Intramolecular methylacridan-methylacridinium complexes with a phenanthrene-4,5-diyl or related skeleton: Geometry-property relationships in isolable C-H bridged carbocations

Article abstract of DOI:10.1002/chem.200801769

The isolation and low-temperature X-ray analyses of a series of intramolecular methylacridan-methylacridinium complexes have been achieved. The two chromophores are in close proximity due to an arylene spacer, such as a phenanthrene-4,5-diyl or biphenyl-2

Full text of DOI:10.1002/chem.200801769

DOI: 10.1002/chem.200801769
Intramolecular Methylacridan–Methylacridinium Complexes with a
Phenanthrene-4,5-diyl or Related Skeleton: Geometry–Property
À
Relationships in Isolable C H Bridged Carbocations
Takanori Suzuki,*[a] Yasuyo Yoshimoto,[a] Takashi Takeda,[a] Hidetoshi Kawai,[a, b] and
Kenshu Fujiwara[a]
Abstract: The isolation and low-tem-
perature X-ray analyses of a series of
intramolecular methylacridan–methyl-
acridinium complexes have been ach-
ieved. The two chromophores are in
close proximity due to an arylene
spacer, such as a phenanthrene-4,5-diyl
or biphenyl-2,2’-diyl unit. These bridg-
ized structure both in solution and in
the solid state. The bridging hydrogen
atom undergoes a facile intramolecular
hydride shift from one carbon atom to
another in solution, and the energy bar-
rier is linearly correlated with the intra-
molecular C···C+ distance in the solid-
state geometry, as determined by
single-crystal X-ray analyses. By ex-
trapolation from the data, the delocal-
Keywords: carbocations · hydride
shift · strained molecules · structure
elucidation · three-center bonds
+
À À
ized three-center bond of [C H C]
would be formed when the C···C+ dis-
tance is less than 2.7 ꢀ.
À
ed carbocations prefer the C H local-
Introduction
array,[6] which is defined by the C···H···C+ angle (q), the
C···C+ distance (D), and two kinds of C···H distances (d1
and d2; Figure 1). However, experimental verification of the
geometry–delocalizability relationship has been hampered
by the instability of the [C···H···C]+ unit, which prevents iso-
lation for X-ray structure analyses.
Multicentered bonds are of interest due to their unique
nature and special properties. Representative examples in-
clude the three-center two-electron (3c–2e) bond of [B H
B], which is ubiquitously observed in uncoordinated bor-
anes.[1] 3c bonds in organic chemistry are much less
common,[2–4] despite the pioneering work by Sorensen,
À À
McMurry and their respective co-workers, who observed the
+
À À
delocalized 3c–2e bond of [C H C] and/or the unsymmet-
+
À
À
Figure 1. Geometrical parameters for [C···H···C]+ bridged carbocations.
ric C H bridged carbocation of [C H···C ] in caged hydro-
carbons in strong acidic media.[5] The delocalizability of the
bridging hydrogen atom in the [C···H···C]+ unit has been
suggested to be sensitive to the geometry of the three-atom
In this context, our recent reports[7] on the intramolecular
+
À
triarylmethane–triarylmethylium complexes [Ar3C H···C
Ar3] are interesting: diaryl(8-diarylmethyl-1-naphthyl)me-
[a] Prof. T. Suzuki, Y. Yoshimoto, Dr. T. Takeda, Dr. H. Kawai,
Prof. K. Fujiwara
À
thyliums are the first isolable C H bridged carbocations for
which the geometries have been successfully analyzed by
crystallography. The central point of the molecular-design
concept is the significant peri interaction of the naphtha-
Department of Chemistry, Faculty of Science
Hokkaido University
Kita 10, Nishi 8, Kita-ku, Sapporo, 060-0810 (Japan)
Fax : (+81)11-706-2714
lene-1,8-diyl-type skeleton, which forces the two chromo-
phores into proximity[8] to facilitate [C H···C ] bridging, de-
+
E-mail: tak@sci.hokudai.ac.jp
À
[b] Dr. H. Kawai
spite the sterically hindered triarylmethyl structure
(Scheme 1). Based on an examination of several derivatives
with the Ar2C+ unit, such as Ph2C+ and N-methylacridini-
um, it was shown that the thermodynamic stability of the
PRESTO, Japan Science and Technology Agency (Japan)
Supporting information for this article (experimental details of new-
compound preparation, ORTEP drawings of 1c–f
redox behavior, and UV/Vis spectra of 1[OTf]) is available on the
WWW under http://dx.doi.org/10.1002/chem.200801769.
