Negligible diradical character for the ultralong C-C bond in 1,1,2,2-tetraarylpyracene derivatives at room temperature

Article abstract of DOI:10.1016/j.tetlet.2009.03.202

There is evidence that the crystalline-state thermochromic behavior of dispiropyracene acr-1 with an ultralong C-C bond [1.791(3) ? at 413 K, 1.771(3) ? at 93 K] might be due to thermal generation of a bond-dissociated triplet diradical. The C₁

Full text of DOI:10.1016/j.tetlet.2009.03.202

Tetrahedron Letters 50 (2009) 3693–3697
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Negligible diradical character for the ultralong C–C bond in
1,1,2,2-tetraarylpyracene derivatives at room temperature
Takashi Takeda a, Hidetoshi Kawai a,b, Rainer Herges c, Eva Mucke c, Yoshitaka Sawai a,
Kei Murakoshi a, Kenshu Fujiwara a, Takanori Suzuki a,
*
a Department of Chemistry, Faculty of Science, Hokkaido University, Sapporo 060-0810, Japan
b PRESTO, Japan Science and Technology Agency, Saitama, Japan
c Insitut für Organische Chemie, Universität Kiel, Otto-Hanh-Platz 4, D-24098 Kiel, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
There is evidence that the crystalline-state thermochromic behavior of dispiropyracene acr-1 with an
ultralong C–C bond [1.791(3) Å at 413 K, 1.771(3) Å at 93 K] might be due to thermal generation of a
bond-dissociated triplet diradical. The C1–C2 bond lengths (d) in the newly prepared 1,1,2,2-tetraarylpy-
racenes 1a–e [1.717(4)–1.761(4) Å at 113/123/153 K] are also much larger than the standard value for
C(sp3)–C(sp3) (1.54 Å). The fact that there is no correlation between d and the radical-stabilizing param-
eter (rꢀ) provides evidence that there is no contribution from a crystallographic artifact caused by con-
tamination of the diradical in the crystal of 1 at room temperature or below. Further indication for the
covalent nature of the ultralong C–C bond in tetraphenylpyracene 1c was attained by Raman spectros-
copy, showing the far red-shifted C1–C2 stretching vibration (638 cmÀ1) and by the theoretical prediction
of very large vertical singlet–triplet energy gap (58.3 kcal molÀ1 at the UB3LYP/6-31G* level). However, at
elevated temperatures bond dissociation might occur forming a triplet diradical.
Received 15 January 2009
Revised 23 March 2009
Accepted 26 March 2009
Available online 2 April 2009
In memory of Professor Fumio Toda (1933–
2008)
Keywords:
Ultralong bond
Strained molecule
Steric repulsion
Thermochromism
Diradical
Ó 2009 Elsevier Ltd. All rights reserved.
X-ray analysis
The C–C bond length (d) is one of the most fundamental param-
eters in organic chemistry,1 and a d value of 1.54 Å is assumed as
the standard length of a C(sp3)–C(sp3) single bond. This explains
why the experimentally determined d can be used to assess the
tion is switched from an intermolecular to an intramolecular
process. Thus, the HPEs with an even longer C–C bond
(d > 1.70 Å) could be isolated [1.712(5)–1.734(5) Å for 3,8-diha-
lo-1,1,2,2-tetraphenylnaphthocyclobutenes by Toda’s group6
and 1.713(2) Å for a multi-benzannulated caged hydrocarbon
by one of us7].
p
-bond order in combination with other standard values [1.33 Å
for C(sp2)@C(sp2); 1.20 Å for C(sp)„C(sp)]. On the other hand,
there have been several reports in which the C–C single bond is
shortened or elongated independent of the bond order,2 and such
elongation is often accompanied by a strained molecular structure
and/or severe steric hindrance.3
Recently, we also reported8a,b that a pyracene-based HPE
derivative acr-1 with two spiro(10-methylacridan) units has
an ultralong C–C bond [d = 1.771(3) Å at 93 K], which is one
of the longest ever reported9 [the d values for other crystallo-
graphically independent molecules of acr-1 are 1.758(3),
1.712(2), and 1.707(2) Å.]. The above X-ray analysis was carried
out with special care, however, there still remains some possi-
Hexaphenylethane (HPE) is a good example of this phenom-
enon; its central C–C bond is elongated [1.67(3) Å]4 primarily
due to steric repulsion among six aryl substituents. According
to the proposed linear relationship between d and bond-dissoci-
ation energy (BDE),5 a smaller BDE is expected for a longer C–C
bond. In fact, some isolable HPE derivatives exist as equilibrated
mixtures containing bond-dissociated trityl radicals in solution.
