76]); 1H NMR (400 MHz, CDCl3, 295 K): d=7.734 (quintet,
0.7 Hz, 1H; H1), 7.604 (d, 7.7 Hz, 1H; H4), 7.531 (quintet, 0.7 Hz, 1H;
H8), 7.495 (d, 7.7 Hz, 1H; H5), 7.245 (dq, 7.7, 0.7 Hz, 1H; H3), 7.147
(dq, 7.7, 0.8 Hz, 1H; H6), 6.420 (brs, 2H; NH2), 2.456 (s, 3H; CH3),
2.416 ppm (s, 3H; CH3).
9-(2,7-Dimethyl-9H-fluoren-9-ylidene)-9H-xanthene (11): 2,7-Dimethyl-
dispiro(9H-fluorene-9,2’-thiirane-3’,9’’-[9H]-xanthene)
(18,
0.11 g,
0.26 mmol) was dissolved in dry benzene (8 mL), and triphenylphosphine
(0.076 g, 0.29 mmol) in benzene (2 mL) was added. The mixture was re-
fluxed for 3 h. The reaction was monitored by TLC (silica, Rf(18)=0.37,
Rf(11)=0.42 (petroleum ether/toluene 4:1)). The solvent was evaporated
under reduced pressure, and the raw product was washed with methanol
and dried (134.8 mg). The black powder was purified by column chroma-
tography (silica, petroleum ether/dichloromethane 95:5–80:20) to give 9-
9-Diazo-2,7-dimethyl-9H-fluorene (16): Diazofluorene 16 was prepared
according to
(0.511 g, 3.60 mmol), yellow mercury oxide (1.608 g, 7.43 mmol), and 2,7-
a
literature procedure.[76] Anhydrous sodium sulfate
Chem. Eur. J. 2006, 12, 3345 – 3354
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3351

I. Agranat et al.
(2,7-dimethyl-9H-fluoren-9-ylidene)-9H-xanthene (11) as a dark-purple
powder (64.6 mg, 0.17 mmol, 67%), m.p. 189–1918C; 1H NMR
(400 MHz, CDCl3, 295 K): d=8.157 (ddd, 7.9, 1.5, 0.5 Hz, 2H; H1, H8),
7.706 (quintet, 0.7 Hz, 2H; H1’, H8’), 7.563 (d, 7.8 Hz, 2H; H4’, H5’),
7.391 (td, 8.2, 6.7, 0.5 Hz, 2H; H3, H6), 7.354 (ddd, 8.2, 1.8, 0.5 Hz, 2H;
H4, H5), 7.125 (td, 8.0, 6.7, 2.1 Hz, 2H; H2, H7), 7.080 (dq, 7.7, 0.7 Hz,
2H; H3’, H6’), 2.277 ppm (s, 6H; CH3); 13C NMR (100 MHz, CDCl3,
295 K): d=154.0 (C4a, C10a), 139.6 (C2’, C7’), 138.1 (C4a’, C4b’), 135.0
(C8a’, C9a’), 131.2 (C9’), 130.2 (C9), 130.1 (C1, C8), 129.8 (C3, C6), 128.4
(C3’, C6’), 124.9 (C8a, C9a), 124.8 (C1’, C8’), 122.5 (C2, C7), 118.8 (C4’,
C5’), 117.6 (C4, C5), 21.7 ppm (CH3); HREI-MS: m/z (%) for C28H20O:
373.16 (29.8) [M+1]+, 372.15 (100) [M]+, 356.12 (8.8) [MÀCH4]+, 342.10
9-(9H-Fluoren-9-ylidene)-9H-xanthene (10): C26H16O, Mr =344.39, mono-
clinic, C2/c, deep purple, a=21.798(6), b=7.834(6), c=10.187(5) Š, b=
106.19(4)8, V=1670.7(16) Š3, Z=4, 1calcd =1.369 gcmÀ3, T=120 K, 3880
total reflections, qmax =358, NicoletP3F diffractometer. The molecules of
10 show an orientational disorder about a crystallographic C2 axis that
runs through the overcrowded fjord regions, perpendicular to the central
double bond (Supporting Information Figure S2). Owing to the close su-
perposition, only average positions for the fjord region carbon atoms C1/
C1’ and C8/C8’ could be refined with anisotropic displacement parame-
ters. The atoms C2, C7, C8a, and C9a are separated by less than 0.4 Š
from the atoms C2’, C7’, C8a’, and C9a’, respectively, of the disordered
molecule (optical resolution 0.7l/2 sin(qmax)=0.43 Š.). Refinement with
anisotropic displacement parameters led to a high correlation between
the parameters. Therefore, these atoms were refined with isotropic dis-
placement parameters. The remaining heavy atoms were refined with ani-
sotropic displacement parameters. Hydrogen atoms were refined as
riding on the respective carbon atoms. The final residual factors are
wR2=0.221, R1=0.089, and GOF=1.023 for 186 parameters and 3659
unique reflections, 2636 reflections with I<2s(I). Residual electron den-
sity: 0.33/À0.28 eŠÀ3 (max/min).