ACHTUNGTREN[NGNU OTf], 2e, and 3e,
AHCTUNGTRENNUNG
cationic part (KR+(methylacridinium)/KR+A
CHTUNGTRENNUNG
(Ph3C+)ꢀ1017)[9]
2210
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2210 – 2216

FULL PAPER
spacer to adopt a more twisted conformation than that in
1d+. Methylacridan–methylacridinium complex 1 f+, with a
diphenyl ether-2,2’-diyl unit, was also investigated; the large
flexibility of the spacer in this case might induce a prefer-
ence for a geometry other than the bridging one.
Herein, we report the generation and isolation of the title
carbocations and their detailed structural features deter-
mined by low-temperature X-ray analyses. The bridged car-
À
bocations prefer the C H localized structure, both in solu-
Scheme 1. Diaryl(8-diarylmethyl-1-naphthyl)carbeniums, the isolable in-
tramolecular triarylmethane–triarylmethylium complexes studied previ-
ously.[7]
tion and in the solid state. The fluctuation behavior of the
bridging hydrogen atom was studied by variable-tempera-
ture (VT) NMR spectroscopy, and a linear correlation was
established between the energy barrier for the H-shift in so-
lution and the C···C+ distance (D) determined by crystallog-
raphy.
does not perturb the dynamic properties of the bridging hy-
drogen atom. By contrast, the [C···H···C]+ bridging distance
was suggested to be a more decisive factor because the fluc-
tuation speed of the bridging hydrogen atom in the intramo-
lecular methylacridan–methylacridinium complexes 1A–C+
was altered just by replacing the acenaphthene-5,6-diyl
(1B+: D=3.00 ꢀ (X-ray data); DG° for hydride (H) shift=
9.6 kcalmolÀ1 (in CD2Cl2)) or acenaphthylene-5,6-diyl (1C+:
D=3.03 ꢀ; DG° =10.1 kcalmolÀ1) skeleton for a naphtha-
lene-1,8-diyl unit (1A+: D=2.95 ꢀ; DG° <8 kcalmolÀ1). To
further assess the detailed relationship between the proper-
ties and the experimentally determined geometries, we plan-
ned to investigate newly designed methylacridan–methyla-
cridinium complexes 1a–f+ with a variety of arylene spacers
(Schemes 2 and 3), which should give a wide variety of geo-
Results and Discussion
+
À
Generation and isolation of [C H···C ] bridged carboca-
tions: We previously found[7] that the naphthalene-1,8-diyl-
+
+
À
type carbocations (1A–C ) with a [C H···C ] unit were
generated by quaternization of the neutral precursors 2 with
a methylacridan–acridine unit. Precursors 2, in turn, were
obtained from arylenedi(9-acridine)s 3 through mono-N-
methylation with CH3OTf followed by hydride addition with
NaBH4. Thus, we postulated that similar procedures would
also work for the new series of carbocations 1a–f+
(Scheme 3). In fact, the less-hindered derivatives 1b–f+
were successfully obtained in this way.
Arylenedi(9-acridine)s 3c–f[11,12] were readily obtained by
the reactions of N-MEM-9-acridone with the corresponding
dilithioarenes generated in situ by halogen–lithium exchange
reactions of 1,11-dibromo-5,7-dihydrodibenzACHTNUTRGNE[NUG c,e]oxepine
Scheme 2. Methylacridan–methylacridinium complexes 1a–e+, with a va-
riety of arylene spacers, studied herein. Ms: methanesulfonyl.
metrical parameters for the [C···H···C]+ bridging unit, such
as different bridging distances (D, d1, and d2). Furthermore,
by incorporating an arylene spacer with a flexible geometry
in the complexes, we could confirm whether [C···H···C]+
bridging provides net stabilization of carbocations 1+ that is
more or less strong than that through p–p interactions.