On the other hand, ‘cross-clamped’ or ‘condensed’ derivatives
of HPE show higher thermodynamic stability because dimeriza-
bility that the concomitant presence of
a bond-dissociated
diradical might have led to the assignment of a larger d than
the intrinsic value.10,13 Our major concern in this Letter is to
shed light on the following points: (a) does a bond-dissociated
diradical exist or can it be generated in the crystal of HPEs with
an ultralong C–C bond; and (b) is the crystallographically deter-
mined very large d ( > 1.75 Å) be intrinsic or an artifact by the
concomitant presence of a bond-dissociated diradical (or dira-
dicaloid) in the crystal.
* Corresponding author. Tel./fax: +81 11 706 2714.
E-mail address: tak@sci.hokudai.ac.jp (T. Suzuki).
0040-4039/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.03.202

3694
T. Takeda et al. / Tetrahedron Letters 50 (2009) 3693–3697
Me
Me
Me
Me
N
N
N
N
(eclipsed)
acr-1
(skewed)
acr-2
Here we found the reversible thermochromic behavior14 of acr-
1 in the single-crystalline state, which might be related to the bond
dissociation of the ultralong bond forming the corresponding
diradical in the crystal, as shown in the first part of this Letter.
Next, we have conducted the X-ray analyses of acr-1 at several ele-
vated temperatures to see if generation/increased contribution of
the diradical at higher temperatures expands the experimentally
determined d value of acr-1. Since the marginal expansion of
0.020 Å over a temperature difference of 320° does not provide
decisive support for the diradical contribution, we have then
turned our attention to the studies on a new series of HPEs,
1,1,2,2-tetraarylpyracenes 1. The experimentally assigned d value
should change with the attachment of radical-stabilizing/non-sta-
bilizing substituents when the coexistence of a bond-dissociated
diradicals affects the apparent bond lengths determined by X-ray
analyses (Scheme 1). If the corresponding diradicals exist in equi-
librium with HPEs in the crystal,14 compounds with radical-stabi-
lizing substituents would exhibit a larger value of d than those
with non-stabilizing or radical-destabilizing substituents. The
low-temperature X-ray structures of the corresponding com-
pounds do not reveal such a relationship. The details of our results
will be shown in the last part of this Letter.
First, we noticed a color change of crystalline samples of acr-1
when they were heated in an ESR tube sealed with Ar. Similar ther-
mochromic behavior was observed by heating the single crystal-
line sample on a glass fiber (Fig. 1), where temperature control
was carried out by a N2 gas flow. At 273 K, the sample is pale yel-
low, whose color gradually turned to dark red (413 K) via orange
(333 K). This observation is not accompanied by decomposition
of acr-1 because the crystal returned to pale yellow upon cooling
to room temperature or 123 K, which still gave X-ray diffraction
data identical to those obtained before the heat treatment. A
reversible phase transition or bond fission to generate a certain
population of diradicals upon heating might be responsible for
the observed thermochromic behavior. Thus, the X-ray analysis
of acr-1 was carried out at several temperatures above
273 K.15,16 No phase transition occurs in the range between
273 K and 413 K, thus the observed thermochromic behavior is
not related to a phase transition.
Figure 1. Single-crystalline thermochromic behavior of acr-1.
most atoms. It is noteworthy that those for C1 and C2 are still small
and round shaped (Fig. 2). The difference electron density map
indicates the presence of bonding electron between C1 and C2,
(Fig. S2) as in the case of the X-ray analysis at 93 K.8a The d value
for the bond in question is 1.791(3) Å, which is only 0.020 Å longer
than that measured at 93 K.17 The X-ray analyses at 333 K and
373 K gave the results with the intermediary structures with the
d value of 1.776(3) Å and 1.786(3) Å, respectively. As shown by To-
da’s and Siegel’s groups, 3,8-dichloro-1,1,2,2-tetraphenylnaphtho-
cyclobutene exhibits
a similar degree of bond expansion
[1.710(2) Å at 90 K,6e 1.720(4) at 298 K6a].