9-(11H-Benzo[b]fluoren-11-ylidene)-9H-xanthene (12): C30H18O, Mr =
394.44, orthorhombic, Pbca, yellow, a=14.508(2), b=14.168(2), c=
19.490(3) Š, V=4006(1) Š3, Z=8, 1calcd =1.308 gcmÀ3, T=293 K, 6455
total reflections, qmax =308, NicoletP3F diffractometer. Hydrogen atoms
were refined with isotropic displacement parameters. The final residual
factors are wR2=0.127, R1=0.115, and GOF=0.977 for 355 parameters
and 5837 unique reflections, 3388 reflections with I<2s(I). Residual elec-
tron density: 0.21/À0.18 eŠÀ3 (max/min).
(12.3) [MÀC2H6]+, 186.07 (7.2) [M]++
.
Single crystals for X-ray crystallography were grown by sublimation. A
small sample of 11 in a glass vial was inserted into the closed lower end
of a quartz glass tube (30 cm long, 10 mm 1). The tube was flushed with
argon through a glass capillary, and the pressure was adjusted to ꢀ1.5–
1.8 Torr by means of a vacuum system and a valve. The closed end of the
glass tube containing the compound was inserted to a depth of ꢀ20 cm
into an electrical oven maintained at 1808C. After 2 d, single purple crys-
tals had formed on the wall of the tube inside the oven, m.p. 193–1958C.
In addition, a small amount of yellow powder and some yellow crystals
had formed on the wall of the tube just outside the oven, m.p. 1458C
(purple melt).
X-ray structural analysis, general procedures: Suitable crystals were se-
lected under a microscope and attached to a glass fiber by means of sili-
cone glue or fast-drying acryl acetate glue. Unit-cell dimensions and dif-
fraction intensities were measured with monochromatic MoKa radiation,
l=0.71073 Š, on a four-cycle diffractometer. Intensities were measured
with w scans (unless stated otherwise) at variable scan speeds and were
corrected for Lorentz and polarization effects. Three check reflections
were measured every 97 reflections and were used to correct for instabili-
ties of the X-ray source and decay of the crystals. No absorption correc-
tion was applied. The structures were solved by direct methods with the
TREFF routine of the program XS.[86, 87] The structures were refined
Acknowledgements
P.U.B thanks the Minerva Foundation for a fellowship and the Deutscher
Akademischer Austauschdienst (DAAD) for
a Kurzstipendium. We
2
based on jFj by means of the full-matrix least-squares method of the
thank Mr. Thomas Hildenbrandt for the measurement of the low-temper-
ature X-ray diffraction data of 10. We thank Dr. Amalia Levy (The
Hebrew University of Jerusalem) for a sample of 12, and Dr. Michal Ra-
chel Suissa (Oslo University College) for a sample of 10.
program XL.[86] The heavy atoms were refined with anisotropic displace-
ment parameters. The hydrogen atoms were located in difference elec-
tron-density maps and were added to the refinement either with inde-
pendent positional parameters and isotropic displacement parameters or
as riding on the carbon atoms with isotropic displacement parameters set
to 1.2” that of the carbon atom (see below). CCDC-283156 (purple 11),
CCDC-283157 (yellow 11), CCDC-283158 (10), and CCDC-283159 (12)
contain the supplementary crystallographic data for this paper. These
data can be obtained free of charge from the Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data request/cif.