The phenanthrene-4,5-diyl skeleton in 1a+ is a rigid ary-
lene unit, which was used in a study of 3c–4e bonding to
identify a superior “proton-sponge” analogue.[10] On the
other hand, the [C···H···C]+ bridging distance in 1d+ with
the biphenyl-2,2’-diyl unit must be larger than that in 1a+,
because 1d+ can adopt a skewed conformation by rotation
around the biaryl axis. If the arylene unit is replaced by
9,10-dihydrophenanthrene-4,5-diyl
(1b+)
or
5,7-
dihydrodibenz
ACHTUNGTRENNUNG
deformation is partly suppressed, whereas 6,6’-dibromo sub-
stitution of the biphenyl unit in 1e+ forces the arylene
Scheme 3. Synthetic route to methylacridan–methylacridinium complexes
1b–f[OTf]. MEM: methoxyethoxymethyl; Tf: trifluoromethanesulfonyl.
AHCTUNGTRENNUNG
Chem. Eur. J. 2009, 15, 2210 – 2216
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2211

T. Suzuki et al.
(4c),[13] 2,2’-diiodobiphenyl (4d),[14] and 2,2’,6,6’-tetrabromo-
biphenyl (4e)[15] or by direct deprotonation in the case of di-
phenyl ether 4 f[16] (Scheme 3). Diacridines 3a and 3b were
derived over several steps by reductive ring closure of 6,6’-
diformylbiphenyl-2,2’-diyldi(9-acridine) (7), which was pre-
pared from 2,2’,6,6’-tetrabromobiphenyl (4e), as shown in
Scheme 4. Diacridines 3a–f thus obtained were smoothly
converted into precursors 2a–f. The less-hindered methyl-
acridan–methylacridinium complexes 1b–f+ were successfully
generated by N-methylation of 2b–f as planned (Scheme 3)
and were isolated as stable OTfÀ salts.
recrystallization of 1 fACHTNUGETRN[UNG OTf]. The diffraction data were col-
lected at low temperature so the structural parameters of
carbocations 1a,c–f+ were determined with sufficient accu-
racy. The bridging hydrogen atom was found on the D-
maps. No positional disorders were observed around the
[C···H···C]+ bridge in any case.
Regardless of the differences in the structures of the ary-
lene spacers, one of the chromophores in 1a,c–f+ is a planar
methylacridinium with an sp2-hybridized C9 atom (sum of
À À
C C C bond angles=359.2–360.08), whereas the other is a
butterfly-shaped methylacridan unit with an sp3-hybridized
C9 atom (334.8–339.18; Table 1). The preferred geometry of
Table 1. Geometrical parameters[a] in methylacridan–methylacridinium
complexes 1a,c–f+, as determined by low-temperature X-ray analyses of
OTfÀ salts, and those of related compounds (2e, 3e). The optimized ge-
ometry of 1a+, as calculated by DFT methods,[b] is also shown.
+
À
À
C
À
À
D
d1
d2
q
f
ꢀC C
ꢀC CH C
1a+
3.14
2.51 1.08 116 29.7 359.2
334.8
336.2
337.8
339.1
335.9
336.1
337.1
336.1
336.0
–
1a+ (calcd) 3.187 2.47 1.09 121 30.5 359.7
1c+
1d+
1e+
1e+
1 f+
2emol1
2emol2
3e
3.26
3.34
3.43
3.45
4.89
3.47
3.56
3.54
2.64 0.99 121 56.2 359.8
2.66 1.00 126 68.8 360.0
2.85 0.92 123 75.4 359.7
2.84 0.99 121 75.0 359.7
mol1
mol2
–
1.01
–
–
359.9
2.91 1.15 108 77.0
3.04 1.11 109 75.4
–
–
–
–
–
–
72.8
Scheme 4. Synthetic route to hindered di(9-acridine)s 3a,b and 1a[I].
DMF: N,N-dimethylformamide.
[a] The estimated standard deviation (esd) for D is less than 0.01 ꢀ in all
cases, whereas those for d1 and d2 are much larger (0.02–0.07 ꢀ). [b] Con-
ducted at the B3LYP/6-31G* level.