The vertical singlet-triplet energy gap (
D
ES-T) is a frequently
used indicator to evaluate the diradical character of molecules.18
We performed calculations at the UB3LYP/6-31G -level of DFT
*
based on the X-ray geometry of acr-1 at 413 K (fixed d of
1.7922 Å) and obtained the
D
ES-T value of 58.99 kcal molÀ1. A quite
similar value (59.04 kcal molÀ1) was obtained for the crystallo-
graphically independent molecule with the shortest C1–C2 bond
measured at 93 K (d = 1.7062 Å). The very large
DES-T values indi-
cate that the diradical contribution to the overall wavefunction
of acr-1 is negligible at lower temperatures (compare with the
D
ES-T of o-, m-, and p-benzyne: 37.5, 21.0 3.8 kcal/mol).19 On the
other hand, we found a triplet minimum structure with a very
large bond distance d of 3.110 Å which energetically is only
8.14 kcal molÀ1 above the singlet structure (UB3LYP/6-31G ,
*
Fig. S3). This species could explain the observed thermochromism
on a theoretical basis. According to TD DFT calculations the longest
wavelength absorption of the singlet state is 362 nm which is in
agreement with the fact that the crystals are almost colorless at
low temperatures. By contrast the triplet state exhibits several
The X-ray structure at 413 K is quite similar to those obtained at
lower temperatures except for the enlarged thermal ellipsoids at
closed-shell species
with an ultralong bond
bond-dissociated diradical
(diradicaliod)
X
X
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
average
?
Ar
X = radical stabilizing / non-stabilizing substituent
observed as longer bond?
Scheme 1.

T. Takeda et al. / Tetrahedron Letters 50 (2009) 3693–3697
3695
We have prepared here the parent tetraphenylpyracene (1c)22
from 5,6-dibromoacenaphthene and benzophenone in three steps
via pyran 3c, as shown in Scheme 2. By starting with the substi-
tuted benzophenones, HPEs with benzyl radical-stabilizing substit-
uents on phenyl rings such as 4-tBu (1a), and 4-Me (1b) could be
similarly prepared along with those with non-stabilizing [3,5-
Me2 (1d)] and destabilizing [4-F (1e)] substituents. The desired
HPEs (1) were isolated as surprisingly stable colorless/pale yellow
solids despite the very large d value for the C1–C2 bond (vide infra).
Another series of HPEs 2 with an acenaphthene p-framework was
also designed to further support the conclusion obtained from 1.
The structurally related acenaphthene derivative acr-28a,8c has a
shorter bond length [1.696(3) Å] than acr-1 because it adopts a
skewed arylene framework to relieve steric repulsion between
two acridan units. Yet, the skeleton might be still suitable enough
to test the validity of our approach adopted here. So, 1,1,2,2-tetra-
arylacenaphthenes 2a(4-tBu),b(4-Me),d(3,5-Me2),e(4-F) were pre-
pared over three steps from 1,8-diiodonaphthalene and the
corresponding benzophenones via pyrans 4 (Scheme 3).22 The
preparation and X-ray structure of tetraphenylacenaphthene (2c)
were reported previously by Gabbaï and coworker.23
Figure 2. ORTEP drawings of acr-1 (top view and side view) at 413 K. Note the
small and round-shaped ellipsoids for C1 and C2.
absorptions within the visible range (strongest bands at 450 and
475 nm; Table S1) which explain the red color of the crystals at
higher temperatures at which the triplet state will be increasingly
populated. Thus, it is clear that the ultralong C–C bond in acr-1 at
room temperature and below is intrinsic and that the concomitant
presence of bond-dissociated triplet diradical becomes important
only at higher temperatures.