[1] J. H. Day, Chem. Rev. 1963, 63, 65–80.
[2] G. Kortüm, Ber. Bunsen-Ges. Phys. Chem. 1974, 78, 391–403.
[3] a) E. D. Bergmann, Prog. Org. Chem. 1955, 3, 81–171; b) H. Bouas-
Laurent, H. Dürr, Pure Appl. Chem. 2001, 73, 639–665.
[4] a) H. Meyer, Ber. Dtsch. Chem. Ges. B 1909, 42, 143–145; b) H.
Meyer, Monatsh. Chem. 1909, 30, 165–177.
[5] D. L. Fanselow, H. G. Drickamer, J. Chem. Phys. 1974, 61, 4567–
4574.
[6] Y. Hirshberg, Compt. Rend. 1950, 231, 903–904.
[7] G. Kortüm, W. Theilacker, V. Braun, Z. Physik. Chem. 1954, NF 2,
179–196.
Purple 9-(2,7-dimethyl-9H-fluoren-9-ylidene)-9H-xanthene (purple 11):
¯
C28H20O, Mr =372.44, triclinic, P1, deep purple, a=6.9232(6), b=
11.9501(8), c=12.3004(8) Š, a=83.314(5), b=76.700(6), g=82.116(6)8,
V=977.3(1) Š3, Z=2, 1calcd =1.266 gcmÀ3, T=293 K, 4364 total reflec-
tions, qmax =258, wÀ2q scans, SiemensP4 diffractometer, crystal dimen-
sions 0.45”0.25”0.15 mm. The aromatic hydrogen atoms were refined
with isotropic displacement parameters. The methyl hydrogen atoms
were found to be disordered with two orientations corresponding to a 608
[8] T. Bercovici, R. Korenstein, K. A. Muszkat, E. Fischer, Pure Appl.
Chem. 1970, 24, 531–565.
À
rotation about the C C bond and were refined as riding on the respective
[9] a) E. Fischer, Chem. Unserer Zeit 1975, 9, 85–95; b) E. Fischer, Rev.
Chem. Intermed. 1984, 5, 393–422; c) K. A. Muszkat in The Chemis-
try of Quinonoid Compounds, Vol. 2 (Eds.: S. Patai, Z. Rappoport),
Wiley, Chichester, 1988, pp. 203–224; d) W. H. Laarhoven in Photo-
chromism, Molecules and Systems (Eds.: H. Dürr, H. Bouas-Lau-
rent), Elsevier, Amsterdam, 1990, pp. 270–313.
[10] a) B. L. Feringa, W. F. Jager, B. de Lange, E. W. Meijer, J. Am.
Chem. Soc. 1991, 113, 5468–5470; b) B. de Lange, W. F. Jager, B. L.
Feringa, Mol. Cryst. Liq. Cryst. 1992, 217, 129–132; c) B. L. Feringa,
W. F. Jager, B. de Lange, Tetrahedron 1993, 49, 8267–8310; d) B. L.
Feringa, W. F. Jager, B. de Lange, J. Chem. Soc. Chem. Commun.
1993, 288–290.
carbon atom. The final residual factors are wR2=0.092, R1=0.041, and
GOF=1.035 for 327 parameters and 3440 unique reflections, 2302 reflec-
À3
tions with I<2s(I). Residual electron density: 0.11/À0.10 eŠ (max/
min).