By contrast, the most-hindered bridged carbocation 1a+
could not be obtained from the methylacridan–acridine pre-
cursor 2a under various conditions, probably because of
steric hindrance. Therefore, the bridged carbocation salt 1b-
1a,c–e+ in the crystal was shown to be the structure with
À
C H localized bridging (observed d1/d2 ratio=2.32–3.10)
+
À À
A
and not the delocalized one with a 3c–2e bond of [C H C]
fonyloxy)dihydrophenanthrene skeleton was converted into
(ideal d1/d2 ratio=1), as detailed below. By contrast, 1 f+
the fully aromatized phenanthrene unit in 1a+ by following
with a highly flexible diphenyl ether spacer does not adopt
À
the same procedure as that applied to 3b to give 3a
the C H bridged geometry in the crystal, and its structure is
more suitable for p–p overlapping between the methylacri-
+
À
(Scheme 4). To our delight, the [C H···C ] unit in carbocat-
ion 1b+ remained intact under the reductive aromatization
conditions (NaI/Zn), and the desired carbocation 1a+ was
finally produced and isolated as a reddish-brown crystalline
salt by Al2O3 chromatography (65% yield of 1a[I] from 1b-
dan and methylacridinium units. As a result, the C9 H part
À
of the methylacridan does not point toward the methylacri-
dinium but rather is directed outward. The solid-state struc-
ture does not always coincide with the most stable geometry
of the species due to the packing force, and this might be
the case with this complex, as shown by its dynamic behav-
ior in solution (see below). In any event, the adoption of the
unbridged structure of 1 f+ in the crystal suggests that the
ACHTUNGTRENNUNG[OTf]). By counterion exchange, 1aACHUTNGTRENN[NGU OTf] could be obtained
from 1a[I]. With a series of stable carbocation salts in hand,
the detailed geometrical features and dynamic properties
were investigated, as described in the following sections.
À
stabilizing effect on the cationic moiety by the C H bridge
Preferred geometry in the solid state (X-ray analyses):
Single crystals of 1a,c–fACTHNUTRGENU[GN OTf] salts suitable for the X-ray
is comparable to that obtained by p–p overlapping
(15 kJmolÀ1 for the benzene dimer[17]).
analyses were successfully obtained by the vapor-diffusion
method. In the case of the 1b+ salt, a specimen of sufficient
quality could not be obtained. Diphenyl ether derivative
1 f+ is labile and was gradually converted into diphenyl
ether-2,2’-diylbis(N-methylacridinium) during the slow crys-
tallization of the salt. The acid-promoted air oxidation of
the acridan unit might be responsible for its instability,
which was suppressed through addition of 1% Et3N upon
Among the bridged carbocations 1a,c–e+, the phenan-
threne-4,5-diyl derivative 1a+ exhibits the closest contacts
of C···C+ (D=3.14 ꢀ) and H···C+ (d1 =2.51 ꢀ) for the [C
À
H···C+] moiety (Figure 2). Due to the phenanthrene skele-
ton, the two chromophores are forced to be located in close
proximity. However, this fused aromatic framework is less
rigid than expected. A skewing deformation of the arylene
spacer is noticeable, with a torsion angle (f) of 29.78 around
2212
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2210 – 2216

Isolable [C···H···C]+ Carbocations
FULL PAPER
As the arylene spacer becomes more skewed, the D and
d1 values gradually increase (3.26 and 2.64 in 1c+; 3.34 and
2.66 in 1d+; 3.43/3.45 and 2.85/2.84 ꢀ in 2 independent mol-
ecules of 1e+, respectively; Table 1 and Figures S1–S4 in the
Supporting Information), which demonstrates that we can
obtain a series of bridged carbocations with a wide variety
of geometrical parameters for the [C···H···C]+ bridging unit.
Although no experimental data are available for 1b+, it is
reasonable to assume that the dihydrophenanthrene deriva-
tive has a geometry that is intermediate between those of
1a+ and 1c+.
X-ray analyses were also conducted on the dibromobi-
phenyl derivatives 2e and 3e (Table 1 and Figures S5 and S6
in the Supporting Information), which are the precursors of
bridged carbocation 1e+. These species adopt a similar
twisted geometry for the arylene spacer (f=72.8–77.0); yet,
among the 3 compounds, the intramolecular separation (D)
between the C9 carbon atoms of the acridine-type chromo-
phore is smallest in 1e+. Thus, the attractive nature of the
+
+
À
short [C H···C ] contact in 1e is stronger than that of the
À
[C H···p]-type (2e) or [p···p]-type (3e) interactions.
Preferred geometry in solution (VT-NMR analyses): The
1H NMR spectrum of acridan–acridinium complex 1a
ACHTUNGTRENNUNG[OTf]
in CD3CN is C2v symmetric at 343 K, with only one N-
methyl resonance at d=3.74 ppm, which can be assigned
+
À À
either to the structure with a 3c–2e [C H C] bond (“delo-
calized form”) or to the unsymmetric bridged structure of
[C H···C ] undergoing rapid H-shift (“localized form”).