To gain further insight into the nature of the ultralong bonds we
set out to crystallographically measure the bond length d in HPEs
other than acr-1. For this purpose, we designed 1,1,2,2-tetraaryl-
pyracene derivatives 1a–e as suitable HPE derivatives from the fol-
lowing viewpoints: first, the pyracene skeleton should act as a
scaffold to induce enough ‘front strain’4 to ensure the presence of
an extremely long C1–C2 bond (d > 1.70 Å), as shown by acr-1; sec-
ond, the system should be designed in such a way that the bond-
dissociated species is a diradical and not (as, e.g., in the case of
the benzocyclobutenes-type HPEs) a closed shell structure (o-
quinodimethane),20 and third, a series of compounds with various
substituents on aryl groups should be accessible via the corre-
sponding dicationic precursors.21
After considerable effort, we finally obtained a single-crystal
specimen of high quality for all nine kinds of HPEs 1a–e/
2a,b,d,e.16 Low-temperature X-ray analyses revealed that all of
them have an extremely long C1–C2 bond [1.717(4)–1.761(4) Å
for 1 and 1.694(3)–1.708(4) Å for 2] (Tables 1 and 2).24 No struc-
tural disorder around the C1–C2 bond was found in any case. As
represented by 4-Me derivative 1b [d = 1.717(4) Å: the smallest
among 1a–e] and 4-F derivative 1e [d = 1.761(4) Å: the largest]
(Fig. 3), the ORTEP drawings show no anomalies in thermal ellip-
soids of the long C1–C2 bond or in electron density maps. The rad-
ical-stabilizing parameters (rꢀ 25
are also summarized in Table 1,
)
which quantify the thermodynamic stability of the benzyl radicals
attached with a various substituent on the benzene nucleus. As
shown by these data, HPEs with radical-stabilizing groups (1a,b/
2a,b; rꢀ > 0) do not necessarily have larger d values than the par-
ent compound (1c/2c) or those with non-stabilizing/destabilizing
groups (1d,e/2d,e: rꢀ < 0). Thus, the radical-stabilizing substitu-
ents do not increase the apparent bond length determined by
X-ray, which leads us to conclude that the contribution from the
bond-dissociated diradical is negligible for the present HPEs 1/2
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
O
Ar
Ar
Br Br
C1 C2
a)
b)
c)
3
1
a : Ar = 4-tBuC6H4; b : Ar = 4-MeC6H4; c : Ar = C6H5; d : Ar = 3,5-Me2C6H3; e : Ar = 4-FC6H4
Scheme 2. Preparation of tetraarylpyracenes 1a–e. Reagents: (a) (1) n-BuLi, then Ar2CO, ether; (2) trifluoroacetic acid, CH2Cl2; (b) TMSClO4, 1,1,1,3,3,3-hexafluoro-2-
propanol; (c) Zn, MeCN.
Ar
Ar
Ar
Ar
Ar
Ar
Ar
Ar
C1 C2
Ar
O
Ar
Ar
Ar
I
I
a)
b)
c)
4
2
a : Ar = 4-tBuC6H4; b : Ar = 4-MeC6H4; c : Ar = C6H5; d : Ar = 3,5-Me2C6H3; e : Ar = 4-FC6H4
Scheme 3. Preparation of tetraarylacenaphthenes 2a,b,d,e. Reagents: (a) (1) n-BuLi, then Ar2CO, ether; (2) trifluoroacetic acid, CH2Cl2; (b) TMSClO4, 1,1,1,3,3,3-hexafluoro-2-
propanol for 4a,b,d; TfOH, 1,1,1,3,3,3-hexafluoro-2-propanol for 4e; (c) Zn, MeCN.

3696
T. Takeda et al. / Tetrahedron Letters 50 (2009) 3693–3697
Table 1
Structural parameters of pyracene-type HPEs 1 determined by low-temperature X-ray
analyses and the radical-stabilizing parameters rꢀ
Compound
1a
1b
1c
1d
1e
X
4-tBu
1.749(4)
0.036
4-Me
1.717(4)
0.015
H
3,5-Me2
1.743(2)
À0.002
4-F
1.761(4)
À0.011
d (Å)
1.754(2)
0.000
rꢀ
Table 2
Structural parameters of acenaphthene-type HPEs 2 determined by low-temperature
X-ray analyses
Compound
2a
2b
2ca
2d
2eb
X
d (Å)
4-tBu
1.708(4)
4-Me
1.694(3)
H
3,5-Me2
1.701(2)
4-F
1.701(3)
1.707(2) 1.705(2)
a
Ref. 23.
b
Two crystallographically independent molecules.