Yellow 9-(2,7-dimethyl-9H-fluoren-9-ylidene)-9H-xanthene (yellow 11):
C28H18O, Mr =372.44, monoclinic, P21/n, yellow, a=13.678(1), b=
7.600(1), c=19.373(2) Š, b=94.281(8)8, V=2008.3(4) Š3, Z=4, 1calcd
=
1.232 gcmÀ3, T=293 K, 3846 total reflections measured, qmax =22.58, Sie-
mensP4 diffractometer, crystal dimensions 0.12”0.06”0.03 mm. Hydro-
gen atoms were refined as riding on the respective carbon atoms. The
final residual factors are wR2=0.098, R1=0.067, and GOF=1.004 for
267 parameters and 2627 unique reflections, 1153 reflections with I<
2s(I). Residual electron density: 0.14/À0.12 eŠÀ3 (max/min).
[11] B. L. Feringa, A. M. Schoevaars, W. F. Jager, B. de Lange, N. P. M.
Huck, Enantiomer 1996, 1, 325–335.
3352
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3345 – 3354

Polymorphism Versus Thermochromism
FULL PAPER
[12] O. Aleksiuk, S. E. Biali, J. Org. Chem. 1996, 61, 5670–5673.
[13] Y. C. Yip, X.-j. Wang, D. K. P. Ng, T. C. W. Mak, P. Chiang, T.-Y.
Luh, J. Org. Chem. 1990, 55, 1881–1889.
[14] A. M. Schoevaars, R. Hulst, B. L. Feringa, Tetrahedron Lett. 1994,
35, 9745–9748.
[15] W. I. Smid, A. M. Schoevaars, W. Kruizinga, N. Veldman, W. J. J.
Smeets, A. L. Spek, B. L. Feringa, Chem. Commun. 1996, 2265–
2266.
[16] M. Riklin, A. von Zelewsky, A. Bashall, M. McPartlin, A. Baysal,
J. A. Connor, J. D. Wallis, Helv. Chim. Acta 1999, 82, 1666–1680.
[17] M. Querol, H. Stoekli-Evans, P. Belser, Org. Lett. 2002, 4, 1067–
1070.
[18] a) M. Jørgensen, K. Lerstrup, P. Frederiksen, T. Bjørnholm, P.
Sommer-Larsen, K. Schaumburg, K. Brunfeldt, K. Bechgaard, J.
Org. Chem. 1993, 58, 2785–2790; b) J. Anders, H. J. Byrne, J. Pop-
lawski, S. Roth, P. Sommer-Larsen, T. Bjçrnholm, M. Jçrgenson, K.
Schaumburg, Synth. Met. 1993, 61, 177–180; c) J. Poplawski, J.
Anders, H. J. Byrne, S. Roth, P. Sommer-Larsen, S. Schaumburg,
Synth. Met. 1993, 56, 1711–1716.
[19] M. L. C. M. Oosterling, A. M. Schoevaars, H. J. Haitjema, B. L. Fer-
inga, Isr. J. Chem. 1996, 36, 341–348.
[20] a) B. L. Feringa, N. P. M. Huck, H. A. v. Doren, J. Am. Chem. Soc.
1995, 117, 9929–9930; b) N. P. M. Huck, W. F. Jager, B. de Lange,
B. L. Feringa, Science 1996, 273, 1686–1688.
[44] a) W. Theilacker, G. Kortüm, H. Elliehausen, Chem. Ber. 1956, 89,
2306–2318; b) G. Kortüm, W. Theilacker, G. Littmann, Naturwissen-
schaften 1957, 44, 114–115; c) R. Korenstein, K. A. Muszkat, S.
Sharafy-Ozeri, J. Am. Chem. Soc. 1973, 95, 6177–6181.
[45] R. B. Woodward, E. Wasserman, J. Am. Chem. Soc. 1959, 81, 5007.
[46] J. F. D. Mills, S. C. Nyburg, J. Chem. Soc. 1963, 308–321.
[47] a) A. Schçnberg, K. Junghans, Chem. Ber. 1965, 98, 2539–2544;
b) R. Lorenz, U. Wild, J. R. Huber, Photochem. Photobiol. 1969, 10,
233–242.
[48] a) G. Kortüm, W. Zoller, Chem. Ber. 1967, 100, 280–292; b) G.
Kortüm, P. Krieg, Chem. Ber. 1969, 102, 3033–3045.