+
À
When the temperature was lowered, several signals, includ-
ing N-methyl peaks, became broad and split (Figure 3). At
233 K, sharp resonances appeared again with two distinct N-
methyl resonances at d=4.50 (acridinium) and 2.99 ppm
(acridan), which can be explained by assuming a C1-symmet-
ric structure similar to the solid-state geometry with a
skewed arylene spacer.[19] These results clearly show that the
preferred geometry of 1a+ in solution is also unsymmetric,
with the bridging hydrogen atom localized on one of the ac-
ridine-type chromophores. In contrast to the frozen geome-
try in the solid state, the bridging hydrogen atom fluctuates
between the two chromophores in solution. Based on the
VT-NMR analysis with a Tc (coalescence temperature) of
278 K for the above-mentioned N-methyl protons at
300 MHz, the energy barrier (DG°) for the H-shift at the Tc
value was determined to be 12.4 kcalmolÀ1, with an error of
0.3 kcalmolÀ1 due to the uncertainity in determining the Tc
value.
Figure 2. ORTEP drawing of methylacridan–methylacridinium complex
1aACHTUNGTRENNUNG[OTf]: a) front view; b) top view; c) side view. The counteranion and
hydrogen atoms, except for the bridging one, are omitted for clarity. The
negatively charged oxygen atom of the counteranion is located 3.50 ꢀ
from the C9 atom of the acridinium in the crystal.
Similar temperature dependence was observed in the
NMR spectra of 1b,cACHTNUGTRNEUNG[OTf] in [D6]DMSO, although the coa-
lescence of N-methyl resonances occurs at much higher tem-
peratures (Tc =363 and 423 K, respectively), which corre-
spond to the larger values of DG° ((16.4Æ0.3) and (19.2Æ
0.3) kcalmolÀ1, respectively, at the Tc value). As suggested
by the X-ray analyses, the skewed geometry with a larger D
value in 1b,c+ must be the reason for their higher activation
energies for the H-shift. In fact, the spectrum did not show
À
À
À
the bay region of C4 C4a C4b C5, which reduces the com-
pression effects to a lesser degree than postulated. This is
the reason for the larger D and d1 values in 1a+ than in
naphthalene-1,8-diyl derivative 1A+ (2.96 and 2.18 ꢀ, re-
spectively).[7] DFT calculations[18] on 1a+ showed that the
experimentally determined geometry is well reproduced by
theoretical optimization (Table 1).
complete coalescence even at 423 K for 1d
ACHTUNGTRENNUNG
Chem. Eur. J. 2009, 15, 2210 – 2216
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org

T. Suzuki et al.
Figure 3. VT-NMR analysis of 1aACHTNUGTRENUNG[OTf] in [D3]acetonitrile.
Figure 4. Deuterium scrambling of [D3]1eACHTNUTGRNEUNG
[OTf], as followed by 1H NMR
spectroscopy in [D3]acetonitrile.
skewed derivative. The Tc value seems to be just above that
temperature, which suggests that DG° is approximately
20 kcalmolÀ1 in 1d+.
spectively; these values are much larger than those in 1a–c+,
as supposed. When methylacridan–acridine complex 2d with
a biphenyldiyl spacer was treated with CD3OTf, only the
For 1e,fACHTUNGTRENNUNG[OTf], the fluctuation of the bridging hydrogen
atom is too slow for us to estimate its activation energy by
VT-NMR spectroscopy. Thus, upon treatment of methylacri-
dan–acridine complexes 2e,f with CD3OTf at room tempera-
ture, labeled carbocation salts [D3]1e,fACTHNUTRGNE[NUG OTf] with deuterium
atoms on the methylacridinium moiety were obtained as the
primary products (Scheme 5); these slowly underwent an H-
isomer mixture of [D3]1dACTHNGUTER[NNUG OTf] was obtained, because the
deuterium scrambling of [D3]1d+ is finished before the N-
methylation of 2d is complete. These results are consistent
with the fact that the DG° value of the H-shift in 1d+ (ap-
proximately 20 kcalmolÀ1) is smaller than those in 1e,f+, as
suggested by VT-NMR analyses.