Figure 4. Experimental (a) and calculated (b) Raman spectra of 1c.
not be assigned to the closed shell structure. Therefore, at room
temperature or below a diradical structure can either be com-
pletely excluded or, if there is an equilibrium, it must be present
in very little concentration in the crystalline sample of 1c.26 Raman
investigations at higher temperatures are in progress.
In conclusion, we have no reason to assume that the extremely
long (d > 1.70 Å) or ultralong (d > 1.75 Å) C–C bond in the present
HPEs gives rise to a noticeable diradical character of the corre-
sponding compounds. However, there is indication that there is a
triplet diradical in a thermal equilibrium with the closed shell
structure at higher temperatures in crystalline acr-1. Despite the
large expansion that may lower the BDE, the present HPEs exhibit
surprising chemical stability: For example, the parent 1c
[d = 1.754(2) Å] can be purified by chromatography using aerated
solvents and can be stored infinitely as solids. It remains intact
upon treatment with Et3SiH or Bu3SnH. Based on our results, stud-
ies to achieve further elongated C–C bonds are warranted with a d
value even longer than the shortest nonbonded C–C contact
(1.80 Å).27
Figure 3. ORTEP drawings of 1b (left) and 1e (right) at 123 K.
in the crystal state even for derivatives with an ultralong C–C bond
with a d value as large as 1.76 Å.
For further validation of the crystallographically determined ex-
tremely long C–C bond, Raman spectroscopy was applied on the
crystalline samples of tetraphenyl derivatives (1c/2c). The stretch-
ing vibration of the long C1–C2 bond studied by Raman spectros-
copy at 298 K gives direct information on the force constant of
the bond. The absorption in pyracene derivative 1c with the high-
est intensity (expt. 638 cmÀ1) in Figure 4a corresponds to the nor-
mal mode with the largest amplitude along the trajectory of the
stretching vibration as verified by DFT calculation (B3LYP/6-
Acknowledgments
This work was partly supported by the Global COE Program
(Project No. B01: Catalysis as the Basis for Innovation in Materials
Science) from MEXT, Japan. T.T. thanks JSPS research fellowship for
young scientists (18-4425).
31G : calcd 650 cmÀ1). There are other absorptions assigned as
*
normal modes that include C1–C2 stretching vibration (470, 621,
718, 812 cmÀ1), for which the calculation well reproduced the fre-
quencies (478, 637, 720, 825 cmÀ1) (Fig. 4b). The observed red-
shifted value compared to the C–C stretching vibration of ethane
itself (995 cmÀ1) clearly shows that the ultralong C–C bond
[dexpt = 1.754(2) Å; dcalcd = 1.759 Å] in 1c has a very small force
constant.
Not only the frequencies of the C1–C2 stretching bands but also
the other experimental spectral lines were well reproduced by the
calculation. In contrast to the X-ray analysis and NMR, Raman is a
very fast spectroscopic method. In a hypothetical equilibrium of a
closed shell species with a large d value and a diradical, both spe-
cies should appear with separate absorption bands and not as a
time-averaged spectrum like in NMR. The agreement of the exper-
imental and theoretical Raman spectrum is very good, and there
are no additional lines in the experimental spectrum, which could
Supplementary data
Crystallographic data (excluding structure factors) for the struc-
tures in this Letter has been deposited with the Cambridge Crystal-
lographic Data Centre as supplementary publication numbers
CCDC 704958–704964, 715129, 715130, 715701–715703). Copies
of the data can be obtained, free of charge, on application to CCDC,
12 Union Road, Cambridge CB2 1EZ, UK [fax: +44 (0)-1223-336033
or e-mail: deposit@ccdc.cam.ac.uk]. Spectral data of new HPEs
(1a–e/2a,b,d,e), ORTEP drawings of 1a,c,d/2a,b,d,e (Fig. S1), the D
map of acr-1 at 413 K (Fig. S2), plots of singlet (d = 1.776 Å) and
triplet (d = 3.110 Å) species obtained by DT-DFT calculation on
acr-1 (Fig. S3), and calculated absorption bands of both singlet
and triplet species of acr-1 (Table S1). Supplementary data associ-
ated with this article can be found, in the online version, at
doi:10.1016/j.tetlet.2009.03.202.