[49] G. Kortüm, W. Zoller, Chem. Ber. 1970, 103, 2062–2076.
[50] G. Kortüm, M. Buck, Z. Elektrochem. 1956, 60, 53–58.
[51] E. Wasserman, J. Am. Chem. Soc. 1959, 81, 5006–5007.
[52] W. Theilacker, G. Kortüm, H. Elliehausen, Z. Naturforschung B
1954, 9, 167–168.
[53] W. Theilacker, G. Kortüm, H. Elliehausen, H. Wilski, Chem. Ber.
1956, 89, 1578–1592.
[54] G. Kortüm, K.-W. Koch, Chem. Ber. 1967, 100, 1515–1520.
[55] a) I. Agranat, M. Rabinovitz, H. R. Falle, G. R. Luckhurst, J. N.
Ockwell, J. Chem. Soc. B 1970, 294–298; b) L. S. Singer, I. C. Lewis,
T. Richterzhagen, G. Vincow, J. Phys. Chem. 1971, 75, 290–291.
[56] R. Korenstein, K. A. Muszkat, E. Fischer, J. Photochem. 1976, 5,
447–456.
[57] K. A. Muszkat, W. Schmidt, Helv. Chim. Acta 1971, 54, 1195–1207.
[58] E. D. Bergmann, H. J. E. Loewenthal, J. Chem. Soc. 1953, 2572–
2574.
[21] T. Bjoernholm, S. Munk, P. Frederiksen, M. Joergensen, Mol. Cryst.
Liq. Cryst. Sci. Technol. Sect. A 1993, 235, 83–88.
[22] A. M. Schoevaars, W. Kruizinaga, R. W. J. Zijlstra, N. Veldman,
A. L. Speck, B. L. Feringa, J. Org. Chem. 1997, 62, 4943–4948.
[23] a) N. Koumura, R. W. J. Zijlstra, R. A. v. Delden, N. Harada, B. L.
Feringa, Nature 1999, 401, 152–155; b) N. Koumura, E. M. Geertse-
ma, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2000, 122,
12005–12006; c) B. L. Feringa, Acc. Chem. Res. 2001, 34, 504–513;
d) N. Koumura, E. M. Geertsema, M. B. v. Gelder, A. Meetsma,
B. L. Feringa, J. Am. Chem. Soc. 2002, 124, 5037–5051.
[24] E. Bergmann, H. Corte, Ber. Dtsch. Chem. Ges. B 1933, 66, 39–43.
[25] W. Theilacker, G. Kortüm, G. Friedheim, Chem. Ber. 1950, 83, 508–
519.
[26] W. T. Grubb, G. B. Kistiakowsky, J. Am. Chem. Soc. 1950, 72, 419–
424.
[27] G. Kortüm, Angew. Chem. 1958, 70, 14–20.
[28] a) Y. Hirshberg, E. Fischer, J. Chem. Soc. 1953, 629–636; b) Y.
Tapuhi, O. Kalisky, I. Agranat, J. Org. Chem. 1979, 44, 1949–1952.
[29] E. Harnik, J. Chem. Phys. 1956, 24, 297–299.
[30] a) E. Wasserman, R. E. Davis, J. Chem. Phys. 1959, 30, 1367; b) G.
Kortüm, G. Bayer, Ber. Bunsen-Ges. Phys. Chem. 1963, 67, 24–28.
[31] R. Korenstein, K. A. Muszkat, E. Fischer, Helv. Chim. Acta 1970,
53, 2102–2109.
[59] a) A. Mustafa, M. E. E.-D. Sobhy, J. Am. Chem. Soc. 1955, 77,
5124–5126; b) A. Mustafa, W. Asker, M. E. E.-D. Sobhy, J. Org.
Chem. 1960, 25, 1519–1525.
[60] S. C. Nyburg, Acta Crystallogr. 1954, 7, 779–780.
[61] G. Shoham, S. Cohen, M. R. Suissa, I. Agranat in Molecular Struc-
ture: Chemical Reactivity and Biological Activity (Eds.: J. J. Stezow-
ski, J.-L. Huang, M.-C. Shao), Oxford University Press, Oxford,
1988, pp. 290–312.