shift to give a 1:1 mixture of two isomeric [D3]1e,f
[OTf]
The chemical shifts of the bridging hydrogen atoms for
1a–e+ are in the range d=4.04–5.29 ppm and are shifted up-
field relative to those of the corresponding neutral precur-
sors 2a–e (Table 2). Although they are very different from
salts with deuterium atoms either on the methylacridinium
moiety or on the methylacridan moiety (Figure 4). From the
kinetics of deuterium scrambling (k=7.76ꢂ10À5 sÀ1 for 1e+
and 4.65ꢂ10À5 sÀ1 for 1 f+ at 318 K in [D3]acetonitrile), the
values of DG° for the H-shift in 1e+ and 1 f+ were deter-
mined to be 24.6Æ0.1 and 25.0Æ0.1 kcalmolÀ1 (318 K), re-
the characteristic values for the bridging hydrogen atoms in
+
À À
aliphatic caged carbocations with 3c–2e [C H C] units
(d=À7 to À3 ppm),[5,20] a less exotic value of d=2.99 ppm
+
À À
was predicted theoretically for the delocalized [C H C]
unit framed in the triarylmethyl-type system.[6b] Thus, the
Table 2. Chemical shift (d) values of the bridging hydrogen atoms in
methylacridan–methylacridinium complexes 1a–f
2a–f.
ACHTUNGTREN[NUNG OTf] and precursors
Compounds 1a–f+ in [D3]MeCN [ppm] 2a–f in CDCl3 [ppm] Dd
1a+/2a
1b+/2b
1c+/2c
1d+/2d
1e+/2e
1 f+/2 f
5.29
4.57
4.25
4.26
4.04
4.86
5.57
4.77
4.58
4.87
4.24
5.22
À0.28
À0.20
À0.33
À0.61
À0.20
À0.36
Scheme 5. Preparation of [D3]1d–fACTHNURGTNE[UNG OTf] for studying the slower H-shift.
2214
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2210 – 2216

Isolable [C···H···C]+ Carbocations
FULL PAPER
+
observed upfield shifts (Dd) in 1a–e+ are indicative of the
higher electron density of the bridging hydrogen atom,
which is consistent with the m-hydrido contribution for the
[C···H···C]+ bridge in the present carbocations. A similar up-
field shift observed in 1 f+ suggests that this carbocation
with a flexible arylene spacer also adopts a bridging geome-
try, at least in part, in solution. This idea is supported by the
fact that the DG° value for the H-shift in 1 f+ is similar to
that in 1e+, which cannot be accounted for by assuming
[C H C] must be formed as isolable species by adopting a
new molecular design concept for the D value to be less
than 2.7 ꢀ.
À À
Conclusion
The present results demonstrate that intramolecular methyl-
acridan–methylacridinium complexes 1a–e+ with a variety
À
that the p–p overlapped structure in the crystal of 1 f
is maintained in solution.
G
of arylene spacers can be best described as C H bridged
À
carbocations with a C H localized geometry that exhibit a
facile 1,6-hydride shift. Modification of the geometry of the
[C···H···C]+ unit (D, d1, d2 and q) drastically changes the dy-
namic behavior of the bridged carbocations. This research
À
Geometry–reactivity relationship: The C H bridged carbo-
cations studied here are interesting not only as unique
À À
À
motifs for studying the organic 3c–2e bonding of [C H C]
shows that studies on isolable C H bridged carbocations
but also as model intermediates for the intramolecular H-
shift in carbocations,[21,22] such as that in the Wagner–Meer-
wein rearrangement. We have successfully obtained the X-
should provide deeper insight into 3c bonds between carbon
and hydrogen atoms, thanks to the finely tuneable geometry,
the details of which can be determined precisely by X-ray
analyses. Studies on species with much smaller C···C+ sepa-
ration (D) are now in progress to design isolable carboca-
À
ray structures of a series of C H bridged carbocations, 1a,c–
e+, which can be considered to be snapshots of 1,6-H-shift
precursors frozen in the crystalline state. Upon dissolution,
the bridging hydrogen atom begins degenerate rearrange-
ment by an H-shift. The energy barriers are in the range of
12.4 to 25.0 kcalmolÀ1 and show a linear correlation with the
separation of C···C+ (D) in the [C···H···C]+ unit, as deter-
mined by X-ray analyses (Figure 5). Although the solid-state
+
[24]
À À
tions with a delocalized 3c–2e bond of [C H C] .