T. Takeda et al. / Tetrahedron Letters 50 (2009) 3693–3697
3697
14. Morita, Y.; Suzuki, S.; Fukui, K.; Nakazawa, S.; Kitagawa, H.; Kishida, H.;
Okamoto, H.; Naito, A.; Sekine, A.; Ohashi, Y.; Shiro, M.; Sasaki, K.; Shiomi, D.;
Sato, K.; Takui, T.; Nakasuji, K. Nat. Mater. 2008, 7, 48–51.
15. Variable temperature X-ray analysis of acr-1 was previously carried out for the
temperature range between 93 K and 233 K. (Ref. 8b). Those measurements
revealed that the phase transition actually occurred at 233 K, whereby four
crystallographically independent molecules at lower temperatures are merged
into two crystallographically independent molecules. No further phase
transition occurred at higher temperature up to 413 K.
16. Crystal data and details on crystallographic studies of 1a–e/2a,b,d,e (at 113/
123/153 K) were given in Supplementary data.
17. In another crystallographically independent molecule, the C1–C2 bond length is
1.719(3) Å; whose expansion degree (0.012 Å) over 320° is similar to the one
with the ultralong C–C bond.
18. Li, X.; Paldus, J. J. Chem. Phys. 2008, 129, 174101.
19. Wenthold, P. G.; Squires, R. R.; Lineberger, W. C. J. Am. Chem. Soc. 1998, 120,
5279–5290.
20. Cycloreversion of 1,1,2,2-tetraphenylbenzocyclobutene to 7,7,8,8,tetraphenyl-
o-quinodimethane at low temperature shows that the conversion to the
thermodynamically stable bond-dissociated valence isomer is a very rapid
process: Quinkert, G.; Wiesdorff, W.-W.; Finke, M.; Opitz, K.; von der Haar, F.-G.
Chem. Ber. 1968, 101, 2302–2325; Tetraaryl-9,10-phenanthraquinodimathane
References and notes
1. Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen, A. G.; Taylor, R. J.
Chem. Soc., Perkin Trans 2 1987, S1.
2. (a) Kaupp, G.; Boy, J. Angew. Chem., Int. Ed. 1997, 36, 48–49; (b) Siegel, J. S.
Nature 2006, 439, 801–802; (c) Tanaka, M.; Sekiguchi, A. Angew. Chem., Int. Ed.
2005, 44, 5821–5823; (d) Huntley, D. R.; Markopoulos, G.; Donovan, P. M.;
Scott, L. T.; Hoffmann, R. Angew. Chem., Int. Ed. 2005, 44, 7549–7553.
3. (a) Suzuki, T.; Takeda, T.; Kawai, H.; Fujiwara, K. Pure Appl. Chem. 2008, 80, 547–
553; (b) Suzuki, T.; Takeda, T.; Kawai, H.; Fujiwara, K. In Strained Hydrocarbon:
Ultralong C–C Bond; Dodziuk, H., Ed.; Wiley-VCH, 2009, pp 70–82.
4. Kahr, B.; Engen, D. V.; Mislow, K. J. Am. Chem. Soc. 1986, 108, 8305–8307.
5. Zavitsas, A. A. J. Phys. Chem. A 2003, 107, 897–898.
6. (a) Toda, F.; Tanaka, K.; Stein, Z.; Goldberg, I. Acta Crystallogr., Sect. C 1996, 52,
177–180; (b) Toda, F.; Tanaka, K.; Watanabe, M.; Tamura, K.; Miyahara, I.;
Nakai, T.; Hirotsu, K. J. Org. Chem. 1999, 64, 3102–3105; (c) Tanaka, K.;
Takamoto, N.; Tezuka, Y.; Kato, M.; Toda, F. Tetrahedron 2001, 57, 3761–3767;
(d) Toda, F.; Tanaka, K.; Takamoto, N. Tetrahedron Lett. 2001, 42, 7979–7982; (e)
Baldridge, K. K.; Kasahara, Y.; Ogawa, K.; Siegel, J. S.; Tanaka, K.; Toda, F. J. Am.