[62] a) P. U. Biedermann, A. Levy, J. J. Stezowski, I. Agranat, Chirality
1995, 7, 199–205; b) P. U. Biedermann, A. Levy, M. R. Suissa, J. J.
Stezowski, I. Agranat, Enantiomer 1996, 1, 75–80.
[63] P. U. Biedermann, J. J. Stezowski, I. Agranat, Eur. J. Org. Chem.
2001, 15–34.
[64] P. U. Biedermann, J. J. Stezowski, I. Agranat in Advances in Theoret-
ically Interesting Molecules, Vol. 4 (Ed.: R. P. Thummel), JAI Press,
Stamford, Conneticut, 1998, pp. 245–322.
[65] J.-S. Lee, S. C. Nyburg, Acta Crystallogr. Sect. C 1985, 41, 560–567.
[66] a) E. Harnik, G. M. J. Schmidt, J. Chem. Soc. 1954, 3295–3302;
b) P. A. Apgar, E. Wasserman, 1978, quoted in ref. [61].
[67] K. S. Dichmann, S. C. Nyburg, F. H. Pickard, J. A. Potworowski,
Acta Crystallogr. B 1974, 30, 27–36.
[32] J. F. D. Mills, S. C. Nyburg, J. Chem. Soc. 1963, 927–935.
[33] a) R. Wizinger, Ber. Dtsch. Chem. Ges. B 1927, 60, 1377–1389; b) R.
Wizinger, Z. Angew. Chem. 1927, 40, 939–945.
[34] A. Schçnberg, O. Schütz, Ber. Dtsch. Chem. Ges. B 1928, 61, 478–
479.
[35] A. Schçnberg, A. Mustafa, M. E. E.-D. Sobhy, J. Am. Chem. Soc.
1953, 75, 3377–3378.
[36] M. R. Padova, Ann. Chim. Phys. 1910, 19, 386.
[68] A. Levy, P. U. Biedermann, I. Agranat, Org. Lett. 2000, 2, 1811–
1814.
[69] a) D. H. R. Barton, E. H. Smith, B. J. Willis, J. Chem. Soc. D 1970,
1226; b) D. H. R. Barton, F. S. Guziec, Jr., I. Shahak, J. Chem. Soc.
Perkin Trans. 1 1974, 1794–1799; c) D. H. R. Barton, Reason and
Imagination: Reflections on Research in Organic Chemistry: Selected
Papers of Derek H. R. Barton, Imperial College Press and World
Scientific, Singapore, 1996, p. 489–492.
[37] Y. Hirshberg, Bull. Res. Counc. Isr. 1951, 1, 123–125.
[38] a) W. Dilthey, J. Prakt. Chem. 1925, 109, 273–323; b) W. Dilthey, R.
Dinklage, Ber. Dtsch. Chem. Ges. B 1929, 62, 1834–1839.
[39] a) A. Schçnberg, W. Asker, Chem. Rev. 1945, 37, 1–14; b) A. Schçn-
berg, A. F. A. Ismail, W. Asker, J. Chem. Soc. 1946, 442–446.
[40] a) E. Bergmann, L. Engel, Z. Phys. Chem. Abt. B 1930, 8, 111–137;
b) E. D. Bergmann, Isomerism and Isomerization of Organic Com-
pounds, Interscience Publishers, New York, 1948, pp. 46–50.
[41] E. Müller, Fortschr. Chem. Forsch. 1949, 1, 325–416.
[42] W. G. Nielsen, G. K. Fraenkel, J. Chem. Phys. 1953, 21, 1619.
[43] a) Y. Hirshberg, E. Loewenthal, E. D. Bergmann, Bull. Res. Counc.
Isr. 1951, 1(1–2), 139–141; b) A. Schçnberg, A. Mustafa, W. Asker,
J. Am. Chem. Soc. 1954, 76, 4134–4136.
[70] A. Schçnberg, M. M. Sidky, J. Am. Chem. Soc. 1959, 81, 2259–2262.