Acknowledgements
This work was supported by the MEXT program called Initiatives for At-
tractive Education in Graduate Schools (“T-type Chemists with Lofty
Ambition”), Department of Chemistry, Graduate School of Science,
Hokkaido University. Financial support from the Global COE Program
by MEXT (“Catalysis as the Basis for Innovation in Materials Science”),
Hokkaido University, is gratefully acknowledged. We thank Prof. Tamot-
su Inabe, Department of Chemistry, Faculty of Science, Hokkaido Uni-
versity, for the use of facilities to analyze the X-ray structures. Elemental
analyses were done at the Center for Instrumental Analysis of Hokkaido
University. Mass spectra were measured by Kenji Watanabe and Dr. Eri
Fukushi at the GC-MS and NMR Laboratory, Faculty of Agriculture,
Hokkaido University.
[1] a) I. R. Epstein, W. N. Lipscomb, Inorg. Chem. 1971, 10, 1921–1928;
b) W. N. Lipscomb, Acc. Chem. Res. 1973, 6, 257–262.
[2] A. J. Arduengo III, S. F. Gamper, M. Tamm, J. C. Calabrese, F. Da-
Figure 5. Linear correlation between the crystallographically determined
C···C+ separation (D) and the DG° value of the H-shift determined in
solution for methylacridan–methylacridinium complexes 1a,c–e
1A–C[OTf].
ACHTUNGTRENN[UNG OTf] and
vidson, H. A. Craig, J. Am. Chem. Soc. 1995, 117, 572–573.
ACHTUNGTRENNUNG
À
À
À
À
À À
À À
[3] Recently, the 3c bonds in the Si H Si, P H P, C F C, and B F B
units have been investigated; see reference [4].
[4] a) R. Panisch, M. Bolte, T. Mꢃller, J. Am. Chem. Soc. 2006, 128,
9676–9682; b) S. A. Reiter, S. D. Nogai, K. Karaghiosoff, H.
Schmidbaur, J. Am. Chem. Soc. 2004, 126, 15833–15843; c) H.
Wang, C. E. Webster, L. M. Pꢄrez, M. B. Hall, F. P. Gabbaꢅ, J. Am.
Chem. Soc. 2004, 126, 8189–8196; d) M. Melaꢅmi, S. Solꢄ, C.-W.
Chiu, H. Wang, F. P. Gabbaꢅ, Inorg. Chem. 2006, 45, 8136–8143.
[5] a) J. E. McMurry, T. Lectka, C. N. Hodge, J. Am. Chem. Soc. 1989,
111, 8867–8872; b) J. E. McMurry, T. Lectka, J. Am. Chem. Soc.
1990, 112, 869–870; c) J. E. McMurry, T. Lectka, Acc. Chem. Res.
1992, 25, 47–53; d) J. E. McMurry, T. Lectka, J. Am. Chem. Soc.
1993, 115, 10167–10173; e) R. P. Kirchen, T. S. Sorensen, J. Am.
Chem. Soc. 1979, 101, 3240–3243; f) T. S. Sorensen, S. M. Whit-
worth, J. Am. Chem. Soc. 1990, 112, 8135–8144; g) C. Taeschler,
T. S. Sorensen, J. Mol. Model. 2000, 6, 217–225; h) C. Taeschler, M.
Parvez, T. S. Sorensen, J. Phys. Org. Chem. 2002, 15, 36–47.
[6] a) R. Ponec, G. Yuzhakov, D. J. Tantillo, J. Org. Chem. 2004, 69,
2992–2996; b) D. J. Tantillo, R. Hoffmann, J. Am. Chem. Soc. 2003,
125, 4042–4043.
geometry does not always represent the structure of the
dominant species in solution, they must be alike for the
present bridged carbocations, 1a,c–e+, based on a consider-
ation of the skeletal rigidity with the congested substruc-
tures.
The regression line, which includes our previous data on
1A–C+, indicates that the DG° value of the H-shift decreas-
es as the C···C+ (D) separation in the [C···H···C]+ unit be-
comes smaller (Figure 5). From an extrapolation, the energy
barrier would disappear when the value of D is less than
2.7 ꢀ. Based on an analogy with very strong intramolecular
[23]
À
O H···O hydrogen bonding (3c–4e), a delocalized single-
well potential would be realized for these cases, which sug-
gests that carbocations with pure organic 3c–2e bonding of
Chem. Eur. J. 2009, 15, 2210 – 2216
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2215

T. Suzuki et al.
[7] a) H. Kawai, T. Takeda, K. Fujiwara, T. Suzuki, J. Am. Chem. Soc.