Chem. Soc. 1998, 120, 6167–6168.
7. Kammermeier, S.; Jones, P. G.; Herges, R. Angew. Chem., Int. Ed. 1997, 36, 1757–
1760.
was isolated as
a stable compound which no longer isomerises into the
8. (a) Kawai, H.; Takeda, T.; Fujiwara, K.; Wakeshima, M.; Hinatsu, Y.; Suzuki, T.
Chem. Eur. J. 2008, 14, 5780–5793; (b) Kawai, H.; Takeda, T.; Fujiwara, K.; Inabe,
T.; Suzuki, T. Cryst. Growth Des. 2005, 5, 2256–2260; (c) Kawai, H.; Takeda, T.;
Fujiwara, K.; Suzuki, T. Tetrahedron Lett. 2004, 45, 8289–8293.
9. The bond of 1.781(15) Å in a silabicyclobutane derivative had been the longest
among the non-ionic organic molecules before this Letter: Fritz, G.;
Wartanessian, S.; Matern, E.; Hönle, W.; Schnering, H. G. v. Z. Anorg. Allg.
Chem. 1981, 475, 87–108.
10. In some earlier examples, the coexistence of a bond-dissociated valence isomer
caused the misassignment of a too-large d value (Ref. 11), since X-ray structure
only gives the time- and position-averaged geometry of molecules in the
crystal. The true value of those cases was later proven to be much smaller by
low-temperature data collection and the careful examination of thermal
ellipsoids (Ref. 12).
corresponding HPE derivative (tetraarylphenanthracyclobutene): Iwashita, S.;
Ohta, E.; Higuchi, H.; Kawai, H.; Fujiwara, K.; Ono, K.; Takenaka, M.; Suzuki, T.
Chem. Commun. 2004, 2076–2077.
21. (a) Suzuki, T.; Nishida, J.; Tsuji, T. Angew. Chem., Int. Ed. 1997, 36, 1329–1331;
(b) Suzuki, T.; Ohta, E.; Kawai, H.; Fujiwara, K.; Fukushima, T. Synlett 2007,
851–869.
22. Spectral data of new compounds 1a–e/2a,b,d,e were given in Supplementary
data.
23. Wang, H.; Gabbaï, F. P. Org. Lett. 2005, 7, 283–285.
24. Some variation in d was observed in the HPE series of both 1 and 2. The larger d
value for the parent (1c) than the 3,5-Me2 derivative (1d) clearly shows that
the steric bulkiness of substituents on the aryl groups have nothing to do with
the C1–C2 bond. We are currently examining the electronic effects of the
substituents affecting the C1–C2 bond length (d) through the high-level
theoretical calculations, whose results will be reported in due course.
25. Dust, J. M.; Arnold, D. R. J. Am. Chem. Soc. 1983, 105, 1221–1227.
26. This is also the case for acenaphthene derivative 2c, which has a shorter C1–C2
bond [d = 1.701(3) Å (Ref. 23)]. Its Raman spectrum measured on the
crystalline sample has an intense band at 655 cmÀ1 (calcd 665 cmÀ1 by
11. Ehrenberg, M. Acta Crystallogr. 1966, 20, 182–186.
12. (a) Battersby, T. R.; Gantzel, P.; Baldridge, K. K.; Siegel, J. S. Tetrahedron Lett.
1995, 36, 845–848; (b) Harada, J.; Ogawa, K.; Tomoda, S. Chem. Lett. 1995, 24,
751–752.
13. One of the reviewers commented that more than 10% of diradical content
is necessary to observe the apparent bond expansion. Another reviewer, in
contrast, assumes that only few % of diradical contamination is enough to
produce the crystallographic artifact. The discrepancy of these experts’
claims suggests that there has been no established discussion on this
matter, for which this Letter can provide the important experimental
results.
*
B3LYP/6-31G ), which is less red shifted than in 1c. This is clear evidence that
1c has a bond with a smaller force constant than 2c, which is consistent with
the crystallographic data shown above.
27. Adcock, J. L.; Gakh, A. A.; Pollitte, J. L.; Woods, C. J. Am. Chem. Soc. 1992, 114,
3980–3981.