[71] M. M. Sidky, M. R. Mahran, L. S. Boulos, J. Prakt. Chem. 1970, 312,
51–54.
[72] N. S. Mills, J. L. Malandra, E. E. Burns, A. Green, K. E. Unruh,
D. E. Kadlecek, J. A. Lowery, J. Org. Chem. 1997, 62, 9318–9322.
[73] A. Levy, P. U. Biedermann, S. Cohen, I. Agranat, J. Chem. Soc.
Perkin Trans. 2 2001, 2329–2341.
[74] M. R. Suissa, PhD Thesis, The Hebrew University of Jerusalem
(Israel), 1988 (in Hebrew).
[75] E. D. Bergmann, G. Berthier, Y. Hirshberg, E. Loewenthal, B. Pull-
man, A. Pullman, Bull. Soc. Chim. Fr. 1951, 669–681.
[76] R. Knorr, A. Schnegg, E. Lattke, E. Räpple, Chem. Ber. 1979, 112,
3490–3514.
Chem. Eur. J. 2006, 12, 3345 – 3354
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3353

I. Agranat et al.
[77] W. Burgert, M. Große, D. Rewicki, Chem. Ber. 1982, 115, 309–323.
[78] S. Scheibye, R. Shabana, S.-O. Lawesson, C. Rømmig, Tetrahedron
1982, 38, 993–1001.
[83] a) T. Suzuki, T. Fukushima, T. Miyashi, T. Tsuji, Angew. Chem.
1997, 109, 2607–2609; Angew. Chem. Int. Ed. Eng. 1997, 36, 2495–
2497; b) J. J. Stezowski, P. U. Biedermann, T. Hildenbrand, J. A.
Dorsch, C. J. Eckardt, I. Agranat, J. Chem. Soc. Chem. Commun.
1993, 213–215; c) A. Beck, R. Gompper, K. Polborn, H.-U. Wagner,
Angew. Chem. 1993, 105, 1424–1427; Angew. Chem. Int. Ed. Eng.
1993, 32, 1352–1354.
[84] J. Bernstein, Polymorphism in Molecular Crystals, Oxford University
Press, Oxford, 2002.
[85] H. Wieland, A. Roseeu, Justus Liebigs Ann. Chem. 1911, 381, 229–
233.
[79] a) M. Walczak, D. G. Stucky, J. Organomet. Chem. 1975, 97, 313–
323; b) N. A. Bailey, S. E. Hull, Acta Crystallogr. B 1978, 34, 3289–
3295; c) J. J. Stezowski, T. Hildenbrand, M. R. Suissa, I. Agranat,
Struct. Chem. 1990, 1, 123–126; d) M. Munakata, L. P. Wu, K. Sugi-
moto, T. Kuroda-Sowa, M. Maekawa, Y. Suenaga, N. Maeno, M.
Fujita, Inorg. Chem. 1999, 38, 5674–5680; e) E. Lee-Ruff, A. Grant,
D. V. Stynes, I. Vernik, Struct. Chem. 2000, 11, 245–248; f) S. Pogo-
din, S. Cohen, P. U. Biedermann, I. Agranat, Enantiomer 2002, 7,
261–269; g) E. Molins, C. Miravitlles, E. Espinosa, M. Ballester, J.
Org. Chem. 2002, 67, 7175–7178.
[86] Programs for Crystal Structure Solution and Refinement, Siemens
SHELXTL Version 5.03 for DOS, Siemens Analytical X-ray Instru-
ments, 1994.
[80] P. U. Biedermann, J. J. Stezowski, I. Agranat, Chem. Commun. 2001,
954–955.
[87] G. M. Sheldrick, Acta Crystallogr. A 1990, 46, 467–473.
[81] J. Fabian, R. Zahradnik, Angew. Chem. 1989, 101, 693–710; Angew.
Chem. Int. Ed. Engl. 1989, 28, 677–694.
Received: September 8, 2005
[82] Y. V. Zefirov, Crystallogr. Rep. 1997, 42, 111–116.
Published online: January 27, 2006
3354
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3345 – 3354