2005, 127, 12172–12173; b) T. Takeda, H. Kawai, K. Fujiwara, T.
Suzuki, Chem. Eur. J. 2007, 13, 7915–7925; c) H. Kawai, T. Nagasu,
T. Takeda, K. Fujiwara, T. Tsuji, M. Ohkita, J. Nishida, T. Suzuki,
Tetrahedron Lett. 2004, 45, 4553–4558.
[16] K. Oita, H. Gilman, J. Am. Chem. Soc. 1957, 79, 339–342.
[17] C. A. Hunter, J. K. M. Sanders, J. Am. Chem. Soc. 1990, 112, 5525–
5534.
[18] All calculations were carried out with the Spartan ꢆ04 software from
Wavefunction Inc. DFT computation at the B3LYP/6-31G(d) level
was used to calculate the optimized geometry.
[8] a) W. B. Schweizer, G. Procter, M. Kaftory, J. D. Dunitz, Helv. Chim.
Acta 1978, 61, 2783–2808; b) V. Balasubramaniyan, Chem. Rev.
1966, 66, 567–641; c) H. E. Katz, J. Am. Chem. Soc. 1985, 107,
1420–1421; d) H. E. Katz, J. Org. Chem. 1985, 50, 5027–5032;
e) H. A. Staab, T. Saupe, Angew. Chem. 1988, 100, 895–909; Angew.
Chem. Int. Ed. Engl. 1988, 27, 865–879; f) P. C. Bell, J. D. Wallis,
Chem. Commun. 1999, 257–258.
[19] Fluctuation of the bridging hydrogen atom and helical inversion are
independent events for the bridged carbocations 1+; yet, the two
processes in 1aACTHNUGTRNEUNG[OTf] coincidently have a similar energy barrier. This
gives a reason for the observed changes in spectral symmetry be-
tween C1 and C2v upon VT-NMR analysis but not between C1 and
C2 (larger barrier for helical inversion) or between C2 and C2v
(smaller barrier for helical inversion).
+
[9] The very large pKR value for 9-phenylacridinium (11.0) warrants
the stability of this cationic unit: J. W. Bunting, W. G. Meathrel,
Can. J. Chem. 1973, 51, 1965–1972.
[20] D. B. DuPrꢄ, J. Phys. Chem. A 2005, 109, 622–628.
[21] M. Saunders, J. J. Stofko, Jr., J. Am. Chem. Soc. 1973, 95, 252–253.
[22] C. I. F. Watt, Adv. Phys. Org. Chem. 1988, 24, 57–112.
[23] a) P. Gilli, V, Bertolasi, L. Pretto, V. Ferretti, G. Gilli, J. Am.
Chem. Soc. 2004, 126, 3845–3855; b) C. L. Perrin, Science 1994, 266,
1665–1668.
[24] CCDC 699789, 699790, 699791, 699792, 699793, 699794, and 699795
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.
[10] a) T. Saupe, C. Krieger, H. A. Staab, Angew. Chem. 1986, 98, 460–
462; Angew. Chem. Int. Ed. Engl. 1986, 25, 451–453; b) ref. [8e];
c) S. T. Howard, J. Am. Chem. Soc. 2000, 122, 8238–8244.
[11] Preparation of diacridine 3d was reported in a preliminary commu-
nication; see reference [12].
[12] T. Suzuki, K. Ohta, T. Nehira, H. Higuchi, E. Ohta, H. Kawai, K.
Fujiwara, Tetrahedron Lett. 2008, 49, 772–776.
[13] R. Schmid, M. Cereghetti, B. Heiser, P. Schçnholzer, H. J. Hansen,
Helv. Chim. Acta 1988, 71, 897–929.
[14] W. Neugebauer, A. J. Kos, P. von R. Schleyer, J. Organomet. Chem.
1982, 228, 107–118.
[15] A. Rajca, A. Safronov, S. Rajca, C. R. RossII, J. J. Stezowski, J. Am.
Chem. Soc. 1996, 118, 7272–7279; F. R. Leroux, R. Simon, N. Nicod,
Lett. Org. Chem. 2006, 3, 948–954.
Received: August 28, 2008
Revised: November 16, 2008
Published online: January 13, 2009
2216
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 2210 – 2216